ABSTRACT
Synchronous cultures of Chlorella, that were obtained with minimum metabolic perturbation by centrifugal selection, reveal that progress through the cell cycle requires no change in the poly(A)÷ mRNA population, although changes do occur during nutritional adaptation. Of the abundant soluble proteins, 93% are synthesized continuously through the cell cycle and those that are discontinuous show similar patterns in control cells. The synthesis of proteins is compared with parallel studies of accumulation of enzyme activity and it is shown that there is no discrepancy in their pattern of accumulation when both are studied under the same culture conditions. The eukaryote cell cycle can allow stable relative rates of synthesis of most proteins and balanced rates of accumulation of most enzyme activities. Macromolecule classes differ in their rates of accumulation throughout the cell cycle: total RNA increases linearly, poly(A)+ RNA accumulation is restricted to G1 phase, but total protein accumulation accelerates smoothly through G1, S and mitosis phases, pausing at cytokinesis. There is no evidence that the cell cycle requires an extensive programme of differential enzyme synthesis. The cycle can therefore proceed with minimum disturbance of metabolism required for growth.
INTRODUCTION
Most reports of enzyme activity in the cell cycle have described discontinuous patterns of increase. Such periodicity is particularly clear in synchronized populations of algae such as Chlorella (John et al. 1973; Schmidt, 1974; Lorenzen & Hess, 1974), and has been the subject of numerous reports in synchronous cultures of Saccharomyces cerevisiae (see reviews by Mitchison, 1969; Halvorson, Carter & Tauro, 1971). Paradoxically, individual proteins in yeast have shown continuous synthesis when non–synchronous cells were pulse–labelled and then segregated into phases of the cycle (Elliott & McLaughlin, 1978). It has therefore been uncertain whether there is a programme of biochemical development extending throughout the cell cycle.
The present study rejected the use of synchronization by periodic illumination, which is commonly used to study the cell cycle of unicellular algae (John et al. 1973; Lorenzen & Hess, 1974), because, although periodic illumination does yield highly synchronous populations of cells, we have recently compared such cultures with others in which cells experienced continuous illumination and a constant degree of self–shading, and we have detected that synchronization that involves changes in illumination initiates metabolic adaptations that are not part of the cell cycle. Change in the amount of light available to individual cells causes fluctuations in photosynthetic capacity, in levels of starch, in rates of respiration and in the accumulation of enzyme activities that occupy a whole cell cycle (John, Lambe, McGookin & Orr, 1981). Similar distortions of enzyme accumulation caused by starvation and by osmotic stress were noted when sucrose density–gradient centrifugation was used to select small cells of Schizosaccharomyces pombe for synchronous culture (Mitchison, 1977).
In the present study synchronous cultures were prepared by selecting small cells by continuous–flow centrifugation from an asynchronous population that was growing under constant environmental conditions. Care was taken that the selected cells were cultured at the same turbidity as they were in the parent culture, since a change in culture density would alter the amount of light reaching individual cells and so alter the growth rate. The potential hazard of such changes in growth rate for the study of enzyme synthesis is indicated by fluctuations in the rate of accumulation of four enzymes in Chlorella, following a 40% change in culture density (John, Cole, Keenan & Rollins, 1980).
MATERIALS AND METHODS
Cell culture
Chlorella strain 211–8p was cultured in mineral salts medium under conditions of illumination and aeration that have been described previously (McCullough & John, 1972). For turbidostat culture a control system similar to that of Myers & Clark (1944) was used. An imbalance between a photocell receiving light through the culture and a reference photocell allowed an inflow of fresh medium, which stabilized culture turbidity. Medium inflow was continually monitored by the automatic recording of liquid level in the medium reservoir. The extent of culture dilution up to each sample time was calculated, taking into account periods when the culture was recovering its volume after sampling and periods when the culture was overflowing (Herbert, Elsworth & Telling, 1956). Data presented in Figs, 1 and 2 have been corrected for culture dilution. They therefore provide a direct indication of growth and show patterns of accumulation that would have been observed in batch culture, if effects of increasing turbidity and increasing mutual shading could have been eliminated.
Cell selection
Small cells were selected from asynchronously dividing culture by continuous–flow centrifugation (Lloyd, John, Edwards & Chagla, 1975). A nylon rotor was made in the form of a shallow cup of 90 mm diameter and 25 mm deep. The sides tapered inwards leaving an opening of 76 mm and the inner face of the rotor contained 11 semi–circular pockets 6 mm deep, into which the larger cells sedimented. Smaller cells passed out of the rotor, over the lip of the rotor cup with the overflowing medium, and were collected in an enclosing dish, from which they drained and were then aerated to await refractionation or inoculation into culture.
Cell number
Cell density was determined using a Coulter Counter BZ.
Protein and DNA
Protein was estimated in extracts of broken cells after precipitation with 5% (w/v) trichloracetic acid and DNA was estimated by diphenylamine reaction, as described previously (McCullough & John, 1972).
RNA isolation and estimation
RNA was extracted in the presence of RNase inhibitors and was estimated by measuring absorbance at 260 nm. Poly(A)+ RNA was purified by binding to oligo(dT)–cellulose and determined by measuring absorbance at 260 nm. Both RNA fractions were prepared as described previously (Lambe & John, 1979). Estimation of relative poly(A)+ RNA levels by hybridization with [3H]poly(U) followed the procedure of Fraser & Carter (1976).
Cell–free protein synthesis
The in vitro protein synthesizing system isolated from wheat germ was employed, as described by Roberts & Paterson (1973), with endogenous protein synthesis reduced below levels detectable by autoradiography, by prior treatment with Ca2+–dependent nuclease (Pelham & Jackson, 1976). The rate of protein synthesis was linear with time for 1 h and with up to 1μg of poly(A)÷ RNA added. In the assays performed for Figs. 3 and 4, 0·5 μg poly(A)+ RNA was added and incubation was for 1 h.
Electrophoresis
One–dimensional electrophoresis in the presence of sodium lauryl sulphate was performed in a 15% (w/v) acrylamide gel according to the method of Laemmli (197o).
Two–dimensional separation, with isoelectric focusing followed by electrophoresis in the presence of sodium lauryl sulphate, according to the method of O’Farrell (1975), was used to fractionate 20 μg samples of extracted protein for the study of in vivo protein synthesis.
RESULTS
A major advantage of preparing synchronous cultures by cell selection is that control experiments can be performed to test for the effects of synchronizing. Fig. 1 shows a control experiment in which asynchronously dividing cells, held at constant turbidity and illumination, were at time o subject to continuous–flow centrifugation, but at a reduced rotor speed so that cells at all stages of the cell cycle remained in the supernatant fraction. After centrifugation the cells were transferred into a smaller vessel, to mimic the treatment of selected cells, and were grown at the same constant illumination as in the larger parent culture. The similarity of environment before and after selection is reflected in the unaltered rates of increase in each class of macromolecules and in cell number (Fig. 1). There is therefore no evidence that changes in the rate of macromolecule accumulation are caused by the cell selection procedure.
The continuous–flow centrifugation procedure can be used to select a homogeneous population of small daughter cells, which form a synchronously dividing culture (Fig. 2) in which 5 phase is completed between 5 h and 10 h and subsequent release of daughter cells occurs between 12 h and 18 h. Cell–cycle events are shown 5h.
The conventional procedure for synchronizing division in algae by periodic illumination does not allow a discrimination between phenomena that are caused by the recurring interruption of growth and those that are part of the cell cycle. Therefore, the comparison that is possible here, between selected cells and control cells, reveals for the first time changes in the rate of macromolecule accumulation that are unambiguously part of the Chlorella cell cycle.
Protein accumulation accelerates throughout G1S and mitosis phases, as is revealed in the semilogarithmic Fig. 2 by the straight line describing the accumulation of protein between o and 12 h. Calculations from the data in Fig. 2 reveal that individual cells, in successive 3–h periods between o and 12 h, achieve the following progressively larger mean increments of protein (in pg); 1·2, i·8, 2·7, 3·3. Accelerating protein accumulation resumes in the subsequent cell cycle beginning at 15 h (Fig. 2). However, when cytokinesis occurs there is an interruption of protein accumulation, which is seen as a pause between 12 h and 15 h. This brief pause is invariably observed in selected cells and also in synchronized cells (Forde & John, 1973). There is evidence that the pause in accumulation is achieved by faster protein breakdown, since rates of incorporation into protein are reduced very little compared with earlier stages in the cycle. Relative to the preceding 5 and mitosis phases, rates of incorporation during the cytokinesis pause, are 93% and 89%, respectively, when measured with ®35SO4 and [3H]leucine, both supplied for 2·5 h period, and 95% when measured with [^S]methionine supplied in 3–h pulses (labelled proteins are described and illustrated in the legend to Fig. 5).
Although protein accumulation accelerates throughout the cell cycle, we consistently observed that total RNA accumulation is linear. Cells establish a rate of RNA accumulation at the beginning of G1 phase and maintain the same rate throughout the cycle, including the period at cytokinesis when protein does not accumulate. The two patterns are revealed in the semilogarithmic plot (Fig. 2), where protein shows a straight line for accumulation up to 1 2 h and again after cytokinesis from 15 h, while for RNA there is a curve, with a rate change at 18 h as daughter cells establish their rate of RNA increase early in G1 phase. In even stronger contrast with protein accumulation is the increase in poly(A) +; RNA, which is restricted to G1 phase. Fig. 2 shows the termination of poly(A) + RNA accumulation at 3 h, as cells commence 5 phase, and its resumption 9 h later in newly formed daughter cells. The accumulation of protein, by contrast, continues to accelerate through S and mitosis phases, between 3 h and 12 h, when the level of mRNA does not increase. Protein synthesis, measured by incorporation of 35SO4 (not shown) also increases smoothly during this period of stationary level of messenger. Poly(A) + mRNA must therefore be present in excess of the amount required for protein synthesis.
To establish whether Chlorella cells have the facility to change the composition of their mRNA population within the time–span of a cell cycle, cells previously grown photosynthetically in batch culture were incubated for a period of less than a third of a cell cycle, under a variety of nutritional conditions, and then their poly(A) +RNA was extracted and translated. Autoradiography of the protein products seen in Fig. 3 reveals that, although some mRNAs are constitutive, Chlorella has an ability to elevate or depress functional levels of many individual mRNAs. Some messengers are present at high levels during faster growth, whether photosynthetic or in heterotrophic growth on glucose. Particular modes of carbon nutrition also influence individual mRNAs and some are specific for inorganic carbon fixation in photosynthesis (Bassham, 1973), others for heterotrophic growth with either glucose or acetate, and others are specific for particular forms of organic carbon: either glucose, which allows glycolysis, or acetate, which requires glyconeogenesis (Syrett, Bocks & Merrett, 1964), while some are specifically repressed by glucose.
Chlorella cells can therefore alter their population of mRNAs when they are synthesizing the different enzymes required for different metabolic patterns (Syrett, Merrett & Bocks, 1963; Bassham, 1973). To test whether similar changes in mRNA are required for progress through the cell cycle, samples were taken from the synchronous cells analysed in Fig. 2. Exactly the same method of analysis showed no comparable change in the levels of individual mRNAs. Samples taken in G1phase at 3 h, in <S phase at 6 h and in cytokinesis at 12 h all show very similar mRNA populations (Fig. 4). The sample at 3 h was taken at the end of a period of poly(A)÷ RNA accumulation when the proportion of mRNA in total RNA was maximal, while the sample at 12 h was taken after a 9 h period during which there was no increase in mRNA level when the proportion of mRNA in total RNA was at a minimum. The similarity between these samples indicates that all the abundant translatable mRNA species follow the same periodic pattern of accumulation. We have not tested directly whether proteins showing the same electrophoretic mobility when coded by different samples of poly(A)+ RNA are the same proteins, but we consider it unlikely that messengers are consistently replaced throughout the cell cycle with others that code for different proteins with identical mobilities that occur in identical amounts.
To investigate whether changing populations of protein are synthesized throughout the cell cycle, successive 200–ml samples were removed from a synchronous culture identical to that described in the legend to Fig. 2 and the cells were pulse–labelled under the same growth conditions for one seventh of a cycle (3 h) with [35S]methionine. The cells were smashed at 0°C in a French pressure cell and the proteins were divided into an insoluble pelleted fraction and a post–ribosomal soluble fraction. Proteins in these fractions were subjected to one–dimensional electrophoresis*under dissociating conditions by the method of Laemmli (1970) and wére detected by autoradiography. Differences in radioactivity of less than twofold are not clearly detected by visual inspection of autoradiographs (Lutkenhaus, Moore, Masters & Donachie, 1979), and with this reservation more than 96% of the proteins in the insoluble fraction and 94% of the proteins in the soluble fraction were synthesized throughout the cell cycle. There is no evidence of periodic cessations of synthesis.
The absence of periodicity is not a failure of resolution since the same procedures, when applied to cells in periodic illumination, showed periodic synthesis in the majority of proteins throughout the cell cycle (John et al. 1981).
To test further whether periodicity of synthesis may have been masked by the co–migration of proteins in one–dimensional electrophoresis, proteins in the soluble fraction were subject to two–dimensional electrophoresis, by the method of O’Farrell (1975)· Examples of the more abundant proteins that were resolved are shown in Fig. 5. In the present study exposure of autoradiographs was terminated before the fainter spots had developed, to prevent fogging of the central area. Even with the improved resolution provided by two–dimensional separation, 93% of the quantitatively predominant proteins were seen to have been synthesized in all phases of the cell cycle. We think it unlikely that there is concealed periodic synthesis within this abundant class of proteins since, to replace one periodically synthesized protein by another without visible effect would require them to have identical iso–electric pH values and identical molecular weights, to be present in identical amounts and to be synthesized at inverse rates.
The periodic synthesis, which is noted here for 7% of the soluble proteins in selected cells, may be a normal part of the cell cycle that has been made evident in the population because of their synchrony, but may alternatively be part of recovery from the selection procedure. If selection did cause their periodic synthesis then non–synchronous cells would show the same periodicity if subjected to the same stress. The control asynchronous cells described in the legend to Fig. 1 were therefore pulse–labelled, exactly as the selected cells were, for the 24 h period after mock selection. In these cells the majority of proteins were synthesized continuously but each of the proteins found to be periodically synthesized in synchronous culture was also periodically synchronized in the control asynchronous culture. Some proteins were synthesized preferentially soon after mock selection, synthesis of others was initially depressed but recovered at 9 h and 12 h and some showed accelerations of synthesis as late as 18 h. Therefore, perturbations affecting a minority of proteins persisted for the whole of a period equal to one cell cycle after selection.
The simplest explanation for the periodicity seen in the small minority of proteins in synchronously dividing cells after selection, is that their periodicity is part of the same response to metabolic disturbance as is seen in control cells. From these data, we cannot exclude the possibility that some of these proteins might also show periodic synthesis in response to progress throughout the cell cycle. However, this possibility can be eliminated by a similar analysis in cells that were made synchronous by periodic illumination but then transferred to constant conditions in a turbidostat for study. In this procedure known metabolic perturbations subside during the first cell cycle (John et al. 1981) and protein synthesis can be studied by pulse–labelling during the second cell cycle. This method has the limitation that no control is available to test for possible perturbations persisting into the second cycle, but it has proved a valuable adjunct to the procedure of cell selection for the present study because all of the proteins that were synthesized periodically following selection by continuous flow centrifugation were found to establish continuous rates of synthesis through the second synchronous division in the turbidostat (M. Rollins & P. C. L. John, unpublished observations). Therefore, we conclude that no abundant soluble protein can be demonstrated to show periodic synthesis in the cell cycle.
DISCUSSION
Observations of macromolecule accumulation that have been made in synchronous cultures of Chlorella prepared by centrifugal selection are significant for two reasons. Firstly, because the cells were grown throughout at constant turbidity and so did not suffer the reductions in photosynthesis and in growth rate that occur in batch culture when increasing cell density causes increasing mutual shading. Secondly, because synchrony was obtained by a means that allowed the performance of control experiments to test for the effect of synchronizing procedures. Such controls are not possible for the conventional procedure of synchronizing with repeated cycles of growth interruption by darkening. When the effect of a dark period is tested on asynchronous cells, significant synchrony is produced and it is therefore uncertain whether subsequent periodicity in metabolism is due to perturbation or is a cell cycle phenomenon made apparent by the synchrony. In the present study the conditions of growth and procedures of cell selection were shown not to cause fluctuations in macromolecule accumulation (Fig. 1). Synchronous cultures, under the same conditions, revealed different patterns of increase throughout the cell cycle in several classes of macromolecule.
Protein accumulation accelerates smoothly throughout the cell cycle up to cytokinesis but there is then a pause (Fig. 2), as has been observed in batch culture (Forde & John, 1973)· Our detection of continued protein synthesis by incorporation of radioactive precursor indicates that the pause is achieved by an acceleration of protein breakdown. There are parallels with this pause in growth at the time of division, since in mammalian cells there is commonly a slowing of growth at mitosis due to reduced protein synthesis (Prescott & Bender, 1962), and in 5. pombe there is a cessation of volume increase, but not of protein synthesis, during cytokinesis (Mitchison, 1963); but in no case is it clear what contribution such pauses make to cell division.
The present data show that there is no obligate correspondence between the patterns of total RNA and protein accumulation through the cell cycle, although previous reports have shown that both total RNA and several enzyme activities accumulate linearly in S. pombe (Fraser & Moreno, 1976) and both total RNA and the rate of protein synthesis increase exponentially in 5. cerevisiae (Elliott & Mc Laughlin, 1978). We consistently observe in Chlorella that protein accumulation accelerates through G1 S and M phases, but in the same cells total RNA accumulation is linear and extends into cytokinesis. Although different patterns of accumulation for RNA and protein have not previously been reported, those seen in Chlorella are consistent with efficient utilization of ribosomal RNA through the cell cycle, since an increase in the rate of protein synthesis in proportion to the linearly increasing RNA could readily account for the accelerating accumulation of protein.
It is also revealed that the eukaryote cell cycle does not universally employ control of protein accumulation by controlling the level of poly(A) + mRNA. Correlations that are consistent with this possibility have been observed in 5. pombe, where a rise in poly(A)÷ RNA levels early in G2 correlates with faster accumulation of several enzymes (Fraser & Moreno, 1976), and in S. cerevisiae, where exponentially increasing poly(A) + RNA levels (Elliott & McLaughlin, 1979) correlate with exponentially increasing protein synthesis (Elliott & McLaughlin, 1978). However, in Chlorella, poly(A) +RNA accumulation is restricted to G1 phase (Fig. 2) and, although its level is constant through late G1, 5 and mitosis phases, the rate of protein accumulation continues to accelerate; therefore, at the end of its periodic accumulation poly(A) + RNA is present in excess.
An unsolved question about the cell cycle concerns the mechanisms by which events leading to division are initiated. Comparison with other developmental processes suggests that the accelerated synthesis of particular proteins could provide the molecular switches (Tuchman, Alton & Lodish, 1974; Linn & Losick, 1976). Some hypotheses of enzyme synthesis in the cell cycle, such as linear reading (Halvorson et al. 1971) and oscillatory repression (Donachie & Masters, 1969), account for a programme of periodic synthesis in a majority of proteins throughout the cell cycle. Periodic synthesis of proteins that might control division could therefore be part of a general programme. This possibility is not supported by our evidence of continuous synthesis of individual proteins (Fig. 5). Our evidence confirms a similar finding in S. cerevisiae (Elliott & McLaughlin, 1978) and extends that evidence by showing that mRNA populations can remain stable throughout the cell cycle (Fig. 4). The correlation between organisms that divide by such different mechanisms as budding and multiple fission, suggests that continuous synthesis of the majority of proteins is a common feature of the cell cycle.
Before the metabolic background to cell–cycle events can be understood, however, it is necessary to resolve the discrepancy between the numerous reports of periodic increase in enzyme activity (see reviews by Mitchison, 1969; Halvorson et al. 1971), and the evidence for continuous synthesis of proteins. Elliott & McLaughlin (1978) suggested that the discrepancy might be due to change in activation state; that is, changes in catalytic activity without a proportionate change in the amount of enzyme protein. This possibility would allow the cell cycle to include numerous changes in metabolic activity, in spite of continuous synthesis of proteins, and therefore it could contain numerous potential mechanisms for the initiation of cell cycle events. However, if this view is correct, sophisticated controls of enzyme activation state would be required (Mitchison, 1981) on a scale for which there has not been any evidence in cells that are simply growing and dividing.
The data presented here suggest a simpler explanation for the discrepancy between the periodic increase in enzyme activity and the continuous synthesis of protein. We have studied the accumulation of enzyme activity in the Chlorella cell cycle under a variety of conditions. Therefore patterns of protein synthesis can be compared with increase in enzyme activity under similar conditions in a way that is not yet possible in 5. cerevisiae.
Under the conditions of growth in batch culture with intermittent illumination, which readily induce synchronous division in unicellular algae, most enzyme activities are seen to increase discontinuously, e.g. in Chlamydomonas (Kates & Jones, 1967) and in Chlorella (John et al. 1973; Schmidt, 1974; Lorenzen & Hess, 1974). In the strain 211–8p, which was employed here, only glutamate oxaloacetate transaminase and nitrate reductase activities increased in proportion to total protein under synchronizing conditions (P. C. L. John, F. Haldane & B. Taylor, unpublished observations), while the following nine enzymes each increased in a single step at different times in the cycle; aconitase, cytochrome oxidase, phosphoenolpyruvate carboxylase, succinate dehydrogenase, ribulose–1,5–bisphosphate carboxylase (Forde & John, 1973), ribulose–5–phosphate kinase (John et al. 1981), isocitric dehydrogenase, fumarase and citrate synthase (John et al. 1980). Under the same synchronizing conditions, pulse–labelling with 35SO4 and with [3H]leucine revealed that different populations of proteins were also being synthesized at different times in the cell cycle (John et al. 1981). Similar changes in protein synthesis have been noted in Chlamydomonas under synchronizing conditions (Howell, Posakony & Hill, 1977). There is therefore a correlation between accumulation of enzyme activity and protein synthesis. However synchronizing conditions result in periodic starvation when cells are darkened, and there are consequent changes in metabolism (John et al. 1981) that affect enzyme synthesis.
To minimize environmental influence, the last six enzymes listed above have been measured in synchronously dividing cells, which either have been transferred to continuous illumination in a turbidostat and allowed to reach a balanced metabolic state (John et al. 1981) or have been made synchronous without exposure to periodic environmental changes by use of centrifugal selection, as in the present study (John et al. 1980). Each of the enzymes then increased exactly in proportion with total protein accumulation. In addition, glycollate dehydrogenase also increased with constant specific activity throughout the cell cycle in turbidostat culture. In comparison with the number of proteins that can be resolved, far fewer enzyme activities have been studied, but the evidence is consistent and we are not aware of any enzyme activity involved in general metabolism that increases out of proportion with cell growth in the cell cycle of Chlorella in the absence of environmental stimulus. There is again, therefore, a correlation between enzyme activity and the continuous synthesis of individual proteins that is described in this paper. Only if the continuous synthesis of individual proteins, which prevails during the cell cycle under conditions of balanced growth, is compared with the periodic increase in enzymes, which results when the environment changes the processes of metabolism throughout the cell cycle, is there an apparent discrepancy.
The periodic increases in enzyme activity, which are frequently observed in synchronous cultures of yeasts, may also have resulted from the use of starvation or sucrose density–gradient centrifugation to obtain synchrony. From the use of suitable controls with S. pombe, Mitchison’s group have shown (Mitchison, 1977) that of the 19 enzymes that have been studied only one shows a periodic increase in activity that is due to progress through the cell cycle, but similar controls for the effects of selection have not yet been reported for 5. cerevisiae. A few enzymes have been studied in 5. cerevisiae using zonal centrifugation to age–fractionate an entire cell population (Carter & Halvorson, 1973; Sebastian, Carter & Halvorson, 1973), but it is difficult to evaluate data from the unusually large and small cells at either end of the population distribution. The discrepancy in 5. cerevisiae between individual protein synthesis and the increase in activity may also be only an apparent one. Therefore, unless periodic increases in activity can be shown to be a common feature of the cell cycle in the absence of metabolic perturbation, it is unnecessary to consider periodic enzyme activation to be a frequent occurrence for the enzymes of general metabolism. The possibility remains that activation of a few key enzymes (Mitchelson, Chambers, Bradbury & Matthews, 1978) is important in the cell cycle.
We cannot conclude that every protein is synthesized continuously throughout the cell cycle. Many proteins could not be measured in the present study and these included any proteins that did not contain methionine, the large number that were insufficiently abundant to be detected by autoradiography, those with isoelectric pH values too extreme to focus with the majority of proteins, and the messenger population studies also excluded any mRNAs that lacked the usual poly(A) tail. For example, histone synthesis was not detected in this study for these last two reasons, but these basic proteins are synthesized periodically at the time of DNA replication (Marks, Paik & Borun, 1973). Therefore, the possibility remains that a minority of proteins is synthesized periodically and may drive the cell cycle.
We conclude that a simplified view of the cell cycle is now appropriate, in which a period of balanced growth, not involving an extensive programme of biochemical differentiation, precedes an eventual commitment to divide. This model is consistent with evidence that the period of cell growth prior to DNA replication can be readily shortened (Tyson, Garcia–Herdugo & Sachsenmaier, 1979) or extended in response to changes in cell size (Hartwell & Unger, 1977; Nurse & Thuriaux, 1977), nutrition (Lorincz & Carter, 1979) or hormone supply (Stiles, Cochran & Scher, 1981).
ACKNOWLEDGEMENTS
We thank the Science Research Council for support.