Background

Saccharomyces cerevisiae is the most common species used in fermentations of alcoholic beverages at industrial level due to its ability to overcome other yeasts. Conversely, S. kudriavzevii species, which has not been related to industrial processes, has been isolated from decayed leaves in Japan [1] as well as from oak barks in Portugal [2] and Spain [3]. Nevertheless, natural hybrids between S. cerevisiae and S. kudriavzevii conducting wine fermentations have been discovered and characterized by genetic approaches [4–8].

The hybridization process among Saccharomyces species has been proposed as an adaptation mechanism to ferment at low temperatures [9–11]. Physiological data suggest that Saccharomyces hybrids might have inherited the ability to grow at high temperatures (30–37 °C) and ethanol tolerance from S. cerevisiae and the ability to grow at low temperatures (10–16 °C) from S. kudriavzevii [12]. These physiological characteristics point to Saccharomyces hybrids as better suited to produce wines in accordance with the new trends in winemaking, such as low temperature fermentations and increased aroma [13–16]. Oenological characterization of S. cerevisiae × S. kudriavzevii hybrid strains has demonstrated that the hybrids are well adapted to ferment at low and intermediate temperatures, producing moderate or higher levels of glycerol and less acetic acid with regard to reference strains of S. cerevisiae and S. kudriavzevii [17, 18]. Similar comparative studies additionally including S. uvarum and a hybrid between S. cerevisiae and S. uvarum indicated that the highest production of glycerol was produced by S. uvarum, S. kudriavzevii and the S. cerevisiae × S. uvarum hybrid [19–21]. Regarding aroma formation, González et al. [18] indicated that hybrids produced the same quantity of aromatic compounds as S. cerevisiae at high temperatures, and the same aromatic intensity as S. kudriavzevii at low temperatures, whereas Gamero et al. [21] found that this trend was only observed in case of fusel alcohol production. Moreover, in the latter study, the best aroma producers at 28 °C were S. cerevisiae strains, whereas S. uvarum and several hybrids excelled at 12 °C. Altogether, these studies pointed to the fact that aroma formation is highly dependent on both yeast strain and fermentation temperature [21].

Higher alcohols, acetate esters and ethyl esters are quantitatively the most important family of compounds forming secondary aroma. These compounds are synthesized by yeasts during alcoholic fermentation as a consequence of its secondary metabolism, a complex biochemical process in which a lot of interconnected reactions are involved. In the formation of these compounds numerous enzymes participate, such as permeases, transaminases, decarboxylases, reductases, acetyltransferases and acyltransferases [13, 22, 23]. Besides, Saccharomyces yeasts can also be involved in primary aroma improvement, for instance, through the release of monoterpenes by the action of glycosidases [24–26].

In addition to aroma compounds, other metabolites affecting sensorial profile of the wine can be formed during winemaking such as ethanol, acetic acid or acetaldehyde. Ethanol is one of the main compounds synthesized in wine fermentation and decreases flavour perception by increasing aromatic compounds solubility in wine thus lowering the volatile fraction [27]. Acetic acid is the main compound constituting wine volatile acidity, giving undesirable odor when present in high concentration. Excessive acetic acid concentration can occur as a consequence of stuck and sluggish fermentations [28]. Acetaldehyde is obtained by pyruvate decarboxylation, being most of it reduced to ethanol. However, when this compound remains in wine in excessive amount contributes to the oxidized perception [28].

Aroma synthesis involves very complex processes where different metabolic pathways are interconnected and several genes participate. After the S. cerevisiae genome sequencing [29], transcriptomic, proteomic, metabolomic, and phenotypic analyses have been conducted. DNA microarrays are one of the most powerful tools to analyze the whole transcriptome in one single analysis. However, all the studies using this technology for better understanding winemaking processes have been done carrying out fermentations with S. cerevisiae [30–37] or comparing the expression of different species of the genus [38]. Only some studies dealing with the hybrid S. pastorianus on beer are available [39, 40].In this way, there is no information about S. cerevisiae × S.kudriavzevii hybrids.

In the present research work, several genes involved in aroma synthesis have been studied at molecular level in two S. cerevisiae × S. kudriavzevii hybrids at the beginning of stationary phase during winemaking at 12 and 28 °C. The particular objectives of this research work were to assess the differences in allele composition and gene expression in the two hybrids and their impact to aroma formation at each temperature. In addition, the contribution of the alleles of each parental species to the aroma formation was explored by RT-PCR.

Results

In the present research work, several genes involved in aroma synthesis have been studied in terms of allele composition and expression in two S. cerevisiae × S. kudriavzevii hybrids in winemaking conditions. The selection of the hybrids was based on their differences in both genotype and phenotype. Both hybrids are allotriploid, containing diploid S. cerevisiae genome and haploid S. kudriavzevii genome; although VIN7 is an almost perfect allotriploid hybrid whereas Lalvin W27 presents several chimerical chromosomes [6–8]. Regarding aroma profile, wine fermentations carried out at 12 °C by Lalvin W27 presented higher amount of higher alcohols compared to VIN7, whereas the latter excelled in the production of both acetate and ethyl esters. On the contrary, at 28 °C, those differences were no so evident [21]. Finally, acommercial non-cryotolerant S. cerevisiae strain was used as a control.

Global analysis of gene expression

Gene expression of the two cryophilic hybrids used in this study, Lalvin W27 and VIN7, was compared to gene expression of reference mesophilic S. cerevisiae strain, Lalvin T73. Table 1 summarizes the number of overlapping and selectively induced genes in the hybrids. Although the number of genes up or down regulated in case of VIN7 seems to be similar at both temperatures, a lower number of genes in Lalvin W27 appeared to be affected by high fermentation temperature. Moreover, there is a higher overlap between the up and down regulated genes at 12 °C than at 28 °C.

Table 1 Global gene expression in hybrids Lalvin W27 and VIN7 at 12 and 28 °C
Full size table

Genes with a fold change in expression greater than 2 (positive or negative) regarding S. cerevisiae Lalvin T73 were taken into consideration for further analysis. Up and down regulated metabolic functions (GO terms) for each hybrid at 12 and 28 °C appear in Tables 2 and 3, respectively. Almost no up regulated functions appeared at 12 °C in any of the hybrids, whereas metabolic functions related to transmembrane transport of sugars were down regulated in both hybrids. However, at 28 °C, the differences in the gene expression between the hybrids were more apparent. Most GO terms were found for down regulated genes in Lalvin W27 whereas in VIN7 only GO terms for up regulated genes were found. Lalvin W27 presented down regulation of most metabolic functions related to transmembrane transport of sugars.

Table 2 Go terms for the up and down regulated genes at 12 °C
Full size table
Table 3 Go terms for the up and down regulated genes at 28 °C
Full size table

Allele composition of aroma genes

Genes related to aroma production were selected among the expressed by the hybrids under fermentation conditions at 12 and 28 °C. With the aim to analyze the potential relationship between gene expression and parental species contribution, the allele composition of aroma related genes was determined for both hybrids (Fig. 1).

Fig. 1
figure 1

Hybrid allele composition and expression fold-change at 12 and 28 °C.*Lalvin W27 and VIN7 hybrid genome composition extracted from Belloch et al. [6] and Peris et al. [7], respectively; C S. cerevisiae allele, K S. kudriavzevii allele, grey colour allele composition of the genes located at chimerical chromosomes, green colour up regulation, red colour down regulation, – no hybridization. Comparison between temperatures is not possible due to the utilization of different controls

Full size image

Lalvin W27 is allotriploid and most of the genes related to aroma production contain two alleles coming from S. cerevisiae and one allele from S. kudriavzevii (CCK). The exceptions to this pattern are some genes located at the chimeric chromosomes IV, V, IX, XIV and XV. Genes located at chromosome IV are composed of either three alleles of S. cerevisiae (CCC), such as genes involved in aminoacid metabolism and pyruvate decarboxylation (ARO1, ARO10, BAP3, PDC2, THI3); or one allele coming from S. cerevisiae and two alleles of S. kudriavzevii (CKK), such as the alcohol dehydrogenase gene SFA1. In chromosome V, genes involved in aldehyde and isoleucine metabolism (ALD5, ILV1) are composed of two alleles coming from S. cerevisiae and two alleles from S. kudriavzevii (CCKK). In addition, at chromosomes IX and XV, the alcohol dehydrogenase gene ADH1 presents pattern (CCC) whereas the rest of the genes involved in aroma formation (ALD4, ATF1, IAH1, SPR1, YIR007W) present the regular pattern CCK. Finally, all genes related to aroma synthesis located in chromosome XIV present the common pattern CCK (LEU4).

Similarly, VIN7 is an almost perfect allotriploid hybrid, reason why most of the genes involved in aroma production presented two alleles coming from S. cerevisiae and one allele from S. kudriavzevii (CCK) (Fig. 1). Only two exceptions to this pattern were found; genes present in chromosome III (ADH7, ILV6, LEU2), which have lost the S. kudriavzevii part and contain two S. cerevisiae alleles (CC) and genes in chromosome VII, which appear to be composed either of three alleles of S. cerevisiae (CCC) (ATF2, BGL2) or present the regular pattern CCK (ADH4, ARO8, LEU1, PDC6).

Expression of genes related to aroma production

The expression of genes associated to aroma production, namely aminoacids, higher alcohols, acetate esters, ethyl esters, ethanol, acetaldehyde and acetate metabolism as well as primary aroma release (glycosidases and glucanases) were investigated by their expression levels at 12 and 28 °C fermentation temperatures in both hybrids. Figure 2 shows changes in gene expression using different colors. Down regulated genes seem to be predominant at 12 °C with very few up regulated genes in both hybrids. No genes appear differently induced in the hybrids. On the other hand, a similar number of down and up regulated genes can be found at 28 °C and BGL2 and PDC1 appear differently induced in both hybrids.

Fig. 2
figure 2

Heat maps depicting the level of expression of the aroma genes at 12 and 28 °C. Green up regulated, red down regulated, black no change in expression, grey no hybridization

Full size image

Expression of genes related to aminoacid and higher alcohol metabolism

Higher alcohols are produced from branched-chain aminoacids (leucine, isoleucine, valine), aromatic amino acids (phenylalanine, tyrosine, tryptophan) and the sulfur-containing amino acid methionine through the action of transaminases, decarboxylases and dehydrogenases. Figure 1 shows that at 12 °C, both hybrids presented clear up regulation in the gene ARO1 whereas LEU2 appeared up regulated exclusively in VIN7. Besides, several alcohol dehydrogenases and genes related to isoleucine biosynthesis were down regulated in both hybrids. The hybrid Lalvin W27 presented down regulation in BAT1 and LEU1. Finally, VIN7 showed down regulation in GAP1. At 28 °C, both hybrids presented differences in expression regarding alcohol dehydrogenases. In addition, down regulation of the permease gene GAP1 and the gene LEU1 could be observed in Lalvin W27, whereas down regulation in the gene ARO1 occurred in VIN7.

Genes related to ester production

Acetate esters are synthesized through the condensation of higher alcohols and acetyl-CoA by the action of acetyltransferases whereas ethyl esters are synthesized by condensation between ethanol and acyl-CoA by acyltransferases. Figure 1 shows that, at 12 °C, the hybrids did not present differences in expression, although Lalvin W27 presented down regulation in the gene YMR210W. At 28 °C, both hybrids presented similar expression, up regulation in ATF2 and EHT1 as well as down regulation in EEB1.

Genes related to ethanol, acetaldehyde and acetate metabolism

In the metabolism of ethanol, acetaldehyde and acetate several enzymes participate, such as alcohol dehydrogenases, aldehyde dehydrogenases, pyruvate decarboxylases and acetyl-CoA hydrolases and synthetases. Figure 1 shows that at 12 °C, both hybrids showed up regulation in most PDC genes as well as down regulation in most ADH and ALD genes. In addition, Lalvin W27 presented down regulation in ACS1. At 28 °C, both hybrids showed up regulation in some ADH and PDC genes as well as down regulation in other ADH and ALD genes. In addition, Lalvin W27 presented down regulation in some PDC genes.

Genes related to glycosidase and glucanase activities

Glycosidases and glucanases can contribute to aroma improvement, through the release of glycosilated terpenes and subsequent varietal aroma increase. Both hybrids showed down regulation of EXG1 at 12 °C. In addition, Lalvin W27 presented down regulation of ABF2 and VIN7 in SPR1 genes. However, VIN7 presented up regulation of YIR007W. At 28 °C, Lalvin W27 presented down regulation of BGL2, EXG1 and SPR1. On the contrary, VIN7 showed up regulation in BGL2.

Real-time PCR expression experiments

Real-time PCR experiments were carried out to investigate the contribution of each allele, S. cerevisiae or S. kudriavzevii, to the global expression showed by the hybrids in aroma genes ARO1, ATF2, BAT1 and EEB1. The selection of genes for RT-PCR was based on the different expression level of these genes among strains and temperatures as well as considering the genome composition differences between W27 and VIN7 (Fig. 1).

The calculated ratios S. kudriavzevii allele expression/S. cerevisiae allele expression (Table 4) showed that in some genes the allele coming from S. kudriavzevii parental was more expressed than its S. cerevisiae homologous. This seems to be the case of ARO1 gene in VIN7 and ATF2 gene in Lalvin W27. Moreover, this occurred at both temperatures, 12 and 28 °C, indicating that the expression of these genes is ruled by the expression of the S. kudriavzevii allele. Interestingly, these alleles were more efficiently expressed at 12 °C compared to 28 °C. In the case of genes BAT1 and EEB1, the alleles coming from the S. cerevisiae parental were more expressed than that of S. kudriavzevii for both strains at both temperatures. Finally, in accordance to allele composition, no expression of the S. kudriavzevii allele was detected in the gene ARO1 of Lalvin W27 as well as in the gene ATF2 of VIN7 in any of the tested temperatures.

Table 4 Allele composition and ratios S. kudriavzevii allele expression/S. cerevisiae allele expression of the four selected genes at both temperatures
Full size table

Discussion

This study focuses on the expression analysis of genes involved in aroma production by cryotolerant S. cerevisiae × S. kudriavzevii hybrids Lalvin W27 and VIN7 at 12 and 28 °C fermentation temperatures.

The selection of the hybrids was based in their genomic and phenotypic differences. Hybrid VIN7 is an almost perfect allotriploid hybrid whereas Lalvin W27 contains several chimerical chromosomes [6–8]. Moreover, oenological characterization of S. cerevisiae × S. kudriavzevii hybrids indicated that these hybrids may have inherited advantageous traits from their parental species, such as efficiency to ferment at low and intermediate temperatures or the production of aromas and glycerol, interesting properties for winemakers [17, 18, 21].

Analysis of global gene expression pointed to differences in the number of overlapping and induced genes (up and down regulated) in the hybrids between temperatures. Comparison of up and down regulated genes showed that there was a higher number of up and down regulated genes shared by both hybrids at 12 °C than at 28 °C, suggesting a similar level of response to cold in both hybrids. Analysis of GO terms pointed out a similar metabolic response at 12 °C consisting in the down regulation of genes related with sugar transmembrane transport activity in both hybrids. In a similar study comparing the global expression of parental species, S. cerevisiae and S. kudriavzevii, the down regulation of sugar transmembrane transporters was observed solely in S. cerevisiae fermenting at 12 °C and not in S. kudiavzevii [38]. This may indicate that the down regulation of these genes in the hybrids correspond to the S. cerevisiae alleles.

Examination of parental genetic structure in the hybrids [6–8] revealed different allele composition in several genes involved in aroma production, what may lead to different levels of expression and different aroma profile in the resulting wines. In case of VIN7, 56 % of the aroma genes diverged from the standard pattern consisting of two copies coming from S. cerevisiae and one copy from S. kudriavzevii, CCK. These differences allowed for changes in expression with respect to the S. cerevisiae reference strain, Lalvin T73, at one or both temperatures tested. Most of the differentially expressed aroma genes were related to higher alcohol and acetate ester metabolism, as well as enzymes involved in varietal aroma release. The down regulation of alcohol dehydrogenases in VIN7 at 12 °C, was found in accordance with the relative low production of higher alcohols showed by this hybrid (Fig. 3) [21]. Similarly, the up regulation in BGL2 in VIN7 at 28 °C would be in accordance with previous studies showing improved terpene release by S. cerevisiae and S. kudriavzevii hybrids respect to S. cerevisiae [41]. Besides, in case of Lalvin W27, this percentage of genes different from the pattern CCK was around 40 % and included genes related to higher alcohol, acetaldehyde and acetate metabolism. However, no correspondence was found between these results and previous studies except in case of 2-phenylethanol (ARO1) (Fig. 3) [21].

Fig. 3
figure 3

Most remarkable correlations between allele composition, gene expression and aroma formation. W27 Lalvin W27, D down regulated with respect to the reference strain Lalvin T73, = no changes in expression with respect to the reference strain Lalvin T73, U up regulated with respect to the reference strain Lalvin T73; aroma amounts extracted from Ref. [21]

Full size image

Data obtained in this study derived from RT-PCR correlated to previously published data regarding the allele composition of the hybrids Lalvin W27 and VIN7 [6–8]. This correlation can be seen in the genes where no S. kudriavzevii allele appears, such as ARO1 in Lalvin W27 and ATF2 in VIN7, since no expression derived from S. kudriavzevii was detected. Real time-PCR data pointed out a higher contribution of the S. kudriavzevii allele to the overall expression of the genes ARO1 and ATF2 in VIN7 and Lalvin W27, respectively. This increase in the expression level of both genes also occurred at both temperatures and interestingly, these alleles were more efficiently expressed at 12 °C compared to 28 °C, which reinforces the hypothesis that S. kudriavzevii is better adapted to ferment at low temperatures.

The up-regulation of ARO1 in Lalvin W27 and ATF2 in VIN7 was in accordance with the respective high levels of 2-phenylethanol and acetate esters produced by these strains (Fig. 3) [21]. Comparison with RT-PCR results indicate that the presence of three S. cerevisiae alleles (CCC) seem to positively influence aroma production. The higher expression of the S. kudriavzevii allele at 12 °C seems to have no effect on the production of 2-phenylethanol or acetate esters by these strains. In accordance with these results, previous studies presented S. kudriavzevii as bad aroma producer at moderate temperatures [21].

Hybridization process in yeasts has been proposed as an adaptation mechanism to ferment at low temperatures [9–11]. Data obtained in this study employing RT-PCR showed that the quantitative contribution of the alleles coming from one parental strain or the other to the overall gene expression was seemingly determined by the fermentation temperature.

Conclusions

The study at molecular level of aroma production carried out by yeasts is of undeniable complexity, especially when dealing with hybrid genomes. In this research work, the study of two S. cerevisiae × S. kudriavzevii hybrids revealed high differences regarding allele composition and gene expression, which resulted in different aroma profiles in the resulting wines. In addition, it has been pointed out the different contribution of the alleles coming from each parental strain to the overall expression, which can be positive or negative in terms of flavour synthesis. Finally, it has been demonstrated that the quantitative contribution to the overall gene expression of the alleles coming from one parental strain or the other clearly was determined by the fermentation temperature for some genes.

Methods

Yeast strains

The yeasts strains used in this study were the commercial S. cerevisiae strain Lalvin T73 used as reference and two hybrids between S. cerevisiae and S. kudriavzevii, Lalvin W27 and VIN7, isolated from wine in Switzerland and South Africa, respectively.

Allele composition of the genes related to aroma production

Genome composition of genes involved in aroma synthesis in the hybrids Lalvin W27 and VIN7 was determined from previously published data [6–8]. The allelic composition of each individual gene was determined searching the chromosome in which the gene was located employing Saccharomyces Genome Database [42].

Total RNA extraction and cDNA labelling

Cells were collected by centrifugation (4000 rpm/min, 5 min) from two independent fermentations at 12 and 28 °C at the beginning of stationary phase, determined when 50 % of reducing sugars were consumed. RNA extraction method was based on consecutive treatments with phenol-tris, phenol–chloroform (5:1) and chloroform-isoamyl alcohol (24:1), and a final precipitation with ethanol and sodium acetate [43]. RNA concentrations and purity were determined using a Nanodrop spectrophotometer ND-1000 (Nanodrop Technologies™, Wilmington, DE). RNA integrity was determined by electrophoresis in 1 % agarose gel. 2–4 μg of total RNA from each sample was linearly amplified using the Low RNA Input Fluorescent Linear Amplification kit (Agilent Technologies™, CA, USA). 2–3 μg of amplified cRNA was used as template for cDNA synthesis. cDNA was marked indirectly with “SuperScript™ Indirect cDNA Labeling System” (Invitrogen™, San Diego, CA). The fluorophores used were Cy3 and Cy5 mono-reactive Dye (Amersham GE Healthcare™, Amersham, UK) and dye incorporation was monitored using a Nanodrop spectrophotometer.

cDNA hybridization

A mixture of 200–300 pmol of the two labeled samples was concentrated in a Concentrator Plus (Eppendorf™, Hamburg, Germany). Competitive hybridization was performed on a Yeast 6.4K Array, PCR-amplified ORFs of yeast S288c strain, (Microarray Centre, UHN, Toronto, ON, Canada) in hybridization chambers AHC (ArrayIt Corporation, CA, USA) at 42 °C overnight. Heterologous conditions according to [38] were employed to assure the hybridization of the S. kudriavzevii genome. Pre-hybridization solution contained 3X SSC, 0.1 % SDS and 0.1 mg/ml BSA; hybridization solution contained 5× SSC, 0.1 % SDS and 0.1 mg/ml of salmon DNA. Microarrays were washed manually with different solutions containing different SSC 20× and SDS 10 % concentrations (Sol.1: 2× SSC-0.1 % SDS; Sol.2: 0.1× SSC-0.1 % SDS; Sol.3: 0.1 SSC; Sol4: 0.01× SSC). Signal intensities of Cy3 and Cy5 were acquired with an Axon GenePix 4100A scanner (Molecular Devices, CA, USA) using GenePix Pro v.6.1 software, at a resolution of 10 µm.

Microarray data analysis

Microarray data were derived from three independent experiments of cDNA hybridization. Raw data with global background subtraction were generated from GenePix pro 6.0. Analyses were done using the Acuity 4.0 software (Molecular Devices, CA, USA). Individual data sets were normalized to a log2 ratio value of 1. After normalization, data were filtered to remove the spots flagged as not found and were manually processed for print tip effect corrections. Only the spots with at least two replicates were considered. Finally, replicates were combined and their medians were calculated. The first cut-off was the selection of the genes presenting at least twofold log2 ratio values, according to the literature [33, 34, 36]. For these genes, a “GO terms” enrichment analysis was done using the GO Term Finder tool in the Saccharomyces Genome Database [42]. Regarding the statistics, a False Discovery Rate (FDR) analysis and a significance level of 99 % (p value <0.01) were applied. Heat maps and hierarchical clustering were done using the Genesis software 1.7.6 (Graz University of Technology, Austria).

The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE30779 http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE30779.

Real-time PCR (RT-PCR)

Expression of genes ARO1, ATF2, BAT1 and EEB1 was investigated by RT-PCR. Gene selection was based on expression differences among the hybrids at the temperatures tested, as observed at microarray data. Gene expression normalization was carried out using ACT1 since its expression remains constant along fermentation. Primer design was achieved using the tool Primer-BLAST (NCBI) and S. cerevisiae and S. kudriavzevii gene sequences deposited in databases [44]. Forward and reverse primers were designed to hybridize with S. cerevisiae or S. kudriavzevii alleles of the selected genes (Table 5). The specificity of the primers as well as their annealing temperature was confirmed by PCR Mastercycler pro (Eppendorf, Germany). RNA extraction and cDNA synthesis were carried out as previously explained. RT-PCR runs were done in triplicate in a LightCycler® 480 Real-Time PCR System (Roche, Switzerland) and analyzed employing the manufacturer software. Relative quantification of gene expression was achieved by comparison with ACT1 expression with kinetic PCR efficiency correction.

Table 5 Primers employed in the RT-PCR experiments for the genes ACT1, ARO1, ATF2, BAT1 and EEB1
Full size table

Gene expression level in the hybrids was expressed as the ratio between the expression of the alleles from S. kudriavzevii and S. cerevisiae.