Considerable amounts of molecular and genetic data indicate that interspecific hybridisation may not be rare among natural strains of Saccharomyces sensu stricto. Although a post-zygotic barrier operating during meiosis usually prevents the production of viable spores, stable hybrids can arise which can even evolve into distinct species. This study was aimed to analyse the genome of a fertile Saccharomyces cerevisiae×S. uvarum hybrid and monitor its changes over four filial generations of viable spores. The molecular genetic analysis demonstrated that the two species did not contribute equally to the formation and stabilisation of the hybrid genome. S. cerevisiae provided the mitochondrial DNA and the more stable part of the nuclear genome. The S. uvarum part of the hybrid nuclear genome became progressively smaller by loosing complete chromosomes and genetic markers in the course of successive meiotic divisions. Certain S. uvarum chromosomes were eliminated and/or underwent rearrangements in interactions with S. cerevisiae chromosomes. Numerous S. uvarum chromosomes acquired S. cerevisiae telomere sequences. The gradual elimination of large parts of the S. uvarum genome was associated with a progressive increase of sporulation efficiency. We hypothesise that this sort of genomic alterations may contribute to speciation in Saccharomyces sensu stricto.
Saccharomyces“sensu stricto” is a group of closely related species capable of mating with each other and producing viable interspecific hybrids (e.g. [1–5]). Their evolutionary separation is ensured by meiotic incompatibility (post-zygotic species barrier) preventing sporulation or the production of viable ascospores in the hybrids. Interspecific hybrids can arise in natural environments (e.g. [6,7]) and can be constructed under laboratory conditions (e.g. [2,8–14]). Natural interspecific mating is supposed to account for the appearance of hybrid yeasts of biotechnological importance such as the wine yeast S6U, the cider yeast CID1 [15,16] and the lager yeasts BRYC 32 and NCYC 1324 . At least two Saccharomyces sensu stricto species have also been proposed to have evolved from interspecific hybridisation. Strains of S. pastorianus (including S. carlsbergensis) were found to be hybrids of S. cerevisiae and either S. bayanus or S. monacensis, an old species grouped now in S. pastorianus[17,18]. The type strain of S. bayanus CBS 380 turned out to be a hybrid of S. cerevisiae and S. uvarum.
The hybrid strains and species usually do not have alloeuploid but chimerical genomes, composed of fractions of genomes of two or even more species , suggesting that the genomes of the founding alloploid zygotes have undergone reduction and rearrangements to attain a stable structure and organisation. The aim of this study was to investigate how the alloploid hybrid genome of the zygote changes to become more stable. Since S. cerevisiae and S. uvarum (frequently confused with S. bayanus due to inconsistencies in taxonomic literature; for a review see ) frequently occur together during wine fermentation and can also form hybrids [6,7,21–23], and the karyotypes of these two species differ markedly (e.g. [19,22–24]), we decided to study the process of post-zygotic genome stabilisation in a S. cerevisiae×S. uvarum hybrid. However, the allodiploid hybrids of these species produce non-viable spores (e.g. ), which reduces their prospects to survive unfavourable environmental changes (e.g. starvation) and thus impairs their chances to persist for longer periods of time. To avoid this post-zygotic barrier we selected a fertile hybrid, supposed to be allotetraploid, and monitored the changes of its genome through four successive generations of ascospores.
2 Materials and methods
2.1 Yeast strains and media
Saccharomyces strains used are listed in Table 1. All strains were maintained routinely on YPGA agar medium (1% yeast extract, 1% peptone, 2% glucose and 2% agar) at 26 °C. The liquid medium YPGL was as YPGA without agar. Melibiose fermentation was tested in YPML (YPGL containing 2% melibiose instead of glucose). Sporulation was examined on acetate sporulation medium (1% K-acetate, 0.1% yeast extract, 0.05% glucose and 2% agar) . The minimal medium MMA (1.5% (NH4)2SO4, 0.1% KH2PO4, 0.05% MgSO4·7H2O, 1% glucose, vitamins and 2% agar) was described by Sipiczki and Ferenczy .
amel: melibiose utilisation; ts: unable to grow at 37 °C.
bYGSC: Yeast Genetic Stock Center, Berkeley, USA. CBS: Centraalbureau voor Schimmelcultures, Utrecht, The Netherlands. ATCC: American Type Culture Collection, Manassas, VA, USA. CLIB: Collection de Levures d'Intérêt Biotechnologique, INRA, Versailles-Grignon, France.
2.2 Mutagenesis and isolation of auxotrophic mutants
Suspensions of sporulating cultures were treated with Zymolyase (ICN) to liberate spores from asci, spread onto YPD plates and subjected to UV irradiation for 10 s (survival rate: 10%). After several days of incubation at 26 °C the colonies produced were replica-plated onto MMA plates. Colonies incapable of growth on MMA were isolated and their growth requirements were determined by the paper–disc method .
2.3 Hybrid isolation
Cultures of the hybridisation partners, grown on sporulation medium at 26 °C for 3 days, were suspended in sterile water and mixed. After 1.5 h of incubation at room temperature, samples were spread onto MMA plates. The plates were incubated at 37 °C for several days. The colonies produced were isolated as putative interspecific hybrids.
2.4 Ascus dissection and tetrad analysis
Samples of the hybrid cultures were streaked on sporulation medium and incubated for 5 days at 26 °C. Four-spored asci were isolated, dissected and their spores were separated on YPGA plates by micromanipulation. The colonies produced from the spores were isolated and their auxotrophic markers were determined by replica-plating on MMA plates supplemented with uracil or leucine. Melibiose fermentation was examined in Durham tubes filled with YPML.
2.5 PCR, PCR-RFLP and sequencing
Yeasts were grown overnight in YPGL with shaking at 26 °C. Genomial DNA was isolated from 5-ml samples of cultures as described by Querol et al. . PCR amplification was performed with the primer pairs listed in Table 2. The amplified fragments were digested with restriction enzymes and the resulting sub-fragments were analysed by electrophoresis in 1.4% agarose gel with 0.5× TBE containing ethidium-bromide. For nucleotide sequence determination, the amplified DNA was sequenced with an ABI PRISM 3700 (Applied Biosystems, Foster City, CA) sequencer using the PCR primers.
Chromosomal DNA was prepared in agarose plugs as described by Nguyen and Gaillardin . Chromosomes were separated in 1% agarose gel (Chromosomal grade, BioRad, Hercules, CA) prepared in 0.5× TBE with a CHEF-Mapper apparatus (Bio-Rad). The running parameters were: 200 V, 14 °C, linear ramping from 40 to 120 s for 24 h. The chromosomal bands were visualised by staining with ethidium-bromide and de-staining in sterile water.
2.7 Southern hybridisation
DNA blotting on GeneScreen (Princeton, NJ) membrane was performed as described [29,30]. Plasmid pJY5 harbouring the Y′ sequence  was labelled with Megaprime labelling kit (Amersham, Little Chalfont, UK) and [α-32P]dCTP. The labelled DNA was hybridised to the membrane overnight at 45 °C in the presence of 50% formamide. After hybridisation the membrane was washed first at 50 °C in 2× SSC, 0.1% SDS and then twice at 65 °C in 0.1× SSC, 0.1% SDS.
2.8 Mitochondrial DNA (mtDNA) extraction and restriction analysis
The mitochondrial DNA was extracted from exponential-phase cells according to the method of Defontaine et al.  with some modifications . The isolated mtDNA was digested with EcoRV (BRL, Carlsbad, CA) or RsaI (New England BioLabs, Frankfurt am Main, Germany), and the fragments were separated by electrophoresis in 0.7% agarose, 0.5× TBE.
3.1 Mutant isolation and interspecific hybridisation
Saccharomyces uvarum and S. cerevisiae differ in two taxonomic traits: the former does not grow at 37 °C and can utilise melibiose as carbon source, whereas the latter is able to grow at 37 °C and cannot ferment melibiose . To have additional distinctive markers for segregational analysis of their hybrids, we did not hybridise wild-type strains but auxotrophic mutants. The S. cerevisiae strain 10-170 was auxotrophic for leucine and methionine, but the latter marker was leaky. To have an S. uvarum partner with complementary auxotrophy, we mutagenised a sporulating culture of the S. uvarum wine strain 10-408, isolated from a Tokaj winery . Nineteen mutants were obtained: 16 ura−, 2 ade− and 1 arg−. Seven of the ura− mutants were resistant to 5′-fluoro-orotic acid, indicating that they were defective in the gene URA3. From these mutants the one (10-522) with the highest sporulation efficiency and spore viability (96%) was chosen for hybridisation with the S. cerevisiae strain 10-170. To confirm the existence of a mutation in 10-522, a genomic fragment was amplified from it with a pair of primers specific for the S. uvarum URA3 gene. The fragment obtained was sequenced in both directions and the sequence was compared with that of the wild-type S. uvarum URA3 gene (MIT_Sbay_C87_6517; WashU_Sbay_Contig645.49 http://db.yeastgenome.org/cgi-bin/FUNGI/nph-showAlign?locus=YEL021W). The alignment revealed a G-to-C nucleotide substitution at position 357 and a T-to-C substitution at position 461.
Upon mass-mating of 10-170 and 10-522 cells, 21 prototrophic colonies were isolated which were melibiose-positive and grew at 37 °C. All isolates were sporulation-proficient and significant proportions of their asci contained four spores. The hybrid (designated H01) with the highest percentage of sporulation (50%) and with high proportion of four-spored asci (30%) was selected for genetic analysis.
3.2 Isolation of generations F1–F4
Ten four-spored asci were isolated from H01 for tetrad analysis. Of the spores deliberated from the asci 70% germinated and produced colonies (F1 spore clones) but only three tetrads had four viable spores. Of the clones developed by the viable spores 50% could sporulate on sporulation medium, and the complete tetrads (No. 1 and 4) contained two sporulating and two non-sporulating clones.
One of the sporulating spore clones of the complete tetrad No. 1 was used for the isolation of F2 tetrads. To obtain the F3 generation, asci were dissected from spore clones of two complete F2 tetrads. Finally, F4 tetrads were obtained from an F3 spore clone. The genealogy of the tetrads analysed is shown in Fig. 1. The sporulation efficiency of the spore clones increased from generation to generation, up to 91% in F4.
The hybrid and its descendants investigated in this study. (A) The genealogy of the tetrads. (B) Sporulation in the hybrid and the F1–F4 generations.
3.3 Marker segregation
When tested for nutritional requirements, 54% of the F1 spore clones were prototrophic and the auxotrophs were all leu−. The complete tetrads (all four spores produced clones) contained 2 leu− and 2 leu+ spores. Since the F2 tetrads were obtained from asci of the leu− spore clone 1c of tetrad No. 1, all F2–F4 spores were leu−.
None of the F1 spores required uracil, indicating that either the ura− marker of the S. uvarum parent got lost during sporulation, or the spores were heterozygous. To test the latter possibility, the hybrid and the spore clones were examined for the presence of the parental alleles of URA3. Using species-specific primers we detected both genes in the hybrid genome (Fig. 2), and the sequence of the fragment amplified with the S. uvarum-specific primers contained both mutations detected in the S. uvarum parent. Both genes were also present in the F1–F4 descendants (data not shown).
PCR-RFLP analysis of URA3. Lane 1: Size marker. Lanes 2 and 5: 10-170. Lanes 3 and 6: hybrid H01. Lanes 4 and 7: 10-522. Lanes 2–4 were amplified with S. uvarum-specific primers; 5–7 were amplified with S. cerevisiae-specific primers.
The temperature sensitivity of the S. uvarum parent did not segregate either. None of the spore clones of any generation were sensitive to 37 °C. In contrast, the sexual types segregated in F1. Of the F1 spore clones 65% sporulated in pure culture. The spores of the complete tetrads analysed produced two sporulating clones and two non-sporulating clones. In tetrad No. 1, the non-sporulating clones were sterile, whereas in tetrad No. 4 they were of mating type. The sporulating clones were most probably homothallic (as the S. uvarum parent) because all descendants of clone 1c, from which F2 was generated, were self-sporulating.
3.4 PCR-RFLP analysis of the segregation of chromosomal genes
To extend the segregation analysis to more regions of the genome, we choose six additional genes or regions (MET2, ITS1-5.8SrDNA-ITS2, NTS-ETS, YCL008c, HIS4 and MET10) located in three different chromosomes in S. cerevisiae (Table 3). Since we had no markers for these genes, we monitored their segregation by PCR-RFLP.
The amplified MET2 genes were digested with EcoRI and PstI. Both enzymes have recognition sites only in one of the MET2 genes [18,34]. EcoRI cuts the S. cerevisiae MET2 fragment but does not cleave the S. uvarum MET2 fragment, whereas PstI cleaves the S. uvarum gene but not that of S. cerevisiae. We found that the hybrid H01 and all its descendants in all generations possessed the MET2 genes of both parents (results not shown).
The ITS1-5.8SrDNA-ITS2 regions of the two species can be differentiated by digestion with HaeIII [13,35,36]. This enzyme cuts the PCR product of S. cerevisiae into four sub-fragments and the PCR product of S. uvarum into three sub-fragments. The hybrid H01 had both the S. cerevisiae and the S. uvarum regions, as did most of its descendants (Fig. 3). The two exceptions were 1a of F1 and 6c of F2. The former clone showed S. cerevisiae pattern, whereas the latter clone had an extra sub-fragment. The new sub-fragment indicates that recombination might have occurred between the S. cerevisiae part and the S. uvarum part of the hybrid genome in the chromosomal region encoding rRNAs. Consistently with the presence of the ITS regions of both species, the adjacent NTS-ETS region  also showed hybrid patterns in the F1 spores when digested with BanI (Fig. 3).
PCR-RFLP analysis of ITS1-5.8S rDNS-ITS2 and NTS2-ETS. Lanes 1–15: ITS1-5.8S rDNS-ITS2 digested with HaeIII. Lane 1: 10-170. Lane 2: Hybrid H01. Lane 3: 10-522. Lanes 4–7: F1 tetrad No. 1. Lanes 8–11: F2 tetrad No. 6. Lanes 12–15: F3 tetrad. Lanes 16–23: NTS2-ETS digested with BanI. Lane 16: 10-522. Lane 17: 10-170. Lane 18: hybrid H01. Lane 19: size marker. Lanes 20–23: F1 tetrad No. 1.
The YCL008c PCR-RFLP can also be used to distinguish between S. cerevisiae and S. uvarum. The PCR product of S. cerevisiae has one recognition site for EcoRV but has no recognition site for PstI. The PCR product of S. uvarum is cut by PstI into three fragments but contains no recognition site for EcoRV. We found both parental patterns in the hybrid. In the F1 generation two spores showed hybrid patterns and two spores had patterns characteristic of the S. cerevisiae parent. Since the F1 clone 1c, from which the F2 generation was isolated, showed S. cerevisiae pattern, all its F2 and F4 descendants had this pattern (Fig. 4).
PCR-RFLP analysis of YCL008c. Digestion with PstI and EcoRV. Lane 1: 10-170. Lane 2: hybrid H01. Lane 3: 10-522. Lanes 4–7: F1 tetrad No. 1. Lanes 8–11: F2 tetrad No. 6. Lanes 12–15: F3 tetrad.
To examine the segregation of the HIS4 genes we used HindIII that has recognition sites in the S. cerevisiae gene but not in the S. uvarum gene . This enzyme cuts the PCR product of S. cerevisiae into three smaller fragments. We found that the hybrid had both parental patterns, whereas all descendants only showed the S. cerevisiae pattern (Fig. 5).
PCR-RFLP analysis of HIS4. Digestion with HindIII. Lane 1: 10-170. Lane 2: hybrid H01. Lane 3: 10-522. Lanes 4–7: F1 tetrad No. 1. Lanes 8–11: F2 tetrad No. 6. Lanes 12–15: F3 tetrad.
We used an S. cerevisiae-specific primer and an S. pastorianus-specific primer together to amplify the MET10 genes . With these primers we obtained a 1.1-kb-long fragment from S. uvarum, a 1.7-kb-long fragment from S. cerevisiae and both types of fragments from the hybrid (Fig. 6). The F1 generation showed a 2:2 segregation of the hybrid pattern and the S. cerevisiae pattern. 1c Produced both fragments and this pattern did not segregate in the progenies.
The MET10 PCR products obtained with mixed primers. Lane 1: 10-170. Lane 2: hybrid H01. Lane 3: 10-522. Lanes 4–7: F1 tetrad No. 1. Lanes 8–11: F2 tetrad No. 6. Lanes 12–15: F3 tetrad.
3.5 Electrophoretic karyotyping and Y′ probe
The parental strains markedly differed in the chromosomal patterns obtained by electrophoretic karyotyping. The S. cerevisiae parent strain showed two doublets: a smaller one composed of the chromosomes I and VI and a larger one which consisted of the chromosomes V and VIII as in many other S. cerevisiae strains. The S. uvarum parent strain exhibited a chromosomal pattern similar to that of the type strain CBS 395T (CLIB 251). As expected, the hybrid possessed the chromosomal bands of both of them (Fig. 7). The karyotypes of the F1 descendants also contained most of the chromosomes of both parents, but certain S. uvarum bands showed 2:2-segregation (Fig. 8). In the F1 tetrad shown in Fig. 8, the chromosomes 2 and 5 of S. uvarum can only be seen in two spore clones (for numbering of chromosomes see Fig. 7). The clones, which did not have chromosome 2, were auxotrophic for leucine, indicating that the S. uvarum LEU2 gene is located in this chromosome. Further segregation was observed in F2. The bands corresponding in size to chromosomes 4 and 5 of S. uvarum were not seen in two spores of the F2 tetrad, shown in Fig. 8. The chromosomal pattern of the F2 clone 6b had an extra band which did not correspond in size to any parental bands. It might have arisen from recombination between non-homologous chromosomes.
Chromosomal patterns of the parental strains and their hybrid compared with S. cerevisiae YNN 295 standard strain and S. uvarum. Sc: YNN 295; 10: S. cerevisiae 10-170; H01: hybrid; m9: S. uvarum 10-522; SuT: S. uvarum CBS 395T. Arrow head indicates the doublet I–VI. Chromosome numbering is according to Nguyen et al. .
Chromosomal patterns and the Y′-hybridisation of the parental strains, their hybrid and the descendants. Lane 1: 10-522. Lane 2: 10-170. Lane 3: hybrid H01. Lanes 4–7: F1 tetrad No. 1. Lanes 8–11: F2 tetrad No. 6. “III” marks chromosome III of S. cerevisiae that gave poor hybridisation signal. Italic numerals “2”, “4” and “5” mark S. uvarum chromosomes 2, 4 and 5 that show 2:2 segregation. Italic numeral “1” marks S. uvarum chromosome 1 which acquired sequences hybridising with the S. cerevisiae telomeric probe. X marks the recombinant chromosome unique in size.
Further chromosomal rearrangements were detected when the karyotypes were probed with the labelled telomeric sequence Y′. This sequence is specific for S. cerevisiae and can be used to differentiate its chromosomes from those of S. uvarum. In our case, the Y′ probe did not hybridise to chomosome III of the S. cerevisiae parent and the hybrid, and only gave faint signals with the corresponding bands of the spore clones (Fig. 8). However, it hybridised to chromosome 1 of S. uvarum in two spores of the F1 tetrad, indicating that in these spores the S. uvarum chromosome had acquired S. cerevisiae telomeric sequences. Acquirement of S. cerevisiae telomeric sequences was also detected in the F2 tetrad: the bands corresponding in size to the segregating chromosome 5 of S. uvarum also hybridised with Y′, although it did not do so in the F1 clones. The extra band of the spore clone 6b also hybridised with the telomeric probe (Fig. 8). No chromosomal segregation or rearrangement was detected in the F4 tetrads isolated from F3 1b (not shown).
3.6 mtDNA analysis
To compare the mitochondrial genomes, we isolated mtDNA from the parental strains, the hybrid and its meiotic descendants. The parents differed in the restriction patterns obtained by digestion with EcoRV and RsaI (Fig. 9). The mtDNA of the hybrid was indistinguishable from that of the S. cerevisiae parent and all its F1–F4 descendents showed the same pattern (two of them are shown in Fig. 9).
Restriction patterns of mitochondrial DNA digested with EcoRV (left panel) or with RsaI (right panel). Lane 1: λ size marker (BstEII digested). Lanes 2 and 7: 10-170. Lanes 3 and 8: 10-522. Lanes 4 and 9: H01. Lanes 5 and 10: Spore clone a of F1 tetrad No. 1. Lanes 6 and 11: Spore clone a of F2 tetrad No. 6.
Hybrids of Saccharomyces sensu stricto species are usually unable to produce viable spores (e.g. [2–5,9]), although numerous authors have reported on interspecific hybrids whose spores germinated and produced vegetative clones (e.g. [8,9,14]). It is hypothesised that the ploidy level of the hybrid determines the viability of spores. Allodiploid hybrids either do not sporulate or their spores cannot germinate, most probably because of the inability of their chromosomes to pair in meiosis I (e.g. ). In the few instances in which high spore viability was detected and genetic analysis of spores was also performed, tetraploid marker segregation was observed [9,14]. It was proposed that viable spores can be produced in an allotetraploid genome because each chromosome has a matching partner for pairing. Such meiosis produces spores of mostly allodiploid chromosomal sets being heterozygous for the parental markers. The interspecific S. cerevisiae×S. uvarum hybrid analysed in this work produced viable spores, indicating that it might also be allotetraploid. It was prototrophic (both parents were auxotrophs), grew at 37 °C, (S. uvarum 10-522 was sensitive to this temperature), could ferment melibiose (S. cerevisiae 10-170 did not ferment melibiose), turned out to be heterozygous for the sexual types and the parental alleles of MET2, HIS4, YCL008c, ITS1-5.8SrDNA-ITS2, NTS2-ETS and MET10, and had a karyotype containing all chromosomal bands of both parents. These findings proved that its genome was alloploid indeed, but did not prove unambiguously its tetraploidy. In principle, it could have been triploid because the cells of the S. cerevisiae parent were heterothallic haploid, but the cells of the S. uvarum parent were homothallic, capable of diploidisation by self-mating. However, diploid cells heterozygous at the MAT locus rarely mate (for a review see ), which makes a haploid S. cerevisiae× diploid S. uvarum conjugation rather unlikely. The 2:2-segregation of certain markers in F1 tetrads also argues against triploidy. Besides, to reduce the probability of triploid mating, we let the S. uvarum partner sporulate (generate haploid spores) before mixing its culture with cells of the S. cerevisiae strain. In such situation mostly allodiploids can be expected with equal participation of both genomes. We propose that H01 might have arisen from an allodiploid zygote that underwent endomitosis (genome duplication) before or during germination. In a recent report Sebastiani et al.  have described fertile allotetraploid S. cerevisiae×S. bayanus hybrids that were also presumed to have been generated by endomitotic events in allodiploid cells. As demonstrated recently, almost all S. bayanus strains are in fact S. uvarum. These results prompted the reinstatement of the old species S. uvarum. In view of these findings, we assume that the hybrids investigated by Sebastiani et al.  were S. cerevisiae×S. uvarum.
Banno and Kaneko  have described S. bayanus×S. cerevisiae allotetraploid hybrids, in which the genomes of the two species did not recombine and the spores retained heterozygosis for the markers of the parental strains, suggesting that there was no physical interaction between the two genomes during meiosis and complete sets of both parental genomes were transmitted into the spores. Consistent with this conclusion, Sebastiani et al.  have found that the ascospores isolated from their S. bayanus×S. cerevisiae allotetraploid hybrids had the same chromosome set as the parental hybrids. In our case certain markers and certain chromosomes segregated and none of the spores had a karyotype identical with that of the hybrid, which indicated that meiosis did not transmit complete copies of the parental genomes into the spores. The spores had marker combinations and karyotypes that hinted at multiple events of recombination and exchange of chromosomal regions between the two genomes. This finding indicates that certain chromosomes of S. cerevisiae could pair with certain chromosomes of S. uvarum (at least in the group of small and medium-size chromosomes), and this “illegitimate” pairing did not foil the production of viable spores. It is pertinent to mention here that the small and medium-size chromosomes of S. cerevisiae wine strains are also prone to undergo intraspecific rearrangements (e.g. [22,25,40–42]). Interestingly, segregation and rearrangement in the hybrid meiosis described in this study also involved chromosomes of the conserved “diagnostic group” used for taxonomic identification of S. uvarum and S. bayanus strains [22,23].
The analysis of complete F1 tetrads produced by the hybrid and the successive three filial generations (isolated from single-spore clones of the previous generations) revealed a process of gradual reduction of the hybrid genome. Certain genetic markers and certain chromosomes of the small and medium-size categories were not transmitted into all spores or changed in size. As a rule, it was usually the S. uvarum component of the hybrid genome that became progressively smaller and smaller by loosing complete chromosomal bands and/or genetic markers. In spite of the preferential reduction of the S. uvarum genome, all spores retained the S. uvarum MET2, URA3 genes and its chromosomal region coding for ribosomal RNA.
In contrast to the presence of both parental sets of chromosomes in the nucleus of the hybrid, only one type of DNA was found in its mitochondria. The restriction pattern of this DNA was indistinguishable from that of the mtDNA of the S. cerevisiae parent, indicating that the mitochondrial genomes of the parents did not recombine in the zygote and only the S. cerevisiae mtDNA was transmitted into the cells produced by the zygote. Similar uniparental inheritance of mitochondrial genome have been revealed in the alloploid lager yeasts BRYC 32 and NCYC 1324  and between laboratory strains .
In conclusion, the results of the analysis of the hybrid H01 demonstrated that the S. cerevisiae genome and the S. uvarum genome did not contribute equally to the formation and stabilisation of the hybrid genome. The former provided the mitochondrial DNA and the more stable part of the nuclear genome. S. uvarum provided a complete set of its chromosomes for the hybrid, but certain of its chromosomes were then eliminated at meiotic divisions or underwent rearrangements in interactions (translocation or recombination) with S. cerevisiae chromosomes. Due to these interactions, some S. uvarum chromosomes acquired S. cerevisiae telomere sequences. The gradual elimination and alteration of large parts of the S. uvarum genome was associated with a progressive increase of sporulation efficiency and karyotype homogeneity in spores, suggesting a causal relationship between the reduction of the S. uvarum components and the stabilisation of the hybrid genome. It is tempting to surmise that this sort of processes may take part in speciation in Saccharomyces sensu stricto by creating the “continuum of genome structures” hypothesised by de Barros Lopes et al. . Nevertheless, when attempting an extrapolation of the conclusions of this study to yeast phylogenesis, one has to bear in mind that only one hybrid and a limited number of its descendants were analysed. An extensive pedigree analysis of further hybrids by examination of more spore clones over more filial generations is already underway in our laboratory.
We thank Iren Pasztor for excellent technical assistance and the EU Framework Programme 6 for a Marie Curie fellowship to Z. A. in Thiverval Grignon. This work was supported by the Grant NKFP 4/0007/2002 provided by the Hungarian Ministry of Education.
(2000) Molecular typing demonstrates homogeneity of Saccharomyces uvarum strains and reveals the existence of hybrids between S. uvarum and S. cerevisiae, including the S. bayanus type strain CBS 380. Syst. Appl. Microbiol. 23, 71–85.
(1997) Two subgroups within the Saccharomyces bayanus species evidenced by PCR amplification and restriction fragment lenght polymorphism of the nontranscribed spacer 2 in the ribosomal DNA unit. Syst. Appl. Microbiol. 20, 286–294.
(1996) Development of a polymerase chain reaction/restriction fragment length polymorphism method for Saccharomyces cerevisiae and Saccharomyces bayanus identification in oenology. FEMS Microbiol. Lett. 138, 239–244.
(2000) RFLP analysis of the ribosomal internal transcribed spacers and the 5.8S rRNA gene region of the genus Saccharomyces: a fast method of species identification and the differentiation of flor yeasts. Anton. Leeuw. 78, 87–97.
(2005) Evolutionary relationships between the former species Saccharomyces uvarum and the hybrids Saccharomyces bayanus and Saccharomyces pastorianus; reinstatement of Saccharomyces uvarum (Beijerinck) as a distinct species. FEMS Yeast Res. 5, 471–483.