Genome-Wide Analysis of Nucleotide-Level Variation in Commonly Used Saccharomyces cerevisiae Strains

doi:10.1371/journal.pone.0000322

PLoS ONE. 2007; 2(3): e322.

Published online 2007 March 28. doi: 10.1371/journal.pone.0000322

PMCID: PMC1829191

Copyright Schacherer et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Genome-Wide Analysis of Nucleotide-Level Variation in Commonly Used Saccharomyces cerevisiae Strains

Joseph Schacherer,^1,² Douglas M. Ruderfer,^1,² David Gresham,^1,³ Kara Dolinski,^1,³ David Botstein,^1,³ and Leonid Kruglyak^1,²^*

¹Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, New Jersey, United States of America

²Department of Ecology and Evolutionary Biology, Princeton University, Princeton, New Jersey, United States of America

³Department of Molecular Biology, Princeton University, Princeton, New Jersey, United States of America

Justin Fay, Academic Editor

Washington University in St. Louis School of Medicine, United States of America

* To whom correspondence should be addressed. E-mail: leonid/at/genomics.princeton.edu

Conceived and designed the experiments: DB LK DG JS. Performed the experiments: DG JS. Analyzed the data: KD DR JS. Wrote the paper: LK DR DG JS.

Received January 8, 2007; Accepted February 28, 2007.

This article has been cited by other articles in PMC.

Abstract

Ten years have passed since the genome of Saccharomyces cerevisiae–more precisely, the S288c strain–was completely sequenced. However, experimental work in yeast is commonly performed using strains that are of unknown genetic relationship to S288c. Here, we characterized the nucleotide-level similarity between S288c and seven commonly used lab strains (A364A, W303, FL100, CEN.PK, ∑1278b, SK1 and BY4716) using 25mer oligonucleotide microarrays that provide complete and redundant coverage of the ~12 Mb Saccharomyces cerevisiae genome. Using these data, we assessed the frequency and distribution of nucleotide variation in comparison to the sequenced reference genome. These data allow us to infer the relationships between experimentally important strains of yeast and provide insight for experimental designs that are sensitive to sequence variation. We propose a rational approach for near complete sequencing of strains related to the reference using these data and directed re-sequencing. These data and new visualization tools are accessible online in a new resource: the Yeast SNPs Browser (YSB; http://gbrowse.princeton.edu/cgi-bin/gbrowse/yeast_strains_snps) that is available to all researchers.

Introduction

For decades, the budding yeast, Saccharomyces cerevisiae, has been used as a model organism for studying eukaryotic molecular and cell biology. Its primacy as a model system and its small and compact genome made it the best eukaryotic candidate for the early sequencing efforts. In 1996, the strain S288c, first isolated in the 1950's through genetic crosses by Robert Mortimer [1], became the first eukaryote to be completely sequenced [2]. The determination of this reference sequence facilitated the construction of the first whole genome microarrays [3] and the systematic deletions of all genes [4], as well as the first whole genome protein and genetic interaction maps [5], [6].

Despite the fact that most of the genomic information has been obtained from S288c, this strain is often not the ideal genetic background for studying particular aspects of biology. In fact, the S288c background has a number of drawbacks, such as low sporulation efficiency [7], an inability to grow on maltose [8] and failure to initiate filamentous growth upon nitrogen starvation [9]. For these reasons, other genetic backgrounds are used for physiological [10], genetic and genomic analyses [11]–[13] in many laboratories. These other genetic backgrounds are sometimes of known close (e.g. A363A and W303) or distant (SK1) genetic relatedness to S288c. However, for some experimentally important strains the relationship is unclear and largely undocumented.

The use of these different strains may contribute to the inconsistencies in some biological results because it is often assumed that the genomic sequence information of S288c can be extrapolated to other strains. In fact, even for strains closely related to S288c, the small number of sequence differences may still have important consequences for different biological pathways and phenotypes. Therefore, understanding genetic differences between strains has become extremely important. Mortimer and Johnson [1] traced a genealogy of the commonly used S. cerevisiae strains from knowledge of their history. More recently, Winzeler et al. [14] determined a subset of DNA sequence variation among a set of 14 S. cerevisiae strains using low-coverage oligonucleotide arrays and discovered 11,115 sites of variation among them. Although this study provided some insight into allelic differences between yeast strains, it was limited by the proportion of the genome covered by the array (~16%) and the 25 bp resolution for localizing these variants.

We recently developed a method for characterizing nucleotide variation in the entire genome using 25mer oligonucleotide microarrays (Affymetrix yeast tiling arrays) that provide complete and redundant coverage of the ~12 Mb S. cerevisiae genome [15]. This design provides for multiple measurements of each nucleotide's contribution to hybridization efficiency and therefore has the ability to detect the presence and location of single nucleotide polymorphisms (SNPs) and deletion events throughout the entire yeast genome with near nucleotide precision. We have employed this approach to characterize the nucleotide-level similarity and divergence between S288c and 7 commonly used lab strains (A364A, W303, FL100, CEN.PK, ∑1278b, SK1 and BY4716). The analyses revealed which genomic regions of each strain are derived from the reference strain and which regions are highly diverged, indicating a different ancestry. The data are presented in an online database, the Yeast SNPs Browser (YSB; http://gbrowse.princeton.edu/cgi-bin/gbrowse/yeast_strains_snps). YSB represents a valuable tool for the yeast community, enabling the development of genomic resources for other yeast strains and informing conventional molecular biology methods such as PCR primer and Southern blot probe design.

Results and Discussion

Frequency of SNPs between strains

For this study, 7 unsequenced laboratory strains, A364A, W303, FL100, CEN.PK, ∑1278b, SK1 and BY4716, were chosen because of their importance and frequent use in research (Table 1). For example, the SK1 background is used for studying sporulation and meiosis [7], [16], ∑1278b is used in pseudohyphal growth studies [17] and A364A is widely used in studies of the cell cycle [18]. As a control, we also included the strain BY4716 that is known to be nearly identical to S288c [19].

Table 1

Strains of S. cerevisiae used in this study

To characterize the genomes of these unsequenced laboratory strains at the nucleotide level, we hybridized genomic DNA from each strain to a high-density Affymetrix yeast tiling microarray and detected potential SNPs using the software package SNPscanner as described [15]. Where possible, we analyzed a prototrophic wildtype clone that was representative of the strain. The provenance of the strains is indicated in Table 1. SNPscanner detects SNPs based on up to 7 measurements of a nucleotide's effect on hybridization efficiency and calculates the log of the likelihood ratio (referred as the prediction signal) for the presence of a SNP at each nucleotide in the genome. The ability to detect a large number of SNPs in an entire diverged genome was tested in our previous study using the two completely sequenced strains, RM11_1a and YJM789. Analysis of a single hybridization of YJM789 with SNPscanner correctly detected 86.9% of 30,303 known SNPs at a prediction signal of 5, with only 7 false positives calls (i.e. a rate of 1 false positive/2 Mb) across the nonrepetitive portion of the genome.

We therefore extended this approach to obtain a nucleotide-resolution comparison of the commonly used laboratory strains with the S288c reference strain by analyzing data from a single hybridization of each strain. Using the same threshold settings (a prediction signal of 5 and heuristic filters) as reported previously [15], we found a moderately low number of predicted SNPs in the two strains closely related to S288c (4,894 calls in A364A and 7,955 calls in W303), whereas a higher number of SNPs was detected in the more distant strains (15,352 calls in FL100, 25,298 calls in ∑1278b and 37,424 calls in SK1) (Table 1). The divergence between the different strains and the reference varies between 0.05% to 0.36% of the genome. As expected, the number of detected SNPs for the BY4716 strain was very low (39 calls in total).

Genome-wide distribution of SNPs

To identify strains that share a common recent ancestry with the reference strain, we examined the genome-wide distribution of variation for each strain. For strains derived from S288c, we expect a small number of SNPs nonrandomly distributed across the genome. By contrast, the strains unrelated to S288c should have a high number of SNPs distributed throughout the genome.

The distribution of SNPs in A364A, W303 (both known to be the result of crosses between S288c and unrelated strains) and CEN.PK is indeed nonrandom. The genomes are mosaics of large regions identical to S288c interspersed with small regions of high sequence divergence (Fig. 1). In the strains closely related to S288c, variation was found in a small fraction of the genome, as previously described [14] (Fig. 2). For example, 80% of the SNPs in A364A, W303 and CEN.PK are found in 10%, 15% and 18% of the genome, respectively. By contrast, variation in SK1 is distributed relatively evenly throughout the genome.

Figure 1

Prediction signal across chromosome 5 for the laboratory strains A364A, W303, Σ1278b and SK1.

Figure 2

Genome-wide distribution of variation.

For the strains that are clearly not close relatives of S288c, we asked if SNPs are unevenly distributed across the chromosomes. In all strains, variation is distributed across the genome, but decreases with proximity to the centromere (Fig. 3). The centromeric regions (within 20 kb of the centromeres) show a lower-than-average variability. This observation is possibly linked to the lack of DNA double-strand breaks (i.e the presence of meiotic recombination coldspots) near the centromeres [20]–[22]. Because of a reduced number of cross-overs and gene conversion events, these regions are less polymorphic, a result that is consistent with population genetic theory and observations in other organisms (e.g. Drosophila) [23].

Figure 3

Distribution of variation by distance from centromere.

By contrast, it is well known that subtelomeric regions that undergo frequent recombination events display increased variability at the sequence level [24]–[26]. However, these regions are rich in redundant sequences that are responsible for high levels of false positives using our strategy. In an effort to reduce false positives, SNPscanner removes all probes that are repeated on the array, thereby removing many of these regions from the initial analysis. With these regions removed, we fail to detect a higher proportion of genetic variability at the regions located near the telomere. When we then include these repeated regions, we see a marked increase in SNP calls near the telomere. However, we are currently unable to determine the percentage of these that are real versus false positives.

Deletion variants in the surveyed strains

As shown in our previous study [15], hybridization of genomic DNA to the yeast tiling arrays can be used to detect deletions and predict the breakpoints with a resolution of a few basepairs. Therefore, these data provide a precise and global view of the deletion events in the laboratory strains examined. To distinguish regions that are deleted from those that are present but highly polymorphic-situations that result in similar SNPscanner profiles-we restricted our analysis to deleted regions that are longer than 500 bp in size (Table 1 and S1).

A total of 113 deletions met this criterion in a comparison between the reference strain and the other strains of this study. The number of deletions varied from 1 in the BY4716 strain to 36 in the SK1 strain. As expected, only the known engineered deletion of the LYS2 gene was detected in the BY4716 strain. The mapped deletions range from 0.5 to 9.1 kb, with the majority falling between 0.5 and 1 kb (Fig. 4).

Figure 4

Size distribution of the deletion variants identified.

We identified all the genes that are involved in these deletion events. In total, 87 genes carry a whole or partial deletion in at least one strain (S1). Twenty-five genes have partial deletions, whereas 62 genes are deleted in full. We compared the function of the deleted genes on the basis of their Gene Ontology (GO) annotations [27] by using the GO Term Finder from the Saccharomyces genome database [28]. We found no significant enrichment with respect to biological function. A large proportion of the deleted genes (~26%) have not yet been assigned a biological function. However, some of them play an important role in metabolism and are involved in transporter activity (SEO1, ARR3, AGP3, MAL11, QCR8, HXT17, OPT2 and SGE1genes). All of these genes except for QCR8 are located within 25 kb of the chromosome ends, consistent with a relationship between adaptation and subtelomeric regions, as previously reported [14].

Sequencing strategy for the related S288c strains

Random shotgun whole genome sequencing is of unquestioned utility for de novo determination of an organism's genome. However, once a reference genome sequence has been determined for a species, applying this approach to re-sequence other closely related members of the population represents a redundant and inefficient approach. Although our method for comparing genomes at nucleotide resolution does not provide the actual DNA sequence of a genome, it does offer one approach to a more efficient resequencing of closely related strains. In this scheme, we suggest that initial determination of regions that are diverged from the reference genome using SNPscanner can then be followed by targeted re-sequencing of these regions.

We propose that this approach is applicable for strains A364A, W303 and CEN.PK, where the majority of variation was found in a small fraction (<20%) of the genome. By selecting and sequencing only the 1 kb regions of high diversity as identified by SNPscanner, we can obtain sequence information for 90% of a yeast strain's genome by sequencing only 1–3 megabases, substantially minimizing cost and effort (Table 2). Although these strains are experimentally important, it is unlikely that they will be of high priority for high coverage sequencing projects. Therefore, we propose that this approach will be informative until whole genome sequencing is greatly reduced in cost. This approach is limited to strains that are clear relatives of S288c, with much of their genomes being identical and the differences clustered into small regions. In addition, this approach will not identify regions of the genome (insertions) that are present in the sample strain but unrepresented in the reference strain.

Table 2

Sequencing strategy for the S288c related strains

The Yeast SNPs Browser

To allow researchers to easily visualize and explore these data we have created a database, the Yeast SNPs Browser, which can be used to view all the SNPs and deletions distinguishing these strains, along with the prediction signal across every base in the genome. We used the Generic Genome Browser (GBrowse), a web-based application for displaying chromosomal features [29], to provide user-friendly access to the data via the web. The Yeast SNPs Browser provides a way to view the putative SNPs in all the strains assayed in the context of the S288C genome. One can search by gene, feature name or oligonucleotide sequence, browse chromosomal regions, and zoom from an overview of an entire chromosome to the individual nucleotide level (Fig. 5). By zooming in to about 100 bp resolution, one can see the nucleotide sequence of the S288C reference strain in the region of the SNP. In addition, the prediction score is available as a graph across every base in the genome (Fig. 5).

Figure 5

Yeast SNPs Browser screenshots.

Conclusion

There are several S. cerevisiae strains that are commonly used in laboratory experiments, some of which are distantly related to the sequenced reference strain S288c. Understanding this variation is important for interpreting biological results from experiments performed with these strains. Using yeast tiling arrays, we have defined the genetic variation between the common laboratory strains of yeast at the nucleotide level in a reliable, cost-effective manner. We have also created a simple and visually pleasing resource for researchers. These data provide a valuable tool for the yeast community. They should inform experimental work that is reliant on knowledge of sequence such as the design of primers for PCR reactions and probes for Southern analysis. For studies of gene expression in the different laboratory strains, these data can also be useful as a source of information to design a “universal” microarray by selecting probes in regions of complete or high sequence similarity, in order to avoid any potential confounding effects on hybridization of strain-specific sequence variation. Finally, these data constitute the first step toward discovering the underlying genetic differences that contribute to the phenotypic differences among these strains.

Materials and Methods

Yeast strains

Strains used in this study are listed in Table 1.

Sample preparation and hybridization on arrays

Yeast strains were grown in yeast extract, peptone, and dextrose (YPD) medium. Total genomic DNA was purified from 30 ml YPD culture using Qiagen Genomic-Tips 100/G and Genomic DNA Buffers as per the manufacturer's instructions. Genomic DNA was digested with DNaseI, labeled and hybrizided to Affymetrix Yeast Tiling Arrays as described in Gresham et al. [15].

Supporting Information

Table S1

Genes carrying a whole or partial deletion in the surveyed strains

(0.02 MB DOC)

Click here for additional data file.^{(24K, doc)}

Acknowledgments

The authors gratefully acknowledge helpful discussions with members of the Botstein, Kruglyak and Dunham laboratories. We thank J. Matese and D. Storton for technical support.

Footnotes

Competing Interests: The authors have declared that no competing interests exist.

Funding: This work was supported by National Institute of General Medical Sciences grant R01 GM046406 and center grant P50 GM071508 (D.B.), NIH grant R37 MH059520 and James S. McDonnell Foundation grant 99-11T (L.K.).

References

1. Mortimer RK, Johnston JR. Genealogy of principal strains of the yeast genetic stock center. Genetics. 1986;113:35–43. [PubMed]

2. Goffeau A, Barrell BG, Bussey H, Davis RW, Dujon B, et al. Life with 6000 genes. Science. 1996;274:546, 563–7. [PubMed]

3. DeRisi JL, Iyer VR, Brown PO. Exploring the metabolic and genetic control of gene expression on a genomic scale. Science. 1997;278:680–686. [PubMed]

4. Winzeler EA, Shoemaker DD, Astromoff A, Liang H, Anderson K, et al. Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science. 1999;285:901–906. [PubMed]

5. Uetz P, Giot L, Cagney G, Mansfield TA, Judson RS, et al. A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature. 2000;403:623–627. [PubMed]

6. Tong AH, Lesage G, Bader GD, Ding H, Xu H, et al. Global mapping of the yeast genetic interaction network. Science. 2004;303:808–813. [PubMed]

7. Deutschbauer AM, Davis RW. Quantitative trait loci mapped to single-nucleotide resolution in yeast. Nat Genet. 2005;37:1333–1340. [PubMed]

8. Charron MJ, Dubin RA, Michels CA. Structural and functional analysis of the MAL1 locus of Saccharomyces cerevisiae. Mol Cell Biol. 1986;6:3891–3899. [PMC free article] [PubMed]

9. Gagiano M, Bauer FF, Pretorius IS. The sensing of nutritional status and the relationship to filamentous growth in Saccharomyces cerevisiae. FEMS Yeast Res. 2002;2:433–470. [PubMed]

10. van Dijken JP, Bauer J, Brambilla L, Duboc P, Francois JM, et al. An interlaboratory comparison of physiological and genetic properties of four Saccharomyces cerevisiae strains. Enzyme Microb Technol. 2000;26:706–714. [PubMed]

11. Primig M, Williams RM, Winzeler EA, Tevzadze GG, Conway AR, et al. The core meiotic transcriptome in budding yeasts. Nat Genet. 2000;26:415–423. [PubMed]

12. Xu H, Boone C, Klein HL. Mrc1 is required for sister chromatid cohesion to aid in recombination repair of spontaneous damage. Mol Cell Biol. 2004;24:7082–7090. [PMC free article] [PubMed]

13. Huang S, Zhou H, Katzmann D, Hochstrasser M, Atanasova E, et al. Rtt106p is a histone chaperone involved in heterochromatin-mediated silencing. Proc Natl Acad Sci U S A. 2005;102:13410–13415. [PubMed]

14. Winzeler EA, Castillo-Davis CI, Oshiro G, Liang D, Richards DR, et al. Genetic diversity in yeast assessed with whole-genome oligonucleotide arrays. Genetics. 2003;163:79–89. [PubMed]

15. Gresham D, Ruderfer DM, Pratt SC, Schacherer J, Dunham MJ, et al. Genome-wide detection of polymorphisms at nucleotide resolution with a single DNA microarray. Science. 2006;311:1932–1936. [PubMed]

16. Ben-Ari G, Zenvirth D, Sherman A, David L, Klutstein M, et al. Four linked genes participate in controlling sporulation efficiency in budding yeast. PLoS Genet. 2006;2:e195. [PubMed]

17. Liu H, Styles CA, Fink GR. Elements of the yeast pheromone response pathway required for filamentous growth of diploids. Science. 1993;262:1741–1744. [PubMed]

18. Hartwell LH. Macromolecule synthesis in temperature-sensitive mutants of yeast. J Bacteriol. 1967;93:1662–1670. [PMC free article] [PubMed]

19. Brachmann CB, Davies A, Cost GJ, Caputo E, Li J, et al. Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast. 1998;14(2):115–32. [PubMed]

20. Lambie EJ, Roeder GS. A yeast centromere acts in cis to inhibit meiotic gene conversion of adjacent sequences. Cell. 1988;52:863–873. [PubMed]

21. Baudat F, Nicolas A. Clustering of meiotic double-strand breaks on yeast chromosome III. Proc Natl Acad Sci U S A. 1997;94:5213–5218. [PubMed]

22. Gerton JL, DeRisi J, Shroff R, Lichten M, Brown PO, et al. Inaugural article: global mapping of meiotic recombination hotspots and coldspots in the yeast Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 2000;97:11383–11390. [PubMed]

23. Hartl DL, Clark AG. Sunderland, Massachusetts: Sinauer Associates; 1997. Principles of Population Genetics.

24. Louis EJ, Haber JE. The subtelomeric Y' repeat family in Saccharomyces cerevisiae: an experimental system for repeated sequence evolution. Genetics. 1990;124:533–545. [PubMed]

25. Pryde FE, Gorham HC, Louis EJ. Chromosome ends: all the same under their caps. Curr Opin Genet Dev. 1997;7:822–828. [PubMed]

26. Liti G, Peruffo A, James SA, Roberts IN, Louis EJ. Inferences of evolutionary relationships from a population survey of LTR-retrotransposons and telomeric-associated sequences in the Saccharomyces sensu stricto complex. Yeast. 2005;22:177–192. [PubMed]

27. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000;25:25–29. [PMC free article] [PubMed]

28. Cherry JM, Adler C, Ball C, Chervitz SA, Dwight SS, et al. SGD: Saccharomyces Genome Database. Nucleic Acids Res. 1998;26:73–79. [PMC free article] [PubMed]

29. Stein LD, Mungall C, Shu S, Caudy M, Mangone M, et al. The generic genome browser: a building block for a model organism system database. Genome Res. 2002;12:1599–1610. [PubMed]

30. Thomas BJ, Rothstein R. The genetic control of direct-repeat recombination in Saccharomyces: the effect of rad52 and rad1 on mitotic recombination at GAL10, a transcriptionally regulated gene. Genetics. 1989;123:725–738. [PubMed]

31. Lacroute F. Regulation of pyrimidine biosynthesis in Saccharomyces cerevisiae. J Bacteriol. 1968;95:824–832. [PMC free article] [PubMed]

32. Grenson M, Mousset M, Wiame JM, Bechet J. Multiplicity of the amino acid permeases in Saccharomyces cerevisiae. I. Evidence for a specific arginine-transporting system. Biochim Biophys Acta. 1966;127:325–338. [PubMed]

33. Kane SM, Roth R. Carbohydrate metabolism during ascospore development in yeast. J Bacteriol. 1974;118:8–14. [PMC free article] [PubMed]

Articles from PLoS ONE are provided here courtesy of
Public Library of Science