17
BioMed Central Page 1 of 17 (page number not for citation purposes) BMC Genomics Open Access Research article Comparative genomics of wild type yeast strains unveils important genome diversity Laura Carreto †1,2 , Maria F Eiriz †1 , Ana C Gomes 2 , Patrícia M Pereira 1 , Dorit Schuller 3 and Manuel AS Santos* 1 Address: 1 Departamento de Biologia & CESAM, Universidade de Aveiro, 3810-193 Aveiro, Portugal, 2 BIOCANT, Centro de Inovação em Biotecnologia, Parque Tecnológico de Cantanhede, Núcleo 04, Lote 3, 3060-197 Cantanhede, Portugal and 3 Centro de Biologia Molecular e Ambiental (CBMA) Universidade do Minho, Braga, Portugal Email: Laura Carreto - [email protected]; Maria F Eiriz - [email protected]; Ana C Gomes - [email protected]; Patrícia M Pereira - [email protected]; Dorit Schuller - [email protected]; Manuel AS Santos* - [email protected] * Corresponding author †Equal contributors Abstract Background: Genome variability generates phenotypic heterogeneity and is of relevance for adaptation to environmental change, but the extent of such variability in natural populations is still poorly understood. For example, selected Saccharomyces cerevisiae strains are variable at the ploidy level, have gene amplifications, changes in chromosome copy number, and gross chromosomal rearrangements. This suggests that genome plasticity provides important genetic diversity upon which natural selection mechanisms can operate. Results: In this study, we have used wild-type S. cerevisiae (yeast) strains to investigate genome variation in natural and artificial environments. We have used comparative genome hybridization on array (aCGH) to characterize the genome variability of 16 yeast strains, of laboratory and commercial origin, isolated from vineyards and wine cellars, and from opportunistic human infections. Interestingly, sub-telomeric instability was associated with the clinical phenotype, while Ty element insertion regions determined genomic differences of natural wine fermentation strains. Copy number depletion of ASP3 and YRF1 genes was found in all wild-type strains. Other gene families involved in transmembrane transport, sugar and alcohol metabolism or drug resistance had copy number changes, which also distinguished wine from clinical isolates. Conclusion: We have isolated and genotyped more than 1000 yeast strains from natural environments and carried out an aCGH analysis of 16 strains representative of distinct genotype clusters. Important genomic variability was identified between these strains, in particular in sub- telomeric regions and in Ty-element insertion sites, suggesting that this type of genome variability is the main source of genetic diversity in natural populations of yeast. The data highlights the usefulness of yeast as a model system to unravel intraspecific natural genome diversity and to elucidate how natural selection shapes the yeast genome. Published: 4 November 2008 BMC Genomics 2008, 9:524 doi:10.1186/1471-2164-9-524 Received: 19 June 2008 Accepted: 4 November 2008 This article is available from: http://www.biomedcentral.com/1471-2164/9/524 © 2008 Carreto et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Comparative genomics of wild type yeast strains unveils important genome diversity

Embed Size (px)

Citation preview

BioMed CentralBMC Genomics

ss

Open AcceResearch articleComparative genomics of wild type yeast strains unveils important genome diversityLaura Carretodagger12 Maria F Eirizdagger1 Ana C Gomes2 Patriacutecia M Pereira1 Dorit Schuller3 and Manuel AS Santos1

Address 1Departamento de Biologia amp CESAM Universidade de Aveiro 3810-193 Aveiro Portugal 2BIOCANT Centro de Inovaccedilatildeo em Biotecnologia Parque Tecnoloacutegico de Cantanhede Nuacutecleo 04 Lote 3 3060-197 Cantanhede Portugal and 3Centro de Biologia Molecular e Ambiental (CBMA) Universidade do Minho Braga Portugal

Email Laura Carreto - lauracarretouapt Maria F Eiriz - franciscaeirizgmailcom Ana C Gomes - acgomesbiocantpt Patriacutecia M Pereira - pereirapuapt Dorit Schuller - dschullerbiouminhopt Manuel AS Santos - msantosuapt

Corresponding author daggerEqual contributors

AbstractBackground Genome variability generates phenotypic heterogeneity and is of relevance foradaptation to environmental change but the extent of such variability in natural populations is stillpoorly understood For example selected Saccharomyces cerevisiae strains are variable at the ploidylevel have gene amplifications changes in chromosome copy number and gross chromosomalrearrangements This suggests that genome plasticity provides important genetic diversity uponwhich natural selection mechanisms can operate

Results In this study we have used wild-type S cerevisiae (yeast) strains to investigate genomevariation in natural and artificial environments We have used comparative genome hybridizationon array (aCGH) to characterize the genome variability of 16 yeast strains of laboratory andcommercial origin isolated from vineyards and wine cellars and from opportunistic humaninfections Interestingly sub-telomeric instability was associated with the clinical phenotype whileTy element insertion regions determined genomic differences of natural wine fermentation strainsCopy number depletion of ASP3 and YRF1 genes was found in all wild-type strains Other genefamilies involved in transmembrane transport sugar and alcohol metabolism or drug resistance hadcopy number changes which also distinguished wine from clinical isolates

Conclusion We have isolated and genotyped more than 1000 yeast strains from naturalenvironments and carried out an aCGH analysis of 16 strains representative of distinct genotypeclusters Important genomic variability was identified between these strains in particular in sub-telomeric regions and in Ty-element insertion sites suggesting that this type of genome variabilityis the main source of genetic diversity in natural populations of yeast The data highlights theusefulness of yeast as a model system to unravel intraspecific natural genome diversity and toelucidate how natural selection shapes the yeast genome

Published 4 November 2008

BMC Genomics 2008 9524 doi1011861471-2164-9-524

Received 19 June 2008Accepted 4 November 2008

This article is available from httpwwwbiomedcentralcom1471-21649524

copy 2008 Carreto et al licensee BioMed Central Ltd This is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommonsorglicensesby20) which permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Page 1 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

BackgroundThe genome of wild-type and laboratory strains of Saccha-romyces cerevisiae (yeast) has significant genetic variabilityIn general natural isolates are often polyploid or aneu-ploid and have high degree of genetic variability and anessentially asexual life cycle [1-4] Indeed environmentalperturbation often selects strains that display local geneamplifications changes in chromosome copy number orgross chromosomal rearrangements such as intra- orinter-chromosomal translocations mediated by transpo-son-related sequences [5-7]

Recent comparative genomics studies showed that wild-type yeast strains cluster according to technological appli-cation rather than geographical distribution [8-10] How-ever the species as a whole is not domesticated andconsists of both wild-type and comercial populations Forexample specialized sake and wine strains were derivedfrom natural populations not associated with alcoholicbeverages rather than the opposite [11] Also yeaststrains are found in diverse habitats namely in oak exu-dates [1213] gut of insects [14] plant leaves and in grapeberries [15] Interestingly damaged grape barriers but notundamaged berries are an important source of yeaststrains [16] The diversity of yeast strains in viticulturalregions is rather high suggesting the occurrence of specificnatural strains associated with particular terroirs [17-21]

Simple sequence repeat (SSR) analysis used to determinephylogenetic relationships between 651 yeast strains iso-lated from 56 worldwide geographical origins [22]showed that macro geographical differentiation of strainsfrom Asia Europe and Africa accounted for only 28 ofthe observed genetic variation suggesting clonal repro-duction and local domestication The close associationbetween vine migration and wine yeast favors the hypoth-esis that yeast may have followed man and vine as a com-mensal member of grapevine micro flora SSRs were alsoused to distinguish populations from vineyards in closegeographical locations and showed that genetic differ-ences among yeast populations were apparent from grada-tions in allele frequencies rather than from distinctivediagnostic genotypes [23]

The continuous utilization of yeast strains for industrialpurposes introduced artificial selective pressure that mayhave also influenced genome features and novel speciali-zation routes In fact yeast has been identified as anemerging human pathogen that can cause clinically rele-vant infections in immune compromised patients[2425] Such pathogenic strains are phylogeneticallyrelated to baking strains grow at higher temperature pro-duce extracellular proteases are capable of pseudohyphalgrowth and may be resistant to antifungal treatment[102627] Also the genome of the pathogenic S cerevi-

siae strain YJM789 has very high percentage of sequencepolymorphisms (60 000 SNPs scattered over thegenome) which may be a primary cause of phenotypicvariation [28]

The wide ecological geographical clinical and industrialdistribution of yeast strains and the genome diversityalready uncovered suggests that it is a good model systemto understand genome diversity in natural populationsand elucidate the relevance of such diversity for adapta-tion to changing environments and to new ecologicalniches One of the first comprehensive studies on geneticvariation of yeast strains carried out using high-densityoligonucleotide arrays containing up to 200000 oligonu-cleotide probes from the yeast genomic sequenceunveiled differences at the level of single nucleotide poly-morphisms and gene copy number alterations [29] Asimilar approach revealed unexpected differences in 288genes between the S288C and CENPK113-7D laboratorystrains involving differential gene amplification geneabsence or sequence polymorphisms [30]

In order to shed new light on the genome diversity of nat-ural populations of yeast we have isolated more than1000 strains genotyped them and identified clusters thatdistinguished the various genotypes We then selected rep-resentatives of these clusters and compared their genomeswith the genomes of clinical and commercial strains Forthis we used spotted DNA microarrays containing probesfor the complete gene set of the S288C reference strainWe compared the genomes of five commercial winemak-ing strains eight strains isolated from winemaking envi-ronments of two wine producing regions in Portugalnamely the Bairrada and Vinho Verde appellations of ori-gin and three clinical strains The laboratorial strainS288C was used as reference for relative genome profilingOur results highlighted differences between the laborato-rial strain and the wild-type isolates linked to sub-telom-eric instability and retrotransposon activity Theseelements shaped differently the genomes of yeast strainsfrom wine and clinical environments The study also iden-tified functional classes of genes where the copy numbervariations were associated with strains from wine- or clin-ical-related environments No correlation was foundbetween geographical origin and relative genome profilein the larger group of wine-related strains

ResultsStrains and overview of genomic variabilityA total of 16 wild-type strains plus the reference S288Cstrain were used in this study (Table 1) The wine strainswere selected amongst isolates of wine cellars and vine-yards of Bairrada and Vinho Verde wine regions in Portu-gal Strains UM218 and UM237 were selected among 300strains isolated from the Vinho Verde Region [19] for their

Page 2 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

distinctive genetic (simple sequence repeats) and enzy-matic (API ZYM system) profiles Strains from the Bair-rada region were selected from 800 isolates collectedduring two consecutive years of grape harvest and wineproduction and represented major strain clusters of inter-delta region PCR genotyping profiles (A C Gomesunpublished) (Figure 1) Commercial yeast strains used inindustrial wine must fermentations were Lalvin EC-1118used in the Vinho Verde region and Lalvin ICV D254IOC 18ndash2007 AEB Fermol Rouge and Davis Lalvin 522commonly used in the Bairrada region These commercialstrains were initially selected from French wine producingregions (Table 1) and are used world-wide Finally threestrains isolated from patients suffering from opportunis-tic fungal infections were also included in this studyThese strains were selected from a clinical strain collectionon the basis of their inter-delta genotyping profiles (Fig-ure 1) The inter-delta profiles of J940915 and J940557were very similar and these strains were used to ascertainhow the inter-delta region relatedness used to genotypethe strain collection correlated with the genomic variabil-ity

Since natural hybridization between species of the Saccha-romyces sensu stricto group can occur [31] and indeed sev-eral S cerevisiae times S kudriavzevii S bayanus times S cerevisiaeand S bayanus times S cerevisiae times S kudriavzevii have beendescribed among wine strains [32-36] all strains weretested for their hybrid nature For this MET2 locus restric-tion fragment analysis was used S cerevisiae-specific pro-files identical to those of the reference strain S288C weredetected in all cases suggesting that the strains selected foraCGH analysis were authentic S cerevisiae These resultswere corroborated by an additional analysis of 10 poly-morphic S cerevisiae-specific simple sequence repeats

(results not shown) Finally PCR-RFLP profiling of theOPY1 KIN82 MET6 KEL2 and CYR1 loci which arelocated on chromosomes II III V VII and X respectively

Table 1 Yeast strains used in this study

Strains Characteristics Sourceorigin

J940047 Wild-type isolate Clinical PortugalJ940557 Wild-type isolate Clinical PortugalJ940915 Wild-type isolate Clinical Portugal06L3FF02 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L1FF11 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L3FF15 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L6FF20 Wild-type isolate Wine cellar Bairrada Wine Region PortugalUM218 Wild-type isolate Vineyard Vinho Verde Wine Region PortugalUM237 Wild-type isolate Vineyard Vinho Verde Wine Region PortugalBB1235 Wild-type isolate Vineyard Bairrada Wine Region PortugalBB2453 Wild-type isolate Vineyard Bairrada Wine Region PortugalLalvin EC-1118 Commercial used in the Vinho Verde Wine Region Champagne FranceLalvin ICV D254 Commercial used in the Bairrada Wine Region Rhocircne Valley FranceIOC 18ndash2007 Commercial used in the Bairrada Wine Region Institut Oenologique de Champagne FranceAEB Fermol Rouge Commercial used in the Bairrada Wine Region Montpellier University FranceDavis Lalvin 522 Commercial used in the Bairrada Wine Region University of California Davis USAS288C MATα SUC2 mal mel gal2 CUP1 flo1 flo8-1 Mortimer amp Johnston 1986 [87]

Inter-delta region profilesFigure 1Inter-delta region profiles The environmental clinical and commercial yeasts used in this study were genotyped by PCR amplification of the inter-delta regions The strains used in this study were selected from our yeast culture collection on the basis of their different inter-delta PCR profiles as shown Inter-delta PCR profiles were obtained by DNA elec-trophoresis on a Labchip HT (Caliper LS) and the data was displayed using the DataViewer software (Caliper LS) Each lane is identified at the top

J940047

J940577

J940915

06L3FF02

06L1FF11

06L3FF15

06L6FF20

UM218

UM237

BB1235

BB2453

Lalvin EC-1118

Lalvin ICV D254

IOC 18-2007

AEB Fermol Rouge

DAvis Lalvin 522

S288C

700

0

490

0

290

0

190

0

110

0

700

500

300

100

15

bp

Page 3 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

confirmed the above data thus unequivocally demon-strating that the strains used had a S cerevisiae-specificprofile and were not hybrids (Additional File 1 FigureS1)

For aCGH analysis genomic DNA from each strain wasfluorescently labeled and competitively hybridized withgenomic DNA from the reference strain S288C withduplicate experiments in reverse Cy-dye labeling (dye-swap) design (see Methods)

Hierarchical cluster analysis of the aCGH data showedhigh genome variability (Figure 2) and the laboratorystrain S288C was clearly differentiated from the other 16strains Clinical strains (Cluster 2) were clearly differenti-ated from the wine strains clusters (1 3 and 4) High sim-ilarity between the genomes of two of the clinical strainsJ940557 and J940915 was expected from their identicalinter-delta region profiles (Figure 1) and microsatellitepatterns (data not shown) and such genome similaritywas confirmed by aCGH profile similarity (Figure 2)Other pairs of strains with distinct inter-delta region pro-files showed the high similarity in the clustering treenamely AEB Fermol Rouge and 06L1FF11 Davis Lalvin522 and 06L6FF20 or UM218 and UM237

Environmental and commercial wine strains formed threeclusters (Cluster 1 3 and 4) and differences between themwere as high as those observed between wine and clinicalstrains Divergence among wine yeast was not correlatedto geographical origin since strains BB1235 and06L6FF20 (Cluster 3 Bairrada region) were more similarto strains UM218 and UM237 (Cluster 4 Vinho Verderegion) than to strains BB2453 06L3FF02 and 06L1FF11(Cluster 1 Bairrada region) Furthermore commercialstrains isolated from French winemaking regions groupedtogether with strains from the Bairrada and Vinho Verderegions There was also no separation of strains isolatedfrom vineyards (BB2453 BB1235 UM218 and UM237)from cellars (06LF3FF02 06L1FF11 06L6FF2006L3FF15) or from commercial strains (AEB FermolRouge Lalvin ICV D254 Davis Lalvin 522 IOC18ndash2007Lalvin EC-1118)

Genome variability is associated with Ty elements and telomeresYeast genomes evolve through dynamic processes andoften contain gene duplications and deletions or evenchromosomal segment rearrangements Genome instabil-ity occurs throughout the genome but is more frequent inparticular regions such as near Ty elements due to recip-rocal translocations or at sub-telomeric regions possiblycaused by high-frequency of ectopic recombination [37]To investigate the presence of regions of increasedgenome variability we used the Cluster Along Chromo-somes (CLAC) method of the CGH Miner software pack-age [38] This software highlighted the occurrence ofclusters of altered data in a set of samples relatively to con-trols using a moving data window to calculate an averagelog ratio value while controlling the False Discovery Rate(FDR) according to user defined parameters Karyoscopemaps that indicated the location of regions with altera-tions in ORF copy number (amplifications and deletions)for each strain were obtained using CGH Miner analysis aspreviously described by Dunn and colleagues [39] withthe chromosomal coordinates of S288C genome Sincesome aCGH probes interrogated duplicated or multicopyORFs namely ORFs of Ty elements those karyoscopemaps display the average dosage of the repeated ORFsrather than the dosage of each of the repeated ORFs Amoving window of three ORFs was chosen as the loweststatistically meaningful interval for averaging the hybridi-zation signal thus defining the resolution of the analysis

The karyoscope maps displaying the relative hybridiza-tion data derived for each strain revealed that the majorityof the genome alterations corresponded to deletions rela-tive to strain S288C while ORF amplifications were rare(Figure 3 and Additional File 2 Figures S2AndashS2P) ORFamplification clusters were found in some strains mostlylocated in sub-telomeric regions (Figure 3A) and within

Hierarchical clustering of aCGH profilesFigure 2Hierarchical clustering of aCGH profiles The strains used in his study were grouped according to their aCGH profiles For this hierarchical clustering analysis using Pear-son correlation with average linkage of the normalized aCGH profiling data was performed The clusters shown identify strains that shared similar ORF copy number altera-tions

BB2453

06L3FF02

06L1FF11

AEB Fermol Rouge

J940047

J940557

J940915

Lalvin ICV D254

06L3FF15

BB1235

06L6FF20

Davis Lalvin 522

IOC 18-2007

Lalvin EC-1118

UM218

UM237

S288C

Clu

ste

r 1C

lus

ter 2

Clu

ste

r 3C

lus

ter 4

Page 4 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

20 Kb of the S288C respective chromosome end but onlyfew corresponded to genes with annotated function Clus-ters of depleted ORFs were found in sub-telomeric regionsand contained large percentage of Ty elements and hypo-thetical ORFs (Figure 4) In the group of wine strains upto one third of the observed gene copy number alterationswere found in sub-telomeric regions (Figure 4A) ndash within50 Kb from the S288C chromosome ends using the crite-rion of Edwards-Ingram and colleagues [40] On averagethe clinical strains showed slightly higher percentage(almost 40) of depleted ORFs localized near the telom-eres However this is mostly explained by the massive lossof chromosomes VII and X right arms in strains J940557and J940915 (see karyoscope map of J940915 in Figure

3B) The depletion of ORFs around the centromericregions- within 20 Kb of the centromere according to thecriterion of Schacherer and colleagues [41]- was reducedand only slightly above average in some of the winestrains

A clear differentiation between clinical and wine relatedstrains was observed when the frequency of transposableelements within the ORFs with depleted copy number(Figure 4B) was considered Ty elements comprised 36of the ORFs absent in all wine strains which meant thatapproximately one third of the ORFs associated with ret-rotransposon activity (42 out of a total of 113) in S288Cwere absent in these strains The genomes of the clinical

Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B)Figure 3Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B) In order to visualize the gene copy number altera-tions along chromosomes and to have a global overview of the alterations detected by aCGH the data was plotted along each chromosome using the annotated ORF coordinates of S288C Vertical bars represent the relative hybridization pattern rela-tively to the genome of strain S288C Red bars correspond to amplified ORFs green bars represent deleted ORFs and grey bars are statistically non-significant alterations (FDR lt0279) The horizontal lines indicate the hybridization ratios in logarith-mic scale Signals of repeated ORFs such as those of Ty elements or repetitive sequences flanking Ty element insertion sites correspond to the average signal rather than the individual signal dosage of that sequence Clusters of altered ORFs contained roughly one third of Ty -ORFs and contributed to localization of the chromosome alterations following the genome coordi-nates of the S288C strain

AA BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 5 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 6 of 17(page number not for citation purposes)

Chromosomal location and classification of ORFs with variable copy numberFigure 4Chromosomal location and classification of ORFs with variable copy number Panel-A shows the distribution of var-iable ORFs indicated by CGH-Miner in terms of chromosome location that is sub-telomeric centromeric or chromosome arms Panel-B shows the distribution of the same ORFs in terms of abundance of Ty elements hypothetical or annotated ORFs In both panels each bar represents the total number of variable ORFs for a given strain

0

100

200

300

400

500

600

J940

047

J940

557

J940

915

06L3

FF02

06L1

FF11

06L3

FF15

06L6

FF20

BB12

35

BB24

53

UM

218

UM

237

Lalvi

n EC-1

118

Lalvi

n ICV D

254

IOC-1

8 20

07

AEB F

erm

ol Rouge

Davi

s Lalv

in 5

22

OR

Fs

ch

an

ge

d

Sub-telomeric Centromeric Chromosome arms

0

100

200

300

400

500

600

J94004

7

J940

557

J9409

15

06L3F

F02

06L1F

F11

06L3F

F15

06L6F

F20

BB12

35

BB24

53

UM

218

UM

237

Lalv

in E

C-1

118

Lalvin

ICV D

254

IOC-1

8 200

7

AEB F

erm

ol Roug

e

Dav

is L

alvi

n 52

2

OR

Fs

ch

an

ge

d

Hypothetical Ty elements Annotated

A

B

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

strains however contained most of the Ty elements iden-tified in the genomic sequence of S288C since only 7 ofthe ORFs absent in all three of them were classified in thiscategory

Highly variable genomic regions in wine and clinical strainsThe differences between wine and clinical strains were fur-ther highlighted with the consensus plots obtained fromthe intersection of the individual karyoscope maps of thethirteen wine fermentation strains (Figure 5A) comparedto the three clinical isolates (Figure 5B) These plots repre-sented the relative amount of samples with copy numberalterations showing the percentage of strains with a givenamplification or deletion Most of the deletion clustersidentified in wine strains co-localized with absent Ty ele-

ments (Figure 5A) while in clinical isolates a significantnumber of depletions were associated with sub-telomericinstability (Figure 5B) Variable regions affected differentsub-functional categories of genes in wine and clinicalstrains For example carbohydrate transporters and gly-cosidases were the main functional groups of genes absentin clinical strains but genes involved in vesicle transporttelomere maintenance and alcohol metabolism were alsoidentified among the lost sub-telomeric genes The func-tional categories most affected by gene copy number vari-ability in wine strains were related to vitamin metabolismDNA recombination polysaccharide metabolism regula-tion of meiotic cell cycle and reproduction and were fre-quently located in the vicinity of Ty element insertionsites

The ORF variability patterns differ between wine and clinical strainsFigure 5The ORF variability patterns differ between wine and clinical strains The consensus karyoscope maps obtained for wine (Panel-A) and clinical (Panel-B) strains showed that ORF variation in wine strains was distributed along chromosomes while in clinical strains such distribution was more frequent in sub-telomeric regions The consensus maps show the relative percentage of strains (percent of samples) within wine (Panel-A) and clinical (Panel-B) environments with copy number altera-tion in a given ORF relatively to the reference strain S288C The data was plotted according to ORF chromosome location Red bars represent percentage of strains with amplifications and green bars represent the percentage of strains with deletions according to the grey-line scale shown above and below the chromosome central line Open circles indicate the position of centromeres and blue arrows identify Ty elements

BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 7 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

BackgroundThe genome of wild-type and laboratory strains of Saccha-romyces cerevisiae (yeast) has significant genetic variabilityIn general natural isolates are often polyploid or aneu-ploid and have high degree of genetic variability and anessentially asexual life cycle [1-4] Indeed environmentalperturbation often selects strains that display local geneamplifications changes in chromosome copy number orgross chromosomal rearrangements such as intra- orinter-chromosomal translocations mediated by transpo-son-related sequences [5-7]

Recent comparative genomics studies showed that wild-type yeast strains cluster according to technological appli-cation rather than geographical distribution [8-10] How-ever the species as a whole is not domesticated andconsists of both wild-type and comercial populations Forexample specialized sake and wine strains were derivedfrom natural populations not associated with alcoholicbeverages rather than the opposite [11] Also yeaststrains are found in diverse habitats namely in oak exu-dates [1213] gut of insects [14] plant leaves and in grapeberries [15] Interestingly damaged grape barriers but notundamaged berries are an important source of yeaststrains [16] The diversity of yeast strains in viticulturalregions is rather high suggesting the occurrence of specificnatural strains associated with particular terroirs [17-21]

Simple sequence repeat (SSR) analysis used to determinephylogenetic relationships between 651 yeast strains iso-lated from 56 worldwide geographical origins [22]showed that macro geographical differentiation of strainsfrom Asia Europe and Africa accounted for only 28 ofthe observed genetic variation suggesting clonal repro-duction and local domestication The close associationbetween vine migration and wine yeast favors the hypoth-esis that yeast may have followed man and vine as a com-mensal member of grapevine micro flora SSRs were alsoused to distinguish populations from vineyards in closegeographical locations and showed that genetic differ-ences among yeast populations were apparent from grada-tions in allele frequencies rather than from distinctivediagnostic genotypes [23]

The continuous utilization of yeast strains for industrialpurposes introduced artificial selective pressure that mayhave also influenced genome features and novel speciali-zation routes In fact yeast has been identified as anemerging human pathogen that can cause clinically rele-vant infections in immune compromised patients[2425] Such pathogenic strains are phylogeneticallyrelated to baking strains grow at higher temperature pro-duce extracellular proteases are capable of pseudohyphalgrowth and may be resistant to antifungal treatment[102627] Also the genome of the pathogenic S cerevi-

siae strain YJM789 has very high percentage of sequencepolymorphisms (60 000 SNPs scattered over thegenome) which may be a primary cause of phenotypicvariation [28]

The wide ecological geographical clinical and industrialdistribution of yeast strains and the genome diversityalready uncovered suggests that it is a good model systemto understand genome diversity in natural populationsand elucidate the relevance of such diversity for adapta-tion to changing environments and to new ecologicalniches One of the first comprehensive studies on geneticvariation of yeast strains carried out using high-densityoligonucleotide arrays containing up to 200000 oligonu-cleotide probes from the yeast genomic sequenceunveiled differences at the level of single nucleotide poly-morphisms and gene copy number alterations [29] Asimilar approach revealed unexpected differences in 288genes between the S288C and CENPK113-7D laboratorystrains involving differential gene amplification geneabsence or sequence polymorphisms [30]

In order to shed new light on the genome diversity of nat-ural populations of yeast we have isolated more than1000 strains genotyped them and identified clusters thatdistinguished the various genotypes We then selected rep-resentatives of these clusters and compared their genomeswith the genomes of clinical and commercial strains Forthis we used spotted DNA microarrays containing probesfor the complete gene set of the S288C reference strainWe compared the genomes of five commercial winemak-ing strains eight strains isolated from winemaking envi-ronments of two wine producing regions in Portugalnamely the Bairrada and Vinho Verde appellations of ori-gin and three clinical strains The laboratorial strainS288C was used as reference for relative genome profilingOur results highlighted differences between the laborato-rial strain and the wild-type isolates linked to sub-telom-eric instability and retrotransposon activity Theseelements shaped differently the genomes of yeast strainsfrom wine and clinical environments The study also iden-tified functional classes of genes where the copy numbervariations were associated with strains from wine- or clin-ical-related environments No correlation was foundbetween geographical origin and relative genome profilein the larger group of wine-related strains

ResultsStrains and overview of genomic variabilityA total of 16 wild-type strains plus the reference S288Cstrain were used in this study (Table 1) The wine strainswere selected amongst isolates of wine cellars and vine-yards of Bairrada and Vinho Verde wine regions in Portu-gal Strains UM218 and UM237 were selected among 300strains isolated from the Vinho Verde Region [19] for their

Page 2 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

distinctive genetic (simple sequence repeats) and enzy-matic (API ZYM system) profiles Strains from the Bair-rada region were selected from 800 isolates collectedduring two consecutive years of grape harvest and wineproduction and represented major strain clusters of inter-delta region PCR genotyping profiles (A C Gomesunpublished) (Figure 1) Commercial yeast strains used inindustrial wine must fermentations were Lalvin EC-1118used in the Vinho Verde region and Lalvin ICV D254IOC 18ndash2007 AEB Fermol Rouge and Davis Lalvin 522commonly used in the Bairrada region These commercialstrains were initially selected from French wine producingregions (Table 1) and are used world-wide Finally threestrains isolated from patients suffering from opportunis-tic fungal infections were also included in this studyThese strains were selected from a clinical strain collectionon the basis of their inter-delta genotyping profiles (Fig-ure 1) The inter-delta profiles of J940915 and J940557were very similar and these strains were used to ascertainhow the inter-delta region relatedness used to genotypethe strain collection correlated with the genomic variabil-ity

Since natural hybridization between species of the Saccha-romyces sensu stricto group can occur [31] and indeed sev-eral S cerevisiae times S kudriavzevii S bayanus times S cerevisiaeand S bayanus times S cerevisiae times S kudriavzevii have beendescribed among wine strains [32-36] all strains weretested for their hybrid nature For this MET2 locus restric-tion fragment analysis was used S cerevisiae-specific pro-files identical to those of the reference strain S288C weredetected in all cases suggesting that the strains selected foraCGH analysis were authentic S cerevisiae These resultswere corroborated by an additional analysis of 10 poly-morphic S cerevisiae-specific simple sequence repeats

(results not shown) Finally PCR-RFLP profiling of theOPY1 KIN82 MET6 KEL2 and CYR1 loci which arelocated on chromosomes II III V VII and X respectively

Table 1 Yeast strains used in this study

Strains Characteristics Sourceorigin

J940047 Wild-type isolate Clinical PortugalJ940557 Wild-type isolate Clinical PortugalJ940915 Wild-type isolate Clinical Portugal06L3FF02 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L1FF11 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L3FF15 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L6FF20 Wild-type isolate Wine cellar Bairrada Wine Region PortugalUM218 Wild-type isolate Vineyard Vinho Verde Wine Region PortugalUM237 Wild-type isolate Vineyard Vinho Verde Wine Region PortugalBB1235 Wild-type isolate Vineyard Bairrada Wine Region PortugalBB2453 Wild-type isolate Vineyard Bairrada Wine Region PortugalLalvin EC-1118 Commercial used in the Vinho Verde Wine Region Champagne FranceLalvin ICV D254 Commercial used in the Bairrada Wine Region Rhocircne Valley FranceIOC 18ndash2007 Commercial used in the Bairrada Wine Region Institut Oenologique de Champagne FranceAEB Fermol Rouge Commercial used in the Bairrada Wine Region Montpellier University FranceDavis Lalvin 522 Commercial used in the Bairrada Wine Region University of California Davis USAS288C MATα SUC2 mal mel gal2 CUP1 flo1 flo8-1 Mortimer amp Johnston 1986 [87]

Inter-delta region profilesFigure 1Inter-delta region profiles The environmental clinical and commercial yeasts used in this study were genotyped by PCR amplification of the inter-delta regions The strains used in this study were selected from our yeast culture collection on the basis of their different inter-delta PCR profiles as shown Inter-delta PCR profiles were obtained by DNA elec-trophoresis on a Labchip HT (Caliper LS) and the data was displayed using the DataViewer software (Caliper LS) Each lane is identified at the top

J940047

J940577

J940915

06L3FF02

06L1FF11

06L3FF15

06L6FF20

UM218

UM237

BB1235

BB2453

Lalvin EC-1118

Lalvin ICV D254

IOC 18-2007

AEB Fermol Rouge

DAvis Lalvin 522

S288C

700

0

490

0

290

0

190

0

110

0

700

500

300

100

15

bp

Page 3 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

confirmed the above data thus unequivocally demon-strating that the strains used had a S cerevisiae-specificprofile and were not hybrids (Additional File 1 FigureS1)

For aCGH analysis genomic DNA from each strain wasfluorescently labeled and competitively hybridized withgenomic DNA from the reference strain S288C withduplicate experiments in reverse Cy-dye labeling (dye-swap) design (see Methods)

Hierarchical cluster analysis of the aCGH data showedhigh genome variability (Figure 2) and the laboratorystrain S288C was clearly differentiated from the other 16strains Clinical strains (Cluster 2) were clearly differenti-ated from the wine strains clusters (1 3 and 4) High sim-ilarity between the genomes of two of the clinical strainsJ940557 and J940915 was expected from their identicalinter-delta region profiles (Figure 1) and microsatellitepatterns (data not shown) and such genome similaritywas confirmed by aCGH profile similarity (Figure 2)Other pairs of strains with distinct inter-delta region pro-files showed the high similarity in the clustering treenamely AEB Fermol Rouge and 06L1FF11 Davis Lalvin522 and 06L6FF20 or UM218 and UM237

Environmental and commercial wine strains formed threeclusters (Cluster 1 3 and 4) and differences between themwere as high as those observed between wine and clinicalstrains Divergence among wine yeast was not correlatedto geographical origin since strains BB1235 and06L6FF20 (Cluster 3 Bairrada region) were more similarto strains UM218 and UM237 (Cluster 4 Vinho Verderegion) than to strains BB2453 06L3FF02 and 06L1FF11(Cluster 1 Bairrada region) Furthermore commercialstrains isolated from French winemaking regions groupedtogether with strains from the Bairrada and Vinho Verderegions There was also no separation of strains isolatedfrom vineyards (BB2453 BB1235 UM218 and UM237)from cellars (06LF3FF02 06L1FF11 06L6FF2006L3FF15) or from commercial strains (AEB FermolRouge Lalvin ICV D254 Davis Lalvin 522 IOC18ndash2007Lalvin EC-1118)

Genome variability is associated with Ty elements and telomeresYeast genomes evolve through dynamic processes andoften contain gene duplications and deletions or evenchromosomal segment rearrangements Genome instabil-ity occurs throughout the genome but is more frequent inparticular regions such as near Ty elements due to recip-rocal translocations or at sub-telomeric regions possiblycaused by high-frequency of ectopic recombination [37]To investigate the presence of regions of increasedgenome variability we used the Cluster Along Chromo-somes (CLAC) method of the CGH Miner software pack-age [38] This software highlighted the occurrence ofclusters of altered data in a set of samples relatively to con-trols using a moving data window to calculate an averagelog ratio value while controlling the False Discovery Rate(FDR) according to user defined parameters Karyoscopemaps that indicated the location of regions with altera-tions in ORF copy number (amplifications and deletions)for each strain were obtained using CGH Miner analysis aspreviously described by Dunn and colleagues [39] withthe chromosomal coordinates of S288C genome Sincesome aCGH probes interrogated duplicated or multicopyORFs namely ORFs of Ty elements those karyoscopemaps display the average dosage of the repeated ORFsrather than the dosage of each of the repeated ORFs Amoving window of three ORFs was chosen as the loweststatistically meaningful interval for averaging the hybridi-zation signal thus defining the resolution of the analysis

The karyoscope maps displaying the relative hybridiza-tion data derived for each strain revealed that the majorityof the genome alterations corresponded to deletions rela-tive to strain S288C while ORF amplifications were rare(Figure 3 and Additional File 2 Figures S2AndashS2P) ORFamplification clusters were found in some strains mostlylocated in sub-telomeric regions (Figure 3A) and within

Hierarchical clustering of aCGH profilesFigure 2Hierarchical clustering of aCGH profiles The strains used in his study were grouped according to their aCGH profiles For this hierarchical clustering analysis using Pear-son correlation with average linkage of the normalized aCGH profiling data was performed The clusters shown identify strains that shared similar ORF copy number altera-tions

BB2453

06L3FF02

06L1FF11

AEB Fermol Rouge

J940047

J940557

J940915

Lalvin ICV D254

06L3FF15

BB1235

06L6FF20

Davis Lalvin 522

IOC 18-2007

Lalvin EC-1118

UM218

UM237

S288C

Clu

ste

r 1C

lus

ter 2

Clu

ste

r 3C

lus

ter 4

Page 4 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

20 Kb of the S288C respective chromosome end but onlyfew corresponded to genes with annotated function Clus-ters of depleted ORFs were found in sub-telomeric regionsand contained large percentage of Ty elements and hypo-thetical ORFs (Figure 4) In the group of wine strains upto one third of the observed gene copy number alterationswere found in sub-telomeric regions (Figure 4A) ndash within50 Kb from the S288C chromosome ends using the crite-rion of Edwards-Ingram and colleagues [40] On averagethe clinical strains showed slightly higher percentage(almost 40) of depleted ORFs localized near the telom-eres However this is mostly explained by the massive lossof chromosomes VII and X right arms in strains J940557and J940915 (see karyoscope map of J940915 in Figure

3B) The depletion of ORFs around the centromericregions- within 20 Kb of the centromere according to thecriterion of Schacherer and colleagues [41]- was reducedand only slightly above average in some of the winestrains

A clear differentiation between clinical and wine relatedstrains was observed when the frequency of transposableelements within the ORFs with depleted copy number(Figure 4B) was considered Ty elements comprised 36of the ORFs absent in all wine strains which meant thatapproximately one third of the ORFs associated with ret-rotransposon activity (42 out of a total of 113) in S288Cwere absent in these strains The genomes of the clinical

Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B)Figure 3Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B) In order to visualize the gene copy number altera-tions along chromosomes and to have a global overview of the alterations detected by aCGH the data was plotted along each chromosome using the annotated ORF coordinates of S288C Vertical bars represent the relative hybridization pattern rela-tively to the genome of strain S288C Red bars correspond to amplified ORFs green bars represent deleted ORFs and grey bars are statistically non-significant alterations (FDR lt0279) The horizontal lines indicate the hybridization ratios in logarith-mic scale Signals of repeated ORFs such as those of Ty elements or repetitive sequences flanking Ty element insertion sites correspond to the average signal rather than the individual signal dosage of that sequence Clusters of altered ORFs contained roughly one third of Ty -ORFs and contributed to localization of the chromosome alterations following the genome coordi-nates of the S288C strain

AA BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 5 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 6 of 17(page number not for citation purposes)

Chromosomal location and classification of ORFs with variable copy numberFigure 4Chromosomal location and classification of ORFs with variable copy number Panel-A shows the distribution of var-iable ORFs indicated by CGH-Miner in terms of chromosome location that is sub-telomeric centromeric or chromosome arms Panel-B shows the distribution of the same ORFs in terms of abundance of Ty elements hypothetical or annotated ORFs In both panels each bar represents the total number of variable ORFs for a given strain

0

100

200

300

400

500

600

J940

047

J940

557

J940

915

06L3

FF02

06L1

FF11

06L3

FF15

06L6

FF20

BB12

35

BB24

53

UM

218

UM

237

Lalvi

n EC-1

118

Lalvi

n ICV D

254

IOC-1

8 20

07

AEB F

erm

ol Rouge

Davi

s Lalv

in 5

22

OR

Fs

ch

an

ge

d

Sub-telomeric Centromeric Chromosome arms

0

100

200

300

400

500

600

J94004

7

J940

557

J9409

15

06L3F

F02

06L1F

F11

06L3F

F15

06L6F

F20

BB12

35

BB24

53

UM

218

UM

237

Lalv

in E

C-1

118

Lalvin

ICV D

254

IOC-1

8 200

7

AEB F

erm

ol Roug

e

Dav

is L

alvi

n 52

2

OR

Fs

ch

an

ge

d

Hypothetical Ty elements Annotated

A

B

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

strains however contained most of the Ty elements iden-tified in the genomic sequence of S288C since only 7 ofthe ORFs absent in all three of them were classified in thiscategory

Highly variable genomic regions in wine and clinical strainsThe differences between wine and clinical strains were fur-ther highlighted with the consensus plots obtained fromthe intersection of the individual karyoscope maps of thethirteen wine fermentation strains (Figure 5A) comparedto the three clinical isolates (Figure 5B) These plots repre-sented the relative amount of samples with copy numberalterations showing the percentage of strains with a givenamplification or deletion Most of the deletion clustersidentified in wine strains co-localized with absent Ty ele-

ments (Figure 5A) while in clinical isolates a significantnumber of depletions were associated with sub-telomericinstability (Figure 5B) Variable regions affected differentsub-functional categories of genes in wine and clinicalstrains For example carbohydrate transporters and gly-cosidases were the main functional groups of genes absentin clinical strains but genes involved in vesicle transporttelomere maintenance and alcohol metabolism were alsoidentified among the lost sub-telomeric genes The func-tional categories most affected by gene copy number vari-ability in wine strains were related to vitamin metabolismDNA recombination polysaccharide metabolism regula-tion of meiotic cell cycle and reproduction and were fre-quently located in the vicinity of Ty element insertionsites

The ORF variability patterns differ between wine and clinical strainsFigure 5The ORF variability patterns differ between wine and clinical strains The consensus karyoscope maps obtained for wine (Panel-A) and clinical (Panel-B) strains showed that ORF variation in wine strains was distributed along chromosomes while in clinical strains such distribution was more frequent in sub-telomeric regions The consensus maps show the relative percentage of strains (percent of samples) within wine (Panel-A) and clinical (Panel-B) environments with copy number altera-tion in a given ORF relatively to the reference strain S288C The data was plotted according to ORF chromosome location Red bars represent percentage of strains with amplifications and green bars represent the percentage of strains with deletions according to the grey-line scale shown above and below the chromosome central line Open circles indicate the position of centromeres and blue arrows identify Ty elements

BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 7 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

distinctive genetic (simple sequence repeats) and enzy-matic (API ZYM system) profiles Strains from the Bair-rada region were selected from 800 isolates collectedduring two consecutive years of grape harvest and wineproduction and represented major strain clusters of inter-delta region PCR genotyping profiles (A C Gomesunpublished) (Figure 1) Commercial yeast strains used inindustrial wine must fermentations were Lalvin EC-1118used in the Vinho Verde region and Lalvin ICV D254IOC 18ndash2007 AEB Fermol Rouge and Davis Lalvin 522commonly used in the Bairrada region These commercialstrains were initially selected from French wine producingregions (Table 1) and are used world-wide Finally threestrains isolated from patients suffering from opportunis-tic fungal infections were also included in this studyThese strains were selected from a clinical strain collectionon the basis of their inter-delta genotyping profiles (Fig-ure 1) The inter-delta profiles of J940915 and J940557were very similar and these strains were used to ascertainhow the inter-delta region relatedness used to genotypethe strain collection correlated with the genomic variabil-ity

Since natural hybridization between species of the Saccha-romyces sensu stricto group can occur [31] and indeed sev-eral S cerevisiae times S kudriavzevii S bayanus times S cerevisiaeand S bayanus times S cerevisiae times S kudriavzevii have beendescribed among wine strains [32-36] all strains weretested for their hybrid nature For this MET2 locus restric-tion fragment analysis was used S cerevisiae-specific pro-files identical to those of the reference strain S288C weredetected in all cases suggesting that the strains selected foraCGH analysis were authentic S cerevisiae These resultswere corroborated by an additional analysis of 10 poly-morphic S cerevisiae-specific simple sequence repeats

(results not shown) Finally PCR-RFLP profiling of theOPY1 KIN82 MET6 KEL2 and CYR1 loci which arelocated on chromosomes II III V VII and X respectively

Table 1 Yeast strains used in this study

Strains Characteristics Sourceorigin

J940047 Wild-type isolate Clinical PortugalJ940557 Wild-type isolate Clinical PortugalJ940915 Wild-type isolate Clinical Portugal06L3FF02 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L1FF11 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L3FF15 Wild-type isolate Wine cellar Bairrada Wine Region Portugal06L6FF20 Wild-type isolate Wine cellar Bairrada Wine Region PortugalUM218 Wild-type isolate Vineyard Vinho Verde Wine Region PortugalUM237 Wild-type isolate Vineyard Vinho Verde Wine Region PortugalBB1235 Wild-type isolate Vineyard Bairrada Wine Region PortugalBB2453 Wild-type isolate Vineyard Bairrada Wine Region PortugalLalvin EC-1118 Commercial used in the Vinho Verde Wine Region Champagne FranceLalvin ICV D254 Commercial used in the Bairrada Wine Region Rhocircne Valley FranceIOC 18ndash2007 Commercial used in the Bairrada Wine Region Institut Oenologique de Champagne FranceAEB Fermol Rouge Commercial used in the Bairrada Wine Region Montpellier University FranceDavis Lalvin 522 Commercial used in the Bairrada Wine Region University of California Davis USAS288C MATα SUC2 mal mel gal2 CUP1 flo1 flo8-1 Mortimer amp Johnston 1986 [87]

Inter-delta region profilesFigure 1Inter-delta region profiles The environmental clinical and commercial yeasts used in this study were genotyped by PCR amplification of the inter-delta regions The strains used in this study were selected from our yeast culture collection on the basis of their different inter-delta PCR profiles as shown Inter-delta PCR profiles were obtained by DNA elec-trophoresis on a Labchip HT (Caliper LS) and the data was displayed using the DataViewer software (Caliper LS) Each lane is identified at the top

J940047

J940577

J940915

06L3FF02

06L1FF11

06L3FF15

06L6FF20

UM218

UM237

BB1235

BB2453

Lalvin EC-1118

Lalvin ICV D254

IOC 18-2007

AEB Fermol Rouge

DAvis Lalvin 522

S288C

700

0

490

0

290

0

190

0

110

0

700

500

300

100

15

bp

Page 3 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

confirmed the above data thus unequivocally demon-strating that the strains used had a S cerevisiae-specificprofile and were not hybrids (Additional File 1 FigureS1)

For aCGH analysis genomic DNA from each strain wasfluorescently labeled and competitively hybridized withgenomic DNA from the reference strain S288C withduplicate experiments in reverse Cy-dye labeling (dye-swap) design (see Methods)

Hierarchical cluster analysis of the aCGH data showedhigh genome variability (Figure 2) and the laboratorystrain S288C was clearly differentiated from the other 16strains Clinical strains (Cluster 2) were clearly differenti-ated from the wine strains clusters (1 3 and 4) High sim-ilarity between the genomes of two of the clinical strainsJ940557 and J940915 was expected from their identicalinter-delta region profiles (Figure 1) and microsatellitepatterns (data not shown) and such genome similaritywas confirmed by aCGH profile similarity (Figure 2)Other pairs of strains with distinct inter-delta region pro-files showed the high similarity in the clustering treenamely AEB Fermol Rouge and 06L1FF11 Davis Lalvin522 and 06L6FF20 or UM218 and UM237

Environmental and commercial wine strains formed threeclusters (Cluster 1 3 and 4) and differences between themwere as high as those observed between wine and clinicalstrains Divergence among wine yeast was not correlatedto geographical origin since strains BB1235 and06L6FF20 (Cluster 3 Bairrada region) were more similarto strains UM218 and UM237 (Cluster 4 Vinho Verderegion) than to strains BB2453 06L3FF02 and 06L1FF11(Cluster 1 Bairrada region) Furthermore commercialstrains isolated from French winemaking regions groupedtogether with strains from the Bairrada and Vinho Verderegions There was also no separation of strains isolatedfrom vineyards (BB2453 BB1235 UM218 and UM237)from cellars (06LF3FF02 06L1FF11 06L6FF2006L3FF15) or from commercial strains (AEB FermolRouge Lalvin ICV D254 Davis Lalvin 522 IOC18ndash2007Lalvin EC-1118)

Genome variability is associated with Ty elements and telomeresYeast genomes evolve through dynamic processes andoften contain gene duplications and deletions or evenchromosomal segment rearrangements Genome instabil-ity occurs throughout the genome but is more frequent inparticular regions such as near Ty elements due to recip-rocal translocations or at sub-telomeric regions possiblycaused by high-frequency of ectopic recombination [37]To investigate the presence of regions of increasedgenome variability we used the Cluster Along Chromo-somes (CLAC) method of the CGH Miner software pack-age [38] This software highlighted the occurrence ofclusters of altered data in a set of samples relatively to con-trols using a moving data window to calculate an averagelog ratio value while controlling the False Discovery Rate(FDR) according to user defined parameters Karyoscopemaps that indicated the location of regions with altera-tions in ORF copy number (amplifications and deletions)for each strain were obtained using CGH Miner analysis aspreviously described by Dunn and colleagues [39] withthe chromosomal coordinates of S288C genome Sincesome aCGH probes interrogated duplicated or multicopyORFs namely ORFs of Ty elements those karyoscopemaps display the average dosage of the repeated ORFsrather than the dosage of each of the repeated ORFs Amoving window of three ORFs was chosen as the loweststatistically meaningful interval for averaging the hybridi-zation signal thus defining the resolution of the analysis

The karyoscope maps displaying the relative hybridiza-tion data derived for each strain revealed that the majorityof the genome alterations corresponded to deletions rela-tive to strain S288C while ORF amplifications were rare(Figure 3 and Additional File 2 Figures S2AndashS2P) ORFamplification clusters were found in some strains mostlylocated in sub-telomeric regions (Figure 3A) and within

Hierarchical clustering of aCGH profilesFigure 2Hierarchical clustering of aCGH profiles The strains used in his study were grouped according to their aCGH profiles For this hierarchical clustering analysis using Pear-son correlation with average linkage of the normalized aCGH profiling data was performed The clusters shown identify strains that shared similar ORF copy number altera-tions

BB2453

06L3FF02

06L1FF11

AEB Fermol Rouge

J940047

J940557

J940915

Lalvin ICV D254

06L3FF15

BB1235

06L6FF20

Davis Lalvin 522

IOC 18-2007

Lalvin EC-1118

UM218

UM237

S288C

Clu

ste

r 1C

lus

ter 2

Clu

ste

r 3C

lus

ter 4

Page 4 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

20 Kb of the S288C respective chromosome end but onlyfew corresponded to genes with annotated function Clus-ters of depleted ORFs were found in sub-telomeric regionsand contained large percentage of Ty elements and hypo-thetical ORFs (Figure 4) In the group of wine strains upto one third of the observed gene copy number alterationswere found in sub-telomeric regions (Figure 4A) ndash within50 Kb from the S288C chromosome ends using the crite-rion of Edwards-Ingram and colleagues [40] On averagethe clinical strains showed slightly higher percentage(almost 40) of depleted ORFs localized near the telom-eres However this is mostly explained by the massive lossof chromosomes VII and X right arms in strains J940557and J940915 (see karyoscope map of J940915 in Figure

3B) The depletion of ORFs around the centromericregions- within 20 Kb of the centromere according to thecriterion of Schacherer and colleagues [41]- was reducedand only slightly above average in some of the winestrains

A clear differentiation between clinical and wine relatedstrains was observed when the frequency of transposableelements within the ORFs with depleted copy number(Figure 4B) was considered Ty elements comprised 36of the ORFs absent in all wine strains which meant thatapproximately one third of the ORFs associated with ret-rotransposon activity (42 out of a total of 113) in S288Cwere absent in these strains The genomes of the clinical

Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B)Figure 3Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B) In order to visualize the gene copy number altera-tions along chromosomes and to have a global overview of the alterations detected by aCGH the data was plotted along each chromosome using the annotated ORF coordinates of S288C Vertical bars represent the relative hybridization pattern rela-tively to the genome of strain S288C Red bars correspond to amplified ORFs green bars represent deleted ORFs and grey bars are statistically non-significant alterations (FDR lt0279) The horizontal lines indicate the hybridization ratios in logarith-mic scale Signals of repeated ORFs such as those of Ty elements or repetitive sequences flanking Ty element insertion sites correspond to the average signal rather than the individual signal dosage of that sequence Clusters of altered ORFs contained roughly one third of Ty -ORFs and contributed to localization of the chromosome alterations following the genome coordi-nates of the S288C strain

AA BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 5 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 6 of 17(page number not for citation purposes)

Chromosomal location and classification of ORFs with variable copy numberFigure 4Chromosomal location and classification of ORFs with variable copy number Panel-A shows the distribution of var-iable ORFs indicated by CGH-Miner in terms of chromosome location that is sub-telomeric centromeric or chromosome arms Panel-B shows the distribution of the same ORFs in terms of abundance of Ty elements hypothetical or annotated ORFs In both panels each bar represents the total number of variable ORFs for a given strain

0

100

200

300

400

500

600

J940

047

J940

557

J940

915

06L3

FF02

06L1

FF11

06L3

FF15

06L6

FF20

BB12

35

BB24

53

UM

218

UM

237

Lalvi

n EC-1

118

Lalvi

n ICV D

254

IOC-1

8 20

07

AEB F

erm

ol Rouge

Davi

s Lalv

in 5

22

OR

Fs

ch

an

ge

d

Sub-telomeric Centromeric Chromosome arms

0

100

200

300

400

500

600

J94004

7

J940

557

J9409

15

06L3F

F02

06L1F

F11

06L3F

F15

06L6F

F20

BB12

35

BB24

53

UM

218

UM

237

Lalv

in E

C-1

118

Lalvin

ICV D

254

IOC-1

8 200

7

AEB F

erm

ol Roug

e

Dav

is L

alvi

n 52

2

OR

Fs

ch

an

ge

d

Hypothetical Ty elements Annotated

A

B

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

strains however contained most of the Ty elements iden-tified in the genomic sequence of S288C since only 7 ofthe ORFs absent in all three of them were classified in thiscategory

Highly variable genomic regions in wine and clinical strainsThe differences between wine and clinical strains were fur-ther highlighted with the consensus plots obtained fromthe intersection of the individual karyoscope maps of thethirteen wine fermentation strains (Figure 5A) comparedto the three clinical isolates (Figure 5B) These plots repre-sented the relative amount of samples with copy numberalterations showing the percentage of strains with a givenamplification or deletion Most of the deletion clustersidentified in wine strains co-localized with absent Ty ele-

ments (Figure 5A) while in clinical isolates a significantnumber of depletions were associated with sub-telomericinstability (Figure 5B) Variable regions affected differentsub-functional categories of genes in wine and clinicalstrains For example carbohydrate transporters and gly-cosidases were the main functional groups of genes absentin clinical strains but genes involved in vesicle transporttelomere maintenance and alcohol metabolism were alsoidentified among the lost sub-telomeric genes The func-tional categories most affected by gene copy number vari-ability in wine strains were related to vitamin metabolismDNA recombination polysaccharide metabolism regula-tion of meiotic cell cycle and reproduction and were fre-quently located in the vicinity of Ty element insertionsites

The ORF variability patterns differ between wine and clinical strainsFigure 5The ORF variability patterns differ between wine and clinical strains The consensus karyoscope maps obtained for wine (Panel-A) and clinical (Panel-B) strains showed that ORF variation in wine strains was distributed along chromosomes while in clinical strains such distribution was more frequent in sub-telomeric regions The consensus maps show the relative percentage of strains (percent of samples) within wine (Panel-A) and clinical (Panel-B) environments with copy number altera-tion in a given ORF relatively to the reference strain S288C The data was plotted according to ORF chromosome location Red bars represent percentage of strains with amplifications and green bars represent the percentage of strains with deletions according to the grey-line scale shown above and below the chromosome central line Open circles indicate the position of centromeres and blue arrows identify Ty elements

BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 7 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

confirmed the above data thus unequivocally demon-strating that the strains used had a S cerevisiae-specificprofile and were not hybrids (Additional File 1 FigureS1)

For aCGH analysis genomic DNA from each strain wasfluorescently labeled and competitively hybridized withgenomic DNA from the reference strain S288C withduplicate experiments in reverse Cy-dye labeling (dye-swap) design (see Methods)

Hierarchical cluster analysis of the aCGH data showedhigh genome variability (Figure 2) and the laboratorystrain S288C was clearly differentiated from the other 16strains Clinical strains (Cluster 2) were clearly differenti-ated from the wine strains clusters (1 3 and 4) High sim-ilarity between the genomes of two of the clinical strainsJ940557 and J940915 was expected from their identicalinter-delta region profiles (Figure 1) and microsatellitepatterns (data not shown) and such genome similaritywas confirmed by aCGH profile similarity (Figure 2)Other pairs of strains with distinct inter-delta region pro-files showed the high similarity in the clustering treenamely AEB Fermol Rouge and 06L1FF11 Davis Lalvin522 and 06L6FF20 or UM218 and UM237

Environmental and commercial wine strains formed threeclusters (Cluster 1 3 and 4) and differences between themwere as high as those observed between wine and clinicalstrains Divergence among wine yeast was not correlatedto geographical origin since strains BB1235 and06L6FF20 (Cluster 3 Bairrada region) were more similarto strains UM218 and UM237 (Cluster 4 Vinho Verderegion) than to strains BB2453 06L3FF02 and 06L1FF11(Cluster 1 Bairrada region) Furthermore commercialstrains isolated from French winemaking regions groupedtogether with strains from the Bairrada and Vinho Verderegions There was also no separation of strains isolatedfrom vineyards (BB2453 BB1235 UM218 and UM237)from cellars (06LF3FF02 06L1FF11 06L6FF2006L3FF15) or from commercial strains (AEB FermolRouge Lalvin ICV D254 Davis Lalvin 522 IOC18ndash2007Lalvin EC-1118)

Genome variability is associated with Ty elements and telomeresYeast genomes evolve through dynamic processes andoften contain gene duplications and deletions or evenchromosomal segment rearrangements Genome instabil-ity occurs throughout the genome but is more frequent inparticular regions such as near Ty elements due to recip-rocal translocations or at sub-telomeric regions possiblycaused by high-frequency of ectopic recombination [37]To investigate the presence of regions of increasedgenome variability we used the Cluster Along Chromo-somes (CLAC) method of the CGH Miner software pack-age [38] This software highlighted the occurrence ofclusters of altered data in a set of samples relatively to con-trols using a moving data window to calculate an averagelog ratio value while controlling the False Discovery Rate(FDR) according to user defined parameters Karyoscopemaps that indicated the location of regions with altera-tions in ORF copy number (amplifications and deletions)for each strain were obtained using CGH Miner analysis aspreviously described by Dunn and colleagues [39] withthe chromosomal coordinates of S288C genome Sincesome aCGH probes interrogated duplicated or multicopyORFs namely ORFs of Ty elements those karyoscopemaps display the average dosage of the repeated ORFsrather than the dosage of each of the repeated ORFs Amoving window of three ORFs was chosen as the loweststatistically meaningful interval for averaging the hybridi-zation signal thus defining the resolution of the analysis

The karyoscope maps displaying the relative hybridiza-tion data derived for each strain revealed that the majorityof the genome alterations corresponded to deletions rela-tive to strain S288C while ORF amplifications were rare(Figure 3 and Additional File 2 Figures S2AndashS2P) ORFamplification clusters were found in some strains mostlylocated in sub-telomeric regions (Figure 3A) and within

Hierarchical clustering of aCGH profilesFigure 2Hierarchical clustering of aCGH profiles The strains used in his study were grouped according to their aCGH profiles For this hierarchical clustering analysis using Pear-son correlation with average linkage of the normalized aCGH profiling data was performed The clusters shown identify strains that shared similar ORF copy number altera-tions

BB2453

06L3FF02

06L1FF11

AEB Fermol Rouge

J940047

J940557

J940915

Lalvin ICV D254

06L3FF15

BB1235

06L6FF20

Davis Lalvin 522

IOC 18-2007

Lalvin EC-1118

UM218

UM237

S288C

Clu

ste

r 1C

lus

ter 2

Clu

ste

r 3C

lus

ter 4

Page 4 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

20 Kb of the S288C respective chromosome end but onlyfew corresponded to genes with annotated function Clus-ters of depleted ORFs were found in sub-telomeric regionsand contained large percentage of Ty elements and hypo-thetical ORFs (Figure 4) In the group of wine strains upto one third of the observed gene copy number alterationswere found in sub-telomeric regions (Figure 4A) ndash within50 Kb from the S288C chromosome ends using the crite-rion of Edwards-Ingram and colleagues [40] On averagethe clinical strains showed slightly higher percentage(almost 40) of depleted ORFs localized near the telom-eres However this is mostly explained by the massive lossof chromosomes VII and X right arms in strains J940557and J940915 (see karyoscope map of J940915 in Figure

3B) The depletion of ORFs around the centromericregions- within 20 Kb of the centromere according to thecriterion of Schacherer and colleagues [41]- was reducedand only slightly above average in some of the winestrains

A clear differentiation between clinical and wine relatedstrains was observed when the frequency of transposableelements within the ORFs with depleted copy number(Figure 4B) was considered Ty elements comprised 36of the ORFs absent in all wine strains which meant thatapproximately one third of the ORFs associated with ret-rotransposon activity (42 out of a total of 113) in S288Cwere absent in these strains The genomes of the clinical

Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B)Figure 3Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B) In order to visualize the gene copy number altera-tions along chromosomes and to have a global overview of the alterations detected by aCGH the data was plotted along each chromosome using the annotated ORF coordinates of S288C Vertical bars represent the relative hybridization pattern rela-tively to the genome of strain S288C Red bars correspond to amplified ORFs green bars represent deleted ORFs and grey bars are statistically non-significant alterations (FDR lt0279) The horizontal lines indicate the hybridization ratios in logarith-mic scale Signals of repeated ORFs such as those of Ty elements or repetitive sequences flanking Ty element insertion sites correspond to the average signal rather than the individual signal dosage of that sequence Clusters of altered ORFs contained roughly one third of Ty -ORFs and contributed to localization of the chromosome alterations following the genome coordi-nates of the S288C strain

AA BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 5 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 6 of 17(page number not for citation purposes)

Chromosomal location and classification of ORFs with variable copy numberFigure 4Chromosomal location and classification of ORFs with variable copy number Panel-A shows the distribution of var-iable ORFs indicated by CGH-Miner in terms of chromosome location that is sub-telomeric centromeric or chromosome arms Panel-B shows the distribution of the same ORFs in terms of abundance of Ty elements hypothetical or annotated ORFs In both panels each bar represents the total number of variable ORFs for a given strain

0

100

200

300

400

500

600

J940

047

J940

557

J940

915

06L3

FF02

06L1

FF11

06L3

FF15

06L6

FF20

BB12

35

BB24

53

UM

218

UM

237

Lalvi

n EC-1

118

Lalvi

n ICV D

254

IOC-1

8 20

07

AEB F

erm

ol Rouge

Davi

s Lalv

in 5

22

OR

Fs

ch

an

ge

d

Sub-telomeric Centromeric Chromosome arms

0

100

200

300

400

500

600

J94004

7

J940

557

J9409

15

06L3F

F02

06L1F

F11

06L3F

F15

06L6F

F20

BB12

35

BB24

53

UM

218

UM

237

Lalv

in E

C-1

118

Lalvin

ICV D

254

IOC-1

8 200

7

AEB F

erm

ol Roug

e

Dav

is L

alvi

n 52

2

OR

Fs

ch

an

ge

d

Hypothetical Ty elements Annotated

A

B

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

strains however contained most of the Ty elements iden-tified in the genomic sequence of S288C since only 7 ofthe ORFs absent in all three of them were classified in thiscategory

Highly variable genomic regions in wine and clinical strainsThe differences between wine and clinical strains were fur-ther highlighted with the consensus plots obtained fromthe intersection of the individual karyoscope maps of thethirteen wine fermentation strains (Figure 5A) comparedto the three clinical isolates (Figure 5B) These plots repre-sented the relative amount of samples with copy numberalterations showing the percentage of strains with a givenamplification or deletion Most of the deletion clustersidentified in wine strains co-localized with absent Ty ele-

ments (Figure 5A) while in clinical isolates a significantnumber of depletions were associated with sub-telomericinstability (Figure 5B) Variable regions affected differentsub-functional categories of genes in wine and clinicalstrains For example carbohydrate transporters and gly-cosidases were the main functional groups of genes absentin clinical strains but genes involved in vesicle transporttelomere maintenance and alcohol metabolism were alsoidentified among the lost sub-telomeric genes The func-tional categories most affected by gene copy number vari-ability in wine strains were related to vitamin metabolismDNA recombination polysaccharide metabolism regula-tion of meiotic cell cycle and reproduction and were fre-quently located in the vicinity of Ty element insertionsites

The ORF variability patterns differ between wine and clinical strainsFigure 5The ORF variability patterns differ between wine and clinical strains The consensus karyoscope maps obtained for wine (Panel-A) and clinical (Panel-B) strains showed that ORF variation in wine strains was distributed along chromosomes while in clinical strains such distribution was more frequent in sub-telomeric regions The consensus maps show the relative percentage of strains (percent of samples) within wine (Panel-A) and clinical (Panel-B) environments with copy number altera-tion in a given ORF relatively to the reference strain S288C The data was plotted according to ORF chromosome location Red bars represent percentage of strains with amplifications and green bars represent the percentage of strains with deletions according to the grey-line scale shown above and below the chromosome central line Open circles indicate the position of centromeres and blue arrows identify Ty elements

BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 7 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

20 Kb of the S288C respective chromosome end but onlyfew corresponded to genes with annotated function Clus-ters of depleted ORFs were found in sub-telomeric regionsand contained large percentage of Ty elements and hypo-thetical ORFs (Figure 4) In the group of wine strains upto one third of the observed gene copy number alterationswere found in sub-telomeric regions (Figure 4A) ndash within50 Kb from the S288C chromosome ends using the crite-rion of Edwards-Ingram and colleagues [40] On averagethe clinical strains showed slightly higher percentage(almost 40) of depleted ORFs localized near the telom-eres However this is mostly explained by the massive lossof chromosomes VII and X right arms in strains J940557and J940915 (see karyoscope map of J940915 in Figure

3B) The depletion of ORFs around the centromericregions- within 20 Kb of the centromere according to thecriterion of Schacherer and colleagues [41]- was reducedand only slightly above average in some of the winestrains

A clear differentiation between clinical and wine relatedstrains was observed when the frequency of transposableelements within the ORFs with depleted copy number(Figure 4B) was considered Ty elements comprised 36of the ORFs absent in all wine strains which meant thatapproximately one third of the ORFs associated with ret-rotransposon activity (42 out of a total of 113) in S288Cwere absent in these strains The genomes of the clinical

Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B)Figure 3Karyoscope maps of strains Lalvin ICV D254 (A) and J940915 (B) In order to visualize the gene copy number altera-tions along chromosomes and to have a global overview of the alterations detected by aCGH the data was plotted along each chromosome using the annotated ORF coordinates of S288C Vertical bars represent the relative hybridization pattern rela-tively to the genome of strain S288C Red bars correspond to amplified ORFs green bars represent deleted ORFs and grey bars are statistically non-significant alterations (FDR lt0279) The horizontal lines indicate the hybridization ratios in logarith-mic scale Signals of repeated ORFs such as those of Ty elements or repetitive sequences flanking Ty element insertion sites correspond to the average signal rather than the individual signal dosage of that sequence Clusters of altered ORFs contained roughly one third of Ty -ORFs and contributed to localization of the chromosome alterations following the genome coordi-nates of the S288C strain

AA BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 5 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 6 of 17(page number not for citation purposes)

Chromosomal location and classification of ORFs with variable copy numberFigure 4Chromosomal location and classification of ORFs with variable copy number Panel-A shows the distribution of var-iable ORFs indicated by CGH-Miner in terms of chromosome location that is sub-telomeric centromeric or chromosome arms Panel-B shows the distribution of the same ORFs in terms of abundance of Ty elements hypothetical or annotated ORFs In both panels each bar represents the total number of variable ORFs for a given strain

0

100

200

300

400

500

600

J940

047

J940

557

J940

915

06L3

FF02

06L1

FF11

06L3

FF15

06L6

FF20

BB12

35

BB24

53

UM

218

UM

237

Lalvi

n EC-1

118

Lalvi

n ICV D

254

IOC-1

8 20

07

AEB F

erm

ol Rouge

Davi

s Lalv

in 5

22

OR

Fs

ch

an

ge

d

Sub-telomeric Centromeric Chromosome arms

0

100

200

300

400

500

600

J94004

7

J940

557

J9409

15

06L3F

F02

06L1F

F11

06L3F

F15

06L6F

F20

BB12

35

BB24

53

UM

218

UM

237

Lalv

in E

C-1

118

Lalvin

ICV D

254

IOC-1

8 200

7

AEB F

erm

ol Roug

e

Dav

is L

alvi

n 52

2

OR

Fs

ch

an

ge

d

Hypothetical Ty elements Annotated

A

B

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

strains however contained most of the Ty elements iden-tified in the genomic sequence of S288C since only 7 ofthe ORFs absent in all three of them were classified in thiscategory

Highly variable genomic regions in wine and clinical strainsThe differences between wine and clinical strains were fur-ther highlighted with the consensus plots obtained fromthe intersection of the individual karyoscope maps of thethirteen wine fermentation strains (Figure 5A) comparedto the three clinical isolates (Figure 5B) These plots repre-sented the relative amount of samples with copy numberalterations showing the percentage of strains with a givenamplification or deletion Most of the deletion clustersidentified in wine strains co-localized with absent Ty ele-

ments (Figure 5A) while in clinical isolates a significantnumber of depletions were associated with sub-telomericinstability (Figure 5B) Variable regions affected differentsub-functional categories of genes in wine and clinicalstrains For example carbohydrate transporters and gly-cosidases were the main functional groups of genes absentin clinical strains but genes involved in vesicle transporttelomere maintenance and alcohol metabolism were alsoidentified among the lost sub-telomeric genes The func-tional categories most affected by gene copy number vari-ability in wine strains were related to vitamin metabolismDNA recombination polysaccharide metabolism regula-tion of meiotic cell cycle and reproduction and were fre-quently located in the vicinity of Ty element insertionsites

The ORF variability patterns differ between wine and clinical strainsFigure 5The ORF variability patterns differ between wine and clinical strains The consensus karyoscope maps obtained for wine (Panel-A) and clinical (Panel-B) strains showed that ORF variation in wine strains was distributed along chromosomes while in clinical strains such distribution was more frequent in sub-telomeric regions The consensus maps show the relative percentage of strains (percent of samples) within wine (Panel-A) and clinical (Panel-B) environments with copy number altera-tion in a given ORF relatively to the reference strain S288C The data was plotted according to ORF chromosome location Red bars represent percentage of strains with amplifications and green bars represent the percentage of strains with deletions according to the grey-line scale shown above and below the chromosome central line Open circles indicate the position of centromeres and blue arrows identify Ty elements

BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 7 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 6 of 17(page number not for citation purposes)

Chromosomal location and classification of ORFs with variable copy numberFigure 4Chromosomal location and classification of ORFs with variable copy number Panel-A shows the distribution of var-iable ORFs indicated by CGH-Miner in terms of chromosome location that is sub-telomeric centromeric or chromosome arms Panel-B shows the distribution of the same ORFs in terms of abundance of Ty elements hypothetical or annotated ORFs In both panels each bar represents the total number of variable ORFs for a given strain

0

100

200

300

400

500

600

J940

047

J940

557

J940

915

06L3

FF02

06L1

FF11

06L3

FF15

06L6

FF20

BB12

35

BB24

53

UM

218

UM

237

Lalvi

n EC-1

118

Lalvi

n ICV D

254

IOC-1

8 20

07

AEB F

erm

ol Rouge

Davi

s Lalv

in 5

22

OR

Fs

ch

an

ge

d

Sub-telomeric Centromeric Chromosome arms

0

100

200

300

400

500

600

J94004

7

J940

557

J9409

15

06L3F

F02

06L1F

F11

06L3F

F15

06L6F

F20

BB12

35

BB24

53

UM

218

UM

237

Lalv

in E

C-1

118

Lalvin

ICV D

254

IOC-1

8 200

7

AEB F

erm

ol Roug

e

Dav

is L

alvi

n 52

2

OR

Fs

ch

an

ge

d

Hypothetical Ty elements Annotated

A

B

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

strains however contained most of the Ty elements iden-tified in the genomic sequence of S288C since only 7 ofthe ORFs absent in all three of them were classified in thiscategory

Highly variable genomic regions in wine and clinical strainsThe differences between wine and clinical strains were fur-ther highlighted with the consensus plots obtained fromthe intersection of the individual karyoscope maps of thethirteen wine fermentation strains (Figure 5A) comparedto the three clinical isolates (Figure 5B) These plots repre-sented the relative amount of samples with copy numberalterations showing the percentage of strains with a givenamplification or deletion Most of the deletion clustersidentified in wine strains co-localized with absent Ty ele-

ments (Figure 5A) while in clinical isolates a significantnumber of depletions were associated with sub-telomericinstability (Figure 5B) Variable regions affected differentsub-functional categories of genes in wine and clinicalstrains For example carbohydrate transporters and gly-cosidases were the main functional groups of genes absentin clinical strains but genes involved in vesicle transporttelomere maintenance and alcohol metabolism were alsoidentified among the lost sub-telomeric genes The func-tional categories most affected by gene copy number vari-ability in wine strains were related to vitamin metabolismDNA recombination polysaccharide metabolism regula-tion of meiotic cell cycle and reproduction and were fre-quently located in the vicinity of Ty element insertionsites

The ORF variability patterns differ between wine and clinical strainsFigure 5The ORF variability patterns differ between wine and clinical strains The consensus karyoscope maps obtained for wine (Panel-A) and clinical (Panel-B) strains showed that ORF variation in wine strains was distributed along chromosomes while in clinical strains such distribution was more frequent in sub-telomeric regions The consensus maps show the relative percentage of strains (percent of samples) within wine (Panel-A) and clinical (Panel-B) environments with copy number altera-tion in a given ORF relatively to the reference strain S288C The data was plotted according to ORF chromosome location Red bars represent percentage of strains with amplifications and green bars represent the percentage of strains with deletions according to the grey-line scale shown above and below the chromosome central line Open circles indicate the position of centromeres and blue arrows identify Ty elements

BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 7 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

strains however contained most of the Ty elements iden-tified in the genomic sequence of S288C since only 7 ofthe ORFs absent in all three of them were classified in thiscategory

Highly variable genomic regions in wine and clinical strainsThe differences between wine and clinical strains were fur-ther highlighted with the consensus plots obtained fromthe intersection of the individual karyoscope maps of thethirteen wine fermentation strains (Figure 5A) comparedto the three clinical isolates (Figure 5B) These plots repre-sented the relative amount of samples with copy numberalterations showing the percentage of strains with a givenamplification or deletion Most of the deletion clustersidentified in wine strains co-localized with absent Ty ele-

ments (Figure 5A) while in clinical isolates a significantnumber of depletions were associated with sub-telomericinstability (Figure 5B) Variable regions affected differentsub-functional categories of genes in wine and clinicalstrains For example carbohydrate transporters and gly-cosidases were the main functional groups of genes absentin clinical strains but genes involved in vesicle transporttelomere maintenance and alcohol metabolism were alsoidentified among the lost sub-telomeric genes The func-tional categories most affected by gene copy number vari-ability in wine strains were related to vitamin metabolismDNA recombination polysaccharide metabolism regula-tion of meiotic cell cycle and reproduction and were fre-quently located in the vicinity of Ty element insertionsites

The ORF variability patterns differ between wine and clinical strainsFigure 5The ORF variability patterns differ between wine and clinical strains The consensus karyoscope maps obtained for wine (Panel-A) and clinical (Panel-B) strains showed that ORF variation in wine strains was distributed along chromosomes while in clinical strains such distribution was more frequent in sub-telomeric regions The consensus maps show the relative percentage of strains (percent of samples) within wine (Panel-A) and clinical (Panel-B) environments with copy number altera-tion in a given ORF relatively to the reference strain S288C The data was plotted according to ORF chromosome location Red bars represent percentage of strains with amplifications and green bars represent the percentage of strains with deletions according to the grey-line scale shown above and below the chromosome central line Open circles indicate the position of centromeres and blue arrows identify Ty elements

BA

I

II

III

IV

V

VI

VII

VIII

IX

X

XI

XII

XIII

XIV

XV

XVI

Page 7 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

ASP3 and YRF1 gene families differentiate wild-type and laboratory strainsThe karyoscope maps constructed by the moving averagewindow algorithm (implemented in CGH miner) high-lighted genes with copy number alterations Of the 634ORFs identified as altered by the CLAC algorithm in atleast one of the strains significance analysis confirmed270 ORF alterations From these 30 corresponded to Tyelements mostly located in depleted hybridization clus-ters in wine strains (Figure 5) Another third of ORFs withcopy number variation were hypothetical ORFs and werecontiguous to deleted Ty elements in wine strains Anno-tated variable ORFs constituted the remaining third of sig-nificantly altered ORFs and the respective aCGH values(M values) were depicted in Figure 6A for further discus-sion

General trends in copy number alterations included bothamplified and depleted genes Genes with the same trendin copy number alteration across all strains were high-lighted in grey Among the genes depleted in all strainsrelative to the reference strain S288C were four copies oftandemly-repeated cell-wall asparaginase genes (ASP3-1ASP3-2 ASP3-3 and ASP3-4) which are induced inresponse to nitrogen starvation [42] Also ORFsYLR161W YLR156W and YLR159W that code for puta-tive proteins of unknown function and of identicalsequence were depleted in all strains These genes arelocated in the right arm of Chromosome XII near a ribos-omal DNA region in a chromosome locus correspondingto a large cluster of depleted ORFs flanking Ty elementsand were indicated in the karyoscope consensus maps ofboth wine and clinical strains (Figure 5A B)

The relative hybridization values of significantly alteredgenes (Figure 6A) identified another set of homologousgenes whose copy number was altered in all wild-typestrains Indeed five genes of the YRF1 family which arelocated in telomeric Y elements and encode DNA heli-cases (YRF1-2 YRF1-3 YRF1-4 YRF1-5 and YRF1-8) werepresent in the genome of S288C but were depleted or atleast partially depleted in environmental clinical andcommercial strains The missing YRF1 genes were part ofthe depleted ORF clusters located at the telomeres of theright arms of chromosomes V VII XII and XV (Figure 5)together with several hypothetical ORFs

Gene YIL014C-A coding for a putative protein ofunknown function was deleted in all strains except inJ940047 The same was observed for the tandem repeatedgenes ENA1 and ENA2 which code for a P-type ATPasesodium pump involved in the efflux of sodium and lith-ium ions required for salt tolerance which were depletedin all strains except 06L3FF02 Gene ENA5 was onlydepleted in strain J940047

A group of genes with increased copy number in almostall strains was also identified (Figure 6A) Among themIMD1 IMD2 PHO11 PHO12 were signaled by the CLACalgorithm as belonging to clusters of sub-telomeric genesamplified in strains Lalvin EC-1118 Lalvin ICV 254 andUM237 but aCGH values suggested that they could alsobe amplified in other strains (Figure 6A) PHO11 andPHO12 genes which code for acid phosphatases and areinduced by phosphate starvation increased their copynumber by a factor of 3 in strain UM237 relative toS288C while in strain Lalvin EC-1118 the fold increase inhybridization signal was more compatible with its dupli-cation The copy number of the IMD2 gene increased by afactor of 2 in Lalvin EC-1118 and UM237 strains as wellas in strain Lalvin ICV D254 although the latter was nothighlighted as duplicated by CGH-Miner In strain LalvinICV D254 RDS1 (a zinc cluster transcription factorinvolved in resistance to cycloheximide) was part of anamplified cluster of genes on Chromosome III and theaCGH signal indicated a 3-fold increase relative to S288CThe aCGH values for this gene (Figure 6A) further sug-gested that it was also amplified in other strains (see PanelD of Figure 6)

Interestingly both IMD1 and IMD2 genes code for inos-ine monophosphate dehydrogenase and confer resistanceto mycophenolic acid which is produced by the fungusPenicillium stoloniferum and inhibits de novo purine synthe-sis These genes together with RDS1 are involved inresistance to compounds that inhibit eukaryotic cell pro-liferation and increased copy numbers point to a survivaladvantage in competitive ecosystems in presence oforganisms that secrete mycophenolic acid or cyclohex-imide as growth inhibitors

Gene copy number alterations and the origin of strainsSeveral genes showed different copy number alterations inwine and in clinical strains (Figure 6 panels B-F) Forexample the genes HXT9 HXT11 and two ORFs of HXT12(YIL170W and YIL171W) were depleted in most winestrains with the exception of UM218 Lalvin EC-1118 andAEB Fermol Rouge but did not show copy number varia-tion in the clinical strains (Figure 6B) These genes codefor putative hexose transporters which are non-functionalin strain S288C Similar copy number changes across theanalyzed strains were identified for HPF1 and FSP2 whichcode for proteins with glucosidase activity HPF1 codes fora haze-protective mannoprotein that reduces the particlesize of aggregated proteins in white wines while FSP2 isinduced under nitrogen limitation Also the gene ENB1which codes for a trans-membrane iron transporter of themajor facilitator superfamily and is expressed under irondeprivation conditions was included in this group Theabsence of these genes was particularly notorious instrains isolated from the Bairrada region and in the com-

Page 8 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Page 9 of 17(page number not for citation purposes)

Gene copy number alterations show significant inter-strain variabilityFigure 6Gene copy number alterations show significant inter-strain variability Several genes were deleted in all environmen-tal clinical and commercial strains while other genes were amplified in almost all strains relative to the laboratory S288C strain (boxed blocks) However most of the genes showed significant copy number variability between strains Panel-A repre-sents the aCGH values (M values) for genes with altered copy number in at least one of the strains The genes with similar copy number alterations in the wild-type strains relatively to the reference laboratorial strain S288C were highlighted in grey (boxed blocks) The coloured bars next to the gene names identify groups of genes whose relative hybridization value is shown in the line graphs of Panels B-G In these panels grey lines represent the aCGH values of genes indicated on the right hand side of each panel (also shown in the color coded map of Panel-A) and the pink line represents the average aCGH values of the genes (lines)

A

B

C

D

G

E

F

ENB1

HPF1

FSP2

HXT12

HXT11

HXT12

HXT9

FDH1

FIT2

FRE5

PHR1

AAD3

AAD10

AAD6

ADH7

RDS1

ARR3

SDL1

COG3

HSP60

DIP5

MSP1

NCE101

SAT4

SPS22

TAZ1

HXT15

AGP3

DAK2

COS5

MPH2

MPH3

SOR1

SOR2

NFT1

AAD15

FLO10

HMLALPHA1

VBA3

YCL073C

J9

40

04

7

J9

40

55

7

J9

40

91

5

06

L3

FF

02

06

L1

FF

11

06

L3

FF

15

06

L6

FF

20

UM

21

8

UM

23

7

BB

12

35

BB

24

53

La

lvin

EC

-11

18

La

lvin

ICV

D2

54

IOC

18

-20

07

AE

B F

erm

ol

Ro

ug

e

Da

vis

La

lvin

52

2

S2

88

C

- 30 00 30

J9400

47

J9405

57

J9409

15

06L

3F

F0

2

06L

1F

F1

1

06L

3F

F1

5

06L

6F

F2

0

UM

218

UM

237

BB

1235

BB

2453

Lalv

inE

C-1

11

8

Lalv

inIC

V D

254

IOC

18

-2007

AE

B F

erm

ol

Ro

ug

e

Davis

Lalv

in522

S288C

ENA5UIP3ENB1HPF1FSP2HXT12HXT11HXT12HXT9ENA1ENA2MAL11MAL13FLO1FLO5FLO9NFT1PRM9YAR028WYAR029WBDS1PRM8COS6KHS1SEO1VTH1VTH2YIL014C-AASP3-4ASP3-3YLR159WASP3-2ASP3-1YLR156WYLR161WAYT1EXG2PLC1HSP33PAU21FIT3TAR1TDH2ARR1MST27MST28DAN2PAU14PAU15NOC4RRN5DIN7YLR164WCNE1YPS6FDH1FIT2FRE5PHR1AAD3AAD10AAD6ADH7RDS1CIC1IMD1SRD1IMD2PHO11PHO12YHR214WCUP1-1CUP1-2RSC30ARR3SDL1COG3HSP60DIP5MSP1NCE101SAT4SPS22TAZ1HXT15AGP3DAK2COS5MPH2MPH3SOR1SOR2NFT1AAD15FLO10HMLALPHA1VBA3YCL073CYRF1-4YRF1-3YRF1-5YRF1-2YRF1-8

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

mercial strain Davis Lalvin 522 (commonly used in Bair-rada wine musts fermentations)

The genes FDH1 FIT2 FRE5 and PHR1 were absent in theclinical isolates J940557 and J940915 and in strain06L3FF02 FDH1 codes for a formate dehydrogenase FIT2codes for a mannoprotein involved in the retention ofsiderophore-iron in the cell wall FRE5 codes for a puta-tive ferric reductase induced by low iron levels and PHRcodes for a DNA photolyase induced by DNA damage(Figure 6 panel C) These genes do not share a commonfunction but belong to a set of contiguous ORFs locatedin the sub-telomeric region of the right arm of Chromo-some V which was deleted in many of the strains (CLACalgorithm) This cluster of deleted ORFs included FIT3which was also deleted in the above mentioned strainsand in 06L1FF11 but showed aCGH values compatiblewith amplification of copy number in many of the naturalwine strains and in J940047 Although redundant consid-ering the existence of other functional homologues varia-bility in copy number of these genes may affect thecapacity of yeast to thrive in environments with high lev-els of exogenous formaldehyde produced from plantmaterial or from degradation of environmental pollut-ants and in environments with low iron availability

A group of genes coding for proteins with alcohol dehy-drogenase activity were missing in strains J940557 andJ940915 but were amplified in many of the wine strains(Figure 6 panel D) For example ADH7 and AAD3 whichbelong to a cluster of amplified ORFs identified in Chro-mosome III in strain Lalvin ICV D254 by CLAC analysis(Figure 3A) showed increased copy number in strainsBB1235 and Davis Lalvin 522 AAD6 and AAD10 weredepleted in strains J940557 and J940915 Since thesegenes code for proteins involved in ethanol tolerancetheir increase in copy number may confer high resistanceto ethanol which is one of the most important pheno-types for grape must fermentation RDS1 was alsoincluded in this group of genes because of the similar copynumber variation profile within the analyzed strainsprobably due to its location in the same cluster of alteredORFs as ADH7 and AAD3 in Chromosome III

A large set of genes were consistently deleted in the threeclinical isolates (Figure 6A panels E and G) The COG3HSP60 MSP1 NCE101 genes (Figure 6 panel E) areimplicated in protein transport and localization whileARR3 and DIP5 are arsenite and amino acid transportersrespectively These genes that were absent in clinicalstrains were amplified in some of the wine strains namely06L3FF02 06L1FF11 and AEB Fermol Rouge The relativeabundance of these genes in the latter strains explained atleast in part why their genome profiles differed fromthose of the other wine strains (Figure 2)

Other genes were absent in clinical and in many winestrains (Figure 6 panel F) For example genes involved incarbohydrate transport (HXT15 MPH2 and MPH3) andsugar metabolism (SOR1 SOR2) belonged to this cate-gory Strains Lalvin EC-1118 and IOC 18ndash2007 were theonly wine strains with a decreased copy number in AGP3which codes for an amino acid permease and DAK2 par-ticipating in the glycerol catabolic process while strain06L6FF20 constituted an exception within the winestrains since it showed identical copy number of all ofthese genes relatively to strain S288C

Gene copy number signatures for wine must fermentationMany of the variable genes coded for transporters per-meases or flocculation proteins contributing to agenomic signature of the wine strains Among these genesthe maltose transporter gene MAL11 and the MAL-activa-tor protein gene MAL13 were particularly interestingbecause they are important for maltose assimilation andMAL13 is non-functional in the laboratory S288C strain[43] These genes were absent in the commercial winestrains and in some of the environmental isolates Alsoonly some of the commercial wine strains (Lalvin EC-1118 and AEB Fermol Rouge) showed aCGH values com-patible with copy number increase of CUP1-2 Genesinvolved in flocculation mediated by cell wall protein-car-bohydrate interactions namely FLO1 FLO5 FLO9 andFLO10 were in general depleted in the wine strains andalso in two of the clinical strains Evidences for variabilityin copy number of the homologous genes were found inalmost all strains with exception of Lalvin ICV D254where the aCGH values indicated that these genes weredeleted Finally copy number variability among winestrains was also found in genes PAU14 PAU15 and PAU21that code for hypothetical proteins with structural similar-ity to the seripauperin family

DiscussionaCGH profiles grouped yeast strains from different geographical originsThe genome variability uncovered within the environ-mental clinical and commercial strains was pronouncedand did not show any correlation between genome char-acteristics and ecosystem or geographical origin Interest-ingly this study unveiled high genomic similaritybetween the commercial strains and isolates from regionswhere these commercial strains were used in wine mustfermentation For instance strains UM218 and UM237from the Vinho Verde region share a similar genomehybridization profile (Figure 2) and grouped with strainLalvin EC-1118 (Cluster 4) which is frequently used inthe production of sparkling Vinho Verde Strain IOC 18ndash2007 which is used in bottle-fermentations is widelyused in the Bairrada sparkling wine production andgrouped with some of the Bairrada isolates (Cluster 3)

Page 10 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

Close associations were also found between AEB FermolRouge and strains 06L1FF11 06L3FF02 and BB2453from cellars and vineyards of the Bairrada region respec-tively (Cluster 1) Strains 06L6FF20 06L3FF15 andBB1235 are more related to Davis Lalvin 522 than to anyother commercial strain AEB Fermol Rouge isolated inFrance and Davis Lalvin 522 obtained from the USA(Table 1) are both used in the Bairrada cellars to fermentwine musts due to their high fermentation performancewith grapes from this region

The continuous use of commercial yeast strains in largequantities and during successive years may lead to somegenetic homogenization of resident strains that belong tothe autochthonous yeast flora This is supported by thefact that strain UM218 was collected in 2001 300 metersfrom a winery where strain Lalvin EC-1118 was used inthat year whereas strain UM237 was collected in 200320 meters from a winery where the same commercialstrain was used since 2001 [44] However since the Bair-rada isolates come from cellars where no commercialstrains were used and the vineyards were not located inthe vicinity of cellars (A C Gomes personal communica-tion) a contamination with commercial yeast does notexplain the similarities between environmental isolatesand some commercial strains Instead the winemakersexperience for the production of wines with the mostdesirable expression of the regions terroir may indirectlybe responsible for this genomic resemblance By favoringphenotypic characteristics in the autochthonous strainswhich adapted during centuries of positive selection byfarming and wine producing activities the wine producermay have selected commercial yeast according to the samerequired phenotypes and this resemblance in phenotypeis reflected to some extent in similar gene copy numbercharacteristics

Large scale genome alterationsS cerevisiae strains that were mainly obtained from wine-making environment are homothallic mostlyhomozygous (65) with low to high (gt 85) sporula-tion ability and are predominantly diploid [3145-47]Aneuploid strains have also been described [4224849]Although no evidence for polyploidy was detected in thestrains analyzed in this study the karyoscope maps of theclinical isolates J940557 and J940915 (see Figure 3B)showed a systematic depletion in hybridization signalthroughout the entire length of chromosomes III VII andX This depletion was only statistically significant at theright arm-ends ruling out the hypothesis that these strainswere aneuploid for the referred chromosomes Also noevidence for aneuploidy was found in the other strains

One explanation for the karyoscopes of strains J940557and J940915 may be the presence of heterologous chro-

mosomes originated from a different strain albeit similarto S cerevisiae S288C The latter hypothesis is supportedby the frequent occurrence of Saccharomyces sp hybridswhich originate from mating of different Saccharomycesspecies that form viable although sterile zygotes [31]These hybrids are sometimes selected for industrial beeror wine fermentations due to phenotypic advantages[325051] but inter-specific hybrids were also found indiverse spontaneous fermentations [3352] However thestrains used in our study including the clinical isolateswere not hybrids (Additional File 1 Figure S1) Thereforedifferences in chromosome copy number that is struc-tural heteromorphisms of chromosomes III VII and Xmay be the explanation for the obtained patterns of rela-tive hybridization

Specific trends of genome instability distinguished wine and clinical isolatesConsensus maps highlighted the most variable regions ofthe yeast genome relatively to the laboratory S288Cstrain High variability was associated with the sub-telom-eric regions of some chromosomes irrespective of thestrains origin In particular a large fraction of strainseither from wine or clinical background had ORF dele-tions located at the right end of Chromosome I left endof Chromosome VI and right end of chromosomes VIIand X (Figure 5) Comparative genome studies performedby Winzeler and colleagues [29] showed that inter-speciesgenome variability is biased toward sub-telomericregions where genes related to carbon source metabolismand transport are located Apparently telomere variabilityis important for adaptation to new environments and todifferent metabolic sources to overcome environmentalstress

Retrotransposons are also known as regions of highgenome diversity between yeast strains and species[5354] Our data further supports the hypothesis thatthey may be selected for generating genomic variability inresponse to environmental stimuli since they wereaffected differently in wine and clinical strains (Figure 5)Wine must fermentation isolates differed dramatically inTy element composition from the reference laboratorialstrain as indicated by the relative absence of about onethird of these elements together with ORFs flanking theirinsertion sites On the other hand clinical strains weresimilar to S288C in composition of Ty element and Ty ele-ment associated ORFs

This raised the question of whether clinical strains andS288C share a common ancestor or whether the reducednumber of Ty elements and the flanking genes in the winestrains could result from selective pressures that affect par-ticular regions of the genome in response to adaptation toparticular environments Genome comparison between

Page 11 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the clinical isolate YJM789 and S288C undertaken by Weiand colleagues [28] showed a close association betweenrepeat sequences sites in S288C and deleted regions inYJM789

A comparison of the consensus plots of the strains used inthis study with the sequences of the chromosomes ofYJM789 showed that the clusters of depleted ORFs asso-ciated with Ty elements in the former (Figure 5A) werealso missing in the YJM789 genome (see Additional File 3Figures S3AndashS3P) On the other hand the consensus plotobtained for the clinical strains coincided with the com-parative analysis carried out by Wei and colleagues [28]In particular in the sub-telomeric variability found in sev-eral chromosomes and in the clusters of deleted ORFsassociated with Ty element insertion sites in chromo-somes X and XII (Figure 5B) Since YJM789 is not a winefermenting strain but is associated to a pathogenic phe-notype the resemblance of genomic alterations related toTy element content between this strain and those isolatedfrom vineyards was surprising and deserves further study

Variability in copy number of transposable elements hasbeen previously reported within the Hemiascomycetousyeast clade [55] and particularly in this group of organ-isms [40] This is in line with previous studies that showedlow copy number of these elements relative to S288C inwild-type wine lager and flor yeast [395657] as wellas in laboratory strains other than S288C [29] The highvariability in Ty element composition observed in thisstudy supports the hypothesis that retrotransposition isrelevant for adaptation at least in the Saccharomyces sensustricto clade Sequences flanking transposable elementsgenerate variability in the yeast genome [6] probably dueto the ectopic recombination involving Ty element repet-itive sequences since Ty elements play a role in the mobi-lization of genome fragments throughout the genomeresulting frequently in chromosomal rearrangements andgene duplications [58] Variability of Ty element contentis supported by retrotransposon activity loss or acquisi-tion in some lineages of Saccharomyces sensu stricto corre-lating with speciation according to Liti and colleagues[59] Interestingly these authors carried out a populationsurvey of LTR-retrotransposons in the Saccharomyces sensustricto complex and showed that Ty elements may beabsent in groups of geographical isolates and are oftenlost or horizontally transferred in an apparent homeo-static control of the total number of repetitive elements

Gene copy number alterations differentiate laboratorial and environmental strainsThe absence of the tandem repeated ASP3 region locatedin Chromosome XII as well as the ENA region of Chro-mosome IV were observed both in wine and clinicalstrains (Figure 6) Nevertheless such deletions were

found in various strains and are not specific of wine orclinical phenotypes [2940566061] Variability in copynumber of the ASP3 genes was found in a comparativehybridization study of 9 strains isolated from ale and lagerbrewing fermentations suggesting that the presence orabsence of these genes could discriminate fermentativeyeast [62] However the absence of these genes in thewine yeasts analyzed in this study including commercialand clinical isolates only discriminated the laboratorialstrain S288C from the environmental isolates indicatingthat the use of asparagine as an alternative nitrogen sourceis not important in natural niches from were the wild-typestrains were isolated

Depletion of the genes of the YRF1 family also differenti-ated the environmental from the laboratorial referencestrain These DNA helicase genes are induced in strainswith deficient telomerase activity as part of a mechanismof telomere rescue [63] These genes are present in sub-tel-omeric Y elements but seem to be dispensable since astudy on the survival and fitness of an S paradoxus isolatewithout telomerase and in the absence of Y elements wassimilar to that of other well characterized strains [59]Although Y elements are conserved between strains andspecies of Saccharomyces [6465] copy number of theYRF1 family of genes varied remarkably in the group ofstrains surveyed in this study (Figure 6)

Variation in fermentation related genesGenes involved in maltose metabolism namely MAL11and MAL13 were depleted relative to S288C in most com-mercial wine strains while some of the non-commercialisolates showed deletion of both or just one of thesegenes In a similar genomic profile study of commercialwine strains Dunn and colleagues [39] found intra-straincopy number variation of MAL11 and MAL13 genes anddid not include them in their commercial wine yeastgenome signature This signature also highlighted thedeletion of two copies of CUP1 relatively to the genomeof S288C However they found variability in copynumber of these genes within isolates of the same com-mercial wine strain Similarly these genes were deleted inmost of the wine related strains studied here but not inBB2453 and AEB Fermol Rouge in which CUP1-2 wasapparently amplified These genes code for a protein thatbinds copper and mediates resistance to high concentra-tions of copper and cadmium and this locus is variablyamplified in different yeast strains [6667] and is notexclusive of wine fermentation strains

Variability of copy number among the wine strains wasalso found in genes of the seripauperin family namelyPAU14 PAU15 and PAU21 These genes are encodedmainly in subtelomeric regions and are strongly regulatedby anaerobiosis during alcoholic fermentation [6869]

Page 12 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

They play a role in sterol lipid transport and hence theirrelevance for ethanol tolerance during anaerobic growthWhile PAU21 was depleted in all wine strains relative toS288C PAU14 and PAU15 copy number increased insome of the wine strains as well as in all clinical isolatesbut decreased in others namely in isolates whosegenomic profiles were similar to those of AEB FermolRouge PAU15 is involved in the response to toxins [70]but no function is yet known for PAU14

ConclusionOur data showed that telomeric recombination and Tyelement insertion were the main genome diversity fea-tures of the clinical commercial and environmental iso-lates used in this study Among the variable genes mostlydepleted were genes involved in metabolic functionsrelated to cellular homeostasis or transport of differentsolutes such as ions sugars and metals Clusters ofdepleted ORFs also contained ribosomal proteins generaltranscription factors and transcription activators transla-tion initiation factors helicases and zinc-finger genes

In the clinical strains genes associated to pathogenesisnamely those involved in pseudohyphal growth and inva-siveness did not show copy number alterations This con-firmed previous studies by Klingberg and colleagues [71]who did not find specific virulence factors separating clin-ical from non-clinical yeast strains Despite this Llanosand colleagues [27] found that clinical isolates have a typ-ical pathogenic phenotype when compared with indus-trial yeasts For example secretion of proteases andphospholipases growth at 42degC and pseudohyphalgrowth are more pronounced in clinical isolates Somestrains isolated from infections may be opportunistic col-onizers of the human environment and not commensalsof humans Indeed a recent survey of 92 yeast invasiveinfections revealed that 50 of them were caused by Sac-charomyces boulardii which is used as a probiotic prepara-tion for the treatment of antibiotic-related diarrhea [72]

The genomic variability found in this study supportedother studies showing duplication and deletion of sub-tel-omeric genes involved in secondary metabolism linked toenvironmental adaptation [73] In this study the sub-tel-omeric genes whose copy number changed namely MALFLO HXT PHO IMD SOR PAU FIT and ARR familygenes did not identify specific alterations of environmen-tal or of commercial wine strains Also our list of geneswith variable copy number showed differences with thecommercial wine yeast signature published by Dun andcolleagues [39] However some of the genes identified inour study confirmed the trend of genome alterations ofwine strains highlighted by the commercial wine yeastsignature For example the IMD and PHO genes wereamplified and the MAL genes were deleted in both stud-

ies Genome variability associated with retrotransposonmobility was characteristic of wine strains Thereforethese variability mechanisms may have a positive impacton the fitness of strains during colonization of new envi-ronment(s) In other words this present study highlightsthe usefulness of yeast as a model system to studygenomic variability in the context of environmental andevolutionary genomics

MethodsStrains and culture conditionsA list of strains used in this study is provided in Table 1together with information about the respective originWine strains were isolated from fermenting musts in winecellars from the Bairrada wine region and from vineyardsof Bairrada and Vinho Verde wine regions according toValero et al [44] Commercial wine strains were kindlyprovided by Adega Cooperativa da Bairrada CantanhedePortugal Clinical isolates were a kind gift of Prof MickTuite from the University of Kent Canterbury UK

Yeast strains were cultivated in 5 ml YEPD (1 yeastextract 2 peptone 2 glucose) at 30degC with 185 rpmagitation until cell density reached 108ndash109 cellsml har-vested and washed three times with distilled water by cen-trifugation for 5 minutes at 3000 g resuspended in 200 μllysis buffer (2 Triton X-100 1 SDS 100 mM NaCl 1mM EDTA 10 mM Tris pH 80) and stored at -80degC untilused for DNA extraction

DNA isolationDNA was isolated as described by Hoffman and Winston[74] with some adaptations For DNA extraction 200 μlof 25241 phenolchloroformisoamyl alcohol wereadded to the cells resuspendend in lysis buffer (seeabove) together with 300 mg of acid-washed glass beads(~500 μm diameter) The mixture was vortexed for 10minutes before the addition of 200 μl of TE buffer (10mM Tris-HCl 1 mM EDTA pH 80) Following a 5 minutecentrifugation at 14000 rpm (Eppendorf centrifuge 5415R) for phase separation the DNA in the aqueous phasewas precipitated with 1 ml ethanol (96 vv) at roomtemperature resuspended in 400 μl of TE containing 60mg RNaseA (GE Healthcare) and incubated at 37degC for 6hours The DNA was precipitated with 10 μl of 4 Mammonium acetate and 1 ml of room temperature abso-lute ethanol and resuspended in 400 μl of TE buffer Pro-teins were removed by treating DNA samples with 10 μgProteinase K (Roche) and incubating overnight at 50degCFinally the DNA was collected by precipitation with 20 μlof 30 M sodium acetate pH 52 and 500 μl of ethanol(96 vv) at -20degC followed by incubation at -80degC for1 hour After centrifugation at 14000 rpm for 20 minutesat 4degC the final DNA pellet was carefully washed with

Page 13 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

500 μl of ethanol (70 vv) left to dry at room tempera-ture and resuspended in 200 μl of TE buffer

Genotyping analysisThe PCRRFLP analysis of the MET2 gene was performedas described by Antunovics et al [75] and of OPY1 KIN82MET6 KEL2 and CYR1 genes as described by Gonzaacutelez etal [3334] PCR analysis of delta sequences was per-formed as described [7677]

Microarray productionFor the production of in-house spotted DNA-microarrays6388 70 mer oligonucleotides targeting the ORFeome ofSaccharomyces cerevisiae (OPERON Yeast AROS v11 col-lection Qiagen) were spotted twice on CodeLink acti-vated slides (GE Healthcare) according to the slidemanufacturers instructions using a MicroGrid CompactII spotter (GenomicSolutions) A set of ten different 70mer probes designed from Escherichia coli genomesequence with less than 70 homology to S cerevisiaegenome was also included in the microarray in order tomonitor non-specific hybridization The array design andspotting protocol were deposited in ArrayExpress data-base [78] under the accession code A-MEXP-1185

Labelling and hybridizationFor labelling 4 μg of gDNA were digested with 20 units ofDpnII (New England Biolabs USA) to yield fragmentsbetween 250 and 3000 bp The fragmented DNA was pre-cipitated with 25 volumes of ethanol (96 vv) at -20degC Genomic DNA was fluorescently labelled using theULS arrayCGH labelling kit from Kreatech (Kreatech TheNetherlands) which is a non-enzymatic protocol thatallows direct labelling of unmodified genomic DNABriefly 1 μl of Cy3-ULS or Cy5-ULS was added to 2 μg ofpreviously digested DNA together with 2 μl of the 10 timeslabelling solution provided with the kit The sample vol-ume was adjusted to 20 μl with DNAse-free water and thelabelling reaction was promoted by incubating the sampleat 85degC for 30 minutes The excess ULS-dye was removedusing the KREA pure columns following the kit manufac-turers instructions The degree of labelling (DoL) corre-sponding to the percentage of labelled nucleotides wasdetermined by measuring the absorbance at 260 nm andat 550 nm for ULS-Cy3 labelled DNA or at 260 nm and650 nm for ULS-Cy5 labelled DNA Samples with DoLbetween 10 and 20 were routinely obtained

For comparative genome hybridization ULS-Cy3 labelledDNA from each of the environmental commercial andclinical strains was combined with ULS-Cy5 labelled DNAfrom strain S288C Dye-swap hybridizations were per-formed for each strain To ensure microarray data baselinerobustness differentially labelled DNA from the S288Cstrain were co-hybridized in a total of six self-self experi-

ments and used as controls The mixture of the requiredCy3 and Cy5 labelled samples was adjusted to 208 μl withDNAse-free water mixed with 52 μl Agilent 10times blockingAgent and 260 μl of Agilent 2times Oligo aCGH HybridizationSolution (Agilent USA) and incubated at 95degC for 3 min-utes After a short spin-down the labelled DNA mixturewas applied to a microarray slide pre-hybridized asdescribed by van de Peppel and colleagues [79] assem-bled in a SureHyb hybridization chamber fitted with agasket slide (Agilent) and incubated for 20 hours at 65degCin a hybridization oven (HIR10M Grant Boekel) with 5rpm rotation speed

Slides were washed as described in the Agilent Oligonu-cleotide Array-based CGH for Genomic DNA Analysisprotocol [80] Briefly the microarray and gasket slidewere disassembled inside a staining dish containing 250ml of Oligo aCGH Wash Buffer 1 and the slides (up to 4)were washed in fresh 250 ml of Oligo aCGH Wash Buffer1 solution at room temperature during 5 minutes withgentle agitation from a magnetic stirrer A second washstep was carried out by immersing the slides in OligoaCGH Wash Buffer 2 solution previously warmed to37degC during 1 minute also with gentle magnetic stirringFinally slides were dried by centrifugation at 800 rpm for3 minutes

Image acquisition and data processingImages of the microarray hybridizations were acquiredusing the Agilent G2565AA microarray scanner The fluo-rescence intensities were quantified with QuantArray v30software (PerkinElmer) Using BRB-ArrayTools v340software [81] manually flagged bad spots were elimi-nated and the local background was subtracted beforeaveraging the replicate features on the array Log2 intensityratios (M values) were then Median normalized to correctfor differences in genomic DNA labelling efficiencybetween samples The raw data as well as the processed(filtered and Median normalized) data for all hybridiza-tions was submitted to the ArrayExpress database and isavailable under the accession code E-MTAB-29

Data analysisThe relative hybridization signal of each ORF was derivedform the average of the two dye-swap hybridizations per-formed for each strain The normalized log2 ratio (Mvalue) was considered as a measure of the relative abun-dance of each ORF relatively to that of the reference strainS288C Deviations from the 11 hybridization ratio weretaken as indicative of changes in DNA copy numberAlthough depleted hybridization may be due to sequencedivergence as well as nucleotide deletions sequence diver-gence of a given ORF relatively to that of S288C wouldhave to be higher than 30 in order to impair the hybrid-ization with 70 mer oligonucleotides [8283] Given that

Page 14 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

the variability usually observed between Saccharomycesgenomes (either within laboratorial strains or natural iso-lates) is much lower than this estimate [113241] weinterpreted statistically significant depletions in hybridi-zation signal as ORF deletions We further confirmed thatthe set of ORFs discussed in Figure 6 were targeted by therespective microarray probe with at least 93 homologyin the genomes of the S cerevisiae strains S288C RM11-1a(wine isolate) and YJM789 (clinical isolate) Thisexcluded the hypothesis of depleted hybridization signaldue to extreme gene variability or partial deletion of thetargeted sequence (Additional File 4 Table S1)

The aCGH hybridization patterns were used to investigatethe genomic relatedness of the wild-type yeast strainsHierarchical cluster analysis (Pearson correlation averagelinkage) was performed using MeV from TM4 softwaresuit [84] with the normalized and dye-swap averagedaCGH profiles The reproducibility of the clustering anal-ysis was confirmed by repeating the paired dye-swaphybridizations for one of the sixteen strains randomlyselected and by verifying that independent assays for thesame strain grouped together in the dendrogram (notshown)

Karyoscope maps were generated for each strain usingCGH-Miner [85] For data smoothing the parameterswere set for BAC analysis to produce a moving window ofthree ORFs for averaging the hybridization signal SixS288C independent self-self hybridizations were used forbase line correction The average data of the six self-selfS288C hybridizations was used to ascertain the baselinenoise in deriving karyoscope maps and is shown in Addi-tional File 2 Figure S2Q Summary plots also denomi-nated consensus plots depicting the relative percentage ofsamples showing a particular alteration were obtained bycombining the required individual karyoscope mapsusing the same software

Multi-class significance analysis (SAM) was done usingthe algorithm implemented in MeV from TM4 softwareThe individual hybridizations were used as the input datain a total of two dye-swap hybridizations for each strainclass except for strain S288C which was represented bythe five least variable self-self hybridizations of the set ofsix performed for CLAC analysis SAM analysis indicatedthe ORFs with significant copy number alteration in atleast one of the strains with a FDR (90th percentile) of0336 Functional annotations and GO terms associationwas done following the Saccharomyces Genome Database(SGD) annotations [86]

AbbreviationsaCGH array Comparative Genome Hybridization FDRFalse Discovery Rate SAM Significance Analysis of Micro-

arrays SGD Saccharomyces Genome Database SNP Sin-gle Nucleotide Polimorphism

Competing interestsThe authors declare that they have no competing interests

Authors contributionsLC performed experiments helped in the experimentaldesign supervised the experiments performed data anal-ysis and wrote the manuscript MFE performed the exper-iments and data analysis ACG isolated identified andgenotyped the environmental Bairrada wine yeast strainsPP contributed to data analysis DS participated in thedesign of the study provided the Vinho Verde strains andperformed the simple sequence repeats analysis MAScoordinated the study wrote and proofread the manu-script All authors read and approved the final manu-script

Additional material

AcknowledgementsThe authors wish to thank Adega Cooperativa da Bairrada Cantanhede Por-tugal for providing the commercial strains The clinical strains were a kind

Additional File 1Hybrid detection by PCR-RFLP analysis PCR-RFLP of five distinct loci located in different chromosomesClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S1pdf]

Additional File 2CGH Miner karyoscope maps Karyoscope maps of the 16 wild-type strains analyzed (S2A-S2P) and the baseline karyoscope obtained for strain S288C (S2Q)Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S2pdf]

Additional File 3Genome alterations between YJM789 and S288C and this study Genome alterations found in the consensus plot obtained for the wine strains are compared of the clinical strain YJM789 whose genome is fully sequencedClick here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S3pdf]

Additional File 4Confirmation of microarray probe targets for a selected group of ORF Specificity and homology of a set of microarray probes using BLAST align-ment against the genomes of S cerevisiae strains S288C RM11-1a and YJM789Click here for file[httpwwwbiomedcentralcomcontentsupplementary1471-2164-9-524-S4xls]

Page 15 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

gift of Prof Mick Tuite from the University of Kent-UK This work was funded by Fundaccedilatildeo para a Ciecircncia e Tecnologia through projects FEDERFCT POCIAGR561022004 and CONC-REEQ7372001

References1 Bidenne C Blondin B Dequin S Vezinhet F Analysis of the chro-

mosomal DNA polymorphism of wine strains of Saccharomy-ces cerevisiae Curr Genet 1992 221-7

2 Carro D Garcia-Martinez J Perez-Ortin JE Pina B Structural char-acterization of chromosome I size variants from a naturalyeast strain Yeast 2003 20171-183

3 Longo E Vezinhet F Chromosomal rearrangements duringvegetative growth of a wild strain of Saccharomyces cerevi-siae Appl Environ Microbiol 1993 59322-326

4 Nadal D Carro D Fernandez-Larrea J Pina B Analysis anddynamics of the chromosomal complements of wild spar-kling-wine yeast strains Appl Environ Microbiol 1999651688-1695

5 Adams J Puskas-Rozsa S Simlar J Wilke CM Adaptation andmajor chromosomal changes in populations of Saccharomy-ces cerevisiae Curr Genet 1992 2213-19

6 Dunham MJ Badrane H Ferea T Adams J Brown PO Rosenzweig Fet al Characteristic genome rearrangements in experimen-tal evolution of Saccharomyces cerevisiae Proc Natl Acad Sci USA2002 9916144-16149

7 Perez-Ortin JE Querol A Puig S Barrio E Molecular characteri-zation of a chromosomal rearrangement involved in theadaptive evolution of yeast strains Genome Res 2002121533-1539

8 Aa E Townsend JP Adams RI Nielsen KM Taylor JW Populationstructure and gene evolution in Saccharomyces cerevisiaeFEMS Yeast Res 2006 6702-715

9 Ayoub MJ Legras JL Saliba R Gaillardin C Application of MultiLocus Sequence Typing to the analysis of the biodiversity ofindigenous Saccharomyces cerevisiae wine yeasts from Leba-non J Appl Microbiol 2006 100699-711

10 Ben-Ari G Zenvirth D Sherman A Simchen G Lavi U Hillel J Appli-cation of SNPs for assessing biodiversity and phylogenyamong yeast strains Heredity 2005 95493-501

11 Fay JC Benavides JA Evidence for domesticated and wild pop-ulations of Saccharomyces cerevisiae PLoS Genet 2005 166-71

12 Naumov GI Naumova ES Sniegowski PD Saccharomyces para-doxus and Saccharomyces cerevisiae are associated with exu-dates of North American oaks Can J Microbiol 1998441045-1050

13 Sniegowski PD Dombrowski PG Fingerman E Saccharomyces cer-evisiae and Saccharomyces paradoxus coexist in a naturalwoodland site in North America and display different levelsof reproductive isolation from European conspecifics FEMSYeast Res 2002 1299-306

14 Suh SO McHugh JV Pollock DD Blackwell M The beetle gut ahyperdiverse source of novel yeasts Mycol Res 2005109261-265

15 Martini A Ciani M Scorzetti G Direct enumeration and isola-tion of wine yeasts from grape surfaces American Journal of Enol-ogy and Viticulture 1996 47435-440

16 Mortimer R Polsinelli M On the origins of wine yeast Res Micro-biol 1999 150199-204

17 Lopes CA van BM Querol A Caballero AC Saccharomyces cere-visiae wine yeast populations in a cold region in ArgentineanPatagonia A study at different fermentation scales J ApplMicrobiol 2002 93608-615

18 Sabate J Cano J Querol A Guillamon JM Diversity of Saccharo-myces strains in wine fermentations analysis for two consec-utive years Lett Appl Microbiol 1998 26452-455

19 Schuller D Alves H Dequin S Casal M Ecological survey of Sac-charomyces cerevisiae strains from vineyards in the VinhoVerde Region of Portugal FEMS Microbiol Ecol 2005 51167-177

20 Valero E Cambon B Schuller D Casal M Dequin S Biodiversity ofSaccharomyces yeast strains from grape berries of wine-pro-ducing areas using starter commercial yeasts FEMS Yeast Res2007 7317-329

21 Frezier V Dubourdieu D Ecology of yeast strain Saccharomycescerevisiae during spontaneous fermentation in a Bordeauxwinery American Journal of Enology and Viticulture 1992 43375-380

22 Legras JL Merdinoglu D Cornuet JM Karst F Bread beer andwine Saccharomyces cerevisiae diversity reflects human his-tory Mol Ecol 2007 162091-2102

23 Schuller D Casal M The genetic structure of fermentativevineyard-associated Saccharomyces cerevisiae populationsrevealed by microsatellite analysis Antonie Van Leeuwenhoek2007 91137-150

24 Hazen KC New and emerging yeast pathogens Clin MicrobiolRev 1995 8462-478

25 Aucott JN Fayen J Grossnicklas H Morrissey A Lederman MMSalata RA Invasive infection with Saccharomyces cerevisiae report of three cases and review Rev Infect Dis 199012406-411

26 Zerva L Hollis RJ Pfaller MA In vitro susceptibility testing andDNA typing of Saccharomyces cerevisiae clinical isolates J ClinMicrobiol 1996 343031-3034

27 de Llanos R Fernandez-Espinar MT Querol A A comparison ofclinical and food Saccharomyces cerevisiae isolates on thebasis of potential virulence factors Antonie Van Leeuwenhoek2006 90221-231

28 Wei W McCusker JH Hyman RW Jones T Ning Y Cao Z et alGenome sequencing and comparative analysis of Saccharo-myces cerevisiae strain YJM789 Proc Natl Acad Sci USA 200710412825-12830

29 Winzeler EA Castillo-Davis CI Oshiro G Liang D Richards DRZhou Y et al Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays Genetics 2003 16379-89

30 Daran-Lapujade P Daran JM Kotter P Petit T Piper MD Pronk JTComparative genotyping of the Saccharomyces cerevisiae lab-oratory strains S288C and CENPK113-7D using oligonucle-otide microarrays FEMS Yeast Res 2003 4259-269

31 Mortimer RK Evolution and variation of the yeast (Saccharo-myces) genome Genome Res 2000 10403-409

32 Masneuf I Hansen J Groth C Piskur J Dubourdieu D New hybridsbetween Saccharomyces sensu stricto yeast species foundamong wine and cider production strains Appl Environ Microbiol1998 643887-3892

33 Gonzalez SS Barrio E Querol A Molecular characterization ofnew natural hybrids of Saccharomyces cerevisiae and S kudri-avzevii in brewing Appl Environ Microbiol 2008 742314-2320

34 Gonzalez SS Barrio E Gafner J Querol A Natural hybrids fromSaccharomyces cerevisiae Saccharomyces bayanus and Saccha-romyces kudriavzevii in wine fermentations FEMS Yeast Res2006 61221-1234

35 Lopandic K Gangl H Wallner E Tscheik G Leitner G Querol A etal Genetically different wine yeasts isolated from Austrianvine-growing regions influence wine aroma differently andcontain putative hybrids between Saccharomyces cerevisiaeand Saccharomyces kudriavzevii FEMS Yeast Res 2007 7953-965

36 Masneuf I Murat ML Naumov GI Tominaga T Dubourdieu DHybrids Saccharomyces cerevisiae times Saccharomyces bayanusvar uvarum having a high liberating ability of some sulfurvarietal aromas of Vitis vinifera Sauvignon blanc wines JournalInternational Des Sciences De La Vigne Et Du Vin 2002 36205-212

37 Liti G Louis EJ Yeast evolution and comparative genomicsAnnu Rev Microbiol 2005 59135-153

38 CGH-Miner calling gains and losses in array CGH data usingthe CLAC method [httpwww-statstanfordedu~wp57CGH-Miner]

39 Dunn B Levine RP Sherlock G Microarray karyotyping of com-mercial wine yeast strains reveals shared as well as uniquegenomic signatures BMC Genomics 2005 653

40 Edwards-Ingram LC Gent ME Hoyle DC Hayes A Stateva LI OliverSG Comparative genomic hybridization provides newinsights into the molecular taxonomy of the Saccharomycessensu stricto complex Genome Res 2004 141043-1051

41 Schacherer J Ruderfer DM Gresham D Dolinski K Botstein DKruglyak L Genome-wide analysis of nucleotide-level varia-tion in commonly used Saccharomyces cerevisiae strains PLoSONE 2007 2e322

42 Bon EP Carvajal E Stanbrough M Rowen D Magasanik B Aspara-ginase II of Saccharomyces cerevisiae GLN3URE2 regulationof a periplasmic enzyme Appl Biochem Biotechnol 1997 63ndash65203-212

Page 16 of 17(page number not for citation purposes)

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References

BMC Genomics 2008 9524 httpwwwbiomedcentralcom1471-21649524

43 Charron MJ Dubin RA Michels CA Structural and functionalanalysis of the MAL1 locus of Saccharomyces cerevisiae MolCell Biol 1986 63891-3899

44 Valero E Schuller D Cambon B Casal M Dequin S Disseminationand survival of commercial wine yeast in the vineyard alarge-scale three-years study FEMS Yeast Res 2005 5959-969

45 Bakalinsky AT Snow R Conversion of Wine Strains of Saccha-romyces cerevisiae to Heterothallism Appl Environ Microbiol1990 56849-857

46 Guijo S Mauricio JC Salmon JM Ortega JM Determination of therelative ploidy in different Saccharomyces cerevisiae strainsused for fermentation and flor film ageing of dry sherry-type wines Yeast 1997 13101-117

47 Barre P Vezinhet F Dequin S Blondin B Genetic improvement ofwine yeasts In Wine Microbiology and Biotechnology Edited by FleetGH London Harwood Academic Publishers 1992265-289

48 Codon AC Benitez T Korhola M Chromosomal reorganizationduring meiosis of Saccharomyces cerevisiae bakers yeastsCurr Genet 1997 32247-259

49 Puig S Querol A Barrio E Perez-Ortin JE Mitotic recombinationand genetic changes in Saccharomyces cerevisiae during winefermentation Appl Environ Microbiol 2000 662057-2061

50 Kishimoto M Fermentation characterstics of hybrids betweenthe cryophilic wine yeast Saccharomyces bayanus and themesophilic wine yeast Saccharomyces cerevisiae J Fermen Bio-eng 1994 77432-435

51 Sinohara TK Saito K Yanagida F Goto S Selection and hybridiza-tion of the wine yeasts for improved winemaking propertiesJ Fermen Bioeng 1994 77428-431

52 Christine lJ Marc L Catherine D Claude E Jean-Luc L Michel A etal Characterization of natural hybrids of Saccharomyces cer-evisiae and Saccharomyces bayanus var uvarum FEMS Yeast Res2007 7540-549

53 Fischer G James SA Roberts IN Oliver SG Louis EJ Chromo-somal evolution in Saccharomyces Nature 2000 405451-454

54 Garfinkel DJ Genome evolution mediated by Ty elements inSaccharomyces Cytogenet Genome Res 2005 11063-69

55 Neuveglise C Feldmann H Bon E Gaillardin C Casaregola SGenomic evolution of the long terminal repeat retrotrans-posons in hemiascomycetous yeasts Genome Res 200212930-943

56 Bond U Neal C Donnelly D James TC Aneuploidy and copynumber breakpoints in the genome of lager yeasts mappedby microarray hybridisation Curr Genet 2004 45360-370

57 Infante JJ Dombek KM Rebordinos L Cantoral JM Young ETGenome-wide amplifications caused by chromosomal rear-rangements play a major role in the adaptive evolution ofnatural yeast Genetics 2003 1651745-1759

58 Dujon B Yeasts illustrate the molecular mechanisms ofeukaryotic genome evolution Trends Genet 2006 22375-387

59 Liti G Peruffo A James SA Roberts IN Louis EJ Inferences of evo-lutionary relationships from a population survey of LTR-ret-rotransposons and telomeric-associated sequences in theSaccharomyces sensu stricto complex Yeast 2005 22177-192

60 Lashkari DA DeRisi JL McCusker JH Namath AF Gentile C HwangSY et al Yeast microarrays for genome wide parallel geneticand gene expression analysis Proc Natl Acad Sci USA 19979413057-13062

61 Primig M Williams RM Winzeler EA Tevzadze GG Conway ARHwang SY et al The core meiotic transcriptome in buddingyeasts Nat Genet 2000 26415-423

62 Pope GA MacKenzie DA Defernez M Aroso MA Fuller LJ MellonFA et al Metabolic footprinting as a tool for discriminatingbetween brewing yeasts Yeast 2007 24667-679

63 Yamada M Hayatsu N Matsuura A Ishikawa F Y-Help1 a DNAhelicase encoded by the yeast subtelomeric Y element isinduced in survivors defective for telomerase J Biol Chem1998 27333360-33366

64 Louis EJ Haber JE The structure and evolution of subtelomericY repeats in Saccharomyces cerevisiae Genetics 1992131559-574

65 Pryde FE Louis EJ Saccharomyces cerevisiae telomeres Areview Biochemistry (Mosc) 1997 621232-1241

66 Karin M Najarian R Haslinger A Valenzuela P Welch J Fogel S Pri-mary structure and transcription of an amplified genetic

locus the CUP1 locus of yeast Proc Natl Acad Sci USA 198481337-341

67 Winge DR Nielson KB Gray WR Hamer DH Yeast metal-lothionein Sequence and metal-binding properties J BiolChem 1985 26014464-14470

68 Rachidi N Martinez MJ Barre P Blondin B Saccharomyces cerevi-siae PAU genes are induced by anaerobiosis Mol Microbiol2000 351421-1430

69 Rossignol T Dulau L Julien A Blondin B Genome-wide monitor-ing of wine yeast gene expression during alcoholic fermenta-tion Yeast 2003 201369-1385

70 Iwahashi H Kitagawa E Suzuki Y Ueda Y Ishizawa YH Nobumasa Het al Evaluation of toxicity of the mycotoxin citrinin usingyeast ORF DNA microarray and Oligo DNA microarrayBMC Genomics 2007 895

71 Klingberg TD Lesnik U Arneborg N Raspor P Jespersen L Com-parison of Saccharomyces cerevisiae strains of clinical andnonclinical origin by molecular typing and determination ofputative virulence traits FEMS Yeast Res 2008 8631-640

72 Enache-Angoulvant A Hennequin C Invasive Saccharomycesinfection a comprehensive review Clin Infect Dis 2005411559-1568

73 Pryde FE Huckle TC Louis EJ Sequence analysis of the right endof chromosome XV in Saccharomyces cerevisiae an insightinto the structural and functional significance of sub-telom-eric repeat sequences Yeast 1995 11371-382

74 Hoffman CS Winston F A ten-minute DNA preparation fromyeast efficiently releases autonomous plasmids for transfor-mation of Escherichia coli Gene 1987 57267-272

75 Antunovics Z Irinyi L Sipiczki M Combined application of meth-ods to taxonomic identification of Saccharomyces strains infermenting botrytized grape must J Appl Microbiol 200598971-979

76 Legras JL Karst F Optimisation of interdelta analysis for Sac-charomyces cerevisiae strain characterisation FEMS MicrobiolLett 2003 221249-255

77 Schuller D Valero E Dequin S Casal M Survey of molecularmethods for the typing of wine yeast strains FEMS MicrobiolLett 2004 23119-26

78 ArrayExpress [httpwwwebiacukmicroarray-asaerae-main[0]]

79 Peppel J van de Kemmeren P van BH Radonjic M van LD HolstegeFC Monitoring global messenger RNA changes in externallycontrolled microarray experiments EMBO Rep 20034387-393

80 Agilent Oligonucleotide Array-based CGH for GenomicDNA Analysis [httpwwwchemagilentcomLibraryusermanualsPublicG4410-90010_CGH_Protocol_v5pdf]

81 Biometric Research Branch-ArrayTools [httplinusncinihgovBRB-ArrayToolshtml]

82 Hughes TR Shoemaker DD DNA microarrays for expressionprofiling Curr Opin Chem Biol 2001 521-25

83 He Z Wu L Li X Fields MW Zhou J Empirical establishment ofoligonucleotide probe design criteria Appl Environ Microbiol2005 713753-3760

84 Saeed AI Sharov V White J Li J Liang W Bhagabati N et al TM4a free open-source system for microarray data manage-ment and analysis Biotechniques 2003 34374-378

85 Wang P Kim Y Pollack J Narasimhan B Tibshirani R A method forcalling gains and losses in array CGH data Biostatistics 2005645-58

86 Saccharomyces Genome Database [httpwwwyeastgenomeorg]

87 Mortimer RK Johnston JR Genealogy of principal strains of theyeast genetic stock center Genetics 1986 11335-43

Page 17 of 17(page number not for citation purposes)

  • Abstract
    • Background
    • Results
    • Conclusion
      • Background
      • Results
        • Strains and overview of genomic variability
        • Genome variability is associated with Ty elements and telomeres
        • Highly variable genomic regions in wine and clinical strains
        • ASP3 and YRF1 gene families differentiate wild-type and laboratory strains
        • Gene copy number alterations and the origin of strains
        • Gene copy number signatures for wine must fermentation
          • Discussion
            • aCGH profiles grouped yeast strains from different geographical origins
            • Large scale genome alterations
            • Specific trends of genome instability distinguished wine and clinical isolates
            • Gene copy number alterations differentiate laboratorial and environmental strains
            • Variation in fermentation related genes
              • Conclusion
              • Methods
                • Strains and culture conditions
                • DNA isolation
                • Genotyping analysis
                • Microarray production
                • Labelling and hybridization
                • Image acquisition and data processing
                • Data analysis
                  • Abbreviations
                  • Competing interests
                  • Authors contributions
                  • Additional material
                  • Acknowledgements
                  • References