Control by Nutrients of Growth and Cell Cycle Progression in Budding Yeast, Analyzed by Double-Tag Flow Cytometry

Abstract

To gain insight on the interrelationships of the cellular environment, the properties of growth, and cell cycle progression, we analyzed the dynamic reactions of individual Saccharomyces cerevisiae cells to changes and manipulations of their surroundings. We used a new flow cytometric approach which allows, in asynchronous growing S. cerevisiae populations, tagging of both the cell age and the cell protein content of cells belonging to the different cell cycle set points. Since the cell protein content is a good estimation of the cell size, it is possible to follow the kinetics of the cell size increase during cell cycle progression. The analysis of the findings obtained indicates that both during a nutritional shift-up (from ethanol to glucose) and following the addition of cyclic AMP (cAMP), two important delays are induced. The preexisting cells that at the moment of the nutritional shift-up were cycling before the Start phase delay their entrance into S phase, while cells that were cycling after Start are delayed in their exit from the cycle. The combined effects of the two delays allow the cellular population that preexisted the shift-up to quickly adjust to the new growth condition. The effects of a nutritional shift-down were also determined.

A budding yeast divides by an asymmetrical process. The degree of asymmetry at division depends on the growth rate, with slow-growing cells dividing more asymmetrically (14, 16, 19–21, 31, 37). In an asynchronously growing S. cerevisiae population, individual cells differ in their positions within the cell division cycle, their genealogical ages (i.e., daughters, parents of first generation, parents of second generation, etc.), and their sizes, although cells of the same size do not necessarily have the same age or cell cycle position (8, 9, 14, 16, 17, 20–22, 28, 31, 34, 35, 38). All of these differences determine the cell size distribution (i.e., the cell protein content distribution) of the growing population. It has been shown that the protein content distribution of a given population in balanced exponential growth is stable and characteristic of each growth condition (1, 3, 4, 24, 27, 29, 35). This homeostasis depends on the mechanism that coordinates cell growth to DNA division cycle and prevents cells from becoming too small or too large (14, 28). Changes in the environment modify the protein content distribution of a given population in steady state, adjusting it, after a transitory period, to the protein distribution characteristic of the new growth condition (1, 3, 4, 29); this behavior indicates that the homeostasis mechanism is operating under constraints different from those active in steady-state environments. These experimental conditions could be very informative in investigating the effects of cell size on the control of cell cycle progression. In order to adequately describe a transitory state it would be useful to determine the relationship between the cell size of the newborn daughter cells (called Po) and the fraction of budded cells. In this study, we describe a set of experiments used to analyze the Po value and the duration of the budded phase during transitory states of growth. Studies on the dynamics of the growth of individual yeast cells and their relationships with factors affecting cell cycle regulation are generally done by cell synchronization procedures (38) or, with a relatively small number of cells, by time-lapse studies (14, 37, 38). Time-lapse studies and synchronous cell population analyses usually require elaborate experimental procedures and tend to perturb the physiological state of the cells. We used a recently developed flow cytometric approach that allows, in asynchronously growing S. cerevisiae populations, the tagging of both the cell age and the cell protein content of individual cells (24–27). This approach permitted characterization of the properties of growth at the single cell level for yeast populations exponentially growing at different specific growth rates.

The findings obtained during transient states of growth (i.e., in a nutritional shift-up, in a nutritional shift-down, and following the addition of cyclic AMP [cAMP]) indicate that the whole population needs several hours to reach the new balanced growth condition. We found that both the nutritional shift-up and the addition of cAMP delay the entrance into S phase and the exit from the cycle, although according to different schedules.

MATERIALS AND METHODS

Strains and growth conditions.

S. cerevisiae S288C (29) and OL214 (7) were used in this study. Cells were grown in flasks by being shaken at 30°C in Difco yeast nitrogen base (YNB, 0.67% [wt/vol]) medium. The carbon sources were 2% (wt/vol) glucose or 2% (vol/vol) ethanol. cAMP was added to a final concentration of 3 mM as previously described (7).

Staining conditions.

Cells growing under conditions of balanced exponential growth (balanced growth was typically observed in a cell concentration range of between 1 × 10⁶ and 6 × 10⁶ cells ml⁻¹) were collected by centrifugation (5 min, 5,000 rpm) and sonicated in order to allow the cell wall staining of each individual cell. Cells were stained by resuspension in precooled fresh YNB-based medium containing 120 μg of conjugated concanavalinA-fluorescein isothiocyanate (ConA-FITC) ml⁻¹ (ca. 3.6 mol of FITC per mol of lectin) at a cell concentration of 2 × 10⁸ cells ml⁻¹. The operations were carried out at 4°C. After 7 min of staining, the cells were recovered by rapid centrifugation and resuspended in a prewarmed fresh medium. We previously showed that the staining procedure does not perturb the normal growth behavior of the population (25). At different times after resuspension and new growth, yeast cells were collected by centrifugation, washed and fixed in 70% (vol/vol) ethanol for 20 min at 4°C. The fixed cells were centrifuged and washed once with phosphate buffer (pH 7.4), and the cell proteins were stained by resuspending the cells in 0.5 M sodium bicarbonate containing 50 μg of tetramethylrhodamine isothiocyanate (TRITC) ml⁻¹. After 30 min, cells were recovered by centrifugation, washed three times, and resuspended in phosphate buffer.

Since the staining procedure slightly sticks the cells to each other, it is very important to repeat the sonication procedure immediately before the flow cytometry analysis (24, 26, 27) in order to avoid the acquisition of cellular aggregates that could invalidate the analyses of the data.

Cell number, fraction of budded cells, budded phase, and Td-d determinations.

Cells were counted after sonication with a Coulter Counter ZBI. The specific growth rate was obtained by fitting the cell number against time. The fraction of total budded cells (FBC) was calculated after microscopic examination (at least 500 cells were scored). The duration of the budded phase during the balanced growth period or the time of budding (Tb), comprising the S, G₂, M, and G₁* cell cycle phases, was determined from the following equation (19):

1

where Td is the duplication period for the whole cell population. Table 1 shows the budded-phase durations for the daughter cells; however, it is generally accepted that, at any given specific growth rate, the budded period is appreciably constant for both daughter and parent cells (3, 19, 25, 35). The duplication times of the daughter subpopulations (Td-d) were calculated with a mathematical model that takes into account four different subpopulations of parent and daughter cells (1, 3, 23, 24, 26, 27) or by using the following equation (19):

2

where FDB represents the fraction of daughter cells that are budded.

TABLE 1.

Duplication time (Td-d), budded- and G₁-phase durations, Po (cell size at birth), Ps (cell size at Start), and Pd (cell size at division) for different cohorts of daughter cells born before, during, and after shift-up

Cohort of daughter cells considered (specific growth rate, h−1)	Duplication time (h)	G1 phase (h)	Budded phase (h)	Po (channel no.)	Ps (channel no.)	Pd (channel no.)
Cells born during balanced growth on ethanol, i.e., before staining; t = −1 h in the figures (0.1556)a	6.56	4.08	2.48	88	166	244
Cells existing at Start at time of staining; t = 0 in the figuresb	7.58	4.08	3.5	88	166	ND
Cells born at t = 0.666 h after resuspension (0.324)c	5.00	3.61	1.39	90	289	455
Cells born at t = 1.000 h after resuspension (0.324)c	4.66	3.27	1.39	100	288	453
Cells born at t = 2.000 h after resuspension (0.324)c	3.66	2.27	1.39	136	284	447
Cells born during the balanced growth phase on glucose, t = 4 to 7 h in the figures (0.324)a	2.80	1.41	1.39	180	283	446
Cells born during exponential growth phase on glucose (i.e., the control) (0.352)a	2.74	1.43	1.31	175	289	458

^aThe duplication times of the daughter cells (Td-d) and G₁ and budded-phase (Tb) durations were calculated by using a model developed for the unequal division of the budding yeast (see text). The duplication times and budded-phase durations were also determined by using equations 1 and 2 (both approaches yield the same values). Since it is generally accepted that the budded-phase durations are equal for daughter and parent cells, the durations of the G₁ phases were simply inferred by subtracting Tb from Td-d. Po sizes were determined experimentally as shown in Fig. 2B. Since we experimentally proved that the cells grew exponentially during both balanced (27) and perturbed conditions (Fig. 3), Ps and Pd were simply estimated from the exponential law as follows: Ps = Poe^μTG₁ and Pd = Poe^μTd-d. 

^bSince this cohort of daughter cells had been stained at Start, the cells have the same G₁ duration and Po and Ps as the cohort of daughter cells described in row 1. The budded phase has been experimentally determined by considering the time of appearance of completely unstained cells (see the text). The duplication time was calculated from the G₁ and budded-phase durations. Since the approach does not allow a determination of the specific growth rate of the cells during the first few minutes (i.e., 30 to 40 min) after resuspension, the exponential law cannot be applied and the Pd value cannot be determined (ND). 

^cThe duplication times were determined as described in Fig. 3 (see the text). Budded-phase and G₁ durations were inferred based on findings reported in the text. Po was experimentally determined as shown in Fig. 2B. Ps and Pd were estimated from the exponential law as follows: Ps = Poe^μTG₁ and Pd = Poe^μTd. Pd − Ps yields the cell size of the newborn daughter cells of the first generation. These values are smaller than those of all of the newborn daughter cells (reported here and in the figures), which include the newborn daughter cells of the first generation (born from daughter cells), the newborn daughter cells of the second generation (born from parent cells with one bud scar), the newborn daughter cells of the third generation (born from parent cells with two bud scars), and so on. 

For the determination of FDB, yeast cells were stained with Calcofluor and scored under a Leitz Dialux fluorescence microscope and, according to the presence of bud scars, they were assigned to one of the following classes: unbudded daughters (cells without bud and without scars), budded daughters (with bud and without scars), unbudded parents (without bud and with one or more scars), and budded parents (with bud and with one or more scars) (26, 27, 35). At least 600 to 800 cells were scored.

Flow cytometric analysis.

FITC and TRITC fluorescence signal intensities were determined with a FACStarplus (Becton Dickinson) equipped with an argon-ion laser (excitation wavelength, 488 nm; laser power, 200 mW) (24, 26, 27). The sample flow rate during analysis did not exceed 500 to 600 cells per s. Typically, 50,000 cells were analyzed per sample. Only raw data have been used to prepare figures.

RESULTS

Analysis of the dynamics of growth during a nutritional shift-up.

Figure 1 shows the growth curve, the FBC, and the cell protein content distributions of an S288C yeast population that was grown on ethanol-YNB, harvested during the exponential phase of growth (specific growth rate, 0.156 h⁻¹; duplication time, 4.42 h), washed, and then resuspended in a glucose-YNB fresh medium at time t = 0. The parameters obtained during the nutritional shift-up indicate a very strong modulation of the FBC (Fig. 1A), the specific growth rate (Fig. 1A), and the distributions of the cellular protein content (Fig. 1B). It is important to note the oscillatory change of the FBC value and the increase of the protein content distribution of the growing population (Fig. 1B). The FBC represents a sensitive determination of the relative rates of the entrance into the S phase-budded phase and of the exit from the mitosis-cell division phase, while the increased size of the cells suggests that the mechanism which coordinates the cell growth with the DNA division cycle has undergone a resetting.

FIG. 1.

Effects of a nutritional shift-up on the FBC value, cell concentration, and distribution of the cellular protein content. (A) S288C yeast cells were grown on ethanol-YNB medium to balanced exponential phase (0.156 h⁻¹), harvested, washed, and inoculated in fresh glucose-YNB medium at t = 0. The specific growth rate and the FBC of the control yeast population exponentially growing on glucose-YNB were 0.352 h⁻¹ and 0.58, respectively. The specific growth rate of the yeast population upon reaching the balanced growth conditions was 0.327 h⁻¹. Symbols: ○, number of cells per milliliter for the control yeast population growing on ethanol-YNB; □, number of cells per milliliter for the control yeast population growing on glucose-YNB; •, number of cells per milliliter for the yeast population growing during the nutritional shift-up from ethanol to glucose; ■, FBC for the yeast population growing during the nutritional shift-up from ethanol to glucose. (B) At different times after resuspension, samples of the culture were withdrawn, and the cell proteins were stained with TRITC and analyzed at t = 0 h (A), t = 1 h (B), t = 2 h (C), t = 3 h (D), t = 4 h (E), and t = 6 h (F). GLU, distribution of the cellular protein content of the control yeast population growing on glucose-YNB.

The transitory state appears to last for about 4 h. After this, the yeast population reaches both the FBC value (Fig. 1A) and the cell protein content distribution characteristic of a population that is exponentially growing on YNB-glucose (i.e., the control population [Fig. 1B]). It is also relevant that only after 4 h of growth does the cell number per milliliter start to increase exponentially, with the yeast population growing at a specific growth rate very close to that observed for a control population exponentially growing on glucose-YNB (0.327 h⁻¹ versus 0.352 h⁻¹, respectively [Fig. 1A]). A transitory state period, which occurs during nutritional shifts from C₂ and C₃ carbon sources to glucose, has been reported (15, 36).

In order to better understand the transitory state of growth, one should analyze and compare different cohorts of cells over time, e.g., existing budded cells and newborn daughter cells. A biparametric flow cytometric procedure recently developed in our laboratory allows the determination of the cell size of cells belonging to the different cell cycle phases of an asynchronously growing yeast population (24–27). Briefly, the procedure is based on labeling the cell wall surface with a lectin (i.e., ConA) conjugated to a fluorescent marker (i.e., FITC). Cells are harvested during exponential growth phase, stained with ConA-FITC, and then allowed to grow in a fresh culture medium. Staining conditions have been developed so the labeling procedure will not perturb the growth behavior and the cells will retain the surface label over the subsequent growth period (25). This first staining is then combined with the determination of the protein content of the individual cells by staining with TRITC (24, 26, 27). The analysis of the double-staining pattern over time has been used to provide direct information on yeast populations growing under balanced growth conditions. The obtained findings indicated that (i) there was an exponential increase of the cell size during growth of the individual cells, (ii) that the daughter and parent subpopulations grow at the same specific growth rate, and (iii) that the average cell size increase rate (CSIR) of each individual cell is identical to the specific growth rate of the overall population. Furthermore, this analysis permits determination of the length of the budded phase of the population, the dimension of the newborn daughter cells, and the cell cycle length for the daughter cell population so as to identify the complex structure of a growing yeast population (24–27).

Identification of newborn daughter cells and determination of their size during the transitory state.

We used the previously described procedure to analyze individual cells growing during the nutritional shift-up depicted in Fig. 1. Figure 2A shows the time course of the double-staining pattern. At time t = 0 (i.e., the time of resuspension) the cells belong to all of the different cell cycle positions and are completely stained with both of the fluorescent tags, ConA-FITC (see insert at the top) and TRITC; each individual cell of the yeast population is represented by a single dot. Since the synthesis of new cell wall material in the budded cells is restricted to the bud (5, 10, 24–27, 32), the first effect of the new cell growth is the production of newborn daughter cells with a gradually decreasing amount of surface stain (i.e., partially stained daughter cells). The analysis of the cytograms clearly indicates a growing tail of newborn partially stained cells and the accumulation of cells with the same degree of staining (i.e., unstained cells). The partially stained cells are newborn daughter cells that originated from cells stained during the budded phase (S-G₂-M-G₁*) of the cell cycle (24–27) (Fig. 2A). Completely unstained cells represent the newborn daughter cells that originated from cells stained while in the unbudded phase (G₁) (24–27). The time of appearance (i.e., after 3.5 h) of this last subpopulation of new daughter cells is a function of the duration of the S-G₂-M-G₁* phase (24–27). Therefore, this analysis allows a direct determination of the length of the budded phase during the transitory state (Table 1) (i.e., being 2.48 h for the budded phase of the yeast population growing on ethanol and 1.39 h for the budded phase during the balanced growth condition on glucose). Examples of selection of newborn daughter cells that originated during the transient state of growth are shown in the Fig. 2A: the gate R1 on the cytogram t = 0.66, the gate R2 on the cytogram t = 1, the gate R3 on the cytogram t = 2, the gate R4 on the cytogram t = 3, and the gate R5 on the cytogram t = 4 indicate the diverse regions where newborn daughter cells were found. In this way, it is possible to identify the newborn daughter cells (i.e., on the abscissas) and to determine their size (i.e., on the ordinates). Figure 2B shows the protein content value of diverse cohorts of newborn daughter cells that originated during the shift-up. The figure also shows the cell protein content of newborn daughter cells that originated during the balanced cell growth period on ethanol and glucose (i.e., the controls). As previously indicated, the analysis of such data also suggests that the yeast population needs about 4 h to reach the new balanced growth condition. After that point, the cell size of the newborn daughter cells appears to be constant and very close to that of the control population growing on glucose.

FIG. 2.

Dynamic of the double tagging over time for cells grown on ethanol-YNB, stained with ConA-FITC, and resuspended in glucose-YNB medium. (A) At different times after resuspension, samples of the culture were withdrawn and cell proteins were stained with TRITC. The cell wall tag (abscissa: ConA-FITC, channel number) and cell size tag (ordinate: TRITC, channel number) signals were acquired with a linear scale. During the acquisition of data at time t = 0, only the cells with lower ConA-FITC signals (i.e., the smaller cells) were acquired (St., completely stained), and the settings were not changed during the experiment. This approach provides evidence of even the smallest differences in the ConA-FITC fluorescence signals between individual cells (26, 27). The evolution of partially stained (P.St.) and then unstained (Un.) cells over time is clearly visible. The diagram at the top of panel A shows the cell cycle phases of a growing yeast cell. (B) Cell protein content value of cohorts of newborn daughter Po cells born at different times after resuspension. S288C yeast cells were grown in ethanol-YNB medium till the exponential phase; they were then stained with ConA-FITC, resuspended in glucose-YNB medium, and processed as described in Fig. 2A. Examples of selection of the regions where newborn daughter cells were located after birth are indicated by gates R1 (t = 0.66), R2 (t = 1.0), R3 (t = 2), R4 (t = 3) and R5 (t = 4) in Fig. 2A. To obtain the data after t = 4, cells from parallel and independent experiments were collected, completely stained with ConA-FITC, resuspended in the same medium, and processed as already indicated. Newborn daughter cells born on ethanol or glucose (controls, open symbols) were selected by applying the same procedure to yeast cells exponentially growing on ethanol or glucose. Data are expressed as the average channel number of the relative protein content distributions (TRITC signals).

Since the partially stained daughter cells originated from cells stained during the budded phase of their cell cycle (27), i.e., from cells that have already passed beyond Start, the increased size of the newborn partially stained daughter cells (Fig. 2B) could depend both on the increased duration of the budded phase (i.e., 3.5 h versus 2.48 h of the yeast population exponentially growing on ethanol) and on an increased specific growth rate.

Determination of the specific growth rate of individual daughter cells during the transitory state.

We previously showed that, since the stained cells retain the surface label over the subsequent growth period (i.e., the dye on the single cells is only diluted by the new growth) (25), it is possible to follow over time the growth dynamics of a cohort of selected daughter cells born at the same time (24–27). In Fig. 2A, an example of the selection of such a cohort of daughter cells over time (gate R1 in the cytograms t = 0.66 through t = 6) is reported. The dynamics of the cell protein content values over time are shown in Fig. 3. Different cohorts of daughter cells, born at different times during the same experiment (i.e., 0.666, 1, and 2 h after resuspension), have been analyzed. Experimental data acquired from birth till t = 5 clearly indicate an exponential increase of the cell size over time, reflecting an exponential growth of the individual cells (Fig. 3, open symbols). The coefficient of correlation for all cases is higher than 0.99. The exponential fitting of the data yields the CSIR, i.e., the specific growth rate (27), of the different cohorts of daughter cells. The CSIR values for the three different cohorts of daughter cells are the same (CSIR = 0.324 ± 0.05 hr⁻¹ from birth till t = 5) and are identical to the specific growth rate of the overall population during the newly reached balanced growth condition as determined by the increase of the cell number concentration (0.327 h⁻¹). Suddenly, 5 h after resuspension, an exponential rate law cannot be utilized any more (Fig. 3, closed symbols). However, it is important to note that the time at which such deviation from the exponential increase occurs is the same for the three different cohorts of cells. At the time of deviation, the cell sizes of the cells belonging to the different cohorts are also the same. We have previously shown for yeast populations exponentially growing with different duplication times that this deviation depends on, and slightly precedes in time, the cellular division of the daughter cells (i.e., 31 min for a yeast population growing with a specific growth rate of 0.25 h⁻¹ and 15 min and 5 to 7 min for specific growth rates of 0.224 and 0.211 h⁻¹, respectively) (27). Figure 3 also reports the cell size increase, from birth to division, of daughter cells born during the balanced cell growth on glucose. It is interesting that the cell size of the dividing daughter cells during balanced growth on glucose (i.e., 441.76) also represents the cell size value reached from the selected daughter cells at 40 min (t = 5.66) after the deviation observed at t = 5. The CISR values as well as the coefficients of correlation for the three cohorts of cells selected do not change following inclusion of these last data, and therefore it is reasonable to assume that the selected daughter cells divide at t = 5.66.

FIG. 3.

Average cell protein content (TRITC signals) of selected cohorts of partially stained daughter cells as a function of time after birth during the transitory state of growth. S288C yeast cells were grown in ethanol-YNB medium until the exponential phase; they were then stained with ConA-FITC, resuspended in glucose-YNB medium, and processed as described for Fig. 2A. (Since, from birth to division, the surface label does not change, the gate R1 in Fig. 2A selects the same cohort of partially stained daughter cells over time.) The figure shows the average cell protein content (TRITC signal) of selected cohorts of daughter cells born at different times (i.e., 0.666 h [◊ and ⧫], 1 h [□ and ■], and 2 h [○ and •]; see arrows) after resuspension. Data represented with the open symbols clearly fit an exponential increase of the cell size over time (in each case the coefficient of correlation is higher than 0.99), while the data represented by the closed symbols show that an exponential rate law cannot be utilized. Starting from t = 0 are also reported data for the cell size increase for daughter cells exponentially growing on ethanol or during the balanced-growth phase on glucose (i.e., the controls). These values have been calculated by taking into consideration the estimated specific growth rate of the daughter populations (0.156 and 0.327 h⁻¹, respectively), the estimated cell size values of the daughter cells at birth, and the calculated generation times of the daughter subpopulations (the generation times of the daughter populations were calculated, by using a previously developed mathematical model, to be 6.56 and 2.74 h for the yeast populations growing on ethanol and glucose, respectively). The same data were obtained by using equation 2 as described in the text). The cell size value of the dividing daughter cells during exponential growth on glucose (i.e., the control) has been used to extrapolate (dashed arrow) the correct cell sizes of the different cohorts of daughter cells at division (data in the circle). Data are expressed as the average channel number of the relative protein content distributions (TRITC signals).

Determination of generation time, duration of the G₁ phase, and duration of the budded phase of diverse cohorts of daughter cells born before, during, and after the transitory state.

By using the experimental procedure previously described in Fig. 2 and 3, the generation times for the daughter cells selected at different times after resuspension were determined (Table 1). These generation times were found to be 5.0 h for daughter cells born at t = 0.66 and grown till t = 5.66; 4.66 h for daughter cells born at t = 1.0 and grown till t = 5.66, and 3.66 h for daughter cells born at t = 2.0 and grown till t = 5.66. It is interesting that 5, 4.66, and 3.66 h are periods of time much longer than the duplication time of the overall population during the steady state of growth on glucose (i.e., 2.11 h = ln2/0.327).

Table 1 also reports the duplication times and the G₁ and budded phase durations of the daughter cells exponentially growing on ethanol both after the shift-up and during the exponential growth phase on glucose. The duplication time (Td-d) and the budded phase duration (Tb) have been estimated by using an algorithm developed for unequally dividing cells (1, 3, 23, 24, 26, 27) (equations 1 and 2 described above yielded the same results, respectively), while the G₁ phase durations were simply inferred by subtracting Tb from Td-d. Furthermore, Table 1 shows the duplication time of the daughter cells born on ethanol and stained with ConA-FITC at Start (Ps) at the time of resuspension. This value has been estimated by considering the time of appearance of completely unstained cells after the resuspension in the fresh medium. As already noted, the time of appearance of this last subpopulation of new daughter cells (i.e., at about 3.5 h) depends on the length of the S-G₂-M-G₁* cell cycle phase. The length of the duplication time for these daughter cells can be determined based on the length of the G₁ phase during exponential growth on ethanol (i.e., 4.08 + 3.5). Findings reported in Table 1 also indicate that the duration of the budded phase, i.e., 2.48 h in ethanol, increases to 3.5 h at the beginning of the shift-up and stabilizes at 1.39 h for cells born 0.666 h after the shift-up. All of the parent and daughter cells that have crossed Start at the time of the shift-up experience an increase in the length of their budded phase, producing newborn partially stained cells of increasing size (Fig. 2B). These new daughter cells have different duplication times (Fig. 3 and Table 1) and are growing exponentially and with the same specific growth rates (Fig. 3). Given the fact that they divide during the new balanced growth condition (t = 5.66), these cells will originate daughter cells of the same size. This consideration can be simply tested by analyzing the data reported in Fig. 2B. After t = 4, the cell size of the newborn daughter cells remains constant over time. Finally, taking into consideration that all of the new materials in the budded cells are restricted to the bud (5, 10, 24–27, 32), these cells share the same budded phase lengths, and this value is equal to that of cells growing under the balanced growth conditions (i.e., 1.39 h).

In conclusion, these data are consistent in indicating that the shift-up induces an early delay in the execution of cell division. The duration of the transitory state correlates with the fact that the value of protein content at birth (Po) of the daughter cells requires about 4 h to reach the value characteristic of the new, richer medium (see both Fig. 2B and Table 1).

Cell size of the newborn daughter cells as related to their G₁-phase length.

The nutritional shift-up produced newborn daughter cells of increasing size (Fig. 2B), growing with the same specific growth rate (Fig. 3) and having different duplication times and different G₁ phase lengths (Fig. 3 and Table 1). More importantly, these new daughter cells share the same budded-phase period and begin to bud and then divide at the same Ps and Pd values (Table 1). This observation suggests that the different cohorts of daughter cells that originated at different times after resuspension are able to actively control their commitment to Start phase and to cell division. A similar process of resetting of the individual cell cycles for daughter cells is also operative during balanced cell growth. In fact, newborn daughter cells which originated from parent cells of different genealogical ages have different sizes at birth but begin to bud and then divide showing similar protein contents (3, 26, 27, 35). A simple explanation is that all daughter cells are endowed at birth of the same amount of an inhibitor of the entrance into S phase that would be inactivated by a growth-dependent protein produced during G₁ phase (2). Figure 4A shows the relationship of the cell size of the newborn daughter cells produced over time during the nutritional shift-up (see also Fig. 3) compared to their G₁ durations (see also Table 1); Fig. 4B shows the relationship between the cell size of a single cohort of daughter cells during cell cycle progression and the remaining G₁ phase duration. The linear correlations found (in both cases the coefficients are higher than 0.99) suggest a relation between the cell size of the daughter cells and the duration of their G₁ phase. The daughter cells under examination in Fig. 4A were produced from cells that crossed Start before the nutritional shift-up. Since all of the cells increased in size at the same specific rate (i.e., the amount of protein synthesized per unit of cell mass per unit of time; see Fig. 3 and text) and grow through the same cell cycle phases, the synthesis rates (i.e., the amount of protein synthesized per unit of cell mass) of the constitutively synthesized proteins per cell may be assumed to be the same. However, if a protein(s) is compartmentalized in the nucleus and it is assumed that the dimension of the nucleus is constant and independent of the cell size, that protein will be inherited at increasing concentrations from the newborn daughter cells produced over time. The increased concentration(s) could determine a proportionally faster commitment to Start phase (Fig. 4). The growth-dependent factor would neutralize the above-mentioned inhibitor and induce a sharp and reproducible commitment to S phase at a given threshold (2). A tentative model that involves the Cln3 protein (and its constant accumulation rate through the cell cycle together with its compartmentalization in the nucleus) as the growth-dependent factor has recently been proposed (12).

FIG. 4.

Cell size of newborn daughter cells produced during the shift-up relative to their G₁ durations. (A) The cell sizes of the newborn daughter cells were determined as described in Fig. 2; their G₁ durations were determined as described in Table 1. (B) Cell sizes during cell cycle progression for the cohort of daughter cells originated at t = 0.666 (Fig. 3) relative to the remaining G₁ phase duration. Data are expressed as the average channel number of the relative protein content distributions (TRITC signals).

Analysis of the dynamics of growth after the addition of cAMP.

The same flow cytometric approach has been used to gain insight into the effects of the addition of cyclic AMP (cAMP) to yeast cells exponentially growing on glucose. In our laboratory it has been previously shown that the addition of cAMP to yeast cells growing on glucose determines an oscillatory behavior in the fraction of budded cells and an increase of the cell size (7). For such investigations, a strain permeable to cAMP is required (strain OL214). Figures 5A and B show the growth curve, FBC, and distributions of the cellular protein contents of a yeast population grown on glucose, harvested in the exponential phase of growth (Td = 1.86 h, with a budded-phase duration of 1.2 to 1.3 h), washed, and then resuspended in a glucose-YNB fresh culture medium containing 3 mM cAMP at time t = 0. The behavior of the FBC value and of the distribution of the cellular protein content is similar to that observed during a nutritional shift-up (see Fig. 1). However, several differences should be noted. Although following a nutritional shift-up the FBC value remains constant for about 0.666 h (Fig. 1), after the addition of cAMP it decreases immediately. A second main difference relies on the fact that after the addition of cAMP, the budded-phase period for cells stained at Start at the time of resuspension does not change compared to the value observed before the addition of cAMP (1.2 to 1.3 h [data not shown]). It is important to note that by 1.2 to 1.3 h after the addition of cAMP, the increase in the number of cells per milliliter is almost blocked (Fig. 5A), indicating that at this time also the rate of exit from the cycle is strongly reduced. On the other hand, confirming data previously obtained by determining the protein and RNA synthesis rates in the whole population (7), we did not detect any difference in the CSIR for cohorts of daughter cells selected before and after the addition of cAMP (data not shown). As a consequence, the size of the newborn daughter cells increases (Fig. 5C).

FIG. 5.

Effect of the addition of cAMP on the FBC value, number of cells per milliliter, distribution of the cellular protein content, and Po cell values. To OL214 yeast cells grown on glucose-YNB medium to balanced exponential phase (duplication time, 1.86 h), cAMP was added at time t = 0 (final concentration, 3 mM). (A) Symbols: □, number of cells per milliliter for the control yeast population growing on glucose; •, number of cells per milliliter for the yeast population after addition of cAMP; ■, FBC for the yeast population after the addition of cAMP. (B) At different times after resuspension, samples of the culture were withdrawn, and the cell proteins were stained with TRITC and analyzed at t = 0 h (A), t = 1.5 h (B), t = 1.83 h (C), t = 3 h (D), t = 3.5 h (E), t = 4.5 h (F), and t = 6 h (G). (C) Behavior of the average cell size of the population (□), the average cell size of the newborn daughter cells, Po (•), and the FBC in the population (■) for the yeast population growing after the addition of cAMP. Po and the average cell size of the population are expressed as the average channel number of the relative protein content distributions (TRITC signals).

Finally, similar to the findings obtained during a nutritional shift-up (Fig. 1), the analysis of the FBC values in Fig. 5A suggests that the population reaches an exponential growth condition after t = 4. However, cells become unable to reach a new balanced growth condition. Figure 5C shows that the average cell size of the whole population and the cell size of the new daughter cells (Po) produced over time also increased after t = 4.

DISCUSSION

Control of the cell cycle progression by nutrients has a key role in the regulation of cell proliferation in all organisms. Mechanisms are operative to coordinate nutrient availability to cell cycle transitions. Therefore, controlled manipulations of the cellular environment represent a useful methodological approach to gaining better understanding of the dynamics and the controls determining the cell cycle progression of a growing population.

In budding yeast, the main control of the cell division cycle takes place late in the G₁ phase at an area called Start (14, 28). At Start, three essential processes begin: budding, DNA replication, and spindle-pole body duplication. The cyclin-dependent kinase Cdc28, which is associated with the Cln3 cyclin, seems to be the most upstream activator of Start and acts as a potent inducer of two distinct transcriptional complexes, SBF and MBF, respectively (11, 12, 18, 30, 33).

The analysis of the findings obtained during the nutritional shift-up shown in Fig. 1 to 4 indicates that during this transition two delays are induced. The cells that at the moment of the nutritional shift-up existed before Start delay their entrance into S phase; cells cycling after Start delay their exit from the cycle. The fast decrease in the FBC (from 47.5 to 22.5%) that was observed during the first 2 h shows that the rate of entrance into S phase is much more reduced than the rate of exit from the cycle. When deprived of essential nutrients, yeast cells are arrested in G₁ phase. It has recently been shown that nitrogen-deprived G₁-phase-arrested cells are able to grow, reaching a volume much larger than the average size required for budding of nonstarved cells and thus suggesting a mechanism that causes the accumulation of cells in G₁ phase even when the critical mass for budding seems to be largely overcome (13). Such as response appears to be related to a faster degradation of Cln3 cyclin as a consequence of the changed nutritional conditions (13). In shift-up there is a prolonged stay in the G₁ phase, although the cell population is actively growing; a diminished availability of Cln3 in shifting cells could be the cause of the observed delay.

The second delay involves the cells shifted and stained after Start at the time of resuspension. The increased duration of the budded-phase period (from 2.48 to 3.5 h [Table 1]) indicates a delay in mitosis and cellular division. The main effect of such as delay is the production of newborn daughter cells of increasing size (Fig. 2B). These daughter cells grow with a specific rate characteristic of the new medium and are all committed to Start for the same Ps value characteristic for the new medium. The combined effects of the two delays allow the cell population that preexisted the shift-up to quickly adjust to the new growth condition.

The experimental approach described here has also been applied to gain insight into the effects of a nutritional shift-down. Cells were grown on fructose-yeast extract-peptone, harvested in the exponential phase of growth, stained with ConA-FITC, and then resuspended in a raffinose-YNB fresh culture medium. The Po cell value changes over time following a pattern similar to but opposite that shown for a nutritional shift-up. On the other hand, the qualitative behavior of the FBC value is similar to that observed during a nutritional shift-up, except that it shifts from higher (71% at time t = 0) to lower values (58% upon reaching the balanced growth condition) (data not shown).

Since the addition of glucose stimulates the level of cAMP in budding yeast cells, it has been proposed that at least part of the effects of glucose on growth and cell cycle progression could be mediated by cAMP, whose connection to the early events of Start is well known (6, 7). The analysis of the transition following the addition of cAMP to permeable yeast cells indicates both similarities and differences. While the shift-up both delays the entrance into S phase and the exit from the cycle for existing cells, the addition of cAMP first delays entrance into S phase and only after 1.2 to 1.3 h delays the exit from the cycle. The nutrients and cAMP modulations of the cell cycle, both at Start and at exit from the cycle, could offer useful experimental approaches for investigating the physiologically relevant events in cell cycle progression.