Figures
Abstract
Gross chromosomal rearrangements have the potential to be evolutionarily advantageous to an adapting organism. The generation of a hybrid species increases opportunity for recombination by bringing together two homologous genomes. We sought to define the location of genomic rearrangements in three strains of Saccharomyces pastorianus, a natural lager-brewing yeast hybrid of Saccharomyces cerevisiae and Saccharomyces eubayanus, using whole genome shotgun sequencing. Each strain of S. pastorianus has lost species-specific portions of its genome and has undergone extensive recombination, producing chimeric chromosomes. We predicted 30 breakpoints that we confirmed at the single nucleotide level by designing species-specific primers that flank each breakpoint, and then sequencing the PCR product. These rearrangements are the result of recombination between areas of homology between the two subgenomes, rather than repetitive elements such as transposons or tRNAs. Interestingly, 28/30 S. cerevisiae- S. eubayanus recombination breakpoints are located within genic regions, generating chimeric genes. Furthermore we show evidence for the reuse of two breakpoints, located in HSP82 and KEM1, in strains of proposed independent origin.
Citation: Hewitt SK, Donaldson IJ, Lovell SC, Delneri D (2014) Sequencing and Characterisation of Rearrangements in Three S. pastorianus Strains Reveals the Presence of Chimeric Genes and Gives Evidence of Breakpoint Reuse. PLoS ONE 9(3): e92203. https://doi.org/10.1371/journal.pone.0092203
Editor: Conrad A. Nieduszynski, The University of Nottingham, United Kingdom
Received: November 7, 2013; Accepted: February 19, 2014; Published: March 18, 2014
Copyright: © 2014 Hewitt 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.
Funding: SKH was funded by a Biotechnology and Biological Sciences Research Council (BBSRC) studentship. This work was also supported by a BBSRC grant to SCL and DD (BB/F007620/1). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Hybridisation in Saccharomycetous yeast occurs readily in natural and industrial environments [1], [2], [3], [4], [5], [6], [7], and may be a swift mechanism for evolutionary innovation. Investigating the genomics of successful natural hybrid species can provide valuable evolutionary insight into how the union of diverged genetic material can sculpt a genome more suited to its new environmental niche. These adaptations may include chromosomal rearrangements such as duplication, translocation, inversion and selective loss of genes or even whole chromosomes. The lager yeast Saccharomyces pastorianus, previously classified as Saccharomyces carlsbergensis, is a natural hybrid between Saccharomyces cerevisiae and a Saccharomyces uvarum-like species [4], [8], [9], [10]. The S. uvarum-like species has most recently been identified as the Argentinean-isolate Saccharomyces eubayanus, which shows 99.5% identity to the non-S. cerevisiae portion of S. pastorianus [11]. S. pastorianus is thought to have arisen by spontaneous hybridisation in brewery conditions, maintained by human selection for colder brewing temperatures, a preference that is conferred by its S. uvarum-like parent [12].
So far, there are only two whole genome lager yeast sequences available, Weihenstephan 34/70 [13] and CCY48–91, which has been recently deposited in Genbank [ID:ALJS00000000.1]. Much of our knowledge of the genome composition of these natural hybrids derives from previous array-based comparative genomic hybridisation studies (array-CGH) performed on 17 strains of S. pastorianus [10]. This particular work identified two groups of lager yeasts: Group 1 strains contain roughly one haploid S. cerevisiae and one haploid S. eubayanus genome with significant loss of S. cerevisiae genes, whereas Group 2 strains contain one haploid S. eubayanus genome and a diploid S. cerevisiae genome. The differences between these two groups suggest that they may have had independent evolutionary origins, a theory given weight by both the aforementioned array-CGH analysis [10] and the differing distribution of transposons between the two groups [14]. Additionally, strains within each group are highly variable in their patterns of chromosomal loss, aneuploidy and gross chromosomal rearrangements, probably reflecting either evolutionary pressure from diverse brewery conditions or random genetic drift [9], [10].
Lager yeast chromosomes have been shown to have undergone recombination, generating chimeric chromosomes composed of genetic material from both parental species [10], [13], [15], [16]. Typically, recombination between chromosomes within a non-hybrid yeast species is thought to be mediated primarily by transposons (Ty elements) [15], [17], [18], [19], tRNAs [17], [18], duplicated genes [18] or, as more recently proposed, origins of replication [20]. However, breakpoint formation in S. pastorianus is thought to be either Ty-mediated [10], [15] or the result of recombination between homologous regions [13]. Studies have also demonstrated the role of high stress brewery conditions in promoting genomic rearrangements, such as localised areas of gene amplification and recombination [21]. Significantly, chromosomal rearrangements have been shown to confer adaptive traits in both wild and industrial yeasts including highly sulphite-resistant wine yeast [22], flor wine yeast [23] and wild copper-tolerant yeast [24]. Furthermore, rearrangements have been shown to contribute to speciation between species of yeast [25].
We sequenced three S. pastorianus strains to both characterise genomic breakpoints and shed further light on their formation and retention. We chose strains that have been used in a previous microarray study to provide a source of validation for our sequencing [10]. These strains have also been pre-classified into one of the two aforementioned groupings of S. pastorianus: two of the chosen strains of S. pastorianus are of Group 1 (DBVPG 6033 and DBVPG 6261) and one is of Group 2 (DBVPG 6257) [10]. These particular strains have the greatest level of differential gene loss and therefore the least amount of redundancy. The latter group is thought to have an independent evolutionary origin from the former group, allowing us to investigate similarities between non-related strains.
We have confirmed the location of many S. cerevisiae- S. eubayanus breakpoints at the single nucleotide level and identified both nearby repetitive elements and regions of homology. Significantly, we found that the majority of genomic breakpoints occurred within protein coding regions, generating chimeric genes. Furthermore, the presence of identical breakpoints in KEM1 and HSP82 is evidence of breakpoint reuse between strains of proposed independent origin.
Results and Discussion
Genome Assembly and Analysis
The genomic DNA of three strains of S. pastorianus, DBVPG 6033 (Saccharomyces carlsbergensis type strain), DBVPG 6261 (Saccharomyces monacensis type strain) and DBVPG 6257 were sequenced using the SOLiD 4 Next Generation Sequencing platform and mapped to S. cerevisiae (sacCer2) and S. uvarum (sacBay MIT), which are representative of the S. pastorianus subgenomes. We used sacBay MIT as the reference genome for S. eubayanus due to its fully available sequence, which is purportedly 7% diverged from S. eubayanus [11]. Visualisation of the S. cerevisiae chromosomes in the UCSC Genome Browser (http://genome.ucsc.edu/) is reported in Figure 1. S. eubayanus reads mapped to contigs were viewed in the Integrative Genomics Viewer (http://www.broadinstitute.org/igv). SOLiD sequencing allowed us to ascertain the approximate chromosomal copy number in each strain using a hierarchical cluster analysis of relative median read depth across multiple regions (Tables S3–S4; Figure S3). In total, DBVPG 6033, 6261 and 6257 are estimated to have 31, 31 and 48 chromosomes respectively (Table 1). These chromosomes map to S. cerevisiae, S. eubayanus or a combination of both S. cerevisiae and S. eubayanus sequence (chimeric chromosomes).
Full set of S. cerevisiae chromosomes is displayed for each strain using the UCSC Genome Browser. Depth of track corresponds to read number. Track depth window is set to a read depth of 300 to accommodate three copies of a chromosome, since S. cerevisiae chromosome copy number in these strains generally varies between zero and three. Regions used in copy number analysis are labelled 1–26 in red.
S. pastorianus shows a high degree of aneuploidy and the chromosomal composition between strains is highly variable. Strains DBVPG 6033, 6261 and 6257 have eight, seven and sixteen complete S. cerevisiae chromosomes, respectively (Table 1). They also have an estimated 14, 10 and 15 complete S. eubayanus chromosomes, and 9, 14 and 17 chimeric chromosomes, composed of both S. cerevisiae and S. eubayanus sequence (Table 1). The approximate number of total chromosomes in DBVPG 6033 and 6261 (Group 1) is 31 each, both roughly equal to a diploid (16×2 = 32). The total number of chromosomes in DBVPG 6257 (Group 2) is 48, which equates to a triploid (16×3 = 48). These data support previous estimates of Group 1 strains generally being diploid-derived and Group 2 strains being triploid-derived [10]. Both Group 1 strains have lost their S. cerevisiae copies of chromosomes VI and XII, i.e. there is no evidence of this sequence, even on a chimeric chromosome. DBVPG 6261 has additionally lost its S. cerevisiae chromosome III and XIV sequences, whereas DBVPG 6033 has lost its S. cerevisiae chromosome XI sequence. There was no detection of S. cerevisiae chromosome V in the Group 2 strain DBVPG 6257. All three strains of S. pastorianus show evidence of homologous recombination between S. cerevisiae and S. eubayanus chromosomes IV, VII, XIII and XVI. Additionally, chromosome VIII and XV are chimeric in both Group 1 strains. Chromosomes IX, X and XIV are also chimeric in Group 2 strain DBVPG 6257. Chromosomes I and VI remain largely stable, showing no evidence of S. cerevisiae- S. eubayanus recombination in any of the sequenced strains. Reciprocal recombination and inversion events could not be identified in this study since these rearrangements are copy-number neutral.
In agreement with previous analysis of these three strains of S. pastorianus [10], we did not detect any S. cerevisiae mitochondrial DNA. The restriction analysis of COX2 in the three strains of S. pastorianus has indicated a S. uvarum-like mitochondrial sequence (data not shown), supporting the widely held notion that lager yeasts tend to inherit and/or retain only their S. eubayanus mitochondria [26]. The 2-micron plasmid maps to S. cerevisiae sequence in strains DBVPG 6033 and DBVPG 6257, but not DBVPG 6261 (Figure S1). It is unknown if any 2-micron plasmids are S. eubayanus-derived.
Chromosomal Rearrangements
We used the UCSC genome browser (http://genome.ucsc.edu/) to identify candidate breakpoints based on variations in S. cerevisiae read copy number across each chromosome (Figure 1). Using this technique, we were able to detect a total of 13 S. cerevisiae- S. eubayanus breakpoints in DBVPG 6033, 13 in DBVPG 6261 and 18 in DBVPG 6257. We used species-specific primers to confirm the presence of each breakpoint by PCR (Figure 2). Each successfully amplified PCR product was sequenced at GATC Biotech (Germany). All of the sequenced breakpoints were then aligned to the S. cerevisiae reference genome and either the S. eubayanus reference genome (where available) or the S. uvarum reference genome (Figure S2). A total of 9/13 S. cerevisiae- S. eubayanus breakpoints were confirmed by PCR in DBVPG 6033, another 11/13 in DBVPG 6261 and 10/18 in DBVPG 6257 (Table 2). In DBVPG 6033, three of these breakpoints were located on chromosome XVI, two on VIII and the remaining four on chromosomes II, IV, VII and XIII. In DBVPG 6261, three were located on chromosome XV, two on chromosome XI, two on chromosome XIII and the remaining four on chromosomes IV, V, VII and XVI. In DBVPG 6257, three breakpoints were located on chromosomes XVI, two on XIII and the remaining five on chromosomes IV, VII, IX, X and XI. It should be noted that the single-read sequencing strategy did not allow us to detect rearrangements that were copy number neutral (e.g. reciprocal recombination).
To establish the position of rearrangements in the three S. pastorianus strains (Sp), species-specific primers (Sc F: S. cerevisiae forward primer, Se F: S. eubayanus forward primer. Sc R: S. cerevisiae reverse primer. Se R: S. eubayanus reverse primer) were designed around putative breakpoints observed using the UCSC Genome Browser. The amplified products were then sequenced to locate the breakpoints at nucleotide level. S. pastorianus strains are labelled as Group 1 or Group 2 according to their previous assignment [10]. Nucleotide level sequence is shown for each rearrangement with the breakpoint region demarcated by a box flanked by the first unique S. cerevisiae nucleotide (blue) and the first unique S. uvarum nucleotide (red, S. eubayanus sequence unavailable).
The majority of the sequenced S. cerevisiae- S. eubayanus breakpoints occur within coding regions (Table 2), despite breakpoints in yeast usually being located in intergenic rather than intragenic regions [27]. A total of seven, eleven and ten intragenic breakpoints were located in DBVPG 6033, 6261 and 6257 respectively. In each strain there were a small number of candidate breakpoints which could not be amplified (Table S1). Four of these unconfirmed breakpoints are in DBVPG 6033 (one of which was within in a coding region), two in DBVPG 6261 (both of which were within in a coding region) and eight in DBVPG 6257 (three of which were within a coding region). Notably, we detected a breakpoint at the MAT locus on chromosome III in strains DBVPG 6033 and 6257. Breakpoints to the right of the MAT locus in chromosome III have been noted previously in many strains of lager yeast [13], [15], [28]. There are two breakpoints that are in close proximity on chromosome X in strain DBVPG 6257, two on chromosome XIV and one on chromosome XII in strain DBVPG 6257, which are in the vicinity of Ty elements. A further three candidate breakpoints, one in each strain (at the far right of each chromosome X) could not be amplified, possibly due to their close proximity to both an AT-rich ARS element and the right telomeric region. The four remaining unsequenced candidate breakpoints are on chromosome XIII and XV in DBVPG 6033, chromosome VIII in DBVPG 6261 and chromosome XIII in DBVPG 6257. These sites of increased copy number, which we were unable to confirm, may represent amplification events rather than translocations. Since we have called the breakpoints based on the depth of S. cerevisiae coverage we cannot distinguish between these two events at the onset, but can confirm any true recombination events via Sanger sequencing. Furthermore, copy number changes of true non-reciprocal translocation events or gene conversion are usually matched in both subgenomes. The unamplified breakpoint on chromosome XII in DBVPG 6257 may constitute a case of amplification since there is an increase of read depth in the S. cerevisiae genome but no change in the S. eubayanus genome (Table S4). Likewise, the pair of closely located unamplified breakpoints on chromosome X in DBVPG 6257 may represent one deletion, since there are no reads mapped to this region of the S. cerevisiae genome, while read depths remains unchanged in the S. eubayanus genome. Interestingly, there are three other pairs of breakpoints that generate small gaps in the S. cerevisiae subgenome sequencing data (ALD2-YMR196w on chromosome XIII of DBVPG 6257; ECM3-INP52 on chromosome XV of DBVPG 6261 and PRP8-intergenic on chromosome VIII of DBVPG 6033). These may represent gene conversion from a small tract in the S. eubayanus to the S. cerevisiae subgenome rather than deletion, since each breakpoint generated a chimeric gene (Table 3).
Chimeric Genes
As a result of homologous recombination between S. cerevisiae and S. eubayanus chromosomes, several chimeric genes were formed (Figure 3; Table 2; see Figure S2 for full sequence alignments). Two of these genes, KEM1, a 5′-3′ exonuclease, and HSP82, a molecular chaperone, are chimeric in all three strains. Interestingly, multiple sequence alignment using Clustal Omega showed that these two breakpoints occur in the same gene position in all three hybrids (Figure 4). The KEM1 S. eubayanus> S. cerevisiae breakpoint occurs within 462–477 bp after the start of the gene and the HSP82 S. eubayanus> S. cerevisiae breakpoint occurs within 1578–1608 bp after the start of the gene. Another chimeric gene, UTP4, is shared between both Group 1 strains, yet the position of the S. eubayanus> S. cerevisiae breakpoint differs between each strain: in DBVPG 6033, the breakpoint occurs within 444–477 bp after the start of the gene, whereas in DBVPG 6261, the breakpoint occurs within 861–876 bp after the start of the gene.
The S. cerevisiae and S. eubayanus portions of the gene are shown in blue and red respectively, and the position at which the breakpoint occurs within each gene is reported.
Panel A: Multiple sequence alignment of the breakpoint region within KEM1 in three strains of S. pastorianus and the parental species S. cerevisiae and S. eubayanus. Panel B: Multiple alignment of the breakpoint region within HSP82 in three strains of S. pastorianus and the parental species S. cerevisiae and S. uvarum (S. eubayanus sequence for this gene is unavailable). Unique nucleotide matches to S. cerevisiae are shown in blue while unique matches to S. eubayanus (KEM1) or S. uvarum (HSP82) are shown in red. The breakpoint region is demarcated by a box. Nucleotides shared between both parental species within the breakpoint region are shaded in grey.
Using regional read depth data across each chromosome (Table S3 and Table S4), we investigated chimeric gene copy number and whether additional complete S. cerevisiae or S. eubayanus copies of each chimeric gene were present in the sequencing data (Table 3). There is only one copy of the majority of chimeric genes across all three strains. However, there are two copies of four chimeric genes and three copies of six chimeric genes. The majority of duplicate and triplicate genes occur towards the ends of the chromosomes (see Figure 1). These have likely formed through either gene conversion involving three chromosomes or duplication of sub-telomeric regions, which are known to be sites of rapid gene expansion [29].
There are no additional complete parental strain copies of seven of the sequenced chimeric genes (HSP82 in all three strains, CDH1, IRR1, KEM1 (DBVPG 6261) and QCR2) (Table 3). Interestingly, IRR1 is an essential gene in S. cerevisiae, as well as being non-redundant. If it is presumed that IRR1 is also essential in S. pastorianus, it can be deduced that IRR1 is a functioning chimeric gene, given that this strain is viable. The remaining 21 chimeric genes have one or more non-chimeric homologues. Although only the chimeric form of HSP82, a molecular chaperone of protein complexes, is present in each strain of S. pastorianus, its homologue HSC82 has been retained in both parental forms. Two chimeric genes, KEM1 in DBVPG 6033 and TDH2 in DBVPG 6257 have additional homologous copies of both S. cerevisiae and S. eubayanus genes present in the genome. A further two chimeric genes, FKS3 and KEM1 in DBVPG 6257 have at least one additional S. cerevisiae homologue remaining in the genome. The final 17 chimeric genes are complemented by at least one additional complete S. eubayanus homologue. The identification of chimeric gene copy number within S. pastorianus is of importance in phylogenetic analysis, since they have the potential to weaken phylogenetic signal and contribute to incongruence [30], [31].
Interestingly, two genes found in the chimeric form are involved in ethanol metabolism, a key biochemical pathway in lager fermentation. ALD2 is involved in the oxidation of ethanol and TDH2 is a component of the tetramer glyceraldehyde-3-phosphate dehydrogenase, which is required for gluconeogenesis. Two further chimeric genes also play a role in energy metabolism. GPH1 is involved in glycogen mobilisation and GAL80 is a repressor of GAL genes in the absence of galactose. With the large genetic redundancy in all three strains of S. pastorianus, especially DBVPG 6257, which contains approximately two S. cerevisiae sets of chromosomes, chimeric gene copies may or may not significantly affect the hybrid organism.
Previous studies on the functionality and fitness of chimeric genes show mixed results. The chimeric gene GPH1 in the lager strain CMBS-33 contains a disruptive base insertion within its initial S. eubayanus sequence, and the resultant gene is not expressed [32]. However, a recent study located a recurrent in-frame breakpoint within MEP2, an ammonium permease, in clones of lab-created hybrids of S. cerevisiae and S. uvarum that were evolved under nitrogen-limiting conditions [33]. The experimentally evolved strains bearing the rearrangement were fitter than the non evolved strains in nitrogen-limiting competition experiments. A number of studies have also examined the fitness effects of rearrangements involving non-homologous genes. The LG-FLO1 gene, involved in flocculation, appears to have been inactivated in non-flocculent lager yeast by a non-reciprocal translocation of S. cerevisiae YIL169c into its C-terminal region in various non-flocculent strains [16]. However, a sulphite-resistant gene found in wine yeast, SSU1-R was generated by recombination between the promoter regions of SSU1 and ECM34 genes and has been found to grant increased sulphite resistance compared to the wild type allele [22], [34].
Comparison of Breakpoints in Different S. pastorianus Strains
Our data were compared to previous studies conducted on other S. pastorianus strains [10], [13], [15], [21]. The majority of the rearrangements detected in our study fall within the low resolution breakpoint regions determined previously (Figure 5). However, we also found new rearrangements in the strain DBVPG 6257 on chromosomes X, XIII and XIV; on chromosomes XVI in DBVPG 6033 and on chromosomes VIII and XIII in DBVPG 6261. A study by Bond and co-workers [15] conducted on Group 2 lager yeasts, namely DBVPG 6701 and CMBS-33, identified several breakpoints in common with the Group 2 strain DBVPG 6257 (Figure 5). However, one breakpoint unique to CMBS-33 was also found in the Group 1 strain DBVPG 6033. A previous whole genome sequencing of the Group 2 Weihenstephan 34/70 strain [13] identified a total of nine breakpoints, eight of which we also detected in our Group 2 strain DBVPG 6257. Moreover, two rearrangements, on chromosome VII (KEM1) and XVI (HSP82) were common to all the S. pastorianus strains analysed in our study (Figure 5). The sequencing of Weihenstephan 34/70 strain also showed a reciprocal breakpoint within TDH2, whereas we found altered copy number at this same site. Either there has been a chromosomal deletion in DBVPG 6257 after a reciprocal recombination event, or the original event was non-reciprocal. Furthermore, since reciprocal recombination events could not be detected in this study, it is also possible that the breakpoint may have occurred twice in this location, once as a reciprocal event and once as a non-reciprocal event.
We noted the presence of absence of similar breakpoints found in S. pastorianus strains analysed in four previous studies, Bond et al. 2004 [15], Dunn and Sherlock 2008 [10], James et al. 2008 [21] and Nakao et al. 2009 [13]. Grey shading indicates the presence of a breakpoint found within the same gene or a nearby/adjacent gene. *Breakpoints in both DBVPG 6033 and DBVPG 6261 fall within UTP4, but in differing locations. **The strains analysed in the study by James et al. 2008 are the product of mutagenesis and a laboratory evolution experiment. Cells labelled “cluster” within the Dunn and Sherlock (2008) analysis refer to breakpoints identified in our study which were not explicitly identified by the aforementioned study, but that fall within breakpoint cluster regions determined by the author across 17 strains of S. pastorianus.
The Reuse of Breakpoints
The reuse of breakpoints is a relatively new hypothesis that challenges the long held random breakage model of chromosomal rearrangements [35] and is gaining momentum in studies of mammalian and fly genomes. Studies of mammalian genomes have indicated that breakpoint regions may be reused throughout evolution at a rate between 7.7% and 20% [36], [37], [38]. The term breakpoint reuse, first coined by Pevzner and Tesler [39], applies to regions of overlapping breakpoints and is not limited to breakpoints shared at the nucleotide level. Although it is unclear whether this overlap in usage is random or non-random, there is increasing evidence of association of these evolutionary breakpoint regions with fragile sites (heritable regions prone to breakage and reorganization) [39], [40], with telomeric and centromeric regions [36], with segmental duplications [36], [37], and gene dense regions [36], [38]. Moreover, fragile sites prone to breakage, rather than functional constraints on genes, are thought to have been instrumental in shaping gene organisation, at least in Drosophila [41]. In our study, we see the reuse of two breakpoints, HSP82 and KEM1 in all three sequenced strains of S. pastorianus (Table 2). These breakpoints may have occurred independently between the separate groups. Alternatively, there could have been genetic exchange between Group 1 and 2, although the differing geographic distributions of each group make this unlikely [10]. Mutagenized lager yeast strains selected under heat stress and high osmotic stress [21] showed a rearrangement in YGL172w, which is adjacent to KEM1. They also showed four other rearrangements in or around breakpoint genes that were identified in our study (Figure 5), a further indication of breakpoint reuse in these regions of the genome. Interestingly, a fully sequenced S. cerevisiae- S. eubayanus breakpoint within GPH1, in the lager strain CMBS-33, differs in location to a breakpoint found in the same gene in DBVPG 6257 in our study [32]. The breakpoint identified in our study occurs within 1449–1464 bp of the start of the gene, whereas the breakpoint in CMBS-33 occurs after the first 330–360 bp of the gene. Similarly in our study, a breakpoint present in both Group 1 strains in UTP4 was resolved into two distinct breakpoints at the nucleotide level, which are in close proximity to a breakpoint approximately 30 kb upstream of this site in the ORF YDR338c of DBVPG 6257 (Table 2, Figure 2). This region on chromosome IV is potentially an additional site of independent evolutionary breakpoint reuse.
Mechanisms of Breakpoint Formation
Chromosomal translocation requires the induction of double-stranded DNA breaks followed by incorrect repair of these breaks using an erroneous homologous and repetitive sequence [42]. A recent study found that the potential of a double stranded break in the genome to cause changes in genome copy number increases when the breakage occurs within non-repetitive DNA rather than repetitive DNA [43]. This effect was far more pronounced in hybrid diploids comparative to non-hybrid diploids. This would suggest that any breakpoint that was to randomly occur within a coding region may be more likely to promote a genomic rearrangement in S. pastorianus than if the break was to occur within a repetitive element.
We looked for the presence or absence of Ty elements, their flanking LTRs, tRNAs and origins of replication in proximity to each identified breakpoint. Using the sequence data mapped to each S. cerevisiae SacCer2 chromosome in UCSC genome browser (http://genome.ucsc.edu/), we manually recorded the nearest repetitive genomic feature to each breakpoint (Table S2). The proximity of our sequenced breakpoints to a repetitive element ranges between 0.6 kb and 39 kb with a mean of 11.4 kb. Five sequenced breakpoints were less than 5 kb from an element; eleven were between 5 and 10 kb away; eight were between 10 kb and 20 kb away and four were father than 20 kb away from a repetitive element. The majority of breakpoints, having occurred within coding regions, were not immediately flanked by repetitive elements. The lack of association between breakpoints and repetitive elements in lager yeast is in agreement with Nakao and co-workers [13] and for some of the breakpoint events studied by Bond and co-workers [15]. However, Dunn and Sherlock [10] have observed clustering of breakpoints near repetitive features in the genomes of lager yeast. It is possible that our sequencing strategy was unable to fully detect any breakpoints that may have occurred within repetitive regions. Furthermore, our analysis of the location of these elements is based on SacCer2, and not the S. cerevisiae progenitor strain, of which we have no information. Additionally, we do not have data concerning the distribution of S. eubayanus repetitive elements and we could not accurately assess the locality of sacBay MIT (S. uvarum) transposons, since this portion of the data is mapped only to contigs. Despite the lack of proximity of repetitive elements to breakpoints as a trend, one breakpoint gene, TDH2 on chromosome XII in DBVPG 6257 is situated adjacent to an ARS, a feature known to promote chromosomal translocation [20]. Additionally we noticed a S. uvarum transposon present on a contig at the site of one breakpoint on chromosome XIII (Table S1).
Large areas of homology are known to induce recombination in yeast, and this mechanism is utilised widely for yeast gene deletion in the laboratory [44]. More recently however, very small areas of microhomology have also been indicated in the formation of chromosomal breakpoints in wine yeast [3], [22]. Since the reference parental species of S. pastorianus are closely related, with an average of 80% nucleotide identity in coding regions [45], we view the induction of recombination via homologous regions in lager yeast a likely hypothesis. Furthermore recombination has occurred more frequently in these coding regions than in non-coding regions, which have an average nucleotide identity of only 62% [45]. We examined the sequence surrounding each breakpoint using multiple alignments to the two parental subgenomes (Figure S2) and identified many cases of local large areas of identity and smaller areas of microhomology that are at the site of each sequenced breakpoint (Table S2).
Whatever the underlying sequence that facilitates breakpoint formation, it is likely that one or all of the following three events is potentiating breakpoint formation: the unstable nature of newly formed hybrids, an increase in the occurrence of double stranded breaks under stressful brewery conditions and/or an evolutionary pressure for recombination. The reoccurrence of known breakpoints in mutagenized lager strains which have been evolved under high stress brewery conditions [21] promotes the existence of breakpoint hotspots in the S. pastorianus genome and gives evidence for the role of stress in promoting and maintaining genomic breakpoints.
Conclusion
Our whole genome sequencing of three strains of S. pastorianus allowed the identification of S. cerevisiae- S. eubayanus chromosomal breakpoints at a single nucleotide resolution. The majority of S. cerevisiae- S. eubayanus breakpoints are located within coding regions and were most likely formed as a result of homology and microhomology between the two parental subgenomes, rather than via repetitive elements in the genome. PCR sequencing of breakpoints enabled the further characterisation of these recombination-generated chimeric genes. The greater resolution granted by PCR sequencing allowed us to verify that the breakpoints within HSP82 and KEM1 have occurred at an identical genomic location in all three strains. We determined that two different breakpoints have occurred within UTP4 in the two Group 1 strains. Although the breakpoints are in different positions, this will still be regarded as an example of breakpoint reuse. Interestingly, we note the presence of a chimeric gene IRR1 in DBVPG 6257 of S. pastorianus that has lost both parental homologues. Since IRR1 is also an essential gene, this indicates that the chimeric gene is efficiently utilised by the hybrid. The presence of chimeric genes in the genome also has the potential to weaken the phylogenetic signal of these genes, which could promote incongruence in phylogenetic analyses [31]. Future studies on the function and fitness of chimeric genes may reveal their evolutionary role in facilitating the adaption of S. pastorianus to high stress brewery conditions.
Materials and Methods
Strains and Media
Saccharomyces pastorianus strains DBVPG 6033 (GSY129), DBVPG 6261 (GSY134) and DBVPG 6257 (GSY132) were obtained from DBVPG Industrial Yeasts Collection, University of Perugia, Italy. Yeast was grown at 25°C, 200 rpm for 20 hours in YPD (1% yeast extract, 2% peptone, 2% glucose) and genomic DNA extracted using Wizard Genomic DNA Purification Kit (Promega).
Genome sequencing reference strain for Saccharomyces cerevisiae, sacCer2 was obtained via the UCSC Genome Browser (http://genome.ucsc.edu/). Genome sequencing reference strain for Saccharomyces uvarum, sacBay MIT, was obtained from the Saccharomyces Genome Database (SGD, http://www.yeastgenome.org).
SOLiD Sequencing
The genomic DNA of three strains of S. pastorianus was sequenced using Next Generation Sequencing Applied Biosystems SOLiD 4 platform to generate 50 bp single-end reads. Using BFAST (http://sourceforge.net/projects/bfast/files/), the reads were mapped to the S. cerevisiae reference genome “sacCer2” obtained from UCSC (http://genome.ucsc.edu/), which includes 16 chromosomes, the mitochondrial genome and the 2 micron plasmid. The ‘-a 3’ flag of the post-process step was used to obtain unique best scoring alignments. The S. cerevisiae ORFs were used to find S. eubayanus consensus ORFs in the S. eubayanus reference strain “sacBay MIT” obtained from SGD (http://www.yeastgenome.org/). BFAST files were filtered to retrieve sets of reads with 0, 0–1 or 0–5 mismatches to each reference genome. Generally, 0 mismatches was found to be the most suitable cut-off value for the S. cerevisiae data and 0–5 mismatches for the S. eubayanus data, having the best agreement to previous microarray data by Dunn and Sherlock (2008).
Chromosomal Copy Number Analysis and Breakpoint Identification
We used the S. pastorianus SOLiD sequence mapped to the pre-annotated S. cerevisiae genome sequence via the UCSC Genome Browser (http://genome.ucsc.edu/) to identify both S. cerevisiae chromosome copy number and potential chimeric chromosomes comprising both S. cerevisiae and S. eubayanus sequence. These candidate breakpoint regions were identified visually by their abrupt and sustained reduction in S. cerevisiae reads along a chromosome. Due to difficulties in mapping and analysing repetitive regions, telomeres were excluded from the analysis. Similarly, changes in read number due to the presence of a yeast transposon (Ty) or other repetitive element were excluded.
To estimate chromosomal copy number we first broke down each genome into regions. The S. cerevisiae genome was broken down into 26 regions (labeled in Figure 1) fully covering either side of each breakpoint based on the mapping to the UCSC Genome Browser (Table S3). Since S. eubayanus genomes is only present in contigs, we chose a representative selection of 47 regions across each chromosome to sample regions either side of each breakpoint observed in the mapped S. cerevisiae data (Figure 1, Table S4). Pre-existing rearrangements between chromosomes within the S. uvarum genome (chromosomes II–IV, VI–X and VIII–XV) [45] were taken into account when estimating copies of S. eubayanus chromosomes. We have assumed the S. eubayanus genome to be collinear with the highly related S. uvarum genome. The ‘coverage’ tool from the Bedtools suite of programs (http://bedtools.readthedocs.org/en/latest/) was used to obtain the read depth for every position in each of the regions under consideration. The analysis used the mapped reads at a cutoff of 0 mismatches for the S. cerevisiae data and 0–5 mismatches for the S. eubayanus data as previously described. For each region the median depth of coverage values were calculated using a custom Peal script excluding 0 coverage values.
To ascertain copy number of each S. cerevisiae and S. eubayanus regions the read depth data of each region were grouped into clusters using hierarchical cluster analysis software [45]. The analyses were performed separately for each strain of S. pastorianus using Ward’s method to generate read depth clusters which are displayed as dendrograms (Figure S3). Using the mean depth of the regions within each cluster the copy number was then assigned (Table S3 and Table S4). The regional copy numbers, in conjunction with breakpoint data, was used to calculate S. cerevisiae, S. eubayanus and chimeric chromosomal copy number.
PCR Amplification and Sequencing of Breakpoints
We used the S. pastorianus SOLiD sequence mapped to the pre-annotated S. cerevisiae genome sequence via the UCSC Genome Browser (http://genome.ucsc.edu/) to identify potential chimeric chromosomes comprising both S. cerevisiae and S. eubayanus sequence. These candidate breakpoint regions were identified by their abrupt and sustained reduction in S. cerevisiae read depth along a chromosome. Species-specific primers were designed to flank each predicted breakpoint area. The S. cerevisiae primers were designed using S. cerevisiae sequence obtained directly from the UCSC Genome Browser (http://genome.ucsc.edu/). The S. eubayanus primers were designed by finding the S. uvarum orthologue of the nearest S. cerevisiae gene using the SGD Synteny Viewer (http://www.yeastgenome.org). This orthologue was then mapped to S. pastorianus to find consensus sequences for the S. eubayanus portion of the S. pastorianus genome. The S. eubayanus-specific primer was then designed within this consensus sequence. Candidate primers were generated for both S. cerevisiae and S. eubayanus sequences using Primer 3 (http://frodo.wi.mit.edu). To circumvent the potential for non-specific binding between the two closely related subgenomes, these primers were then carefully selected for species-specificity using the Fungal BLAST tool in SGD (http://www.yeastgenome.org). For ease of amplification, primers were designed to anneal no more than a few thousand base pairs apart but with sufficient sequence either side of the breakpoint for clear identification of each subgenome. Primer sequences are available in Table S5.
PCR conditions were optimised for each breakpoint to obtain pure homogeneous chimeric sequence. The PCR product was separated by electrophoresis on 1% (w/v) agarose gel. PCR products were purified prior to sequencing using QIAquick PCR Purification Kit (Qiagen, UK). The purified PCR products were Sanger sequenced at GATC Biotech (Germany).
Multiple Alignment of Sanger Sequenced Breakpoints to Parental Subgenomes
Each sequence covering each breakpoint in each S. pastorianus strain was aligned to S. cerevisiae (Scer S288C, Saccharomyces Genome Database) and S. eubayanus (FM318, http://hittinger.genetics.wisc.edu/index.html), where available, or S. uvarum (MIT_Sbay or WashU_Sbay, Saccharomyces Genome Database) using Clustal Omega (http://www.clustal.org/). Low quality ends of breakpoint sequences were trimmed before alignment. The breakpoint region in each sequence was determined as the area of identical nucleotides between the parental species flanked by S. cerevisiae-like sequence on one side and S. eubayanus-like sequence on the other side of the region. The start of S. cerevisiae-like sequence was determined by nucleotide match to S. cerevisiae but mismatch to either S. eubayanus or S. uvarum. The start of S. eubayanus-like sequence was determined by nucleotide match to S. eubayanus or S. uvarum but mismatch to S. cerevisiae.
Analysis of Chimeric Gene Length and Reading Frame
The proposed length of each chimeric gene was determined by merging S. cerevisiae sequence and S. eubayanus or S. uvarum sequence at the junction of the pre-determined breakpoint region. Reading frame was checked using Expasy Translate (http://web.expasy.org/translate/).
Analysis of Sequence Identity
Percentage nucleotide identity between each subgenome and S. pastorianus was calculated using S. cerevisiae and either S. eubayanus (obtained from http://hittinger.genetics.wisc.edu/, where available) or S. uvarum sequences obtained from SGD (http://www.yeastgenome.org/) and Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/) Amino acid identity was calculated similarly. Chimeric nucleotide sequences were first converted to protein sequences using Expasy Translate Tool (http://web.expasy.org/translate/).
Data Deposition
Raw reads from this study have been deposited at the European Nucleotide Archive under the accession number PRJEB4654 at http://www.ebi.ac.uk/ena/data/view/PRJEB4654; Sanger sequenced reads covering the breakpoint regions have been submitted to the European Nucleotide Archive under the accession numbers HG803141–HG803169 at http://www.ebi.ac.uk/ena/data/view/HG803141-HG803169.
Supporting Information
Figure S1.
Mapping of the 2-micron plasmid to S. cerevisiae sequence. The 2-micron plasmid DNA from each strain of S. pastorianus is mapped to S. cerevisiae sequence using the UCSC Genome Browser. The scale on the Y axis is capped at a read depth of 4500.
https://doi.org/10.1371/journal.pone.0092203.s001
(TIF)
Figure S2.
Multiple alignment of each S. pastorianus breakpoint sequence to the parental species. The region sequenced over each breakpoint in each S. pastorianus strain was aligned with S. cerevisiae (Scer, Saccharomyces Genome Database) and either S. eubayanus (FM318, http://hittinger.genetics.wisc.edu/index.html) or S. uvarum (MIT_Sbay or WashU_Sbay, Saccharomyces Genome Database) ORF sequences, using Clustal Omega. The two intergenic breakpoints were aligned using nucleotide sequence upstream or downstream from the nearest ORF. The ORF sequences obtained from the Saccharomyces Genome Database are taken from Cliften et al. [46] and Kellis et al. [45]. Any low quality ends of each breakpoint sequence were trimmed before alignment. Breakpoint area is demarcated by underlined sequence. Nucleotides shared between both parental species are highlighted in grey and are flanked by the first unique S. cerevisiae nucleotide (shown in blue) and the first unique S. eubayanus or S. uvarum nucleotide (shown in red). All gene sequences are 5′-3′.
https://doi.org/10.1371/journal.pone.0092203.s002
(PDF)
Figure S3.
Hierarchical cluster analysis of read depth of S. pastorianus chromosomes. The median read depth for 26 regions covering the S. cerevisiae-like chromosomes (shown in Figure 1) and 47 regions across the S. eubayanus-like chromosomes (sample contigs across each chromosome) were clustered independently and for each strain of S. pastorianus. Results from each hierarchical cluster analysis using Ward’s method are shown as a dendrogram. The blue boxes indicate the copy number assigned to that cluster.
https://doi.org/10.1371/journal.pone.0092203.s003
(PDF)
Table S1.
Breakpoints which were not successfully amplified by PCR.
https://doi.org/10.1371/journal.pone.0092203.s004
(DOC)
Table S2.
Analysis of the breakpoint region in each strain of S. pastorianus.
https://doi.org/10.1371/journal.pone.0092203.s005
(DOC)
Table S3.
Analysis of the copy number of S. cerevisiae-derived chromosomes in each strain of S. pastorianus using median read depth across multiple regions.
https://doi.org/10.1371/journal.pone.0092203.s006
(XLSX)
Table S4.
Analysis of the copy number of S. eubayanus-derived chromosomes in each strain of S. pastorianus using median read depth across multiple regions.
https://doi.org/10.1371/journal.pone.0092203.s007
(XLSX)
Table S5.
Primer sequences for amplification of breakpoint regions.
https://doi.org/10.1371/journal.pone.0092203.s008
(DOC)
Acknowledgments
The authors wish to thank Dr Casey Bergman for help in designing the bioinformatics approach and suggesting analysis of breakpoint reuse.
Author Contributions
Conceived and designed the experiments: DD SKH. Performed the experiments: SKH. Analyzed the data: SKH IJD SCL DD. Contributed reagents/materials/analysis tools: SCL DD. Wrote the paper: SKH IJD SCL DD.
References
- 1. de Barros Lopes M, Bellon JR, Shirley NJ, Ganter PF (2002) Evidence for multiple interspecific hybridization in Saccharomyces sensu stricto species. FEMS yeast research 1: 323–331.
- 2. Naumova ES, Naumov GI, Masneuf-Pomarà ¨de I, Aigle M, Dubourdieu D (2005) Molecular genetic study of introgression between Saccharomyces bayanus and S. cerevisiae. Yeast (Chichester, England) 22: 1099–1115.
- 3. Belloch C, Perez-Torrado R, Gonzalez SS, Perez-Ortin JE, Garcia-Martinez J, et al. (2009) Chimeric Genomes of Natural Hybrids of Saccharomyces cerevisiae and Saccharomyces kudriavzevii. Appl Environ Microbiol 75: 2534–2544.
- 4. Martini AV, Martini A (1987) Three newly delimited species of Saccharomyces sensu stricto. Antonie van Leeuwenhoek 53: 77–84.
- 5. Masneuf I, Hansen J, Groth C, Piskur J, Dubourdieu D (1998) New Hybrids between Saccharomyces Sensu Stricto Yeast Species Found among Wine and Cider Production Strains. Applied and Environmental Microbiology 64: 3887–3892.
- 6. Groth C, Hansen J, Piškur J (1999) A natural chimeric yeast containing genetic material from three species. International Journal of Systematic and Evolutionary Microbiology 49: 1933–1938.
- 7. Lopes CA, Barrio E, Querol A (2010) Natural hybrids of S. cerevisiae×S. kudriavzevii share alleles with European wild populations of Saccharomyces kudriavzevii. FEMS Yeast Research 10: 412–421.
- 8. Hansen J, Kielland-Brandt MC (1994) Saccharomyces carlsbergensis contains two functional MET2 alleles similar to homologues from S. cerevisiae and S. monacensis. Gene 140: 33–40.
- 9. Rainieri S, Kodama Y, Kaneko Y, Mikata K, Nakao Y, et al. (2006) Pure and Mixed Genetic Lines of Saccharomyces bayanus and Saccharomyces pastorianus and Their Contribution to the Lager Brewing Strain Genome. Appl Environ Microbiol 72: 3968–3974.
- 10. Dunn B, Sherlock G (2008) Reconstruction of the genome origins and evolution of the hybrid lager yeast Saccharomyces pastorianus. Genome Res 18: 1610–1623.
- 11.
Libkind D, Hittinger CT, Valério E, Gonçalves C, Dover J, et al.. (2011) Microbe domestication and the identification of the wild genetic stock of lager-brewing yeast. Proceedings of the National Academy of Sciences.
- 12. Belloch C, Orlic S, Barrio E, Querol A (2008) Fermentative stress adaptation of hybrids within the Saccharomyces sensu stricto complex. International Journal of Food Microbiology 122: 188–195.
- 13. Nakao Y, Kanamori T, Itoh T, Kodama Y, Rainieri S, et al. (2009) Genome Sequence of the Lager Brewing Yeast, an Interspecies Hybrid. DNA Research 16: 115–129.
- 14. Liti G, Peruffo A, James SA, Roberts IN, Louis EJ (2005) Inferences of evolutionary relationships from a population survey of LTR-retrotransposons and telomeric-associated sequences in the Saccharomyces sensu stricto complex. Yeast (Chichester, England) 22: 177–192.
- 15. Bond U, Neal C, Donnelly D, James TC (2004) Aneuploidy and copy number breakpoints in the genome of lager yeasts mapped by microarray hybridisation. Current genetics 45: 360–370.
- 16. Sato M, Maeba H, Watari J, Takashio M (2002) Analysis of an inactivated Lg-FLO1 gene present in bottom-fermenting yeast. Journal of bioscience and bioengineering 93: 395–398.
- 17. Dunham MJ, Badrane H, Ferea T, Adams J, Brown PO, et al. (2002) Characteristic Genome Rearrangements in Experimental Evolution of Saccharomyces cerevisiae. Proceedings of the National Academy of Sciences of the United States of America 99: 16144–16149.
- 18. Fischer G, James SA, Roberts IN, Oliver SG, Louis EJ (2000) Chromosomal evolution in Saccharomyces. Nature 405: 451–454.
- 19. Rachidi N, Barre P, Blondin B (1999) Multiple Ty-mediated chromosomal translocations lead to karyotype changes in a wine strain of Saccharomyces cerevisiae. Molecular and General Genetics MGG 261: 841–850.
- 20. Di Rienzi SC CD, Raghuraman MK, Brewer BJ (2009) Fragile genomic sites are associated with origins of replication. Genome Biol Evol 1: 350–363.
- 21. James TC, Usher J, Campbell S, Bond U (2008) Lager yeasts possess dynamic genomes that undergo rearrangements and gene amplification in response to stress. Current genetics 53: 139–152.
- 22. Perez-Ortin JE, Querol A, Puig S, Barrio E (2002) Molecular characterization of a chromosomal rearrangement involved in the adaptive evolution of yeast strains. Genome Res 12: 1533–1539.
- 23. Infante JJ, Dombek KM, Rebordinos L, Cantoral JM, Young ET (2003) Genome-Wide Amplifications Caused by Chromosomal Rearrangements Play a Major Role in the Adaptive Evolution of Natural Yeast. Genetics 165: 1745–1759.
- 24. Chang S-L, Lai H-Y, Tung S-Y, Leu J-Y (2013) Dynamic Large-Scale Chromosomal Rearrangements Fuel Rapid Adaptation in Yeast Populations. PLoS Genet 9: e1003232.
- 25. Delneri D, Colson I, Grammenoudi S, Roberts IN, Louis EJ, et al. (2003) Engineering evolution to study speciation in yeasts. Nature 422: 68–72.
- 26. Rainieri S, Kodama Y, Nakao Y, Pulvirenti A, Giudici P (2008) The inheritance of mtDNA in lager brewing strains. FEMS Yeast Research 8: 586–596.
- 27. Gerton JL, DeRisi J, Shroff R, Lichten M, Brown PO, et al. (2000) Global mapping of meiotic recombination hotspots and coldspots in the yeast Saccharomyces cerevisiae. Proceedings of the National Academy of Sciences 97: 11383–11390.
- 28.
Bond U, Allen IL, Sima S, Geoffrey MG (2009) Chapter 6 The Genomes of Lager Yeasts. Academic Press. 159–182.
- 29. Brown C, Murray A, Verstrepen K (2010) Rapid expansion and functional divergence of subtelomeric gene families in yeasts. Curr Biol 20: 895–903.
- 30. Hao W, Palmer JD (2011) HGT turbulence: Confounding phylogenetic influence of duplicative horizontal transfer and differential gene conversion. Mob Genet Elements 1: 256–261.
- 31. Salichos L, Rokas A (2013) Inferring ancient divergences requires genes with strong phylogenetic signals. Nature 497: 327–331.
- 32. Usher J, Bond U (2009) Recombination between homoeologous chromosomes of lager yeasts leads to loss of function of the hybrid GPH1 gene. Applied and environmental microbiology 75: 4573–4579.
- 33. Dunn B, Paulish T, Stanbery A, Piotrowski J, Koniges G, et al. (2013) Recurrent Rearrangement during Adaptive Evolution in an Interspecific Yeast Hybrid Suggests a Model for Rapid Introgression. PLoS Genet 9: e1003366.
- 34. Yang R, Wek SA, Wek RC (2000) Glucose limitation induces GCN4 translation by activation of Gcn2 protein kinase. Mol Cell Biol 20: 2706–2717.
- 35. Nadeau JH, Taylor BA (1984) Lengths of chromosomal segments conserved since divergence of man and mouse. Proc Natl Acad Sci U S A 81: 814–818.
- 36. Murphy WJ, Larkin DM, der Wind AE-v, Bourque G, Tesler G, et al. (2005) Dynamics of Mammalian Chromosome Evolution Inferred from Multispecies Comparative Maps. Science 309: 613–617.
- 37. Larkin DM, Pape G, Donthu R, Auvil L, Welge M, et al. (2009) Breakpoint regions and homologous synteny blocks in chromosomes have different evolutionary histories. Genome Res 19: 770–777.
- 38. Ma J, Zhang L, Suh BB, Raney BJ, Burhans RC, et al. (2006) Reconstructing contiguous regions of an ancestral genome. Genome Res 16: 1557–1565.
- 39. Pevzner P, Tesler G (2003) Human and mouse genomic sequences reveal extensive breakpoint reuse in mammalian evolution. Proc Natl Acad Sci U S A 100: 7672–7677.
- 40.
Ruiz-Herrera A, Castresana J, Robinson TJ (2006) Is mammalian chromosomal evolution driven by regions of genome fragility? Genome Biol 7.
- 41. von Grotthuss M, Ashburner M, Ranz JM (2010) Fragile regions and not functional constraints predominate in shaping gene organization in the genus Drosophila. Genome Res 20: 1084–1096.
- 42. Szostak JW, Orr-Weaver TL, Rothstein RJ, Stahl FW (1983) The double-strand-break repair model for recombination. Cell 33: 25–35.
- 43. Hoang ML, Tan FJ, Lai DC, Celniker SE, Hoskins RA, et al. (2010) Competitive Repair by Naturally Dispersed Repetitive DNA during Non-Allelic Homologous Recombination. PLoS Genet 6: e1001228.
- 44. Baudin A, Ozier-Kalogeropoulos O, Denouel A, Lacroute F, Cullin C (1993) A simple and efficient method for direct gene deletion in Saccharomyces cerevisiae. Nucleic Acids Res 21: 3329–3330.
- 45. Kellis M, Patterson N, Endrizzi M, Birren B, Lander ES (2003) Sequencing and comparison of yeast species to identify genes and regulatory elements. Nature 423: 241–254.
- 46. Cliften P, Sudarsanam P, Desikan A, Fulton L, Fulton B, et al. (2003) Finding Functional Features in Saccharomyces Genomes by Phylogenetic Footprinting. Science 301: 71–76.