Vineyard yeast project

VINEYARD YEAST PROJECT

University of Padova (Padua), Department of Biology, Department of Agronomy, Food, Natural Resources, Animals and the Environment (DAFNAE)

Final results are reported in these papers

Treu L, Toniolo C, Nadai C, Sardu A, Giacomini A, Corich V, Campanaro S. The impact of genomic variability on gene expression in environmental Saccharomyces cerevisiae strains. Environ Microbiol. 2013".

Treu L*, Campanaro S*, Nadai C, Toniolo C, Nardi T, Giacomini A, Valle G, Blondin B, Corich V. Oxidative stress response and nitrogen utilization are strongly variable in Saccharomyces cerevisiae wine strains with different fermentation performances. Appl Microbiol Biotechnol. 2014 Apr 4. [Epub ahead of print] (*equally contributing)

Sardu A, Treu L, Campanaro S. Transcriptome structure variability in Saccharomyces cerevisiae strains determined with a newly developed assembly software. BMC Genomics. 2014 Dec 15:1045.

Nadai C, Treu L, Campanaro S, Giacomini A and Corich V. Different mechanisms of resistance modulate sulfite tolerance in wine yeasts. Applied Microbiology and Biotechnology. 2015;215:49-56.

Genome sequences of the vineyard S. cerevisiae yeast strains reported in the first paper (P283, R008, R103, P301) were deposited in the NCBI database and are freely accessible:

S. cerevisiae P283 contigs row_reads

S. cerevisiae R008 contigs row_reads

S. cerevisiae R103 contigs row_reads

S. cerevisiae P301 contigs row_reads

Data described in this web page were obtained from the poster presented at the "25th INTERNATIONAL CONFERENCE ON YEAST GENETICS AND MOLECULAR BIOLOGY 11th-16th July 2011, Olsztyn-Kortowo, Poland".

A COMPARATIVE GENOMIC AND TRANSCRIPTOMIC STUDY ON ECOTYPICAL S. cerevisiae STRAINS

(1) Department of Agricultural Biotechnology, University of Padua, Viale dell'Università, 16, Legnaro (PD) 35020, ITALY; (3) CRIBI Biotechnology Centre, Department of Biology - University of Padua, phone: +39 049 8276306, mail: stefano.campanaro@unipd.it; (2) Centro Interdipartimentale per la Ricerca in Viticoltura ed Enologia (CIRVE) - University of Padua

INTRODUCTION

Natural selection combined with domestication applies selective pressures on yeast genome producing a large number of different strains with specialized phenotypes. During the last decades thousand of strains have been phenotypically characterized but correlation between phenotype and genotype is not yet completely unveiled. Genome sequences analysis is a crucial step to obtain a general description of genes content and to highlight differences between yeasts. In this work four ecotypical Saccharomyces cerevisiae strains isolated from Venetian vineyards have been sequenced using Next Generation sequencing producing high-quality assemblies and a variety of tools have been implemented to solve the complex task of genome finishing. Furthermore a global analysis of gene expression have been performed using SOLiD RNA-seq followed by a comparison between differences in promoter regions and their downstream effect on expression. Results show a higher influence of tandem repeat variability respect to mutations on transcription factor binding sites. Finally, a general overview of gene expression in different winemaking and laboratory strains (performed at 6g/l and 45g/l of CO2 produced) revealed a transcriptional fingerprint characterizing oenological strains adaptation to their stressful environment.

RESULTS

Statistics

Genome sequencing was performed using 454-FLX with a mixed approach of shotgun sequencing and 8 kb paired-end reads. Assembly have been performed using Newbler (454 Roche) software and then refined with a bioinformatic finishing (local reassemblies). Gene finding and annotation was performed using a mixed approach: firstly we transferred the genes from the S288c annotation using the RATT software, secondly genes absent in the reference strain S288c were identified using a de-novo procedure (GeneMark software) and added to the previous annottaion (Table 1). Genomes were aligned using MAUVE software and results has been used to visualize genome comparisons. From this alignment the translocations and the genomic regions absent in non-vineyard yeast strains were identified (Figure 1) and the number of SNPs between strains were calculated to determine a phylogenetic tree (Figure 2).

Table 1.

Whole genome alignments and rearrangements

Figure 1.

The more relevant translocations identified are discussed in a separate web page

Neighbor joining tree representing genetic distances between strains were determined from SNPs present in whole genome alignments using Phylip package and reported in Figure 2. Ecotypical strains (purple) are clustered together and with oenological strains. Brewing (yellow), bioethanol (green) and laboratory (blu) strains are shown. Phylogenetic distances were determined more accurately from the MAUVE output considering also the heterozygous genomic positions and results were reported in the final paper.

Figure 2.

Fermentation kinetics

Numerous phenotypic parameters were determined for the strains. The most important phenotypic traits are those related to the fermentationprocess. Strains have been grown in synthetic wine media on controlled bioreactors in winemaking condition to provide gene expression profiles during fermentation (Figure 3). Behaviors are strongly different for the laboratory strains and the strain R103 that are slow fermenting, strain R008 that has intermediate performance and strains P283, P301 and EC1118 that are fast fermenting strains. RNA has been extracted from samples collected at specific point of the curve, corresponding at 6 and 45 g/l of CO2 produced. Sampling points are reported in Figure 3. Growth curves were determined in different media, in Figure 4 results obtained in synthetic wine must are reported. The more relevant compounds for the winemaking were determined at three fermentation steps using HPLC (6, 45 and 80 g/l of CO2 produced) and are reported in Figure 5.

Figure 3.

Physiological validation

Extensive phenotypic test (i.e. growth curves on various media -Figure 4- and HPLC analysis -Figure 5-) have been performed to confirm data obtained from genomic comparison and transcription data.

Figure 4.

Figure 5.

SOLiD RNA-seq

In the two steps of the fermentation process RNA was extracted and the transcriptome was sequenced using SOLiD technology. The aim was to determine the influence of the genomic variability on gene expression and to determine the final effect on the phenotypic characters determined. In the Figure 6 the reads obtained from RNA-seq were aligned on the genomic region of one strain with PASS and visualized with Artemis software. Notice the strong expression differences among genes.

Figure 6.

Figure 7.

Gene expression comparison of 6000 orthologuos features in the six considered strains led to the identification of an enormous variability in their expression, but interesting only 140 genes have a strong difference in gene expressed in oenological strains respect to s288c. In Figure 7 the behaviour of these genes in the six strains in the two conditions analyzed was represented using Jexpress.

Gene Ontology analysis, performed with GOMiner, indicates an over-representation of molecular functions typical of the winemaking process. In fact deletion of those genes over-expressed in oenological strains in S288c led to a reduction of efficiency in processes relevant for fermentation, as listed here:

Table 2.

Comparison analysis

Differences in promoter regions including mutations in transcription factor (TF) binding sites and tandem repeat (TR) sequences. TF binding sites mutating more frequently than expected were calculated with a hypergeometric distribution and genes regulated by them were typically enriched in genes required for adaptation to selective pressures (p<0.05). Mutated TRs are quite similar among sequences with different TR unit lengths, anyway those composed by repetitions of di- and tri-nucleotides resulted significantly more mutated in intergenic regions than in other genomic positions.

Figure 8.

Figure 9.

Correlation between mutations and differences in transcription profiles of the regulated genes. Distributions of log2 expression ratios calculated for differentially expressed genes among classes created considering promoter sequences mutations show similar means. Expected true differences in variance are difficult to detect due to low sample sizes of some classes. Statistical tests suggests a higher influence of TR variability respect to mutation on TF binding sites on differential gene expression (p<0.05). Moreover, percentages of highly differentially expressed genes in classes regulated by tandem TRs with 10% or higher differences in repeat length are higher than those of classes with less mutated sequences.