Mechanism for Restoration of Fertility in Hybrid Zygosaccharomyces rouxii Generated by Interspecies Hybridization

ABSTRACT

The mechanism of whole-genome duplication (WGD) in yeast has been intensively studied because it has a large impact on yeast evolution. WGD has shaped the genomic architecture of modern Saccharomyces cerevisiae; however, the mechanism for restoring fertility after interspecies hybridization, which would be involved in the process of WGD, has not been thoroughly elucidated. In this study, we obtained a draft genome sequence of the salt-tolerant yeast Zygosaccharomyces rouxii NBRC110957 and revealed that it is a hybrid lineage of Z. rouxii (allodiploid) with two subgenomes equivalent to NBRC1876. Because this allodiploid yeast can mate with other allodiploid strains and form spores, it can be a good model of restoring fertility after interspecies hybridization. We observed that NBRC110957 and NBRC1876 contain six mating-type-like (MTL) loci. There are no large deletions or deleterious mutations in MTL loci, except for several-base-pair deletions in the X region in certain MTL loci. We also assigned only one mating-type (MAT) locus that exclusively determines mating types from six MTL loci. These results suggest that it is possible to recover mating competence regardless of whether cells lose one MAT locus through random gene loss by mitotically dividing after interspecies hybridization. Moreover, we propose that perturbation of gene expression and substantial breakdown of MAT heterozygosity caused by chromosomal rearrangement at MTL loci play roles in restoring the mating competence of allodiploids. This scenario can provide a mechanism for restoring fertility after interspecies hybridization that is compatible with random gene loss models and suggests genomic plasticity during WGD in yeast.

IMPORTANCE A whole-genome duplication occurred in an ancestor of the baker's yeast Saccharomyces cerevisiae. The origins of this complex and multifaceted process, which requires intra- or interspecies hybridization followed by dysfunction of one mating-type (MAT) locus to regain mating competence, has not been thoroughly elucidated. In this study, we provide a mechanism for regaining fertility in an interspecies hybrid, Zygosaccharomyces rouxii. The draft genome sequence analysis and mating test showed that the Z. rouxii strain used in this study is an intact interspecies hybrid, suggesting that it is possible to recover fertility regardless of whether cells lose one MAT locus.

INTRODUCTION

Polyploidy is a major evolutionary process in eukaryotes, particularly in plants and, to a lesser extent, in fungi and animals, in which several past and recent whole-genome duplication (WGD) events have been described (1). The double genetic content of polyploidy may have originated from a single species (autopolyploidy) or from different ones that are generally closely related (allopolyploidy) (1). One of the most characterized WGD events occurred in the lineage leading to the baker's yeast Saccharomyces cerevisiae (2–4). Although WGD in yeast was initially thought to occur by autopolyploidization at approximately 100 million years ago (2), a recent study showed that WGD in yeast is likely the result of allopolyploidization (5). Allopolyploidization would have conferred an immediate selective advantage to the newly created cell by fusing two organisms with slightly different characteristics.

Given that allopolyploidization occurs by mating and subsequent genome doubling, the mating type determination system is key to understanding WGD in yeast. The two versions of the mating-type (MAT) locus, MATa and MATα, specify three types of cell identity: haploid a, haploid α, and diploid a/α (6). In S. cerevisiae, the MAT locus information can be replaced with silent cassette, HML and HMR loci, that carry a- or α-specific information, and these silent cassettes are epigenetically silenced through chromatin modification mediated by the cis-acting silencer sequence E and I adjacent to these loci (6). This mating-type switching is facilitated by HO endonuclease, which creates a double-strand break at the MAT locus that is repaired using HML or HMR with the help of two sequence regions, X and Z, located at both ends of the idiomorph-specific region Y in the three mating-type-like (MTL) loci, the MAT, HML, and HMR loci (6). This fundamental mechanism of sex determination mediated by mating-type switching is thought to be shared with various yeast species, including not only S. cerevisiae but also the genus Zygosaccharomyces.

Previous studies have provided a scenario for WGD following interspecies hybridization (see Fig. S1 in the supplemental material) (5, 7). First, two haploid cells of different species, which are closely related, mate and form an allodiploid. The resulting allodiploids are usually sterile (unable to produce viable spores) or infertile (unable to sporulate). Some mechanisms have been proposed to account for the sterility and infertility: incompatibility between a nuclear-encoded mitochondrial regulatory protein and its mitochondrially encoded target gene (8) and dissimilarity of each chromosome derived from two different species that prevents proper pairing of homologous chromosomes during meiosis (9) or triggers the mismatch repair system (10). This postzygotic reproductive isolation usually prevents the evolution of allodiploid genomes, making interspecies hybridization an evolutionary dead end in most case (11). Conversely, allodiploids are still able to divide mitotically for many generations; therefore, the chromosome and/or one allele from every locus that is not haploinsufficient starts to be lost. Second, occasional loss of the MAT locus or a chromosome that carries a MAT locus leads to the formation of an allodiploid or alloaneuploid with mating competence (7, 11, 12). Third, mating-type switching provides the population with an organism of the opposite mating type that allows the possibility of sexual reproduction (7). Therefore, genome doubling could occur by mother-daughter mating between allodiploid MATa and MATα.

Soy sauce yeast, Zygosaccharomyces rouxii, was originally thought to be a heterothallic haploid yeast based on the observation that the heterothallic pair of mating strains NRRL Y2547 (NBRC1876, mating type MTa) and NRRL Y2548 (NBRC1877, mating type MTα) was isolated from Z. rouxii (13). However, recent studies have shown that these strains seemed to be the allodiploid yeast with two subgenomes as in ATCC 42981 (Table 1): the T subgenome derived from Z. rouxii CBS732 and the P subgenome derived from NCYC3042 (informally called Z. pseudorouxii) (14–19). These results indicate that NBRC1876 and NBRC1877 could also be allodiploid yeasts (Table 1) with mating competence and that the allotetraploid generated by mating between these two allodiploids would regain fertility. Thus, these allodiploid strains could be a good model to understand a fertility-regaining process during allopolyploidization. Here, we refer to the allodiploid strains as hybrid Z. rouxii or allodiploid in a context-dependent manner. Recently, Bizzarri et al. investigated the sex determination system in ATCC 42981 and concluded that ATCC 42981 is a sterile allodiploid without mating competence (20). They also argued that the sequence divergences in the chimeric a1-α2 heterodimer could be involved in the generation of negative epistasis, contributing to the allodiploid sterility and dysregulation of cell identity (20). However, our results obtained from this study suggest that JCM22060 (ATCC 42981) can mate with another allodiploid strain to generate a fertile allotetraploid (see Results). It is only recently that Ortiz-Merino et al. showed that Zygosaccharomyces parabailii, an interspecies hybrid that was formed by mating between 2 relatives, regained fertility when one of the two MAT loci was disrupted by chromosomal rearrangement (21). They also suggested that genome doubling after the disruption of one MAT locus is a real evolutionary process that occurs in natural interspecies hybrids, enabling them to resume mating and meiosis (21). Given that the mechanism of resuming mating and meiosis in the interspecies hybrid Z. parabailii is generalizable to other yeasts, one of the two MAT loci in hybrid Z. rouxii must also have been disrupted.

TABLE 1.

Strains used in this study

Strain name	Genotype	Other collectionsa	Genome size (Mb)/ploidy ratiob	Source (reference[s])
CBS732T	MATα		9.8/1.3	Grape must
Z1	1-D MTLαP (= MATα)		NDc	Soy sauce mash (50)
NBRC110957	1-D MTLaP (= MATa)		19.3/ND	Soy sauce mash (52)
NBRC1876	1-D MTLαP (= MATα)	CBS4837, NRRL Y-2547	19.4/1.96	Miso (13, 41)
NBRC1877	1-D MTLaP (= MATa)	CBS4838, NRRL Y-2548	ND/1.90	Miso
JCM22060	1-D MTLαP (= MATα)	ATCC 42981	ND/2.1	Miso
NBRC1876 ura3Δ	ura3Δ::loxP-KanMX-loxP/ura3Δ::loxP-ZeoMX-loxP 1-D MTLαP (= MATα)			This study
NBRC1877 ura3Δ	ura3Δ::loxP-KanMX-loxP/ura3Δ::loxP-ZeoMX-loxP 1-D MTLaP (= MATa)			This study
NBRC110957 ura3Δ 1-D MTLaP	ura3Δ::loxP-KanMX-loxP/ura3Δ::loxP-ZeoMX-loxP 1-D MTLaP (= MATa)			This study
NBRC1877 ade2Δ	ade2Δ::loxP-KanMX-loxP/ade2Δ::loxP-ZeoMX-loxP 1-D MTLaP (= MATa)			This study
NBRC110957 ura3Δ 1-D MTLαP	ura3Δ::loxP-KanMX-loxP/ura3Δ::loxP-ZeoMX-loxP 1-D MTLαP (= MATα)			This study
NBRC110957 ura3Δ 1-D MTLaT	ura3Δ::loxP-KanMX-loxP/ura3Δ::loxP-ZeoMX-loxP 1-D MTLaT (= MATa)			This study
NBRC110957 ura3Δ 1-D MTLαT	ura3Δ::loxP-KanMX-loxP/ura3Δ::loxP-ZeoMX-loxP 1-D MTLαT (= MATα)			This study

^aStrains in other culture collections which are same isolates and pooled at different culture collection centers.

^bPloidy data were obtained from references 18 and 51.

^cND, not determined.

In this study, we report the analysis of the draft genome sequence of hybrid Z. rouxii NBRC110957 obtained from soy sauce mash to further understand the mating type determination system in allodiploid Z. rouxii. Recently, Sato et al. also reported the analysis of the draft genome sequence of hybrid Z. rouxii NBRC1876 (22). These genome sequence analyses revealed that NBRC110957 and NBRC1876 have six MTL loci that would be derived from two ancestors possessing three MTL loci. Unexpectedly, although we could not find any evidence that these MTL loci have been disrupted, NBRC110957 can mate with NBRC1876 and form a fertile allotetraploid. Eventually, we found that only one of six MTL loci exclusively determines mating type in hybrid Z. rouxii. These results suggest that it is possible to recover mating competence regardless of whether one MAT locus is lost by random gene loss or chromosome loss during mitotic division after interspecies hybridization. The fertility-regaining model of hybrid Z. rouxii established in this study is a good complement to the model of Z. parabailii for understanding yeast evolution related to hybridization and WGD.

RESULTS

Draft genome sequence of NBRC110957.

We first conducted genome sequence analysis of the hybrid Z. rouxii NBRC110957, which is thought to be an allodiploid yeast (Table 1). The de novo assembly of the reads using the Platanus assembler resulted in 954 scaffolds. To reconstruct the genome sequence of NBRC110957, we adopted a similar approach, which was used to assemble the genomes of other hybrid yeast strains (23–26). Synteny and the sequence conservation of the gene between NBRC110957 and CBS732 make it possible to map 270 scaffolds (those of the 954 scaffolds that are larger than 500 bp) to the reference genome of CBS732 (27). The junction points between scaffolds were checked by PCR to confirm correct scaffold positioning, and the existing gaps were closed by sequencing the amplification product. The resulting 132 scaffolds were annotated manually and automatically using the Yeast Genome Annotation Pipeline (28) and the Microbial Genome Annotation Pipeline (29) to detect and annotate the gene. The genome assembly statistics and the general features are summarized in Tables 2 and 3, respectively. Apart from our work, analysis of the draft genome sequence of NBRC1876, which is also a hybrid Z. rouxii strain, has been reported recently (Table 1) (22). In the literature, the total genome size of NBRC1876 is estimated to be 19.4 Mb. Overall, the genome sizes of NBRC110957 and NBRC1876 and the coding sequence number of NBRC110957 are approximately twice those of CBS732, which is consistent with a previous karyotype analysis using flow cytometry (19). Moreover, we mapped reads to the reconstructed genome and reference genome of CBS732 to detect duplicated regions; for example, if reads are not mapped on a locus in the reference genome of CBS732 and 2-fold signals are detected on homologous loci of the reconstructed P subgenome (see below), then segmental duplication occurred at this locus. In this way, we reconstructed the NBRC110957 genome structure as shown in Fig. 1. Apart from a few exceptions, we found that NBRC110957 has two subgenomes that are comparable in size to the genome of CBS732, suggesting that NBRC110957 is apparently allodiploid. In addition, dot plot analyses between the genomes of CBS732 and allodiploids indicated that not only NBRC110957 but also NBRC1876 has two subgenomes that are compatible in size to the genome of CBS732 (see Fig. S2 and S3 in the supplemental material). Based on the nblast search results, the genomes of the allodiploids seemed to be divided into two types of subgenomes in almost all regions: the T subgenome, with approximately 99% identity to the CBS732 genome, and the P subgenome, with approximately 80 to 90% identity to the CBS732 genome, in accordance with a previous study (15) (see Table S1 in the supplemental material). To compare the genetic relatedness among the subgenomes, the average nucleotide identity (ANI) was calculated (30). The ANI among the genomes of CBS732 and the T subgenomes of NBRC110957 and NBRC1876 revealed approximately 100% identity, as observed between the P subgenomes of NBRC110957 and NBRC1876 (see Fig. S4 in the supplemental material). Previously reported genes of Z. pseudorouxii NCYC3042 (e.g., HIS3, ADE2, and SOD2) (14) show >99% identity with the corresponding sequence of the P subgenomes in NBRC110957 and NBRC1876 (data not shown). However, the sequence identity of the entire genome between the P subgenome and the Z. pseudorouxii genome is unclear because the genome sequence of Z. pseudorouxii has not yet been investigated. Note that we assume that there are strong, but not complete, collinearities among the genomes of CBS732 and the T and P subgenomes of hybrid Z. rouxii. There may be some additional chromosomal rearrangements, which are not visible due to short-read assembly in NBRC110957. For example, physical linkage between scaffold 5 and scaffold 8 in NBRC110957 has not been confirmed, suggesting that unexpected chromosomal rearrangement may have occurred.

TABLE 2.

Genome assembly statistics for strain NBRC110957

Parameter	Value
Sequence coverage (fold)	262
Assembled sequence length (bp)	19,269,454
Number of scaffolds	132
Scaffold length (bp)
Maximum	1,474,404
Minimum	506
N50a (bp)	1,286,020

^aN50, minimum contig length needed to cover 50% of the genome.

TABLE 3.

General features of NBRC110957 and CBS732

Strain	No. of chromosomes	Genome size (Mb)	Avg GC content (%)	Total no. of:
				Coding sequences	tRNA genes
NBRC110957	NDa	19.3	39.7	10,038	539
CBS732	7	9.8	37.1	4,992	272

^aND, not determined.

FIG 1.

Map of the scaffold of hybrid Z. rouxii NBRC110957 aligned on the chromosome structure of Z. rouxii CBS732. The black horizontal bars indicate the chromosome of Z. rouxii CBS732 with tick marks for every 100 kbp. The colored and hatched boxes indicate the T subgenome and P subgenome of hybrid Z. rouxii NBRC110957, respectively. The open boxes and lines connected to the box indicate a putative translocation. The numbers indicate the scaffold number. Arrows indicate the direction of the segment.

Figure 2 shows a schematic representation of the MTL organization in NBRC110957 and NBRC1876 and the inferred process of local chromosomal rearrangement during hybridization and subsequent genome stabilization. In species that had branched off from the ancestor of S. cerevisiae before the WGD (non-WGD species), MAT and HML are usually on the same chromosome and HMR is often on a different chromosome (31). Z. rouxii NBRC1130 maintains a conserved MAT organization, DIC1-MAT-SLA2, among non-WGD species (19). Given that this fundamental architecture is retained in two ancestors of the hybrid Z. rouxii, NBRC110957 and NBRC1876 probably experienced chromosomal rearrangement by reciprocal translocations among MTL loci, as shown in Fig. 2. Since it is unknown which locus corresponds to an active MAT locus due to chromosomal rearrangement by reciprocal translocation, we refer to MAT, HML, and HMR in both NBRC110957 and NBRC1876 as MTL in this study. Moreover, scaffolds 52 and 15 of NBRC110957 and NBRC1876, respectively, share the same structural features: the sequences corresponding to chromosome G and chromosome A of CBS732 are present in both scaffolds (Fig. S2 and S3). Considering the genetic relatedness (Fig. S4), NBRC110957 and NBRC1876 seemed to have diverged from a common ancestor through different types of reciprocal translocations. In addition, we found that six MTL loci are contained in NBRC110957 and NBRC1876 and that three of six MTL loci in NBRC110957 are identical to the MTL loci in CBS732, while two of six MTL loci in NBRC1876 are identical to the MTL loci in CBS732 (Fig. 3; see Fig. S5 to S8 in the supplemental material). Solieri et al. have proposed that Zygosaccharomyces sapae is a new species belonging to the genus Zygosaccharomyces (17). They found that Z. sapae carries two copies of several genes (18), like the hybrid Z. rouxii. Z. sapae thus may also be allodiploid like NBRC110957 and NBRC1876. Consistent with this assumption, our phylogenetic analyses of mating-type genes (the a1, a2, α1, and α2 genes) indicated that copy 1 of the α1 or α2 gene in Z. sapae ABT301 shares >99% identity to a sequence derived from the T subgenome; copy 2 of the α1 or α2 gene in Z. sapae ABT301 shares >99% identity to a sequence derived from the P subgenome (Fig. S5 and S6). For convenience, the MTL locus corresponding to CBS732 is referred to as MTLa^T or MTLα^T, and the other is referred to as MTLa^P or MTLα^P based on its information coded in the Y region in this study. Hybrid Z. rouxii NBRC110957 and NBRC1876 possess two divergent HO genes, the same as ATCC 42981 (20) and Z. sapae (32). These HO endonuclease sequences can be divided into a sequence derived from the T subgenome (100% identical to CBS732 HO) and a sequence derived from the P subgenome (92% identical to CBS732 HO) (see Fig. S9 and S10 in the supplemental material). Because these two HO genes are actively transcribed and upregulated under hypersaline stress in allodiploid ATCC 42981 (20), they are thought to be able to facilitate mating-type switching in hybrid Z. rouxii.

FIG 2.

Schematic representation of the MTL organization in NBRC110957 and NBRC1876, the inferred process of local chromosomal rearrangement during hybridization, and the subsequent genome stabilization. Filled triangles indicate the primers used for Fig. 3. The circles and diamonds indicate the DIC1 and SLA2 genes, respectively. The solid line and dotted line indicate the T and P subgenomes, respectively. The green and black boxes indicate MTL (MAT, HMR, or HML) loci, and X, Y, and Z indicate the X, Y, and Z regions at each MTL locus, respectively.

FIG 3.

Mating behavior and information encoded at each MTL. (A) Mating behavior of various hybrid Z. rouxii strains. Three tester strains with uracil auxotrophy and drug resistance were mated to various strains on mating medium. Allotetraploids were selected on SCD−ura medium supplemented with G418 and Zeocin and photographed. (B) PCR amplification of the MTL locus from Z1, NBRC110957, NBRC1876, NBRC1877, and JCM22060. All primer pair combinations were assayed (see Fig. 2). Based on the PCR amplification pattern of the MTL locus, the hybrid Z. rouxii strains used in this study were divided into two MTL locus patterns. (C) PCR-RFLP analysis of the MTL locus. The amplicon obtained from panel B was digested with HindIII and/or XhoI. Based on the PCR-RFLP pattern of the MTL locus, the information encoded at each MTL locus was determined (a^T, a^P, α^T, or α^P). (D) Calculated lengths of the restriction enzyme-digested MTL loci.

Presumed active MAT locus from six MTL loci.

To gain insight into the cell identity and MTL organization, we performed mating tests and PCR analyses using specific primer sets that anneal to the MTL locus flanking region (outside the X and Z regions, as indicated by the filled triangles in Fig. 2) (Fig. 3A and B). We found that the allodiploid strains used in this study apparently can mate with each other (Fig. 3A). Interestingly, we also detected only a few colonies at some intersections (e.g., at the intersection between NBRC110957 and NBRC1877) (see “Mating-type switching and gene expression at MTL loci in hybrid Z. rouxii” below). PCR analysis showed that the strains used in this study could be divided into two MTL locus patterns: pattern A, the PCR products of which were detected using primer sets 1-D, 2-B, 3-A, 4-C (except for Z1), 5-E, and 6-F, and pattern B, the PCR products of which were detected using primer sets 1-D, 2-D, 3-C, 4-B, 5-E, and 6-F (Fig. 3B). The difference between these patterns would be caused by the differences in MTL organization depending on strain-specific translocation events (Fig. 2). This result indicates the presence of an evolutionary relationship among hybrid Z. rouxii: NBRC110957 is related to Z1, whereas NBRC1876 is more related to NBRC1877 and JCM22060.

To elucidate the information encoded by each MTL locus, the PCR products were digested with HindIII and/or XhoI, and the resulting fragment length was compared with the calculated length for four types of idiomorphs encoded by six MTL loci in NBRC110957 (Fig. 3C and D). This PCR-restriction fragment length polymorphism (PCR-RFLP) analysis resulted in assignment of the MTLa^T, MTLα^T, MTLa^P, or MTLα^P genotype to each MTL locus. Although we could not detect the MTLa^T idiomorph in Z1, all the other strains retained six MTL loci and four types of idiomorphs.

To investigate how cell identity is determined in hybrid Z. rouxii, we compared the information obtained from the MTL loci (Fig. 3C) to the results of their mating tests (Fig. 3A). Given that cell identity is determined by a single MTL locus and is common among the allodiploids used in this study, the MTL locus amplified by the 1-D primer sets (for convenience, this MTL locus is referred to as the 1-D MTL locus) is thought to be the only MTL locus accountable for the mating test results shown in Fig. 3A (e.g., NBRC1876 and NBRC1877 can mate, and only the genotype of the 1-D MTL locus is different among their six MTL loci). In addition, the 1-D MTL locus also maintains the conserved MAT organization DIC1-MAT-SLA2 among non-WGD species (Fig. 2). Taking these results into account, the 1-D MTL locus might be an active MAT locus in the allodiploid.

Mating-type switching and gene expression at MTL loci in hybrid Z. rouxii.

We confirmed whether the 1-D MTL locus is an active MAT locus using two approaches. First, we screened strains with a 1-D MTLα locus from the NBRC110957 ura3Δ background with a 1-D MTLa^P locus to obtain allodiploid strains with the same genetic background, excluding the genotype of the 1-D MTL locus (Fig. 4A). From approximately 5,000 colonies, we succeeded in obtaining three candidates using auxotrophic complementation as a selectable marker. PCR-RFLP analysis showed that only the 1-D MTL locus is switched from a^P to α^P in these screened strains (for convenience, these strains are referred to as “switched strains”) (Fig. 4B). In the case of these switched strains, the 5-E MTL locus must be used as a template for switching from a^P to α^P because it only carries α^P information among six MTL loci. In addition, the switched strains showed mating behavior opposite to that of the parent strain (e.g., NBRC110957 mated with NBRC1876 but not NBRC1877 and NBRC110957, while the switched strains mated with NBRC1877 and NBRC110957 but not NBRC1876) (Fig. 4C). These results show that the mating-type switching mechanism still functions in hybrid Z. rouxii and suggest that the 1-D MTL locus would be an active MAT locus in the allodiploid. This presumption also suggests that the stray colonies detected in Fig. 3A are probably generated by mating-type switching followed by mating. To test this hypothesis, we performed PCR analysis of the 1-D MTL locus in the stray colonies obtained from the intersection between NBRC110957 and itself. The PCR amplicons at the 1-D MTL locus of these colonies showed a heterozygous fragment length corresponding to a and α, suggesting that these stray colonies could be allotetraploids generated by mating-type switching followed by mating (data not shown).

FIG 4.

Mating-type switching and gene expression at the MTL loci in the allodiploid. (A) Schematic diagram of the mating-type switching experiment. Allotetraploid strains with prototrophy were generated by crossing with NBRC1877 ade2Δ 1-D MTLa^P and NBRC110957 ura3Δ 1-D MTLa^P. Corresponding colonies on a master plate in NBRC110957 ura3Δ were obtained as candidate strains in which the idiomorph encoded by the 1-D MTL locus was converted to the opposite one (switched strain). (B) PCR-RFLP analysis of the MTL locus in switched strains obtained from NBRC110957 ura3Δ 1-D MTLa^P. The amplicon obtained from PCR analysis using various primer sets was digested with HindIII and/or XhoI. (C) Mating behavior of the switched strains obtained from NBRC110957 ura3Δ 1-D MTLa^P. The switched strains with uracil auxotrophy and drug resistance were mated to NBRC1876, NBRC1877, and NBRC110957 on mating medium. Allotetraploids were selected on SCD−ura medium supplemented with G418 and Zeocin and photographed. (D) Gene expression at the MTL loci of various strains. cDNA was amplified from total RNA extracted from yeast cells cultured in mating medium. + or −, cDNA synthesis reaction mixtures with or without reverse transcriptase, respectively.

Second, in addition to the mating test, we performed RT-PCR analysis in various hybrid Z. rouxii strains to assess which of the four possible mating-type genes (a^T, α^T, a^P and α^P) were expressed at each MTL locus. Apart from the background expression of the a^T2, α^T1, and α^T2 genes observed in NBRC110957, all the strains expressed the a^P and/or α^P gene (Fig. 4D). To interpret this result, we focused on the gene expression pattern of NBRC110957 (Fig. 4D, lanes 3 and 4) and its switched strain (Fig. 4D, lanes 5 and 6). NBRC110957 expressed only the a^P gene, which could be expressed by the 1-D and/or 3-A MTL locus. Conversely, the switched strains derived from NBRC110957 expressed both α^P and a^P genes; the α^P gene could be expressed by the 1-D and/or 5-E MTL locus, and the a^P gene could be expressed only by the 3-A MTL locus (Fig. 4B and D). The 5-E MTL locus must be a silent cassette, because α^P gene expression was not observed in NBRC110957 (Fig. 4D, lanes 3 and 4). Taking into account the results for the gene expression patterns of other allodiploid strains with MTL locus pattern A and the information for the MTL loci (Fig. 4D and 3C), it seemed that the 1-D and 3-A MTL loci would be active loci in strains with MTL locus pattern A. Moreover, we also assumed that the 1-D and 2-D MTL loci would be active in strains with MTL locus pattern B in the same manner as in strains with MTL locus pattern A. Interestingly, both MTL locus patterns A and B share the 1-D MTL locus as a candidate for the active locus. Taken together, these results strongly suggest that the 1-D MTL locus would be an active MAT locus in the allodiploid.

Analysis of the allotetraploid and its gametes.

To exclude the possibility that the 2-D MTL locus is involved in determining cell identity in hybrid Z. rouxii, we further investigated the genotype and gene expression patterns of gametes derived from allotetraploids generated by mating between the hybrid Z. rouxii JCM22060 and NBRC1877 ura3Δ (Fig. 5). The 1-D and 2-D MTL loci of JCM22060 encode α^P, and those of NBRC1877 ura3Δ encode a^P, while the other four MTL loci encode the same idiomorph between these strains (Fig. 3C). This fact allows the assessment of four types of gametes with the same genotype, excluding the 1-D and 2-D MTL loci: a^P and a^P, a^P and α^P, α^P and a^P, or α^P and α^P. We mated JCM22060 with NBRC1877 ura3Δ, and the resulting allotetraploid was incubated on a sporulation plate to form a spore. After confirmation of spore formation by microscopic analysis, we selected gametes on a 5-fluoroorotic acid (5-FOA) plate using random spore analysis (Fig. 5A and B). PCR-RFLP analysis revealed that four types of gametes with the same genotype excluding the 1-D and 2-D MTL loci were successfully isolated, as expected (Fig. 5C). This result suggests that meiotic division occurred in the allotetraploid generated by mating JCM22060 and NBRC1877 ura3Δ. In agreement with Fig. 4D, expression of the α^P and a^P genes was detected from the 1-D and 2-D MTL loci (Fig. 5D). Moreover, the cell identity was determined, depending on the information for the 1-D MTL locus excluding the no. 2 gamete (Fig. 5E). The no. 2 gamete carrying the a^P and α^P genes at the 1-D and 2-D MTL loci, respectively, could not mate with either NBRC1876 with the 1-D MTLα^P locus or NBRC1877 with the 1-D MTLa^P locus, suggesting that the no. 2 gamete is a nonmater (Fig. 5E). The nonmater phenotype was also observed when other tester strains or other gametes were used (data not shown).

FIG 5.

Genotypes and phenotypes of the gametes obtained from allotetraploids. (A) Schematic diagram of the mating and sporulation experiment. JCM22060 1-D MTLα^P was mated to NBRC1877 1-D MTLa^P with uracil auxotrophy and drug resistance, and the resulting allotetraploids were plated on sporulation medium. After confirmation of spore formation by microscopic analysis, the sporulated culture was incubated with Zymolyase and plated on 5-FOA medium. The resulting colonies were analyzed. (B) Drug resistance and auxotrophy of the gamete. The parent strain, the allotetraploid, and its gametes were inoculated on several selective media and incubated. (C) PCR-RFLP analysis of the MTL locus of the gamete. The amplicon obtained from PCR analysis using various primer sets was digested with HindIII and/or XhoI. (D) Gene expression at MTL loci of the gamete. cDNA was amplified from total RNA extracted from yeast cells cultured in mating medium. + or −, cDNA synthesis reaction mixtures with or without reverse transcriptase, respectively. (E) Mating behavior of the gamete. The gamete was mated to NBRC1876 and NBRC1877. Allotetraploids were selected on SCD−ura medium supplemented with G418 and Zeocin and photographed.

Taking the data together, despite a few exceptions, it seems more likely that the 1-D and not the 2-D MTL locus is the only active MAT locus, because cell identity is determined exclusively by the information derived from the 1-D MTL locus and mating-type genes expressed for the allodiploid cell are consistent with the 1-D MTL locus information.

MTL idiomorphs derived from the T subgenome are also functional.

Unfortunately, the 1-D MTL loci of all strains used in this study carry the a^P or α^P gene but not the a^T or α^T gene. Thus, we are not able to discriminate whether the MTL idiomorphs derived from the T subgenome are functional. Although there are no large deletions and truncations in the a^T or α^T sequence in comparison to the a^p or α^p sequence (Fig. S5 to S8), we cannot exclude the possibility that a few amino acid substitutions triggered dysfunction of the a^T or α^T idiomorphs. To address this issue, we converted 1-D MTLα^P of NBRC110957 ura3Δ to 1-D MTLa^T or α^T, as shown in Fig. 6A. We succeeded in obtaining 1-D MTL locus-converted strains as expected and performed a mating test, followed by confirmation of the mating-type gene expression (Fig. 6B, C, and D). As a result, the converted strain with the 1-D MTLa^T locus expressed the a^T gene and mated with NBRC1876, while the other converted strain with the 1-D MTLα^T locus expressed the α^T gene and mated with NBRC1877 (Fig. 6C and D). These results suggest that the a^T and α^T idiomorphs are functionally redundant to the a^P and α^P idiomorphs.

FIG 6.

Functional confirmation of the T-subgenome-type MTL. (A) Schematic diagram of the MTL conversion experiment. NBRC110957 1-D MTLα^P ura3Δ was transformed with the converting construct, and the resulting transformants were cultured on 5-FOA medium. The obtained colonies were analyzed. (B) PCR-RFLP analysis of the MTL loci of converted strains with T-subgenome type 1-D MTL. The amplicon obtained by PCR analysis using various primer sets was digested with HindIII and/or XhoI. (C) Gene expression at MTL loci of the converted strains with T-subgenome type 1-D MTL. cDNA was amplified from total RNA extracted from yeast cells cultured in mating medium. + or −, cDNA synthesis reaction mixtures with or without reverse transcriptase, respectively. (D) Mating behavior of converted strains with T-subgenome type 1-D MTL. The converted strains were mated to NBRC1876 and NBRC1877. Allotetraploids were selected on SCD−ura medium supplemented with G418 and Zeocin and photographed.

DISCUSSION

Infertility is a common characteristic of hybrid yeast strains due to the difference between the parental set of chromosomes, which prevents precise chromosome pairing during meiosis (9, 33, 34). One way to avoid this situation is to lose one MAT allele and experience mating-type switching and mother-daughter mating, allowing each chromosome to pair with an identical partner. Indeed, experimental inactivation of the MAT locus or elimination of the chromosome that carries a MAT locus in diploids was conducted to construct allotetraploids that were meiotically fertile (11, 35). A recent study also shows that the natural interspecies hybrid Z. parabailii regained fertility by genome doubling after interspecies mating as a consequence of damage to one copy of its MAT locus (21). This inactivation of the MAT locus or chromosome elimination abolishes the heterozygosity of MATa/α and results in the formation of allodiploids or alloaneuploids with either MATa or MATα. In plants, it is also known that allotetraploid formation through the fusion between diplogametes (2n pollen and egg) derived from different species serves as an important mechanism for the restoration of hybrid fertility (36), similar to the situation obtained for meiotically fertile allotetraploid yeasts. In this study, we showed that NBRC110957 and NBRC1876 are allodiploids that have both mating competence and six MTL loci derived from two ancestor strains. This result apparently shows that allodiploid yeast can regain mating competence regardless of whether one MAT allele is lost by random gene loss or by loss of the MAT chromosome after interspecies hybridization. We propose that chromosome translocation between MTL loci gives rise to perturbation of gene expression at the MAT locus, leading to a substantial MAT loss-like effect. In haploid Z. rouxii, it is known that the translocation of chromosomes often occurs via ectopic exchanges between MTL loci (19). Similarly, in hybrid Z. rouxii, chromosomal translocations were observed between MTL loci (Fig. 2). This chromosomal rearrangement frequently occurred at MTL loci because six MTL loci, excluding the Y region, share almost 100% sequence identity with each other, while approximately 80 to 90% sequence identity is shared between the two parental genomes. This translocation event will lead to unlocking of the silencing of cryptic mating-type loci, HML and HMR, via cis-acting silencers known as E and I, and/or to silencing of the action of the MAT locus. A model to recover mating competence via chromosomal translocation between MTL loci based on the genome organization of NBRC110957 is shown in Fig. 7. Two closely related haploid cells mate to form an allodiploid. The allodiploid cannot undergo meiosis due to the dissimilarity of the two copies of each chromosome, but they are still able to divide mitotically. If the allodiploid cell replicated mitotically for many generations, there would be nothing to prevent chromosomal translocation between the MTL loci. Perturbation of the gene expression at the MAT loci, which is caused by chromosomal translocation, may allow allodiploid cells to behave like a haploid and to switch their mating type, undergo mother-daughter mating, and form the allotetraploid. The allotetraploid can undergo meiotic division and return to the allodiploid. During many generations of sexual and asexual reproduction, redundant genes would be lost and the allodiploid/allotetraploid sexual reproduction status may gradually return to an (allo)haploid/(allo)diploid sexual reproduction status. The difference in the mechanism of regaining fertility between hybrid Z. parabailii and hybrid Z. rouxii concerns the mechanism by which one of the two MAT loci is inactivated. Z. parabailii inactivates the MAT locus through physical breakage of one of the two MAT loci, whereas the hybrid Z. rouxii substantially inactivates the MAT locus through perturbation of the gene expression at MAT locus. It is noteworthy that two different interspecies hybrids adopt two different ways to regain fertility, suggesting that the retrieval of fertility in interspecies hybrids is not uniform.

FIG 7.

Model of the restoration of fertility after interspecies hybridization. A detailed explanation is provided in Discussion. The squares represent yeast cells. Genes are denoted by numbers, and E and I indicate the putative cis-acting silencer sequence. The red boxes are the MAT loci, and the pink boxes are the MTL loci, which do not determine the cell identity despite their partially active gene expression.

A remaining question is why hybrid Z. rouxii cells can exhibit an a or α cell identity even though both idiomorphs are expressed in a or α cells. For example, NBRC110957 with the 1-D MTLα^P locus expresses both a^P genes at the 3-A MTLa locus and α^P genes at the 1-D MTLα locus, but it can mate with NBRC1876 and JCM22060 (Fig. 3 to 6). Given that the gene-silencing mechanism in MTL loci is almost the same in Z. rouxii and S. cerevisiae, one possible explanation is that gene expression at the 3-A MTLa locus might differ from cell to cell because the silencing mechanism is epigenetic and the E silencer sequence is maintained around the 3-A MTL locus (Fig. 7). We infer that a cell expressing both idiomorphs behaves as a nonmater in the clonal cell population, while other cells undergoing epigenetic silencing at 3-A MTL or 2-D MTL behave as a or α cells. Recent single-cell studies show that some silencing mutants with an intermediate phenotype, such as sir1 mutants, consist of two distinct populations: one repressed and another derepressed HMR and/or HML. This result suggests that the HMR and HML loci behave independently within a single cell, demonstrating that heterochromatin formation is locus autonomous (37). Unfortunately, however, the analysis to measure the degree of gene silencing at HMR or HML in single cells with only the E silencer has not yet been performed in hybrid Z. rouxii. It is also unclear why the 1-D MTL locus could not determine the cell identity of the no. 2 gamete as shown in Fig. 5E. This finding suggests that the mating competence acquired by the perturbation of gene expression at the MAT locus is unstable when cells have recently undergone alloplolyploidization.

Although the model proposed in this study using the natural hybrid Z. rouxii (Fig. 7) remains to be established experimentally in synthetic interspecies hybrids between Z. rouxii CBS732 and Z. pseudorouxii NCYC3042, it seemed to be quite compatible with the experimental results obtained in Saccharomyces hybrids. It is known that an infertile allodiploid can give rise to a fertile allotetraploid through an endomitotic event (38) or mating between allodiploids with MAT locus homozygosity (MATa/a or MATα/α). It is known that an allodipolid (MATa/a or MATα/α) can be generated by loss of heterozygosity at the MAT locus in yeast (39). Allotetraploids (MATa/a/α/α) generated by mating between MATa/a and MATα/α allodiploids can produce viable spores with allodiploid chromosome sets. Half of the cells derived from spores with an allodiploid chromosome set are unable to mate with each other due to heterozygosity at the MAT locus (MATa/α), while they enter meiosis directly in response to appropriate signals (e.g., starvation); however, they cannot perform normal meiotic divisions, and even if they can produce spores, the spores will not be viable due to incompatibility between the two parental genomes (11). Taking the data together, allotetraploids generated by endomitotic events or mating between MATa/a and MATα/α allodiploid produce viable spores, but the resulting spore clones cannot produce viable spores. This kind of sterility is referred to as F1 sterility to distinguish it from the sterility of allodiploid hybrids (33). Therefore, elimination of the MAT locus in the allodiploid is the central element for regaining fertility without F1 sterility (11, 12, 35, 40). These results convincingly show that the mechanism of reproductive isolation of the species and overcoming the species barrier seems to be similar in the genera Saccharomyces and Zygosaccharomyces. It is quite reasonable to suppose that similar mechanisms can operate simultaneously in both genera. Taking into account the experimental results for Saccharomyces and Zygosaccharomyces, fundamental principles of reproductive isolation established in the genus Saccharomyces may be generalized across genera.

It is also important to discuss the contradiction between the results in this study and those of Bizzarri et al. (20). The significant differences involve the mating behavior and the MTL information for ATCC 42981 and JCM22060, which should be same isolate. First, Bizzarri et al. did not find any evidence that ATCC 42981 cells can mate with NBRC1876 or NBRC1877 tester strains based on a microscopic observation of a mixed culture on malt extract agar plates. They interpreted this result to indicate that the ATCC 42981 did not respond to Z. rouxii pheromone signaling or that there was an imbalance resulting in deficient organization of the MTL loci (20). Conversely, our data here confirmed that JCM22060 apparently mates with NBRC1877 (Fig. 3A), casting doubt on the negative results observed by Bizzarri et al. Mori and Onishi found that hybrid Z. rouxii can neither conjugate nor sporulate on commonly used sporulation media (41). They reported that a suitable medium for conjugation and sporulation is 5% NaCl-containing Shoyu-koji extract agar, as used in the present study. Therefore, the discrepancy between our data and those of Bizzarri et al. could be explained by the lower microscopic sensitivity leading to the oversight of conjugation asci and spores, differences in the medium used to induce a mating response, or differences in the stocks used. Second, Bizzarri et al. also identified 7 MTL loci in ATCC 42981 using a PCR-based cloning strategy (20). They cloned the α2 and α1 gene copies 1 (which are substantially identical to α^T2 and α^T1, respectively), α2 and α1 gene copies 2 (which are substantially identical to α^P2 and α^P1, respectively), a2 gene copy 1, a2 gene copy 2, a2 gene copy 3, and a1 gene (which is substantially identical to a^T1) (see Fig. S5 to S8 in the supplemental material). The sequence differences among the a2 gene copies 1, 2, and 3 are mainly at the C-terminal end, which is derived from beyond the X region and variable region, depending on the cassette location. The sequence derived from inside the X region is identical among a2 gene copies 1, 2, and 3; therefore, the sequences are substantially identical to that of a^T2 (see Fig. S8 in the supplemental material). In summary, Bizzarri et al. cloned 1 MTLα^T, 2 MTLα^P, and 4 MTLa^T from ATCC 42981, whereas we cloned 1 MTLα^T, 3 MTLα^P, 1 MTLa^T, and 1 MTLa^P from JCM22060 (Fig. 3B and C). This discrepancy must be caused by differences in the stocks used in each study. To resolve these profound discrepancies, a mating test in ATCC 42981 using an auxotrophic marker is needed. It would be better to recognize that the same hybrid Z. rouxii strains obtained from different stocks could be substantially different, because similar problems have been found in other strains (e.g., NBRC1130 and CBS732 were originally the type strain of the species in each stock, but their genotypes are now different) (19).

MATERIALS AND METHODS

Strains, medium, and genetic methods.

The strains used in this study are listed in Table 1. Strains were maintained and cultivated in YPD (1% yeast extract, 2% peptone, and 2% glucose) or YPDU (YPD supplemented with 0.17% uracil) medium. Mating medium (10% Shoyu-koji extract, 5% NaCl, and 5% glucose) and sporulation medium (10% Shoyu-koji extract and 5% NaCl) were used to induce mating and sporulation (41). The gene knockout in yeast was performed by PCR-mediated gene replacement, and it was checked by PCR analysis. pUG6 and pUZ6, possessing G418 and Zeocin resistance genes surrounded by two loxP sequences, respectively, were used as PCR templates for the gene deletions. The transformations of hybrid Z. rouxii were performed using electroporation (42). The bacterial transformations and bacterial DNA manipulation were performed using standard methods (43). YPDU medium and synthetic complete dextrose (SCD) medium lacking uracil (SCD−ura) containing 200 mg/liter G418 and/or 100 mg/liter Zeocin were used to select the hybrid Z. rouxii transformants. 5-FOA medium (SCD supplemented with 0.17% 5-fluoroorotic acid) was used for counterscreening of URA3 gene-lacking transformants.

Genome sequence.

The genome sequence of strain NBRC110957 was performed by TaKaRa Bio (Shiga, Japan) by means of a whole-genome shotgun approach that explored paired-end sequences using HiSeq 2000 (Illumina, San Diego, CA, USA) (inserts with 348 100-base reads). Approximately 6 Gb of reads was obtained; the genome coverage is approximately 300-fold if the genome size of NBRC110957 is approximately 20 Mb.

Genome assembly and annotation.

Preprocessing and assembly were performed using the DDBJ Read Annotation Pipeline (https://p.ddbj.nig.ac.jp/pipeline/Login.do) (44, 45). The preprocessing parameters are indicated below. The Qv threshold of base trimming was 24, and the Qv threshold and percent threshold of read removal were 14 and 20%, respectively. Short reads were assembled using Platanus 1.2.1, which was developed to reconstruct genomic sequences of highly heterozygous diploids (46). The resulting scaffolds were sequentially ordered based on their level of synteny with the genomes of Z. rouxii CBS732. The junction points between scaffolds were checked by PCR to confirm correct scaffold positioning, and gaps were closed by Sanger sequencing of the amplification product. Gene identification and annotation were carried out manually and through the Yeast Genome Annotation Pipeline (http://wolfe.ucd.ie/annotation/) (28), the Microbial Genome Annotation Pipeline (http://www.migap.org/) (29), and the Yeast Gene Order Bowser (http://ygob.ucd.ie/) (47). Dot plot analysis was performed using YASS (48). MTL loci of NBRC110957 and NBRC1876 were identified based on sequence similarity with that of the reference strain CBS732. The ANI was estimated using the ANI calculator with default parameters (30).

Mating and sporulation.

Mating and sporulation were conducted as described previously (41) with a few modifications. Briefly, mating was performed as by (i) streaking one strain on mating medium, (ii) cross-streaking another strain onto the same medium, (iii) incubating at 25°C for 5 days, (iv) replicating the streaks onto selective medium (SCD−ura supplemented with 600 mg/liter G418 and 600 mg/liter Zeocin), and (v) incubating the samples at 30°C for 2 to 3 days. Positive colonies growing at the intersection of the streaks were isolated on selective medium. The isolated colonies were then cultured on sporulation medium at 25°C for 5 days to form spores. Spore formation was confirmed by microscopy. Spore isolation was performed by random spore analysis. The culture on sporulation medium was incubated with 0.5 mg/ml Zymolyase-20T (Nacalai Tesque, Kyoto, Japan)-containing 100 mM Tris-HCl buffer (pH 7.5) at 30°C for 1 h, followed by plating on 5-FOA medium. After incubation, the colonies with the preferable genotype were selected by colony PCR.

Mating-type switching.

Fifty microliters of full-growth culture of NBRC1877 ade2Δ 1-D MTLa^P was plated evenly on mating medium, and a moderately diluted culture of NBRC110957 ura3Δ 1-D MTLa^P was plated on another mating medium to generate approximately 200 to 300 colonies (master plate). The colonies of the latter were replicated on the plate of the former and incubated at 25°C for 5 days. The incubated plate was replicated onto SCD medium without uracil and adenine, followed by incubation at 30°C for 2 to 3 days. After confirming the emergence of positive hybrid colonies, the corresponding candidate colonies that may have undergone mating-type switching were picked from the master plate, and the genotype was checked by PCR analysis. To obtain NBRC110957 ura3Δ 1-D MTLa^T or MTLα^T, the MTLa^T or MTLα^T sequence was amplified by PCR from the NBRC110957 genomic DNA using the 4-C and 2-B primer sets, respectively. The primers used in this study are listed in Table 4. The PCR products were purified and used as a template for nested PCR. The nested PCR was performed using the a^T_cloning_Fw and a^T_cloning_Rv and the α^T_cloning_Fw and α^T_cloning_Rv primer sets to append the restriction site at both ends of the PCR products. The PCR products were digested with EcoRI and SalI or XbaI and cloned into the same restriction site of pEU, which was constructed from pZEU (49) by ClaI digestion and self-ligation to exclude the sequence from pSRI. The resulting plasmid was linearized by inverse PCR using the Inv_Fw and Inv_Rv primer sets. NBRC110957 ura3Δ 1-D MTLα^P was transformed using the linearized plasmid and plated on SCD−ura medium. After the resulting transformants were incubated with agitation in YPDU, they were transferred to 5-FOA medium, followed by incubation at 30°C for 1 week. The genotype of the resulting colonies was checked by PCR and PCR-RFLP analysis.

TABLE 4.

Primers used in this study

Primer name	Sequence (5′ to 3′)	Corresponding figure(s)
1	CAACATGGTCATGGTCAACA	3B and C, 4B, 5C, 6B
2	TGAATCTTCTCTGAAATATCCAC	3B and C, 4B, 5C, 6B
3	AGCTGTAAACGGTAAAGCTG	3B and C, 4B, 5C, 6B
4	TTGGTTAAACGAGGTGTGTG	3B and C, 4B, 5C, 6B
5	TACAAGGTGTTGTGAGAGCT	3B and C, 4B, 5C, 6B
6	TGTATTGACCAGCTTCGTTTGA	3B and C, 4B, 5C, 6B
A	CCAACCAGTTAGTGTGTTATC	3B and C, 4B, 5C, 6B
B	ATCAGTACCAGAAGTGGTCT	3B and C, 4B, 5C, 6B
C	AGAGAGTCAAACGATGCCTA	3B and C, 4B, 5C, 6B
D	TCATCTGACGTAAGACTTGC	3B and C, 4B, 5C, 6B
E	GGGCAACAATTTTGGTATAATAAC	3B and C, 4B, 5C, 6B
F	ATGGACTACACGTACCACAA	3B and C, 4B, 5C, 6B
aT_cloning_Fw	AAAAGAATTCGGCGATGGTTTTTTCTTGGG	6A
aT_cloning_Rv	AAAAGTCGACAACACAATCCATCGCCACTG	6A
αT_cloning_Fw	AAAATCTAGAGGCGATGGTTTTTTCTTGGG	6A
αT_cloning_Rv	AAAATCTAGAAACACAATCCATCGCCACTG	6A
Inv_Fw	TTCCAAATTGGCGATGAGCG	6A
Inv_Rv	GATATCCTAATACTCCCCC	6A
a1T_Fw	GTCGGACGAAACCTTTAATTCTCT	4D, 5D, 6C
a1T_Rv	CCATATTCTAACCTGAATCGGTGTC	4D, 5D, 6C
a1P_Fw	GTCTGACAAGACTTTCAATTCTCCA	4D, 5D, 6C
a1P_Rv	ATTCTTACTTGAAGTGGTGTCAAGC	4D, 5D, 6C
α1T_Fw	AAAGTTGAAAAGAAGGGCAGGT	4D, 5D, 6C
α1T_Rv	TGTTGCTCGGGATGAGAATG	4D, 5D, 6C
α1P_Fw	AGAATCGACCCAGACACCAA	4D, 5D, 6C
α1P_Rv	GATGTGAATGCCACTCTGCA	4D, 5D, 6C
a2T_Fw	ATTTAGCCAAGTGGGCGATT	4D, 5D, 6C
a2T_Rv	AAGGGAAGTTCAAAGGAAACCA	4D, 5D, 6C
a2P_Fw	ATTGGCGATGAGCGAAGAAG	4D, 5D, 6C
a2P_Rv	CGTAAATACGGAGAGTCACAAGGA	4D, 5D, 6C
α2T_Fw	CGTTCAATTCTTCAGTACCCTGTT	4D, 5D, 6C
α2T_Rv	TCCTCCGTCTTTTCCTTGCT	4D, 5D, 6C
α2P_Fw	GGAACTTTTGTTGAGGTAAGGATG	4D, 5D, 6C
α2P_Rv	CGAAGCTACCAGCAGAGAATG	4D, 5D, 6C

MTL locus PCR and PCR-RFLP analysis.

PCR was performed using KOD Fx neo (Toyobo, Osaka, Japan) polymerase for 35 cycles with an annealing temperature of 59°C. Colony PCR was performed by heating the cells in 10 μl of KOD Fx neo buffer at 95°C for 10 min, followed by addition of the PCR reagent mixture. DNA fragments amplified by PCR were digested with HindIII and/or XhoI, run on an agarose gel, and typed based on the resulting fragment length.

RNA extraction and RT-PCR analysis.

RNA extraction was performed as previously described (50). Following DNase I (TaKaRa Bio) treatment, cDNA was synthesized from total RNA using the PrimeScript RT-PCR kit (TaKaRa Bio) according to the manufacturer's instructions. For RT-PCR, amplification was performed using KOD Fx neo polymerase for 32 to 35 cycles with an annealing temperature of 59°C.

Accession number(s).

The nucleotide sequences of the hybrid Z. rouxii NBRC110957 genome have been deposited in DDBJ under accession numbers BDGX01000001 to BDGX01000132. The accession number of the raw reads is DRR065857. This paper describes the first version of the genome. The accession numbers of the MTL loci of hybrid Z. rouxii NBRC1876 are LC221835 to LC221840.