Analysis of CNBr fragments and other peptides from human liver cytoplasmic aldehyde dehydrogenase enabled determination of the complete primary structure of this protein. The monomer has an acylated amino terminus and is composed of 500 amino acid residues, including I1 cysteine residues. N o evidence of any microheterogeneity was obtained, supporting the concept that the enzyme is a homotetramer. The disulfiram-sensitive thiol in the protein, earlier identified through its reaction with iodoacetamide, is contributed by a cysteine residue at position 302, while the cysteine which in horse liver mitochondrial aldehyde dehydrogenase is reactive with coenzyme analogs appears to correspond to either Cys-455 or Cys-463. Analysis of glycine distribution and prediction of secondary structures to localize PEP regions typical for coenzyme-binding are not fully unambiguous, but suggest a short region around position 245 as a likely segment for this function. In this region, sequence similarities to parts of a bacterial aspartate-P-semialdehyde dehydrogenase and a mammalian alcohol dehydrogenase were noted. Otherwise, no extensive similarities were detected in comparisons with characterized mammalian enzymes of similar activity or subunit size as aldehyde dehydrogenase (glyceraldehyde-3-phosphate dehydrogenase and glutamate dehydrogenase, respectively).Aldehyde dehydrogenases have a wide substrate specificity and participate in ethanol metabolism through clearance of acetaldehyde. In human liver, two major isoenzymes of aldehyde dehydrogenase are present and have been ascribed to the cytoplasm and mitochondria, as in the livers of other mammalian species [ 1 -51. Structurally, the isoenzymes behave like different proteins in the sense that hybrid molecules are not (but see [6]) usually found [7]. Furthermore, a comparison of short segments from horse cytoplasmic and mitochondrial aldehyde dehydrogenases indicated a high extent of substitutions (about 50%) between the two forms [8]. Immunologically, however, they appear to be homologous [5,9] and the two forms are also similar in many enzymic properties. One notable exception is the much greater sensitivity of the cytoplasmic isoenzyme to inactivation by disulfiram [I, 5, 10, 111. This compound is also used clinically for alcohol aversion therapy, acting through the effects of acetaldehyde accumulation [12].Only recently have detailed molecular aspects of aldehyde dehydrogenase been studied. In the human cytoplasmic isoenzyme, a cysteine-containing segment has been implicated in the reaction with disulfiram because reaction of one of its cysteine residues with iodoacetamide is blocked by disulfiram [I 31. Another segment from horse mitochondrial aldehyde dehydrogenase, also containing a reactive cysteine residue, has been implicated in NAD(H) binding from its reaction with a coenzyme analog [14]. Although not finally proven, the possi- bility of a position close to the active site for one of these cysteine residues appears likely [I 31; this would be compatible with the ...

The recently determined primary structure of glucose dehydrogenase from ~a~i~~as megater~~m was scanned by computerized comparisons for similarities with known polyol and alcohol dehydrogenases. The results revealed a highly significant similarity between this glucose dehydrogenase and ribitol dehydrogenase from Klebsiella aerogenes. Sixty-one positions of the 262 in glucose dehydrogenase are identical between these two proteins (23% identity), fitting into a homology alignment for the complete polypeptide chains. The extent of similarity is equivalent to that between other highly divergent but clearly related dehydrogenases (two zinc-containing alcohol dehydrogenases, 25%; sorbitol and zinc-containing alcohol dehydrogenases, 25%; ribitol and non-zinc-containing alcohol dehydrogenases, 20%), and suggests an ancestral relationship between glucose and ribitol dehydrogenases from different bactera. The similarities fit into a previously suggested evolutionary scheme comprjsing short and long aIcohol and polyol dehydrogenases, and greatly extend the former group to one composed of non-zinc-containing alcoholpolyol-glucose dehydrogenases.

Three different size classes of cDNA clones coding for the &subunit of human alcohol dehydrogenase (ADH) were characterized from a human liver cDNA library. Clones were identified by hybridization with synthetic oligodeoxyribonucleotides.A total of 2530 nucleotides were determined, covering an ADH-coding region of 1122 nucleotides, a preceding 72-nucleotide segment and 3 types of 3'-non-coding region. The coding nucleotide sequence is in full agreement with the amino acid sequence of the &subunit. Of 8 clones identified, 6 had a short, 213-nucleotide 3'-non-coding region; 1 an intermediate, 590-nucleotide 3'-region; and 1 a long, 1330-nucleotide 3'-region. In addition, 2 unused polyadenylation signals were found. These results suggest that human liver /I-ADH mRNAs occur in several size classes, and that in addition to the consensus sequence AATAAA further signals are important for 3'-end formation.Alcohol dehydrogenase cDNA

A comparison of the structure of class II human liver alcohol dehydrogenase (alcohol:NAD+ oxidoreductase, EC 1.1.1.1) (containing nT subunits) with those of the human class I isozymes (containing a, (3, and y subunits) reveals differences at about 40% of all positions. Variations are large for active-site regions, the segment around the second zinc atom, and for segments involved in subunit interactions. The two classes of alcohol dehydrogenase have diverged to exhibit structural differences to about half the extent of those between alcohol and polyol dehydrogenases. Hence, the two classes of alcohol dehydrogenase represent steps in enzyme rather than isozyme divergence. An evolutionary scheme that relates different types of zinc-containing mammalian dehydrogenases to one another encompasses at least three levels of gene duplication subsequent to the early step(s) of assembly of building unit(s). The first level of duplication results in the formation of now clearly different enzymes. The second level concerns the various classes of alcohol dehydrogenase, forming steps between typical enzymes and isozymes. The third level encompasses recent and multiple duplications in isozyme evolution of alcohol dehydrogenases. This scheme, linking zinc-containing dehydrogenases at different levels, resembles that in other protein families and reflects general patterns in protein relationships.Isozyme differences of mammalian alcohol dehydrogenases (alcohol:NAD' oxidoreductase, EC 1.1.1.1) were initially defined structurally for the horse liver enzyme (1). The enzyme from this source is still the only alcohol dehydrogenase for which the tertiary structure has been analyzed directly by x-ray crystallography (ref. 2; recent review in ref.3). However, progress in the knowledge of alcohol dehydrogenase in general has been rapid. The yeast, plant, and mammalian enzymes are related to the horse enzyme (4-16), all depending on functional zinc at the active site. A completely different type of alcohol dehydrogenase in insects has been differentiated in a scheme that separates alcohol dehydrogenases into two types that differ in zinc content, size, and enzymatic mechanisms (17). Furthermore, sorbitol dehydrogenases are known to be related to the group of large, zinc-containing alcohol dehydrogenases (17,18), while bacterial ribitol and glucose dehydrogenases are related to the line of small alcohol dehydrogenases that lack zinc (17,19).Knowledge about the human alcohol dehydrogenase has increased particularly rapidly. This enzyme constitutes a complex system composed ofthree classes, I, II, III (20). The class I enzymes constitute the "typical" alcohol dehydrogenases composed of three types of subunit, a, 13, and -y, and originating from three loci (21) located on chromosome 4 (22).All of these class I structures have been characterized (summaries in refs. 14 and 23) at both the protein and cDNA levels. Genomic structures (24) and their organizational relationships to other alcohol dehydrogenases (25) have also been studied. Fu...

No abstract

The amino acid sequence of a recently isolated camel milk protein rich in half-cystine has been determined by peptide analyses. The 11 7-residue protein has 16 half-cystine residues, concluded to correspond to disulfide bridges and suggesting a tight conformation of the molecule.Comparisons of the structure with those of other proteins reveal several interesting relationships. The camel protein is clearly homologous with a previously reported rat whey phosphoprotein of possible importance for mammary gland growth regulation, and with a mouse protein of probable relationship to neurophysins. The camel, rat and mouse proteins may represent species variants from a rapidly evolving gene. Residue identities in pairwise comparisons are 40% for the camel/rat proteins and 33% for the camel/mouse proteins, with 38 positions conserved in all three forms. The camel protein also reveals an internal repeat pattern similar to that for the other two proteins.The homology between the three milk whey proteins has wide implications for further relationships. Thus, previously noticed similarities, involving either of the milk proteins, include limited similarities to casein phosphorylation sites for the camel protein, to neurophysins in repeat and half-cystine patterns for the mouse and rat proteins, and to an antiprotease for the rat protein. These similarities are reinforced by the camel protein structure and the recognition of the three whey proteins as related. Finally a few superficial similarities with the insulin family of peptides and with some other peptides of biological importance are noticed. Combined, the results relate the camel protein in a family of whey proteins, and extend suggestions of relationships with some binding proteins.

No abstract

No abstract