Quantitative analysis of wine yeast gene expression profiles under winemaking conditions

YeastYeast 2005; 22: 369–383.Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/yea.1217

Research Article

Quantitative analysis of wine yeast gene expressionprofiles under winemaking conditions

Cristian Varela1#, Javier Cardenas1, Francisco Melo2 and Eduardo Agosin1*1Departamento de Ingenier ıa Quımica y Bioprocesos, Facultad de Ingenier ıa, Pontificia Universidad Catolica de Chile, Santiago, Chile2Departamento de Genetica Molecular y Microbiolog ıa, Facultad de Ciencias Biologicas, Pontificia Universidad Catolica de Chile, Santiago,Chile

*Correspondence to:Eduardo Agosin, PontificiaUniversidad Catolica de Chile,Casilla 306 Correo 22, Santiago,Chile.E-mail: [email protected]

#Current address: The AustralianWine Research Institute, PO Box197, Glen Osmond, Adelaide, SA5064, Australia.

Received: 17 November 2004Accepted: 23 December 2004

AbstractWine fermentation is a dynamic and complex process in which the yeast cell issubjected to multiple stress conditions. A successful adaptation involves changes ingene expression profiles where a large number of genes are up- or downregulated.Functional genomic approaches are commonly used to obtain global gene expressionprofiles, thereby providing a comprehensive view of yeast physiology. We usedSAGE to quantify gene expression profiles in an industrial strain of Saccharomycescerevisiae under winemaking conditions. The transcriptome of wine yeast wasanalysed at three stages during the fermentation process, mid-exponential phase,and early- and late-stationary phases. Upon correlation with the yeast genome,we found three classes of transcripts: (a) sequences that corresponded to ORFs;(b) expressed sequences from intergenic regions; and (c) messengers that did notmatch the published reference yeast genome. In all fermentation phases studied,the most highly expressed genes related to energy production and stress response.For many pathways, including glycolysis, different transcript levels were observedduring each phase. Different isoenzymes, including hexose transporters (HXT ), weredifferentially induced, depending on the growth phase. About 10% of transcriptsmatched non-annotated ORF regions within the yeast genome and could correspondto small novel genes originally omitted in the first gene annotation effort. Up to 22% oftranscripts, particularly at late-stationary phase, did not match any known locationwithin the genome. As the available reference yeast genome was obtained from alaboratory strain, these expressed sequences could represent genes only expressed byan industrial yeast strain. Further studies are necessary to identify the role of thesepotential genes during wine fermentation. Copyright 2005 John Wiley & Sons, Ltd.

Keywords: wine yeast; fermentation; SAGE; gene expression

Introduction

Yeasts are subjected to multiple and changingstress conditions during alcoholic fermentation,which is a dynamic and complex process. Theseconditions include hyperosmotic shock, nutrientlimitation and starvation, temperature variationsand ethanol toxicity. Wine yeasts have evolvedmechanisms to sense and respond to environmen-tal changes and thus maintain metabolic activ-ity and cellular integrity (Bauer and Pretorius,2000). The stress conditions that occur during

wine fermentation affect the yeast’s metabolism,producing a metabolic response that in turn enablesthe cell to adapt to the new environmental condi-tions. A successful adaptation implies a metabolicreorganization in order to maintain cellular activ-ity (vitality). Sometimes, this reorganization leadsto a decrease in the fermentation rate, generatingsluggish fermentations (Walker, 1998). When theadaptation mechanisms fail, the reduced viabilityis followed by a decrease in vitality, resulting instuck fermentations. A successful adaptation alsoinvolves changes in gene expression profiles where

Copyright 2005 John Wiley & Sons, Ltd.

370 C. Varela et al.

a large number of genes are up- or downregulated.Therefore, obtaining the specific expression pro-files of wine yeast during wine fermentation wouldallow understanding, at the molecular level, ofboth the biological process of wine fermentationand the regulation of gene expression in responseto changes in the environment (Perez-Ortin et al.,2002).

Among molecular biology techniques that enablethe whole transcriptome to be studied, only DNAmicroarrays have been used to describe the tran-scriptome of wine yeast studied in standard lab-oratory culture media, chemically defined winemedia and grape must (Backhus et al., 2001; Eras-mus et al., 2003; Hauser et al., 2001; Marks et al.,2003; Rossignol et al., 2003). Unfortunately, DNAmicroarrays are not fully quantitative and resultsfrom different experiments cannot be compareddirectly. From a global point of view of cellmetabolic regulation, where genomic informationcould be complemented with gene expression pro-files, total protein concentration or data from intra-cellular metabolites, it is of key importance to usemethods that enable quantification of all the tran-scripts present within the yeast cell. Serial analysisof gene expression (SAGE), a technique developedby Velculescu et al. (1995), enables the analysis ofthousands of expressed genes and a total quantifi-cation of each transcript.

SAGE is an experimental technique designed togain a direct and quantitative measure of geneexpression (Velculescu et al., 1995). This tech-nique has already been used to quantify the yeasttranscriptome (Kal et al., 1999; Velculescu et al.,1997), as well as to study the expression pro-files in different cellular types (Blomberg andZuelke, 2004; Chen et al., 1998; Norman et al.,2004). SAGE is based mainly on two princi-ples: representation of mRNAs by short (9–10 bp)nucleotide sequences (tags); and linkage of thesetags for cloning, which allows an efficient sequenc-ing analysis (Stollberg et al., 2000; Yamamotoet al., 2001). The expression level of each tran-script is quantified by the frequency that a par-ticular tag is observed. Unlike DNA microarrays,SAGE does not require prior knowledge of thegenes to be analysed; indeed, SAGE allows effi-cient identification of novel transcripts or novelgenes in the genome (Chen et al., 2000; Saha et al.,2002).

In this work, we used SAGE to quantify thegene expression profiles of the wine yeast Sac-charomyces cerevisiae EC1118 under winemakingconditions. To the best of our knowledge, this is thefirst report where the transcriptome of an industrialstrain of S. cerevisiae is quantitatively assessed.The differential expression of genes related to stressresponse, the discovery of potential novel genesand the expression of oxygen-dependent pathwaysunder anoxic conditions, are discussed.

Material and methods

Yeast strain and culture medium

An isolate of S. cerevisiae EC1118 (Lalvin, Zug,Switzerland), a widely used commercial winestrain, was used throughout this study. Initial seedcultures were grown in YPD medium at 28 ◦Cunder aerobic conditions. MS300 artificial must,which simulates a standard grape juice, was used inbioreactor fermentations (Salmon and Barre, 1998).MS300 medium was modified by increasing totalsugar concentration to 240 g/l and including equalparts of glucose and fructose (Varela et al., 2004).Assimilable nitrogen content of MS300 mediumwas 300 mg N/l supplied as ammonia and aminoacids.

Growth conditions

A 50 l Bioengineering bioreactor (Bioengineering,Wald, Switzerland) with a 35 l working volume wasinoculated to an initial density of 1 × 106 cells/ml.Cells were washed with 0.9% NaCl to eliminateany remaining nitrogen from the rich media priorto inoculation. The temperature was maintained at28 ◦C and the pH at 3.5. Nitrogen was used tosparge the medium for 30 min (250 ml/min) beforeinoculation to eliminate any oxygen. Agitation waskept at 100 rpm to hold cells in suspension. Carbondioxide production, in addition to nitrogen sparg-ing and weak agitation, ensured anaerobic condi-tions throughout the experiment. Three independentexperiments were carried out to obtain fermentationprofiles.

Analytical techniques

Carbon dioxide evolution in the bioreactors wasdetermined with a Gallus 1000 volumetric flux

Copyright 2005 John Wiley & Sons, Ltd. Yeast 2005; 22: 369–383.

Gene expression ‘in’ wine fermentation 371

transductor (Schlumberg, Buenos Aires, Argen-tina). Culture samples were taken periodically toestablish the fermentation status. These sampleswere analysed to determine the dry cell weight,the cell number, and the concentrations of glu-cose, fructose, organic acids, amino acids, ammo-nia and free amino acid nitrogen. Dry cell weightwas estimated by filtering cells and washing themtwice with distilled water and then drying thepreparation to a constant weight at 85 ◦C. Cellnumbers were estimated microscopically using aNeubauer chamber (Brand, Wertheim, Germany).Glucose and fructose concentrations were mea-sured by high-performance liquid chromatography(HPLC), using a Waters high-performance car-bohydrate cartridge (Varela et al., 2004). Organicacids, glycerol and ethanol concentrations weremeasured by HPLC, using a Bio-Rad HPX-87Hcolumn (Varela et al., 2003). Amino acids werederivatized with Waters AccQ Fluor reagent, andthen the concentrations were measured by HPLC,using an AccQ Tag amino acid analysis columnin accordance with the instructions of the man-ufacturer (Waters, Milford, MA). The ammoniaconcentration was measured enzymatically, usingglutamate dehydrogenase (Sigma, St. Louis, MO).The concentration of free amino acid nitrogenwas determined by using the σ -phthaldehyde/N -acetyl-L-cysteine spectrophotometric assay (NOPA)procedure (Dukes and Butzke, 1998). Viability wasmeasured by using a LIVE/DEAD yeast viabilitykit (Molecular Probes, Eugene, OR) as previouslydescribed (Varela et al., 2004).

mRNA extraction and SAGE procedure

The yeast transcriptome was analysed at threestages of the fermentation process: mid-exponentialphase, and early- and late-stationary phases. TotalRNA was extracted from the samples using Trizolreagent (Life Technologies, Carlsbad, CA) follow-ing the manufacturer’s instructions. Total RNA wasused to obtain the SAGE libraries using the I-SAGE kit (Invitrogen, Carlsbad, CA) as describedby the manufacturer. Briefly, mRNA was convertedinto double-stranded copy DNA (cDNA) using abiotinylated oligo(dT) primer. cDNA was digestedwith NlaIII and 3′ cDNAs fragments were isolated,using streptavidin paramagnetic beads. 3′ cDNAswere split into two pools and SAGE linkers Aand B were ligated to pools 1 and 2, respectively.

SAGE tags were released with BsmF1 and blunt-ended with T4 polymerase. The tags from pools1 and 2 were then ligated to each other. The lig-ation product was amplified with 28 cycles of apolymerase chain reaction (PCR) and digested withNlaIII. Ditags were isolated from a 12% polyacry-lamide gel, concatemerized and cloned into pZerodigested with SphI. Clones containing at least 500bp inserts were sent for sequencing to AgencourtBioscience (Beverly, MA).

SAGE data analysis

SAGE data analysis involved four parts: (a) deter-mination of tag abundance; (b) identification ofeach observed tag (genome mapping); (c) determi-nation of statistically significant differences inexpression levels; and (d) cluster analysis. Fortag abundance determination, analysis of sequenc-ing data was performed using the software pack-age SAGE 2000 (Velculescu et al., 1995; Zhanget al., 1997). For tag identification we generated adatabase with tags from the whole yeast genome(manuscript in preparation). The annotated tagsdatabase contained all the 14 bp sequences in theyeast genome that started with the NlaIII restrictionsite (5′-CATG-3′). All potential tags were linkedto the existing gene annotations from the NCBIdatabase. Then, the observed tags were mergedwith the constructed database and the expressionof different genes was determined. The criterionused to map the tags was 100% identity matcheswithin the gene or up to 500 bases downstream(towards the 3′ end), unless an ORF was annotatedbefore that downstream length was accomplished.When different tags originating from the same genewere found, these tags were added to calculateexpression levels. It is noteworthy that the result-ing database allowed the annotation of 47% moretags than those originally classified as NORFs andNIDs, when employing the tag annotated databaseavailable from S accharomyces Genome Database(SGD) (Cherry et al., 1998), which contains onlyexperimentally obtained sequence tags.

Determination of statistically significant differ-ences was performed using the SAGEstat soft-ware, which allows to compare differences betweentwo libraries (Kal et al., 1999; Ruijter et al.,2002). Hence, the phases: mid-exponential–early-stationary, early–late-stationary and mid-exponen-tial–late-stationary were analysed statistically as



pairs. After this analysis, the total numbers of tagsfrom the three SAGE libraries was normalized to15 000 mRNAs to obtain the number of transcriptsper cell (Kal et al., 1999).

Cluster analysis was performed using Seqexpress1.2.1. This software allows the determination ofthe optimal number of clusters by applying theexpectation maximization algorithm (Boyle, 2004).Transcripts were clustered according to the expres-sion profile and expression level. Therefore, 13clusters were defined for representing upregulation,downregulation and constant expression at severalexpression levels. SAGE data sets are available onthe web (www2.ing.puc.cl/∼sage).

Metabolic fluxes and gene expression

Flux distributions at mid-exponential phase andearly- and late-stationary phases were determinedby metabolic flux analysis and have been reportedpreviously (Varela et al., 2004). This analysisenabled quantification of fluxes through the meta-bolic pathways which describe the anaerobic meta-bolism of S. cerevisiae. Then, metabolic fluxes ofevery biochemical reaction were compared withthe expression levels of the gene involved in thatparticular reaction. Two ‘entry gates’ or groupsof reactions that feed a particular pathway were

defined. The glycolysis entry gate grouped glucoseand fructose transport reactions and the reactionsresponsible for the conversion of both sugarsinto phosphate sugars. The TCA cycle entry gategrouped the reactions involved in the formation ofisocitrate from pyruvate: pyruvate dehydrogenase,citrate synthase and aconitase.

Results

Fermentation profiles

The average time to reach dryness (<4 g sugar/lwine) was 6 days (170 h) (see Figure 1A in Varelaet al., 2004). S. cerevisiae EC1118 consumed glu-cose in preference to fructose, which is a result ofdifferences in the transporters’ affinities for thesesugars (Reifenberger et al., 1997). The biomassgrowth curve showed an exponential phase and astationary phase that began at 48 h, at which pointthe biomass concentration was 5.8 g/l. Ammo-nia was depleted after 24 h of culturing, and thiscoincided with the highest specific growth rate(0.2 h−1). Assimilable nitrogen was depleted fromthe medium at 48 h, which resulted in cell growtharrest. Even though all assimilable nitrogen wasconsumed, the cell viability remained greater than

100

10

1

0.1(2-10) (11-20) (21-40) (41-60) (61-80) (81-100) (101-150) (151-200) (201-500) (>500)

Expression level ranges[Transcripts/cell]

Fra

ctio

n of

tota

l tag

s[%

]

Figure 1. Relative distribution of gene expression levels at different stages of wine fermentation. Mid-exponential phase(white columns), early-stationary phase (grey columns) and late-stationary phase (black columns). Circles representexpression levels at log phase of S. cerevisiae YPH499 grown in YPD medium (Velculescu et al., 1997). Total tags obtained byphase: 4881 for mid-exponential; 3483 for early-stationary; 4327 for late-stationary; and 20 184 for log phase in rich medium



Table 1. Fermentation values at different yeast growthphases

Mid-exponential

phase

Early-stationary

phase

Late-stationary

phase

Time (h) 20 48 96DCW (g/l) 2.0 ± 0.2 5.8 ± 0.1 5.8 ± 0.1Residual sugar (g/l) 204 ± 5.1 100 ± 3.8 16 ± 3.9Yeast assimilablenitrogen (mg N/l)

76 ± 5.9 7 ± 2 7 ± 2

Ethanol (g/l) 9.6 ± 3.3 50.7 ± 3.1 94.2 ± 3.7

Data obtained from three independent experiments.

97% until all the sugar was depleted. Ethanolsynthesis occurred mainly in the stationary phaseand resulted in a final ethanol concentration of12.7% (v/v). Besides ethanol, the yeast produceda number of other products during the fermenta-tion. The final concentration of glycerol, quantita-tively the second most important product of winefermentation, was 7.8 g/l. Other significant com-pounds produced by the yeast were succinic acidand acetic acid, whose final concentrations were1.8 and 1.0 g/l, respectively.

Gene expression profiles were examined at mid-exponential phase, and early- and late-stationaryphases (Table 1). At mid-exponential phase (20 h),sugar concentration was very high (∼200 g/l),whereas ammonia was almost completely con-sumed (∼5 mg/l). At early-stationary phase (48 h),cell growth was arrested, since all assimilablenitrogen was depleted from the medium. At late-stationary phase (96 h), sugar was still present inthe medium (∼16 g/l), whereas ethanol concentra-tion had reached 12% (v/v).

Yeast transcriptome

The total number of tags obtained was 12 691(including 4881 for mid-exponential phase, 3483from early-stationary phase and 4327 from late-stationary phase). Analysis of the relative dis-tribution of gene expression levels showed thatonly 0.4% of the messengers were expressed at100 copies/cell or more, 11.4% at more than 10copies/cell, whereas most transcripts (88.6%) werepresent at 10 copies/cell or less (Figure 1). Thishad already been observed for yeast cells grownin rich medium (Velculescu et al., 1997). For sub-sequent analyses, only tags present at least twice

Table 2. Tag numbers under winemaking conditions

Mid-exponential

phase

Early-stationary

phase

Late-stationary

phase

Total tags 4881 3483 4327Unique tags 614 449 625ORF tags1 449 (73%) 344 (77%) 354 (57%)NORF tags2 62 (10%) 42 (9%) 136 (22%)Non-identified tags3 103 (17%) 63 (14%) 135 (21%)

1 Tags that correspond to a defined ORF in the yeast genome.2 Tags that match intergenic regions within the yeast genome and donot correspond to an annotated ORF.3 Tags that do not match to any location within the yeast genome.

were considered, since those present only once arenot reliable (Table 2).

Upon correlation with the yeast genome, threedifferent kinds of tags were found: (a) tags thatcorresponded to an annotated open reading frame(ORF tags); (b) tags that corresponded to intergenicregions (NORF tags); and (c) tags that did notmatch the published reference yeast genome (non-identified or NID tags). At both mid-exponentialand early-stationary phases, about 10% and 15%of the total tags corresponded to NORF andNID tags, respectively (Table 2). At late-stationaryphase NORF and NID tags increased to 22% of thetotal, i.e. 136 and 135 tags, respectively. Both kindsof tags showed differential expression at early-and late- stationary phases as compared to mid-exponential phase. Thirty-three NORF tags and 10NID tags were shown to be statistically signifi-cantly upregulated by at least five times at late-stationary phase, reaching a maximum of 134 and51 transcripts/cell, respectively. Since these twokinds of tags do not correspond to any knownORF, they are not available in the existing DNAmicroarrays and therefore have not been reported tobe expressed under winemaking conditions. Hence,NORF and NID tags could represent potentialnovel genes that are required to be expressed duringwine fermentation.

In the original annotation of yeast genes, ORFsencoding proteins with less than 100 amino acidswere omitted from the annotation unless evidencefor the presence of a gene had been found byother direct means (Goffeau et al., 1996). Cur-rently, only 224 known genes (3.5% of the yeastgenome) encode these small proteins (Oshiro et al.,2002). Since, in other sequenced organisms, genes



encoding for small proteins correspond up to 10%of their genomes, Oshiro et al. (2002) suggestedthat there might be an additional 400 genes encod-ing small proteins in the yeast genome. SAGEallows detection and identification of expressedsequences originating from any part of the genomethat makes this method very useful for identifyingnew genes (Saha et al., 2002).

In this study, we found 197 different NORFtags expressed during wine fermentation. Although72 tags have been reported in other yeast studiesusing SAGE (Kal et al., 1999; Oshiro et al., 2002;Velculescu et al., 1997), 70% of the NORF tagsobtained here, especially at late-stationary phase,have not been described previously. Similarly, we

found 245 different NID tags expressed throughoutthe fermentation.

Grouping genes according to expression profilesand expression levels, we defined 13 distinct clus-ters; however, more than 99% of the observed tagswere grouped by the first nine clusters (Figure 2).Tags upregulated at late-stationary phase weregrouped into the first three clusters, represent-ing 24.2% of total tags. These clusters exhib-ited different expression levels. Thus, cluster 1showed an expression increase from medium level(20–60 transcripts/cell) to high level (60–100 tran-script/cell) and cluster 2 an increase from low level(0–20 transcripts/cell) to medium level. Cluster3 grouped low expressed tags that increased at

Figure 2. Cluster analysis. The nine most representative clusters. Clusters were defined according to expression profile andexpression levels of observed tags. (A) Clusters 1–3 grouped tags upregulated at late-stationary phase (96 h). (B) Clusters4–6 represented tags downregulated at late-stationary phase. (C) Clusters 7 and 8 grouped tags expressed at constantlevels, whereas cluster 9 grouped tags upregulated at early-stationary phase (48 h)



late-stationary phase (Figure 2A). Tags downregu-lated at late-stationary phase were grouped inclusters 4–6, accounting for 8.3% of total tags(Figure 2B). In cluster 4, expression levels decrea-sed from high to low levels, whereas in cluster5 expression was reduced from medium to lowlevels. Cluster 6 grouped low expressed tags thatdecreased at late-stationary phase. Tags expressedat constant levels throughout the fermentationwere grouped in clusters 7 and 8. Cluster 7grouped tags expressed at medium level, whilstcluster 8 grouped those tags expressed at lowlevel. These two clusters represented 61.1% ofthe total tags (Figure 2C). Tags expressed at lowlevel at mid-exponential and late-stationary phasesbut at higher level at early-stationary phase weregrouped in cluster 9, representing 6.1% of total tags(Figure 2C). Tags upregulated and downregulatedat late-stationary phase showing other expressionlevels were grouped into clusters 10–13. Theseclusters represented 0.7% of the observed tags andeach cluster grouped less than four tags (data notshown). The clustered patterns observed here couldbe used to find common promoters that regulategene expression at different stages of fermentationand even at different levels. Therefore, it could bepossible to ‘switch on’ and ‘switch off’ exogenous(homologous and heterologous) genes at specificstages during wine fermentation.

As the target genes of most yeast transcriptionalregulators have been established (Lee et al., 2002),we used this information to determine whethergenes grouped in the same cluster are regulatedby the same transcription factors. In clusters 1

and 2, 21% of the genes had target promoterregions for the stress response transcription factorsHsf1p and Skn7p, whereas in cluster 3 thesegenes accounted for 7%. The transcription factorsRap1p, Fhl1p and Gcr1p, which bind to regulatoryregions of genes encoding ribosomal proteins andglycolytic enzymes, were associated to 26% of thegenes grouped in clusters 4 and 5. In cluster 6,12% of the genes were related to Fhl1p. Thus, asignificant proportion of co-regulated genes in bothup- and downregulated clusters were associatedwith the same transcription factor. Other clusterswere related to several transcriptional regulatorsand the relevance of specific promoter regions wasless evident.

At mid-exponential and early-stationary phases,the most highly expressed genes corresponded tothe glycolytic genes: aldolase, 3-phosphoglyceratekinase and glyceraldehyde 3-phosphate dehydro-genase I. At late-stationary phase, however, themost highly expressed genes corresponded to genesencoding proteins related to stress response, suchas heat shock proteins, metallothioneins regulatedby heat shock elements (Liu and Thiele, 1996) andchaperone proteins (Table 3). Among the top 30most highly expressed genes, a NORF tag, whichcorresponds to an expressed sequence tag (EST)from chromosome 12, and a NID tag, were foundto be expressed at greater than 50 transcripts/cellat late-stationary phase. Although the function ofthese putative genes remains to be clarified, theirexpression at the end of the wine fermentationcould be relevant.

Table 3. Most highly expressed genes at late-stationary phase

GeneExponential

(transcripts/cell)Early-stationary(transcripts/cell)

Late-stationary(transcripts/cell) Description

HSP26 128 98 512†∗ Heat shock protein 26HSP82 33 30 294†∗ Heat shock protein 82TEF1/2 61 59 199†∗ Translational elongationPGK1 224 339§ 186∗ 3-Phosphoglycerate kinaseFBA1 506 610 162†∗ AldolaseCUP1-1/1-2 82 77 161†∗ Copper ion bindingHSP30 18 42 134†∗ Heat shock protein 30NORF1 24 25 134†∗ EST from chromosome 12TIP1 64 25§ 131†∗ Cold and heat shock proteinSEC3 6 <4 121†∗ Cellular transport

1 Tag that matches an intergenic region within the yeast genome and does not correspond to an annotated ORF.§†∗ Statistically significant difference in expression levels (p < 0.05): § comparing mid-exponential and early- stationary phases;† comparing early- and late-stationary phases; ∗ comparing mid-exponential and late-stationary phases.



Yeast metabolism

Many biosynthetic pathways were downregulatedas fermentation progressed, confirming the resultsof Rossignol et al. (2003). Genes related to cellwall organization and biogenesis, protein biosyn-thesis, RNA processing and protein metabolismshowed moderate to low expression levels (below60 transcripts/cell) at the beginning of the fermen-tation. The expression of these genes decreasedthroughout the fermentation. The downregulationof biosynthetic genes is thought to allow the cellto save energy and to adapt to new conditions(Gasch et al., 2000). Genes related to nucleotidemetabolism showed constant expression levels andwere classified into cluster 8.

Sugars

Many glycolytic genes were downregulated at theend of fermentation and therefore classified intoclusters 4–6 (Table 4). Other genes (HXK1 andPGI1 ) showed the same expression level through-out the fermentation and were classified in clus-ter 7. Genes encoding glycolytic enzymes exhib-ited differences in expression levels of up to 150times at the same phase. Similarly, Velculescuet al. (1997) reported differences of up to 70 timesfor glycolysis-related transcripts for glucose-grownyeast cells. The phosphofructokinase gene wasthe lowest expressed gene in glycolysis, <4 tran-scripts/cell, indicating that expression of this geneis tightly regulated. Surprisingly, FBA1, whichencodes the following enzymatic reaction in gly-colysis, was the most expressed gene under theconditions used in this study (Table 4). Genes

encoding enzymes of the TCA cycle, except forACO1 (aconitase), MDH1 (malate dehydrogenase)and SDH4 (succinate dehydrogenase), showedvery low expression levels throughout fermenta-tion. These three genes were expressed at 30–40transcripts/cell at early-stationary phase and thendownregulated at late-stationary phase.

Expression levels of hexose transporter genes(HXT ) were determined during wine fermentation(Table 5). As observed for genes involved in gly-colysis, expression levels of HXT genes decreasedat late-stationary phase, suggesting a possible co-regulation of glycolysis and sugar transport. HXT2showed low expression levels throughout fermen-tation. HXT3 was the most highly expressedHXT gene, showing a pattern consistent with ear-lier results (Luyten et al., 2002). Unexpectedly,HXT6/7, described as high-affinity carriers (Reifen-berger et al., 1997), were expressed throughout fer-mentation. However, expression of HXT6, clonedfrom an oenological strain into a hxt1-7� mutantstrain, could not restore growth on a mediumwith low glucose concentration, whereas HXT7-only cells were shown to grow only weakly onthe same medium. In contrast, expression of HXT6or HXT7 from a laboratory strain in the hxt1-7�

mutant enabled the yeast to grow well on low glu-cose (Luyten et al., 2002). This result illustratesthat, despite high sequence homology, hexose car-riers from oenological and laboratory yeast strainshave dissimilar behaviours and could play differentroles during wine fermentation.

In addition to the HXT genes, TDH1 and TDH2encoding two glyceraldehyde-3-phosphate dehy-drogenase isoenzymes were found to be expressed

Table 4. Expression of glycolytic genes

GeneExponential



HXK1 24 12 13 HexokinasePGI1 <4 12 13 PhosphoglucoisomerasePFK2 <4 <4 <4 PhosphofructokinaseFBA1 506 610 162†∗ AldolaseTDH1 98 163§ 103 Glyceraldehyde 3-phosphate dehydrogenase ITDH2 109 46§ <4†∗ Glyceraldehyde 3-phosphate dehydrogenase IIPGK1 224 339§ 186† Phosphoglycerate kinaseGPM1 49 64 17†∗ Phosphoglycerate mutaseENO2 46 30 <4†∗ EnolaseCDC19 12 47§ 34 Pyruvate kinase

§†∗ Statistically significant difference in expression levels (p < 0.05): § comparing mid-exponential and early-stationary phases; † comparing early-and late-stationary phases; ∗ comparing mid-exponential and late-stationary phases.



Table 5. Expression of HXT genes

Gene

Exponentialphase (tran-scripts/cell)

Early-stationary phase(transcripts/cell)

Late-station-ary phase

(transcripts/cell)

HXT2 <4 8 <4HXT3 30 21 <4∗HXT6/7 18 8 <4∗

§†∗ Statistically significant difference in expression levels (p < 0.05):§ comparing mid-exponential and early-stationary phases; † comparingearly- and late-stationary phases; ∗ comparing mid-exponential andlate-stationary phases.

at different levels (Table 4). Similarly, isoenzymesof glycerol 3-phosphate dehydrogenase (encodedby GPD1 and GPD2 ), pyruvate decarboxylase(encoded by PDC1 and PDC5 ), alcohol dehydro-genase (encoded by ADH1/2 and ADH6 ) and alde-hyde dehydrogenase (encoded by ALD2/3, ALD4and ALD6 ) were also differentially expressed,depending on growth phase (data not shown).

Nitrogen and amino acids

GAP1, which encodes the general amino acid per-mease, showed a constant expression level through-out the fermentation. Genes encoding amino acidvacuolar transporters (AVT3, AVT4 and AVT6 )were mainly expressed during mid-exponentialphase and then downregulated. This expression pat-tern is consistent with the amount of assimilablenitrogen present in the medium at each growthphase. PUT1, which encodes proline oxidase, wasexpressed at high levels (about 89 transcripts/cell)during the first two phases and downregulated atlate-stationary phase. However, proline was notremoved from the must at any growth phase.Although proline is able to induce expression ofPUT1 in anaerobic conditions, oxygen is essentialto allow proline degradation (Wang and Brandriss,1987). PUT2, encoding �1-pyrroline 5-carboxylatedehydrogenase, the next step of the proline uti-lization pathway, increased significantly at early-stationary phase but was undetectable later on.

Expression of many genes related to amino acidbiosynthesis, including the transcriptional activatorGCN4, was downregulated at late-stationary phase(cluster 5). However, GDH2 and GDH3, whichencode two glutamate dehydrogenase isoenzymes,HOM3 encoding aspartate kinase and MET30 aregulator of amino acid metabolism, were upregu-lated at late-stationary phase.

Trehalose and glycogen

Genes related to trehalose biosynthesis did notchange significantly, whereas genes involved intrehalose degradation (ATH1 and NTH2 ) exhib-ited different expression patterns. ATH1 encodesan acid trehalase, whereas NTH2 encodes aneutral trehalase. Expression of ATH1 increasedthroughout fermentation, whilst NTH2 expressionincreased in the early-stationary phase and thendecreased. We also observed the downregulation atthe end of the wine fermentation of GLC3, involvedin glycogen biosynthesis, whereas GPH1, involvedin glycogen degradation, was expressed at a lowand constant level throughout fermentation.

Sterols

The yeast cell cannot produce ergosterol in theabsence of oxygen. However, many genes encodingproteins involved in ergosterol biosynthesis wereexpressed at mid-exponential and early-stationaryphases and downregulated at the end of fermenta-tion. A similar expression pattern was observed byRossignol et al. (2003).

Electron transport chain

Several genes related to the electron transport chain(QRC, COX, CYC and CYT genes), inactive dur-ing wine fermentation, were also expressed. Wealso found a differential expression of COX6 andCOX13, encoding subunit VI and VIa, respec-tively, of cytochrome c oxidase. Whereas COX13was downregulated at stationary phase (half of theexpression found at mid-exponential phase), COX6was upregulated at late-stationary phase. Althoughthe expression of these genes under anoxic con-ditions could seem contradictory, genes encodingdifferent subunits of cytochrome c oxidase (COX1,COX2, COX4, COX5A, COX6, COX7, COX8 andCOX9 ) were shown to be expressed in the absenceof oxygen (Dagsgaard et al., 2001). Moreover,most cytochrome c oxidase subunits, including sub-units I and II, which comprise the catalytic core,are present in anoxic cells and they are localizedin the promitochondria (Dagsgaard et al., 2001). Asno cytochrome c oxidase activity has been foundin anaerobiosis (Rogers and Stewart, 1973), thefunction of this protein complex, if any, is stillunknown.



Stress response

All environmental changes that trigger an adaptiveresponse are denominated as stress. Genes regu-lated by stress or that respond in a similar way tomany stresses have been referred to as environmen-tal stress response (ESR) or common environmen-tal response (CER) genes, respectively (Rossignolet al., 2003). The proteins encoded by the genesinduced by stress are involved in many cellu-lar processes. Stress-responsive heat shock genes(HSP26, HSP42, HSP60, HSP78, HSP82, SSE2 ),which encode proteins related to protein folding,were strongly induced (Table 6). Although thesegenes showed different expression levels, all wereinduced at late-stationary phase. HSP26 was foundto be the most highly expressed gene of the genomeat late-stationary phase (Table 3). HSP104 andHSP12, the latter previously described as glucose-repressed in a laboratory strain of S. cerevisiae (deGroot et al., 2000), were expressed at constant lev-els throughout the fermentation (Table 6).

Genes encoding proteins involved in protonhomeostasis, HSP30 and PMP2 (coding for a neg-ative and a positive regulator of the plasma mem-brane H+-ATPase, respectively) were upregulated

at the end of the fermentation (Table 6). Simultane-ous induction of positive and negative regulators ofplasma membrane H+-ATPase has been describedas consistent with the fine-tuning of H+ extru-sion activity (Alexandre et al., 2001). The plasmamembrane H+-ATPase, encoded by PMA1/2, wasexpressed from the beginning of the fermentationat similar levels (Table 6).

Genes which encode proteins with unknownmolecular functions and that respond to stress,according to SGD (NCE102, SPI1 ), were alsoupregulated at late-stationary phase (Table 6).Some genes related to stress response (PST1,PWP1, SIW14, SSC1, SSE1, TIR1, YRO2 ) showedsimilar expression levels throughout the fermenta-tion (Table 6). Other genes, e.g. YGP1, HSP31 andHSP32/33, were downregulated at late-stationaryphase, indicating a response to an early stress nolonger present at the end of fermentation.

Correlation between metabolic fluxes and geneexpression

In a recent study (Varela et al., 2004), we reportedthe redistribution of metabolic fluxes in S. cere-visiae EC1118 during wine fermentation, using

Table 6. Expression of genes related to stress response

GeneExponential



HSP12 9 <4 6 Protein foldingHSP26 128 98 512†∗ Protein foldingHSP30 18 42 134†∗ H+ homeostasisHSP42 6 8 34†∗ Protein foldingHSP60 9 4 55†∗ Protein foldingHSP78 <4 <4 31†∗ Protein foldingHSP82 33 30 294†∗ Protein foldingHSP104 <4 17 <4 Protein foldingNCE102 12 21 31∗ Function unknownPMA1/2 21 25 13 Membrane ATPasesPMP2 33 42 83†∗ H+ homeostasis (putative)PST1 <4 8 <4 Function unknownPWP1 <4 4 10 Function unknownSIW14 9 12 <4 Tyrosine phosphatase activitySPI1 18 8 91†∗ Function unknownSSC1 24 17 16 Protein foldingSSE1 6 8 20 Protein foldingSSE2 <4 4 24†∗ Protein folding (putative)TIR1 18 <4 <4 Constituent of cell wallYGP1 30 30 6†∗ Function unknownYRO2 24 30 10 Function unknown

§†∗ Statistically significant difference in expression levels (p < 0.05): § comparing mid-exponential and early-stationary phases;† comparing early- and late-stationary phases; ∗ comparing mid-exponential and late-stationary phases.



120

100

80

80

60

60

40

40

20

200

0

50

40

30

20

10

00.0 0.2 0.4 0.6

Flux [Cmmol gDCW-1 h-1]

Flux [Cmmol gDCW-1 h-1]

R2 = 0.848

R2 = 0.928

Correlation for glycolysis gate point

Correlation for TCA cycle gate point

Exp

ress

ion

leve

l[tr

ansc

ripts

/cel

l]E

xpre

ssio

n le

vel

[tran

scrip

ts/c

ell]

GLC G6P

F6PFRUC

CARBOHYDRATES

CO2

CO2 CO2

CO2

CO2

CO2

DNA

RNA

LIPIDS

R5P

E4PDHAP

GLY3P

GLYCSER

G3P

3PG

PEP

OAA

ASP

GLN

GLU

NH4+

NH4+

AKG

SUC

SUCEXPROTEINS

FUM

ISOCITOAA

PYR

ACCOAAC

ADE ETOH

CYTOSOL

MITOCHONDRIA

Figure 3. Correlation between gene expression and metabolic fluxes at different yeast growth phases. Entry gatesto glycolysis (black squares) and the TCA cycle (grey squares). The stoichiometric model included glycolysis, pentosephosphate pathway, the TCA cycle, amino acid metabolism and macromolecule biosynthesis pathways

the same experimental conditions described here.Although the expression of some glycolytic genesincreased, we found that metabolic fluxes through-out glycolysis diminished. Genes involved in theTCA cycle were slightly upregulated, while fluxeswithin the cycle decreased. Entry gates to gly-colysis, considering sugar transport and the hex-okinase reaction, and to the TCA cycle, consid-ering the reactions from pyruvate to isocitrate,showed the best correlations between gene expres-sion and metabolic fluxes (r2 = 0.928 and r2 =0.848, respectively) (Figure 3).

Discussion

Stress conditions throughout the wine fermentationprocess, such as nutrient limitation, starvation, tem-perature variations and ethanol toxicity, affect theyeast’s metabolism, eliciting a metabolic response.

This response, which enables the cell to adapt to thenew environmental conditions, implies a metabolicreorganization in which a great number of genesare up- or downregulated. In this work, SAGEwas used to quantify gene expression profiles inthe commercial strain S. cerevisiae EC1118 underwinemaking conditions. We detected the expres-sion of genes that have not been reported in otherstudies using SAGE (Kal et al., 1999; Velculescuet al., 1997), suggesting that winemaking condi-tions impose a substantially different environmentfor yeast growth.

Since the total number of tags obtained inthis study was not high enough, we were unableto detect very low-abundance mRNA molecules(<2 copies/cell). The total number of tags alsoaffects the confidence level for detecting low-abundance transcripts. However, analysis of the rel-ative distribution of gene expression levels revealedthat low-abundance mRNA classes were highly



similar to those described previously using agreater number of tags obtained from a differentyeast strain (Velculescu et al., 1997). Therefore, weconsidered the data of low-abundance transcriptsreliable.

SAGE enabled the identification of novel tran-scripts (NORFs and NIDs) within the yeast tran-scriptome. Since we compared the transcriptomeof a wine yeast with the published reference yeastgenome, obtained from a laboratory strain of S.cerevisiae, several factors could explain the pres-ence of non-identified tags, e.g. differences in thesequence of known genes between both yeastsand novel genes present only in the wine yeast.Indeed, a high frequency of polymorphisms hasbeen described in S. cerevisiae (Oefner, 2002). Fur-ther research is necessary to assess the presence ofnew genes in S. cerevisiae EC1118.

Many of these novel transcripts were expressedat high levels and differentially regulated depend-ing on the growth phase. Particularly at late-stationary phase, where ethanol concentration was12% (v/v) and nitrogen had been already depletedfor 2 days, we observed a high increase in the num-ber of new transcripts. This harsh environment,very different to situations normally studied underlaboratory conditions, probably triggers the expres-sion of genes that encode critical functions for theyeast cell and that had not been discovered untilnow. Oshiro et al. (2002) investigated the expres-sion and translation of several NORFs previouslyidentified by SAGE under laboratory conditions.After growing yeasts in different growth condi-tions, which included treatments with hydroxyurea,nocodazole, methyl methane sulphonate, ultravioletlight, along with heat and cold shocks, they pro-posed the addition of 62 new genes to the yeastgenome. This illustrates that in order to increasethe chance of observing expression of new genes,unusual growth conditions are often necessary.

Our results have indicated that ‘new’ genes couldhave a significant role under winemaking condi-tions. Although expression patterns do not presentconclusive evidence of gene function, they indi-cate directions for further research. Indeed, a strongfunctional connection between transcription factorsand metabolic pathways has been described, sug-gesting that uncharacterized ORFs can be assigneda functional link based on co-expression with spe-cific metabolic pathways (Ihmels et al., 2004).Characterization of these novel genes could lead to

further understanding of the yeast physiology andthe wine fermentation process.

Many genes involved in pathways and reactionsthat require oxygen for their function, such as theERG, PUT, COX and OLE genes, were expressedin absence of oxygen. Post-transcriptional mecha-nisms are probably responsible for the regulation ofthe corresponding proteins, which would enable thecell to rapidly respond when oxygen is again avail-able. An interesting case is the differential geneexpression of subunits VI and VIa of cytochromec oxidase. Similarly to subunits Va and Vb, whichchange cytochrome c oxidase activity in responseto oxygen availability (Kwast et al., 1998), subunitVIa affects the activity of cytochrome c oxidasein response to ionic strength and ATP concentra-tion (Beauvoit et al., 1999; Taanman and Capaldi,1993). As there is no activity of cytochrome c oxi-dase under anoxic conditions (Rogers and Stewart,1973), differential regulation of cytochrome c oxi-dase subunits with different properties by the yeastcell remains to be studied.

Within the yeast transcriptome, we found differ-ent transcript levels for genes involved in the samemetabolic pathway (glycolysis, TCA cycle, ergos-terol biosynthesis, pentose phosphate pathway, etc.)at the same growth phase. In glycolysis, differencesin transcript levels were up to 150 times at early-stationary phase between two genes, which codefor sequential steps within the pathway. In addition,protein levels and protein activities for glycolyticgenes have been also shown to differ greatly atthe same growth phase (Fraenkel, 2003). However,despite differences in transcript and protein levels,carbon flux through glycolysis is highly coordi-nated under winemaking conditions, as recentlyshowed by Varela et al. (2004). Hence, these differ-ences show that, in order to present a fully func-tional metabolic pathway, the yeast cell requiresdifferent mRNA levels of the genes involved inthat pathway. Probably, mRNA stability or post-transcriptional mechanisms are responsible for dis-similar transcript levels in the same pathway.

New genes introduced and overexpressed in S.cerevisiae are normally under the control of theconstitutive promoter pPGK1 (Lilly et al., 2000;Verstrepen et al., 2003). However, expression ofthe FBA1 gene is two-fold higher than the PGK1gene during the first two stages of fermentation(Table 4). Although this difference could be strain-dependent, the FBA1 promoter might be used



to overexpress genes at greater levels and thusmake the phenotype of the new strain stronger. Inaddition, the use of promoters specific for differentgrowth stages could be advantageous to obtain adifferential expression of new genes in wine yeasts.

Many isoenzymes involved in pathways of thecentral metabolism were expressed differentially,depending on the growth phase. In addition to dif-ferent physiological roles, the presence of isoen-zymes in the S. cerevisiae genome probably enablesthe yeast cell both to manage efficiently the poolof co-factors, such as NADH and NADPH, and tocontrol the gene expression by different physiolog-ical triggers rather than different pathways in orderto respond competently to new environmental con-ditions (Ansell et al., 1997). Alternatively, isoen-zymes could be dedicated to distinct processes,using a common reaction, and thus reduce cross-talk and unwanted interactions between separatemetabolic pathways (Ihmels et al., 2004).

We observed different patterns of expression forgenes related to stress response, which illustratesthe temporal impact of the various stresses on theyeast cell during wine fermentation. Immediatelyafter inoculation, yeast faces hyperosmotic con-ditions that trigger a fast and transient osmoticstress response (Perez-Torrado et al., 2002). Asthe first sample in our experiments was probablytaken after yeast adaptation, we did not observeany response to the high osmotic pressure presentin the must. Although other individual responsesare more difficult to evaluate because many genesare expressed in response to different environ-mental stimuli (Bauer and Pretorius, 2000), ourresults show that several stress-related genes areexpressed from mid-exponential phase onwards.Some of these genes, many of them with unknownfunctions, were upregulated by growth arrest, butthe highest transcript levels were found at late-stationary phase, indicating that yeasts are sub-jected to extreme conditions during this phase.As mentioned before, at late-stationary phase, weobserved the highest number of NORFs and NIDs,suggesting that wine fermentation conditions couldbe appropriate for studying the function of bothgenes with unknown functions and novel genes.Probably the high amount of ethanol present inthe medium at this phase is responsible for thesephysiological changes. Although more research isnecessary, the different responses of genes relatedto stress response could be used as a gene reporter

system to assess the presence of early stresses thatcould affect a ‘normal’ fermentation profile andimpair the quality of the final product. Indeed, thereis an inverse correlation between stress resistanceand stuck fermentations (Ivorra et al., 1999).

Comparing metabolic fluxes and gene expres-sion levels, many discrepancies were found, espe-cially for genes involved in glycolysis and in theTCA cycle. Such divergences have already beenreported in other integrative approaches with S.cerevisiae and other microorganisms (Bro et al.,2003; Oh and Liao, 2000). Nevertheless, entrygates to glycolysis and to the TCA cycle showedthe best correlations between metabolic fluxes andtranscript levels. Detailed analysis of co-regulationin central metabolic pathways has shown that theco-expressed enzymes are often arranged in a lin-ear order, corresponding to a metabolic flow thatis directed in a particular direction (Ihmels et al.,2004). Although fluxes through central metabolicpathways in S. cerevisiae are not primarily con-trolled at the transcriptional level (Daran-Lapujadeet al., 2004), pathway entry gates probably con-trol carbon flux throughout a specific pathwayand could be regulated at both transcriptional andmetabolic levels. This would in turn enable theyeast cell to optimize the flux and thus maximizemetabolic efficiency under different conditions(Stelling et al., 2002). A good correlation betweenmetabolic fluxes and gene expression was alsofound for pathway entry gates in Corynebacteriumglutamicum (Kromer et al., 2004), suggesting thatthis regulation mechanism is probably commonto microorganisms. Further research is necessaryto evaluate the contribution of post-transcriptionalmechanisms, post-translational modification, andregulation by intracellular concentrations of sub-strates, products, and effectors to the regulation offluxes of central metabolic pathways.

Acknowledgements

We would like to thank Professor Sakkie Pretorius andDr Florian Bauer for their support in early stages of thiswork, Dr Jan Ruijter for kindly providing the SAGEstatsoftware, and Ana Maria Molina and Alicia Zuniga fortheir contribution to this work. This work was supported byFondo Nacional para el Desarollo Cientıfico y Tecnologicode Chile (FONDECYT), Grants 2010087 and 1010959.Cristian Varela was supported by a doctoral fellowship fromConsejo Nacional de Investigacion Cientıfica y Tecnologicade Chile (CONICYT).



References

Alexandre H, Ansanay-Galeote V, Dequin S, Blondin B. 2001.Global gene expression during short-term ethanol stress inSaccharomyces cerevisiae. FEBS Lett 498: 98–103.

Ansell R, Granath K, Hohmann S, Thevelein J, Adler L. 1997.The two isoenzymes for yeast NAD+-dependent glycerol 3-phosphate dehydrogenase encoded by GPD1 and GPD2 havedistinct roles in osmoadaptation and redox regulation. EMBO J16: 2179–2187.

Backhus L, DeRisi J, Brown P, Bisson L. 2001. Functionalgenomic analysis of a commercial wine strain of Saccharomycescerevisiae under differing nitrogen conditions. FEMS Yeast Res1: 111–125.

Bauer F, Pretorius I. 2000. Yeast stress response and fermentationefficiency: how to survive the making of wine — a review. SAfr J Enol Viticult 21: 27–51.

Beauvoit B, Bunoust O, Guerin B, Rigoulet M. 1999. ATP-regulation of cytochrome oxidase in yeast mitochondria — roleof subunit VIa. Eur J Biochem 263: 118–127.

Blomberg A, Zuelke K. 2004. Serial analysis of gene expression(SAGE) during porcine embryo development. Reprod Fertil Dev16: 87–92.

Boyle J. 2004. SeqExpress: desktop analysis and visualizationtool for gene expression experiments. Bioinformatics 20:1649–1650.

Bro C, Regenberg B, Lagniel G, et al. 2003. Transcriptional,proteomic, and metabolic responses to lithium in galactose-grown yeast cells. J Biol Chem 278: 32 141–32 149.

Chen H, Centola M, Altschul S, Metzger H. 1998. Characteriza-tion of gene expression in resting and activated mast cells. J ExpMed 188: 1657–1668.

Chen J, Rowley J, Wang S. 2000. Generation of longer cDNAfragments from serial analysis of gene expression tags for geneidentification. Proc Natl Acad Sci USA 97: 349–353.

Cherry J, Adler C, Ball C, et al. 1998. SGD: SaccharomycesGenome Database. Nucleic Acids Res 26: 73–79.

Dagsgaard C, Taylor L, O’Brien K, Poyton R. 2001. Effects ofanoxia and the mitochondrion on expression of aerobic nuclearCOX genes in yeast: evidence for a signaling pathway fromthe mitochondrial genome to the nucleus. J Biol Chem 276:7593–7601.

Daran-Lapujade P, Jansen M, Daran J, et al. 2004. Role oftranscriptional regulation in controlling fluxes in central carbonmetabolism of Saccharomyces cerevisiae — a chemostat culturestudy. J Biol Chem 279: 9125–9138.

de Groot E, Bebelman J, Mager W, Planta R. 2000. Very lowamounts of glucose cause repression of the stress-responsivegene HSP12 in Saccharomyces cerevisiae. Microbiology 146:367–375.

Dukes B, Butzke C. 1998. Rapid determination of primary aminoacids in grape juice using an o-phthaldialdehyde/N -acetyl-L-cysteine spectrophotometric assay. Am J Enol Viticult 49:125–134.

Erasmus D, van der Merwe G, van Vuuren H. 2003. Genome-wideexpression analyses: metabolic adaptation of Saccharomycescerevisiae to high sugar stress. FEMS Yeast Res 3: 375–399.

Fraenkel D. 2003. The top genes: on the distance from transcript tofunction in yeast glycolysis. Curr Opin Microbiol 6: 198–201.

Gasch A, Spellman P, Kao C, et al. 2000. Genomic expressionprograms in the response of yeast cells to environmentalchanges. Mol Biol Cell 11: 4241–4257.

Goffeau A, Barrell B, Bussey H, et al. 1996. Life with 6000 genes.Science 274: 546, 563–567.

Hauser N, Fellenberg K, Gil R, et al. 2001. Whole genomeanalysis of a wine yeast strain. Comp Funct Genom 2: 69–79.

Ihmels J, Levy R, Barkai N. 2004. Principles of transcriptionalcontrol in the metabolic network of Saccharomyces cerevisiae.Nature Biotechnol 22: 86–92.

Ivorra C, Perez-Ortin J, del Olmo M. 1999. An inverse correlationbetween stress resistance and stuck fermentations in wine yeasts.A molecular study. Biotechnol Bioeng 64: 698–708.

Kal A, van Zonneveld A, Benes V, et al. 1999. Dynamics of geneexpression revealed by comparison of serial analysis of geneexpression transcript profiles from yeast grown on two differentcarbon sources. Mol Biol Cell 10: 1859–1872.

Kromer J, Sorgenfrei O, Klopprogge K, Heinzle E, Wittmann C.2004. In-depth profiling of lysine-producing Corynebacteriumglutamicum by combined analysis of the transcriptome,metabolome, and fluxome. J Bacteriol 186: 1769–1784.

Kwast K, Burke P, Poyton R. 1998. Oxygen sensing and thetranscriptional regulation of oxygen-responsive genes in yeast.J Exp Biol 201: 1177–1195.

Lee T, Rinaldi N, Robert F, et al. 2002. Transcriptional regulatorynetworks in Saccharomyces cerevisiae. Science 298: 799–804.

Lilly M, Lambrechts M, Pretorius I. 2000. Effect of increasedyeast alcohol acetyltransferase activity on flavor profiles of wineand distillates. Appl Environ Microbiol 66: 744–753.

Liu P, Thiele D. 1996. Oxidative stress induced heat shockfactor phosphorylation and HSF-dependent activation of yeastmetallothionein gene transcription. Genes Dev 10: 592–596.

Luyten K, Riou C, Blondin B. 2002. The hexose transporters ofSaccharomyces cerevisiae play different roles during enologicalfermentation. Yeast 19: 713–726.

Marks V, van der Merwe G, van Vuuren H. 2003. Transcriptionalprofiling of wine yeast in fermenting grape juice: regulatoryeffect of diammonium phosphate. FEMS Yeast Res 3: 269–287.

Norman B, Davis J, Piatigorsky J. 2004. Postnatal gene expressionin the normal mouse cornea by SAGE. Invest Ophthalmol VisSci 45: 429–440.

Oefner P. 2002. Sequence variation and the biological functionof genes: methodological and biological considerations.J Chromatogr B Anal Technol Biomed Life Sci 782: 3–25.

Oh M, Liao J. 2000. Gene expression profiling by DNAmicroarrays and metabolic fluxes in Escherichia coli . BiotechnolProg 16: 278–286.

Oshiro G, Wodicka L, Washburn M, et al. 2002. Parallel identifi-cation of new genes in Saccharomyces cerevisiae. Genome Res12: 1210–1220.

Perez-Ortin J, Garcia-Martinez J, Alberola T. 2002. DNA chipsfor yeast biotechnology. The case of wine yeasts. J Biotechnol98: 227–241.

Perez-Torrado R, Carrasco P, Aranda A, et al. 2002. Study of thefirst hours of microvinification by the use of osmotic stress-response genes as probes. Syst Appl Microbiol 25: 153–161.

Reifenberger E, Boles E, Ciriacy M. 1997. Kinetic characteriza-tion of individual hexose transporters of Saccharomyces cere-visiae and their relation to the triggering mechanisms of glucoserepression. Eur J Biochem 245: 324–333.



Rogers P, Stewart P. 1973. Respiratory development in Saccha-romyces cerevisiae grown at controlled oxygen tension. J Bac-teriol 115: 88–97.

Rossignol T, Dulau L, Julien A, Blondin B. 2003. Genome-widemonitoring of wine yeast gene expression during alcoholicfermentation. Yeast 20: 1369–1385.

Ruijter J, Van Kampen A, Baas F. 2002. Statistical evaluation ofSAGE libraries: consequences for experimental design. PhysiolGenom 11: 37–44.

Saha S, Sparks A, Rago C, et al. 2002. Using the transcriptome toannotate the genome. Nature Biotechnol 20: 508–512.

Salmon J, Barre P. 1998. Improvement of nitrogen assimilationand fermentation kinetics under enological conditions byderepression of alternative nitrogen-assimilatory pathways inan industrial Saccharomyces cerevisiae strain. Appl EnvironMicrobiol 64: 3831–3837.

Stelling J, Klamt S, Bettenbrock K, Schuster S, Gilles E. 2002.Metabolic network structure determines key aspects offunctionality and regulation. Nature 420: 190–193.

Stollberg J, Urschitz J, Urban Z, Boyd C. 2000. A quantitativeevaluation of SAGE. Genome Res 10: 1241–1248.

Taanman J, Capaldi R. 1993. Subunit VIa of yeast cytochrome coxidase is not necessary for assembly of the enzyme complexbut modulates the enzyme activity. Isolation and characterizationof the nuclear-coded gene. J Biol Chem 268: 18 754–18 761.

Varela C, Agosin E, Baez M, Klapa M, Stephanopoulos G. 2003.Metabolic flux redistribution in Corynebacterium glutamicumin response to osmotic stress. Appl Microbiol Biotechnol 60:547–555.

Varela C, Pizarro F, Agosin E. 2004. Biomass content governsfermentation rate in nitrogen-deficient wine musts. Appl EnvironMicrobiol 70: 3392–3400.

Velculescu V, Zhang L, Vogelstein B, Kinzler K. 1995. Serialanalysis of gene expression. Science 270: 484–487.

Velculescu V, Zhang L, Zhou W, et al. 1997. Characterization ofthe yeast transcriptome. Cell 88: 243–251.

Verstrepen KJ, Van Laere SD, Vanderhaegen BM, et al. 2003.Expression levels of the yeast alcohol acetyltransferase genesATF1, Lg-ATF1, and ATF2 control the formation of a broadrange of volatile esters. Appl Environ Microbiol 69: 5228–5237.

Walker G. 1998. Yeast Physiology and Biotechnology. Wiley: NewYork.

Wang S, Brandriss M. 1987. Proline utilization in Saccharomycescerevisiae: sequence, regulation, and mitochondrial localizationof the PUT1 gene product. Mol Cell Biol 7: 4431–4440.

Yamamoto M, Wakatsuki T, Hada A, Ryo A. 2001. Use of serialanalysis of gene expression (SAGE) technology. J ImmunolMethods 250: 45–66.

Zhang L, Zhou W, Velculescu V, et al. 1997. Gene expressionprofiles in normal and cancer cells. Science 276: 1268–1272.


Documents

Quantitative analysis of wine yeast gene expression profiles under winemaking conditions