Current Microbiology

, Volume 58, Issue 3, pp 211–218

Polymorphisms of Saccharomyces cerevisiae Genes Involved in Wine Production

Authors

  • Ileana Vigentini
    • Dipartimento di Scienze e Tecnologie Alimentari e MicrobiologicheUniversità degli Studi di Milano
  • Daniela Fracassetti
    • Dipartimento di Scienze e Tecnologie Alimentari e MicrobiologicheUniversità degli Studi di Milano
  • Claudia Picozzi
    • Dipartimento di Scienze e Tecnologie Alimentari e MicrobiologicheUniversità degli Studi di Milano
    • Dipartimento di Scienze e Tecnologie Alimentari e MicrobiologicheUniversità degli Studi di Milano
Article

DOI: 10.1007/s00284-008-9310-x

Cite this article as:
Vigentini, I., Fracassetti, D., Picozzi, C. et al. Curr Microbiol (2009) 58: 211. doi:10.1007/s00284-008-9310-x

Abstract

The setting up of new molecular methods for Saccharomyces cerevisiae typing is valuable in enology. Actually, the ability to discriminate different strains in wine making can have a benefit both for the control of the fermentation process and for the preservation of wine typicity. This study focused on the screening of single-nucleotide polymorphisms in genes involved in wine production that could evolve rapidly considering the selective pressure of the isolation environment. Preliminary screening of 30 genes in silico was performed, followed by the selection of 10 loci belonging to 8 genes. The sequence analysis showed a low polymorphism and a degree of heterozygosity. However, a new potential molecular target was recognized in the TPS1 gene coding for the trehalose-6-phosphate synthase enzyme involved in the ethanol resistance mechanism. This gene showed a 1.42% sequence diversity with seven different nucleotide substitutions. Moreover, classic techniques were applied to a collection of 50 S. cerevisiae isolates, mostly with enologic origin. Our results confirmed that the wine making was not carried out only by the inoculated commercial starter because indigenous strains of S. cerevisiae present during fermentation were detected. In addition, a high genetic relationship among some commercial cultures was found, highlighting imprecision or fraudulent practices by starter manufacturers.

Introduction

Saccharomyces cerevisiae have been used for a long time in food and alcoholic beverage production. In the enologic field, many studies have reported the development of molecular tools for typing yeasts aimed at recognizing a specific strain unambiguously. Currently, discrimination at the strain level becomes a strategic activity for the wine industry because it can link territory, environment, and final products for wine valorization [21]. Moreover, the characterization at the molecular level improves knowledge concerning yeast biodiversity and the dynamics of microbial population during wine production, verifies the starter domination, and allows fast detection of potential spoiling microorganisms [5, 8, 19].

Among the techniques with high discrimination power, multilocus sequence typing (MLST) permits identification at the strain level throughout the analysis of single-nucleotide polymorphisms (SNPs) in loci present in housekeeping genes [11] and characterized by low evolutionary rates [6]. The setting up of allelic profiles in databases shared on Web sites would allow a rapid and reliable recognition of genotypes.

To date, only few reports have discussed the application of this method for enologic microorganisms belonging to the Oenococcus oeni and S. cerevisiae species [13, 10, 15]. Although the perspectives are encouraging, databases still are poorly represented [15]. Therefore, accessibility and comparison of results by the Internet, representing the major advantage, are not possible.

For a deep investigation of the phylogenetic relationships among strains belonging to the same species, some authors have proposed analysis of the loci enclosed in genes that can evolve more rapidly than the housekeeping genes. A molecular target more sensitive to the environmental pressure allows a better determination of the differences among the evolutionary lines. This is an advantage as a decrease in the number of loci to be analyzed [4]. In particular, S. cerevisiae genome investigation is a current practice used to search for a new polymorphic molecular target that allows strain characterization by simple protocols [15].

This study aimed to explore the nucleotide variability of S. cerevisiae genome by screening of SNPs in loci selected on genes encoding for functional proteins involved in wine making, considering the potential mutagen environment. A collection of starters normally distributed on the Italian market and wild yeasts isolated in the Franciacorta area (Brescia, Italy) were investigated. Results from classic molecular methods for typing strains were used to compare the relevant discriminatory power with the one obtained from analysis of single nucleotide polymorphism.

Materials and Methods

Yeast Strains and Media

To isolate and maintain all the isolates YPD agar medium [18] supplemented with 0.1 g/l chloramphenicol was used. Yeasts were conserved at −80°C in YPD medium with 20% (v/v) glycerol. The collection, composed of 50 yeast isolates from different sources (Table 1), included 33 commercial starters, 14 wine isolates, and 3 laboratory yeast strains. In particular, 21 yeast isolates belonging to the collection and listed in Table 1 from 1 to 79E were collected during the sparkling wine production in 10 different wineries of the Franciacorta area. Eight of these were reisolated from commercial starters usually used in these wineries, and 13 came from wine samples at the end of alcoholic fermentation 3 to 4 weeks later.
Table 1

Detail of Saccharomyces cerevisiae isolates used in this study

Strain

Geographic origin

Type of culture

mtDNA-RFLP profilea

δ Sequence amplificationb

Genotype

CLS1

France

CS

26

11

1

CLS2

California

CS

24

1

2

CLS4

France

CS

22

10

3

CLS6

Chile

CS

2

21

4

CLS8

France

CS

22

10

3

CLS9

Uruguay

CS

13

8

5

CLR12

France

CS

1

8

6

CLS13

France

CS

1

8

6

CLS15

Unknown

CS

14

8

7

CLS16

France

CS

4

23

8

CLS17

Unknown

CS

12

15

9

CLS18

South Africa

CS

6

23

10

CLS19

South Africa

CS

10

20

11

CLS20

South Africa

CS

10

20

11

CLS21

Unknown

CS

10

22

12

CLS23

South Africa

CS

14

22

13

CLS25

South Africa

CS

14

22

13

CMR40

France

CS

25

3

14

CMR45

New Zealand

CS

3

6

15

CMR46

France

CS

5

22

16

CMR47

France

CS

16

24

17

CMR48

Unknown

CS

14

17

18

CMR49

Unknown

CS

21

32

19

CMR50

Unknown

CS

20

4

20

CMR52

Unknown

CS

20

5

21

GRF18

Unknown

LY

25

19

22

W303

Unknown

LY

12

17

23

CENPK

Unknown

LY

25

18

24

CMR3

Italy

WI

3

16

25

1

Italy

CS* (cellar A,B)

24

30

26

2

Italy

CS* (cellar C)

18

31

27

3

Italy

CS* (cellar D, E)

17

7

28

4

Italy

CS* (cellar F)

15

27

29

5

Italy

CS* (cellar G)

12

27

30

6

Italy

CS* (cellar A, H, I)

23

27

31

7.1

Italy

CS* (cellar A, L)

27

29

32

7.2

Italy

CS* (cellar L)

19

14

33

3E

Italy

WI (cellar A)

8

27

34

5G

Italy

WI (cellar C)

8

27

34

13H

Italy

WI (cellar D)

7

12

35

13I

Italy

WI (cellar D)

8

28

36

28A

Italy

WI (cellar H)

8

28

36

28B

Italy

WI (cellar H)

7

2

37

32A

Italy

WI (cellar E)

7

12

35

32E

Italy

WI (cellar E)

9

13

38

18C

Italy

WI (cellar L)

9

9

39

18U

Italy

WI (cellar L)

9

33

40

51C

Italy

WI (cellar F)

11

26

41

78C

Italy

WI (cellar B)

9

25

42

79E

Italy

WI (cellar G)

9

25

42

CS commercial starter, LY laboratory yeast strain, WI wine isolate, CS* commercial starter used in the Franciacorta area

aProfile obtained by mtDNA-RFLP analysis (Fig. 1a)

bProfile obtained by δ sequence amplification (Fig. 1b)

Gene Selection

The following 30 genes were subjected to preliminary analysis in silico for assessment of their genetic variability: ACO1/YLR304C (aconitase); ADH1/YOL086C and ADH2/YMR303C (alcohol dehydrogenase); ADR1/YDR216W (alcohol dehydrogenase regulator); ALD6/YPL061 W (aldehyde dehydrogenase); ATH1/YPR026 W (acid trehalase); ATP1/YBL099 W (energy production); BTN2/YGR142 W (arginine uptake); CAR1/YPL111 W (arginine degradation); CWP2/YKL096 W-A (pH resistance); GPD1/YDL022W (glycerol-3-phosphate dehydrogenase); HSP26/YBR072 W,HSP30/YCR021C, and HSP104/YLL026W (stress responsive proteins); HXT1/YHR094C (hexose transporter); IntAY (genomic organization [2]); MDH2/YOL126C (malate dehydrogenase); MET4/YNL103 W (sulfur amino acid pathway); MPR1/YFR004 W (N-acetyl transferase); NTH1/YDR001C and NTH2/YBR001C (neutral trehalase); PDC1/YLR044C (pyruvate decarboxylase); PGM2/YMR105C (phosphoglucomutase); RHR2/YIL053W (DL-glycerol-3-phosphatase); SED1/YDR077W (cell wall organization and biogenesis, mitochondrial genome maintenance); SUC2/YIL162 W (sucrose invertase); TPS1/YBR126C (trehalose-6-phosphate synthase); TPT1/YOL102C (tRNA replication); UBI4/YLL039C (ubiquitin); and YGR012W (cysteine synthase).

Because few sequences of the same genes from different strains are deposited online, we made alignments with identified orthologous genes of other yeast species phylogenetically close to S. cerevisiae such as S. paradoxus, S. mikatae, S. bayanus, S. castelli, S. kluyveri, and S. kudriavzevii. The related amino acid sequences were compared to find the most variable regions throughout ClustalW (www.yeastgenome.org) [7, 16].

DNA Extraction and Polymerase Chain Reaction Amplification

Yeasts were grown overnight in liquid YPD medium with shaking. From a 5-ml culture, DNA was extracted as described by Querol et al. (1992) [22] using 500 μg/ml of Zymolyase 100T (USBiological, Swampscott, MA, USA) as the lytic enzyme. The primers used in the amplification reaction and the respective annealing temperatures are listed in Table 2. The concentrations of the reagents were 10 ng of DNA, 0.1 μmol of each primer, 200 μmol of dNTPs, 1× reaction buffer (MgCl2 free), 2.5 mmol of MgCl2, and 1 U of Taq polymerase (Expand High Fidelity System, Roche, Germany) in a final volume of 25 μl.
Table 2

List of genes, primers, and genetic characteristics of the 10 loci

Locus

Systematic name

Gene product

Primer Forward Reverse

Primer sequence (5′–3′)

Annealing temp. (ºC)

Expected amplicon sizes (bp)

Sequenced fragment (bp)

No. of polymorphic sites (nt position)

No. of heterozygous sites

ScADR1

YDR216W

Alcohol dehydrogenase regulator

961

ATTCTTCCGGCTCACCAACC

57

606

513

5 (1,147, 1,239, 1,251, 1,285, 1,365)

5

1567

TGGCTTGAACATCTCTGGTC

3448

GTATGCTTGTCCCTGGTT

51

494

417

0

0

3942

AGTGTGGAAATCACTTGCC

ScGPD1

YDL022W

Glycerol-3-phosphate dehydrogenase

169

GTAAGGGATACCCAGAAG

51

568

414

2 (558, 564)

0

737

CAACCTAAGGCAACAACG

ScHSP30

YCR021C

Heat-shock protein

57

CCACATGGCCTGGATATGCAC

62

883

660

3 (148, 240, 531)

0

939

TCCAGATGCTCTTGGGCTCTC

ScHSP104

YLL026W

Heat-shock protein

110

TGCCTTCATTGAAACGCC

46

547

345

3 (273, 480, 510)

0

657

GGTCTTACCGATACCT

1214

CGCAAGAGATTCTAAGCC

51

648

480

1 (1516)

1

1862

TTTTACCGGAACCGGAC

ScPGM2

YMR105C

Phospho-glucomutase

677

GAACGGTGTAACTGGACCATAC

64

538

423

0

0

1215

GGCCAAGATGTTCAACCACGCC

ScRHR2

YIL053W

DL-glycerol-3-phosphatase

137

CGAACACGTTATTCACATCTCTCAC

54

556

514

0

0

669

ATCGGTTTCAGCGTTGTATTCACCG

ScTPS1

YBR126C

Trehalose-6-phosphate synthase

283

TCCATTACCATCCTGGTGAG

48

567

492

7

0

888

TCTGAGGCACACCTTTGATG

YGR012W

Cysteine synthase

628

CAGGCCCAGAAATTGCCCATC

58

480

374

0

0

1080

CACAATGTTTGAGCCGTGAG

Amplification reactions consisted of an initial denaturation for 2 min at 95°C followed by 30 cycles of 1 min each at 95°C, 45 s at a temperature depending on the set of the primer (Table 2), 1 min at 72°C, and a final elongation for 10 min at 72°C. Polymerase chain reaction (PCR) products were visualized on agarose gels 1% (w/v), then purified and sequenced on both ends by Primm (Milan, Italy). Alignments by ClustalX software (ftp://ftp.ebi.ac.uk/pub/software/unix/clustalx) (EBI tools) were carried out to evaluate differences between S. cerevisiae strains. The frequency value of nucleotide substitution was calculated as the ratio between the number of SNPs and the total number of screened nucleotide, whereas the heterozygosity frequency was calculated as the ratio between the number of heterozygous SNPs and the total number of SNPs. The sequences for each strain and gene were deposited in GenBank (www.ncbi.nlm.nih.gov/Genbank/) with accession numbers from UE924646 to UE924743.

Mitochondrial and δ-Sequence Fingerprinting

Genomic DNA was restricted as indicated by Comi et al. [9] using 20 U of HinfI endonucleases (Fermentas International Inc., Burlington, Canada) in a final volume of 20 μl according to the manufacturer. Then DNA fragments were completely separated at 100 V for 3 h in a 0.7% (w/v) agarose gel. The δ-sequence fingerprinting was performed as described by Ness et al. [20]. The generated restriction and amplified patterns were visualized on agarose gels at 1% (w/v) and elaborated by Gel Compare software version 3.5 (Applied Maths; Sint-Martens-Latem, Belgium) using the unweighted pair group method with arithmetic (UPGMA) mean. Restriction profiles (mtDNA-RFLP) and amplification patterns (δ-PCR) were normalized to those of the CMR45 isolate used as a reference. The repeatability threshold of the protocols was determined as the percentage value of similarity joining the electrophoretical patterns of the replicates obtained from various independent assays with DNA samples of the same isolate (CMR45).

Results and Discussion

New Molecular Targets for Yeast Typing

The genetic biodiversity of a S. cerevisiae collection was assessed using traditional protocols for strain typing and investigation of new genetic markers with single-nucleotide polymorphisms (SNPs). Recent papers report the usefulness of the latter analysis for studying the evolution of a microbial population [13]. The housekeeping genes are considered stable genetic targets because no nonsynonymous mutations usually are accumulated. However, the nucleotide analysis of other genes that evolve faster due to the exposition to selective environments may permit a more accurate subtype discrimination [4]. Moreover, Aa et al. [1] and Jubany et al. (2008) [15] demonstrated that genes with accepted enologic significance can well describe the population structure of the natural isolates of S. cerevisiae.

A preliminary bioinformatic analysis showed that the 30 genes screened are sufficiently preserved. Subsequently, only the genes showing long regions with variability into the amino acid sequence were selected to increase the probability of finding differences in the nucleotide strings.

Single-nucleotide polymorphisms were investigated considering 10 loci in the ADR1, GPD1, HSP30, HSP104, PGM2, RHR2, TPS1, and YGR012 W genes (Table 2), which also included constitutively expressed genes. We designed primers to amplify a region 300 to 600 bp in length. In particular, two couples of primers were drawn for the longest genes, HSP104 and ADR1. The DNA amplicons, corresponding to a total length of 4,632 bp, were sequenced, aligned, and analyzed. Only eight strains in the S. cerevisiae collection were subjected to preliminary analysis: five commercial starters (CLS2, CLS21, CMR3, CMR40, CMR52) and three laboratory strains. The three laboratory strains (CENP K, GRF18, W303) were used as reference haploid yeasts.

A total of 21 SNPs found within the 10 gene fragments were tested, and 6 of these were heterozygous sites. Both the frequency of the nucleotide substitution (0.45%) and the heterozygosity state of alleles detected at the polymorphic sites (28.6%) were lower than what was already reported for S. cerevisiae (~2–3%) and C. albicans (~60–80%) [2, 23].

The distribution of the heterozygous polymorphisms sites among the pool of the investigated genes was nonhomogeneous. They were focused only in HSP104 and ADR1 loci (Table 2) and for no more than two different isolates, CLS21 and CMR52, respectively. The maintenance of biologic protein activity is related to the probability that neutral (or silent) mutations can be accumulated in the genome.

Most of the SNPs led to synonymous changing of the amino acids, with three exceptions: (1) in the HSP104 sequence, a switch between arginine (basic) and isoleucine (hydrophobic) was the result of the polymorphism at position 1,516 only for the isolate CLS21; (2) in the ADR1 gene, isoleucine was substituted by leucine (hydrophobic) (position 1,147) and (3) serine (hydrophilic) by arginine (position 1,365) only for the isolate CMR52. Disaccharide trehalose is the major factor in the protection of endocytoses from ethanol in S. cerevisiae, and mutant cells unable to synthesize it did not develop resistance [17].

The ClustalW alignment showed that the locus into the TPS1 gene was the most variable out of the eight strains analyzed. When only the forward direction was sequenced, an overestimated number of polymorphic sites was recognized, although a high-fidelity Taq polymerase was used. The real polymorphism along the locus consisted of seven different nucleotide substitutions (1.42%) (Table 2).

Using the TPS1locus as a molecular target, the study was extended to the whole collection. The alignment of the sequences gathered the strains into four clusters (Table 3). Genotype 1 grouped together the three reference haploid strains and S288C (www.yeastgenome.org). Genotypes 2 and 3 were the most represented, with 23 and 22 isolates, respectively. Genotype 4 consisted of only a single isolate (CMR40). Extending the research to other genes (not defined as housekeeping), it may be possible to type the yeast collection at the strain level, and in the future to set up a S. cerevisiae database devoted to wine strains [15].
Table 3

Position of polymorphic nucleotide sites and genotypes identified for locus TPS1a

Genotype

543

588

615

681

690

705

813

1 (4)

G

C

G

C

A

C

A

2 (23)

G

T

A

T

G

T

A

3 (22)

G

T

A

C

G

T

A

4 (1)

A

C

A

C

G

C

G

aThe numbers of isolates with the same genotype are indicated in brackets

Molecular Characterization of Yeast Collection

In a second approach, the genetic similarities among the 50 isolates were investigated combining mitochondrial DNA restriction analysis (mtDNA-RFLP) and amplification of δ-sequences (δ-PCR) (Fig. 1). Throughout a comparison of the banding profiles of the same isolate (CMR45) obtained in different electrophoretic gels, a repeatability of 98.8% was calculated for mtDNA-RFLP and that of 98.6% for δ-PCR. The isolates showing a similarity level higher than the repeatability value were put together and marked with numbers (Table 1). The mtDNA-RFLP showed a similarity level among the isolates ranging from 67% to 100% (Fig. 1a), whereas a higher level of homology was detected by δ-PCR (>75%) (Fig. 1b), confirming the high discriminatory power of these techniques [22].
https://static-content.springer.com/image/art%3A10.1007%2Fs00284-008-9310-x/MediaObjects/284_2008_9310_Fig1_HTML.gif
Fig. 1

Dendrograms established by the GelCompare (Applied Maths, Sint-Martens-Latem, Belgium) software package basis on a mtDNA-RFLP profiles and b δ-PCR profiles of Saccharomycescerevisiae yeast isolates

The collection was grouped into 27 clusters by mtDNA-RFLP and into 33 clusters by δ-PCR analyses. Table 1 shows the genotypes obtained by cluster combination, which corresponded to a potential single strain. In the current study, 42 genotypes were recognized. In particular, eight genotypes, each consisting of two isolates, were considered clones of the same strain. Genotypes 3 and 6 included isolates collected in France but distributed by the same company for different technology aims. The same was found for genotypes 10 and 12 even if the identical ecologic origin was declared.

Because the analysis did not discriminate each commercial starter as a single strain, we supposed that a high genetic relation among yeast cultures furnished by the same producers can occur or else inaccuracy and fraudulent practices are widespread by companies, as mentioned by Fernándes-Espinar et al. [12]. A risk for a gradual leveling of wine quality and a loss of biodiversity are the drawbacks that occur when the same strains often named with different marks are used. Sensory properties of wine were shown to be enhanced by the presence of indigenous yeast species and naturally occurring strains of S. cerevisiae during fermentation [13]. Moreover, it could not be assumed that inoculated strains of S. cerevisiae would inhibit the growth of indigenous strains of S. cerevisiae that also occur in the wine must [14].

Actually, our study suggests that the typicity of sparkling wine produced in the Franciacorta area could be related to the presence of different yeast strains. Franciacorta DOCG wineries control the wine-making process through the use of commercial starters. Yet, we ascertained that at least a strain different from those added at the beginning of the process was found at the end of fermentations in all the samples.

Acknowledgment

We are grateful to the wineries registered at “Consorzio per la tutela del Franciacorta DOCG” that collaborated in this study.

Copyright information

© Springer Science+Business Media, LLC 2008