Crossbreeding of Yeasts Domesticated for Fermentation: Infertility Challenges

Sexual reproduction is almost a universal feature of eukaryotic organisms, which allows the reproduction of new organisms by combining the genetic information from two individuals of different sexes. Based on the mechanism of sexual reproduction, crossbreeding provides an attractive opportunity to improve the traits of animals, plants, and fungi. The budding yeast Saccharomyces cerevisiae has been widely utilized in fermentative production since ancient times. Currently it is still used for many essential biotechnological processes including the production of beer, wine, and biofuels. It is surprising that many yeast strains used in the industry exhibit low rates of sporulation resulting in limited crossbreeding efficiency. Here, I provide an overview of the recent findings about infertility challenges of yeasts domesticated for fermentation along with the progress in crossbreeding technologies. The aim of this review is to create an opportunity for future crossbreeding of yeasts used for fermentation.


Introduction
Sexual reproduction is ubiquitous in eukaryotes and expands the genetic diversity by producing progeny that resemble their parents but are NOT identical to them [1]. The benefits of sexual reproduction include the purging of deleterious mutations from the genome. Due to genetic recombination and diversification, sexual reproduction can provide a recombinant progeny, well-adapted to a changing environment [2], although there is a twofold cost of sex in relation to asexual reproduction [3][4][5][6]. It is rational that numerous species in nature maintain sexual reproduction for survival.
It is known that many species can reproduce both asexually and sexually. Budding yeasts also exhibit asexual and sexual reproductive cycles, whereas the most common mode of vegetative growth is asexual reproduction in both haploid and diploid cells. In haploid cells, there are two cell types determined by the co-dominant a and α alleles, at the MAting Type (MAT) locus [7,8]. While haploid cells can be annihilated under high-stress conditions such as nutrient starvation, diploid cells can endure harsh environments by undergoing sporulation during sexual reproduction (meiosis). Through a series of meiotic divisions, the four resulting haploid progenies are packaged into individual spores [9]. Opposite mating type haploid cells mate with each other to reform diploid cells, following spore germination. Similar to that of other fungi, many yeast species preserve sexual reproduction although it is not obligatory [10].
Fermentation of carbohydrate sources into ethanol using yeast species is one of the oldest technologies. The earliest evidence about the existence of a fermented beverage dates to the Neolithic period [11]. As is widely known, various yeast species are still used for many essential biotechnological processes including the production of beer, wine, and biofuels. Saccharomyces cerevisiae is one of the most extensively distinguished yeast species and the preferred host for many bioprocesses Table 1. S. cerevisiae genes described in this review.

Sex Determination and Behaviour in Budding Yeasts
Mating of the opposite mating type haploid cells (MATa and MATα) results in the formation of a new diploid strain (MATa/α). Phenotypic differences arise from the different gene expression patterns regulated by the MAT loci. The differences in gene expression are summarized in Table 2. Genes involved in mating are divided into three categories that include a-type-specific genes (asg), α-type-specific genes (αsg), and haploid-specific genes (hsg) [32]. MATa haploids express the genes a1 whereas MATα haploids express the genes α1 and α2. The synthesized protein α1 is the activator of αsg, and α2 functions as a repressor of asg. While the synthesized a1 protein does not influence the expression of mating-related genes, the formation of a1-α2 complex suppresses the expression of α1 and hsg in MATa/α diploid cells. In S. cerevisiae, the process of mating is regulated by a single canonical G-protein signaling pathway [33]. Most of components in this pathway are shared between MATa and MATα cells ( Figure 1). MATa cells express the G-protein coupled receptor (GPCR) Ste2, which detects the α-factor mating pheromone expressed from the MFα1 gene in MATα cells. Reciprocally, MATα cells express the receptor Ste3, which binds with the a-factor pheromone expressed from the MFA1 gene in MATa cells. Pheromone stimulation leads to the activation of heterotrimeric G-proteins comprising Gpa1 (Gα), Ste4 (Gβ) and Ste18 (Gγ) through the GPCR. The activated G-protein subsequently dissociates into Gα and Gβγ complex subunits, and the Gβγ complex induces activation of the mitogen-activated protein kinase (MAPK) cascade [19] followed by transcriptional activation for cell fusion (Figure 1). While the G-protein subunit genes GPA1, STE4, and STE18 are hsg, the pheromone and receptor genes are asg or αsg [18,34]. Expression of these genes is regulated by the MAT gene, resulting in mating behaviors between the MATa and MATα cells. In the generated MATa/α diploid cells, pheromone response signal transduction pathway does not function due to the lack of components.
MATa/α diploid cells can undergo meiosis under starvation conditions, while MATa and MATα haploid cells cannot undergo meiosis when exposed to such an environment. Meiosis is regulated by the expression of the Inducer of MEiosis 1 (IME1) gene, commonly known as the master initiator of meiosis [20]. In nutrient-rich environments, IME1 is suppressed by two major nutrient sensing signaling pathways [21,35] (Figure 2). Target of rapamycin complex I (TORC1) and protein kinase A (PKA) activities are suppressed in the absence of nitrogen and glucose. IME1 expression is activated by starvation signaling and repressed by the gene products of the Regulator of MEiosis 1 (RME1) and IME1 Regulatory Transcripts (IRT1) [21]. The difference in gene expression (Table 1) is a result of differential gene regulation by the a1-α2 complex. In both MATa and MATα haploid cells, the absence of the a1-α2 complex induces expression of the RME1 gene, a member of hsg. Rme1 is a zinc finger protein that activates expression of a long noncoding RNA, the product of IRT1, which in turn represses IME1. Meiosis is not induced by starvation signaling in haploid cells, as the expression of the IME1 gene is repressed. In contrast, the a1-α2 complex directly represses the expression of the RME1 gene, and the gene product of IME1 activates the meiosis pathway in MATa/α diploid cells in response to starvation. It should be noted that the a1-α2 complex is the most important factor distinguishing MATa and MATα haploids from MATa/α diploids. cells. Pheromone stimulation leads to the activation of heterotrimeric G-proteins comprising Gpa1 (Gα), Ste4 (Gβ) and Ste18 (Gγ) through the GPCR. The activated G-protein subsequently dissociates into Gα and Gβγ complex subunits, and the Gβγ complex induces activation of the mitogen-activated protein kinase (MAPK) cascade [19] followed by transcriptional activation for cell fusion (Figure 1). While the G-protein subunit genes GPA1, STE4, and STE18 are hsg, the pheromone and receptor genes are asg or αsg [18,34]. Expression of these genes is regulated by the MAT gene, resulting in mating behaviors between the MATa and MATα cells. In the generated MATa/α diploid cells, pheromone response signal transduction pathway does not function due to the lack of components. Figure 1. Signal transduction pathway of budding yeasts in response to mating pheromones. Mating pheromones, a-factor and α-factor, are secreted from MATa and MATα cells, respectively. Binding of the mating pheromones to a seven-transmembrane, G-protein-coupled receptor (GPCR) on the cellsurface, leads to activation of heterotrimeric G-proteins composed of Gα, Gβ, and Gγ. The activated G-proteins subsequently dissociate into Gα and a Gβγ dimer, and then Gβγ dimer induces activation of the mitogen-activated protein kinase (MAPK) cascade, resulting in mating-responsive gene expression.
MATa/α diploid cells can undergo meiosis under starvation conditions, while MATa and MATα haploid cells cannot undergo meiosis when exposed to such an environment. Meiosis is regulated by the expression of the Inducer of MEiosis 1 (IME1) gene, commonly known as the master initiator of meiosis [20]. In nutrient-rich environments, IME1 is suppressed by two major nutrient sensing signaling pathways [21,35] (Figure 2). Target of rapamycin complex I (TORC1) and protein kinase A (PKA) activities are suppressed in the absence of nitrogen and glucose. IME1 expression is activated by starvation signaling and repressed by the gene products of the Regulator of MEiosis 1 (RME1) and IME1 Regulatory Transcripts (IRT1) [21]. The difference in gene expression (Table 1) is a result of differential gene regulation by the a1-α2 complex. In both MATa and MATα haploid cells, the absence of the a1-α2 complex induces expression of the RME1 gene, a member of hsg. Rme1 is a zinc finger protein that activates expression of a long noncoding RNA, the product of IRT1, which in turn represses IME1. Meiosis is not induced by starvation signaling in haploid cells, as the expression of the IME1 gene is repressed. In contrast, the a1-α2 complex directly represses the expression of the RME1 gene, and the gene product of IME1 activates the meiosis pathway in MATa/α diploid cells in response to starvation. It should be noted that the a1-α2 complex is the most important factor distinguishing MATa and MATα haploids from MATa/α diploids.  While IME1 expression is activated by starvation signaling, it is also repressed by the gene products of the Regulator of MEiosis 1 (RME1) and IME1 Regulatory Transcripts (IRT1). In both haploid cells (MATa and MATα), the absence of the a1-α2 complex permits expression of the RME1 gene. Then Rme1 transcribes a long noncoding RNA, the product of IRT1, which in turn represses IME1. The a1-α2 complex directly represses the expression of the RME1 gene, and the gene product of IME1 activates the meiosis pathway in MATa/α diploid cells in response to starvation.

Mating Type Switching and Autodiploidization
Whereas cell type determination in many multicellular organisms is an irreversible process, the mating types of haploids (a and α) of the Saccharomycetaceae yeast family can interconvert in a reversible manner by a programmed DNA-rearrangement process called mating type switching [18,22]. Following spore germination, many haploid cells mate with the nearby cells of the opposite mating type. In contrast, to mate and become diploid cells, haploid cells that cannot contact the mating partners convert sister cells from one mating type to the other. The endonuclease Ho is responsible for the mating type switching in S. cerevisiae. Following DNA cleavage at the MAT locus, replacement of a or α DNA sequences is induced by the opposite sequences derived from one of the two silent donor loci, HMRa and HMLα [36]. It is known that loss-of-function mutations in the HO gene (ho allele) prevent mating type switching of the haploid cells [37]. Yeast strains containing the ho allele are categorized into heterothallic yeasts, whereas homothallic yeasts maintain the HO gene inducing mating type switching in the haploid cells.
The HO gene is a member of hsg and is repressed by the a1-α2 complex in the diploid cells. In addition, mating type switching occurs exclusively in the mother cells, and never in the daughter cells ( Figure 3), due to the difference in HO gene expression. Mother/daughter asymmetric HO gene expression is induced by the product of the HO-specific repressor gene, known as the Asymmetric Synthesis of HO (ASH1) [23]. Delivery of the ASH1 mRNA to the daughter cells is achieved by the products of the Swi5p-dependent HO Expression (SHE) genes. She2 binds to ASH1 mRNA in the nucleus and mediates export to the cytoplasm, after which She3 associates with the ribonucleoprotein particle (RNP). She proteins act as adapters for the type V myosin Myo4 (also called She1), transporting the RNP complex along the actin fiber to the bud tip [24]. ASH1 mRNA is finally translated in the daughter cells, and Ash1 binds its consensus sequences within the upstream repression sequence 1 (URS1) of the HO gene in the following G1 phase. Meanwhile the endonuclease Ho is exclusively synthesized in the mother cells (only during late G1), resulting in mating type switching. Using this programmed DNA-rearrangement, yeasts can proceed with sexual reproduction despite beginning from just a single haploid cell.

Mating Type Switching and Autodiploidization
Whereas cell type determination in many multicellular organisms is an irreversible process, the mating types of haploids (a and α) of the Saccharomycetaceae yeast family can interconvert in a reversible manner by a programmed DNA-rearrangement process called mating type switching [18,22]. Following spore germination, many haploid cells mate with the nearby cells of the opposite mating type. In contrast, to mate and become diploid cells, haploid cells that cannot contact the mating partners convert sister cells from one mating type to the other. The endonuclease Ho is responsible for the mating type switching in S. cerevisiae. Following DNA cleavage at the MAT locus, replacement of a or α DNA sequences is induced by the opposite sequences derived from one of the two silent donor loci, HMRa and HMLα [36]. It is known that loss-of-function mutations in the HO gene (ho allele) prevent mating type switching of the haploid cells [37]. Yeast strains containing the ho allele are categorized into heterothallic yeasts, whereas homothallic yeasts maintain the HO gene inducing mating type switching in the haploid cells.
The HO gene is a member of hsg and is repressed by the a1-α2 complex in the diploid cells. In addition, mating type switching occurs exclusively in the mother cells, and never in the daughter cells ( Figure 3), due to the difference in HO gene expression. Mother/daughter asymmetric HO gene expression is induced by the product of the HO-specific repressor gene, known as the Asymmetric Synthesis of HO (ASH1) [23]. Delivery of the ASH1 mRNA to the daughter cells is achieved by the products of the Swi5p-dependent HO Expression (SHE) genes. She2 binds to ASH1 mRNA in the nucleus and mediates export to the cytoplasm, after which She3 associates with the ribonucleoprotein particle (RNP). She proteins act as adapters for the type V myosin Myo4 (also called She1), transporting the RNP complex along the actin fiber to the bud tip [24]. ASH1 mRNA is finally translated in the daughter cells, and Ash1 binds its consensus sequences within the upstream repression sequence 1 (URS1) of the HO gene in the following G1 phase. Meanwhile the endonuclease Ho is exclusively synthesized in the mother cells (only during late G1), resulting in mating type switching. Using this programmed DNA-rearrangement, yeasts can proceed with sexual reproduction despite beginning from just a single haploid cell. Figure 3. Asymmetric expression of the HO gene between mother and daughter cells. Mating type switching occurs exclusively in the mother cells, and never in the daughter cells. Mother/daughter asymmetric HO gene expression is induced by the product of the HO-specific repressor gene, Asymmetric Synthesis of HO (ASH1). ASH1 mRNA is delivered to the daughter cells by the products of the Swi5p-dependent HO Expression (SHE) genes. She proteins bind to ASH1 mRNA in the nucleus, mediate export to the cytoplasm, and transport them along the actin fiber to the bud tip. In Figure 3. Asymmetric expression of the HO gene between mother and daughter cells. Mating type switching occurs exclusively in the mother cells, and never in the daughter cells. Mother/daughter asymmetric HO gene expression is induced by the product of the HO-specific repressor gene, Asymmetric Synthesis of HO (ASH1). ASH1 mRNA is delivered to the daughter cells by the products of the Swi5p-dependent HO Expression (SHE) genes. She proteins bind to ASH1 mRNA in the nucleus, mediate export to the cytoplasm, and transport them along the actin fiber to the bud tip. In the daughter cells, Ash1 binds its consensus sequences within the upstream repression sequence 1 (URS1) of the HO gene in the following G1 phase. Meanwhile the endonuclease Ho is exclusively synthesized in the mother cells, resulting in mating type switching.

Past and Current Situations of Yeast Sporulation
In 1996, the genome of S. cerevisiae was completely sequenced as the first among the eukaryotic organisms [38]. For several decades, yeast has served as a model for eukaryotic organisms, which has contributed significantly to the progress of cell biology. Compared to the non-essential genes, yeast essential genes exhibited more homologs in the other organisms [39,40]. Importantly, tetrad spore analysis revealed 1105 essential genes of S. cerevisiae, based on whether it is possible to acquire haploid cells with the target genes [40,41]. In fact, yeast strains without sporulation deficiency have been utilized for these analyses.
In industrial use, sporulation invites great attention as a means of crossbreeding. Unfortunately, many industrial S. cerevisiae strains are known to exhibit poor ability for sexual reproduction. It was reported that >80% of the lager, >60% of sake, >50% of ale, and >15% of wine yeast strains did not undergo sporulation, and more than half of the spores were inviable after dissection in >80% of the lager, >30% of sake, >70% of ale, and >50% of wine yeast strains undergoing sporulation [17]. Many species of the Saccharomyces genus other than S. cerevisiae are used for industrial fermentation (including the production of lager beer, wine, and cider), some of which are interspecies hybrids [42][43][44][45] such as the lager yeast S. pastorianus derived from the two parental species S. cerevisiae and S. eubayanus. While these different species can mate, hybrid offspring are almost completely sterile, producing <1% viable spores [46]. The hybrid genomes can undergo extensive modifications including loss of chromosomes and various types of recombination events, resulting in aneuploidies, massive copy number variations [47]. Hybrid sterility seems to be mainly caused by the instability of the chromosomes [43].
Regardless of the defects in sexual reproduction, the generated crossbreds (mainly diploid cells) can be continued for use in ethanol fermentation by asexual proliferation. In addition, mutagenesis is also one of the conventional and effective approaches used for modifying certain traits; hence, there is a possibility that numerous industrial yeast strains have been derived from a few strains with defects in sexual reproduction. In modern human societies, industrial yeasts do not need to withstand harsh environmental changes if they maintain excellent characteristics during fermentation. In fact, yeast strains with higher fermentation rates have been historically selected to shorten the fermentation periods and spontaneously inhibit the growth of undesirable microorganisms. Recently, it was found that gene mutations in nutrient-regulated genes are related to high fermentation rates [48].

Loss-of-Function Mutations in Nutrient-Regulated Genes
Sake yeast strains produce significantly more ethanol during fermentation than any other type of S. cerevisiae strains. Gene expression profiling has revealed that the stress-responsive transcription factors Msn2 and Msn4 (Msn2/4) are inactivated in sake yeast strains during fermentation [25,26]. It is known that sake yeast strains are more sensitive to ethanol stress and heat shock than the laboratory strains. Dysfunction of Msn2/4 increases the fermentation rates through the modification of glucose metabolism, including reduced synthesis of storage and structural carbohydrates [48]. Although a specific loss-of-function mutation was found in the MSN4 gene (msn4 C1540T ) of sake yeast strains, this mutation does not appear to be solely responsible for the inactivation of Msn2/4 [43]. In addition, deletion of the MSN4 gene in the S. cerevisiae laboratory strains leads to only a modest increase in their fermentation rates [26]. The inactivation of Msn4 is inadequate to achieve the high fermentation rate observed in sake yeasts, suggesting the existence of an alternative factor that inhibits Msn2/4 activities and contributes to the brewing properties.
In S. cerevisiae, nutrient starvation causes inactivation of the nutrient-sensory kinases such as PKA and TORC1, resulting in activation of the PAS kinase Rim15 (Figure 4). RIM15 was identified as one of functionally related genes (Regulator of IME2) that stimulate early meiotic gene expression in yeast [27]. In response to nutrient starvation, RIM15 contributes to gene expression of the IME1 gene [49] and helps IME1 activate the downstream sporulation genes such as IME2 [28,29]. Meanwhile, Rim15-responsive transcripts are involved in stress resistance (essentially heat shock and oxidative stress resistance), carbohydrate metabolism, and respiration [50]. Most of these genes are regulated by the transcription factors, Gis1 binding to the post-diauxic shift (PDS) element, and Msn2/4 recognizing the stress response element (STRE) [30]. Rim15 has been known to be involved in the activation of Msn2/4 ( Figure 4)  In S. cerevisiae, nutrient starvation causes inactivation of the nutrient-sensory kinases such as PKA and TORC1, resulting in activation of the PAS kinase Rim15 (Figure 4). RIM15 was identified as one of functionally related genes (Regulator of IME2) that stimulate early meiotic gene expression in yeast [27]. In response to nutrient starvation, RIM15 contributes to gene expression of the IME1 gene [49] and helps IME1 activate the downstream sporulation genes such as IME2 [28,29]. Meanwhile, Rim15-responsive transcripts are involved in stress resistance (essentially heat shock and oxidative stress resistance), carbohydrate metabolism, and respiration [50]. Most of these genes are regulated by the transcription factors, Gis1 binding to the post-diauxic shift (PDS) element, and Msn2/4 recognizing the stress response element (STRE) [30]. Rim15 has been known to be involved in the activation of Msn2/4 ( Figure 4)   In sake yeast strains, the insertion of a single adenine nucleotide was found immediately after position 5067, compared to that in S. cerevisiae laboratory strains. The frameshift mutation at nucleotide position 5068 (rim15 5055insA ) generates a premature stop codon that shortens the RIM15 gene product by 75 amino acids in the C-terminal region [48], which severely impairs the function of Rim15. Since sake yeast strains contain both copies of the rim15 5055insA mutations, they exhibit a high fermentation rate, stress sensitivity, and severe defects in meiosis. It might be inevitable that infertility challenges are broadly seen in industrial yeasts selected by using fermentation characteristics as the indicator. By introduction of the wild type RIM15 in the sake yeast strains, the sporulation efficiency significantly increased; however, they produced few viable spores [16].

Chromosome Recombination Defect in Meiosis
In normal meiosis, the diploid genome is reduced to produce haploid gametes through two rounds of nuclear division that follow a single round of DNA replication ( Figure 5). Homologous parental chromosomes pair and separate at the first cell division (meiosis I), while sister chromatids segregate during the second division (meiosis II). One of the main reasons for poor spore viability in In sake yeast strains, the insertion of a single adenine nucleotide was found immediately after position 5067, compared to that in S. cerevisiae laboratory strains. The frameshift mutation at nucleotide position 5068 (rim15 5055insA ) generates a premature stop codon that shortens the RIM15 gene product by 75 amino acids in the C-terminal region [48], which severely impairs the function of Rim15. Since sake yeast strains contain both copies of the rim15 5055insA mutations, they exhibit a high fermentation rate, stress sensitivity, and severe defects in meiosis. It might be inevitable that infertility challenges are broadly seen in industrial yeasts selected by using fermentation characteristics as the indicator. By introduction of the wild type RIM15 in the sake yeast strains, the sporulation efficiency significantly increased; however, they produced few viable spores [16].

Chromosome Recombination Defect in Meiosis
In normal meiosis, the diploid genome is reduced to produce haploid gametes through two rounds of nuclear division that follow a single round of DNA replication ( Figure 5). Homologous parental chromosomes pair and separate at the first cell division (meiosis I), while sister chromatids segregate during the second division (meiosis II). One of the main reasons for poor spore viability in industrial yeast strains is a chromosome recombination defect in meiosis. Since meiotic recombination is an essential step for linkage of the homologous parental chromosomes, its absence leads to uneven inheritance of the chromosomes (nondisjunction) leading to aneuploidy and inviable gametes [16]. Spo11 is a meiosis-specific protein that initiates meiotic recombination by catalyzing the transesterification reaction in DNA double-strand breaks. The deletion of the SPO11 gene completely abolishes meiotic recombination, and prevents linkage of the homologous chromosomes, resulting in random segregation of homologous parental chromosomes at meiosis I [31,52]. industrial yeast strains is a chromosome recombination defect in meiosis. Since meiotic recombination is an essential step for linkage of the homologous parental chromosomes, its absence leads to uneven inheritance of the chromosomes (nondisjunction) leading to aneuploidy and inviable gametes [16]. Spo11 is a meiosis-specific protein that initiates meiotic recombination by catalyzing the transesterification reaction in DNA double-strand breaks. The deletion of the SPO11 gene completely abolishes meiotic recombination, and prevents linkage of the homologous chromosomes, resulting in random segregation of homologous parental chromosomes at meiosis I [31,52]. Compared to that in S. cerevisiae laboratory strains, there are several mutations in the SPO11 gene of sake yeast strains. The single nucleotide substitution of C73T (spo11 C73T ) that results in the R25W missense mutation is found in sake yeast strains with significantly poor spore viability. Introduction of spo11 C73T to the laboratory strain causes sporulation deficiency and spore nonviability, whereas the transformants with other mutations except for C73T show a high sporulation efficiency and produce many viable colonies [16]. As described above, many sake yeast strains also contain loss-of-function mutations in the nutrient-regulated genes. Therefore, in these strains, introduction of the intact SPO11 cannot completely resolve the infertility challenges. In fact, introduction of both intact RIM15 and SPO11 leads to high sporulation efficiency and spore viability [16].

Difficulty and Contradiction in Conventional Crossbreeding
Whole genome and transcriptome analyses have revealed genetic differences between strains of the same species. It was reported that mutations in >200 genes affect yeast sporulation [53,54]. They have been divided into the following five categories according to gene functions [55][56][57]: (i) mitochondria/metabolism, (ii) vacuolar and autophagy, (iii) meiosis/early sporulation, (iv) spore formation, and (v) undefined role in sporulation. Mutants in the first two categories arrest the process prior to the meiotic divisions. Sporulation is triggered by nitrogen starvation and requires the presence of a nonfermentable carbon source [58]. The nonfermentable carbon source is metabolized Figure 5. The role of chromosomal recombination in meiosis. Homologous parental chromosomes pair and separate at meiosis I, while sister chromatids segregate during meiosis II. Spo11 is a meiosis-specific protein that initiates meiotic recombination, and the deletion of SPO11 gene completely prevents linkage of the homologous chromosomes, resulting in random segregation of homologous parental chromosomes at meiosis I. Compared to that in S. cerevisiae laboratory strains, there are several mutations in the SPO11 gene of sake yeast strains. The single nucleotide substitution of C73T (spo11 C73T ) that results in the R25W missense mutation is found in sake yeast strains with significantly poor spore viability. Introduction of spo11 C73T to the laboratory strain causes sporulation deficiency and spore non-viability, whereas the transformants with other mutations except for C73T show a high sporulation efficiency and produce many viable colonies [16]. As described above, many sake yeast strains also contain loss-of-function mutations in the nutrient-regulated genes. Therefore, in these strains, introduction of the intact SPO11 cannot completely resolve the infertility challenges. In fact, introduction of both intact RIM15 and SPO11 leads to high sporulation efficiency and spore viability [16].

Difficulty and Contradiction in Conventional Crossbreeding
Whole genome and transcriptome analyses have revealed genetic differences between strains of the same species. It was reported that mutations in >200 genes affect yeast sporulation [53,54]. They have been divided into the following five categories according to gene functions [55][56][57]: (i) mitochondria/metabolism, (ii) vacuolar and autophagy, (iii) meiosis/early sporulation, (iv) spore formation, and (v) undefined role in sporulation. Mutants in the first two categories arrest the process prior to the meiotic divisions. Sporulation is triggered by nitrogen starvation and requires the presence of a nonfermentable carbon source [58]. The nonfermentable carbon source is metabolized to produce energy via the Krebs cycle and to provide precursors of macromolecules such as nucleic acids, lipids, and polysaccharides required for spore formation. Therefore, respiratory-incompetence blocks entry into sporulation [59,60]. Similarly, because new protein synthesis during sporulation requires the recycling of preexisting proteins, many genes involved in delivery of proteins to the vacuole or in autophagy contributes to entry into the meiotic divisions.
In natural selection, these mutations might be purged from the yeast genome. Unfortunately, however, it is not easy to remove the deleterious mutations for sporulation from industrial yeast strains. To begin with, screening methods corresponding to each specific phenotype of interest are usually required for the introduction of a specific change at the target genes. In the case of a phenotype involved in meiosis, observation of individual sporulation consumes immense time to evaluate the clones that have been generated and isolated from the cell population. In addition, some meiosis-related genes such as RIM15 interfere with the fermentation characteristics that need to be improved mainly in the breeding of industrial yeasts. Removal of mutations in these genes inevitably causes a decline in productivity of the fermentation processes without any improvement in the product quality.
Spore viability in interspecific hybrids further declines due to nucleotide divergence [61,62]. The Saccharomyces genus is composed of eight species (S. arboricola, S. cerevisiae, S. eubayanus, S. jurei, S. kudriavzevii, S. mikatae, S. paradoxus and S. uvarum) [63][64][65][66], and some studies demonstrated that interspecific hybridization can also occur in nature [67,68]. While industrial Saccharomyces hybrids inherit good fermentation performance such as growth ability at lower temperatures [69][70][71][72][73], hybridization gives rise to instability of the chromosomes, which in turn results in spore unviability [74,75]. It has been also reported that sporulation efficiency is significantly anticorrelated with the fraction of the genome associated with large (>20 kb) amplifications and deletions [76]. Aneuploidies and massive copy number variations are also seen in inter-strain hybrids of S. cerevisiae. A ploidy study revealed that sake yeast strains (belonging to S. cerevisiae) are mostly diploids, but some are aneuploid [77]. Regarding the effect of chromosomal aneuploidy on the brewing characteristics, it has been reported that trisomy of chromosomes XI and XIV leads to the pyruvate underproduction [78]. Unfortunately, although a variety of yeast strains have been generated using it, the technique of conventional crossbreeding has reached its limit in trait modification of yeast strains with favorable brewing characteristics. To meet the needs of future yeast strain development in industrial use, alternative techniques to bypass sporulation are essentially required to acquire mating-competent yeast cells.

Artificial Mating Type Conversion Using the Ho Endonuclease
The Ho endonuclease expressed in haploid cells of homothallic yeasts has been used for artificial mating type conversion [79]. Through forced expression of the HO gene, replacement of a or α DNA sequences at the MAT loci is induced by opposite mating-type sequences derived from one of two silent donor loci (HMRa and HMLα), also in MATa/α diploid cells. By the action of the Ho endonuclease, MATa/a and MATα/α diploid cells are generated from parental MATa/α diploid cells (Figure 6), which possess the same mating ability as either MATa and MATα haploids generated via sporulation. Since autopolyploidization impedes the isolation of mating-type-converted (MATa/a and MATα/α) cells, hybrid tetraploid strains generated during cocultivation along with the artificial mating type conversion have been selected from the cell population by using marker genes introduced into each parental strain [80,81]. To avoid autopolyploidization and isolate the mating-competent cells generated through artificial mating type conversion, the a1-α2 complex is artificially formed by introducing synthetic gene expression circuits (Figure 7) into the parental MATa/α diploid cells, using episomal vectors [36,82]. MATa/a or MATα/α diploids generated through artificial mating type conversion are easily isolated from the cell population since the abundance ratio of target cells in the cell population reaches much beyond 20% by preventing autopolyploidization [36]. The mating type of the isolated cells can be identified by a mating assay using a and α tester strains [82][83][84]. Hybrid tetraploid strains are generated through cocultivation of MATa/a and MATα/α diploids derived from different strains, and selected from the cell population by using marker genes on episomal vectors introduced into each parental strain. By the removal of episomal vectors, the introduced synthetic gene expression circuits can be rendered non-functional in yeast cells, and there is no detection of exogenous DNA throughout the entire genome of crossbreeds [12]. To avoid autopolyploidization and isolate the mating-competent cells generated through artificial mating type conversion, the a1-α2 complex is artificially formed by introducing synthetic gene expression circuits (Figure 7) into the parental MATa/α diploid cells, using episomal vectors [36,82]. MATa/a or MATα/α diploids generated through artificial mating type conversion are easily isolated from the cell population since the abundance ratio of target cells in the cell population reaches much beyond 20% by preventing autopolyploidization [36]. The mating type of the isolated cells can be identified by a mating assay using a and α tester strains [82][83][84]. Hybrid tetraploid strains are generated through cocultivation of MATa/a and MATα/α diploids derived from different strains, and selected from the cell population by using marker genes on episomal vectors introduced into each parental strain. By the removal of episomal vectors, the introduced synthetic gene expression circuits can be rendered non-functional in yeast cells, and there is no detection of exogenous DNA throughout the entire genome of crossbreeds [12]. The MAT locus consists of five regions (W, X, Y, Z1, and Z2) based on the sequences shared with the two silent copies, HMLα and HMRa [85]. The difference between the a and α type DNA sequences exists in the Y region (Ya or Yα). The nucleotide sequences recognized by the Ho endonuclease reside within the Z1 regions of the MAT loci [86]. It is known that a single nucleotide substitution (called stuck mutation) in the Z1 region prevents mating type switching by severely reducing the Ho endonuclease cleavage. To clarify the presence or absence of stuck mutations, the MAT gene sequences of target strains should be confirmed in advance. Importantly, the mechanism of mating type switching is common among the Saccharomyces genus [22]. There is variation in the sequence of the HO genes between species [87,88], synthetic interspecies hybrids, S. cerevisiae × S. eubayanus, S. cerevisiae × S. kudriavzevii, and S. cerevisiae × S. uvarum, were generated as synthetic lager, Belgian, and cider strains, respectively, by introducing and expressing the HO gene derived from S. cerevisiae [80]. Furthermore, Saccharomyces allopolyploids of six species were similarly generated through mating type conversion of synthetic interspecies hybrids [81], which implies the applicability of these approaches to a wide range of industrial yeast strains (including natural interspecies hybrids).

Chromosomal Aberrations Related with Mating Type
The MAT genes are located on chromosome III, which is the most unstable of all 16 S. cerevisiae chromosomes [89]. This instability irregularly produces other sexual reproduction routes derived from chromosomal aberrations occurring during the mitotic division. There are several kinds of chromosomal aberrations such as loss of heterozygosity (LOH) and mitotic chromosome loss. In diploid cells, LOH is a naturally occurring process that generates homozygous loci through chromosomal rearrangement of the heterozygous loci [90]. LOH occurring at the MAT loci within MATa/α diploid cells produces either a MATa/a or a MATα/α diploid cell, whose spontaneous frequency is less than 1 × 10 −4 [91]. Mitotic chromosome loss is another naturally occurring event where diploid cells lose single or multiple chromosomes [92]. The loss frequency of chromosome III in yeast diploid cells has been reported at 5 × 10 −5 [89], and yeast cells having lost one of the two copies of chromosome III acquire either a or α type mating ability. While the spontaneous frequencies of such events are quite low, polyploids and aneuploids can be naturally generated not by sporulation but by chromosomal aberration during mitotic division.
Rare mating based on chromosomal aberration during mitotic division can offer a way to obtain hybrid strains [93]. In a rare mating, MATa/α diploid cells can hybridize with a haploid cell of the opposite mating type following chromosomal aberration during cocultivation. The isolation of The MAT locus consists of five regions (W, X, Y, Z1, and Z2) based on the sequences shared with the two silent copies, HMLα and HMRa [85]. The difference between the a and α type DNA sequences exists in the Y region (Ya or Yα). The nucleotide sequences recognized by the Ho endonuclease reside within the Z1 regions of the MAT loci [86]. It is known that a single nucleotide substitution (called stuck mutation) in the Z1 region prevents mating type switching by severely reducing the Ho endonuclease cleavage. To clarify the presence or absence of stuck mutations, the MAT gene sequences of target strains should be confirmed in advance. Importantly, the mechanism of mating type switching is common among the Saccharomyces genus [22]. There is variation in the sequence of the HO genes between species [87,88], synthetic interspecies hybrids, S. cerevisiae × S. eubayanus, S. cerevisiae × S. kudriavzevii, and S. cerevisiae × S. uvarum, were generated as synthetic lager, Belgian, and cider strains, respectively, by introducing and expressing the HO gene derived from S. cerevisiae [80]. Furthermore, Saccharomyces allopolyploids of six species were similarly generated through mating type conversion of synthetic interspecies hybrids [81], which implies the applicability of these approaches to a wide range of industrial yeast strains (including natural interspecies hybrids).

Chromosomal Aberrations Related with Mating Type
The MAT genes are located on chromosome III, which is the most unstable of all 16 S. cerevisiae chromosomes [89]. This instability irregularly produces other sexual reproduction routes derived from chromosomal aberrations occurring during the mitotic division. There are several kinds of chromosomal aberrations such as loss of heterozygosity (LOH) and mitotic chromosome loss. In diploid cells, LOH is a naturally occurring process that generates homozygous loci through chromosomal rearrangement of the heterozygous loci [90]. LOH occurring at the MAT loci within MATa/α diploid cells produces either a MATa/a or a MATα/α diploid cell, whose spontaneous frequency is less than 1 × 10 −4 [91]. Mitotic chromosome loss is another naturally occurring event where diploid cells lose single or multiple chromosomes [92]. The loss frequency of chromosome III in yeast diploid cells has been reported at 5 × 10 −5 [89], and yeast cells having lost one of the two copies of chromosome III acquire either a or α type mating ability. While the spontaneous frequencies of such events are quite low, polyploids and aneuploids can be naturally generated not by sporulation but by chromosomal aberration during mitotic division.
Rare mating based on chromosomal aberration during mitotic division can offer a way to obtain hybrid strains [93]. In a rare mating, MATa/α diploid cells can hybridize with a haploid cell of the opposite mating type following chromosomal aberration during cocultivation. The isolation of outcrossed hybrids is often achieved by using a respiratory-deficient and an auxotrophic parental strain, making rare hybrids easily selectable by their prototrophy and respiratory proficiency [94][95][96]. In addition, the isolation of both a and α type mating-competent cells derived from MATa/α diploid cells is achieved by using the expression of the marker gene in a mating type-specific manner [83,97]. This permits the efficient selection of a or α type cells from within a cell population, despite the significantly low frequency of chromosomal aberration. Meanwhile, dual-dye fluorescent staining was carried out to isolate hybrid cells without using marker genes [98,99]. By labeling each parental strain with a respective fluorescent dye prior to mating, mated cells can be enriched by fluorescence-activated cell sorting (FACS). When applied to rare mating, FACS-based selection of dual-stained cells allowed efficient enrichment of interspecies Saccharomyces hybrids without requiring selectable hybrid phenotypes, both for homothallic and heterothallic strains [99].

Conversion of the MAT Gene in Haploid Cells
Since most of the strains used in industrial use are homothallic [17,100], a haploid in the spore progeny remains only transiently, despite the success in sporulation. In wine yeast strains, conversion of homothallic strains to heterothallism was achieved by introduction of the ho allele [100]. Meanwhile, heterothallic yeasts containing the ho allele are also used in brewing. It is known that most of the modern sake yeast strains are heterothallic yeasts [101]. While facing difficulties in sporulation, many kinds of haploid strains have been generated in the process of crossbreeding. As a matter of course, pairs of the same mating type cannot mate with each other, and conversion of the mating types of either parent is a prerequisite for crossbreeding.
Similar to diploid cells, the Ho endonuclease has been used for the artificial mating type conversion of haploid cells [36,79]. However, it is difficult to convert mating types of yeast strains with stuck mutation at the MAT gene as described above. Recently, the MAT gene of haploid cells was successfully substituted with a synthetic DNA containing MATa or MATα, through homologous recombination [84]. Mating type alteration of yeast cells was performed by suppressing the mating ability of parental cells in the same manner shown in Figure 7. Due to the absence of cell-cell interaction between MATa and MATα cells, the mating type-altered cells independently exist in the cell population. Regardless of the stuck mutations, all haploid strains of heterothallic yeasts can be utilized for crossbreeding in any combination.
In addition, CRISPR/Cas9 system [102,103] was used for artificial mating type switching through a double-strand break at the MAT locus. According to gRNA sequences, Cas9-mediated double strand breaks (DSBs) are generated at either Ya or Yα region [104]. Unlike mating type conversion using Ho endonuclease, it is possible to switch mating type of yeast strains lacking HMRa and HMLα by addition of the donor DNA segment. Similar to haploid cells, diploid cells were switched to the specified mating type (MATa/a or MATα/α) using this approach [104]. As described above, mating-competent yeast cells can be acquired without going through sporulation (Figure 8). It might be a potential public concern that some of these methods require genetic modifications, unlike conventional crossbreeding. However, including genome editing with CRISPR/Cas9 system, all genetic modifications can be introduced into the yeast cells using episomal vectors, which allows exclusion of the exogenous DNA throughout the entire genome of crossbreeds after its removal.

Figure 8.
Variations of yeast crossbreeding. Since yeast strains domesticated for fermentation often have infertility challenges, it takes tremendous time and effort to generate hybrid strains in conventional crossbreeding. Meanwhile, mating-competent yeast cells can be acquired through alternative processes such as double strand breaks (DSBs) at the MAT loci and DNA replication errors. The common purpose, which is to generate hybrid strains, would be achieved by using either method.

Conclusions
Infertility challenges pose a significantly high challenge for yeast trait modification using conventional crossbreeding. There are numerous valuable yeast resources that cannot be fully exploited due to sporulation deficiency. In natural environments, deleterious mutations for sexual reproduction would have been purged from the yeast genome. Industrial yeasts used in modern human societies have acquired excellent characteristics in fermentation with partial sacrifice of sporulation efficiency, due to protection from harsh environmental changes. Although a variety of yeast strains have been generated during the long human history, the technique of conventional crossbreeding used for yeast trait modification based on the molecular mechanism has reached its limit. To meet the needs of future yeast strain development for each custom-engineered fermentation process, alternative techniques to bypass sporulation are essentially required to acquire matingcompetent yeast cells. In this review, I have described several alternative methods to generate matingcompetent cells from parental yeasts exhibiting severe deficiencies in sporulation. Even if there is a need to use exogenous genes, all genetic modifications described here can be introduced into the yeast cells using episomal vectors, which allows complete removal of the modifications following the isolation of crossbreds. These methods will enable us to improve yeast resources beyond the limitation of conventional crossbreeding by facilitating the generation of new yeast strains exhibiting desirable properties for industrial applications.

Conflicts of Interest:
The author declares that he is one of inventors on Japanese patent No. 6112548 and 6284139.
Abbreviations asg a-type-specific genes αsg α-type-specific genes DSBs double strand breaks FACS fluorescence-activated cell sorting GPCR G-protein coupled receptor hsg haploid-specific genes Figure 8. Variations of yeast crossbreeding. Since yeast strains domesticated for fermentation often have infertility challenges, it takes tremendous time and effort to generate hybrid strains in conventional crossbreeding. Meanwhile, mating-competent yeast cells can be acquired through alternative processes such as double strand breaks (DSBs) at the MAT loci and DNA replication errors. The common purpose, which is to generate hybrid strains, would be achieved by using either method.

Conclusions
Infertility challenges pose a significantly high challenge for yeast trait modification using conventional crossbreeding. There are numerous valuable yeast resources that cannot be fully exploited due to sporulation deficiency. In natural environments, deleterious mutations for sexual reproduction would have been purged from the yeast genome. Industrial yeasts used in modern human societies have acquired excellent characteristics in fermentation with partial sacrifice of sporulation efficiency, due to protection from harsh environmental changes. Although a variety of yeast strains have been generated during the long human history, the technique of conventional crossbreeding used for yeast trait modification based on the molecular mechanism has reached its limit. To meet the needs of future yeast strain development for each custom-engineered fermentation process, alternative techniques to bypass sporulation are essentially required to acquire mating-competent yeast cells. In this review, I have described several alternative methods to generate mating-competent cells from parental yeasts exhibiting severe deficiencies in sporulation. Even if there is a need to use exogenous genes, all genetic modifications described here can be introduced into the yeast cells using episomal vectors, which allows complete removal of the modifications following the isolation of crossbreds. These methods will enable us to improve yeast resources beyond the limitation of conventional crossbreeding by facilitating the generation of new yeast strains exhibiting desirable properties for industrial applications.