Genomic signatures of adaptation to wine biological ageing conditions in biofilm-forming flor yeasts.

The molecular and evolutionary processes underlying fungal domestication remain largely unknown despite the importance of fungi to bioindustry and for comparative adaptation genomics in eukaryotes. Wine fermentation and biological ageing are performed by strains of S. cerevisiae with, respectively, pelagic fermentative growth on glucose and biofilm aerobic growth utilizing ethanol. Here, we use environmental samples of wine and flor yeasts to investigate the genomic basis of yeast adaptation to contrasted anthropogenic environments. Phylogenetic inference and population structure analysis based on single nucleotide polymorphisms revealed a group of flor yeasts separated from wine yeasts. A combination of methods revealed several highly differentiated regions between wine and flor yeasts, and analyses using codon-substitution models for detecting molecular adaptation identified sites under positive selection in the high-affinity transporter gene ZRT1. The cross-population composite likelihood ratio revealed selective sweeps at three regions, including in the hexose transporter gene HXT7, the yapsin gene YPS6 and the membrane protein coding gene MTS27. Our analyses also revealed that the biological ageing environment has led to the accumulation of numerous mutations in proteins from several networks, including Flo11 regulation and divalent metal transport. Together, our findings suggest that the tuning of FLO11 expression and zinc transport networks are a distinctive feature of the genetic changes underlying the domestication of flor yeasts. Our study highlights the multiplicity of genomic changes underlying yeast adaptation to man-made habitats and reveals that flor/wine yeast lineage can serve as a useful model for studying the genomics of adaptive divergence.

Dynamics and quantitative analysis of the synthesis of fermentative aromas by an evolved wine strain of Saccharomyces cerevisiae.

We performed a dynamic and quantitative analysis of the synthesis of fermentative aromas by an aromatic wine yeast, ECA5, obtained by adaptive evolution. During fermentation at pilot scale on synthetic and natural musts, ECA5 produced volatile compounds (higher alcohols and their acetates, ethyl esters) at higher rates than the ancestral strain, with the exception of propanol. Marked differences in the chronology of synthesis of several compounds were observed between the two strains. Overproduction of phenyl ethanol occurred mainly during the growth phase for ECA5, consistent with its higher flux through the pentose phosphate pathway, which plays a key role in biosynthetic processes. The kinetics of production of isobutanol and isoamyl alcohol were differently affected by different media (synthetic or natural must) and, in particular, according to the nature of the sterols in the media (ergosterol or phytosterols). We also observed differences in the chronology of synthesis of ethyl acetate and isoamyl acetate or ethyl esters, suggesting that the regulation of the synthesis of these compounds in the evolved strain differs from that in the ancestral strain. This study shows that a dynamic analysis of volatile compounds, using high acquisition frequency online gas chromatography, can provide novel insights into the synthesis and regulation of aromas and is thus a potentially powerful tool for strain characterization.

Glycerol formation during wine fermentation is mainly linked to Gpd1p and is only partially controlled by the HOG pathway.

Glycerol 3-phosphate dehydrogenase, a key enzyme in the production of glycerol, is encoded by GPD1 and GPD2. The isoforms encoded by these genes have different functions, in osmoregulation and redox balance, respectively. We investigated the roles of GPD1, GPD2 and HOG1-the kinase involved in the response to osmotic stress-in glycerol production during wine fermentation. We found that the deletion of GPD2 in a wine yeast-derived strain did not affect growth or fermentation performance and reduced glycerol production by only 20%. In contrast, a gpd1delta mutant displayed a prolonged lag phase, and produced 40% less glycerol than the wild-type strain. The deletion of HOG1 resulted in a slight decrease in growth rate and a 20% decrease in glycerol production, indicating that the HOG pathway operates under wine fermentation conditions. However, the hog1delta mutant was not as severely affected as the gpd1delta mutant during the first few hours of fermentation, and continued to express GPD1 strongly. The hog1delta mutant was able to increase glycerol production in response to high sugar concentration (15-28% glucose), to almost the same extent as the wild-type, whereas this response was totally abolished in the gpd1delta mutant. These data show that Gpd1p plays a major role in glycerol formation, particularly during the first few hours of exposure to high sugar concentration, and that GPD2 is only of little significance in anaerobic fermentation by wine yeast. The results also demonstrate that the HOG pathway exerts only limited control over GPD1 expression and glycerol production during wine fermentation.

Glycerol export and glycerol-3-phosphate dehydrogenase, but not glycerol phosphatase, are rate limiting for glycerol production in Saccharomyces cerevisiae.

Glycerol, one of the most important by-products of alcoholic fermentation, has positive effects on the sensory properties of fermented beverages. It was recently shown that the most direct approach for increasing glycerol formation is to overexpress GPD1, which encodes the glycerol-3-phosphate dehydrogenase (GPDH) isoform Gpd1p. We aimed to identify other steps in glycerol synthesis or transport that limit glycerol flux during glucose fermentation. We showed that the overexpression of GPD2, encoding the other isoform of glycerol-3-phosphate dehydrogenase (Gpd2p), is equally as effective as the overexpression of GPD1 in increasing glycerol production (3.3-fold increase compared to the wild-type strain) and has similar effects on yeast metabolism. In contrast, overexpression of GPP1, encoding glycerol 3-phosphatase (Gpp1p), did not enhance glycerol production. Strains that simultaneously overexpress GPD1 and GPP1 did not produce higher amounts of glycerol than a GPD1-overexpressing strain. These results demonstrate that GPDH, but not the glycerol 3-phosphatase, is rate-limiting for glycerol production. The channel protein Fps1p mediates glycerol export. It has recently been shown that mutants lacking a region in the N-terminal domain of Fps1p constitutively release glycerol. We showed that cells producing truncated Fps1p constructs during glucose fermentation compensate for glycerol loss by increasing glycerol production. Interestingly, the strain with a deregulated Fps1 glycerol channel had a different phenotype to the strain overexpressing GPD genes and showed poor growth during fermentation. Overexpression of GPD1 in this strain increased the amount of glycerol produced but led to a pronounced growth defect.

The potential of genetic engineering for improving brewing, wine-making and baking yeasts.

The end of the twentieth century was marked by major advances in life technology, particularly in areas related to genetics and more recently genomics. Considerable progress was made in the development of genetically improved yeast strains for the wine, brewing and baking industries. In the last decade, recombinant DNA technology widened the possibilities for introducing new properties. The most remarkable advances, which are discussed in this Mini-Review, are improved process performance, off-flavor elimination, increased formation of by-products, improved hygienic properties or extension of substrate utilization. Although the introduction of this technology into traditional industries is currently limited by public perception, the number of potential applications of genetically modified industrial yeast is likely to increase in the coming years, as our knowledge derived from genomic analyses increases.

Characterization of Schizosaccharomyces pombe malate permease by expression in Saccharomyces cerevisiae.

In Saccharomyces cerevisiae, L-malic acid transport is not carrier mediated and is limited to slow, simple diffusion of the undissociated acid. Expression in S. cerevisiae of the MAE1 gene, encoding Schizosaccharomyces pombe malate permease, markedly increased L-malic acid uptake in this yeast. In this strain, at pH 3.5 (encountered in industrial processes), L-malic acid uptake involves Mae1p-mediated transport of the monoanionic form of the acid (apparent kinetic parameters: Vmax = 8.7 nmol/mg/min; Km = 1.6 mM) and some simple diffusion of the undissociated L-malic acid (Kd = 0.057 min(-1)). As total L-malic acid transport involved only low levels of diffusion, the Mae1p permease was further characterized in the recombinant strain. L-Malic acid transport was reversible and accumulative and depended on both the transmembrane gradient of the monoanionic acid form and the DeltapH component of the proton motive force. Dicarboxylic acids with stearic occupation closely related to L-malic acid, such as maleic, oxaloacetic, malonic, succinic and fumaric acids, inhibited L-malic acid uptake, suggesting that these compounds use the same carrier. We found that increasing external pH directly inhibited malate uptake, resulting in a lower initial rate of uptake and a lower level of substrate accumulation. In S. pombe, proton movements, as shown by internal acidification, accompanied malate uptake, consistent with the proton/dicarboxylate mechanism previously proposed. Surprisingly, no proton fluxes were observed during Mae1p-mediated L-malic acid import in S. cerevisiae, and intracellular pH remained constant. This suggests that, in S. cerevisiae, either there is a proton counterflow or the Mae1p permease functions differently from a proton/dicarboxylate symport.

Global gene expression during short-term ethanol stress in Saccharomyces cerevisiae.

DNA microarrays were used to investigate the expression profile of yeast genes in response to ethanol. Up to 3.1% of the genes encoded in the yeast genome were up-regulated by at least a factor of three after 30 min ethanol stress (7% v/v). Concomitantly, 3.2% of the genes were down-regulated by a factor of three. Of the genes up-regulated in response to ethanol 49.4% belong to the environmental stress response and 14.2% belong to the stress gene family. Our data show that in addition to the previously identified ethanol-induced genes, a very large number of genes involved in ionic homeostasis, heat protection, trehalose synthesis and antioxidant defence also respond to ethanol stress. It appears that a large number of the up-regulated genes are involved in energy metabolism. Thus, 'management' of the energy pool (especially ATP) seems to constitute an ethanol stress response and to involve different mechanisms.

Engineering of the pyruvate dehydrogenase bypass in Saccharomyces cerevisiae: role of the cytosolic Mg(2+) and mitochondrial K(+) acetaldehyde dehydrogenases Ald6p and Ald4p in acetate formation during alcoholic fermentation.

Acetic acid plays a crucial role in the organoleptic balance of many fermented products. We have investigated the factors controlling the production of acetate by Saccharomyces cerevisiae during alcoholic fermentation by metabolic engineering of the enzymatic steps involved in its formation and its utilization. The impact of reduced pyruvate decarboxylase (PDC), limited acetaldehyde dehydrogenase (ACDH), or increased acetoacetyl coenzyme A synthetase (ACS) levels in a strain derived from a wine yeast strain was studied during alcoholic fermentation. In the strain with the PDC1 gene deleted exhibiting 25% of the PDC activity of the wild type, no significant differences were observed in the acetate yield or in the amounts of secondary metabolites formed. A strain overexpressing ACS2 and displaying a four- to sevenfold increase in ACS activity did not produce reduced acetate levels. In contrast, strains with one or two disrupted copies of ALD6, encoding the cytosolic Mg(2+)-activated NADP-dependent ACDH and exhibiting 60 and 30% of wild-type ACDH activity, showed a substantial decrease in acetate yield (the acetate production was 75 and 40% of wild-type production, respectively). This decrease was associated with a rerouting of carbon flux towards the formation of glycerol, succinate, and butanediol. The deletion of ALD4, encoding the mitochondrial K(+)-activated NAD(P)-linked ACDH, had no effect on the amount of acetate formed. In contrast, a strain lacking both Ald6p and Ald4p exhibited a long delay in growth and acetate production, suggesting that Ald4p can partially replace the Ald6p isoform. Moreover, the ald6 ald4 double mutant was still able to ferment large amounts of sugar and to produce acetate, suggesting the contribution of another member(s) of the ALD family.

Re-assessment of the influence of yeast strain and environmental factors on glycerol production in wine.

Increasing glycerol production is of concern for wine-makers in improving the quality of certain wines. We have compared the impact of strain and relevant environmental factors influencing glycerol production under the same conditions, i.e. standardized conditions simulating enological fermentation. The glycerol production of 19 industrial wine strains ranged from 6.4 to 8.9 g l-1 and varied significantly between strains. The production of acetate and succinate was also found to differ substantially depending on the strain but no significant strain-dependent variation was observed for acetaldehyde. Interestingly, high glycerol production was not correlated to high production of acetate or acetaldehyde, which are undesirable in wine. A detailed study with two low or two high glycerol-producing strains showed that temperature and the initial concentration of nitrogen had little effect on the amount of glycerol formed, although agitation or a nitrogen source composed mainly of ammoniacal nitrogen slightly enhanced glycerol production. The influence of environmental factors remained minor while the predominant factor for glycerol variability in wine was attributed to the strain. Taking into account wine-making constraints, the results indicate that achieving a high glycerol content in wine requires the selection or improvement of yeast strains rather than the control of growth and cultivation conditions.

Glycerol overproduction by engineered saccharomyces cerevisiae wine yeast strains leads to substantial changes in By-product formation and to a stimulation of fermentation rate in stationary phase

Six commercial wine yeast strains and three nonindustrial strains (two laboratory strains and one haploid strain derived from a wine yeast strain) were engineered to produce large amounts of glycerol with a lower ethanol yield. Overexpression of the GPD1 gene, encoding a glycerol-3-phosphate dehydrogenase, resulted in a 1.5- to 2.5-fold increase in glycerol production and a slight decrease in ethanol formation under conditions simulating wine fermentation. All the strains overexpressing GPD1 produced a larger amount of succinate and acetate, with marked differences in the level of these compounds between industrial and nonindustrial engineered strains. Acetoin and 2,3-butanediol formation was enhanced with significant variation between strains and in relation to the level of glycerol produced. Wine strains overproducing glycerol at moderate levels (12 to 18 g/liter) reduced acetoin almost completely to 2,3-butanediol. A lower biomass concentration was attained by GPD1-overexpressing strains, probably due to high acetaldehyde production during the growth phase. Despite the reduction in cell numbers, complete sugar exhaustion was achieved during fermentation in a sugar-rich medium. Surprisingly, the engineered wine yeast strains exhibited a significant increase in the fermentation rate in the stationary phase, which reduced the time of fermentation.

Modulation of glycerol and ethanol yields during alcoholic fermentation in Saccharomyces cerevisiae strains overexpressed or disrupted for GPD1 encoding glycerol 3-phosphate dehydrogenase.

The possibility of the diversion of carbon flux from ethanol towards glycerol in Saccharomyces cerevisiae during alcoholic fermentation was investigated. Variations in the glycerol 3-phosphate dehydrogenase (GPDH) level and similar trends for alcohol dehydrogenase (ADH), pyruvate decarboxylase and glycerol-3-phosphatase were found when low and high glycerol-forming wine yeast strains were compared. GPDH is thus a limiting enzyme for glycerol production. Wine yeast strains with modulated GPD1 (encoding one of the two GPDH isoenzymes) expression were constructed and characterized during fermentation on glucose-rich medium. Engineered strains fermented glucose with a strongly modified [glycerol] : [ethanol] ratio. gpd1delta mutants exhibited a 50% decrease in glycerol production and increased ethanol yield. Overexpression of GPD1 on synthetic must (200 g/l glucose) resulted in a substantial increase in glycerol production ( x 4) at the expense of ethanol. Acetaldehyde accumulated through the competitive regeneration of NADH via GPDH. Accumulation of by-products such as pyruvate, acetate, acetoin, 2,3 butane-diol and succinate was observed, with a marked increase in acetoin production.

Membrane potential-generating malate (MleP) and citrate (CitP) transporters of lactic acid bacteria are homologous proteins. Substrate specificity of the 2-hydroxycarboxylate transporter family.

Membrane potential generation via malate/lactate exchange catalyzed by the malate carrier (MleP) of Lactococcus lactis, together with the generation of a pH gradient via decarboxylation of malate to lactate in the cytoplasm, is a typical example of a secondary proton motive force-generating system. The mleP gene was cloned, sequenced, and expressed in a malolactic fermentation-deficient L. lactis strain. Functional analysis revealed the same properties as observed in membrane vesicles of a malolactic fermentation-positive strain. MleP belongs to a family of secondary transporters in which the citrate carriers from Leuconostoc mesenteroides (CitP) and Klebsiella pneumoniae (CitS) are found also. CitP, but not CitS, is also involved in membrane potential generation via electrogenic citrate/lactate exchange. MleP, CitP, and CitS were analyzed for their substrate specificity. The 2-hydroxycarboxylate motif R1R2COHCOOH, common to the physiological substrates, was found to be essential for transport although some 2-oxocarboxylates could be transported to a lesser extent. Clear differences in substrate specificity among the transporters were observed because of different tolerances toward the R substituents at the C2 atom. Both MleP and CitP transport a broad range of 2-hydroxycarboxylates with R substituents ranging in size from two hydrogen atoms (glycolate) to acetyl and methyl groups (citromalate) for MleP and two acetyl groups (citrate) for CitP. CitS was much less tolerant and transported only citrate and at a low rate citromalate. The substrate specificities are discussed in the context of the physiological function of the transporters.

Metabolic analysis of S. cerevisiae strains engineered for malolactic fermentation.

A complete malolactic fermentation was achieved using Saccharomyces cerevisiae strains coexpressing the genes mleS and mae1 coding for the Lactococcus lactis malolactic enzyme and the Schizosaccharomyces pombe malate permease under the control of yeast promoters. The expression level of mae1 greatly influences the kinetics of the reaction by controlling the rate of malate uptake meanwhile a high expression level of mleS induces a partial consumption of malate derived from glucose by the malolactic enzyme. A strain expressing several copies of mae1 and one copy of mleS degrades 3 g/l of malate almost exclusively through the malolactic pathway in 4 days under enological conditions, without metabolic side effects.

Malolactic fermentation by engineered Saccharomyces cerevisiae as compared with engineered Schizosaccharomyces pombe.

The ability of yeast strains to perform both alcoholic and malolactic fermentation in winemaking was studied with a view to achieving a better control of malolactic fermentation in enology. The malolactic gene of Lactococcus lactis (mleS) was expressed in Saccharomyces cerevisiae and Schizosaccharomyces pombe. The heterologous protein is expressed at a high level in cell extracts of a S. cerevisiae strain expressing the gene mleS under the control of the alcohol dehydrogenase (ADH1) promoter on a multicopy plasmid. Malolactic enzyme specific activity is three times higher than in L. lactis extracts. Saccharomyces cerevisiae expressing the malolactic enzyme produces significant amounts of L-lactate during fermentation on glucose-rich medium in the presence of malic acid. Isotopic filiation was used to demonstrate that 75% of the L-lactate produced originates from endogenous L-malate and 25% from exogenous L-malate. Moreover, although a small amount of exogenous L-malate was degraded by S. cerevisiae transformed or not by mleS, all the exogenous degraded L-malate was converted into L-lactate via a malolactic reaction in the recombinant strain, providing evidence for very efficient competition of malolactic enzyme with the endogenous malic acid pathways. These results indicate that the sole limiting step for S. cerevisiae in achieving malolactic fermentation is in malate transport. This was confirmed using a different model, S. pombe, which efficiently degrades L-malate. Total malolactic fermentation was obtained in this strain, with most of the L-malate converted into L-lactate and CO2. Moreover, L-malate was used preferentially by the malolactic enzyme in this strain also.

The Saccharomyces cerevisiae FLO1 flocculation gene encodes for a cell surface protein.

The sequencing of a 6619 bp region encoding for a flocculation gene previously cloned from a strain defined as FLO5 (Bidard et al., 1994) has revealed that it was a FLO1 gene. The FLO1 gene product has been localized at the cell surface of the yeast cell by immunofluorescent microscopy. The Flo1 protein contains four regions with repeated sequences which account for about 70% of the amino acids of this protein. A functional analysis of the major repeated region has revealed that it plays an important role in determining the flocculation level. A gene disruption experiment has shown that FLO5 strain STX 347-1D contains at least two flocculation genes of the FLO1 type but that they are supposed to be inactive and do not contribute to its flocculation. However, enzyme-linked immunosorbent assays performed on intact cells have revealed that a protein expressed at the cell surface of the FLO5 strain STX 347-1D is antigenically related to Flo1p. A deletion analysis of the 5' region of the FLO1 gene has shown that the expression is submitted to controls which depend on the genetic background of the strain.

Cloning and analysis of a FLO5 flocculation gene from S. cerevisiae.

A yeast flocculation gene was isolated from a genomic library of an FLO5 strain of S. cerevisiae on the basis of its ability to trigger flocculation in a non-flocculent strain. Characterization of the cloned gene by restriction mapping, Southern analysis, and chromosome mapping have shown that it corresponds to a FLO5 gene previously located on chromosome I and that this gene is related to the already described FLO1 gene. A study of gene expression in different yeast strains has indicated that, while this gene is dominant, its expression can be suppressed in some genetic backgrounds. A Northern-blot analysis has demonstrated that the same 5000-nt transcript was present in an FLO5 and an FLO1 strain. A gene disruption experiment has led to the conclusion that another flocculation gene is present and can be active in the FLO5 strain we used.

Mixed lactic acid-alcoholic fermentation by Saccharomyces cerevisiae expressing the Lactobacillus casei L(+)-LDH.

We describe the construction of a Saccharomyces cerevisiae strain expressing the gene encoding the L(+)-lactate dehydrogenase [L(+)-LDH)] from Lactobacillus casei. The recombinant strain is able to perform a mixed lactic acid-alcoholic fermentation. Yeast cells expressing the L(+)-LDH gene from the yeast alcohol dehydrogenase (ADH1) promoter on a multicopy plasmid simultaneously convert glucose to both ethanol and lactate, with up to 20% of the glucose transformed into L(+)-lactate. Such strains may be used in every field where both biological acidification and alcoholic fermentation are required.

Cloning, sequence and expression of the gene encoding the malolactic enzyme from Lactococcus lactis.

Many lactic acid bacteria can carry out malolactic fermentation. This secondary fermentation is mediated by the NAD- and Mn(2+)-dependent malolactic enzyme, which catalyses the decarboxylation of L-malate to L-lactate. The gene we call mleS, coding for malolactic enzyme, was isolated from Lactococcus lactis. The mleS gene consists of one open reading frame capable of coding for a protein with a calculated molecular mass of 59 kDa. The amino acid sequence of the predicted MleS gene product is homologous to the sequences of different malic enzymes. Bacterial and yeast cells expressing the malolactic gene convert L-malate to L-lactate.

Analysis of the chromosomal DNA polymorphism of wine strains of Saccharomyces cerevisiae.

Wine yeast strains are characterized by a high chromosomal DNA polymorphism. This can be explained partly by a size difference of different variants of specific chromosomes. This difference can reach up to 45% of the size of the chromosome in question. Two strains, SB1 and Eg8, have a very complex chromosomal pattern and show one band hybridizing with probes from two different chromosomes derived from a reference strain. This is an indication of the presence of "hybrid" chromosomes in these wine strains. The most astonishing result concerns chromosome VIII, frequently present in wine strains in two variant forms. The first normal form has a size of about 580 kb while the second is around 1000 kb. These two forms segregate at meiosis and recombine with a normal chromosome VIII from a laboratory strain. Wine yeasts are thus very different from haploid laboratory strains.

DNA sequences in chromosomes II and VII code for pyruvate carboxylase isoenzymes in Saccharomyces cerevisiae: analysis of pyruvate carboxylase-deficient strains.

A gene encoding pyruvate carboxylase has previously been isolated from Saccharomyces cerevisiae. We have isolated a second gene, PYC2, from the same organism also encoding a pyruvate carboxylase. The gene PYC2 is situated on the right arm of chromosome II between the DUR 1, 2 markers and the telomere. We localized the previously isolated gene, which we designate PYC1, to chromosome VII. Disruption of either of the genes did not produce marked changes in the phenotype. However, simultaneous disruption of both genes resulted in inability to grow on glucose as sole carbon source, unless aspartate was added to the medium. This indicates that in wild-type yeast there is no bypass for the reaction catalysed by pyruvate carboxylase. The coding regions of both genes exhibit a homology of 90% at the amino acid level and 85% at the nucleotide level. No appreciable homology was found in the corresponding flanking regions. No differences in the Km values for ATP or pyruvate were observed between the enzymes obtained from strains carrying inactive, disrupted versions of one or other of the genes.