Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
There has been considerable interest in modifying glycerol metabolism in Saccharomyces cerevisiae because of the sensory importance of glycerol in wine and other beverages (reviewed by Scanes et al., 1998). Although glycerol does not contribute directly to wine aroma, it can have a positive influence on the taste and mouth feel of wines. Glycerol is typically produced at a concentration of 4–10 g/l in dry table wines and is, apart from ethanol and carbon dioxide, the most abundant product of fermentation (Rankine and Bridson, 1971; Ough et al., 1972). Furthermore, increasing glycerol formation by S. cerevisiae can divert the flow of carbon away from ethanol during fermentation (Michnick et al., 1997; Remize et al., 1999; de Barros Lopes et al., 2000; Prior et al., 2000). This is potentially important, as it is a common practice to harvest fully ripened grapes, which can lead to the production of wines that have, in addition to enhanced varietal character, a high concentration of alcohol and which can be perceived by some as being out of balance.
In S. cerevisiae, glycerol is synthesized from the glycolytic intermediate dihydroxyacetone phosphate in two steps, reduction to glycerol 3-phosphate byglycerol 3-phosphate dehydrogenase (Gpd) followed by dephosphorylation by glycerol 3-phosphatase. Two isoenzymes of glycerol 3-phosphate dehydrogenase, encoded by GPD1 and GPD2, have been identified. The major role of Gpd1 activity is osmoadaptation (Albertyn et al., 1994), whereas Gpd2 activity appears to relate to maintaining redox balance, primarily during anaerobic growth (Ansell et al., 1997).
By overexpressing either the GPD1 or GPD2 gene in S. cerevisiae, which greatly increases the production of glycerol, wines with a decreased ethanol concentration have been made (Michnick et al., 1997; Remize et al., 1999; de Barros Lopes et al., 2000). Concomitant with the overproduction of glycerol, however, the concentration of acetic acid is two- to three-fold higher than in wine made with the non-modified strains. In some cases, the concentration of acetic acid can be high enough to have a negative sensory impact on the wine. The biosynthesis of acetic acid by the oxidation of acetaldehyde is catalysed by aldehyde dehydrogenases (AlDHs). Five genes encoding AlDH have been identified in the S. cerevisiae genome (Cherry et al., 1999; Navarro-Aviño et al., 1999). These genes have a role in the generation of acetyl-CoA from acetaldehyde, the metabolism of toxic aldehydes and the balance of cellular redox potential. Two of the genes, ALD4 and ALD5, are thought to encode mitochondrial enzymes, while ALD2, ALD3 and ALD6 encode cytoplasmic proteins. The ALD2 and ALD3 genes, as well as ALD5, are stress-inducible and glucose-repressed (Kurita and Nishida, 1999; Navarro-Aviño et al., 1999). The ALD6 gene product, which is activated by Mg2+ and preferentially uses NADP+, is functional in both ethanol- and glucose-grown cells (Meaden et al., 1997; Wang et al., 1998). This gene, together with ALD4, appears to contribute to the oxidation of acetaldehyde during fermentation (Remize et al., 2000).
In order to exploit wine yeasts that have increased glycerol and decreased ethanol production (from overexpression of GPD2) for winemaking, the problem of excessive acetic acid production must be overcome. In this study, the effect of decreased AlDH activity on the fermentation properties of wild-type and GPD2-OP laboratory strains grown in synthetic medium was determined. In addition, ‘metabolic snapshots’ (Raamsdonk et al., 2001) of the strains, determined by GC–MS, were used to study the formation of aroma and flavour compounds.
Materials and methods
Strains with a deletion in the ALD6 gene (YPL061W) were obtained by the sporulation and dissection of the ALD6/ald6Δ heterozygous diploid strain 22767 (MATa/MATα his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 LYS2/lys2Δ0 met15Δ0/MET15ura3Δ0/ura3Δ0, Research Genetics, USA). Both ald6Δ and ALD6 haploid strains were transformed with either the pAAH5 control plasmid or the GPD2 overexpressing plasmid, pAAH5–GPD2. The plasmid pAAH5 is a high copy number 2µ plasmid and contains the LEU2 marker gene for selection. Plasmid pAAH5–GPD2 was constructed by cloning the GPD2 gene downstream of the strong constitutive ADH1 promoter in plasmid pAAH5. The GPD2 gene was amplified from the 15DUα genome by polymerase chain reaction using the forward primer 5′-GCACTCAAAGATGACTGCTC-3′, the reverse primer 5′-GAAAAAGAGGCAACAGGAAA-3′, and proof-reading Pfu DNA polymerase (Stratagene, USA). The forward primer anneals to the start codon of the GPD2 gene and the reverse primer anneals to a region 46–66 bp downstream of the GPD2 gene (de Barros Lopes et al., 2000). Yeast strains were transformed using the lithium acetate method (Ausubel et al., 1994).
Media and fermentations
Yeast were grown in a synthetic leucine-free dropout medium (Ausubel et al., 1994) which consisted of 0.67% yeast nitrogen base without amino acids (Difco, USA), an amino acid mix lacking leucine and 8% D-glucose, at 28°C and 200 rpm. Overnight cultures were grown aerobically to early stationary phase and used to inoculate 50 ml medium at 1×106 cells/ml. Each strain was grown anaerobically (nitrogen-purged medium contained in a flask fitted with a water airlock) in duplicate. Cell numbers were determined spectrophometrically, approximate sugar concentrations were obtained by refractometry and the Clinitest® assay (Miles Inc., USA), and the samples were stored at −20°C for further analysis.
Concentrations of glucose, glycerol, acetic acid, succinic acid and ethanol were measured using enzymatic test kits (Boehringer Mannheim, Germany). Gas chromatography–mass spectrometry (GC–MS) was used to determine the relative ratio of selected volatile compounds, generally by peak area of the total ion chromatograms. For GC–MS, 2-ethyl hexanol was added as an internal standard. The samples (1 ml) were extracted at room temperature with an equal volume of diethyl ether and analysed with a Hewlett-Packard (HP) 6890 GC. The GC was fitted with a 30 m×0.25 mm fused silica capillary column DB-1701, film thickness 0.25 µm (J&W Scientific, USA). The oven temperature was set at 50°C and held at this temperature for 1 min before being increased to 250°C at 12°C/min and held at this temperature for a further 20 min. The injector was held at 220°C and the transfer line at280°C throughout the experiment. The sample volume injected was 1 µl. The carrier gas was helium at a flow rate of 1.2 ml/min. The splitter, at 30:1, was opened after 36 s in the splitless/split runs, run in purge splitless mode. Positive ion electron impact spectra at 70 eV were recorded in the range m/z 50–400 on an HP 5973 MS.
Several volatile metabolites were quantified using headspace GC–MS. Sodium chloride (1 g) and methyl heptanoate (10 µl, internal standard) were added, the sample (5 ml) was heated at 60°C for 15 min with intermittent vigorous shaking, and 1 ml of the headspace was analysed with a Hewlett-Packard (HP) 6890 GC fitted with an HP 7694 headspace sampler. The GC was fitted with a 30 m×0.25 mm BP20 capillary column, film thickness 0.25 µm (SGE, USA). The oven temperature was set at 40°C and held at this temperature for 4 min before being increased to 100°C at 20°C/min and then to 210°C at 25°C/min. The injector was held at 200°C and the carrier gas was helium at a flow rate of 1.0 ml/min. The split ratio was 15:1. Positive ion electron impact spectra at 70 eV were recorded on an HP 5973 MS.
Results and discussion
Decreasing acetic acid formation by strains overexpressing GPD2
As previously shown for a number of S. cerevisiae strains, the increased expression of GPD1 or GPD2 leads to the increased biosynthesis of glycerol during fermentation (Michnick et al., 1997; Remize et al., 1999; de Barros Lopes et al., 2000). In this study, the concentration of glycerol in the medium was increased from 5.1 g/l for the control strain (S288C laboratory strain background) to 13.4 g/l for the GPD2 overexpressing strain (GPD2-OP) when the cells were grown anaerobically in a synthetic medium containing 8% glucose (Table 1). The increased biosynthesis of glycerol was coupled to a decrease in the ethanol concentration and an increase in the acetic acid concentration. The changes in ethanol and acetic acid concentration are thought to be due to the sequestering of carbon away from ethanol production and the need to regenerate NADH, in this case by oxidizing acetaldehyde. The ethanol concentration in the medium was decreased by 24%, from 34.1 g/l for the wild type strain to 26.0 g/l for the GPD2-OP strain. The acetic acid concentration was more than doubled, increasing from 0.66 g/l to 1.42 g/l (see Table 1).
Table 1. Concentration of metabolites for wild-type, GPD2 and ALD6 modified strains
The following compounds are expressed as % of maximum# (rounded to nearest %):
Rancid, pungent, goaty
Sweet, woody, balsamic
4-Hydroxybenzene ethanol (tyrosol)3
Bees wax, honey-like
Rancid, soapy, fatty
Strategies to decrease the amount of acetic acid formed in the GPD2-OP strain were investigated. As the ALD4 and ALD6 genes have been shown to have a role in the oxidation of acetaldehyde to acetic acid (Meaden et al., 1997; Wang et al., 1998; Remize et al., 2000), these were chosen as primary targets for modification. Deletion of the ALD4 gene had no obvious effect on the biosynthesis of acetic acid in GPD2-OP cells when grown anaerobically (unpublished results). Loss of the ALD6 gene function affected the concentration of acetic acid in the medium, however. In an otherwise wild-type strain, the deletion of ALD6 led to a three-fold reduction in the concentration of acetic acid. Importantly, the concentration of acetic acid produced when the cells overexpress the GPD2 gene was also diminished when the ALD6 gene was deleted, with the concentration of acetic acid produced by the GPD2-OP ald6Δ strain (0.36 g/l) being lower than that obtained with the original non-modified strain. The biosynthesis of glycerol remained elevated (16.3 g/l), with a concomitant reduction in ethanol production (27.3 g/l).
The growth rate and fermentation rate were similar for the modified and wild-type strains (Figure 1), which is important if GPD2-OP ald6Δ strains are to be of use for winemaking, although the fermentation rate for the GPD2ald6Δ strain was slightly less than that of the other strains from 24 h onwards. The lower fermentation rate for the GPD2ald6Δ strain was not reflected in a lower growth rate, which was indistinguishable from that of the other strains. This finding contrasts with studies on ALD6, which have shown that the growth rate of ald6Δ strains was significantly less than that of the parental strain on glucose (Meaden et al., 1997; Navarro-Aviño et al., 1999). The phenotype might be strain-dependent, however, as other reports concur with our findings (Wang et al., 1998; Remize et al., 2000).
Formation of secondary metabolites by modified yeasts
There is a potential for aroma and flavour modification during winemaking as a result of altering glycerol and acetic acid metabolism. To test this, the effect on the biosynthesis of secondary fermentation products by GPD2-OP and GPD2-OP ald6Δ strains was examined. These results are summarized in Table 1. For the GPD2-OP strain, the results obtained for acetaldehyde, 2,3-butanediol, acetoin and succinic acid concentrations generally confirm previous findings for strains that produce more glycerol (Michnick et al., 1997; Remize et al., 1999; de Barros Lopes et al., 2000; Prior et al., 2000). The change in the acetaldehyde concentration in the medium is one of the most notable, increasing more than 10-fold when GPD2 is overexpressed. Increased acetaldehyde production was observed for both GPD2-OP ALD6 and GPD2-OP ald6Δ strains, indicating that it was independent of the ALD6 gene. This result suggests that metabolism to acetic acid by Ald6 is not a major pathway for the extra acetaldehyde that is produced by the glycerol overproducing strains.
GC–MS was used to identify compounds that were influenced by increased Gpd activity. The concentrations of the cis- and trans- isomers of 4-hydroxymethyl-2-methyl-1,3-dioxolane and 5-hydroxy-2-methyl-1,3-dioxane were increased more than 10-fold in the strains overexpressing GPD2. These compounds are ethylidene glycerols (acetals of glycerol and acetaldehyde) and have been previously found in wines that contain a high concentration of acetaldehyde, such as sherry (Williams and Strauss, 1978; Etievant, 1979). It is unknown whether these compounds are produced by enzymatic reactions in the yeast cell or their production occurs in the medium during or after fermentation. The concentration of a number of compounds, including 4-hydroxybenzaldehyde, 4-hydroxybenzene ethanol (tyrosol), isoamyl alcohol and isobutanol, was lower in medium fermented with the GPD2 overexpressing strains. Of these compounds, the change in the concentration of isobutanol was the most prominent, with the concentration decreasing by 70–77%. The change in concentration of these compounds, which are products of amino acid catabolism, was probably due to the cells' response to the alteration in redox balance created by increased glycerol biosynthesis (van Dijken and Scheffers, 1986). The concentration of these compounds is higher in medium fermented by ald6Δ strains, where their formation may partially replace that of acetic acid as a ‘redox sink’.
Many of the metabolites in the synthetic medium that were altered by the modification in GPD2 and ALD6 are important aroma contributors in wine (see Table 1). For example, an acetic acid concentration exceeding approximately 0.7 g/l can lead toa vinegar odour in some wine styles, and this sensory property was the primary reason for investigating the influence of ALD genes in the GPD2-OP yeast. Acetaldehyde can impart oxidized, nutty or bruised apple aromas and the substantial variation in the concentration of this metabolite (>10-fold in this study) could strongly influence the sensory character of wine if the concentration was above the sensory threshold. Isoamyl acetate, which was only detected in the GPD2 ald6Δ ferments, has a banana or pear aroma. The other important sensory metabolites changed to a lesser extent, but the change might contribute to a modification of the sensory property of a wine because, even when they are present in sub-threshold concentrations, variations in the relative amounts of compounds can have a subtle and complexing effect on wine aroma. For example, ethyl acetate is the main ester occurring in wines and, depending on its concentration, imparts a fruity or solvent (varnish) odour. The aroma of isoamyl alcohol has been described as marzipan-like. The concentration of some phenolic compounds was also shown to change by the modification of glycerol and acetic acid metabolism. Some of these compounds (e.g. tyrosol) are believed to have protective effects against several pathologies, which is mainly attributed to their antioxidant properties (Frankel et al., 1993).
Strains of S. cerevisiae that produce higher concentrations of glycerol alter the production of additional metabolites to offset the changes in carbon and NADH use (Michnick et al., 1997; Remize et al., 1999; de Barros Lopes et al., 2000; Prior et al., 2000). One of the main compensatory reactions that has important consequences for winemaking is the increased biosynthesis of acetic acid. It is shown here that the excessive amounts of acetic acid are not produced in a GPD2 overexpressing strain when one of the AlDH genes, ALD6, is deleted. These modifications may provide a useful strategy to increase the glycerol concentration, and thereby decrease the ethanol concentration, in wines produced from high sugar juices and musts. These results have been obtained with laboratory strains of S. cerevisiae growing in synthetic medium, and further experiments using wine strains and grape juice are required to determine the sensory impact of these changes in wine.
The use of ‘metabolic snapshots’ in yeast for the analysis of mutant phenotypes is a relatively new concept (Raamsdonk et al., 2001). Analysis of yeast metabolomes is particularly useful when the phenotype is silent by traditional measurements, such as growth rate. In our study, investigation of the ald6Δ and GPD2-OP mutants was extended by analysis of the mutant metabolome, using GC–MS to assign the identity of metabolites. This approach has proved useful for investigating the effect of ald6Δ and GPD2-OP genotypes on whole-cell metabolic networks, rather than the isolated pathways involved in glycerol and acetic acid biosynthesis.
We are grateful to On Sin and Robert Smillie (Carlton & United Breweries Ltd, Abbotsford, Victoria, Australia) for headspace GC–MS analysis of the ferments. We thank Mark Sefton for assistance with interpretation of the GC–MS results and Peter Høj for encouraging this work and for critical reading of the manuscript. This work was funded by the Grape and Wine Research and Development Corporation and the Cooperative Research Centre for Viticulture.