Identification and characterization of novel mammalian alcohol dehydrogenases

The vertebrate medium chain alcohol dehydrogenases are dimeric zinc metalloenzymes that catalyze the reversible oxidation/reduction of alcohols/aldehydes using NAD+/NADH as coenzyme. In mammals, six classes of ADH have been defined (ADH1-ADH6) whereof five have been identified in humans. These enzymes are further divided into isozymes, and the presence of allelozymes, which have been detected for a number of gene loci, adds additional multiplicity to this enzyme system. Only a few ADHs have been assigned specific metabolic functions, e.g. glutathione dependent formaldehyde oxidation by ADH3, whereas other ADHs seem to contribute in the general defense against xenobiotics and endogenously formed alcohols and aldehydes. This thesis aims to provide insights into ADH structure and function, with emphasis on the identification and characterization of novel ADH forms. A cDNA coding for a mouse ADH2 was cloned, which showed high structural similarity with the rat ortholog but lower identity with other species variants. Interestingly, the mouse and rat ADH2 forms share the same residue replacements, as compared to other ADHs, at positions important for catalysis and substrate specificity. Kinetic measurements displayed that rodent ADH2s are low activity enzymes, but the activity could be restored by substituting a unique Pro47 for Ms. Large substrate isotope effects for octanol oxidation showed that hydride transfer is rate-limiting for turnover. Altogether, these results indicate that the rodent enzymes form an ADH2 subgroup within the ADH family. The mechanisms of mammalian ADH2 enzymes were studied further and an ordered mechanism with coenzyme binding as the first reactant was suggested for both human and mouse ADH2. The enzymes display strong pH dependence with pK values for kcat and kcat/Km several pH units higher than fot the ADH1 enzymes. Additional replacements at position 47 in the mouse enzyme (Pro47Ala and Pro47Gln) also increased oxidative activity and we propose that the rigid ring structure of Pro47 causes coenzyme binding that is unfavorable for efficient catalysis. This also shows that a His at position 47 does not act as a catalytic base in the deprotonation of the alcohol substrate. Experiments with deuterated substrates and transient kinetic measurements indicate that dissociation of coenzyme is rate-limiting for human ADH2 whereas hydride transfer is rate-limiting for all mouse ADH2 forms. Polymorphisms in both the ADH2 and the ADH3 genes were found by single strand conformational polymorphism (SSCP) analysis. The ADH2 polymorphism affects the coding region and results in an Ile308Val substitution. Homology modeling located position 308 in the subunit interface of the molecule and in the vicinity of the active site pocket entrance. Characterization of the two allelozymes showed that the 308Val substitution decreases protein stability as compared to the 308Ile variant and that Km-values, for a number of model substrates, were higher for the 308Val enzyme. The ADH3 polymorphisms were detected within the 5'-flanking region. A reporter gene assay showed a significant reduction in promoter activity for a rare C+9 g T+9 transition, which could arise from decreased binding of nuclear proteins, as indicated by electrophoretic mobility shift assays. ADH5 and ADH6 have been identified only at the nucleic acid level in human and deermouse, respectively. These enzymes share 67% structural identity and could possibly belong to the same diverged class. Previous reports indicated that human ADH5 differs from other ADHs in exon/intron organization, lacking the last exon and thus should be expressed in a truncated form. However, we were able to identify full-length ADH5 transcripts and the presence of a ninth exon was also confirmed by sequencing of genomic DNA. Northern blot analyses established the full-length variant as the major transcript with the strongest signal from adult liver. We conclude that the ADH5 gene contains a composite internal/terminal exon, which can be differentially processed as a result of competition between polyadenylation and splicing. In addition, a new ADH was identified in rat which showed 78% identity with the deermouse ADH6 variant. No soluble ADH5 or ADH6 protein could be recombinantly expressed in E. coli, but using a protocol of refolding insoluble proteins from inclusion bodies, milligram amounts of ADH-GST fusion proteins of both human ADH5 and rat ADH6 could be isolated. The purified proteins were not stable without the GST tag and no activity was observed with a number substrates, which could be a result of an incorrect fold of the ADH part of the fusion protein. In vitro translation experiments and expression of ADH-GFP fusion protein in COS cells indicate that soluble ADH5 and ADH6 protein could be produced in mammalian cells