Jun 29, 1984 - Human alcohol dehydrogenase: Structural differences between the. ,8 and y subunits suggest parallel duplications in isoenzyme evolution and ...
Proc. Natd. Acad. Sci. USA Vol. 81, pp. 6320-6324, October 1984 Biochemistry
Human alcohol dehydrogenase: Structural differences between the ,8 and y subunits suggest parallel duplications in isoenzyme evolution and predominant expression of separate gene descendants in livers of different mammals (isoenmzyme gene/amino acid sequence/structure-funetion relationship)
ROLF BUHLER*t, JOHN HEMPELt, RUDOLF KAISERt, JEAN-PIERRE AND HANS JORNVALLt
VON
WARTBURG*, BERTIL. VALLEEt,
*Medizinisch-Chemisches Institut der Universitat, Buhistrasse 28, CH-3000 Bern 9, Switzerland; tDepartment of Chemistry I, Karolinska Institutet, S-104 01 Stockholm, Sweden; and tCenter for Biochemical and Biophysical Sciences and Medicine, Harvard Medical School, 250 Longwood Avenue, Boston, MA 02115
Contributed by Bert L. Vallee, June 29, 1984
ABSTRACT libman alcohol dehydrogenase (ADH; alcohol:NAD+ oxidoreductase, EC 1.1.1.1) occurs in multiple forms, which exhibit distinct electrophoretic mobilities and enzymatic properties. The homogeneous isoenzymes I3P, and y1yi were isolated from livers of Caucasians with "tyrpical" ADH phenotype by double ternary complex affinity chromatography and ion exchange chromatography. The differences between the PI and Vi subunits were determined by structural analysis of all tryptic peptides from the carboxymethylated proteins. The human f3I and Vi chains differ at 21 of the 373 positions (5.6%). Ten tryptic peptides account for the differences. All residue substitutions are compatible with one-base mutations and result in largely unaltered properties, but five lead to charge differences. Sixteen substitutions are at positions corresponding to the catalytic domain of the well-known horse enzyme; five correspond to the coepzyme-binding domain. Substitutions adjacent to important regions may correlate with differences in coenzyme binding, substrate specificities, and active-site relationships. The residue replacements between the PI and yV subunits of human ADH are not identical to the known substitutions between ethanol-active (E) and steroid-active (S) subunits of horse ADH. Thus, the duplication leading to human pi and yi subunits is separate and different from that leading to equine E and S subunits. Both duplications are likely to have occurred after the ancestral separation of human and equine ADH. Of the 21 residues that are different between P1I/ yi, 13 in y'but only 6 in flu are identical to those of the horse E chain. This suggests a closer relationship between yV and E, although .81 in mah and E in the horse are the subunits recovered in highest yield from liver ADH preparations. Consequently, in these two mammalian species, relative activities of genes for an isoenzyme family appear to be
liver in the order ( > y > a. Allelic forms at both the ADH2 and ADH3 loci code for fB (typical) and (2 (atypical; see ref. 8) chains, and for yV and 'y chains, respectively. Atypical 02subunits have been differently reported from livers of Caucasians and Asians (12-lern and f2-Oriental) (9, 10), but both of these 2 types are now known to be structurally identical (11, 12), differing from typical (1 subunits by a single His/Arg exchange at position 47. Another type of ADH, ADHIndianapolis, is probably derived from a further variant of the ADH2 locus (13). Finally, two additional gene loci are likely to code for the more different ir- and X-isoenzymes (14-16). Horse liver ADH has been studied most extensively, but much less is known of the human enzyme, in part because of its complicated isoenzyme pattern that initially rendered the purification of homogeneous forms of its isoenzymes difficult. Introduction of double ternary complex affinity chromatography (17) has greatly simplified this problem. Thus, the cathodic, pyrazole-sensitive human alcohol dehydrogenase forms can then be isolated free of contaminating proteins in an isoenzyme mixture composed of products from the ADH!, ADH2, and ADH3 loci (18). Subsequent CM-cellulose ion-exchange chromatography separates the isoenzyme mixture into its individual forms. It now is possible to obtain homogeneous human ADH-isoenzymes of known subunit composition in quantities sufficient for further characterization (16, 20, 21). The enzymatic properties of the various isoenzymes differ. The well-known horse liver isoenzymes EE (for the dimer of ethanol-active subunits) and SS (for steroid-active subunits) exhibit differences in substrate specificity (1) that do not correspond to those among the rat or human ADH isoenzymes (22). However, the human isoenzymes also have characteristic physical and functional properties that allow for differentiation. Thus, ,81(31 is less stable in vitro than ylyi. These isoenzymes also differ in relation to chloral hydrate reduction, n-pentanol oxidation, and inhibition by isobutyramide (6). The activity toward benzyl alcohol is particularly striking; ythy oxidizes benzyl alcohol rapidly, while (31i(1 is almost inactive with this substrate (16, 20). Furthermore, vyly also oxidizes ethanol much better than 831if3 (16, 19, 20). Characterization of the structural differences may provide correlations with the functional properties. Also, the relationships among the various ADH isoenzymes in different species could illustrate isoenzyme developments, genetic organization, and differential gene activations. Thus, recent studies of the (2 atypical human ADH variant and compari-
different. Human and horse alcohol dehydrogenases (ADH; alcohol:
NAD' oxidoreductase, EC 1.1.1.1) are structurally closely related and both exhibit multiple isoenzymes, which can be differentiated by their electrophoretic mobilities (1). Two polypeptide chains of the horse liver enzyme, E and S (ethanol-active and steroid-active, respectively), combine randomly to form the homodimeric isoenzymes, EE and SS (with EE apparently the most abundant form), as well as the heterodimeric form, ES (2, 3). The human liver ADH isoenzyme pattern is more complex. At least three gene loci, ADH!, ADH2, and ADH3, code for the polypeptide chains a, (i, and y, respectively (4-7), with apparent amounts in the The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviation: ADH, alcohol dehydrogenase.
6320
Biochemistry:
BOhler et at
Proc. Natl. Acad. Sci. USA 81 (1984)
son with the corresponding yeast ADH counterparts have provided structural explanations for the functional differences and have also illustrated the presence of parallel mutational exchanges (11). We now describe differences in the amino acid sequence of human ADH 3 and y polypeptides coded by the ADH2 and ADH3 loci of Caucasian livers with normal phenotype. These differences are then compared with those known for other ADH isoenzymes in horse (23), yeast (24), and man (25). The results show that human PI and yi ADH chains differ in another way than the horse E and S ADH chains. Evolutionarily, the human y' chain is closely related to the horse E chain, whereas the human pi and horse S chains appear to have evolved differently. Thus, subunits PI in man and E in horse are structurally nonequivalent, suggesting that the predominant polypeptide of liver ADH corresponds to different gene products in these species.
MATERIALS AND METHODS Purification of Human Liver ADH Isoenzymes. Caucasian human livers were obtained at autopsy, frozen 10-20 hr after death, and stored at -20'C. The mixtures of the pyrazolesensitive class I ADH isoenzymes (26) were isolated by means of a minor modification of the affinity-chromatography method described (18). Crude ADH derived from livers was loaded on a CapGapp-Sepharose column (3 x 35 cm) in 50 mM sodium phosphate, pH 8/0.37 mM NAD. The column was washed with this buffer, after which the ADH was eluted with 50 mM sodium phosphate, pH 7.5/500 mM ethanol. Homogeneous isoenzymes were isolated by subsequent ionexchange chromatography on CM-cellulose CM-52 in 50 mM
Tris HCl (pH 9). The isoenzymes were eluted from the gel with a linear gradient of 0-50 mM NaCl in 50 mM Tris HCl
(pH 9) (20).
Purity of the isoenzymes was tested by NaDodSO4 slab gel electrophoresis in 12% polyacrylamide (27) and by starch gel electrophoresis at pH 7.7 (4). Isoenzymes were identified by their mobilities on starch gel electrophoresis (4). Their subunit composition was verified by monomerization and hybridization with the horse liver ADH isoenzyme EE, followed by starch gel electrophoresis (20). Structure Determination. After CM-cellulose chromatography, pools containing the homodimeric isoenzymes /31,31 and ViYi were dialyzed extensively against distilled water and then lyophilized. The material was dissolved in 6 M guanidinium hydrochloride, reduced, and carboxymethylated with iodo[2-14C]acetate (25). The labeled proteins were then fragmented with trypsin in 0.1 M ammonium bicarbonate (protein concentration, 3 mg/ml; trypsin:protein, 1:100; 370C for 4 hr). The fragments were separated into groups by exclusion chromatography on Sephadex G-50 in 0.1 M ammonium bicarbonate and purified further by reversed-phase high-performance liquid chromatography (28). The isolated fragments were checked for purity by analysis of total compositions and end groups. Manual sequence degradations were performed by the DABITC method (29) using by-products to assist the identifications (30). Some peptides were degraded by the dansyl-Edman method, using polyamide thin-layer chromatography for identification (31, 32). Liquidphase sequencer degradations were carried out in a Beckman 890C sequencer, modified with a cold trap, new valves, and a conversion cell (33), using glyrine-precycled Polybrene and a 0.1 M Quadrol peptide program (34). B
A
0.15S
-
6321
1
2
3
4
5
6
7
8
7500
9
0.115
2
2
3
5
4
6
7
8
9
75)00
a) _>
E c xc
K-ent _
00 .
E
O.110
5000 ';'
0.10
2
C0._
._
2500
200
W.
0.10-
Ala33-Arg37
0.75
50 m
0.15p[ Ec x:
j
300
400
D
75
0.500.25
200
,
-
i=
25
75
0.75
Ala33-Arg37
E
0.10 *a 0.50
50
CIA
0.05
25
0.25S
C I!L\_ X 0
10
-oo
Volume, ml
.l
0.05
251
)5
0
E c x
0.C
400
300 Volume, ml
C 0.15
c)
50
20 0.05F
E
30
Volume, ml
50
10
v 30
50
Volume, ml
FIG. 1. Purification of tryptic peptides from human ADH subunits f,3 and yl. (A and B) Sephadex G-50 chromatography of tryptic peptides of PI (A) and y, (B). Horizontal bars at the top denote the fractions pooled and further purified by high-performance liquid chromatography. -, A28onm; ----, radioactivity. (C and D) Purification by high-performance liquid chromatography of the peptides in pool 7 of A and B from ,1 (C) and y} (D). Elution was by a gradient (%B) of acetonitrile in 0.1% trifluoroacetic acid. The peptide showing the largest difference in the patterns is indicated (Ala-33 to Arg-37, containing Tyr-34 in PI but His-34 in yi), as well as peptides corresponding to adjacent peaks (Phe-229 to Lys231, unchanged between ,81/yl; and Phe-130 to Arg-137 present only in f31, because of exchange of Arg-137 to Ser-137 in yl). The last two small peaks in B correspond to a low extent of P,3 contamination in the yV preparation. ----, A28nm; -, A2l4nm; -*-, %B.
Biochemistry:, Buhfler
6322
et
aL
NatL.
Proc.
Acad. Sci. USA 81 (1984)
Table 1. Amino acid sequences of tryptic peptides containing all the amino acid exchanges between I8i and Vi Structure Isoenzyme Position AAVLWEVKKPFSIEDVEVAPPKAVEVR Pi1 AAVLWELKKPFSIEEVEVAPPKAHEVR Yi
fPi
MVAVGICRTDDHVVSGNLVTPLPVILGHEAAGIVESVGEGVTTVK MVAAGICRSDEHVVSGNLVTPLPVILGHEAAGIVESVGEGVTTVK
Yi
A3
1013CRVCKNPESNYCLK 1013CRICKNPESNYCLK
Yi
A3
FITCRGKPIHHFLGTSTFSQYTVVDENAVAK
Vi
1019FTCSGKPIHHFVGVSTFSQYTVVDENAVAK
13i
VCLIGCGFSTGYGSAVNVAKVTPGSTCAVFGLGGVGLSAVMGCK
Yi
1922VCLIGCGFSTGYGSAVKVAKVTPGSTCAVFGLGGVGLSVVMGCK
/3i
LDTMMASLLCCHEACGTSVIVGVPPASQNLSINPMLLLTGR 2232LDTMVASLLCCHEACGTSVIVGVPPDSQNLSINPMLLLTGR
Vi
813
GAVYGGFKSKEGIPK 3630GAIFGGFKSKESVPK
Yi
f31
FSLDALITHVLPFEKINEGFDLLHSGK 339-366
Vi Peptides
are
denoted
FSLDALITNILPFEKINEGFDLLRSGK
according to their piositions in the protein chain. Deviating residues are represented by standard one-letter abbreviations.
are
printed
in boldface. Amino acids
RIESULTS
performance liquid chromatography to produce all the tryppeptides. Fig. 1 C and D shows the elution profiles of pool differentiate /13 and yi already at the pre7, which fractionation stage. In pool 7, the large difference in the absorbance (Fig. 1 A and B) and in the pattern from' high-performance liquid chromatography (Fig. 1 C and D) corresponds to a Tyr/His exchange at position 34 (see below). Table 1 shows the primary structures of all peptides that tic
14C-carboxymethylated Pi and yi prodigested with trypsin, -and the pejnides obtained were separated into groups by chromatography on and yi dige~ts Sephadex G-50. The elution profiles of the are similar (Fig. 1 A and B), suggesting that the positiorns of the basic amino acid residues along the two' protein' chains are closely related. However, several pools differ between the and profiles already at this level of resolution. Thus, 14C label or absorbance at 280 nm of pools 3 and 7 in particdlar differ in 1ifrom the patterns in yi (Fig. 1). On this basis, the isoenzymes would seem to be largely similar but still derived from protein chains with several amino acid replacePeptide Maps.
tein chains
'Visually
The
were
813
differ between
and
yi.
tween
yi
and
Structural
Analyses. After chromatography on Sephadex G-50, all pools were purified further by reversed-phase highAmino acid 11
Isozyme
composition of peptides
18
19
Si
Y1
32
33
Yi
40
Si
Y1
47
also shows the most
100
Y1
positions,
as
the E and S subunits of the horse
coviering positions at which subunits P, and
37
21 differences in 10 pep-
consistent with one-base mutations. Ta-
are
ble 3 summarizes these
Segment
are
pairs, and Table 2 shows the amino acid compositions supporting the sequence data. Analysis of remaining peptides gives the complete 813 and structures (35, 36). Amino Acid Substitutions. All 21 residue replacements be-
ments.
Table 2.
There
tide
169
113
el
Y
61
188
316
-yA
the six in which
as
residue at each
differ (23). It
position.
At 13
differ 323
Si
Yl
common
well
iso'en'zymes
326
330
81
Yl
339
355
354
81
y1
366
SI
Y1
Yl
Composition
Cys
(CM)
0.9
11.1
Asx
(1)
1.0
(1)
(1)
2.6
(3)
2.9
(3)
2.1
(2)
2.2
(2)
Thr
Ser
1.1
(1)
1.0
(1)
Pro
1.0
(1)
0.9
(1)
L1.9
(2)
3.0
(3)1
2.9
(3)
3.0
(3)
1.0
(1)
0.9
(1)
Gly Ala
2.0
(2)
2.1
(2)
1.0
(1)
1.0
(1)
1.1
(1)
1.0
(1)
Val
1.9
(2)
1.1
(1)l
2.0
(2)
1.9
(2)
1.1
(1)
1.0
(1)
Met
Ilie Leu
0.9
(1)
1.1
(1)
b1(
1.1
(1,)
(1)
2.2
(2)
0.9
(1)
0.6
(1)
1.0
(1)
0.9
(1)I
11.0 (1 ) 2.72(2)] 11.1
Tyr Phe
Trp
0.7
(1)
0.9
(1)
Lys
0.9
(1)
1.1
(1)
1.0
(1)
1.0
(1)
1.8
(2)
2.0
(2)
Arg
0.8 8
10
14
(1)
-
1.1
(1)
1.0
5
(2)
2.1
(2)
I
-
1.1
(1)
1.0
(1)
1.7
(2)
1.7
(2)
5
(1)
2.0
(2)l
1.1
(1)
0.9
(1)
(1) 8
1.1
(1) 8
(2)
1.9
(2)
(1)
1.0
(1)
-
0.9
(1)I
1
1
(1)
1.0
(1)
1.1
(1)
0.9
(1)
(1)
1.0
(1)
1.0
(1)
1.1
(1)
1.0
(1)
1.1
(1)
1.2
(1)
1.0
(1)
1.0
(1)
1.2
(1)
1.0
(1)
1.1
(1)
1.2
(1)
1.0
(1)
L1.1
(1)
2.2
(2)
2.2
(2)
1.1
(1)
1.0
(1)
~
3.9
(4)
4.0
(4)
2.9 (3)
3.2
(3)
2.0
(2)
2.1
(2)
1.0
(1)
1.0
(1)
3)
3.1
(3)
.-
1.0
(71)
1.1
(1)
0.9
(1)
1.0
(1)
0.9
(1)
1.0
(1)
1.0
(1)
1.0
(1)
1.0
(1)
(2)
2.0
(2)
1.0
(1)
1.0
(1)
1.1
(1)
1.1
(1)
11.1
(1)
1.9
(21)
I-
-
[1.0 (1) L1.0 (1) 1.0
(1)
1.0
(1)l
1.9
(2)1
1.0
(1)
-I
1.[11 (1i)
-_
0.9
1.0 (1)
0.9
(1)
0.9
(1)
1.0
(1)
2.9
(3)
2.7
(3)
2.0
(2)
1.9
(2)
2.0
(2)
2.0
(2)
1.0
(1)
1.0
(1)
(2)
1.9
(2)
1.8
10.9 1.0
2.1
1.0
(1)l (1)
11.1
1.1
2.3
His
Sum
1.1
11.0
1.7
11i2 (1)
0.9
(1) 14
0.8
(1)
20
20
8
8
5
5
16
16
1.1 (1)
(1)-
(1) 14
1.0 (1)
0.9
-
12
1.0
-I
(1)l 12
For segment 100-113, the values given are the sum of two peptides (100-104 plus 105-113 for 81; iOO-101 plus 102-113 for Vi) because of differential tryptic cleavage. Similarly, for segments 169-188 and 355-366, the Vi values are the sums 'of two peptides because of substitutions in Vi at positions 185 and 363 to Lys and Arg, respectively. Differences are boldface and 'enclosed in boxes. Only peptides with 20 or fewer amino acids are listed and cover 13 of the P/3i/V substitutions.
Biochemistry: Bfihler et aL
Proc. NatL Acad. Sci. USA 81 (1984)
Table 3. Positions with known isoenzyme differences in ADH from either man or horse Man
Position 17 25 34 43 48
50 94 101 102 110 115 133 141 143 185 207 276 297 318 319 327 328 348 349 363 366 E differences
Horse
/31
lY1
Val Asp Tyr Val Thr Asp Thr Arg Val Tyr Asp Arg Leu Thr Asn Ala Met Ala Val Tyr
Leu Glu His Ala Ser
Gly lie
Glu Thr Arg Ile Tyr Asp Ser Val Val Lys Val Val Asp Ile Phe Ser Val Asn
E Glu Glu His Thr Ser Asp Thr Arg Val Phe Asp Ser Leu Thr Lys Val Val Asp Ile Phe Ser Val His Val Arg
His Val His Lys
Ile Arg Lys
Glu
13
6
2
S Gln
ND ND
lie Ser Val Leu (Ser) ND ND ND ND
Most
abundant alternative None Glu His None Ser Asp Thr Arg Val Tyr Asp Ser Leu Thr Lys Val Val Asp
lie
ND ND
Lys
Phe Ser Val His Val Arg Lys
4
L16 identities1 9 identities-
P1I and yi denote the two human protein chains presently analyzed; E and S are the two equine protein chains (23). Empty positions in the S column denote those positions not reported for the SS isoenzyme but are derived from peptides with unaltered properties in a fingerprinting system (23) and, therefore, probably represent no or minor E/S differences. ND denotes positions in larger peptides that were not analyzed. Positional numbers refer to the primary structure of the horse E chain and correspond to the numbers in the human A3 chains except for a one-residue shift because of a gap at position 128 in man versus horse (35). Differences at positions 17-115 and 319-366 are in the catalytic domain and those at 185-318 are in the coenzyme-binding domain (1). Residues in boldface type indicate the larger similarities between y and E than between any other isoenzyme pairs.
of the 21 positions in which yi differs from ,1, the human 'Y alternative is identical to that in the horse E subunit of ADH. However, residues at only 6 of the 21 positions are identical between 813 and E (at the remaining two positions, P1, yi, and E all differ). Thus, where 81 and y' differ, the yj chain is the one most closely related to E, whereas the deviations in 61 do not relate significantly to either E or S.
DISCUSSION Isoenzyme Differences. The primary structures of human ADH subunits ,81 and differ at 21 of 373 positions (5.6%). All of these differences are compatible with one-base mutations, and for most of them, the resultant properties are largely unchanged (6 Val/Leu/Ile exchanges, 2 Asp/Glu, and 1 His/Arg). However, some differences in charge result, His-34 in yi is substituted by Tyr in f1, Ser-133 by Arg, Lys185 by Asn, Asp-297 by Ala, and Asn-348 by His. Among the two human isoenzymes, the alternative found in yi is usually
6323
the one that is identical to the horse enzyme, concerning both conservative and radical substitutions. The data also reveal that an isoenzyme mixture studied previously (25), without full knowledge of the isoenzyme complexity, was a mixture of y and 81 chains, because the amino acid exchange then detected was Val/Ala at position 43, in agreement with present results (Table 3). The data further establish that the two human ADH isoenzyme subunits ,31 and Yi are closely related, corresponding to a comparatively recent ADH2/ADH3 gene duplication, followed by accumulation of a limited number of point mutations. Functional Correlations. Most of the amino acid differences between the P1 and yV chains occur in the substratebinding domain, as judged from comparisons with the tertiary structure of the horse E chain (37) (16 of 21 differences; cf. legend to Table 3), and only 5 occur in the coenzymebinding domain. One of the latter presumably alters the coenzyme-binding properties, explaining the higher activity of yi in relation to f81 (16, 20), because NADH dissociation is the rate-limiting step (1). One substitution in the coenzymebinding domain results in a charge difference, Lys-185 in yv instead of Asn-185 in P1. The charge lost in 81 is not at a position interacting with the coenzyme but corresponds to one near the end of the first a-helix (aA) in the coenzymebinding domain of the horse enzyme (37). The substitution of Ile-318 in y, (and in horse E) by Val-318 in 81 has a position lining the active-site pocket (37), suggesting that it could possibly influence substrate relationships of the isoenzymes. One important residue that is probably involved in a proton relay system is Ser-48 of the horse enzyme (37, 38). Significantly, this residue is conserved in the yi chain but is substituted by Thr-48 in P31. This mutation still allows an unaltered proton relay system [in analogy with a similar difference in the yeast enzyme (39)], but it could affect space relationships close to the catalytic zinc atom, possibly explaining the changes in substrate specificity (cf. ref. 40). The difference at position 48 may therefore be associated with the low enzyme activity of P1i1 against large substrates such as benzyl alcohol (16, 20). Evolution of Isoenzymes and Activation of Genes. The residue substitutions between human ,/1 and yi subunits are different from those between the horse E and S subunits. Consequently, the duplication leading to the f/y chains appears to be separate and different from that which led to the E/S chains. The two sets of isoenzymes must, therefore, have arisen subsequent to the ancestral separation of the human Ancestor 48
Horse 6 E
Man 21
S
Yi
13,
FIG. 2. Separate but parallel gene duplications in the develop-
ment of mammalian ADH isoenzymes. Length of arrows is a measure of the known number of deviations of each isoenzyme chain from the most common alternative (cf. Table 3). Thick arrows denote the subunits that are recovered in highest yield from liver. As
shown, they represent different lines. Numbers give degree of positional differences between characterized present-day forms, representing for man/horse the /8/E differences [48 substitutions (35) after complete analysis of both forms], for horse the E/S differences [6 substitutions hitherto known (23) after complete analysis of E and incomplete analysis of S], and for man the yl/,Bl differences (21 substitutions now reported after complete analysis of both forms).
6324
Biochemistry: Bfihler et aL
and equine lines. This example shows that isoenzymes in mammalian systems can develop by different and independent, but parallel, gene duplications. Considering the residues exchanged, not just the positions, it is obvious that the human y' subunit and the horse E subunit are the ones most closely related, suggesting that they have the most direct relationship to the ancestor. Therefore, the 81 and S chains appear to be the ones that are evolutionarily "new." The relationships are summarized in Fig. 2. Finally, it is of particular interest that the liver ADH forms that predominate in horse and man are not equivalent (although relative subunit abundance is difficult to estimate). Thus, the E chain of horse ADH is both abundant in liver and closely related to the ancestral form. In man, however, P1 is recovered in larger amounts than yi from liver ADH preparations, but it is not the subunit most closely related to the ancestral form. Thus, in these two mammalian species, the relative activities of genes for one isoenzyme family appear to differ. This work was supported by grants from the Swedish Medical Research Council (Project 13X-3532), the Swiss National Science Foundation (Grant 3.355-0.82), and the Knut and Alice Wallenberg Foundation. R.B. was a recipient of EMBO short-term fellowships and J.H. of a fellowship from the Endowment for Research in Human Biology (Boston), and the Samuel Bronfman Foundation, with funds provided by Joseph E. Seagram and Sons. 1. Branden, C.-I., Jornvall, H., Eklund, H. & Furugren, B. (1975) in The Enzymes, ed. Boyer, P. D. (Academic, New York), 3rd Ed., Vol. 11, pp. 103-190. 2. Pietruszko, R. & Theorell, H. (1969) Arch. Biochem. Biophys. 131, 288-298. 3. Lutstorf, U. M. & von Wartburg, J.-P. (1969) FEBS Lett. 5, 202-206. 4. Smith, M., Hopkinson, D. A. & Harris, H. (1971) Ann. Hum. Genet. 34, 251-271. 5. Smith, M., Hopkinson, D. A. & Harris, H. (1972) Ann. Hum. Genet. 35, 243-253. 6. Smith, M., Hopkinson, D. A. & Harris, H. (1973) Ann. Hum. Genet. 36, 401-414. 7. Smith, M., Hopkinson, D. A. & Harris, H. (1973) Ann. Hum. Genet. 37, 49-67. 8. von Wartburg, J.-P., Papenberg, J. & Aebi, H. (1965) Can. J. Biochem. 43, 889-898. 9. Berger, D., Berger, M. & von Wartburg, J.-P. (1974) Eur. J. Biochem. 50, 215-225. 10. Yoshida, A., Impraim, C. C. & Huang, I.-Y. (1981) J. Biol. Chem. 256, 12430-12436. 11. Jornvall, H., Hempel, J., Vallee, B. L., Bosron, W. F. & Li, T.-K. (1984) Proc. Natl. Acad. Sci. USA 81, 3024-3028. 12. Buhler, R., Hempel, J., von Wartburg, J.-P. & Jornvall, H. (1984) FEBS Lett., in press.
Proc. Natl. Acad. Sci. USA 81 (1984) 13. Bosron, W. F., Li, T.-K. & Vallee, B. L. (1980) Proc. Natl. Acad. Sci. USA 77, 5784-5788. 14. Bosron, W. F., Li, T.-K., Dafeldecker, W. P. & Vallee, B. L. (1979) Biochemistry 18, 1101-1105. 15. Pares, X. & Vallee, B. L. (1981) Biochem. Biophys. Res. Cornmun. 98, 122-130. 16. Wagner, F. W., Burger, A. R. & Vallee, B. L. (1983) Biochemistry 22, 1857-1863. 17. Lange, L. G. & Vallee, B. L. (1976) Biochemistry 15, 46814686. 18. Lange, L. G., Sytkowski, A. J. & Vallee, B. L. (1976) Biochemistry 15, 4687-4693. 19. Buhler, R. & von Wartburg, J.-P. (1982) FEBS Lett. 144, 135139. 20. Bosron, W. F., Magnes, L. J. & Li, T.-K. (1983) Biochemistry 22, 1852-1857. 21. Jornvall, H. (1980) in Advances in Experimental Medicine and Biology: Alcohol and Aldehyde Metabolizing Systems, ed. Thurman, R. G. (Plenum, New York) Vol. 4, pp. 67-76. 22. Burger, A. R. & Vallee, B. L. (1981) Fed. Proc. Fed. Am. Soc. Exp. Biol. 40, p. 1886 (abstr.). 23. J6rnvall, H. (1970) Eur. J. Biochem. 16, 41-49. 24. Wills, C. & Jornvall, H. (1979) Eur. J. Biochem. 99, 323-331. 25. Jomvall, H. & Pietruszko, R. (1972) Eur. J. Biochem. 25, 283290. 26. Strydom, D. J. & Vallee, B. L. (1982) Anal. Biochem. 123, 422-429. 27. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 28. von Bahr-Lindstrom, H., Carlquist, M., Mutt, V. & Jornvall, H. (1982) in Methods in Protein Sequence Analysis, ed. Elzinga, M. (Humana Press, Clifton, NJ), pp. 455-462. 29. Chang, J. Y., Brauer, D. & Wittmann-Liebold, B. (1979) FEBS Lett. 93, 205-214. 30. von Bahr-Lindstrom, H., Hempel, J. & Jcrnvall, H. (1982) J. Protein Chem. 1, 257-262. 31. Woods, K. R. & Wang, K.-T. (1967) Biochim. Biophys. Acta 133, 369-370. 32. Jornvall, H. (1970) Eur. J. Biochem. 14, 521-534. 33. Wittmann-Liebold, B. (1981) in Chemical Synthesis and Sequencing of Peptides and Proteins, eds. Liu, T., Schechter, A., Heinrikson, R. & Condliffe, P. (Elsevier/North-Holland, Amsterdam), pp. 75-110. 34. Jornvall, H. & Philipson, L. (1980) Eur. J. Biochem. 104, 237247. 35. Hempel, J., Buhler, R., Kaiser, R., Holmquist, B., de Zalenski, C., von Wartburg, J.-P., Vallee, B. L. & Jornvall, H. (1984) Eur. J. Biochem., in press. 36. Buhler, R., Hempel, J., Kaiser, R., de Zalenski, C., von Wartburg, J.-P. & Jornvall, H. (1984) Eur. J. Biochem., in press. 37. Eklund, H., Nordstrom, B., Zeppezauer, E., Soderlund, G., Ohlsson, I., Boiwe, T., Soderberg, B.-O., Tapia, O., Branden, C.-I. & Akeson, A. (1976) J. Mol. Biol. 102, 27-59. 38. Eklund, H., Plapp, B. V., Samama, J.-P. & Branden, C.-I. (1982) J. Biol. Chem. 257, 14349-14358. 39. Jornvall, H. (1977) Eur. J. Biochem. 72, 425-442. 40. Eklund, H., Branden, C.-I. & Jornvall, H. (1976) J. Mol. Biol. 102, 61-73.
