YEAST

VOL. 6: 367-382

(1990)

The Chromosomal Constitution of Wine Strains of Saccharomyces cerevisiae ALAN T. BAKALINSKY* AND RICHARD SNOW? *Department of Food Science and Technotogy, Oregon State University, Corvallis. OR 97331-6602, U.S.A. *Genetics Department, University of California, Davis, CA 95616, U.S.A.

Received 20 July 1989; revised 22 January 1990

A general procedure is described for determining the chromosomal constitution of industrial strains of Saccharomyces cerevisiae based on analysis of segregation frequencies for input markers among random spore progeny of industrial-

laboratory strain hybrids. The multiply auxotrophic haploid testers used carried a dominant erythromycin-resistance marker, allowing hybrids to be selected in mass matings with spores produced by the wild-type industrial strains. Analysis of a number of independent crosses between the haploid testers and an unselected population of spores of each wine strain distinguished between disomic, trisomic and tetrasomic chromosomal complementsin the parents. Possible explanations for a significant class of aberrant segregation frequencies are discussed. Results of the analysis indicate that UCD Enology 522 (Montrachet) is diploid and possibly trisomic for chromosome VII; 522X is diploid; UCD Enology 505 (California Champagne) is disomic for chromosome XVI, trisomic for chromosomes I, 11,111, VI, VIII, IX, X, XII, XV, tetrasomic for chromosomesIV, XI, XIII, XIV and either trisomic or tetrasomic for chromosomes V and VII; and that UCD Enology 595 (Pasteur Champagne) is disomic for chromosomes I, 11,111,IX, XVI, trisomic for chromosomes IV, VI, X, XII, XIV, XV, tetrasomic for chromosomes V, VIII, XI, XI11 and either disomic or tetrasomic for chromosome VII. KEY WORDS - Yeast

genetics; wine yeast; Saccharomyces cerevisiae;chromosomes; karyotyping; aneuploidy.

INTRODUCTION Like most

industrially

important strains of

Succhuromyces cerevisiae, wine yeasts are poorly

characterized genetic entities. Present-day winemakers have inherited strains that have been domesticated and valued for their exceptional ethanol-tolerance, their resistance to sulfur dioxide and their ability to ferment grape juice containing 20-25% sugar rapidly and efficiently without producing off-flavors or aromas. As untamed genetic organisms they share with related baking, brewing and distilling strains certain difficulties that make them awkward subjects for genetic analysis and improvement. They are homothallic, sporulate inefficiently, produce few viable spores of which most are unable to mate, possess unknown chromosomal constitutions, a great deal of genetic heterogeneity and generally lack selectable genetic markers (Snow, 1983; Spencer and Spencer, 1983; Beckerich et ul., 1984; Subden, 1987; Rank et al., 1988). All of *Addressee for correspondence. 0749-503X/90/050367-16 $08.00 0 1990 by John Wiley & Sons Ltd

these difficulties greatly complicate efforts to introduce genetic improvements into the strains as well as attempts to conduct genetic analyses of important enological traits. Circumstantial evidence supporting the widespread occurrence of aneuploidy and/or polyploidy among industrial strains includes observations of poor spore viability, great variability in growth rates among spore progeny, and a very low frequency of mating-competent meiotic segregants. Why the strains should be aneuploid at all may be due in part to the well-known tolerance S. cerevisiae has for this condition (Parry and Cox, 1970). So long as meiosis is not a regular part of their life cycle-which one would expect to be true of industrial production strains-the maintenance of a stable but unbalanced chromosome complement may be advantageous. Some advantage may be gained by possession of extra sets or incomplete sets of chromosomes which provide the strains with a degree of protection from the effects of spontaneously arising recessive lethal or otherwise deleterious mutations. Extra copies ofchromosomes

368 also provide a means of increasing gene copy number, which in turn may give rise to more vigorous or faster-fermenting strains. This report describes experiments undertaken to determine the chromosomal constitution of three widely used wine strains and one derivative strain. Since the yeast chromosomes are too small to be counted reliably in stained preparations under the light microscope with conventional techniques, we have used a genetic method to estimate their number based on the segregation frequency of input markers among spore progeny of laboratory-wine strain hybrids. This approach has proven fruitful where it has been undertaken in conjunction with tetrad analysis of industrial-laboratory strain hybrids that retained high spore viability (Gunge, 1966; Sakai and Takahashi, 1972; Thornton and Eschenbruch, 1976; Takahashi, 1978; Cummings and Fogel, 1978; Keiding, 1985; Oda and Ouchi, 1989). In principle, the method involves massmating a laboratory haploid carrying known mutations on each of several different chromosomes with an unselected population of wild-type wine strain spores of unknown ploidy. The resultant hybrids are selected, grown and sporulated, and segregation frequencies for markers introduced by the laboratory parent are determined among random spore progeny. The frequencies reflect the number of wild-type copies of the marker geneand the chromosome with which it is associatedthat are introduced by the spore parents. The chromosomal constitution of the wine strain from which the spores are derived is inferred from the frequencies obtained from several independent crosses between different spores of the same parent and the laboratory tester strains. Expectations for disomic, trisomic and tetrasomic parents are different, and thus distinguishable. Our analysis of random spore progeny rather than complete tetrads has allowed us to overcome the otherwise considerable difficulty that poor spore viability poses, while use of a positive selection to obtain hybrids (Spencer and Spencer, 1980) has eliminated the need to isolate mating-competent spores. The method is not foolproof, but is more sensitive to the consequences of aneuploidy than methods based on determinations of DNA content (Leusch et al., 1985; Talbot et al., 1988) or cell volume. The latter procedures are more applicable to closely related euploid strains in which the number of chromosomes differs by factors of at least 2, rather than some fractional value which may be of the same order of magnitude as the variance

A. T. BAKALINSKY A N D R. SNOW

encountered in making the measurements (Snow, 1983). MATERIALS AND METHODS Media

Formulations are given in terms of recipes described previously (Sherman et al., 1986). YEPD and M are equivalent to YPD and SD, respectively. Petite is YPG+0.025% glucose, M-P+ER is SDG 0.4% erythromycin (Sigma) added after autoclaving, and SM is Synthetic Complete Medium containing required amino acids and bases at the prescribed concentration, except for uracil, which was added to a final concentration of 10 pg/ml. Diagnostic plates are SM lacking single nutrients or to which cycloheximide (Sigma) or Lcanavanine sulfate (Calbiochem) are added-the latter to an arginine-free medium-to final concentrations of 3 or 60 pg/ml, respectively. KAc is 2% potassium acetate and YEPA is 1% Difco yeast extract, 2% Difco peptone and 2% potassium acetate.

+

Sporulation

Cells grown overnight at 30°C on YEPD were transferred to either YEPA or KAc and held as long as 2 weeks at room temperature. The sporulated cultures were eventually transferred to liquid potassium acetate (0.2 M, pH 5 . 5 ) and stored as long as several weeks at 4°C until needed. Generation of random spores

The ether treatment of Dawes and Hardie (1974) was used. The 1 : 1 spore suspension : ether mixture was agitated at room temperature for 2&70 min before being pelleted at 400-500 x g , washed twice with water, resuspended in 2 ml of water, and stored at 4°C. No glusulase was used. Appropriate plating and dilutions were carried out to yield less than 160 colonies on any single YEPD plate. Due to the significant heterogeneity in clonal growth rates, the random spore plates were incubated from 4 to 8 days at 30°C to allow the most slowly growing colonies to emerge. All colonies except those that appeared misshapen or sectored were sampled from the plates chosen so that the total spore sample from any given hybrid consisted of 75-1 60 progeny.

3 69

CHROMOSOMAL CONSTITUTION OF WINE STRAINS

Table 1.

List of yeast strains.

Strains

Genotype

2785-7 2785-28 281 1-176 2806-61 2790-20 2807- 1 2789-3 2780-96 2945-38 X2 180-I A 2407'

a ade2 cyh2 his6 met14 trpl [ER] a ade2 cyh2 his6 met14 trpl [ER] a ade2 ura4 met8 his2 lys7 canl-100 [ER] a ade2 ura4 met8 his2 lys7 cyh2 cad-100 [ER] a ade2 arg4 cyh2 leu2-3 leu2-112 lys9 ura3 [ER] a ade2 arg4 cyh2 leu2-3 leu2-I12 lys9 ura3 [ER] a adel leul ura2 radl c a d - I 0 0 [ER] a adel leul ura2 radl c a d - I 0 0 [ER] a his4 ura2 adel trp.5 radl canl-100 [ER] a SUC2 ma1 me1 gal2 CUP1 aja SUC2jSUC2 mallma1 mellmel ga12jga12

AC39-1 2794d

a / a tyrlltyrl ajala TYRlltyrlltyrl SUC2/?/?mall?/? mel/?/?

505K 522A 522XA 595A

gal2/.?/.?CUPI/?/? ala HOIHO' ala HO/H@ aja HOIHO aja H O l H P

Source This study" This study" This study" This study" This study" This study" This study" This study" This study" YGSCb This study

CUPIICUPI YGSC This study

g g g, h f, g f, f9

dTesterstrains constructed for use in this study. Mutant alleles were derived from strains ofthe YGSC except for radl, which was originally designated uvs9 (Snow, 1967) and leu2-3 leu2-112 (Botstein et al., 1979). byeast Genetic Stock Center, Department of Molecular and Cellular Biology, Division of Biophysics and Cell Physiology, U.C. Berkeley, Berkeley, C A 94720. '2407 is our designation for a spontaneously diploidized (a/a) isolate from X2180-I A. d2794is a cross between 2407-la and AC39-1. "Tentatively assigned genotype at outset of this study. 'Obtained from the wine yeast collection at the Department of Viticulture and Enology, University of California, Davis, CA 95616. Single colony isolates from the respective parental wine strains: U C D Enology 505, U C D Enology 522 (Montrachet), 522X or U C D Enology 595. h522Xis a second-generation spore isolate of U C D 522. U C D 522 was sporulated, asci were dissected and a selected spore was grown. Among spore progeny of the latter clone, a single isolate was designated 522X (Snow, 1979).

in buffer containing 0.5 pg/ml ethidium bromide for Pulsed-field gel electrophoresis Cells were prepared by the agarose bead encapsu- 30 min and destained overnight. lation method essentially asdescribed by Overhauser and Radic (1 987). Construction and use of the haploid tester strains The electrophoretic apparatus (Bio-Rad ChefDR 11) has been described (Chu et al., 1986). Yeast strains are listed in Table 1. Tester strains Standard conditions were a 30-h run (23 h with a 60- carrying five to seven chromosomal markers each, s switching interval followed by 7 h with 100-sinter- covering the 16 well-defined yeast chromosomes vals) at 200 V, 1.8 liters of 0.5XTBE buffer, 65 ml of were constructed and scored by standard genetic I % agarose (Ultra pure DNA grade of Bio-Rad) methods (Mortimer and Hawthorne, 1969). Their cast in a 5" x 5.5" mold, and a buffer temperature important features were: (1) the presence of the maintained at 12°C by constant recycling through a dominant cytoplasmic marker, ER;(2) the inclusion cooling bath. Samples of beads, 35 pl, were loaded of recessive drug markers, either canl-100 or cyh2 to into wells which were subsequently sealed by allow random spore selection by methods other addition of a few drops of molten 1% low gelling- than an ether treatment, and adel and ade2 to pertemperature agarose (Bio-Rad). Gels were stained mit visual confirmation of hybridization among

370 progeny; and (3) the absence of more than one marker conferring the same phenotype within the same strain. Complementation tests were not required nor desired since most of the segregants were unable to mate because they were homothallic, were disomic for chromosome I11 and heterozygous for MAT, carried some undetermined mutant allele causing sterility, or for any combination of these possibilities. Mass matings, either spore-to-cell or cell-to-cell, were carried out between the homothallic wine strain spores or the mating-competent progeny of the heterothallic triploid, 2794, and the haploid testers. Actively growing haploid tester cells from an overnight M-P+ER or YEPD plate incubated at 30°C were suspended in YEPD and mixed with either a 2794 segregant or a sporulated wine strain pretreated with ether for 20-25 min to kill all unsporulated vegetative cells and a significant fraction of spores. No glusulase was used. The sporecell or cellkell mixtures were mixed well and 20 p1 aliquots placed on YEPD plates for overnight incubation at 30°C. The following day, cells from each aliquot were streaked onto M-P + ER plates, which were incubated for 3-6 days at 30°C. Cells taken from isolated colonies representing independent hybrids were sporulated and the spores subjected to random spore analysis.

A. T. BAKALINSKY AND R. SNOW

triploid-derived monosomic spore and the haploid tester should produce a hybrid whose random spores will segregate 1 + : 1 - for the chromosomal marker in question. On the other hand, a cross between a triploid-derived disomic spore and the haploid tester should produce a trisomic hybrid whose random spores will segregate 5 + : 1 - to 3.8+ : 1 - for the marker in question, depending on the marker-centromere distance. This assumes trivalent formation at meiosis I followed by regular two-by-one disjunction. The basis for these values as diagrammed in Figure 1 has been described by

I

00

Analysis of segregation frequencies Our indirect genetic estimate of ploidy was based on mass-mating marked haploid strains with a genetically undefined and unselected population of wine yeast spores. The resultant hybrids were selected, grown and sporulated and their random spores scored for segregating markers. The segregation frequencies indicated the ploidy of the spore parent. The ploidy of the wine strain from which the spore was derived was inferred by making assumptions about meiotic chromosome pairing and disjunction in the hybrid. Expectations for diploids, triploids and tetraploids are described below. If one assumes that a wine strain is diploid and homologous to the laboratory haploid tester strains, then its spores will be haploid and a laboratory-wine strain hybrid will be diploid. One would expect a 1 + : 1- segregation frequency for any given input marker among the hybrid’s random spore progeny. If one assumes that the wine strain is triploid, then on average half its spores will be monosomic and half disomic for any given chromosome. Again assuming homology, a cross between a

01

m

00

Figure 1 . Trivalent formation, two-by-one disjunction. Marker-centromere interval is equal to or greater than 50 cM.

Burnham (1962) in his analysis of ‘chromosome segregation’ and ‘maximum equational segregation’. These two segregation patterns represent theoretical extremes. In the case of ‘chromosome segregation’, crossovers are presumed never to occur between the marker of interest and its centromere, while for ‘maximum equational segregation’, a single such crossover is presumed always to occur. Further, crossovers are allowed to occur simultaneously between any two of the three synapsed homologs. This is a common cytological observation in polyploid plants (Fincham et al., 1979) and genetic evidence for its occurrence in a tetra-

37 1

CHROMOSOMAL CONSTITUTION OF WINE STRAINS

Table 2. Summary of some possible outcomes of meiosis in a +/+/- trisome with respect to the segregation frequency of an input recessive marker among random spore progeny. Wild-type : mutant I. Trivalent formation, two-by-one disjunction A. Disomes stable B. Disomes unstable 1. Random chromosome loss 2. Non-random chromosome loss a) Loss of homolog bearing the wild-type allele b) Loss of homolog bearing the mutant allele 11. Bivalent-univalent formation A. Random pairing 1. Loss of univalent 2. No loss of univalent a) Disomes stable b) Disomes unstable i) Random chromosome loss ii) Non-random chromosome loss 1. Loss of homolog bearing the wild-type allele 2 . Loss of homolog bearing the mutant allele B. Non-random pairing of chromosomes sharing greatest homology 1. Loss of unpaired non-homologous chromosome 2 . No loss of univalent a) Disomes stable b) Disomes unstable i) Random chromosome loss ii) Non-random chromosome loss 1. Loss of homolog bearing the wild-type allele 2. Loss of homolog bearing the mutant allele

ploid strain of S. cerevisiae trisomic for chromosome I11 has been obtained by Riley and Manney (1978). In accepting a segregation frequency ranging from 3.8 + : 1 - to 5 : 1 - for a marker as evidence of trisomy regardless of its proximity to a centromere, we are admitting a level of uncertainty that we justify by our lack of information concerning the occurrence of crossover interference and the unknown degree of homology shared by the three homologs. An alternative model, which for mathematical simplicity allows only two, but any two, homologs to take part in a given recombination event has been used to predict various ascal classes in trisomic diploids with reasonable success (Shaffer et al., 1971; Culbertson and Henry, 1973). Finally, if one assumes that the wine strain is tetraploid, then all its spores ought to be diploid. In crosses between the diploid spores and the homologous haploid testers, hybrid triploids will be formed. Segregation frequencies for input markers among the hybrid’s random spores will be identical to the case described

+

3‘8-5 : 1

2: 1 1-1.8 : 1 3.8-5 : 1

2:1 5:1

2:1 1:l 5:1

4:O 8:O 3: 1 1:1 8:O

for crosses between the disomic spores of a triploid and the haploid testers. Thus in principle, the three possibilities-disomy, trisomy and tetrasomy--can be distinguished by analysing the progeny of several independent hybrids. If the segregation frequency for a given marker is always 1 : 1 -, then the spores tested are most probably monosomic with respect to the chromosome in question, and the wine strain parent disomic. If the segregation frequency always ranges from 5 + : 1 - to 3.8+ : 1 -, then the spores tested are probably disomic and the wine strain parent tetrasomic. And if one obtains the former and latter results in about equal numbers, then the spores tested are probably derived from a trisomic parent. Table 2 summarizes several complicating possibilities.

+

Limitation to the analysis One limitation to the method is that only spores capable of mating are represented. Sterile and non-

372

A. T. BAKALINSKY A N D R. SNOW

Table 3 . Estimate of maximum spore killing by ether Strain"

Initial spores

Post-treatment viabilityb

29096

2.9 x 108/ml 5.5 x 107/ml

7.1 x lo3spores/ml 4.4 x lo4spores/ml

2863

Survival" 2.4 x 8.0 x

P( 2 two survivors)d

3.5 x 10-9 3.8 x

"29096 is a cross between 2789-3 and a 522XA spore. 2863 is a cross between 2785-28 and the laboratory triploid-derived spore, 2794-2a. b2909Gspores were treated for 70 min, 2863 spores for 20 min. 'These values presume initial 100% spore viability, complete killing of all vegetative cells, and that each colony formed on YEPD arose from a single spore. dThis term is the probability that two or more spores from the same ascus survived, calculated by summing the appropriate terms of the binomial expansion: 6p2q2+ 4p3q+p4, where p = frequency of survivors and q= 1 - p . No correction was made to take into account the observation that on average, fewer than four spores were present in a given ascus.

mating spore progeny are excluded unavoidably. While random segregation of alleles causing sterility ought to reduce the frequency of spores able to mate, it should not introduce a systematic bias against the recovery of unaffected chromosomes. In contrast, the inability of MATaIMATa disomes to mate is expected to introduce a systematic bias against the recovery of disomic spores from wine strains trisomic for chromosome 111. Although all chromosome 111 monosomes derived from such parents are expected to be competent maters, all disomes derived from the same parents are not. Thus, one would expect to detect more monosomes than disomes by use of our method. The same bias does not apply in the case of disomic spores produced by a chromosome I11 tetrasome. If we assume that disjunction at meiosis I is exclusively two by two, only disomes are recovered. Assuming heterozygosity at MAT, the frequency of homozygous chromosome I11 disomes depends on the arrangement of MAT alleles in the tetrasome. Thus, the frequency of mating-competent spores will be reduced but the exclusive detection of chromosome I11 disomes will accurately reflect the chromosomal constitution of the tetrasomic parent. RESULTS Haploid tester strains Tetrad analysis of crosses carried out to confirm the genotypes of the haploid tester strains listed in Table 1 yielded a preponderance of 2 : 2 - segregations for the chromosomal markers. In addition, non-significant X2-valuesfor I + : 1 - segregations among random spores confirmed that the ether treatment and spore sampling method generated

+

random progeny. Aberrant segregations for ade2 among tetrad and random spore progeny of the cross involving 2806-61 were detected suggesting the presence of a second ade mutation, and thus segregation data for ade2 involving this tester in crosses with wine strain spores or progeny of the heterothallic triploid were not used. Segregation data for ade2 among progeny of crosses between 522XA spores and 2785-28 were not used for the same reason. Random spores of the cross involving the tester 2945-38 were not analysed. Ether treatment A 20-min treatment was found sufficient to kill 3 4 x lo9 vegetative cells from early stationary phase cultures. Since this number was in excess of the number of vegetative cells present in any of the sporulated cultures subjected to treatment by a factor of 3 to 10, a nominal 20-min treatment was felt sufficient to kill any viable, unsporulated vegetative cells present among spores. The effect of the ether treatment on spores of two hybrids is shown in Table 3.2909G was presumed to produce viable spores at a frequency approximately equivalent to that of 522X-95% (Bakalinsky and Snow, 1990) and 2863 at a frequency similar to that measured for the triploid 2 7 9 6 1 4 % . An essential effect of the ether treatment, in addition to killing unsporulated vegetative cells, was the killing of a significant fraction of spores. This eliminated the need for enzymatic digestion of asci and subsequent sonication. For 2863, the maximum killing-which assumes initial 100% spore viability-would appear to be an overestimate of the actual killing since one of the parents of 2863 is derived from a triploid. If

373

CHROMOSOMAL CONSTITUTION OF WINE STRAINS

one assumes that the initial viability of spores produced by 2863 was closer to 10 than to loo%, the percent survival is still only 0.8 after a 20-min treatment. The corresponding probability of more than one spore per ascus surviving as calculated by use of the binomial expansion described in footnote ‘e’ of Table 3 is still reassuringly small. As noted elsewhere, we also observed a significant induction of petites by the ether treatment (p. 42 in Sherman et ul., 1986).

canavanine and for requirements for any of the amino acids or bases shared by our auxotrophic testers. While spores of UCD 522A, 522XA and UCD 595A carried no recessive mutations, a single leucine auxotroph was found among spores of UCD 505K. Consequently, segregation data for a leucine requirement were monitored but not used to make predictions about the ploidy of UCD 505K.

A test of the genetic predictions Mitotic segregation among spores

The heterothallic triploid, 2794, was constructed Random spore colonies that appeared misshapen specifically to test our assumptions concerning or sectored, presumably due to mitotic segregation triploid meiosis. One unavoidable limitation to the resulting from recombination or chromosome loss, use of a model triploid constructed from laboratory generally constituted a significant minority. Since strains was that its three sets of chromosomes probour interest was in evaluating meiotic events, we ably shared greater homology than those of the attempted to minimize the contribution of mitotic wine-laboratory strain hybrids. The principal findings, based on segregation data segregation by at least avoiding clones in which the consequences were visible. With the exception of presented in Tables 4 and 5, are that both trivalent spontaneous mutations arising in a single clone and bivalent-univalent formation occurs during within a random spore-derived colony, most mitotic meiotic pairing in trisomes; that disjunction at the segregation probably resulted from +/- + +/+, end of the first meiotic division is two by one; that - 1 - , - , or events. With respect to our recessive among mating-competent spores both monosomes auxotrophic markers, segregation resulting in a and disomes were recovered for every chromosome; change from wild-type to mutant would never have that segregation frequencies appear to vary with been detected on our diagnostic plates if a signifi- centromere-marker gene distances in a linear cant number of cells in the colony remained wild- fashion; and that among a significant and heterotype. The same is true in the case of a spontaneously geneous class of aberrant segregation frequencies, arising recessive mutation. In the case of the recess- two sub-classes can be explained-though not the occurrence of random, ive drug resistance markers, cyh2 and canl, it is exclusively so-by possible that a colony derived from a heterozygous post-meiotic chromsome loss and preferential pairsensitive spore (+/ - ) could have been mistakenly ing of chromosomes sharing greatest homology, respectively. scored as resistant if mitotic segregants ( - or - / -) arose early enough or grew at a significantly faster rate than the original clone on the YEPD plate. This Segregation frequencies consistent with monosomy would seem to be in part an unavoidable conse- and disomy. The majority of segregation frequencies quence of the prolonged incubation required for fit either a 1 : 1 - or a 3.8-5 : 1 - ratio, supportthe smallest aneuploid spores colonies to attain ing our model for meiotic trivalent formation and sufficient size to be visible. random two by one disjunction at the first division. While the frequencies that fit a 5 + : 1 - ratio are Testing .for naturally occurring recessive mutations also consistent with our model for bivalentunivalent pairing-as well as that of Shaffer et al. carried by the wine strains (1 97 1)-the model does not predict any second Since the wine strains are known to be genetically division segregation-dependent variation, and heterogeneous (Snow, 1983; Subden, 1987), we hence does not account for the segregation ratios tested for the presence of naturally occurring and that are less than 5 + : 1 -. Among the 70 segrepotentially interfering recessive mutations confer- gations that fit the 3.8-5 + : 1 - category, 59 (84%) ring the same phenotype as the markers carried by were consistent with a more precise prediction based our tester strains. Random spores generated from on the assumption that for markers 0-50 cM from the four wine strains were tested for sensitivity to their respective centromeres, the phenotypic freUV irradiation, resistance to cycloheximide and quencies ought to vary proportionally between t he

+

+

+

374

A. T. BAKALINSKY A N D R. SNOW

Table 4. Pooled" segregation data for random spore progeny of independent hybrids formed by crossing spores of the heterothallic triploid 2794 with haploid testers.

Chromosome

1 2 2 3 4 5 5 6 7 7 8 9 10 11

12 13 14 15 16 Total YOof total

Marker ade 1 met8 tyrl leu2 trpl can1 ura3 his2 leu1 cyh2 arg4 his6 ura2 met14 ura4 lys 7 lys9 ade2 radl

Number of segregations for which X*-valuesfitted the indicated ratio Segregations 1+ : 1 - 3.8-5+ : 1 - 2 + : 1 -

2 2 4 6 2 6 3 6 2 4 3 2 6 2 6 5 5 13 7 86 35.2

10 10 25 10 10 20 10 10 10 23 10 10 10 10 10 10 10 26 10

244 100

5

2 11 1

4 8 4 3 3 6 3 3 2 4 1 I 2 5 2 70 28.7

0 2 3 2 2 3 1 1 3 3 2 3 1 1 2 2 1

5 1 38 15.6

Otherb

3 4 7 1 2 3 2 0 2 10 2 2 1

3 1 2 2 3 0 50 20.5

"Data are based on a total of40crosses between 10mating-competent spores of 2794 and four different testers carrying markers covering 16 chromosomes. bAberrant ratios pooled in this category are sub-divided in Table 5.

Table 5. Classes of aberrant segregation frequencies among random spore progeny of hybrids formed in crosses between laboratory haploids and spores of 2794 or wine strains.

Spore parent

Total

2794 UCD 505K UCD 522A UCD 522XA UCD 595A

88 69 59 70 73

% of total aberrant segregations 2+ : I - 3 : 1 - 8.8-95+ : 1 -

No segregation

0 0 0 0 1.4

6.8 0 0 0 0

+

0.1 143.72+ : 1 - 1.3-t : 1 6.8 5.8 15.2 8.6 6.8

3.4 0 5.1

2.9 2.7

two theoretical extremes. This is expressed mathematically by the formula + : - = 5 - (x/50)( 1.2) : 1, where x = the marker-centromere distance in cM. For markers unlinked to a centromere, 50 is substituted for the actual value of x. A surprisingly good fit was obtained in spite of the fact that no correction was applied either to account for undetectable two-

43.2 63.8 79.1 88.6 65.8

39.8 30.4 0 0 23.3

strand double or higher order crossover events, or for interference. As expected of random two by one disjunction, both monosomes and disomes were recovered for every chromosome among each of the ten triploidderived spores, though not in equal frequency. However, among the total number of segregations

375

CHROMOSOMAL CONSTITUTION OF WINE STRAINS

that fit either the 1 + : 1 - or 3.8-5+ : 1 - classes, statistically equivalent numbers were found, 86 and 70, respectively. Contradictory ratios. Contradictory results were obtained in seven independent crosses involving chromosomes, V, VII and XV in 2794-1b, XV in 2794-3a, V in 2794-3b, VII in 2794-4a and I1 in 27944b where ratios indicated simultaneous monosomy and disomy for the same chromosome in the triploidderived spore parent. Since our testers carried only four to six markers each, with canl, cyh2 and ade2 carried in more than one, four crosses were required to determine the status of all 16 chromosomes in a given uncharacterized strain. Thus, it is possible though unlikely that different mitotic segregants of the uncharacterized strain, those having undergone chromosome loss, / -+ , may have been crossed to the set of haploid testers.

++ +

Aberrant segregation frequencies. Approximately one third of the segregation frequencies did not fit 1 + : 1 - or 3.8-5 + : 1 - ratios and are indicated in Table 4 in the ‘2+ : 1-’ and ‘other’ columns and are further sub-divided in Table 5. The largest class of aberrant frequencies fit a 2 + : 1 - ratio and was detected for all markers except adel, while the second most significant class ranged from 9.6-95 : 1 - and was detected at least once for all markers except arg4. The few cases (six) in which no segregation was detected for an input marker involving tyrl, canl, ura3, arg4 and met14 might be accounted for by pre-meiotic gene conversion, pre-meiotic mitotic crossing over, premeiotic chromosome loss of the homolog bearing the mutant allele (3n-+3n- l), meiotic gene conversion, post-meiotic loss of the homolog bearing the mutant allele (2n+2n- l), or exclusive nonrandom pairing and disjunction of ‘like homologs’ -those derived from the diploid parent of 2794. Four of six such ratios were detected in the progeny of hybrids derived from the same triploid-derived spore, 2794-2b. While we have no rational basis to determine the exact extent to which each of these mechanisms may play a role, gene conversions are sufficiently rare so that the occurrence of four such events involving different loci in crosses with the same triploid-derived parent ought to be highly improbable. Two smaller classes of frequencies, 0.44-0.69+ : 1 - and 1.37-1.45+ : 1 -,constituted 10% of all aberrant segregations and neither can be easily derived from the possibilities proposed in Table 2.

+

Chromosomal constitution of the wine strains Segregation data for the wine strains presented in Tables 5 and 6 are summarized in Table 7. Since the results are based on single cell isolates taken from bulk cultures known to harbor a great deal of genetic heterogeneity, we cannot conclude that their chromosomal constitutions are necessarily representative of every cell in the parental cultures. Further, it should be noted that since the wine strains are homothallic, hybrids formed between their spores and the laboratory haploids are necessarily heterozygous at HO. Among random spore progeny of such hybrids, the fraction arising from MAT homozygotes carrying HO consisted of clones that had diploidized. While this should not have directly altered their phenotype relative to the progenitor spores whose genotype we were interested in determining, it is very likely to have stimulated mitotic chromosome loss due to their doubled aneuploid genetic complement. We observed not only a great deal of variability in colony size, ranging over a factor of ten, but also a significant number of sectored or misshapen colonies presumably due to mitotic segregation. In contrast to the results obtained for the laboratory triploid 2794, where all of the chromosomal markers segregated in a manner indicating trisomy, these segregations indicate simultaneous disomy, trisomy and tetrasomy for various of the 16 chromosomes. Aberrant segregation frequencies constituted approximately one third of the total segregations and no new significant classes were detected relative to what was observed for the crosses involving the laboratory triploid-derived spores. Segregation frequencies consistent with monosomy and disomy. Of 14 of the 16 UCD 505K chromosomes whose copy number was unambiguously identified, nine appear to be trisomic, four tetrasomic and one disomic. In 83% (80 out of 96) of the segregations where the ratio fit 3.8-5 :.I - ,the data were in agreement with a more precise prediction based on marker-centromere distances as described for the analogous class of segregants among the triploid (2794)-derived spore progeny. Segregation data for UCD 595A hybrids indicate that it is apparently disomic for five, trisomic for six and tetrasomic for four chromosomes. In 87% (80 out of 92) of the segregations where the ratio fit 3.85 + : 1 -, the data were in agreement with the more precise prediction based on marker-centromere distances as previously described. Segregation for

+

adel met8 his4 1euZb trpl can 1 ura3 his2 trp5 cyh2 leu1 arg4 his6 ura2 met14 ura4 lys 7 lys9 ade2 radl

1 0 1 0 1 1 0 1 0 2 0 1 0 1 0 1 1 2 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 3 0 1 0 222

1 0 1 0 0 1 0 1 0 2 0 1 0 1 0 0 2 0 10 1 0 1 0 1 0 1 0 1 0 1 0 1 0 2 0 1 0 210

100

1 0 1 0 0 1 0 1 0 2 0 1 0 1 0 0 2 0 10 1 0 1 0 1 0 1 0 1 0 1 0 1 0 3 0 1 0 220

1 0 4 1 0 5 7 4 0 0 1 2 0 5 1 0 1 0 2 8 9 7 4 0 0 1 2 0 0 2 2 0 0 5 0 0 0 1 0 0 0 3 0 0 0 1 1 0 7 7 6 7 3 0 0 0 0 1 3 3 0 1 0 0 3 9 2 7 0 0 3 3 7 1 3 0 2 0 0 1 5 1 5 0 1 7 0 0 1 3 0 5 5 0 3 1 0 4 1 0 7 0 4 0 0 9 1 0 3 0 1 1 0 4 8 8 3 2 0 0 2 3 2 2 5 1 0 0 5 0 0 0 1 0 0 0 4 0 0 0 0 2 0 0 8 1 4 0 1 0 5 2 1 4 0 6 4 3 2 0 8 4 6 0 0 0 0 0 2 6 3 0 10 1 0 1 5 4 0 6 0 0 5 2 5 5 2 1 5 7 1 5 0 0 7 3 4 1 2 0 2 1 0 5 2 0 0 3 2 0 5 2 0 6 9 5 1 0 0 5 7 0 7 0 0 8 0 5 2 0 3 1 0 2 3 0 0 3 1 1 2 4 3 8 3 7 1 0 1 0 0 9 7 0 5 0 0 8 1 1 3 2 4 1 0 0 9 6 3 5 0 0 5 5 1 3 1 0 3 0 7 2 2 1 2 5 1 0 0 2 9 1 1 6 5 1 5 2 7 8 3 7 0 0 0 1 1 1 7 0 2 1 0 220 57 156 136 55 96 5 4 92 44 47 62 48 25 26 71 64 25 43 2 2 42 20 21 30 22 11

505

1 0 2 0 0 2 1 12 6

1

1 0 0

0

0

0 2 0 2 0 0 0

0 8 4

1 1

11

2 25

1

1 0 I

0 0 3 1 3 0 0 2

1

0

2 7

1

1 0 0 0 0 0 0 0 0 0 1 2 0 1 0 0

0 1 0

Number of segregations for which X2-valuesfitted the indicated ratio 1+:13.8-5+ 1 2+:1Othera 522 522X 595 505 522 522X 595 505 522 522X 595 505 522 522X 595

,'Aberrant ratios pooled in this category are sub-divided in Table 5. hSegregationdata for leu are tabulated, but were not used in determining the ploidy of UCD 505K since an unidentified requirement for leucine had segregated among its random spores.

I2 13 14 15 16 Total YOof total

10 11

1 2 3 3 4 5 5 6 7 7 7 8 9

Chromosome

Segregations Marker 505 522 522X 595

Table 6. Summary of segregation data for random spore progeny of independent hybrids formed by crossing spores of wine strains with haploid testers.

>

e

P

T 12U

~1

2

F

w

9

377

CHROMOSOMAL CONSTITUTION OF WINE STRAINS

Table 7. Number of chromosome copies in the wine strains based on genetic evidence."

Chromosome I 2 3 4 5 6 7 8 9 10 11

12 13 14 15

16

UCD 505

UCD 522

3 3 3

3

2 2 2 2 2 2

3or4

2or3

3 3 3 4 3 4 4 3 2

2 2 2 2 2 2 2 2 2

4 301-4

522X

UCD 595

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

2 2 2 3 4 3 201-4 4 2 3 4 3 4 3 3 2

'Conclusions are based on segregation frequencies presented in Table 5. If statistically equivalent numbers of segregations belonging to the I + : I - and 3-8-5f : I - classes were obtained. the parental wine strain was presumed to be trisomic for the chromosome in question. If between the two classes, segregations belonging only to the former or latter were obtained. the strain was presumed to be disomic or tetrasomic, respectively

results were obtained for chromosome VII, where cyh2 and trp5 were used as markers in crosses with spores of UCD 505K and cyh2 and leul in crosses with spores of UCD 522A and UCD 595A. For UCD 505K, trp5 segregations indicated trisomy whereas segregation for cyh2 indicated tetrasomy. For UCD 595A, segregation for leul indicated disomy and for cyh2, tetrasomy. In the case of UCD 522A, segregations for leul indicated disomy while segregations for cyh2 indicated trisomy. Aberrant segregationfrequencies. As for the laboratory triploid, the largest class of aberrant frequencies detected for the wine strains was 2 : 1 - ,while the second most significant class fit 8.87-53.5 : 1 -. The possibility that the latter class derives from preferential pairing of 'like homologs' within a trisome is strengthened by the observation that it was never detected for markers associated with chromosomes shown to be present in only two copies in the parental strains. Insignificant numbers of other segregation frequencies were detected and are classified in Table 5.

+

+

Electrophoretic karyotype of the strains

An profile Of the wine and laborstory strain chromosomes is shown in Figure 2. The

all chromosomal markers in 522XA hybrids was 1 + : 1 - indicating that the strain is diploid and confirming a previous finding for 13 of its chromosomes (Cummings and Fogel, 1978). Segregations for markers in UCD 522A hybrids were 1 : 1 except for one of two chromosome VII markers, which gave approximately equal numbers of 1 + : 1 - and 3.8-5 + : 1 - ratios indicating that the strain is diploid and possibly trisomic for chromosome VII. This contrasts with data for a different single cell isolate of UCD 522 (Montrachet) indicating tetrasomy for chromosome VIII (Fogel and Welch, 1987).

+

Contradictory ratios. Data for chromosome V among progeny of UCD 505K hybrids and for VII among progeny of UCD 505K and UCD 595A hybrids are contradictory. can1 and ura3 were used as markers for chromosome V though not in the Same hap1oid tester' Data based On segregations indicated tetrasomy whereas segregation for ura3 indicated trisomy. Similar contradictory

Figure 2. Electrophoretic profiles of laboratory and wine strains. Lanes A and H, 2407-la (laboratory haploid); B, 2407 (laboratory diploid); C, UCD 522A; D, 522XA; E, UCD 505K; F, UCD 595A;G , 2794 (laboratory triploid).

378 karyotype of the laboratory strains (lanes A, B, G and H) is virtually identical to that obtained by Carle and Olson (1985) and hence we presume that the identities of the bands are also the same. In the profile of the UCD 522A (lane C), band 1 migrates faster than the laboratory counterpart and although band 4 corresponds well to band 3 (chromosome 111) in the laboratory strains, an additional faster-migrating band is also present. The doublet corresponding to band 5 in the laboratory strains migrates faster as does the UCD 522A version of laboratory band 7. Laboratory band 1 1, which is a doublet, is partially resolved in UCD 522A. The laboratory form of band 12 (chromosome IV) is not present. Instead, a faster-migrating band is seen, which just merges with the partially resolved doublet corresponding to laboratory band 11. The karyotype of the genetically purified and presumably homozygous clone 522XA (lane D) has a different profile than its progenitor, UCD 522A. The band migrating just ahead of laboratory band 3 is absent. The laboratory doublet band 11, partially resolved in UCD 522A, is completely resolved. While one would expect all bands of 522XA to be present in UCD 522A-but not vice versa-the band corresponding to 12 in the laboratory strains is absent in UCD 522A. Profiles of single cell isolates of UCD 505K (lane E) and UCD 595A (lane F) reveal greater differences than seen between the laboratory strains and either UCD 522A or 522XA. The profile of UCD 505K indicates the presence of laboratory band 3, the faster-migrating band seen in the UCD 522A profile and a unique band migrating between laboratory bands 4 and 5. The faster-migrating form of laboratory band 7 seen in UCD 522A and 522XA has a diffuse counterpart in UCD 505K, which may be a doublet. Laboratory bands 8 and 9 co-migrate in the UCD 505K profile. Laboratory bands 10, 11 and 13 have approximate counterparts in UCD 505K but band 12 is absent. The profile of UCD 595A indicates a very faint counterpart to laboratory band 1, perhaps three merged bands migrating between laboratory bands 2 and 3, and a partially resolved doublet corresponding to laboratory band 5. Laboratory bands 6 through 9 have diffuse and approximate counterparts in this wine strain, which may indicate the presence of additional and unresolved chromosomes. As in UCD 505K, laboratory bands 10, 11 and 13 have counterparts in this strain while band 12 is absent. Instead, a faint band migrates slightly

A. T. BAKALINSKY AND R. SNOW

more slowly than the UCD 595A form of laboratory band 11. The range of variation observed between the electrophoretic karyotypes of the laboratory strains and those of the wine strains does not appear to exceed that commonly encountered between laboratory strains (Carle and Olson, 1985; Johnston and Mortimer, 1986; Ono and Ishino-Arao, 1988). The doublet bands of the wine strains whose intensities are inconsistent with their location relative to neighboring bands in the same lane may represent co-migrating chromosomes as demonstrated in laboratory strains, multiple copies of single chromosomes, or multiple copies of one or more of the co-migrating chromosomes. Given that the number of doublet bands appears to be at least the same in both wine and laboratory strain karyotypes, that most bands in the laboratory strain profiles have at least approximate if not precise counterparts in the wine strains, and that UCD 522A, UCD 505K and UCD 595A all have unique bands, these profiles suggest that the wine strains carry multiple versions of some of the yeast chromosomes. DISCUSSION A genetic estimate of the chromosomal constitution of previously uncharacterized or partially characterized industrial strains of S. cerevisiae has been undertaken. Initially, the study required a set of haploid tester strains carrying markers for the 16 wellcharacterized yeast chromosomes and a dominant mutation to permit selection of hybrids. The decision to analyse segregation frequencies for input markers among random spore rather than tetrad progeny was imposed by the impracticality of obtaining complete tetrads. In order to distinguish between disomes, trisomes and tetrasomes, ten independent hybrids were analysed from the mass matings carried out between spore progeny of the wine strains and each of the haploid testers. Distinctions were based on the expectation that disomes produce monosomic spores exclusively, tetrasomes produce disomic spores exclusively, and trisomes produce both in equal frequency. Since the majority of segregation frequencies for input markers among progeny of hybrids formed between the aneuploid spores produced by the triploid 2794 and laboratory haploids conformed to our expectations, we did not find it necessary to construct a tetraploid to demonstrate that all its spore progeny would be disomic. The choice of an ether treatment to generate random spores for analysis from the wine-laboratory

379

CHROMOSOMAL CONSTITUTION OF WINE STRAINS

strain hybrids allowed us to eliminate the usual glusulase and sonication steps and permitted use of YEPD plates on which the spore clones grew faster than they would have on minimal-based media to which canavanine or cycloheximide would have been added. Moreover, use of either drug would have selected against CANllcanl and CYH2lcyh2 disomes, respectively. Our decision to sample every random spore colony that arose on the YEPD plates (if fewer than 160 were present) was felt to be a less biased means of obtaining a random sample than an arbitrary choice among colonies. Since many clones were exceedingly slow-growing, we incubated the YEPD plates at 30°C for as long as 8 days. Even so, we undoubtedly selected against the most slowly growing since in many cases colonies not visible to the unaided eye could be seen under 30X magnification after prolonged incubation. Campbell et al. (1981) and Campbell and Doolittle (1987) incubated dissection plates containing tetrad progeny of hybrids formed between triploid-derived spores and a haploid for 16 days and observed that the viability (85%) of tetrad progeny of a laboratory triploid was no different than that of a diploid (that is, the spores germinated as frequently), but that the majority failed to given rise to visible colonies presumably due to the growth-retarding effects of multiple disomy.

Ratios consistent with predictions f o r monosomy or disomy The majority of segregation frequencies were consistent with our simplest assumptions. Our findings suggest that chromosome pairing may tolerate a degree of heterology as suggested by the chromosome length polymorphisms seen in the electrophoretic profiles of the strains. Based on the behaviour of small artificial chromosome derivatives, Dawson et al. (1 986) have shown that chromosome pairing at meiosis in yeast can occur in the absence of homology and crossing over. Further evidence in support of pairing and normal disjunction in the absence of extensive homology is suggested by the high spore viability retained by some hybrids formed in crosses between laboratory haploids and haploids in which single homoeologous chromosomes have been substituted from a lager strain of S . carlsbergensis (Kielland-Brandt et al., 1983; Nilsson-Tillgren et al., 1986; Pedersen, 1986a; Petersen et al., 1987).

Aberrant segregation frequencies

Aberrant segregations which fit neither 1 + : 1 nor 3.8-5 : 1 - ratios constituted a significant and unexpected heterogeneous class.

+

2 + :1 - .The largest sub-class of frequencies among all strains was 2 + : 1 - . This ratio is predicted by random post-meiotic chromosome loss in trisomes where pairing is trivalent or bivalent-univalent (Table 2). Suggestive evidence for its occurrence among hybrids formed between 2794 spore progeny and the haploid testers includes our observation that all 16 chromosomes tested were sometimes detected as monosomes and sometimes as disomes and that the minimal average estimate of multiple disomy among the ten triploid-derived spore parents was five, less than the eight expected if such a condition were stable (data not shown). Our estimate is minimal because we did not obtain interpretable data for all the chromosomes tested. Such a mechanism has been proposed to explain identical results among tetrads produced by hybrids formed in crosses between triploid-derived spores and a haploid tester (Campbell et al., 1981; Campbell and Doolittle, 1987). In the case of radl for the UCD 505K-derived data, adel, met8, leu2 and leu1 for the UCD 595Aderived data and all markers except met8 in 522XA and cyh2 in UCD 522A, where segregations indicated disomy for the associated chromosomes, 2 + : I - ratios were also detected. We previously had reasoned that the 2 + : 1 - class of frequencies was due to random, post-meiotic chromosome loss in a trisome (2n 1 -2n). No evidence is available to support the possibility that these chromosomes are present in more than two copies in the parental strains. Hence, although this class of frequencies was also detected for markers on chromosomes shown to be present in three copies in the hybrids, the explanation does not account for a significant number of segregations.

+

8.87-95+ :I - . These frequencies can be derived from models of meioses involving a significant but not exclusive degree of bivalent pairing among chromosomes sharing greatest homology in a trisome (Table 2). The diploid parent of 2794 (AC39- 1) arose from a spontaneously diploidized haploidand thus twoofits threesets ofchromosomes were almost certainly isogenic (R. Contopoulou, personal communication). Significantly,we detected these frequencies only for markers that segregated

380

A. T. BAKALINSKY A N D R. SNOW

in a manner indicating trisomy or tetrasomy for the associated chromosomes in the wine strains. To account for frequencies of various tetrad arrangements obtained from diploids trisomic or tetrasomic for chromosome I, James and Inhaber (1975) estimated that ‘like homologs’ paired in a trisome at a frequency approaching 0.5 rather than 0.33 predicted by chance. Contradictory ratios One explanation for the contradictory ratios detected is that linkage relationships among the chromosomes of the wine strains are not identical to those in the laboratory testers. A translocation may have placed one of the two or three genes found on a single chromosome V or VII in laboratory strains on different chromosomes present in different copy number in the wine strains. Another is that different copies of chromosomes V and VII in the wine strains may not be identical. Some may lack genes encoded by the laboratory versions as observed in a distiller’s strain tetrasomic for chromosome 111. Based on an analysis of restriction fragment length polymorphisms, four different versions were found, three lacking H M L and one H M R (Keiding, 1985). If we assume, for example, that UCD 505K is tetrasomic for chromosome V based on segregation for canl and that only three of the four homologs carry a functional URA3 gene, then the genotype of such a strain with respect to URA3 would be / / / - . Upon sporulation, both +/+ and +/- disomes would be produced, though not in equal frequency. In crosses with the ura3 tester, both +I+/- and + / - / - trisomes would result, the former producing random spores in a 3.8-5+ : 1 - ratio and the latter in a ratio very close to 1 + : 1 - . This reasoning would also explain the contradictory results obtained for chromosome VII in UCD 505K where segregations for trp2 and cyh2 indicated trisomy and tetrasomy, respectively. However, it would not explain the results obtained for chromosome VII in UCD 595A or UCD 522A. In both strains, segregation for leul indicated disomy while for cyh2, tetrasomy was indicated in UCD 595A and trisomy in UCD 522A. An additional possibility that we have tested and ruled out is that our ura3, canl, cyh2, trp5 and leul markers were misidentified and in fact represented some other mutations conferring identical phenotypes (linkage data not shown).

+++

Analysis of the electrophoretic proJiles While the electrophoretic profiles of UCD 522A, UCD 505K and UCD 595A provide qualitative

evidence for aneuploidy, a more useful quantitative analysis leading to a direct determination of chromosome copy number based on densitometric scanning of an ethidium bromide-stained gel or an X-ray film of a probed blot is complicated by at least four factors: the difficulty of loading DNA samples corresponding to equivalent numbers of cells per strain; variability in chromosome length observed for otherwise apparently homologous chromosomes; the possibility of translocations or deletions, which interfere with band identification if probes containing the rearranged sequence are used; and the possibility of sequence heterology of functionally homologous genes resulting in differential hybridization. This has been observed for probes derived from laboratory strains of S. cerevisiae used in hybridizations with a distilling strain (Keiding, 1985), lager strains (Kielland-Brandt et al., 1983; Pedersen, 1986b; Casey, 1986; Nilsson-Tillgren et ai., 1986; Peterson et al., 1987; Casey and Pedersen, 1988) and a lager strain in which both S . cerevisiaeand S. carlsbergensis-derived probes were used (Casey and Pedersen, 1988).At a minimum, obtaining quantitative data from these profiles will require inclusion of an internal standard such as an artificial chromosome (Burke et af., 1987) carrying a marker conferring a known additive phenotype and some means of correcting for differential staining due to the complicating factors described above. Conclusions Our genetic method of determining the chromosomal constitution of wine strains ought to be applicable to other industrial strains of S. cerevisiae. While its most serious flaw appears to be the significant background of aberrant 2 : 1 - segregations that cannot be explained in all cases, most of the data conform to accepted models of meiosis. An analysis of random spores instead of tetrads make it possible to overcome the poor spore viability that characterizes production strains and use of an ether treatment minimizes some of the tedium associated with obtaining progeny for analysis. While this analysis does not explain why aneuploidy is SO widespread among industrial strains of S. cerevisiae, it does provide a basis for further studies that might provide answers. Is it simply a matter of S. cerevisiae being exceptionally tolerant of the spontaneous loss or gain of chromosomes? Does a near-triploid or near-tetraploid genetic constitution provide significantly greater stability than a diploid constitution? As more genetic alterations are introduced into production strains, they will necessarily acquire a more

+

CHROMOSOMAL CONSTITUTION OF WINE STRAINS

domesticated genetic disposition. Part of this domestication will be reflected in improved sporulation efficiency, mating capacity and a reduction in the frequency of inviable spores produced. In cases where these improvements result from the strains becoming diploid or near-diploid, it will be possible to test directly the question of whether the former aneuploid or polyploid chromosome complement provided the strains with a detectable advantage with respect to fermentation properties. ACKNOWLEDGEMENTS We thank Norman Eaton for critically reading the manuscript and gratefully acknowledge Nicole Weinzettl and Wendy Ma for their expert and tireless technical assistance. This work was supported in part by the Agricultural Experiment Station at the University of California, Davis. REFERENCES Bakalinsky, A. T. and Snow, R. (1990). Conversion of wine strains of S . cerevisiae to heterothallism. App. Env. Micro. 56,849-857. Beckerich, J. M., Fournier, P., Gaillardin, C., Heslot, H., Rochet, M. and Treton, 9. (1984). Yeasts. In Ball, C. (Ed.), Genetics and Breeding of Industrial Microorganisms. CRC Press, Boca Raton, Florida, pp. 115-157. Botstein, D., Falco, S. C., Stewart, E., Brennan, M., Scherer, S., Stinchcomb, D. T., Struhl, K. and Davis, R. (1979). Sterile host yeast (SHY): a eukaryotic system of biological containment for recombinant DNA experiments. Gene 8, 17-24. Burke, D. T., Carle, G. F. and Olson, M. V. (1987). Cloning of large segments of exogenous DNA into yeast by means of artificial chromosome vectors. Science 236,806-8 12. Burnham. C. R. (1962). Discussions in Cytogenetics. Burgess Publishing Co., Minneapolis, Minnesota, pp. 184-191. Campbell, D. A., Doctor, J. S., Feuersanger, J. H. and Doolittle, M. M. (1981). Differential mitotic stability of yeast disomes derived from triploid meiosis. Genetics 98,239-255. Campbell, D. A. and Doolittle, M. M. (1987). Coincident chromosomal disomy in meiotic dyads from triploid yeast. Curr. Genet. 12,569-576. Carle, G . F. and Olson, M. V. (1985). An electrophoretic karyotype for yeast. Proc. Natl Acad. Sci. U.S.A. 82, 3756-3760. Casey, G . P. (1986). Molecular and genetic analysis of chromosomes X in Saccharomyces carisbergensis. Carlsberg Res. Commun. 51,343-362. Casey, G. P. and Pedersen, M. B. (1988). DNA sequence polymorphisms in the genus Saccharomyces. V.

38 1 Cloning and characterization of a LEU2 gene from S. carisbergensis. Carlsberg Res. Commun. 53,209-2 19. Chu, G., Vollrath, D. and Davis, R. W. (1986). Separation of large DNA molecules by contourclamped homogeneous electric field. Science 234, 1582-1585. Culbertson, M. R. and Henry, S. A. (1973). Genetic analysis of hybrid strains trisomic for the chromosome containing a fatty acid synthetase gene complex Vh-sl) in yeast. Genetics 75,441-458. Cummings, J. and Fogel, S. (1978). Genetic homology of wine yeasts with Saccharomyces cerevisiae. J . lnst. Brew. 84,267-270. Dawes, I. W. and Hardie, I. D. (1974). Selective killing of vegetative cells in sporulated yeast cultures by exDosure to diethyl ether. Mol. Gen. Genet. 131,281-289. Dawson, D. S., Murray, A. W. and Szostak, J. W. (I 986). An alternative pathway for meiotic chromosome segregation in yeast. Science234,713-717. Fincham, J. R. S., Day, P. R. and Radford, A. (1979). Fungal Genetics. Botanical monographs, vol. 4. 4th edn. University of California Press, p. 139. Fogel, S . and Welch, 3. W. (1987). Yeasts in biotechnology. In Stewart, G. G., Russell, I., Klein, R. D. and Hiebsch, R. R. (Eds), Biological Research on Industrial Yeasts, vol. I. CRC Press, Boca Raton, Florida, pp. 99-1 10. Gunge, N. (1966). Breeding of bakers’ yeastdetermination of the ploidy and an attempt to improve practical properties. Japan. J . Genetics. 41,203-214. James, A. P. and Inhaber, E. R. (1975). Evidence of preferential pairing of chromosomes at meiosis in aneuploid yeast. Genetics 79,561-571. Johnston, J. R. and Mortimer, R. K. (1986). Electrophoretic karyotyping of laboratory and commercial strains of Saccharomyces and other yeasts. Int. J . Sys. Bact. 36,569-572. Keiding, A. K . (1985). Genetic and molecular characterization of a distiller’s yeast. Carlsberg Res. Comniun. 50,95-125. Kielland-Brandt, M. C., Nilsson-Tillgren, T., Litske Petersen, J. G., Holmberg, S. and Gjermansen, C. (1983). Approaches to the genetic analysis and breeding of brewer’s yeast. In Spencer, J. F. T., Spencer, D. M . and Smith, A. R. W. (Eds), Yeast Genetics, Fundamental and Applied Aspects. Springer-Verlag, New York, NY, pp. 421-437. Leusch, H. G., Hoffmann, M. and Emeis, C. C. (1985). Fluorometric determination of the total DNA content of different yeast speciesusing 4,6-diamidino-2-phenylindol-dihydrochloride. Can. J. Micro. 31, 1164-1 166. Mortimer, R. K. and Hawthorne, D. C. (1969). Yeast genetics. In Rose, A. H . and Harrison, J . S. (Eds), The Yeasts. I. Biology of Yeasts. Academic Press, Sandiego, California, pp. 385-460. Nilsson-Tillgren, T., Gjermansen, C., Holmberg, S., Litske Petersen, J. G. and Kielland-Brandt, M. C. (1986). Analysis of chromosome V and the ZLVI gene

382 from Saccharomyces carlsbergensis. Carlsberg Res. Commun. 51,309-326. Oda, Y. and Ouchi, K . (1989). Genetic analysis of haploids from industrial strains of baker’s yeast. App. Env. Micro. 55, 1742-1 747. Ono, B. and Ishino-Arao, Y. (1988). Inheritance of chromosome length polymorphisms in Saccharomyces cerevisiae. Curr. Genet. 14,413-418. Overhauser, J. and Radic, M. Z. (1987). Encapsulation of cells in agarose beads for use with pulsed-field gel electrophoresis. Focus 9(3), 8-9. (A proprietary newsletter published by Bethesda Research Laboratories.) Parry, E. M. and Cox, B. S. (1970). The tolerance of aneuploidy in yeast. Genet. Res., Cumb. 16,333-340. Pedersen, M. B. (1986a). DNA sequence polymorphisms in the genus Saccharomyces. IV. Homoeologous chromosomes I11 of Saccharomyces bayanus, S. carlsbergensis, and S. uvarum. Carlsberg Res. Commun. 51,185-202. Pedersen, M. B. (1986b). DNA sequence polymorphisms in the genus Saccharomyces. 111. Restriction endonuclease fragment patterns of chromosomal regions in brewing and other yeast strains. Carlsberg Res. Commun. 51,163-183. Petersen, J. G. L., Nilsson-Tillgren, T., Kielland-Brandt, M. C., Gjermansen, C. and Holmberg, S. (1987). Structural heterozygosis at genes f L V2 and fLVS in Saccharomyces carlsbergensis. Curr. Genet. 12, 167-1 74. Rank, G . H., Casey, G. and Xiao, W. (1988). Gene transfer in industrial Saccharomyces yeasts. Food Biotech. 2, 1-41. Riley, M. I. and Manney, T. R. (1978). Tetraploid strains of Saccharomyces cerevisiae that are trisomic for chromosome 111. Genetics 89,667-684.

A. T. BAKALINSKY A N D R. SNOW

Sakai, K. and Takahashi, T. (1972). Estimation of ploidies in brewery yeasts. Bull. Brew. Sci. 18,29-36. Shaffer, B., Brearley, I., Littlewood, R. and Fink, G. R. (1971). A stable aneuploid of Saccharomyces cerevisiae. Genetics 67,483-495. Sherman, F., Fink, G. R. and Hicks, J. (1986). Laboratory Course Manual for Methods in Yeast Genetics. Cold Spring Harbor Laboratory, New York. Snow, R. (1967). Mutants of yeast sensitive to ultraviolet light. J. Bact. 94,571-575. Snow, R. (1979). Toward genetic improvement of wine yeast. Am. J . Vit. Enol. 30,33-37. Snow, R. (1983). Genetic improvement of wine yeast. In Spencer, J. F. T., Spencer, D. M. and Smith, A. R. W. (Eds), Yeast Genetics, Fundamental and Applied Aspects. Springer-Verlag, New York, NY, pp. 439459. Spencer, J. F. T. and Spencer, D. M. (1980). The use of mitochondria1 mutants in the isolation of hybrids involving industrial yeast strains. Molec. Gen. Genet. 177,355-358. Spencer, J. F. T. and Spencer, D. M. (1983). Genetic improvement of industrial yeasts. Ann. Rev. Micro. 37, 121-142. Subden, R . E. (1987). Current developments in wine yeasts. CRC Crit. Rev. Biotech. 5,49-65. Takahashi, T. (1978). Genetic analysis of a German wine yeast. Bull. Brew. Sci. 24,39-47. Talbot, N. J., Rawlins, D. and Coddington, A. (1988). A rapid method for ploidy determination in fungal cells. Curr. Genet. 14,Sl-52. Thornton, R. J. and Eschenbruch, R. (1976). Homothallism in wine yeasts. Antonie van Leeuwenhoek 42,503-509.