JOURNAL OF BACTROLOGY, Oct. 1973, p. 323-330Copyright i 1973 American Society for Microbiology

Vol. 116, No. 1Printed in U.S.A.

Kinetics of the Mating-Specific Aggregation inSaccharomyces cerevisiae

D. A. CAMPBELL'Department of Biology, University of Chicago, Chicago, Illinois 60637

Received for publication 15 June 1973

Procedures are described for obtaining efficient mating (zygote formation) inthe heterothallic yeast Saccharomyces cerevisiae. The methods separate opera-tionally the initial, mating-specific aggregation from subsequent steps inconjugation. This cell-cell interaction has been characterized. The data supportthe conclusions that (i) aggregation in liquid suspension is a random collisionprocess, and (ii) only non-budded cells participate in aggregation. A mathemati-cal model for the kinetics of aggregation in liquid suspension has been developedwhich is in good agreement with the experimental data.

Under appropriate conditions, vegetativehaploid cells of the heterothallic yeast Sac-charomvces cerevisiae fuse cytologically to formzygotes, which may in turn give rise to stablediploid clones. This conjugational process is anexample of cellular differentiation at both ge-netic and physiological levels. Conjugation iscontrolled, in part, by a single genetic locus thatdetermines haploid mating type. In geneticcrosses the two mating types, designated a anda, segregate as alleles of the mating type locus.An intimate cell-cell interaction must neces-

sarily precede cell fusion (conjugation). Thisinteraction may be mating specific, but needonly be of sufficient strength to maintain cells ofopposite mating type in direct contact until cellwall fusion is initiated or effected, or both. Thisreport describes a cell-cell interaction (termedaggregation) that fulfills these properties, anddescribes methods by which aggregation may bedetected. The principal result of this work is akinetic description of aggregation of cells inliquid suspension under conditions where subse-quent steps in zygote formation have beenprevented or delayed.

MATERIALS AND METHODSMedia. Ingredients are given in amounts per liter of

demineralized water. Media were solidified with 15 gof agar (Difco) per liter. YPD consisted of D-glucose,20 g; peptone (Difco), 20 g; yeast extract (Difco), 10 g.Minimal medium consisted of: D-glucose, 20 g; yeastnitrogen base without amino acids (Difco), 6.7 g.Synthetic complete medium (SC) consisted of mini-mal medium to which are added the following supple-

'Present address: Department of Genetics, University ofCalifornia, Berkeley, Calif. 94720.

ments as a filter-sterilized aqueous solution (.ug/mfinal concentration): adenine-SO4, 10; L-argininehydrochloride, 50; L-histidine-hydrochloride, 20; Lisoleucine, 50; L-leucine, 50; L-lysine-hydrochloride,50; L-methionine, 20; L-phenylalanine, 50; L-threo-nine, 300; L-tryptophan, 50; L-tyrosine, 50; uracil, 20.Sporulation medium consisted of: potassium acetate,20 g; D-glucose, 1 g; yeast extract (Difco), 2.5 g;supplemented with 1.5 times the amounts of nutriliteslisted for SC medium. Sporulation medium wasadjusted to pH 7 by addition ofKOH before autoclav-ing.

Yeast strains. Genotypes of the principal haploidmutant strains of Saccharomyces cerevisiae employedin this work are for S142, a; leul, trp5, metl3, tyr3,lys5, ade5, 7; and for Z133.9d, a; adel; ade2. Twoadditional haploid strains were employed in tests ofthe mating specificity of aggregation. These areA364A (a; gall, lys2, tyrl, his7; adel; ade2; ural) andM140 (a; ade3). For these tests the "crosses" were (i)A364A x S142 (a x a) and (ii) Z113.9d x M140 (a x a) .In both cases the haploid strain listed first was theminority parent. Genetic map locations of markerslisted for the four strains are given by Hawthorne andMortimer (5). The strains employed in this work werechosen for reasons other than their ability to aggregateefficiently -specifically, for their non-clumpinessduring normal growth in YPD medium and for theirknown genetic constitution with respect to nutritionalrequirements. The strains aggregated efficiently inmating mixtures in all pairwise (a x a) combinations.None of the strains were observed to agglutinate, i.e.,to give a flocculent, precipitating clumping reactionwhen cells of opposite mating type were mixed to-gether (2, 8, 9).

Preparation of cells. Haploid cells were streakedfrom stock cultures onto YPD medium and incubatedfor 2 days at 30 C. A single-colony suspension wasthen inoculated at about 105 cells/ml into YPD brothsupplemented with the same concentrations of nutri-lites listed above for SC medium. In addition, tetracy-

323

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

CAMPBELL

cline (Sumycin Oral Suspension, Squibb) was addedto the broth to a final concentration of 0.8 gliters/mlto minimize bacterial contamination. The broth sus-pension of cells was incubated in a flask on a rotaryshaker for 24 h at 30 C, at which time the cells were inthe late exponential-early stationary phase of growth.Cells were then harvested, washed by three cycles ofcentrifugation and resuspension in water, with finalresuspension in 0.15 M NaCl. Cell suspensions weremaintained in ice until the experiment was begun.The proportions of budded cells in the suspensionswere determined by hemocytometer counts.

Aggregation procedures: (i) Sedimented aggre-gation method. Equal volumes (0.5 ml) of strainZ113.9d in 0.15 M NaCl at about 2 x 106 cells/ml, andstrain S142 in 0.15 M NaCl at about 2 x 107 cells/ml,were mixed in small tubes; the cells were immediatelysedimented by brief (5 min) low-speed centrifugationand incubated at 30 C. Zero time was arbitrarilyestablished at the end of the centrifugation period. Atintervals thereafter the mating mixtures were resus-pended gently, diluted, and plated on SC medium.

In measurements of average pairing ratios, suspen-sion of cells of the two mating types at about 2 x 107cells/ml in 0.15 M NaCl were mixed in small tubesand sedimented by brief low-speed centrifugation.The relative amounts of the cell suspensions in themixtures were adjusted to give different input ratiosin a constant volume of 1.0 ml and at a constant total,concentration of about 2 x 107 cells/ml. In all casesstrain Z113.9d was the minority parent and strainS142 was the majority parent. The sedimented cellswere incubated for 90 min at 30 C. The several matingmixtures were then resuspended gently, diluted, andplated on SC medium.

(ii) Liquid aggregation method. To monitor ag-gregation in liquid suspension, suspensions of cells ofthe two mating types at about 2 x 107 cells/ml in 0.15M NaCl were mixed in a ratio of 8.0: 1 (S142: Z113.9d)in a total volume of 20 ml in a 50-ml flask. The flaskwas placed immediately on a rotary shaker main-tained at 30 C and approximately 100 rpm. Zero timewas established at the time of mixing. At periodicintervals thereafter, 0.1-ml samples were taken fromthe flask, diluted gently, and plated on SC medium.

In all experiments precautions were taken to pipetand dilute the resuspended aggregation mixturesslowly and gently to minimize shear forces. Duplicateplating assays performed in parallel were quite re-producible with respect to the recovery of zygoticcolonies, suggesting that the handling employed wasbelow the threshold for aggregate destruction (see be-low). All platings were by the melted, soft agar (0.7%)overlay method, and all plates were incubated at 30 C.

Sporulation and ascus dissection. Diploid cellswere streaked from stock cultures on YPD mediumand incubated for 2 days at 30C. Single colonies werethen spotted on sporulation medium and incubated afurther 5 days. Asci were dissected by micromanipu-lation on the surface of SC plates by standardmethods (6).

RESULTSThis report describes the results of experi-

ments designed to elaborate the initial proc-

esses in mating in the heterothallic yeast S.cerevisiae. The first step is a mating-specificaggregation of cells of opposite mating type.The term aggregation is used advisedly, for it isimportant to distinguish this cell-cell interac-tion as a necessary prelude to conjugation fromthe dramatic flocculent precipitation of yeastcells of opposite mating type (agglutination)observed in other systems (2). In Sac-charomvces, by contrast, agglutination is weakor delayed (8, 9). The methods described in thisreport, therefore, necessarily embody precau-tions designed to minimize shearing forces thatcould obscure detection of the mating-specificaggregation.

Mating-specific aggregation. Mating mix-tures prepared as described in Materials andMethods and plated on SC medium produce,after 4 to 6 days of incubation, three types ofcolonies (Fig. 1). The small (1-mm diameter)colonies are morphologically and geneticallyidentical to those of the two haploid parents.The large (3-mm diameter) colonies are zygoticcolonies. That these colonies consist wholly ofdiploid cells is supported by the following re-sult. After 48 h of incubation, zygotic coloniesare prominently visible on the plates, whereashaploid colonies are still quite small. Seventylarge 48-h zygotic colonies were picked, resus-pended, and replated on SC medium. Since thegenetic markers carried by the two haploidparent strains complement each other, the re-sultant diploid should be phenotypically wildtype. These secondary colonies were thus testedphenotypically for nutritional requirements byreplica-plating to minimal medium. Among19,075 secondary diploid colonies examined,none contained nutritional requirements char-acteristic of those carried by the haploid par-ents. It is tentatively estimated that the propor-

FIG. 1. Appearance of colonies produced by platingan aggregated mating mixture on SC. The two typesof small colonies (light: S142; dark: Z113.9d) derivefrom unmated haploid cells. The large colonies arezygotic colonies. The scale bar represents 1 cm.

324 J. BACTERIOL.

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

YEAST AGGREGATION KINETICS

tion of haploid cells per zygotic colony is lessthan 0.3%.The cells in zygotic colonies were additionally

judged to be diploid by their ability to sporulateand by the pattern of segregation in dissectedtetrads of genetic markers present in the hap-loid parents. No cells of ploidy greater thandiploid were detected among 77 primary isolatessubjected to segregational analysis. The propor-tion of cells of ploidies greater than diploid perzygotic colony is tentatively estimated to be lessthan 2%.Enumeration of the three types of colonies, as

well as the unequal parental input employed inthese experiments, permits an estimate to bemade of the overall efficiency of aggregationand subsequent zygote formation. The matingefficiency is calculated as [zygotic colonies/(zy-gotic colonies + minority parent colonies)]. Im-plicit in this treatment is the assumption thateach entity that ultimately forms a zygoticcolony contained no more than one minorityparent cell. This assumption will be examinedcritically and justified in a later section of theseResults.

Calculated data of this type are presented inFig. 2 for the two experimental conditions. Inthe upper curve, data from sedimented mix-tures (method i) are plotted as a function of thetime after mixing at which the mixture wasresuspended and assayed. The efficiency ofmating is nearly 80%, and there is no change inthis value with time of incubation. The lowercurve shows the result obtained when cells aremixed in liquid suspension (method ii). There isan initial rapid increase in the efficiency ofmating, followed by a slower, more gradualincrease at later times. The absolute concentra-tion of zygotic colony-forming entities roughlyfollows these same kinetics (see Fig. 4).Even after extended incubation times (24 h)

microscope inspection of resuspended mixtures(method i) revealed no classic zygotic figures.This finding is, perhaps, not surprising giventhe absence of exogenous nutrients. It leads tothe contention that the entities responsible forthe formation of zygotic colonies on the platesare not zygotes per se, but, rather, boundaggregates of cells of opposite mating type. Thiscontention is reinforced by two additional ex-perimental observations. First, sedimentedmixtures were subjected to strong agitation(15-s Vortex mixing at maximal speed setting)before being assayed. Zygotic colony-formingentities were reduced to 11.7% (mean of fivedeterminations) of the value obtained by thestandard gentle handling. This value did notchange with time of incubation prior to resus-pension of the sedimented mixtures. Second,

Cl)w4w 1.0cr.

zW 0.1

4

0r

o

z 0.001

0z0

U. 0.0010 1 2 3

HOURS AFTER MIXING4

FIG. 2. Kinetics of appearance of zygotic colonies(aggregation) under two experimental conditions. Themating efficiency, calculated by the formulation givenin the text, is plotted as a function of time aftermixing. The upper curve (A) shows results obtainedfrom sedimentation of the mating mixture by centrif-ugation (method i). The lower curve (0) shows resultsobtained from aggregation in liquid suspension(method ii).

sedimented mixtures were formed by mixingcells of two different strains possessing the samemating-type allele (i.e., a x a; a x a). Aggre-gation was monitored by taking advantage of thered pigment accumulation in ade2 mutantstrains (7). The plated mixtures were thusexamined for sectored (red-white) colonies asevidence of non-mating-specific aggregation.The frequency of sectored colonies under theseconditions was less than 2% of the frequency ofzygotic colonies found in the comparable mat-ing-specific (a x a) combinations.

In summary, sedimentation of cells in amating mixture into a pellet promotes a mat-ing-specific aggregation of cells of opposite mat-ing type. This aggregation is unique to a x amixtures and occurs as well in liquid suspen-sion, though at a reduced rate. Zygote forma-tion must rarely occur at this stage, since nearly90% of potential zygotic colony-forming entitiescan be destroyed by vigorous agitation of themating mixture. Aggregation is, nevertheless,necessary for subsequent formation on theplates of zygotes and zygotic colonies, sinceplating of mixtures of a and a cells without theaggregation step results in no zygotic colonyformation. In effect, the methods employed hereseparate operationally the mating-specific ag-

VOL. 116, 1963 325

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

CAMPBELL

gregation of cells of opposite mating type fromsubsequent steps involved in conjugation andzygote formation.

Restriction of aggregation to non-buddedcells. The efficiency of mating, as calculatedhere, was observed always to be less than 100%.Efficiencies of mating for several experimentsperformed by method (i) are listed in Table 1.Also listed for each experiment are the parentalinput ratio and the fraction of budded cells inthe minority cell population. The sum of thebudded cell fraction and the mating efficiencyapproximates unity in all experiments, and themean of all 12 summations is very close to unity.Though this datum is circumstantial, it sug-gests that budded cells may not participate inmating-specific aggregation. This suggestionhas also been reached independently by Sena etal. (10). The observation also suggests that theextent of aggregation observed may be close tothe maximum possible under the conditionsemployed. Thus, the fraction of budded cells inthe minority cell population may place a limiton the overall efficiency of mating.The data in Table 1 represent crosses in

which a (strain Z113.9d) was the minorityparent. Less extensive data from crosses inwhich a (strain S142) was the minority parent(not included in Table 1) suggest that the samesummation holds in this configuration as well.

TABLE 1. Evidence that aggregation is limited tonon-budded cellsa

Input Fraction MatingExpt ratio budded cells efficiency (A + B)

(A) (B)

1 6.0 0.16 0.78 0.942 9.4 0.17 0.82 0.993 9.3 0.17 0.76 0.934 6.9 0.19 0.72 0.915 9.2 0.19 0.74 0.936 9.5 0.21 0.74 0.957 8.0 0.22 0.77 0.998 11.4 0.25 0.65 0.909 6.5 0.26 0.73 0.9910 7.2 0.27 0.75 1.0211 10.9 0.30 0.89 1.1912 11.9 0.33 0.71 1.04

a For each independent experiment, performed bymethod (i), the parental input ratio (majority to mi-nority), the fraction of budded cells in the minorityhaploid cell population (A), and the efficiency ofmating as calculated by the formulation given in thetext (B), are listed. In the last column the sum of thebudded cell fraction and the mating efficiency (A + B)is given. The mean of all 12 experimental summationsis: 0.98 + 0.078 (standard deviation).

The suggestion that budded cells do notaggregate may be understood in the followingterms. Haploid cells of both mating types se-crete diffusible mating factors which arrest cellsof the opposite mating type in the unbuddedstage of the cell cycle (1, 3, 4, 12). Activelygrowing mixtures of a and a cells becomeentrained with respect to the cell cycle, andnon-budded cells accumulate (4). If only at thisstage are cells capable of conjugation, andhence aggregation, then in an asynchronouspopulation containing cells representing allstages of the cell cycle the process of aggregationitself will in effect select out the appropriatenon-budded subpopulation. The ability of cellsto aggregate to cells of the opposite mating typeappears thus to be a temporal function of thecell cycle (4, 10).Composition of aggregates. It was assumed

in calculations of the efficiency of mating (Fig. 2and Table 1) that each aggregate contained nomore than a single minority haploid cell. Thisassumption is directly testable by a quantita-tive consideration of the average composition ofaggregates. Since both aggregates and free (i.e.,non-aggregated) cells make colonies that maybe enumerated separately, experiments wereperformed by method (i) in which the parentalinput ratio was varied over a wide range. Themean number of cells per aggregate, both ma-jority and minority, was calculated from thedifferences observed between the concentra-tions of cells originally placed in the aggregationmixture and the concentrations of free haploidcells recovered after aggregation. The assump-tion in this treatment was that any excess ofhaploid input over output would be representedby aggregates containing more than a single cellof each mating type. By this method a calcu-lated mean pairing ratio was derived (Fig. 3).The ratio of majority to minority haploid cellsper aggregate (the pairing ratio) is plotted as afunction of the parental input ratio. The pairingratio increases as a function of the input ratiobut lags considerably behind the latter. Thisfunctional behavior may be understood if it isrecalled that a finite limit must exist for thenumber of cells that can aggregate to a singlecell. Geometrical considerations indicate thatthis limit is approximately 12 cells per cell. Thedata are not intended to suggest that all aggre-gates at any input ratio are composed of exactlythe same numbers of majority and minorityhaploid cells. Rather, the values shown mustrepresent the mean of a distribution of aggre-gates of different sizes.

Since the data in Fig. 3 represent ratios of

J. BACTERIOL.326

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

YEAST AGGREGATION KINETICS

04

(93 _

z2i

0 5 10 15 20PARENTAL INPUT RATIO

FIG. 3. Characterization of aggregate composition.The mean ratio of majority to minority cells peraggregate (the pairing ratio) is plotted as a function ofthe haploid parental input ratio. Each point is theresult of an experiment performed by method (i).

majority to minority cells, they tend to obscurethe actual numbers of cells per aggregate. Atinput ratios near unity, the average number ofminority cells per aggregate is in the range 1.4 to2.6. For input ratios greater than about 5,

however, this number is not significantly differ-ent from 1.0, and the pairing ratios shown inFig. 3 denote directly the average number ofmajority cells per aggregate. Thus the assump-tion made earlier that each aggregate consists ofa single minority cell and one or more majoritycells is justified for input ratios in excess of 5

(majority:minority). This justification appliesstrictly only to aggregation by sedimentation(method i). As will be shown in the followingsections, unrestricted aggregation in liquid sus-

pension (method ii) can result in aggregatescontaining more than a single minority haploidcell.

Qualitative description of aggregation ki-netics. The liquid suspension aggregation ki-netics (method (ii) presented in Fig. 2 are

complex, but they do yield to simple mathemat-ical treatment. In this section, factors thatcould influence aggregation under the condi-tions employed in this work are considered froma qualitative point of view. In the followingsection, a quantitative model is developed fromthese considerations and tested against thedata.

It is intuitively apparent that each aggregatethat is subsequently resolved as a zygotic colonymust have consisted of at least one haploid cellof each mating type. However, mean pairingratios in excess of one majority cell per aggre-gate have been shown to occur under theconditions of unequal parental input employed(method i) (Fig. 3). This finding may be reason-ably extended to include aggregation in liquidsuspension (method ii) as well, from which onemay infer that larger aggregates may also arisethere, especially at late times. If haploid cellscan aggregate to aggregates previously formed,rather than contributing to new aggregates, therate of aggregate formation should eventuallydecline. This would be especially true if mino-rity haploid cells could aggregate to aggregatesalready formed. This reasoning is not intendedto contradict the conclusion made earlier that insedimented mixtures (method i) aggregatesconsist of a single minority cell and one or moremajority cells. In that case the rapid packing ofcells by centrifugation in effect fixes the pairingratio for any given input ratio. In the liquidsuspension conditions considered here, cell-cellcollisions are unrestricted encounters, and it isreasonable to assume that both majority andminority cells contribute to many-celled aggre-gates. It may be noted in support of thisassumption that the secondary rise in the effi-ciency of mating after 0.5 h as calculated by theformula presented earlier (Fig. 2) is due, in part,to a loss of free minority haploid cells and notwholly to an increase in the absolute concentra-tion of aggregates (compare with Fig. 4).A substantial fraction of minority haploid

cells participates in aggregation, and this frac-tion increases as a function of time. If theinstantaneous rate of aggregate formation isdependent upon the concentrations of free hap-loid cells remaining at any time, the rate cannotremain constant, but must decline at latertimes.

Qualitatively, then, aggregation in liquid sus-pension is envisaged as involving at least twofactors which act in concert to diminish the rateof aggregation as a function of time. These are(i) the aggregation of free cells to aggregatesalready formed, and (ii) a diminution in theconcentrations of free haploid cells with time.

Quantitative description of aggregation ki-netics. If aggregation can be thought of as abimolecular rate reaction, the rate of aggregateformation (dA/dt) may be written as

dA (4(MI) (M2) (1)dt

VOL. 116, 1963 327

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

CAMPBELL

where M1 and M2 are the free concentrations ofthe two types of haploid cells. This formulationis correct provided M1 and M2 do not changeappreciably in the course of the reaction. Whilethis is approximately true for majority haploidcells, minority haploid cells are consumed in thereaction to a considerable degree. Hence theconcentration of free minority cells is itself afunction of time. With these considerationsequation 1 may be simplified to

dA=kM (2)

dtwhere k, is a constant with units per hour, andM is the concentration of free minority cells atany time t. M may be scored directly in theexperiments: to a first approximation the datafit a linear regression of the form M = a(t) +MO, where M. is the concentration of freeminority cells at time 0, and a is a constant.Equation 2 fails to take account of aggrega-

tion of free minority cells to aggregates alreadyformed. It is not possible from the informationat hand to derive directly the algebraic formthat such second order aggregation must take,but it is reasonable to assume that it is propor-tional to, and limited by, the aggregate concen-tration and the free minority cell concentrationat any time t. Applying these corrections toequation 2 gives

dA k M - kA M (3)dt

where k2 is a constant with units of cubiccentimeters per hour. The expression for aggre-gation of free minority cells to aggregates al-ready formed (k2 A M) embodies the approxi-mation that k2 is independent of the distributedsize of aggregates. k2 must be a function of thecollision cross section of aggregates, a parame-ter that itself must be proportional to aggregatesize, and a function of time. A more preciseexpression of the form Ei kiAM would bepossible only if the functional form of thedistribution of aggregate sizes were known.After substitution and rearrangement, the solu-tion of equation 3 is

fot (k1 - k2A)-IdA = fot (at + MO) dt (4)and resolution of the definite integrals gives

In k2At-k = - (M + M0) (5)k2AO - k1 2

where At and AO are the aggregate concentra-tions at times t and 0. Additional rearrange-ment of equation 5 yields a final equation of theform

At = AO e-X + k I (I - e-X)k 2

(6)

where x = [(k2t)/2][M + Mo]. Since it is notpossible to estimate k2 by independent means,equation 6 generates a family of theoreticalcurves for different values of k2. At early times,when M a Mo and t is small, equation 6simplifies to give A, as a linear function of time(At = Ao + k1 MO t). This expectation of themodel is borne out: the initial aggregationkinetics are in fact linear to t < 0.5 h, thuspermitting ki to be estimated directly from thedata.

In applying equation 6 to the actual data, Mand MO were estimated from the least-squareslinear regression of the free minority cell con-centrations, and k1 and AO were estimated fromthe least squares regression of the initial linearportion of the aggregation kinetics. In Fig. 4 theexperimental raw data are plotted (aggregateconcentration versus time), and a selection oftheoretical curves, calculated by equation 6 forseveral values of k2, is drawn for comparison.The theoretical curves are judged to be in goodagreement with the experimental data.

In conclusion, a model for the aggregation ofhaploid cells in liquid suspension has beendeveloped in which (i) aggregation is envisagedas a bimolecular rate reaction dependent on theconcentrations of free haploid cells, and (ii)competition for free haploid cells between ag-gregates already formed and new aggregatescontributes in large measure to the decline inthe rate of aggregation at late times. The closeagreement between the experimental data andthe direct quantitative expression of these pa-rameters provides strong support for the valid-ity of the model.

DISCUSSIONThis work establishes methods by which hap-

loid cells of the yeast S. cerevisiae may bemated in mass culture. The initial cell-cellinteraction in mating (aggregation) has beenstudied kinetically under non-nutrient condi-tions. The procedures separate operationallythe initial aggregation from subsequent steps inconjugation. The experimental results may besummarized as follows. (i) Aggregation is mat-ing specific; in the present system only cells ofopposite mating type aggregate. (ii) Aggrega-tion appears to be limited to cells in theunbudded stage of the cell cycle. (iii) Aggrega-tion appears to be the consequence of randomcellular collisions since the average compositionof aggregates may be altered by experimental

J. BACTERIOL.328

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

YEAST AGGREGATION KINETICS

10b The results reported here complement obser-_- , _- - - vations made on the action of two classes of___----8-- bt mating-specific factors secreted by vegetative

-X*,-<i------ ---.; - haploid cells (1, 3, 4, 11, 12). These factorsinduce morphological expansion in cells of theopposite mating type, and in the case of oneclass, result in an accumulation of cells at the

t 105 unbudded stage of the cell cycle (1, 4). TheX _.t1 _ apparent restriction of effective aggregation toeI) the non-budded cells originally present in theW mating mixture suggests that in the absence of

exogenous growth requirements, the diffusibleWr factors are not produced, or else that budding

cells of the opposite mating type are metaboli-cally or temporally prevented from responding

104 l to these factors. The data do not distinguishbetween these two alternatives. They do sug-

_ gest, however, that exposure of cells to theI diffusible mating factors may not be a necessaryO 2 3 prerequisite for aggregation.

The distinction drawn in this work betweenHOURS AFTER MIXING aggregation and agglutination merits further

FIG. 4. Kinetics of aggregation in liquid suspen- consideration. Sexual agglutination is weak insion: a comparison of experimental results with the Saccharomyces (8) by comparison with thatkinetic model developed in this work. Points are raw observed in the unrelated yeast Hansenuladata (aggregates per milliliter) plotted as a function of wingei (2). In Saccharomyces, moreover, exoge-time after mixing. Error bars represent a 1 standard nous nutrients required for vegetative growthdeviation. The theoretical lines were calculated by are necessary for sexual agglutination (10).equation 6 (see text) for three values of k2, In all Aggregation, by contrast, occurs and is de-calculations k, was set at the experimentally derived tected in the total absence of all exogenousvalue (0.382/h). Curve a: k2 = 0.45 x 10-6 cm8/h;Curve b: k2 = 0.65 x 10- cm3/h; Curve c: k2 = 0.85 growth requirements. Recently, two classes ofx 10-6cm3/h. Saccharomyces strains have been distinguished

on the basis of agglutinability (9), those whichagglutinate only after a lag period, and those

manipulation of the haploid input ratio. (iv) A which begin agglutinating immediately uponkinetic model for the aggregation of haploid mixing. A comparative study of aggregationcells in liquid suspension, incorporating the employing these two classes of haploid strainsfeatures of random collision and mating speci- may help to clarify the relationship(s) betweenficity, has been developed which closely re- aggregation and agglutination.produces the experimental data. The kinetic model for aggregation in liquidOne inference from this work is that vegeta- suspension developed in this work embodies a

tive haploid cells are "sticky" with respect to number of explicitly stated assumptions andcells of the opposite mating type. The mating- simplifying approximations. The demonstratedspecific binding described here is a weak one, predictive success of the model, however, doesand special handling precautions must be taken not imply that all of these are necessarilyto insure its detection. Aggregation is distinct correct. A critical test of the model's validityfrom the precipitating agglutination reaction would lie in a study of the kinetics of aggrega-observed in other systems (2, 8, 9). That aggre- tion under a variety of ambient conditions, or ingation appears to be limited to the unbudded its predictive ability when applied to othercells in the vegetative population suggests, cases of cell-cell interaction.moreover, that this surface stickiness is a tem- In summary, experiments have been pre-poral feature of the cell cycle. The finding that sented in which the initial cellular aggregationmany-celled aggregates arise under both experi- preceding zygote formation in yeast has beenmental conditions (but especially method i) characterized. Aggregation occurs efficientlymay tentatively be interpreted to mean that under non-nutrient conditions, and appears tomating-specific binding is a general feature of be restricted to the non-budded cells present inthe cell surface. the vegetative cell population. An exact de-

VOL. 116, 1963 329- e-

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

330 CAMPBELL

scriptive model has been developed whichclosely reproduces the kinetics of aggregation inliquid suspension.

ACKNOWLEDGMENTSI thank R. E. Esposito and M. S. Esposito for advice and

helpful discussions. I am especially indebted to L. B. Rowe forencouragement and for valuable mathematical suggestionsduring development of the model.

This work was supported by National Science Foundationgrant GB-27688 and Public Health Service grant1-TI -00174-05

LITERATURE CITED1. Biicking-Throm, E., W. Duntze, L. H. Hartwell, and T.

R. Manney. 1973. Reversible arrest of haploid yeastcells at the initiation of DNA synthesis by a diffusiblesex factor. Exp. Cell Res. 76:99-110.

2. Crandall, M. A., and T. D. Brock. 1968. Molecular basisof mating in the yeast Hansenula wingei, Bacteriol.Rev. 32:139-163.

3. Duntze, W., V. MacKay, and T. R. Manney. 1970.Saccharomyces cerevisiae: A diffusible sex factor.Science 168:1472-1473.

4. Hartwell, L. H. 1973. Synchronization of haploid yeast

J. BACTERIOL.

cell cycles, a prelude to conjugation. Exp. Cell Res.76:111-117.

5. Hawthorne, D. C., and R. K. Mortimer. 1968. Geneticmapping of nonsense suppressors in yeast. Genetics60:735-742.

6. Johnston, J. R., and R. K. Mortimer. 1959. Use of snaildigestive juice in isolation of yeast spore tetrads. J.Bacteriol. 78:272.

7. Roman, H. 1956. A system selective for mutations affect-ing the synthesis of adenine in yeast. C. R. Lab.Carlsberg Ser. Physiol. 26:299-314.

8. Sakai, K., and N. Yanagishima. 1971. Mating reaction inSaccharom,vces cerevisiae I. Cell agglutination relatedto mating. Arch. Mikrobiol. 75:260-265.

9. Sakai, K., and N. Yanagishima. 1972. Mating reaction inSaccharomvces cerevisiae. II. Hormonal regulation ofagglutinability of a type cells. Arch. Mikrobiol.84:191-198.

10. Sena, E. P., D. N. Radin, and S. Fogel. 1973. Synchro-nous mating in yeast. Proc. Nat. Acad. Sci. U.S.A.70:1373-1377.

11. Throm, E., and W. Duntze. 1970. Mating-type-depend-ent inhibition of deoxyribonucleic acid synthesis inSaccharomvces cerevisiae. J. Bacteriol. 104:1388-1390.

12. Yanagishima, N. 1969. Sexual hormones in Sac-charomyces cerevisiae. Antonie van Leeuwenhoek J.Microbiol. Serol. 35 Suppl. Yeast symposium1969:C9-C10.

on May 25, 2018 by guest

http://jb.asm.org/

Dow

nloaded from