The alcohol dehydrogenases of Saccharomyces cerevisiae: a comprehensive review

Abstract

Alcohol dehydrogenases (ADHs) constitute a large family of enzymes responsible for the reversible oxidation of alcohols to aldehydes with the concomitant reduction of NAD⁺ or NADP⁺. These enzymes have been identified not only in yeasts, but also in several other eukaryotes and even prokaryotes. The ADHs of Saccharomyces cerevisiae have been studied intensively for over half a century. With the ever-evolving techniques available for scientific analysis and since the completion of the Yeast Genome Project, a vast amount of new information has been generated during the past 10 years. This review attempts to provide a brief summary of the wealth of knowledge gained from earlier studies as well as more recent work. Relevant aspects regarding the primary and secondary structure, kinetic characteristics, function and molecular regulation of the ADHs in S. cerevisiae are discussed in detail. A brief outlook also contemplates possible future research opportunities.

Saccharomyces cerevisiae, alcohol dehydrogenase, kinetics, localization, molecular regulation, function

Introduction

Saccharomyces cerevisiae is, without a doubt, the most important microorganism commercially exploited by humans. No other microorganism has been more intimately associated with the progress and wellbeing of the human race than S. cerevisiae and its closely related species. Its contribution to human progress has been due largely to its capacity for the ethanolic fermentation of carbohydrate feedstocks. Two major pathways are involved in the energy metabolism of S. cerevisiae, namely glycolysis and aerobic respiration. Ethanol is a key metabolite in energy metabolism, being an end product of glycolysis and ethanolic fermentation while also serving as a carbon substrate during aerobic respiration, with the alcohol dehydrogenases (ADHs) catalysing the interconversion of acetaldehyde and ethanol.

The ADH systems of various organisms have been investigated thoroughly in respect to their molecular structure, mode of catalysis and, especially in S. cerevisiae, physiological significance. During the last century, biochemical analysis of yeast ADH has largely been promoted by the easy availability of yeast cells. Physiologically, the ADH reaction in S. cerevisiae and in related species plays a dual and quite critical role in sugar metabolism. Almost all of the carbohydrate is used fermentatively, regardless of the availability of oxygen, and a specific ADH isozyme serves to regenerate the glycolytic NAD⁺, thereby restoring the redox balance, through the reduction of acetaldehyde to ethanol. Under aerobic conditions, respiration of the accumulated ethanol occurs after depletion of the fermentable sugar, again by the action of specific ADH isozymes. Thus, in S. cerevisiae, the ADH reaction links fermentative and respiratory (oxidative) carbon metabolism, allowing the optimal use of the sugar carbon.

Before the sequencing of the S. cerevisiae genome, completed in April 1996, only five ADHs were known. The genome sequence, however, revealed the existence of c. 6000 ORFs, including a number of ADHs or sequences possibly related to these enzymes. Since then, ADH6 and ADH7 have also been identified and their enzyme products have been characterized. Considering all the new data available on the ADHs and the wealth of knowledge gathered over the past century, a need exists for a comprehensive and extensive overview regarding the primary and secondary structure, mechanisms of function and regulation of these enzymes. This review constitutes a compilation of the information gathered since the 1960s pertaining to the molecular biology and physiological role of the S. cerevisiae ADHs identified to date.

Classification of the ADHs

ADHs (E.C. 1.1.1.1) are oxidoreductases that catalyse the reversible oxidation of alcohols to aldehydes or ketones, with the concomitant reduction of NAD⁺ or NADP⁺. ADHs constitute a large group of enzymes that can be subdivided into at least three distinct enzyme superfamilies: medium-chain (MDR) and short-chain dehydrogenases/reductases and iron-activated ADHs (Reid & Fewson et al., 1994; Kallberg et al., 2002). The MDR superfamily is divided into different enzyme families with regard to structural and functional relationships (Jornvall et al., 2001). The family consists of enzymes with a subunit size of c. 350 amino acid residues, dimeric or tetrameric, with two domains in each subunit: one catalytic and one responsible for binding with the nucleotide NAD⁺ or NADP⁺. Many enzymes of the MDR family have zinc in their active site and have a sequence motif known as the zinc-containing ADH signature: GHEX₂GX₅(G,A)X₂(I,V,A,C,S) (Persson et al., 1993).

The genes encoding classical ADHs include ADH1, ADH2, ADH3, ADH4 and ADH5 (Lutstorf & Megnet et al., 1968; Ciriacy, 1975a; Walton et al., 1986; Feldmann et al., 1994). Other genes include SFA1, BDH1 (Gonzalez et al., 2000), BDH2 (YAL061W), SOR1 putative SOR2 (YAL246C), XDH1, the cinammyl ADH CDH1 and the recently proposed ADH6 (YMR218C) and ADH7 (YCR105W) (Wehner et al., 1993; van Iersel et al., 1997; Gonzalez et al., 2000; Larroy, 2002a, b).

Localization

Eukaryotic cells are organized into a complex network of membranes and compartments, which are specialized for various biological functions. A comprehensive knowledge of the location of proteins within these cellular microenvironments is critical for understanding their functions and interactions. In the late 1960s, the presence of at least three NAD-dependent ADHs in baker's yeast was known (Heick et al., 1969). Two of these enzymes were located in the cytosol (Adh1p and Adh2p) and fractionation experiments placed Adh3p in the mitochondrial matrix (van Loon & Young et al., 1986).

Drewke & Ciriacy (1988) suggested that it was unlikely that Adh4p is proteolytically processed as has been demonstrated for Adh3p (Pilgrim & Young et al., 1987) and that Adh4p was, therefore, a cytoplasmic enzyme. To date, no other reports to support this have been published. Furthermore, while studying the mechanisms of mitochondrial oxidation of cytosolic NADH, Overkamp (2000) and Bakker (2000) obtained evidence that ADH3 was not the sole mitochondrial ADH. At that time, Adh4p and Ydl114W were the only probable candidates with no assigned function or localization. Huh (2003) described the construction and analysis of a collection of yeast strains expressing full-length, chromosomally tagged green fluorescent fusion proteins (GFP). The use of the GFP tag and co-localization with red fluorescent protein (RFP)-tagged reference proteins allowed them to resolve many related subcellular compartments with confidence, especially in the case of proteins for which little functional data existed. They established that Adh5p was located in the cytoplasm and that not only Adh3p but Adh4p was also a mitochondrial enzyme. If this were the case, an explanation is yet to be provided for the absence of the amino terminal leader peptide sequence in Adh4p.

Currently, the data as to the localization of Adh6p and Adh7p are not conclusive, as most information available on the characterization of these two enzymes was derived from overexpression studies (Larroy, 2002a, b; Valencia et al., 2003, 2004) or resulted from monitoring expression at the mRNA level (Petersson et al., 2006).

Primary and secondary structure

Most of the work on yeast ADH genetics is due to the masterful analysis of Michael Ciriacy (Ciriacy et al., 1997). Genetic analysis performed on several haploid adh mutants led to the identification of four unlinked genes, namely ADH1, ADH2, ADH3 and ADH4. Identification of ADH5 followed later (Feldmann et al., 1994) and ADH6 and ADH7 with gene exploration opportunities as a result of the S. cerevisiae genome project (Gonzalez et al., 2000, Larroy, 2002a, b).

ADH1 (YOL086C) encoding Adh1p

Yeast ADH was the first pyridine nucleotide-dependent dehydrogenase to be crystallized. Determination of the primary structure of Adh1p in S. cerevisiae marked the first case of gene cloning by functional complementation. Physico-chemical methods revealed that the protein had a molecular weight of 150 kDa, that the active enzyme contained four identical reactive sites and that in all probability it consisted of four similar, if not identical, polypeptide chains (Harris et al., 1964). Yeast Adh1p seems to be a good model for the study of the stability of complex enzymes, because effects of the environment on the structure can easily be tested by measuring the inactivation of reduced and oxidized Adh1p. Ca²⁺ stabilizes S. cerevisiae Adh1p by preventing the dissociation of the reduced form of the enzyme and by preventing the unfolding of the oxidized form of the enzyme. It is also possible that Adh1p binds Mg²⁺in vivo (de Bolle et al., 1997).

ADH2 (YMR303C) encoding Adh2p

By the early 1980s, the structural genes encoding both Adh1p (ADH1) and Adh2p (ADH2) were identified genetically (Ciriacy, 1975a), cloned (Williamson et al., 1980, 1981) and their DNA sequences were determined (Bennetzen & Hall et al., 1982; Russell, 1983b). High homology was evident at both the nucleotide sequence level (90%) and in their amino acid sequences (95%). The amino acid sequences of Adh1p and Adh2p had only 22 differences out of 347 residues, with no differences in the groups directly involved in catalysis (Ganzhorn et al., 1987).

ADH3 (YMR083W) encoding Adh3p

The presence of Adh3p in respiratory-deficient mutants proved that nuclear DNA encodes the synthesis of this mitochondrially bound enzyme (Wiesenfeld et al., 1975). Young & Pilgrim (1985) isolated and sequenced the ADH3 gene, and the nucleotide sequence indicated a 73% and 74% identity with ADH1 and ADH2, respectively. Wiesenfeld (1975) assigned Adh3p its tetrameric structure and the amino acid similarity of the predicted Adh3p polypeptide to Adh1p and Adh2p was 79% and 80%, respectively. All the active site, cofactor-binding and noncatalytic zinc-binding residues identified in S. cerevisiae Adh1p by Jornvall (1977) are conserved in both Adh2p and Adh3p.

The ORF-encoding Adh3p has a highly basic 27 amino acid amino terminal extension relative to Adh1p and Adh2p. Gene fusion showed that the amino terminus of Adh3p contained the information for targeting the protein to, and transporting it into the mitochondrion (van Loon & Young et al., 1986). The amino acid sequence and secondary structure(s) of the leader sequence, as well as the adjacent sequences in the mature amino terminus of Adh3p, all contribute information for normal mitochondrial binding, importing and processing (Mooney et al., 1990).

ADH4 (YGL256W) encoding Adh4p

ADH4 is the most distal marker on the left arm of chromosome VII and both restriction and genetic analysis of the chromosome copy of ADH4 indicated that it was situated near a telomere (Walton et al., 1986). Paquin & Williamson (1986) described ADH4 as a new ADH isozyme, or that it regulated the expression of an ADH. Sequencing analysis of ADH4 revealed a long ORF that was not homologous to other yeast ADHs and only distantly related to other characterized eukaryotic ADHs (Paquin & Williamson et al., 1986). Analysis of the sequence of the hypothetical Adh4p protein showed that it did not contain structurally or functionally important amino acid residues that were conserved between yeast Adh1p and horse liver ADH. The hypothetical ADH4 gene product did, however, show a strong homology to the iron-activated ADH from the bacterium Zymomonas mobilis. This homology suggested that ADH4 did encode an ADH, but one distinct from any other that has been described in eukaryotes (Williamson & Paquin et al., 1987).

The presence of a TATAA sequence upstream of the transcription start and the moderate codon bias suggested that it may be a functional yeast gene. Its location near the end of chromosome VII is interesting in view of the fact that many of the genes located at the ends of chromosomes in S. cerevisiae encode enzymes involved in fermentation and glycolysis (Mortimer & Schild et al., 1985). Unlike Adh1p, Adh2p and Adh3p, which are thought to function as tetramers (Leskovac et al., 2002), Adh4p is a dimeric protein that normally occurs in low concentrations in laboratory strains (Drewke & Ciriacy et al., 1988). Contrary to the suggestion of Williamson & Paquin (1987), Adh4p is activated by zinc ions, like the other yeast ADH isozymes, and not by ferrous ions as is the case with the structurally similar ADH from Z. mobilis (Drewke & Ciriacy et al., 1988).

ADH5 (YBR145W) encoding Adh5p

ADH5 was first identified through sequencing of chromosome II (Feldmann et al., 1994) and shares a 76%, 77%, and 70% sequence identity with ADH1, ADH2 and ADH3, respectively (Feldmann et al., 1994; Ladriere et al., 2000). No information is currently available regarding characterization of the putative ADH5 gene product.

ADH6 (YMR218C) encoding Adh6p

The ADH6 (Gonzalez et al., 2000) gene product was the first NADPH-dependent medium-chain ADH to be characterized in S. cerevisiae (Larroy, 2002b). It is also the first cinammyl ADH and member of the MDR superfamily whose three-dimensional crystal structure has been determined (Valencia et al., 2004). The heterodimeric enzyme Adh6p consists of two 40 kDa subunits: one in the apo conformation and the second in the holo conformation (Valencia et al., 2004). Adh6p exhibits conservation of the zinc-signature, as well as amino acid sequences in the substrate and coenzyme-binding domains characteristic of the zinc-containing MDR enzymes (Gonzalez et al., 2000; Larroy, 2002b). Adh6p showed only a 26% sequence identity to the tetrameric enzyme family (Valencia et al., 2003) and exhibited a strict specificity for NADPH (Larroy, 2002b).

ADH7 (YCR105W) encoding Adh7p

Adh7p has a 64% sequence identity with Adh6p and the purified ADH7 gene product is a homodimer whose reductase activity is about fivefold that of the dehydrogenase activity (Larroy, 2002a, b). The phylogenetic tree constructed from the MDRs identified in the genomes of S. cerevisiae, Escherichia coli, Drosophila melanogaster and Caenorhabditis elegans placed Adh7p in a family of enzymes structurally related to cinnamyl ADHs (Larroy, 2002a, b).

Tertiary and quaternary structure

Active centres

Relationships known from tertiary structures of dehydrogenases show that the constituent monomers are separated into two domains, namely a ‘coenzyme-binding’ domain and a ‘catalytic’ domain. The three-dimensional structure of the active site of the S. cerevisiae enzyme revealed the presence of a hydrogen-bonded proton relay system (Leskovac et al., 1998). On the basis that each individual chain contains one reactive sulphydryl group and, by binding one atom of zinc and 1 mol of NAD⁺/NADP⁺, it is capable of forming an independent ‘active centre’ within the quaternary structure of the active tetramer (Harris et al., 1964). Adh1p has a methionine residue at position 294 (numbered as in the horse liver enzyme), whereas isozymes Adh2p and Adh3p have leucine. Apart from these differences, the active sites of the S. cerevisiae enzymes are the same (Ganzhorn et al., 1987). It has been proposed that the shape or the accessibility of the catalytic pocket appears to be different in the yeast and horse liver enzymes and that it is possible to alter the specificity of the enzyme without sacrificing catalytic power (Green et al., 1993). Such approaches are limited by the lack of data on the tertiary and quaternary structure of tetrameric S. cerevisiae ADH. Crystallization of Adh1p has been reported repeatedly, but the crystals are seemingly not very useful in X-ray diffraction studies (Ciriacy et al., 1997). The size and shape of the Adh6p active site appears to be adapted to the bulky and hydrophobic substrates of cinammyl ADHs. The crystal structure of this enzyme showed that its specificity towards NADP(H) is achieved mainly by tripod-like interactions of the cofactor terminal phosphate group with certain side chains (Valencia et al., 2004).

Coenzymes and metal-binding sites

All MDR enzymes utilize NAD(H) or NADP(H) as a cofactor and have one zinc ion with a catalytic function at the active site (Branden et al., 1975). Early studies showed that the active enzyme binds 4 mol of NAD⁺ (Karlovic et al., 1976) and four atoms of zinc. Karlovic (1976) also demonstrated that binding of the coenzymes was linear over a wide temperature range, both at the level of binary and ternary complexes, and thermodynamic parameters showed no close similarity between heat and entropy changes associated with NAD and NADH binding. Zinc atoms are essential for maintaining the quaternary structure of the enzyme and both zinc and the coenzyme are bound at, or near to, each of the four reactive cysteines (Harris et al., 1964). Zinc is one of the principal trace elements in biology, with structural or enzymatic roles in hundreds of proteins (Vallee & Falchuk et al., 1993). Saccharomyces cerevisiae Adh1p, Adh2p and Adh3p contain one catalytic zinc atom and a second zinc atom, which plays a prominent conformational role, probably through stabilization of the tertiary structure. The second zinc is located at the periphery of the molecule and the external localization of this structural zinc affects local conformations of the enzyme (Magonet et al., 1992). The different coenzyme-binding domains have extensive similarities, are composed of two ‘mononucleotide-binding units’ and resemble the tertiary structures of kinases and some other proteins (Rossman et al., 1975)

The important role of the zinc atom in alcohol oxidation is to stabilize the alcoholate ion for the hydride transfer step in the reverse direction. Zinc functions as an electron attractor, which gives rise to an increased electrophilic character of the aldehyde, consequently facilitating the transfer of a hydride ion to the aldehyde. Thus, the proposed mechanism is essentially electrophilic catalysis mediated by the active site zinc atom (Leskovac et al., 2002).

Kinetic characteristics

The first biochemical data on Adh1p and Adh2p showed that the kinetic properties of both enzymes favoured alcohol production. Under the conditions of a high ethanol concentration and the efficient removal of acetaldehyde, both enzymes could function in the oxidation of ethanol (Heick et al., 1969). Wills (1976) later stated that Adh1p was normally constitutive under laboratory conditions, had a high K_m value for ethanol (17 000–20 000 μmol L⁻¹) (Thomson et al., 2005) and, therefore, seemed chiefly responsible for the production of ethanol during anaerobic growth. If levels of intracellular ethanol were low, Adh2p would produce acetaldehyde and NADH at a faster rate than Adh1p (Wills et al., 1982). Kinetic investigation of commercially available ADH showed it to be capable of oxidizing all primary alcohols with chain lengths of between two and 10 carbon atoms (Schopp & Aurich et al., 1976), and the activity of Adh1p decreased with increasing chain length of the primary alcohols (Ganzhorn et al., 1987). The substrate specificity of Adh1p is restricted to primary unbranched aliphatic alcohols and any branching decreased the activity and efficiency of the enzyme (Leskovac et al., 2002). Cyclic alcohols (benzyl alcohol, cyclohexanol) were not oxidized in detectable amounts (Drewke & Ciriacy et al., 1988) and thiol compounds exerted no effect on this isozyme (Cheng & Lek et al., 1992). It was also reported that overexpressed Adh1p reduced formaldehyde (FA) to methanol in vivo (Grey et al., 1996) and was able to provide a considerable degree of protection against cadmium (Yu et al., 1991).

Adh2p has a low K_m for ethanol (600–800 μmol L⁻¹) and is found only in aerobically grown yeast cells (Wills et al., 1976; Thomson et al., 2005). These findings concur with its role as a major ethanol oxidizer. For all alcohols, normalized reaction rates with Adh2p were about threefold faster than with Adh1p (Leskovac et al., 2002). Some contradiction is found in the literature regarding the kinetic characteristics of Adh1p and Adh2p. Some reports state remarkably similar normalized reaction activities for both Adh1p and Adh2p under conditions of low substrate concentration, even though the kinetic characteristics of the enzymes are very different (Dickinson & Dack et al., 2001).

The mitochondrial enzyme Adh3p (Bakker et al., 2000) showed great affinity for alcohols with double bonds conjugated to the alcohol function (Wiesenfeld et al., 1975). The methionine (Adh1p) or leucine (Adh2p and Adh3p) at position 294 by itself had no interaction with ethanol or propanol (Ganzhorn et al., 1987).

Adh4p has different substrate specificity and pH profiles compared with other ADH isozymes (Drewke & Ciriacy et al., 1988). Although Adh2p and Adh4p differ remarkably in almost all the kinetic parameters (Lutstorf & Megnet et al., 1968), the latter resembles Adh1p in its kinetic constants. Adh4p is seemingly more specific, however, because only ethanol and n-propanol are oxidized, whereas isomers of aliphatic alcohols, secondary alcohols and glycerol are not used (Drewke & Ciriacy et al., 1988). No data on the kinetic characteristics of Adh5p are presently available.

Adh6p accepts a wide range of compounds as substrates, including linear and branched-chain primary alcohols and aldehydes, substituted cinnamyl alcohols and aldehydes as well as substituted benzaldehydes and their corresponding alcohols. It is able to produce 2,3-butanediol from acetoin during fermentation (Gonzalez et al., 2000; Larroy, 2002a, b).

In general, the substrate specificity of Adh7p is quite similar to that of Adh6p (Larroy, 2002a). It showed the same activity towards linear and branched-chain alcohols, but much higher catalytic efficiencies towards the oxidation of cinnamyl alcohols and aliphatic alcohols (Larroy, 2002b).

Regulation

Regulation of ADH2 mediated by Adr1p

The earliest study of regulation of the ADH genes was documented in the mid-1970s and provided the first defining proof of a controlling site involved in carbon catabolite repression in a eukaryote (Ciriacy, 1975b, 1976). The first gene associated with this function was ADR1, a positive regulatory gene specifically activating the expression of the structural gene ADH2 under derepressed conditions (Ciriacy et al., 1979). ADR1 encodes the trans acting protein (Adr1p) containing two zinc fingers and an adjacent region on the amino-terminus side, which together are essential for DNA binding (Blumberg et al., 1987). Initiation of transcription from most known eukaryotic promoters is positively regulated by the binding of specific transcriptional activator proteins to the enhancer region of the upstream activation sequence (UAS) of a promoter (Hope & Struhl et al., 1985).

Two unusual features upstream of the ADH2 promoter, a 22-bp perfect dyad sequence and a (dA)₂₀ tract, were identified (Russell, 1983a). The ADH2 promoter may normally be in an inactive conformation in the yeast chromosome and derepression requires positive activation by Adr1p that is mediated through the 22-bp perfect dyad (UAS1) (Beier et al., 1985). The Adr1p monomers are able to form one of two complexes: complex I corresponds to the binding of one molecule to the cis acting element UAS1 and complex II corresponds to the binding of two molecules to UAS1 (Thukral et al., 1991). A GC rich, 20 bp sequence (UAS2) upstream of UAS1 was identified and proposed to act synergistically as a binding site for a protein that interacts with Adr1p to activate the expression of ADH2 (Yu et al., 1989). In vitro binding data to this cis-acting element suggest that Adr1p binds with a low affinity to UAS2, most likely at the AGGAGA sequence. Other interpretations, such as an indirect effect or a direct protein–protein interaction, are also possible but seem less likely (Donoviel et al., 1995).

Synthesis of Adr1p is 10–16-fold greater during growth on ethanol than during growth on glucose. This derepression of ADR1 protein translation was found to occur within 40–60 min of glucose depletion. Glucose, therefore, represses ADH2 expression by considerably decreasing the rate of Adr1p synthesis. Cook & Denis (1993) established that a 510 bp untranslated leader sequence of the ADR1 mRNA played a role in the increased rate of ADR1 mRNA degradation under conditions of growth on glucose as compared with growth on ethanol. Other positive factors influencing ADH2 expression were also proposed, because excess Adr1p could not overcome a three- to fourfold inhibition in ADH2 transcription caused by multiple promoters on a multicopy vector (Irani et al., 1987).

Other elements participating in ADH2 regulation

Genetic and biochemical analysis showed that expression of the Adh2p structural gene was under the control of at least 24 other unlinked genetic elements or proteins, most of which influence ADH2 expression mainly in a direct fashion (Table 1). Several of the genes appear likely to do so through control of Adr1p, whether by mRNA translation, phosphorylation or protein interaction. In contrast to the situation with most genes that are subject to catabolite repression, there is no evidence for an MIG1-binding site in the ADH2 promoter (Denis & Audino et al., 1991).

1

Elements other than Adr1 involved in the regulation of the ADH2 structural gene

Element	Postulated mode of action	References
ADR3	Cis-acting regulatory locus required for glucose regulation of the ADH2 gene	Beier & Young (1982)
CCR1	Encodes a constitutive protein that acts, probably posttranscriptional, in concert with or through Adr1p	Denis & Gallo (1986)
ADR4	Regarded as a negative regulatory factor because of its complete recessiveness. It is not yet established whether ADR4 acts specifically on the ADR1 or ADH2 gene or whether its gene product constitutes a regulatory element with a pleiotropic action spectrum	Ciriacy (1979), Denis (1981)
ADR6	Contains an ORF encoding a protein that has a potential metal-binding finger domain. It appears to require an ADH2 sequence located downstream to or including the TATAA box and may act subsequent to Adr1p, but before translation of ADH2 mRNA	Taguchi & Young (1987), O'Hara (1988)
CRE1 & 2	Affect transcription independently of the upstream regulatory sequences. They are not specific to the regulation of carbon metabolism and function as general effectors of transcriptional processes	Denis (1984), Denis & Malvar (1990)
CCR4	Also affects transcription independently of the upstream regulatory sequences. Ccr4p contains two distinct glutamine and leucine rich regions that presumably play an important role in protein interactions that mediate the regulatory role of Adh2p	Malvar (1992)
NOT1-4	Encode a group of factors involved in repressing the transcription of HIS3 from a noncanonical TATA. MS and two hybrid analyses identified Not1-Not5 as proteins associated with the Ccr4p complex, implying that it should also be positively involved in gene transcription	Collart & Struhl (1994)
SRB9-11	Encode proteins that coimmunoprecipitate the Ccr4p and Not proteins. Defects in these Srb proteins affect expression at the ADH2 locus in a manner similar to that observed for defects in Ccr4p complex components	Liu (2001)
cAPK	Although cAPK inactivates ADH2 expression by inhibiting Adr1p function through phosphorylation, either directly or of another protein required for Adr1p activity, it does not appear responsible for the glucose to ethanol transition in controlling ADH2 expression	Denis (1992)
Sch9	Appears to activate ADH2 expression by turning off Adr1p function independently of cAPK. SCH9 derepression does not operate exclusively through Adr1p. Because Sch9p does not act through UAS1, it may control factors that act through other activation sequences or through factors controlling the general transcriptional machinery	Denis & Audino (1991)
ADR7-9	Are trans-acting elements unlinked to Adr1p that affect ADH2 expression under both repressing and derepressing growth conditions. The gene products allow ADH2 expression to efficiently derepress in the absence of Adr1p as do cre1 and cre2	Karnitz (1992)
SAF1-3	Required for maintaining high levels of ADR1 RNA when growing on glucose, but have little effect during growth on ethanol. Whereas the physiological role of the SAF genes is unclear, they appear to be expressed with both ethanol and glucose as carbon sources and no direct evidence exists that they influence genes other than ADR1 and ADH2	Cook & Denis (1993)
REG1	Encodes a nuclearly localized protein whose expression is not regulated by glucose availability. REG1 is the first negative genetic element identified that affects ADH2 expression through an ADR1 dependent pathway via an additional step in transcriptional activation that does not involve the phosphorylation–dephosphorylation of Adr1p	Niederacher & Entian (1991)
BCY1	Encodes the regulatory subunit of cAPK and influences ADH2 expression in an Adr1p-dependent manner	Donoviel (1995)
Cat8	Has an essential role during the adaptation of yeast on ethanol by controlling the induction of many genes in response to glucose depletion. It contributes to transcriptional derepression of ADH2 synergistically with the Adr1p dependent UAS1 element	Haurie (2001), Walther & Schuller (2001), Tachibana (2005)
MoT1	Strains defective in this gene allow SNF1 dependent constitutive activation of ADH2 expression by Adr1	Voronkova (2006)

Influence of chromatin remodelling on ADH2 expression

Yeasts have to respond very rapidly to environmental changes. Therefore, their chromatin is maintained in a state of ‘transcriptional readiness’, even when the cells are transcriptionally inactive. The ADH system serves as an ideal model to detect localized differences in chromatin structure, which can reflect changes in transcriptional activity. The possibility exists that a decrease in transcriptional activity changes the structure of chromatin in the whole ADH2 region, for instance, by a change of nucleosome spacing (Sledziewski & Young et al., 1982). Under repressing conditions, UAS1 and UAS2 are a nucleosome-free region, the TATA box and RNA initiation sequence (RIS) is protected by nucleosomes −1 and +1 and a 20 bp poly (dA−dT) tract is included in the DNA wrapped around nucleosome −1. Nucleosome mapping data show that glucose exerts its inhibitory effect by keeping the relevant promoter sequences (TATA box and RIS) in a nucleosomal configuration, thus precluding their engagement with the transcription machinery (Verdone et al., 1997).

Chromatin remodelling that occurs at the S. cerevisiae ADH2 promoter upon derepression is characterized by two distinct structural alterations. The existence of two steps in the process of chromatin remodelling suggests that at least two functions can be attributed to Adr1p. First, the protein reconfigures nucleosomes in the immediate vicinity of its binding site, allowing the basal promoter elements to assume the most appropriate structure for the subsequent activation. Second, the protein recruits the transcription machinery through its activation domain, allowing mRNA accumulation (Di Mauro et al., 2000).

Three possible states of the ADH2 promoter have been proposed: structurally and functionally inactive, structurally derepressed but functionally inactive and fully derepressed and functionally active. The three possible states of the promoter, due to the absence or the presence of different Adr1p portions, can be considered to be an ordered sequence of events occurring at the ADH2 locus during derepression (Di Mauro et al., 2000).

Verdone (2002) analysed the in vivo chromatin structure and the kinetics of transcriptional activation of the S. cerevisiae ADH2 promoter as a function of genetically modified histone acetylation levels. By genetically altering the steady-state pattern of histone acetylation at the repressed ADH2 promoter, the structure of the nucleosome containing the TATA box is destabilized, the promoter becomes accessible to Adr1p and, when the cells are shifted to derepressing conditions, the kinetics of mRNA accumulation is faster. Histone deacetylation/acetylation is, therefore, directly involved in altering the chromatin structure at the ADH2 promoter, influencing the binding of the major transcriptional activator, with a concomitant effect on the kinetics of mRNA accumulation.

Xella (2006) later established the potential requirement for the chromatin remodelling factors Isw1p, Isw2p and Chd1p in the regulation of the ADH2 gene. They concluded that these factors contributed to the kinetics of activation of ADH2 in response to glucose depletion and were crucial in establishing the correct chromatin structure across the ADH2-coding region, but seemed largely dispensable for nucleosome organization at the promoter.

Regulation of ADH1, ADH3, ADH4 and ADH5

Initially, it was stated that regulatory genes affecting Adh2p mRNA expression appeared to have no effect on ADH1 expression (Denis et al., 1983). It was later established that deletion of the genes CCR4 and CAF1 not only influenced the repression of ADH2, but brought about a phenotype with enhanced expression of ADH1 (Liu et al., 2001). A conserved sequence UAS_RPG, present in various glycolytic genes and ribosomal protein genes, also exists in the upstream region of ADH1 and appears to be essential for efficient transcription (Santangelo & Tornow et al., 1990). However, this seems to be true only for cells in the exponential growth phase, because the efficiency of the promoter is virtually indistinguishable with or without the UAS_RPG during growth on ethanol. In glucose-grown cells, the region between −700 and −412 bp containing the UAS_RPG is required for the start of ADH1 promoter activity during the early exponential growth phase. The promoter region 750 bp upstream of the ADH1 promoter features the presence of an Adr1p-binding site. It is, therefore, tempting to speculate that the downregulation of the long ADH1 promoter is indeed due to activation of the upstream promoter by Adr1p. Thus, it is possible that the presence of ethanol in the culture broth rather than the absence of glucose decreased the activity of the long ADH1 promoter in glucose-grown cells (Ruohonen et al., 1995). Bird (2006) demonstrated that during zinc starvation, Zap1p was required for the repression of ADH1 expression. Zap1p binds upstream of the activator Rap1p and induces an intergenic RNA transcript (ZRR1). ZRR1 expression leads to the transient displacement of Rap1p and Gcr1p from the promoter, resulting in ADH1 repression.

The complex regulation of ADH2 has made it an attractive model system for the genetic dissection of its transcriptional control, while the regulation of the remainder of the ADHs has not received as much attention. Scant data are available regarding the up- or downregulation of these genes and no evidence exists regarding genetic elements associated with this action. Similar to the behaviour of enzymes of the tricarboxylic acid cycle (Heick et al., 1969), ADH3 is repressed by glucose, although its repression is not as severe as that of ADH2 (Ciriacy, 1975a; Young & Pilgrim et al., 1985), and ADH3 is derepresssed after glucose is depleted from the medium. Repression of ADH3 through expression of an intergenic transcript under conditions of zinc starvation will likely occur by a mechanism similar to that of ADH1, a mechanism that may act to conserve zinc during a limitation of this nutrient (Bird et al., 2006).

ADH4 expression is upregulated by lithium, a compound that is toxic to yeast cells grown on galactose, but is downregulated by dimethyl sulphoxide (DMSO) (Bro et al., 2003; Zhang et al., 2003). Under conditions where Adh1p is nonfunctional, spontaneous chromosomal amplification of ADH4 was able to rescue the mutant phenotype (Dorsey et al., 1993). ADH4 was also stringently regulated by zinc without an observable phenotype (Yuan et al., 2000). ADH5 expression was unaffected by DMSO (Zhang et al., 2003) and its transcription was significantly increased in an S. cerevisiae mutant strain able to grow anaerobically on d-xylose, a carbon source not normally utilized by yeasts in the absence of oxygen (Sonderegger et al., 2004).

Verified and probable functions

Kinetically, Adh1p seems to be chiefly responsible for the production of ethanol from acetaldehyde in cells grown anaerobically or in the presence of a glucose excess, whereas Adh2p primarily functions to convert ethanol accumulated during aerobic growth to acetaldehyde, with the concomitant reduction of NAD⁺. Adh1p can also accomplish this task, though presumably less efficiently. These two enzymes contribute to the economy of the cell by stabilizing the NAD⁺–NADH ratio. In the case of these two isozymes the possibility of inter-substitution seems to exist; for instance, cells lacking Adh2p activity can grow on ethanol as a carbon source under aerobic conditions (Wills et al., 1976; Wills et al., 1982).

Wenger & Bernofsky (1971) reported that mitochondrial Adh3p was responsible in part for the coupled respiration of yeast mitochondria with ethanol as a substrate. Disruption mutants of ADH3 showed no discernible mutant phenotype other than the lack of Adh3p activity (Young & Pilgrim et al., 1985). Bakker (2000) demonstrated that Adh3p was involved in the shuttling of mitochondrial NADH to the cytosol, where it is deoxidized by the external NADH dehydrogenases. Their results supported the hypothesis that Adh3p forms part of the ethanol–acetaldehyde shuttle that is necessary for the reoxidation of mitochondrial NADH under anaerobic conditions.

The Adh4p protein has been characterized as a zinc-dependent ADH that was thought to be minimally expressed, if at all (Drewke & Ciriacy et al., 1988). Williamson & Paquin (1987) were unable to detect ADH4 transcripts on Northern blots of laboratory strains grown on glucose or ethanol. Activity was only observed upon insertion of a Ty transposon at ADH4 or amplification of ADH4. It may be a ‘cryptic’ gene with no function or simply a gene that is not necessary under laboratory conditions. Genetic and physiological analysis showed that disruption of ADH4 did not influence the viability of the yeast cell, nor was the enzyme responsible for the decrease in ethanol production in adh1–adh4 quadruple deletion mutants (Drewke et al., 1990). It was suggested recently that ADH4 encoded the major cytosolic ADH in the two brewing yeast strains 2DGR19 and NCYC1245 in which elevated ADH4 mRNA levels could be detected under conditions, of high ethanol production (Mizuno et al., 2006). Yuan (2000) hypothesized that the induction of ADH4 expression under low-zinc conditions suggested that under these conditions the protein functioned as a back-up for Adh1p. Bird (2006) later also demonstrated the same regulation pattern during zinc starvation and suggested two possible models for induction: as Adh4p binds less zinc per subunit than Adh1p, the Adh4p enzyme could be a more efficient ADH during zinc limitation or Adh4p might bind iron when zinc levels are limiting.

Drewke (1990) found that a deletion mutant strain lacking adh1 to adh4 was still able to produce ethanol when grown on glucose as a carbon source. It is, therefore, reasonable to believe that in this case Adh5p might have been the enzyme capable of producing ethanol from acetaldehyde.

A search was undertaken to discover the genes and enzymes used by S. cerevisiae in the catabolism of leucine to isoamyl alcohol (Dickinson et al., 1997), valine to isobutanol (Dickinson et al., 1998) and isoleucine to active amyl alcohol (Dickinson et al., 2000). As long as the yeast had one functional enzyme out of Adh1p–Adh5p or Sfa1p, it was viable and any one of these six enzymes was sufficient for the final stage of amino acid catabolism, namely the conversion of an aldehyde to a long chain or a complex alcohol (Dickinson et al., 2003).

The specificity of the substrate and cofactor strongly supports the physiological involvement of Adh6p in aldehyde reduction rather than in alcohol oxidation and under oxidative conditions allows the yeast to use 2,3-butanediol as a carbon and energy source (Larroy, 2002a). The potential role of Adh6p in S. cerevisiae is not easy to ascertain when simply considering its structural similarity to plant cinnamyl ADHs. It may afford the yeast the capacity to live in ligninolytic environments where products derived from lignin biodegradation may be available. Another potential function may include the biosynthesis of fusel alcohols (Larroy, 2002a) and most certainly NADP(H) homeostasis (Larroy et al., 2003). Recently, genome-wide transcription analysis also identified the ADH6 gene as encoding NADPH-dependent 5-hydroxymethyl furfural (HMF) reduction activity (Petersson et al., 2006). It is also plausible that manipulation of the levels of Adh6p and Adh7p could be used by the fermentation industry to alter the organoleptic properties of fermented beverages (Larroy, 2002b).

Outlook

Even though the ADHs of S. cerevisiae have been intensively investigated, many aspects remain that warrant further research. The possibility of functional substitution among the different enzymes remains an interesting concept. Further data on this would contribute to assessing the in vivo roles of the enzymes. These investigations should be conducted not only under standard growth conditions or conditions of carbon repression, but also under conditions of other nutrient limitations such as a zinc limitation. Biochemical analysis of Adh5p has not yet received much attention. Information on Adh5p could provide an insight regarding the function of this enzyme in, for instance, amino acid metabolism. It is also appealing to consider the prospect of chimeric enzymes with improved or dual functions.

Author notes

Editor: Patrizia Romano