Synchronous cell growth occurs upon synchronizing the two regulatory steps of the Saccharomyces cerevisiae cell cycle

Experimental

Cell Research

151 (1984) 542-556

Synchronous Cell Growth Occurs upon Synchronizing the Two Regulatory Steps of the Saccharomyces cerevisiee Cel I Cycle SUSAN

A. MOORE

Deparrmenr of Chemistry, University of Guelph, Guelph, Ontario, Canada NIG 2~1, and Department of Genetics, University of Washington, Seattle, WA 98195, USA

There are two known asynchronous steps in the budding yeast Saccharomyces cerevisine cell cycle, where an asynchronous step is one which is completed in different lengths of time by different cells in an isogenic population. It is shown here that elimination of the asynchrony due to cell size by preincubation of cells with the mating pheromone a-factor, and decreasing the asynchrony in the cdc28 ‘start’ step by lowering the pH, yields highly synchronous cell growth measured as the time period between the emergence of buds. In one experiment, cell budding for 92 % of cells occurred within a 12-min period for at least two generations. Under identical conditions, cell number increase is not as synchronous as bud emergence indicating that there is a third asynchronous step, which is concluded to be at cell separation. These results are consistent with there being two-and only two-asynchronous steps in the cell cycle, measured from bud emergence to bud emergence. Surprisingly, these two steps are also the two major regulatory steps of the cell cycle. It is concluded that asynchrony may be a general feature of cell cycle regulatory steps. The asynchrony in the completion of the cdc28 ‘start’ step which occurs in the first cell cycle after a-factor washout is shown here to be almost or entirely eliminated for the second passage through this step after a-factor washout. The ‘true’ time between the onset of budding and the point where SO% of cells have budded (called rs,BE) is 17 and 62 min for the first and second budding, respectively, after a-factor washout. The cell cycle models requiring a transition probability, or asynchrony, at ‘start’ for every cell cycle are therefore incorrect.

Why are certain steps of the cell cycle synchronous, and others asynchronous? An asynchronous step is defined as one which is completed in different lengths of time by different cells in an isogenic population. In the broadest sense an asynchronous step can be identified as anything from a single gene-controlled event, to the cell cycle itself measured from division to division, which occurs in different periods of time for different isogenic cells. However, it is likely that the individual asynchronous steps combine to yield the overall asynchrony of the cell cycle. Those steps of the cell cycle that are asynchronous must, by definition, occur at different rates in individual cells. These steps identify the first level of cell cycle timing or control because different cells pass through the cell cycle at different rates due to these steps. I undertook to identify which steps of the yeast Copyright @ 1984 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827184 $03.00

Cell synchrony

and cell cycle regulation

543

cell cycle are asynchronous because, based on the above considerations, these steps are likely candidates to be involved in cell cycle control (see also ref. [l]). In this paper evidence is presented which indicates that the asynchronous steps of the yeast Saccharomyces cereuisiue cell cycle are few in number and are identical to the previously identified regulatory steps of the cell cycle. Several techniques-are available for obtaining synchronously dividing yeast cell populations (see ref. [2-71 and those in [I, 8, 91). However, there is no comprehensive analysis available, of which this author is aware, which describes the name and number of the asynchronous steps which are made synchronous in order to achieve the net cell division synchrony. The yeast S. cereuisiae cell cycle can be described as occurring in two stages, called Growth and Division (scheme I). The separation of the Growth and

‘.A

I COR

GROWTH to a CSR

DIVISION

(Stage

(Stage

I)

2)

Division portions of the cell cycle was reported several years ago [8]. More recent evidence has connected these two portions of the cell cycle by means of a critical size requirement for the cell division ‘start’ event, as follows (see also

[lOI). The stage 1 portion of the cell cycle occurs in Gl and involves growth of the newly formed cell to a critical size requirement (CSR) [l l-151. There is a variation or asynchrony in the stage 1 growth time requirement to reach the critical size among daughter cells in an asynchronousy growing population [ 13, 161. This is because the daughter cell sizes at completion of cell division vary, and are always or nearly always smaller than the CSR. In contrast, parent cell sizes at completion of cell division are larger and more nearly equal to the critical size for cell division [ll, 13, 161. The stage 1 growth time for parent cells is correspondingly short, and similar among parent cells [I 1, 161. Therefore, there is an asynchrony in the stage 1 time, the major source of which occurs among daughter cells in an exponentially growing culture. Significantly, the stage 1 growth time can be considered to be a regulated step of the yeast cell cycle because it is modulated by nutrient availability [17, 181, and compounds which alter the protein synthesis rate [l I]. Concomitant with or shortly after completion of the critical size requirement the genetically defined cdc28 ‘start’ step of the S. cereuisiae cell division is initiated [ll, 131. We cannot yet say whether the attainment of the critical size requirement and ‘start’ are consecutive or identical events [13]. It is known that the ‘start’ step is also mediated in an interdependent manner by the cdc28,36,37, Exp Cell Res 151 (1984)

544 S. A. Moore and 39 genes [19]. Stage 2 (scheme I) begins with the cdc28 ‘start’ step and is followed by the dependent sequence of events, including the visual events, of cell division per se [20]. Stage 2 ends with the completion of cell division, i.e., with cytokinesis and cell separation. Cell growth continues throughout stage 2 as evidenced in part by the emergence and growth of the bud. The stage 2 time is much less sensitive to various parameters of growth compared to stage 1. For instance, the stage 2 time from the cdc28 ‘start’ step to cell separation is much less sensitive to inhibition by the protein synthesis inhibitor cycloheximide [ 111, and by nutrient starvation [ 131, compared with stage 1. However, bud growth is inhibited by cycloheximide and the daughter cell size at the completion of division is smaller in the presence of cycloheximide compared with its absence [ll]. The time of stage 2 is relatively constant (i.e. synchronous) for all ceils in an asynchronously growing population, in contrast to the large variation (i.e. asynchrony) in the stage 1 time among cells [ll, 16-181. Therefore, stages 1 and 2 show a correlation between their degree of synchrony and their ability to be regulated by growth factors. The cdc28 mediated step has been termed ‘start’ because it controls the onset of cell division per se [20]. Preventing completion of the cdc28 step causes all cells to arrest as unbudded, mononucleated cells with an unduplicated spindle pole body [lo, 21, 32, 381, while growth continues normally [22, 321. The cdc28 step may be described as a regulatory step of the cell cycle because it controls the ability of the cell to undergo various differentiative pathways. Once the cdc28 ‘start’ step is completed, the cell is committed to undergo division events of stage 2. After the completion of start the cell cannot undergo conjugation to form diploid cells [22], and it does not enter into the stationary phase GO state [lo, 131, until stage 2 has been completed. In addition to its regulatory properties, the cdc28 ‘start’ step has been reported to occur asynchronously after its inhibition by a-factor, or by restrictive temperature using temperature-sensitive ‘start’ mutants [23-261. Therefore, the two known regulatory steps of the yeast cell cycle (i.e., growth to a critical size and ‘start’) are also the two reported asynchronous steps of the cell cycle, at least under certain conditions. Using this as a starting point I undertook to define how many and what individual steps of the cell cycle are asynchronous. The strategy used was to first eliminate the asynchrony in those steps known to display asynchrony, and then to determine the degree of asynchrony remaining in the overall cell cycle. This process could be repeated until all of the asynchronous steps of the cell cycle had been identified and the net cell cycle was synchronous. Surprisingly, it was found that eliminating the asynchrony in the two reported asynchronous steps was sufficient to yield synchronous cell growth measured from bud emergence to bud emergence. The strategy used to synchronize individual steps was to increase their rates of completion. Thus the asynchronous steps were accelerated such that the total time that all cells took to go Exp

Cell

Res

151 (1984)

Cell synchrony

and cell cycle regulation

545

through them was extremely short compared to the time required to complete the remainder of the cell cycle. EXPERIMENTAL

PROCEDURES

Saccharomyces cerevisiae haploid strain X2180-1A MATa [27] was used throughout. YMl medium was made as previously described [28] which yielded a final pH of 5.8. This medium was made to pH 7.0 by the addition of concentrated sodium hydroxide, or to pH 2.7 by the addition of concentrated hydrochloric acid. It is important how the medium is made to the desired pH because slightly different results were obtained when medium was made to pH 2.7 by varying the original concentration of sodium hydroxide, presumably because this method lowers the final salt concentration. Purified afactor was made as previously described [29] and showed a specific activity of 1.1 X lo4 units/Azso in the standard N,/Na cell division arrest assay [29]. Commercial products included sterile particle free 0.14 M sodium chloride (Travenol), 37 % formaldehyde containing 10-15 % methanol as preservative (Fischer Scientific Co.), Millipore filters (Millipore Corp., Bedford, Mass.), and noble agar (Difco). Cell wall digestive enzymes were sulfatase (23000 units/g of solid) and /Lglucuronidase (500000 units/g of solid) which were contained in type HS sulfatase from Helix pomatia obtained from Sigma Chemical Co. Photography supplies (Eastman Kodak Co.) were Tri-X Pan film (ASA-400) D-76 developer, Rapid Fixer, and Photo-Flo. Distilled water was used throughout. Quenching solution was sterile, particle free 0.14 M sodium chloride containing 2-4 % formaldehyde. This isotonic formaldehyde solution completely stopped cell growth and bud emergence.

Cell Treatment Cells were stored and cloned monthly to insure the purity of the yeast strains as previously described [29]. Exponentially growing cells were used in all experiments and were produced from stationary phase cells as previously described [29]. During experiments in liquid the cultures were incubated in Erlenmeyer flasks at GO.4 vol capacity with vigorous rotary shaking. Filtration of cells was done using a 0.45 pm Millipore filter. These cells were washed by passing approx. 25 ml of medium through the cells on the filter. Solutions were sonicated for l&30 set at 50-100 W of power to disperse cell clumps and cell numbers were obtained by particle count using a Coulter Counter [30]. Buds were counted on singlets, doublets, and quadruplets for first, second, and third bud emergence cycles, respectively, without additional sonications by rotating the cells under the coverslip during microscopic examination. A large fraction of quadruplet cells were broken open and killed, but quadruplets were not broken apart, upon repeated sonication. Such tight cell adhesion accounts for the variable, and initially less than 2-fold increase in cell number per generation time in fig. 6 of the results. In figs 1 and 5, the line representing bud emergence at
Time-lapse Photomicroscopy

on Agar Surface

Cells were photographed using a Canon OM-1 camera attached to a Zeiss phase contrast microscope containing a NG9 neutral filter (46-78-26, Zeiss) between the cells and light source. Cell growth was inhibited by the intense light of the microscope in the absence of the NG9 filter, possibly from a temperature effect. Agar slabs for time-lapse photomicroscopy on agar surface were prepared by modification of a previous procedure [I I] by pouring 10 ml of hot 1% noble agar over a glass slide in a Petri dish, cooling, and removing the slide with a thin agar slab on top of it. A drop of lightly sonicated log-phase cells at 106-10’ cells/ml grown in liquid YM1+2 % glucose, pH 5.8, 23”C, was rolled down the agar surface, a nylon grid was placed over the cells onto the agar, and a glass coverslip placed on a greased plastic ring surrounding and well above the agar slab. A drop of sterile water was placed on the slide adjacent to the agar slab on occasion to retard dehydration. At approx. 10 min intervals the slide was placed under a microscope, the coverslip removed, and the cells located and photographed. Photographed sections of the agar slabs were selected randomly and all cells within a section were monitored and included in the reported data. Exp Cell Res 151 (1984)

546 S. A. Moore The nearness of the cell to budding for the second bud was taken as the first bud time plus 110 min (see fig. 2 and the text for the rational behind this). This allowed the time of second bud emergence to be pinpointed with great precision, and several photographs were taken within a 5-7 min period at this time of calculated and actual bud emergence. The r5aBE values which were determined from timelapse photomicroscopy for cells on agar are in excellent agreement the values determined for cells in liquid (see fig. 1 and Results) which demonstrates the accuracy of this parameter which was determined by time-lapse photomicroscopy.

RESULTS Elimination

of Asynchrony

in the Stage 1 (Gl) Growth Time

The mating pheromone a-factor arrests cell division by preventing completion of the c&28 ‘start’ step [31] without preventing cell growth; cells grow up to thirty times as large as the normal cell size at division [32]. Fig. 1 shows that upon initial release of cells from 6 h of a-factor arrest, there is a lag followed by an asynchronous rate of bud emergence, as previously described [23-261. The % unbudded data for fig. 1 were obtained for cells in liquid (open circles) and on an agar surface (closed circles). From this data a value, called tsoBE, can be defined as the time point after the lag where 50% of the cells have budded after a-factor washout. These initial kinetics have been claimed to be first order, in which case the tssBE is a t1/2 value [23, 241, but there has been some controversy over this [33]. The term tssBE is therefore used in this report. The values of tSoBE are 17 and 18 min for the first and second buddings respectively when cells were grown in liquid. The corresponding tssBE values for cells resting on an agar surface and monitored by time-lapse photomicroscopy are 17 and 19 min. Thus the tsoBE values are identical within experimental error for cells allowed to bud in liquid or on the agar surface. The lag on the other hand was extended slightly by 7 min (first budding) to 14 min (second budding) for cells on agar compared with liquid. This may be due in part to the fact that new buds must grow slightly larger to be detected in the photographs of cells on the agar surface, compared with new buds formed in liquid and monitored under the microscope. Fig. 1 shows that the kinetics of bud emergence are identical for the first and second buddings after a-factor washout. This requires that the cells have an average generation time of 112 min, which is derived from the distance between the first and second bud emergence curves of fig. 1 at any % unbudded value. Despite the average generation time of 112 min, the data are consistent with cells having individual generation times ranging from 60 to 170 min. However, this variation in generation times is not what actually occurs, as can be seen in fig. 2. Fig. 2 displays the generation times for individual cells after a-factor washout. Note that the time between first and second bud emergence is the generation time. Fig. 2 shows that of the 36 cells monitored in fig. 1 (closed circles), 92% have first generation times after a-factor washout in the range of 104-116 min. This corresponds to a generation time of 1lo&6 min for 92 % of the cells. These Exp Cell

Res 151 (19841

Cell synchrony

and cell cycle regulation

547

1 0

SO

100 150 Ttme after elf wash-out

200 ml" Time

between

1st and

2nd

bud

emergence

Fig. 1. The kinetics of bud emergence after a-factor washout for two consecutive buddings. X2180-1A cells at lo4 cells/ml were incubated in 100 ml of YM1+2% glucose, pH 5.8, 23”C, containing 9.9 nM a-factor for 6 h. The cells were quickly filtered, washed, resuspended in 0.5 ml of medium lacking afactor, and sonicated lightly. For (0) a drop of the resuspended cells was placed on a 1% noble agar slab made in YMl+2 % glucose, pH 5.8 medium and a plastic grid was placed on top to allow location of individual cells as previously described [l 11. Bud emergence was monitored by time-lapse photomicroscopy in which the agar-grid structure was placed under a microscope at approx. 10 min intervals, and the cells were located and photographed. The times of bud emergence of individual cells were determined from the developed film strips. The total number of cells monitored was 36. For (0) the lightly sonicated cells which had been resuspended in 0.5 ml of YM1+2% glucose, pH 5.8, were incubated at 23°C. Aliquots of 0.01 ml were added to 0.02 ml of quenching solution and 300+ 100 cells were assayed by microscopic inspection for percent unbudded at various times after a-factor washout. Fig. 2. The actual variation in the second bud emergence after a-factor washout. The times of first and second bud emergence on individual cells were determined from the developed film strips which were obtained in fig. 1 (0). Therefore the total number of cells monitored was 36, and these were the same cells which make up the data in fig. 1 (0). The time between first and second bud emergence which is plotted on the x-axis is identical with the generation time Ill]. The y-axis value represents the percent cells which have a greater than x-axis generation time. The inset represents this curve plotted in semilog fashion. A kinetic constant called fsOis obtained from the inset, which represents the time since first bud emergence that it takes 50 % of the cells to undergo second bud emergence. This value is equal to 4 min.

36 cells monitored in fig. 1 (closed circles) and fig. 2 represents cells that were parents (i.e., cells which have budded at least once) and daughters (i.e., cells which have never produced a bud) at the time of a-factor addition. However, upon a-factor washout all of the cells become parents because they all produce a bud. The observed generation time of llOf6 min for these cells therefore represents the first parent generation time after a-factor washout. The generation times for ~95 % of the daughter cells formed on these parent cells after a-factor washout are also identical with one another and equal to 110+6 min. This was determined as follows. When cells are monitored for their second budding on agar or in liquid, it is observed that doublets, consisting each of a parent and a daughter cell, are converted directly to quadruplets (shown schematically in the upper portion of fig. 1). There are fewer than 5 % triplets at Exp

Cell

Res

I51 (1984)

548 S. A. Moore

any time in the population. This means that parent and daughter cells within parent-daughter doublets have identical generation times, measured from bud emergence to bud emergence, which is therefore 1 10+6 min. The parent generation time is defined as the time between the first and second bud formed on the parent cell; the daughtergeneration time is defined as the time between the first appearance of the daughter bud and the formation of a bud on the daughter cell

[Ill. Synchronous Bud Emergence a-Factor Washout

Occurs

during

the second Cell Cycle after

A significant conclusion which arises from the data is that the asynchrony in the rate of first bud emergence after a-factor washout is almost or entirely eliminated for the second bud emergence. Fig. 2 is a plot of the percentage of cells remaining with a greater than x-axis generation time. The kinetic constant obtained from this graph is tsoBE=4 min. Fig. 2 allows the conclusion to be made that the true half-maximal time for the second bud emergence after a-factor washout is ~4 min, as follows. The variation which was observed in the generation time and which is plotted in fig. 2 includes the variation due to the experimental error of the measurement by agar surface photomicroscopy (see the Experimental section for the preciseness of these measurements), as well as the variation due to the true t,aBE value for the second budding (called tSOBEzCtruej). On the other hand, the variation in the generation time shown in fig. 2 excludes the variation due to the first budding (t5,BE=17 min, fig. 1) because this is omitted in the measurement of the time between the first and second buddings. Fig. 2 therefore gives an upper limit for the true value of tsoBEz, which is ~4 min (fig. 2, inset). This is an upper limit because the observed value of 4 min includes the variation which arises from the experimental error. The hBE2(true) can be estimated in yet another way from the kinetics of bud emergence in liquid medium as: t5oB%true)

= tsoB%otx)-

f5oB%obs)

are the observed t50 values for first and t50BE2(ot,s) and t50BEl(obs) second bud emergence, respectively, which are obtained from fig. 1 (open circles). It was demonstrated at pH 5.8 that each individual cell has a generation time of 1 IO+6 min (fig. 2). This means that the cells which bud early and late for the first bud, go on to bud early and late, respectively, for the second bud in fig. 1. There is no rearrangement of the order of budding. This is an experimental fact. This means that the asynchrony in the first bud emergence (t50BEl(obs)=17 min) will be held over such that it appears in the second bud emergence kinetics, and therefore in the t50BE2(obs) value of 18 min (fig. 1, open circles). Thus, subtracting 17 min from the t50BE2~obs~ value for this ‘held over’ asynchrony yields the much smaller tSOBESCtruejof 1 min. The value of 1 min is

where

Exp

Cell

Res

151 (1984)

Cell synchrony

and cell cycle regulation

549

slightly smaller than the experimental error of the data in liquid, S2 min being a more rigorous value. It is noteworthy that t,,BE values which were identical within 1 min were also found for the first and second buddings after a-factor washout at pH 2.7 (fig. 5), and pH 7.0. In summary, it is concluded that the tSOBEzCtruej is at least ~4 min, and almost certainly ~2 min. This means that the asynchrony in bud emergence which appears in the first budding is almost or entirely lost for the second consecutive budding after a-factor washout, within the experimental error of the data. Relationship of Shortened Generation Other Cell Cycle Parameters

Time after a-Factor

Washout

to

After a 6-h preincubation in a-factor, the cells behave as ifthey have lost their stage 1 growth time requirement for cell division. This is based on the observation that the individual cell generation time of 1 lOf6 min after a-factor washout is identical with the calculated time of stage 2 (scheme I) in the untreated asynchronously growing population. The stage 2 time is calculated as follows. The time between bud emergence and completion of cell division can be calculated (see [l I]) to be 94f5 min from the generation time of the untreated asynchronously growing culture of 165 min and the percentage of unbudded cells in this population of 35+3%. To this time is added 20 min, which is the estimated time between the cdc28 ‘start’ step and bud emergence. This yields the total time of stage 2 of 114+5 min under the conditions of fig. 1. The time of 20 min between ‘start’ and bud emergence was determined as the lag in the time for the % UB value to increase upon the addition of a-factor to an exponentially growing cell population of X2180-1A cells in liquid medium under the conditions of fig. 1 (data not shown, see refs [29, 341). In addition an identical value of 20 min was obtained (data not shown) as previously described [23, 25, 261 from the time difference between: (a) bud emergence after a-factor washout and (6) the development of insensitivity to arrest by readdition of a-factor. The identical time of 20 min between ‘start’ and bud emergence determined using these two assays strongly suggests that 20 min is the actual duration of this segment of the cell cycle. It suggests that the time it takes cells to respond to the various treatments is insignificant (that is to cell division arrest by a-factor for the first assay and to both recover upon a-factor washout, and arrest cell division in a-factor in the second assay). Similar values in the range of 15-30 min have been obtained using an a-factor analog [26] or restrictive temperature arrest of temperature-sensitive ‘start’ mutants [23, 251. Achieving

Synchronus

Bud Emergence

after Alpha Factor

Washout

The value of t,,BE is pH-dependent (fig. 3). In contrast, both the population doubling time, and the time of the budded period of the cycle in untreated logphase cultures, showed less than 10% variation at pH values of 2.7, 5.8, and 7.0 Exp Ceil

Res I51 (1984)

550 S. A. Moore Time

_

(min.)

(min.)

t50BE : O

w 2

3 4

6’

PH

8

OrI’

2

2

4

’

’

6

4 8

PH

Fig. 3. The pH dependence of the q,BE after o-factor washout. Final conditions were X2180-1A cells in YMl+2 % glucose, 23°C and the pH indicated. Log phase cells at lo5 cells/ml were incubated in 100 ml of liquid medium containing 10-30 nM a-factor for 6 hours after which they were 295% unbudded. Cells were then filtered, washed, and resuspended in 10 ml of the original medium lacking a-factor. This solution was sonicated for 10 set at 3&50 W of power. Aliquots were removed at intervals, added to quenching solution, and assayed under a microscope for percent unbudded cells (% UB). The % UB values were plotted versus time after a-factor washout. This plot always showed a lag followed by a precipitous drop in % UB. The time between the end of the lag and the point where 50 % of the cells had budded is defined as the t,,BE for the first budding, and is plotted in the figure. Fig. 4. The pH dependence of the budded period and generation time, in the absence of a-factor. X2180-1A cells at 1~10~ cells/ml were grown in YM1+2% glucose, 23°C at the pH indicated. At various times aliquots were removed, added to quenching solution, sonicated, and Coulter-counted for cell number. The generation time (0) was obtained from a plot of log cell number vs time. The time of the budded period of the cell cycle (0) was obtained from the equation: rg = T[ln(l+Foa)~T/0.693) where ra is the budded period time, F or, is the fraction of unbudded cells in the culture, and T is the generation time [II]. The fraction of unbudded cells in the culture was obtained from the quenched, sonicated aliquots by assaying 300 cells under the microscope for buds.

(fig. 4). The value of ts,BE was found to be independent of the time of cell division arrest in a-factor between 4 and 10 h, at pH values of 2.7, 5.8, and 7.0. The value of rwBE is 6 min at pH 2.7. This results in 80% of the cells budding within a period of 14 min for first bud emergence, and 17 min for second bud emergence (fig. 5). In one experiment 92% of the cells budded within a 1Zmin period under conditions identical to those in fig. 5. This variation is probably due to minor variations in medium and temperature among experiments. It is concluded that the synchrony of bud emergence after a-factor washout is increased by reducing the pH. Asynchronous Cell Separation

It can be predicted that if cell number increase is as synchronous as bud emergence at pH 2.7 (see fig. 5), then 80 % of cells should double their number within a 14-17 min period for at least two generations. That is, the asynchrony observed in the bud emergence pattern should correspond precisely to the asynchrony in the subsequent cell number increase, because the time from the c&28 ‘start’ step to cytokinesis is identical [ll] among cells (see also fig. 2). Fig. 6 shows that an 80 % increase in cell number after a-factor washout takes approx. 40 min for both the first and second generations, rather than the predicted 14-17 min. This means that there must be an additional asynchronous step in the yeast Exp

Cell

Res

IS1 (1984)

Cell synchrony

IO6

L.

0

Time

after

o!f wash-out

551

and cell cycle regulation

(min.)

2

Time aflsr

4

6

6I 8

hours

Clf rash-out

Fig. 5. The kinetics of bud emergence after o-factor washout at pH 2.7. The data was obtained as described in fig. 3 and at pH 2.7. Fig. 6. The kinetics of cell number increase after a-factor washout. X2180-IA cells at 2~ 10’ cells/ml were grown in 100 ml of YMl+2 % glucose, pH 2.7, 23”C, containing 30 nM a-factor for 7.5 h, after which 295 % of the cells were unbudded. The cells were filtered and washed on a 0.45 urn Millipore filter, resuspended in 5 ml of medium minus a-factor, pH 2.7, sonicated lightly, and then resuspended in an additional 15 ml of medium minus a-factor, pH 2.7. Aliquots of 0.2 ml were removed at various times, added to 1.8 ml quenching solution, sonicated lightly, and Coulter-counted for cell number.

cell cycle-which is probably at cell separation, because the asynchrony appears in the cell number but not the bud emergence data after a-factor washout. Verification that cell separation is asynchronous among cells comes from the following observations. First, when exponentially growing cells are sonicated such that only budded and unbudded cells are present, and these cells are then allowed to grow for several generations, it is found that the majority of cells now exist in clumps of three to greater than ten cells. Thus, cells that have completed division adhere together. These cells probably separate asynchronously in order to produce clumps containing the varying numbers of cells. Secondly, it was found in preliminary experiments that increased sonication, or the addition of 0.5 mg/ml cell wall digestive enzymes (sulfatase + glucuronidase) at the time of afactor washout, both caused a greater synchrony in the cell number increase data for X2180-IA cells under conditions identical to fig. 6. Third Bud Emergence Cycle after a-Factor Washout

The kinetics of bud emergence for the third budding cycle after a-factor washout were identical with those for the first and second in an experiment identical with that in fig. 1 except that the temperature was 34°C and cells were incubated in 150 nM a-factor (obtained from Sigma Chemical Co.) for 4 h. The observed fsoBE at 34°C was 7 min and identical for the first, second, and third budding cycles. This is consistent with a true tSOBE of 7 min for the first budding, and
552 S. A. Moore It is noteworthy that the generation and stage 2 (scheme I) times in cultures treated with a-factor, at 34°C as above, are 82 and 57 min respectively. generation time (g) after cc-factor washout for both gl and g2 was ,58 min identical with the stage 2 time. The gl and g2 values were determined from average distance between, respectively, the first and second, and the second third bud emergence curves after a-factor washout (data not shown).

not The and the and

DISCUSSION Synchronization

of the Two Known Regulatory

Steps

An asynchronous step may be made synchronous in at least two ways. Cells may be allowed to complete an asynchronous step during arrest at a block, resulting in the elimination of the asynchronous step from the cell cycle for one or more generations after release from the block. Secondly, creating conditions which accelerate the rate of completion of an asynchronous step is tantamount to making that step more synchronous. One or both of these techniques was achieved in this study. The two known asynchronous steps which are required in the yeast S. cereuisiae cell cycle were synchronized among ceils resulting in overall synchronous cell growth. These two asynchronous steps are, surprisingly, identical with the regulatory steps or phases of the cell cycle. It is plausible to suggest that the regulatory steps of a given cell cycle may generally be characterized by asynchronous behaviour. The first and major asynchrony in the S. cereuisiae cell cycle occurs in the time that newly divided cells take to complete the stage 1 (Gl) growth requirement and reach the critical size. The stage 1 (Gl) growth time was greatly accelerated or already completed by preincubating cells in a-factor. The asynchrony in this step was correspondingly reduced or eliminated. The generation time for nearly all cells after a-factor washout at 23°C was 110+6 min, and was therefore shortened to the time it takes to go from the c&28 ‘start’ step to cell separation in the normal untreated culture (scheme I). This same phenomenon occurred at 34°C. A decrease in the Gl (or unbudded) phase of the cell cycle after the arrest of cells in a-factor [43] or upon growth in hydroxyurea 144, 451 has been reported previously. The acceleration or elimination of a cell cycle program after arrest at a block has been observed (see ref. [S]). An interesting and well studied example is during synchronous meiosis in the fungus Coprinus where an arrest prior to premeiotic DNA synthesis results in the acceleration or elimination of the entire diplotene phase after recovery [35]. In this case the chromosomal nicking program is carried out during arrest, and this apparently obviates the need for a lengthy diplotene phase upon recovery. The second asynchrony in the S. cereuisiae cell cycle is the asynchronous budding which occurs upon a-factor washout (fig. 1). Further analysis has shown that the asynchrony appears in the kinetics of the completion of ‘start’, which is Exp Cell

Res 151 (1984)

Cell synchrony

and cell cycle regulation

553

followed by a constant time to bud emergence among cells [23,26]. However, the data reported are consistent with the actual asynchrony occurring either at or before the cdc28 ‘start’ step. a-Factor certainly arrests cells at the gene-controlled event which is the cdc28 ‘start’ step [31]. Upon recovery from this arrest the kinetics will include any steps required for recovery from the arrest process, as well as the completion of the cdc28 ‘start’ step. One or several of those steps required for recovery from arrest, which occur prior to completion of ‘start’, could be the asynchronous step which is revealed in the recovery kinetics. Therefore it is concluded from the data presented in this study that lowering the pH accelerates the regulatory process (or, more specifically, the recovery aspect of the regulatory process) which occurs at ‘start’ (fig. 3), but lowering the pH may or may not accelerate the cdc28 ‘start’ step itself. The identification of an asynchrony with a regulatory step holds true for ‘start’, regardless of the above considerations, because the observed asynchrony (fig. 1) is either at ‘start’, or brought about by the regulation of the ‘start’ step by a factor. Asynchrony in Growth to a Critical Size Requirement (CSR) Is in a Different Step Compared with the Asynchrony at or before ‘Start’

After cells are grown large in a-factor, and then released from a-factor arrest, the requirement for growth to a critical size is functionally absent (see Results). In actual fact it is either eliminated or made to occur extremely rapidly. The asynchrony in bud emergence which is normally associated with this step is also absent after a-factor washout. The ‘start’ event, on the other hand, is prevented from occurring in the presence of a-factor, and an asynchrony in bud emergence which is at or before ‘start’ [23, 261 is observed upon release from a-factor arrest. Therefore, the growth to a critical size and the ‘start’ step behave oppositely in response to a-factor prearrest. Significantly, it has been shown that there is no correlation of the time of bud emergence on a cell after a-factor washout, with the size of the cell at the time of a-factor washout for daughter cells [44], or with the size of the cell at the time of a-factor addition for parent and daughter cells combined as measured by the technique [43] of perfusion photomicroscopy (S. A. Moore, unpublished observation). These data, taken together, demonstrate that the bud emergence asynchrony after a-factor washout, which is at or before ‘start’, is in a different step compared with the asynchrony in the growth to a critical size. The above data do not rule out the hypothesis that the CSR and ‘start’ are identical events. However, this is consistent with the data only if (1) the hypothetical single ‘CSR-start’ event is not completed during a-factor arrest, but is completed extremely rapidly upon a-factor washout; and (2) the asynchrony at or before ‘start’ which is observed upon a-factor washout is in fact before ‘start’, and is a result of the a-factor arrest process. 36-848334

Exp Cell

Res I51 (1984)

S. A. Moore

554

Cell Separation

is Asynchronous

A third asynchrony in S. cerevisiae was found which does not occur in, and is extraneous to, the cell cycle measured from bud emergence to bud,emergence. This asynchrony appeared in the cell number increase but not the bud emergence kinetics, and is at cell separation. Cell separation is regulated insofar as it is accelerated by such external factors as shear force and the presence of exogenous cell wall digestive enzymes. Therefore, cell separation is a regulated step which has the general feature of asynchronous behaviour. This is consistent with the conclusion of this study that asynchrony may be a general feature of cell cycle regulatory steps. In nature, the ability of cells to transform from clumps of cells to single cells in response to certain external factors may confer a survival advantage on the cells. Such regulation is precluded under the conditions where cell separation is already rapid, and therefore essentially synchronous. Cell Cycle Model Requiring

a Transition

Probability

at ‘Start’

Is Incorrect

A model for the yeast cell cycle has been proposed in which cells grow to a critical size, after which they initiate cell division [ll-151. A second model has been proposed [23, 241 in which cells must grow to a critical size, after which they enter the so-called A state of transition probability theory [16, 361 from which their exit into cell division is probabilistic and displays first-order kinetics. This has been called the tandem (critical size-transition probability) model [ 16, 23, 241. It was proposed in this second model that an asynchronous completion of the cdc28 ‘start’ step occurs in the normal cell cycle and is furthermore identical with the above-mentioned probabilistic event of the normal cell cycle [23, 241. This proposal was based exclusively on the conclusion that completion of the cdc28 ‘start’ step after its arrest is asynchronous and follows first-order kinetics. However, it was pointed out above that the previously reported data [23, 241 are consistent with the observed asynchrony occurring not just at, but also before the cdc28 ‘start’ step, under the conditions where the asynchrony is observed. Secondly, this asynchrony is seen only after prearrest of cells at ‘start’, [23, 261; it has not been demonstrated to occur in the normal untreated cell cycle. Firm evidence against a required transition probability, or asynchrony, in the cell cycle at ‘start’ was obtained in this study. The asynchrony in bud emergence (fig. 1) is in fact greatly reduced or eliminated for the second consecutive budding after a-factor washout (cf fig. 2). The results obtained at 34°C are consistent with the initial asynchrony being absent in both the second and third bud emergence patterns after a-factor washout. The asynchrony in tirst bud emergence after afactor washout has been resolved into an asynchrony which occurs at or before the ‘start’ event [23, 261. Therefore the data described above allow the conclusion to be made that the asynchrony at ‘start’ is greatly reduced or eliminated for the second, and almost certainly for the third consecutive cell cycle after a-factor Exp

Cell

Res

I.71 (1984)

Cell synchrony

washout. Therefore,

an asynchrony

and cell cycle regulation

at ‘start’ is not an obligatory

555

part of the cell

cycle. This result is consistent with the observed asynchrony occurring at a step which is prior to the cdc28 ‘start’ step, and which is therefore involved in recovery from the a-factor arrest process. Such an asynchronous step would have to be followed by a constant time to the “start” step. The ‘start’ step in turn would have to occur synchronously among cells for its first as well as its subsequent completions after a-factor washout. Alternatively, the data can be accounted for if the observed asynchrony occurs at ‘start’, and the abnormally large cells reset the rate of this step after their first but before their second passage through it after a-factor washout. In either case the tandem (critical sizetransition probability) model of the cell cycle [16, 23, 241 is incorrect as it stands, and must be modified or eliminated to account for the asynchrony at ‘start’ not being an obligatory part of the cell cycle. It is noteworthy that the critical size model for cell cycle control, for which there is a great deal of experimental evidence [ 1 I, 161, sufficiently accounts for all of the data on the cell cycle except that concerning release from ‘start’ after a block (see ref. [16]). It seems that the critical size model is the most strongly supported by the data at this time, as previously suggested [16, 33, 371. Comparison

with Other Reports

a-Factor prearrest has been reported to produce synchronous cell number increase [7, 391 although most reports show complete asynchrony in this parameter 119, 40-421. These differences in data are probably due to variations in the adhesion of cells to one another, even after sonication, because bud emergence is synchronous for several generations after a-factor washout (figs 1, 5), whereas cell number increase is not (fig. 6). The author is grateful to Professor Michael J. Gresser for support during the writing of the manuscript, to Professor Leland H. Hartwell for helpful discussions, laboratory space and equipment, and to Janice Abbott, Kathi Kennedy and Dianne Mallette for help in the preparation of the manuscript. S. A. M. was supported by postdoctoral fellowships from Damon Runyon-Walter Winchell (DRG-171-F), and the NIH (I-F32-GM07581-01) and by research grants from the Banting Foundation and the University of Guelph Research Board. Professor Hartwell was supported by NIH (grant GM-17709).

REFERENCES 1. Brooks, R F, The cell cycle (ed P C L John) p. 35. Cambridge University Press, London/New York (1981). 2. Williamson, D H & Scopes, A W, Nature 193 (1962) 256. 3. Kramhoft, B & Zeuthen, E, Methods in cell biology (ed D M Prescott) vol. 12, p. 373. Academic Press, New York (1976). 4. Dwek, R D, Kobrin, L H, Grossman, N & Ron, E Z, J bacterial 144 (1980) 17. 5. Lloyd, D, John, L, Edwards, C & Chagla, A H, J gen microbial 88 (1975) 153. 6. Gordon, C N & Elliott, S G, J bacterial 129 (1977) 97. 7. Newlon, C S, Petes, T D, Hereford, L M & Fangman, W L, Nature 247 (1974) 32. Exp

Cell

Res 151 (1984)

556 S. A. Moore 8. Mitchison, J M, The biology of the cell cycle. Cambridge University Press, London/New York (1971). 9. Zeuthen, E, Synchrony in cell division and growth. Wiley, New York (1964). 10. Pringle, J R 8~ Hartwell, L H, Molecular biology of the yeast Saccharomyces. Life cycle and inheritance (ed J N Strathem, E W Jones & J R Broach) p. 97. Cold Spring Harbor, N.Y. (1981). 11. Hartwell, L H & Unger, M W, J cell biol 75 (1977) 422. 12. Fantes, P & Nurse, P, Exp cell res 107 (1977) 377. 13. Johnston, G C, Pringle, J R & Hartwell, L H, Exp cell res 105 (1977) 79. 14. Johnston, G C, Ehrharst, C W, Lorincz, A & Carter, B L A, J bacterial 137 (1979) 1. 15. Fantes, PA, Grant, W D, Pritchard, R H, Sudbery, P E & Wheals, A E, J theor biol50 (1975) 213. 16. Lord, P G & Wheals, A E, J cell sci 50 (1981) 361. 17. Jagadish, M N & Carter, B L A, Nature 269 (1978) 145. 18. Carter, B L A & Jagadish, M N, Exp cell res 112 (1978) 373. 19. Reed, S I, Genetics 95 (1980) 561. 20. Hartwell, L H, Culotti, J, Pringle, J R & Reid, B J, Science 183 (1974) 46. 21. Byers, B, Molecular biology of the yeast Saccharomyces. Life cycle and inheritance (ed J N Strathem, E W Jones L J R Broach) p. 59. Cold Spring Harbor, N.Y. (1981). 22. Reid, B J & Hartwell, L H, J cell biol 75 (1977) 355. 23. Shilo, B, Shilo, V & Simchen, G, Nature 264 (1976) 767. 24. - Ibid 267 (1977) 647. 25. Shilo, B, Simchen, G & Pardee, A B, J cell physio197 (1978) 177. 26. Samokhin, G P, Lizlova, L V, Bespalova, J D, Titov, M I & Smimov, V N, Exp cell res 131 (1980) 267. 27. Hartwell, L H, J cell biol 85 (1980) 811. 28. - J bacterial 93 (1967) 1662. 29. Moore, S A, J biol them 258 (1982) 13849. 30. Hartwell, L H, J bacterial 104 (1970) 1280. 31. Hereford, L M & Hartwell, L H, J mol biol 84 (1974) 445. 32. Throm, E & Duntze, W, J bacterial 104 (1970) 1388. 33. Wheals, A E, Nature 267 (1977) 647. 34. Manney, T R, Duntze, W & Betz, R, Sexual interactions in eukaryotic microbes (ed D H O’Day & P A Horgan) p. 21. Academic Press, New York (1981). 35. Lu, B C, Molecular basis of genetics processes, vol. 3. p. 305. Proc 14th int congr Moscow, MIR, Moscow (1981). 36. Smith, J A & Martin, L, Proc natl acad sci US 70 (1973) 1263. 37. Nurse, P & Fantes, P, Nature 267 (1977) 647. 38. Byers, B & Goetsch, L, Cold Spring Harbor symp quant biol 38 (1974) 123. 39. Shulman, R W, Methods in cell biology, vol, 20. p. 35. Academic Press, New York (1978). 40. Stotzler, D, Betz, R & Duntze, W, J bacterlol 132 (1977) 28. 41. Chart, R & Otte, C, Mol cell bio12 (1982) 11. 42. Chan, R, J bacterial 130 (1977) 766. 43. Moore, S A, J biol them 259 (1984) 1004. 44. Lord, P G & Wheals, A E, J cell sci 59 (1983) 183. 45. Singer, R A & Johnston, G C, Proc natl acad sci US 78 (1981) 3030. Received May 12, 1983 Revised version received November 24. 1983

Exp

Cell

Res

151 (1984)

Printed

in Sweden