A Simple and Reliable Method for Hybridization of Homothallic Wine Strains of Saccharomyces cerevisiae

Abstract

A procedure was developed for the hybridization and improvement of homothallic industrial wine yeasts. Killer cycloheximide-sensitive strains were crossed with killer-sensitive cycloheximide-resistant strains to get killer cycloheximide-resistant hybrids, thereby enabling hybrid selection and identification. This procedure also allows backcrossing of spore colonies from the hybrids with parental strains.

Hybridization is the first method to be considered for improvement of diploid industrial yeast strains. For heterothallic strains, one can select hybrids by micromanipulating the zygotes formed between meiotic segregants with complementary mating types. For homothallic strains, the most frequently used of the industrial yeasts (1, 5, 8, 11, 13, 16), different strategies have to be used. Some approaches use laboratory haploid heterothallic strains with appropriate markers to be crossed with haploid cells from spores of homothallic industrial strains. In these cases, hybrids are easy to detect, and improvement of some industrial yeasts has been described (5, 7, 8, 15). However, backcrossing is needed to regenerate the industrial strain properties that are lacking in laboratory strains (7, 15). For two homothallic strains, hybridization can be accomplished by mixing sporulated cultures (13, 19). Cell fusion can occur between spore germination and diploidization. However, hybrids are obtained with lower frequency, and they are difficult to identify. In this paper, we describe a method for homothallic wine yeast spore hybridization that allows hybrid selection and identification, so that a greater number of hybrids can be obtained easily. The method takes advantage of the fact that the killer phenotype is very frequent among wine yeasts (3, 6, 14, 17). By changing the culture conditions, one can make yeasts conjugate or kill each other.

The killer phenotype in Saccharomyces cerevisiae is determined by double-stranded RNA molecules. Several situations that lead to the loss of this phenotype have been described. Among them are high-temperature growth and the presence of certain compounds, such as ethidium bromide, 5-fluorouracil, or cycloheximide, in the culture medium (2, 4, 10, 18). We isolated cycloheximide-resistant (CYH^R) spontaneous mutants in YEPD-CYH medium (1% Bacto-yeast extract, 2% Bacto-peptone, 2% glucose, 2% Bacto-agar, 2 μg of cycloheximide per ml) from killer K2 (K⁺) cycloheximide-sensitive (CYH^S) diploid homothallic wine yeasts. These mutants lose the killer phenotype, becoming killer-sensitive (K⁻) strains. All of the mutants isolated bear heterozygous dominant chromosomal mutations (JP73R and JP85R in Table 1).

TABLE 1.

Characteristics of the S. cerevisiae strains used in this study

Strain	Origin	Genotype (description)	% Spore viability	Spore/colony size segregationa	CYHR/CYHS segregation
JP73	Enological	MATα/MATa HO/HO cyhS/cyhS [k2+] (killer, immune)	38	2l:2nv (1l:3nv)b	0:4
JP85	Enological	MATα/MATa HO/HO cyhS/cyhS [k2+] (killer, immune)	100	2l:2s	0:4
JP88	Enological	MATα/MATa HO/HO cyhS/cyhS [k2+] (killer, immune)	100	2l:2s	0:4
JP73R	Spontaneous mutant from JP73	MATα/MATa HO/HO CYHR/cyhS [k2o] (nonkiller, sensitive)	86	4l:0 (3l:1nv)c	2:2
JP85R	Spontaneous mutant from JP85	MATα/MATa HO/HO CYHR/cyhS [k2o] (nonkiller, sensitive)	100	2l:2s	2:2
854D	Spore clone from JP85	MATα/MATa HO/HO cyhS/cyhS [k2+] (killer, immune)	100	4l:0	0:4
881A	Spore clone from JP88	MATα/MATa HO/HO cyhS/cyhS [k2+] (killer, immune)	100	4l:0	0:4
73R11D	Spore clone from JP73R	MATα/MATa HO/HO CYHR/CYHR [k2o] (nonkiller, sensitive)	100	4l:0	4:0
85R4A	Spore clone from JP85R	MATα/MATa HO/HO CYHR/CYHR [k2o] (nonkiller, sensitive)	100	4l:0	4:0
3AR	Hybrid from 854D × 85R4A	MATα/MATa HO/HO CYHR/cyhS [k2+] (killer, immune)	100	4l:0	2:2
7AR	Hybrid from 881A × 85R4A	MATα/MATa HO/HO CYHR/cyhS [k2+] (killer, immune)	100	4l:0	2:2
15CR	Hybrid from 854D × 73R11D	MATα/MATa HO/HO CYHR/cyhS [k2+] (killer, immune)	100	4l:0	2:2

^aColony size: l, large or normal; s, small; nv, nonviable. 

^b50% of the tetrads are 2l:2nv, and 50% are 1l:3nv. 

^c55% of the tetrads are 4l:0, and 45% are 3l:1nv. 

The K⁻ CYH^R spontaneous mutants JP73R and JP85R and the wild-type prototrophic yeasts JP85 and JP88 (K⁺ CYH^S) were sporulated, and tetrad analysis was performed (12 to 14 tetrads for each strain). Standard yeast genetic procedures were used for sporulation of cultures and dissection of asci (9). The spore viability and the segregation ratio of the colony size are given in Table 1. In order to obtain homozygous strains free from growth-retarding alleles, single-spore cultures having large colony size were sporulated again, and 12 to 14 tetrads were analyzed. The procedure was repeated until 100% of the viable spores and a uniform large colony size were obtained for all of the spore colonies in YEPD (one repeat for JP85, JP88, and JP85R and three repeats for JP73R). The single-spore cultures chosen from JP73R and JP85R progenies were always CYH^R. Finally, we chose four homozygous single-spore cultures for breeding: 854D from JP85, 881A from JP88, 85R4A from JP85R, and 73R11D from JP73. All of these spore cultures showed must fermentation kinetics and enological properties that were very similar to those of the original yeasts (data not shown). Consequently, all of these homozygous strains can be considered equally suitable for industrial fermentation.

The homozygous K⁻ CYH^R (854D and 881A) and K⁺ CYH^S (85R4A and 73R11D) spore cultures were crossed to obtain K⁺ CYH^R hybrids. The following crosses were performed: 854D × 85R4A, 881A × 85R4A, and 854D × 73R11D. Since all of the strains are homothallic, it is necessary to mix spores to achieve conjugation of haploid cells before they become diploid as a consequence of a mating-type switch. The parental yeasts were inoculated in sporulation medium (1% potassium acetate, 0.1% Bacto-yeast extract, 0.05% glucose, 2% Bacto-agar) and incubated at 25°C until more than 50% of the tetrads (from 10 to 30 days) were obtained. A small dab of each sporulated culture was treated with Zymolyase to digest the ascus wall (9). Tetrads were picked up with the microneedle of a micromanipulator, each ascus was broken separately, and the four spores were mixed thoroughly with the four spores from another tetrad of a different parent. As controls, eight spore mixtures were created with two different tetrads from the same parental strain. Each cross was repeated five times (five mixtures). Since we accomplished three crosses with different parental strains, the total number of eight-spore mixtures was 35: 15 with spores from different parental strains and 20 with spores from the same parental strain (controls). All of the mixtures were mixed in YEPD plates and then incubated for 4 days at 30°C. Under these conditions, the K2 killer toxin is inactive, so that the killer cells do not kill the sensitive cells which can conjugate. Subsequently, the mixture cultures from the YEPD plates were inoculated in low-pH (pH 4.4) blue plates (9) and incubated for 4 to 5 days at 20°C. Under these conditions, the K2 killer toxin is active and kills the sensitive cells of the K⁻ CYH^R parents. However, the toxin kills cells neither from the K⁺ CYH^S parental strains nor from the K⁺ CYH^S and K⁺ CYH^R hybrids that originated by conjugation. Single-cell colonies of the culture mixtures grown on low-pH blue plates were isolated by spreading samples of the cultures over YEPD plates followed by incubation at 30°C. Four colonies of each mixture (20 of each cross) were replica plated to YEPD-CYH (2 μg of cycloheximide per ml) and incubated at 30°C. As expected, all of the colonies from the control crosses of K⁻ CYH^R parental strains (85R4A and 73R11D) grew in YEPD-CYH. On the contrary, no colony from the controls of K⁺ CYH^S parental strains (854D and 881A) grew. The results of the crosses between different parental strains were variable. In some mixtures, none of the four colonies was resistant to cycloheximide. Probably, there was no conjugation between cells from different parental strains, and all of the K⁻ CYH^R cells died in low-pH blue plates. Also, it is possible that we did not catch any hybrid among the four colonies chosen. However, in other mixtures, it was possible to obtain several CYH^R colonies (from 1 to 4). We were able to isolate CYH^R colonies from more than one mixture of each cross performed with different parental strains. These resistant yeasts should be K⁺ CYH^R hybrids.

To confirm this, we analyzed the killer phenotype of all 20 of the colonies isolated from each cross, both resistant and not resistant to cycloheximide. As expected, all of the CYH^S colonies from the controls of K⁺ CYH^S parental strains (854D × 854D and 881A × 881A) were K⁺, and all of the CYH^R colonies from the controls of K⁻ CYH^R parental strains (85R4A × 85R4A and 73R11D × 73R11D) were K⁻. With respect to the crosses of different parental strains (854D × 85R4A, 881A × 85R4A, and 854D × 73R11D), all of the CYH^S colonies were K⁺ (Table 2). These colonies are from cells of the K⁺ CYH^S parental strain that have not conjugated or from haploid cells of the same parent that have conjugated with a haploid sister cell of a different mating type. All of the CYH^R colonies from the crosses 854D × 85R4A and 854D × 73R11D and 50% of those from the cross 881A × 85R4A were K⁺. Therefore, all of these CYH^R colonies are hybrids obtained by conjugation of cells from different parental strains. The rest (50%) of the CYH^R colonies from the cross 881A × 85R4A were K⁻. They may correspond to K⁺ CYH^R parent cells that, for some reason, have resisted the killer toxin in low-pH blue plates or to K⁺ CYH^R hybrids that do not express the killer phenotype under our test conditions.

TABLE 2.

Crossbreeding of K⁺ CYH^S (854D and 881A) with K⁻ CYH^R (85R4A and 73R11D) homozygous homothallic strains to obtain killer K2 cycloheximide-resistant hybrids (K⁺ CYH^R)

Cross	No. of 8-spore mixtures	No. of colonies analyzed	% CYHR colonies	% CYHR colonies with killer phenotype
854D × 85R4A	5	20	30	100
881A × 85R4A	5	20	30	50
854D × 73R11D	5	20	20	100

To double-check the new hybrids, we analyzed the progeny of one K⁺ CYH^R hybrid from each cross (3AR from 854D × 85R4A, 7AR from 881A × 85R4A, and 15CR from 854D × 73R11D). After sporulation and dissection of 12 to 14 tetrads, the spore colonies were replica plated to YEPD-CYH with different concentrations of cycloheximide. The segregation ratio was 2CYH^R:2CYH^S for the three hybrids in cycloheximide concentrations equal to or lower than the MIC for the original CYH^R parental strain (Table 3). The MIC for all of the CYH^R spore colonies was the same as that for the corresponding parental strain. This shows that these heterozygous strains are actually hybrids that originated by the conjugation of haploid cells from spores of different parental strains.

TABLE 3.

Cycloheximide resistance segregation ratio of the K⁺ CYH^R hybrids

Hybrid	MIC for the parental CYHR strain (μg/ml)	CYHR segregation ratio at cycloheximide concn (μg/ml) ofa:
		2	50	100	2,000
3AR	2,000	2R:2S	2R:2S	2R:2S	0R:4S
7AR	2,000	2R:2S	2R:2S	2R:2S	0R:4S
15CR	100	2R:2S	2R:2S	0R:4S	0R:4S

^aR, resistant; S, sensitive. 

Vinification trials in sterile must with the hybrids 3AR, 7AR, and 15CR and with four K⁺ CYH^R spore colonies from each hybrid were performed as described by Regodón et al. (12). As a control, vinification trials with the original yeasts (JP85, JP88, JP73R, and JP85R) were performed in parallel. The fermentation kinetics and the quality of the resulting wines were analyzed. In all cases, the results obtained with the hybrids and the meiotic spore segregants were equal to or better than those of the original wine yeasts (data not shown). These results suggest that our method for homothallic yeast hybridization could be very useful for industrial yeast improvement. In our case, we obtained new hybrids with a killer phenotype that are very easy to monitor in industrial fermentations, just by replica plating to YEPD-CYH. Because cycloheximide inhibits mammalian cells, it is not currently used in human disease treatment. Moreover, cycloheximide resistance is widespread in wild yeasts; some of these yeasts (such as Brettanomyces or Dekkera and Zygosaccharomyces) are involved in spontaneous must fermentation and therefore are ingested by humans without toxic effects.

This hybridization method could also be used to enable genetic studies of natural wild yeast populations without the need for crossing with laboratory domesticated strains. Back crossing of K⁺ CYH^R spore cultures from the hybrids with the K⁺ CYH^S parental yeasts is also possible if the killer spore culture cells are cured by growing them in YEPD-CYH. Care should be taken, because CYH^R strains do not lose the killer phenotype as easily as CYH^S strains (2). Cells should be spread over YEPD-CYH plates in order to produce single-cell colonies, and these should be tested for killer phenotype loss. We obtained killer phenotype losses of 20 to 80%, 28 to 40%, and 15 to 33% among different K⁺ CYH^R single-spore cultures from 3AR, 7AR, and 15CR, respectively.