Population Analysis of the Deinduction Kinetics of Galactose Long-Term Adaptation Mutants of Yeast

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 7, No. 3, pp , March 1973 Population Analysis of the Deinduction Kinetics of Galactose Long-Term Adaptation Mutants of Yeast (ethidium bromide/heterogeneous induction units/endogenous inducer) SHINJI TSUYUMU AND BRUCE G. ADAMS Department of Microbiology, University of Hawaii, Honolulu, Hi Communicated by Herschel L. Roman, January 15, 1973 ABSTRACT By use of a selective galactose agar medium containing ethidiuni bromide, a population analysis of the deinduction kinetics of yeast galactose long-term adaptation mutants (gal 3) has been done. It was first determined that the gal 3 mutation is specific to the yeast galactose system and that induced cultures of gal 3 strains are capable of growth on galactose agar medium containing ethidium bromide, whereas noninduced cultures are not. Population analyses of induced gal 3 strains undergoing deinduction in the absence of galactose demonstrate that a minimum number of five induction units per cell are required for induction of the galactose system. It is concluded that: these induction units are actively synthesized only in the presence of inducer and are diluted out through cell division; they are stable under nongrowing conditions; they are heterogeneous in nature; at most two of the five minimum units are products of the gal 2 locus; and the other units may be three of one type, one of one type and two of another, or one each of three different types. That the induction of an enzyme system could be clonally distributed was first demonstrated with the galactose longterm adaptation mutants of yeast, which were originally isolated by Winge and Roberts (1) and extensively investigated by Spiegelman and coworkers (2-6). By using the maintenance effect, Novick and Weiner (7) and Cohn and Horibata (8) demonstrated the same phenomenon in the bacterial lactose system. In recent years population analyses have not been extensively used in the study of induction processes. Differential methods previously used to determine the percentage of induced cells in a population include measurement of the enzymatic activity of the descendants of a single cell under maintenance conditions (8) and colony differentiation on an indicator medium such as eosin-methylene blue plates (5). These methods are incapable of precisely determining low percentages of induced cells; if either induced or noninduced cells comprise a progressively smaller percentage of the population, it becomes impossible to determine that percentage. The galactose long-term adaptation mutation (gal 3*) in Saccharomyces cerevisiae is characterized by a long lag period in galactose medium after which growth in galactose commences at a normal rate (9). Neither the function nor the nature of the Gal 3 gene product is known. However, it is known that the combination of the gal 3 mutation and the * Genetic nomenclature. The symbols used were those adopted at the Osaka Yeast Genetics Conference (Microbial Genetics Bulletin no. 31, 1969). 919 respiratory-deficient petite (p -) mutation acts as a conditional lethal mutation with regard to growth on galactose (9). In this communication, we report the observation that gal 3, p+ strains adapted to galactose continue to grow on galactose when converted to p - strains by treatment with ethidium bromide, which specifically induces the respiratory petite mutation without attendant lethality (1). This property of induced gal 3 strains, that of being capable of growth on galactose in the petite condition while noninduced strains are not, allows one to selectively grow only induced gal 3 cells on galactose plates containing ethidium bromide. This circumstance allows us to precisely determine the number of induced cells in a population consisting predominantly of noninduced cells. Using this selective medium, we have studied the kinetics of the deinduction process of induced gal 3 strains of yeast in the absence of inducer. MATERIALS AND METHODS Yeast Strains. The haploid Saccharomyces cerevisiae strains 18-3C (a, gal+, ura 1, trp 1), 476-5D (a, gal+, his 1), 17-1C (a, gal 3, ade 2, his 1, trp 1, met), 17-1D (a, gal 3, ura 1, his 1, thr 1, trp 1, met), and 381-1C (a, gal 2, his 1, ade 6, trp 1) were kindly provided by Howard C. Douglas. Strains X18-D and D17 are the diploid strains resulting from the crosses 18-3C x 476-5D and 17-IC x 17-1D, respectively, and were isolated by prototrophic selection. The haploid double mutant strain 23-B (gal 2, gal 3) was obtained by the complimentation test for each dissected tetrad obtained from the cross 17-1D x 381-1C (11). It is important to note that ho cell-tocell adhesion or clumping whatever was observed in any growth phase of the diploid strain D17 in the different growth media used. Cultivation. Cells were grown aerobically at 3 on a rotary shaker in Wickerham's complete defined medium (abbreviated WC) (12) or in YET broth medium (1 g of tryptone and 5 g of yeast extract per liter of medium). Solid media contained 2 g/liter of agar. Carbon sources were sterilized separately and added to the medium at a concentration of 2 g/liter unless specified otherwise just before use. The selective agar-plate medium was made by addition of ethidium bromide to a concentration of io ug/ml into YET agar medium containing either 1 g/liter of galactose (abbreviated EB+1 galactose) or 1 g/liter of galactose (abbreviated EB+1 galactose). Cell concentrations in liquid culture were monitored with a Coulter model Zb particle counter with a 1-,um aperture as described (13).

2 92 Genetics: Tsuyumu and Adams TABLE 1. Nonpleiotropic nature of gal 3 strains Cultures adapted in YET medium + Galac- Plating medium Glucose Lactate Melibiose tose YET+ glucose EB+ glucose YET+ maltose EB+ maltose YET+ melibiose EB+ melibiose YET+ galactose LA LA LA + EB+ galactose + Cultures of strains 17-1D, 17-1C, and D17 were each twice subcultured and grown out in medium glucose, containing lactate, melibiose, or galactose. The cultures were diluted to contain 13 cells per ml, and.1 ml was spread onto YET and EB plates containing glucose, maltose, melibiose, or galactose. + = immediate growth; LA = long-term adaptation; = no growth. z C-)~~~~~~~~~~~~~~~~) PHASE I PHASEHI ' FIG. 1. Kinetics of retention of inducibility of gal+ and gal 3 strains after removal from galactose medium. Fully adapted cultures of strain X18-D (gal+) growing in YET+ galactose and strain D17 (gal 3) growing in YET+ galactose and in WC+ galactose broth media were harvested, washed four times with YET or WC medium minus a carbon source, and suspended in fresh medium at the desired cell concentration. The cultures were aerated at 3 for 5 min, glucose (2 g/liter) was added, and aeration was continued. At the indicated intervals samples were removed, diluted, and spread onto three plates each of YET+ glucose and EB+ 1 galactose agar. The plates were incubated at 3 and observed for colony growth. Representative EB+ 1 galactose plates were replicated onto YET+ glucose plates, and the resultant colonies were replicated onto fresh EB+ 1 galactose plates to screen for the presence of constitutive mutants in the population. Phase I and Phase II of the growth of strain D17 is explained in the text. Viable counts of strain X1OS-D on YET+ glucose (OH-O) and on EB+1 galactose (AE -A) of samples from YET+ glucose broth. Viable counts of strain D17 on YET+ glucose (A- ) and on EB+ 1 galactose (A--A) of samples from YET+ glucose broth, and on YET+ glucose (- *) and on EB+1 galactose (Eli-fl) from WC+ glucose broth. Proc. Nat. A cad. Sci. USA 7 (1973) Chemicals. Ethidium bromide (2,7-diamino-1-ethyl-9- phenyl-phenanthridium bromide), B grade, was purchased from Calbiochem, Los Angeles, Calif. "Baker Grade" a->galactose was purchased from J. T. Baker Chemical Co., Phillipsburg, N.J., and contained less than.1 mg/1 g of glucose as determined with the "Glucostat" glucose assay kit (Calbiochem). All other chemicals were of reagent grade obtained from commercial sources. RESULTS That the requirement for respiratory competency in both haploid and diploid gal 3 strains is not a general requirement for the induction of catabolic systems for fermentable carbon sources is shown by the data in Table 1. Cultures of gal 3 strains grown on glucose, lactate, melibiose, or galactose are able to grow when plated onto ethidium bromide YET agar plates with glucose, maltose, or melibiose as a carbon source. Such cultures, however, are not capable of growth on EB+ galactose plates unless they have been adapted to galactose before plating. Adapted gal 3 strains grew as well as the wildtype strains on EB+ galactose plates. The plate counts of gal 3 strains pregrown in galactose were the same on EB+ glucose, EB+ galactose, YET+ glucose, and YET+ galactose plates. These data demonstrate that a Gal 3 gene product is required for the initial induction process, but is not required after the cells have been induced for growth on galactose. Furthermore, this product is not a prerequisite for the induction of the catabolic systems for maltose or melibiose. The kinetics of the deinduction of the galactose system were studied after the transfer of strains D17 (gal 3) and X18-D (gal+) growing in YET+ galactose broth and WC+ galactose medium into YET+ glucose broth and WC+ glucose medium, respectively (Fig. 1). From the number of colonies growing on EB+1 galactose and YET+ glucose plates of samples of cell suspension removed from YET+ glucose and WC+ glucose cultures, the conversion from induced (positive) to noninduced (negative) cells is observed only in strain D17 as the number of colonies on the EB+1 galactose plates decreases (Phase I in Fig. 1) while the total colony counts on YET+ glucose increase exponentially. It should be pointed out that the observed kinetics in Phase I are a function of dilution by cell division and independent of the growth medium used; the deinduction in YET+ glucose and WC+ glucose (Fig. 1) or in YET+ lactate (data not shown) showed identical kinetics in a plot of log F against the number of generations of growth. We have also confirmed this observation by determining that the fraction of induced cells remains constant for at least 72 hr under nongrowing conditions (data not shown). The concept that an all or none response of an inducible system to inducer was due to the number of "induction units" per single cell was prevalent in the arguments of other investigators (5, 8). The positive cells were assumed to be cells that contained more than a certain minimum number of induction units. The basic assumption for evaluation of the minimum number of units required for induction and the average number of units possessed by a fully induced cell was that the units are stable, their disappearance only being effected by dilution through cell division. Using a differential method rather than a selective method, previous workers estimated the minimum number of units required for induction to be about one (5, 8). That the mini-

3 Proc. Nat. Acad. Sci. USA 7 (1973) Analysis of Deinduction Kinetics 921 TABLE 2. Requirement of a functional petmease for growth of galactose mutants on EB+1 galactose plates Growth Growth on on Adaptation EB+ 1 EB+ 1 Strain Genotype carbon source galactosegalactose 17-1C gal 3 Glucose 17-1C gal 3 Galactose D gal 2 Glucose D gal 2 Galactose + 23-B gal 2, gal 3 Glucose 23-B gal 2, gal 3 Galactose + Cultures were each twice subcultured and grown out in YET medium containing glucose or galactose. The cultures were diluted to contain 13 cells per ml, and.1 ml was spread onto EB+1 galactose and E13+1 galactose plates. = no growth; + = immediate growth. mum number of units must be greater than unity is obvious from the data in Fig. 1, because the absolute number of positive cells in the population declines after reaching a maximum value rather than remaining constant as would be expected if but one unit was sufficient for induction. Having established that the minimum number is greater than unity, the question must be asked whether or not the units are homogeneous as assumed by previous workers (5, 8) or whether they are heterogeneous. Cohn and Horibata (8) determined that the induction unit for the lactose system of Escherichia coli was the 3-galactoside permease. In order to determine if the induction units for the yeast galactose system are heterogeneous and if one or all of these units is the galactose-specific permease specified by the Gal 2 locus (14), we constructed the haploid double mutant strain 23-B (gal 2, gal 3). Table 2 shows that strain 23-B grows on EB+1 galactose plates when it is preinduced with galactose (1 g/liter) but exhibits no growth on EB+1 galactose plates whether preinduced or not, and also that single gal 2 mutants are incapable of growth on 1% galactose. Therefore, gal 3 strains of yeast plated onto EB+1 galactose agar medium require a functional Gal 2 gene product for growth. This characteristic makes it possible to analyze the deinduction kinetics of gal 3 strains by determining the fraction of induced cells in a population that is dependent on the permease and the fraction that is independent of the permease. The following relationship between these fractions can be derived on the assumption that the gene product of the Gal 2 locus is of one kind: co F = F'* E (xpk/k!) * exp(-xp) k =mp where F is the fraction of positive cells dependent on the permease, F' is the fraction of positive cells independent of the permease, xp is the average number of permease units at the nth generation, and mp is the minimum number of permease -units required to be a positive cell. The theoretical curves of the plot of log F/F', n, and log xp are shown in Fig. 2. The plate counts on EB+1 galactose/ YET+ glucose represent the fraction, F, accounting for the disappearance of the permease, and the plate counts on EB+1 galactose/yet+ glucose represent the fraction, F', FIG. 2. Theoretical curves of the plot of log F/F', n, and log xp and experimental plot of log F/F'. In order to visualize the relationships among F/F', xp, and 77, the third axis, xp, is drawn such that log xp = log F/F'. The theoretical curves for each value of mp were drawn by reading the theoretical values from the table of the cumulative terms of the Poisson distribution corresponding to representative values of xp in the plot of log xp against log F/F'. Experimental values of F/F' were determined by transferring an induced culture of strain D17 into YET+ glucose broth as described in Fig. 1. Samples were removed, diluted, and spread onto three plates each of YET+ glucose, EB+1 galactose, and EB+1 galactose. () Experimental values of F/F'. The value of 2 units was obtained by extrapolation of log xp from mp = 2. that is independent of the permease. The experimental data obtained in this manner are also plotted in Fig. 2. The experimental values of F/F' at each generation correspond to the theoretical curve of mp = 2. It may be concluded that two units of galactose permease and some number of other units are minimally required for inducibility. These two units, however, cannot definitely be identified as "permease" molecules per se. The data in Table 2 establish only that a Gal 2 gene product is required for growth on 1% galactose, but does not eliminate the possibility that an additional growth requirement is alleviated by increasing the galactose concentration. Furthermore, the molecular mechanism of galactose transport in yeast has not been clearly defined and the nature of the deleted unit in gal 2 mutants is thus unknown. The term, permease unit, therefore, must be considered as a provisional description of a functional unit that is replaceable by high galactose concentrations. It is pertinent to note here that if we can assume that increasing the galactose concentration serves no other significant function than to alleviate a transport deficiency, one can calculate from the intersection of the log xp axis with the log F/F' axis in Fig. 2 the average number of galactose permease units in a fully induced cell to be 2. The change in the fraction of positive cells due to the units other than the permease, measured under conditions of independence of the permease (determined by plating onto EB+1 galactose), can be used for an analysis of induction units other than the galactose permease. These other units could be either heterogeneous or homogeneous, or a combination of the two. In the first case the theoretical curves of F' for each value of m (number of different kinds of units) can be obtained from the following equation on the assumption

4 922 Genetics: Tsuyumu and Adams F 1 1' /-. ' FIG. 3. Theoretical curves and experimental plot of the fraction of positive cells (F') independent of the galactose permease. The experimental values of F' obtained with strain D17 plated onto Eb+ 1 galactose plates, as described in Fig. 2, are plotted (EJ) along with the theoretical curves for homogeneous units (- -), heterogeneous units (-), and intermediate units that the average number of each kind of unit is about the same: F' = [1-exp(-x)]m In the second case, the theoretical curve of F' for each number of units (M) can be obtained from the cumulative terms of the Poisson distribution: F' = xk/k! * exp(-x) k=m The theoretical curves for these two possibilities exhibit a close relationship with the experimental data obtained by plating onto EB+1 galactose (Fig. 3). In either case, the number of units other than the permease is shown to be three. The third possible combination of units, heterogeneous and requiring one of one kind and two of another kind of three required units, is described by the following equation: co F' = E x1/k! * exp(-x) [1 - exp(-x)] k = 2 The theoretical curve for this combination is also plotted in Fig. 3. Although the number of units can be established as three in any case, the closeness of the theoretical curves makes it impossible to conclude which of the theoretical curves corresponds to the experimental data. However, it can be determined from the intersection of the log F' axis and the x axis that: if the other units are homogeneous, the number of these units per fully induced cell is about 12; if the other units are all heterogeneous, this number should be about 6 for each unit; and if the third possibility holds true, there should be about 75 of each unit. DISCUSSION The studies presented in this communication demonstrate that agar plates containing ethidium bromide with galactose as the only source of carbon and energy are highly selective Proc. Nat. Acad. Sci. USA 7 (1978) for induced cells of yeast galactose long-term adaptation mutants. The applicability of this selective method for the analysis of the deinduction kinetics of the galactose system in these mutants was shown. That a fully induced diploid yeast cell may contain 2 permease units is quite reasonable when one considers that the number of permease units for the E. coli lactose system has been estimated to be between 1 and 2 (8) for a cell having a total surface area of about 6 sim2. A yeast cell of oval 4,um by 5/um dimensions has a total surface area of about 63 MAm2, indicating a similar distribution of permease units on both cell surfaces. It was concluded from the population analysis of the lactose system of E. coli that the f3-galactoside permease was the sole type of induction unit (8). Our data obtained from the population analysis of the deinduction process with gal 3 strains of yeast, however, indicate a Gal 2 gene product to comprise at most two out of five induction units. Thus, gal 3 mutants lack some component(s) required for initiation of induction, and this component is completely lacking only under conditions of respiratory deficiency. The yeast aerobic system, then, is capable of directly or indirectly influencing the synthesis of this component. Since we have demonstrated that gal 3 strains require respiratory function only for the initial induction processes, this component must be either required for the removal of an inhibitory condition and the onset of induction diminishes this inhibition or it is a product of the galactose-specific pathway such that the onset of induction continuously supplies this component. If the latter is the case, this component must be capable of being synthesized in the absence of inducer only in respiratory-competent cells, since gal 3 strains exhibit long-term adaptation when in the respiratory competent condition and the absence of visible growth of negative cells in the respiratory deficient state. The constant low fraction of positive cells observed in Phase II of Fig. 1 has been observed by previous workers who reported 5% of these colonies to be gal+ lines (9). We have observed that colonies transferred from EB+1 galactose plates in Phase II through glucose medium retain a typical gal 3 phenotype (data not shown). The hypothesis that the gal 3 mutants exhibit a long-term adaptation period because of a deficiency in the synthesis of the prerequisite endogenous compound is in accord with studies on the endogenous induction of the galactose operon in E. coli. Wu and Kalckar (15) have shown that the constitutive synthesis of the remaining galactose pathway enzymes in galactokinaseless mutants of E. coli was dependent upon the presence of UDPGlc-pyrophosphorylase and UDPGlc- 4-epimerase activity. They concluded that the constitutive nature of the galactokinaseless mutants was due to the generation of an endogenous inducer through an unimpaired UDPGlc/UDPGal pathway and that a likely candidate for this inducer would be a galactosyl compound derived from UDPGal. The yeast gal 3 phenotype, then, could be due to a severely impaired production of some required coinducer that, once accumulated in sufficient quantity, allows induction to occur. Induction of the galactose pathway would then allow for the synthesis of the requisite compound(s) from exogenous galactose. The remaining three induction units that are diluted out by cell division may, in fact, be the Leloir pathway enzymes.

5 Proc. Nat. Acad. Sci. USA 7 (1973) NOTE ADDED IN PROOF Since we submitted this manuscript, further studies on the pattern of fractional changes of double mutants homozygous for the gal 3 loci but heterozygous for the other mutant loci of the galactose system have shown that such mutant combinations involving the gal 1, gal 7, or gal 1 loci, but not the gal 5 locus, exhibit a significant shift in the deinduction kinetics similar to that seen in Fig. 3. We thank Dr. H. C. Douglas for his generosity in supplying the yeast strains used in this study and Dr. E. W. Jones for her assistance with the tetrad analysis. We also thank Dr. C. E. Folsome for his interest in this study and his critical review of the manuscript. This investigation was aided by Public Health Service Grant AM from the National Institute of Arthritis and Metabolic Diseases and an intramural research grant from the University Research Council, University of Hawaii. This paper is part of a thesis to be submitted by S.T. to the Graduate Division of the University of Hawaii in partial fulfillment of the requirements for the Ph.D. degree in Microbiology. 1. Winge,. & Roberts, C. (1948) C. R. Trav. Lab. Carlsberg. Ser. Physiol. 24, Analysis of Deinduction Kinetics Spiegelman, S. (1951) Cold Spring Harbor Symp. Quant. Biol. 16, Spiegelman, S. (1954) "Virology panel," Proc. Nat. Cancer Conf. 2nd Spiegelman, S. & DeLorenzo, W. F. (1952) Proc. Nat. Acad. Sci. USA 38, Spiegelman, S., DeLorenzo, W. F. & Campbell, A. M. (1951) Proc. Nat. Acad. Sci. USA 37, Spiegelman, S., Sussman, R. R. & Pinska, E. (195) Proc. Nat. Acad. Sci. USA 36, Novick, A. & Weiner, M. (1957) Proc. Nat. Acad. Sci. USA 43, Cohn, M. & Horibata, K. (1959) J. Bacteriol. 78, Douglas, H. C. & Pelroy, G. (1963) Biochim. Biophys. Acta 68, Goldring, E. S., Grossman, L. I. & AMarmur, J. (1971) J. Bacteriol. 17, Johnston, J. R. & MIortimer, R. K. (1959) J. Bacteriol. 78, Wickerham, L. J. (1946) J. Bacteriol. 58, Adams, B. G. (1972) J. Bacteriol. 111, Douglas, H. C. & Condie, F. (1954) J. Bacteriol. 68, Wu, H. C. & Kalckar, H. M. (1966) Proc. Nat. Acad. Sci. USA 55,