THE CELL CYCLE DURING THE VEGETATIVE STAGE OF DICTYOSTELIUM DISCOIDEUM AND ITS RESPONSE TO TEMPERATURE CHANGE

Transcription

1 J. Cell Set. 3a, 1-20 (1978) Printed in Great Britain Company of Biologists Limited 197S THE CELL CYCLE DURING THE VEGETATIVE STAGE OF DICTYOSTELIUM DISCOIDEUM AND ITS RESPONSE TO TEMPERATURE CHANGE IRENE M.ZADA-HAMES* AND J. M. ASHWORTH Department of Biology, University of Essex, Colchester, Essex, CO4 3SQ, England SUMMARY The cell cycle in amoebae of Dictyostelium discoideum has been analysed in cells growing asynchronously in axenic medium. For cells growing at the optimum growth temperature of 22 C with a culture doubling time of 8 h the average times for the cell cycle phases are as follows: G 1( i # 5 h; S, 2 - i h; G t, 4-4 h; M, 15-2 min. When amoebae are grown at temperatures below 22 C, culture doubling time increases and the cell cycle phases are altered in ways characteristic for each phase. G t is the most variable period and may occupy up to 70 % of the total cell cycle time; S and G x are the least affected, increasing by only 20% when the cell generation time is doubled. When cells which have reached the stationary phase of growth in liquid medium are washed and reinoculated into fresh medium they divide synchronously after a lag period of 5 h. By following cell number increase and nuclear DNA synthesis in these cultures we have shown that stationary phase cells are arrested in the G x phase of the cell cycle. Finally, although more than 97 % of amoebae grown on a bacterial food source are uninucleate, when grown axenically up to 35 % of the cell population may become multinucleate. Our results suggest that these cells probably arise through the failure of cytokinesis to follow karyokinesis. Multinucleate cells appear to have a slightly longer G, period than mononucleate cells. INTRODUCTION The cellular slime mould, Dictyostelium discoideum, is a soil amoeba which can be grown easily in the laboratory either axenically (Watts & Ashworth, 1970) or in association with bacteria (Sussman, 1966). Its life cycle is relatively simple, consisting of 2 distinct phases. During the vegetative stage the solitary amoebae feed, grow, and then divide by binary division. On exhaustion of nutrients the developmental phase is initiated during which the cells aggregate and go through a series of morphogenetic stages resulting in the formation of a fruiting body (Loomis, 1975). To date, very little work has been done on the cell cycle of D. discoideum during vegetative growth. Our present knowledge has been gained from induction synchrony studies using a temperature-sensitive mutant (Kate & Bourguignon, 1974). Since the latter stages of development in this mutant are somewhat abnormal we have attempted to use induction synchrony methods on the developmentally normal strain AX2. Despite numerous attempts to induce synchrony by a variety of means Author to whom all correspondence should be addressed, at the Department of Biochemistry, 9 Hyde Terrace, University of Leeds, Leeds, Yorkshire, England.

2 2 /. M. Zada-Hames and J. M. Ashtoorth we have failed to do so reproducibly; standard methods are either ineffective or result in partial synchrony which can be largely attributed to the induction and subsequent break-up of multinucleate cells (Yalovsky, Zada-Hames & Ashworth, unpublished results). However, as reported here, we have been able to determine the timing of the phases of the cell cycle in vegetative amoebae of D. discoideum using asynchronously growing cells which have the advantage that possible artifacts introduced by synchroni2ing techniques are absent (Mitchison, 1971). Using this technique we have examined the variation in the lengths of the cell cycle phases of D. discoideum growing at different temperatures with different specific growth rates. By pinpointing the exact stages in the cell cycle where changes in temperature or other environmental factors affect the growth rate, we may gain an insight into the events involved in the regulation of growth and how these relate to the developmental phase. We have also examined cells which have reached the stationary phase of cell growth and by observing their behaviour upon reinoculation into fresh medium shown that these cells are arrested in the G 2 phase of the cell cycle. MATERIALS AND METHODS Growth D. discoideum amoebae of strain AX2 were grown axenically in ml of HL-5 medium which contained per litre, Bacteriological peptone (Oxoid), 14-3 g; yeast extract (Oxoid), 7-14 g; Na 2 HPO 4.i2H,O, 1-28 g; KH,PO 4) 049 g; D-glucose, g; at a ph of 67. The flasks were shaken on an orbital shaker at 160 rev/min. To ensure that cells were in steady state of growth cultures were kept at the experimental temperature for 3 days prior to use and maintained in the exponential phase of growth (below 6 x 10* cells per ml). For cell cycle analysis cultures at 3-4 x io* cells per ml were routinely used, grown at temperatures from 17 to 22 C. When bacterially grown cells were required strains AX2 and NC4 were grown on agar plates in association with KlebsieUa aerogenes as described by Sussman (1966). Plating efficiency was determined as described by Sussman (1966). Counting and sizing of amoebae Amoebae were diluted 1:100 with isotonic saline (0-7 % (w/v) NaCl) and counted using a model Z B Coulter Counter with a ioo-/tm orifice tube. Sizing was done with a Coulter model P64 size distribution analyser attachment. Latex beads of known size were used for calibration. Labelling of cells Exponential phase cultures. From exponentially growing cultures at 3-4 x 10* cells per ml, two 20-m] samples were removed. [Me 3 H]thymidine (sp. act. 47 Ci/mmol, the Radiochemical Centre, Amersham) was added to one of these samples at a final concentration of 50 /ici/ml to specifically label those cells synthesizing DNA, whilst to the other was added an equivalent volume of sterile distilled water. Both cultures were then treated identically. The unlabelled culture serves as a control to monitor whether the presence of radioisotope has any effect on the growth kinetics of the cells. The cultures were shaken at the experimental temperature for 12 min, centrifuged at 250 g for 3 min, and the supernatant carefully decanted. Each amoebal pellet was washed 3 times in 'conditioned' medium (see below) which had been kept at the experimental temperature. After the final wash, each pellet was resuspended in ml of conditioned medium at a density of 1 i-6 x 10* cells per ml and shaken at the experimental temperature. To follow the fate of the pulse-labelled cells, samples

3 Cell cycle of D. discoideum 3 were removed at various times after labelling and prepared for autoradiography. A i-o-ml sample was removed at each point. This was centrifuged at 250 g for 3 min at room temperature and washed once in 3 ml of 0-4% (w/v) NaCl. The pellet was fixed and prepared for autoradiography as described below. Cell washing was done in hypotonic saline in order to produce cells which were slightly swollen. After fixation this resulted in intact cells in which nuclei could be readily distinguished. Water could not be used when preparing cells for observation of intact cells because this resulted in extensive cell spreading and rupture of cell membranes such that individual cells could not be distinguished. However, when preparing cells for chromosome observation water could be used for cell washing and resuspension, the increased cell spreading facilitating observation of mitotic figures. Conditioned medium was obtained by centrifugation from exponentially growing cultures which had been grown at the experimental temperature to the same cell density as the final cell density of the cells after labelling, i.e. I-I-6XIO* cells per ml. The supernatant was examined microscopically to ensure that all cells had been removed. All manipulations were carried out under sterile conditions. Stationary phase cultures. To detect resumption of nuclear DNA synthesis in stationary phase cultures that had been reinoculated into fresh medium [Me 'H]thymidine (sp. act. 47 Ci/mmol) was added either as a 12-min pulse label or as a continuous label in parallel cultures. For the continuous label a final concentration of 50 fici/ml was used in a total volume of 20 ml medium. For the pulse label, 0-5-ml samples were removed hourly from a 20-ml culture and mixed with 0-5 ml of medium containing radioisotope to give a final concentration of so/ici/ml. After 12 min the radioisotope was removed by washing twice with fresh medium and once with 0-4 % NaCl. From the continuously labelled culture, o-5-ml samples were removed hourly and washed in the same way. The cell pellets from the final washes were fixed and prepared for autoradiography as described below. Fixation of amoebae for autoradiography Fixation for intact cells. Cell pellets were resuspended in a few drops of 0-4 % NaCl and 1-2 ml of freshly prepared ethanol/glacial acetic acid (3:1, v/v) fixative was added drop wise, while vortexing. A few drops of this cell suspension were placed on to a dry methanol-cleaned glass slide and allowed to air dry without agitation. It was important not to centrifuge the cells through the fixative at any time; this resulted in rupture of the cell membrane and intact cells could not be observed. Fixation for chromosomes. Cell pellets were resuspended in a few drops of distilled water or 0-4% NaCl and 1-2 ml of freshly prepared ethanol/glacial acetic acid (3:1, v/v) fixative was added dropwise, while vortexing. Cell suspensions were centrifuged at 250 g for 3 min and the fluffy pellets resuspended in the small amount of fixative left after decanting the supernatants. A further 2-3 ml of fresh fixative were added dropwise, while vortexing end the samples centrifuged as before. The pellets were gently resuspended in about 0-5 ml of fixative and a few drops of cell suspension were spread rapidly across the surface of a dry methanolcleaned slide and allowed to air dry. Staining of slides To prevent cytoplasmic staining due to RNA obscuring nuclear and chromosome morphology, all slides were treated with ribonuclease (type i-a from bovine pancreas, Sigma) 0-2 mg/ml in M Sorensen's phosphate buffer, ph 6-8, at 37 C for 2 min and then washed in distilled water. Slides were stained in 10% (v/v) Gurr's Giemsa stain (Improved R66) as previously described (Zada-Hames, 1977). Slides for autoradiography were treated with RNase prior to dipping but were stained through the emulsion. Autoradiography Slides were dipped in Kodak NTB2 liquid emulsion, air dried in the vertical position, and stored at 4 C in the dark in the presence of silica gel. After an exposure time of 21 days the slides were developed at 20 C as follows: Kodak D-19 developer, 3 min; distilled water rinse; stop bath, 30 s; Kodak Rapid Fixer, 1-5 min; running tap water, 30 min; distilled water rinse. The slides were then air dried.

4 4 /. M. Zada-Hames andj. M. Ashworth Microscopic examination of cells Slides were systematically scanned with a Leitz Ortholux microscope using iooo x magnification and green bright-field illumination. Slides that had been prepared to demonstrate intact cells were scored for the percentage of cells labelled, the percentage mitotic index and the frequency of multinucleate cells. More than iooo cells were scored for each parameter for each time point. Slides that had been prepared for chromosome analysis were scored for the percentage of mitoses that were labelled and whether the labelled mitoses were mono- or multinucleate. The low mitotic index of these cultures necessitated scoring between 3000 and 4000 nuclei for each time point. Any nucleus with 5 or more silver grains above it was considered to be labelled. RESULTS Multinuclearity in vegetative cells When vegetative cells were fixed and stained to retain cellular integrity the number of nuclei per cell could be counted. In cultures of NC4 and AX2 strains grown in association with K. aerogenes more than 97% of the cells were mononucleate. The remaining 3% were mostly binucleate and probably represented cells in the last stages of mitosis just prior to daughter-cell separation. However, when AX2 was grown axenically in shaken liquid culture the percentage of mononucleate cells decreased to 65-80% and the remaining 20-35% consisted, in decreasing order of frequency, of bi-, tri- and tetranucleate cells (Table 1). More rarely, cells with 5-20 nuclei have been observed (Fig. 1). The proportion of cells that were multinucleate was variable and depended on the batch of peptone and yeast extract used, the composition of the medium, and the growth temperature. When glucose was omitted from HL-5 medium, the percentage of multinucleate cells increased. Similarly, when AX2 was grown at temperatures below the optimum of 22 C, the culture doubling time increased and so did the proportion of cells that were multinucleate (Table 1). Nuclei within any single cell entered and progressed through mitosis synchronously (Fig. ij). At the end of mitosis multinucleate cells often split up to give various combinations of mono- and multinucleate cells (Fig. 1K, L). The cell cycle in vegetative amoebae growing axenically at 22 C The optimum temperature for growth of D. discoideum in axenic medium was 22 C, giving cultures with a doubling time of 8 h. In cells growing asynchronously in exponential phase the mitotic period can be determined by scoring the percentage of the cells that are in mitosis (the mitotic index) at any one time. However, conversion of the fraction of a population in a certain phase to the duration of that phase requires an understanding of the cell age distribution. Because in an expanding, exponentially growing culture, a single 'old' cell becomes 2 young daughter cells as it passes through mitosis there are always twice as many young as old cells. Between these 2 extremes the frequency of cells of any particular age decreases exponentially. Therefore, because of the age distribution, the proportion of cells in any particular phase is not equal to the proportion of the generation time spent in

5 Cell cycle of D. discoideum 5 that phase. If the mitotic index is small we can use a standard equation to determine the mitotic time (Mitchison, 1971). Hence, t m = Mx 1-44x T. Where t m = the mitotic time, M = the mitotic index, and T = the total cell cycle time. Strain NC4 AX2 AX2 AX2 AX2» AXa» AX2«Table 1. Effect of growth conditions and temperature on the nuclearity of vegetative cells of D. discoideum Food source K. aerogenes K. aerogenes HL-s HL-s (no glucose) HL-s HL-s HL-s Growth temperature, Culture, doubling r C 22 time, h These determinations were carried extract different from those above IS % of cells with no. of nuclei per cell: out on cells grown on batches of peptone and yeast S S 4 0 O < > < 1 < 1 The mitotic index of cells growing at 22 C was 2-2%, giving a time of 15-2 min for the mitotic period. Chromosome appearance changed in the classical sequence (Fig. 2) except that at anaphase the daughter chromatids of individual chromosomes did not always separate synchronously (Fig. 2F). Moens (1976) in an electronmicroscope study of mitosis in D. discoideum has similarly observed that partner kinetochores of individual chromosomes do not always separate synchronously. The durations of the other phases of the cell cycle were analysed in asynchronous cultures by following autoradiographically the fraction of mitoses labelled as a function of time after a pulse label of [Me- 3 H]thymidine (Sisken, 1964; Cleaver, 1967). Labelled interphase and mitotic nuclei could be distinguished quite easily, despite the small size of D. discoideum chromosomes (Fig. 3). Any sudden change in cellular environment, such as centrifugation or change of medium, may induce delays in the progress of cells through the cell cycle resulting in a burst of synchronous mitoses that could affect the results of the analysis. To check for such perturbations the mitotic index was determined at all points used to construct the percentage labelled mitoses curve. Immediately following centrifugation to remove radioisotope the mitotic index fell to i-o% but recovered rapidly and by the subsequent sample at 1 h had returned to its previous value and then remained constant throughout the rest of the experiment (Fig. 4A). There was no lag in cell number increase. Since multinucleate cells represented 35% of the population a change in their proportions could produce apparent changes in the growth rate (as determined by cell counts) that were not due to true mitotic cell division. The percentage multinuclearity was monitored and remained constant at 35% throughout the analysis. < 1 < 1 < 1

6 6 1A /. M. Zada-Hames and J. M. Ashworth B _. C D H 10 //m Fig. i. Multinucleate cells of D. discoideum from axenically growing cultures. Scale bar (10 /tin) applies to all photographs, A-I, single cells containing from 1-20 nuclei, j, a tetranucleate cell in which all the nuclei are progressing through telophase synchronously. K, cytokinesis of an originally binucleate cell giving rise to four mononucleate daughter cells. Cytoplasm is not distributed evenly between the daughter cells. L, cytokinesis of an originally binucleate cell giving rise to 2 mononucleate and 1 binucleate daughter cells.

7 Cell cycle of D. discoideum 7 Pulse-labelled cultures multiplied with the same cell kinetics as control cultures which had been treated identically except for the omission of the radioisotope (Fig. 4B) and the population doubling time was the same during the experiment as it had been prior to the labelling period. Since plating efficiencies were greater than 95% dead non-cycling cells could be disregarded. On the basis of the above criteria the cells could be said to be in a steady state of growth. Theoretically, if there were no variations in cycle or phase times between individual cells the time for the first labelled mitotic cell to appear after the labelling would represent the duration of G 2 - However, in practice cells do not have identical cycle times and the variations in the time of G 2 causes the initial rise in the fraction of labelled mitoses to be more gradual than in the ideal case. The curve should reach a maximum when most of the cells which had been in S at the time of labelling enter mitosis. Ideally the curve should approach a maximum value of 100% labelled mitoses, but when G 2 is long compared to the rest of the cell cycle variations in cycle times will tend to lower the maximum level attained. Usually, the average S period is taken as the time between the 2 points on the curve where 50 % of the mitoses are labelled, whereas the time between the pulse label and the first of these points is the average G 2 period (Sisken, 1964). However, it should be mentioned that the value calculated for S from the 50% points on the curve can be difficult to interpret (Steel & Hanes, 1971), and that when the first peak of labelled mitoses does not reach 100% the value of S may bear little relation to the actual mean value of the duration. Nevertheless, the times estimated for the phase durations by this method are of value for comparative purposes. Since in this study the maximum level of labelled mitoses attained was less than 100% the points on the curve where 50 % of the peak value for labelled mitoses occurred rather than the points where 50% of the mitoses were labelled were used to determine the G 2 ~ and S-phase lengths. G x was determined by difference from the total doubling time. Fig. 4C shows the results of such a study on cells of AX2 grown axenically at 22 C with a culture doubling time of 8 h. No labelled mitoses appeared until 4 h after the pulse label and thus this represents the minimum time for G 2. The curve then rose to a peak value of 35 % at 5 h and descended to zero at 8 h (the time of completion of a whole culture doubling). The average cycle and phase lengths were thus shown to be: total cycle time, 8 h; G x, 1-5 h; 5, 2-1 h; G 2, 4-4 h. At each point on the labelled mitoses curve it was noted whether each labelled mitosis was mono- or multinucleate. In a preliminary account of this work (Zada- Hames & Ash worth, 1977) it was reported that in cultures growing at 24 C with a doubling time of 9 h, labelled multinucleate cells appeared slightly later than labelled mononucleate cells. This was found to be true also for cells growing at 22 C although the delay of multinucleate cells was not as great as that reported earlier for cells growing at 24 C (Fig. 4D). Effect of temperature on the cell cycle Cell cycle analysis was carried out on cells grown axenically at 19 C (culture doubling time 12 h) and 17 C (culture doubling time 15 h). In both cases there

8 8 2A /. M. Zada-Hames and J. M. Ashworth B 'jr*» 7 H fs- 2/im

9 Cell cycle of D. discoideum 3A B 10 //m Fig. 3. Autoradiographs of-d. discoideum cells and chromosomes showing localization of silver grains over nuclei and chromosomes. Scale bar (10 /tm) for both figures. A, a labelled mononucleate and an unlabelled binucleate cell, B, a binucleate cell showing labelling above prophase mitotic figures. were no growth perturbations during the analysis; labelled cultures multiplied with the same cell kinetics as control cultures, population doubling times during the experiment were identical to those seen before the labelling period, the mitotic indices and percentage multinuclearity remained constant throughout and plating efficiencies were greater than 95%. The results for all three growth temperatures are shown in Table 2. As for cells grown at 22 C, the G 2 period at 19 and 17 C occupies the greatest part of the cell cycle and is the phase most affected by changes in temperature, increasing by 141 % between 22 and 17 C when culture doubling Fig. 2. Appearance of D. discoideum chromosomes during mitosis. Scale bar (2 fim) applies to all photographs, A, interphase; chromosomes are completely decondensed. B, early prophase; chromosome condensation just beginning, chromosomes appearing as fine, continuous threads, c, early metaphase; chromosomes condensed and visible as 7 discrete units, each of which has a distinct banding pattern, D, late metaphase; chromosomes well condensed and showing daughter chromatid separation, E, synchronous anaphase; daughter chromatids have separated and are moving to opposite poles F, asynchronous anaphase in a binucleate cell; daughter chromatids of the individual chromosomes have not separated synchronously, resulting in chromatid lag. G, early telophase; daughter chromatids are at opposite poles; chromosomes still condensed but no longer visible as discrete units. H, late telophase; chromosomes have decondensed and appear very similar to interphase nuclei. Cytokinesis has not yet occurred.

10 10 I. M. Zada-Hames andj. M. Ashworth X <D D C o toti E a?? X "F ( - - _ ", * I I I I A B S/l "3 u c I. 1 i i c led 0).Q «tos E \ / V 1 T t \ */ \ J \ T f i i D led 09.Q _CD in (A o %mi _ - - m A /\ / \ / \ / \ L \ T *-" k - r - i T 8 Time, h / ^ y \ P \ ~i i Fig. 4. Cell cycle analysis of D. discoideum AX2 grown at 22 C with a culture doubling time of 8 h. The culture was pulse-labelled with [Me J H]thymidine, washed, and allowed to resume growth at o h in conditioned medium, A, the mitotic index in the pulse-labelled culture, B, growth kinetics of pulse-labelled (#) and control (O) cultures, c, appearance of labelled mitoses in all cells of the culture after the pulse label. The phase durations are represented by the lengths of the arrows between the points of 50 % of maximum mitotic labelling. C, culture doubling time; G t, the period after DNA synthesis and before mitosis; G x, the period after mitosis and before DNA synthesis; S, the period of DNA synthesis. D, appearance of labelled mitoses in mononucleate ( ) and multinucleate (O) cells.

11 Cell cycle of D. discoideum 11 time is almost doubled. At 22 C G 2 occupies 55 % of the total cell cycle time, whilst at 17 C it occupies 71 %. S and G x are the least affected by temperature, the proportion of the cell cycle they occupy decreasing with increasing generation time. In terms of absolute time both S and G x increase by only 20 % or so while there has been almost a doubling in total cell cycle time. As would be expected, the percentage of cells that are labelled during the 12-min pulse of phjthymidine decreases from 10-5% at 22 C to 6-5% at 17 C. Table 2. Effect of temperature on the cell cycle phases of D. discoideum Growth Culture, * v tern- doubling % S G % M* perature, time, total cell C h h cycle h % h % min % i ss i io Because of the methodology used to determine the values of G lt S and G, the time for mitosis is incorporated within these values. The mitotic period (M) is intermediate in sensitivity to changes in temperature, increasing by about 50% between 22 and 17 C. The absolute time for mitosis reaches a plateau level of about 23 min at temperatures between 19 and 17 C but whether this holds true at lower temperatures and thus represents a maximum for the mitotic period is not known. It is also noteworthy that whereas both S and G x increase linearly as doubling time increases (but decrease in terms of the proportion of the total cell cycle time they occupy), G a increases in a non-linear way, occupying proportionately more of the total cell cycle time (Fig. 5). Although G 2 increases continuously in terms of absolute time as doubling time increases the proportion of the cell cycle it occupies tends to level off at around 70 % in slow-growing cells. Use of conditioned medium In all the cell cycle analysis experiments conditioned medium from cells grown at the same temperature as that under examination, was used to resuspend cells after pulse labelling. This was found to be very important since when cells growing at 17 C with a population doubling time of 15 h were washed and placed into fresh (not conditioned) medium at 17 C, the doubling time was, rather surprisingly, reduced to 8 h. Cell cycle analysis of these cultures showed that instead of a G 2 period characteristic of cells growing at 17 C with a doubling time of 15 h, the G 2 period was reduced to that characteristic of cells grown at 22 C with a doubling time of 8 h. The S period remained the same as that in cultures grown at 17 C with a doubling time of 15 h. The mitotic index in fresh medium also changed. Whereas prior to transfer to fresh medium it had been 1-7%, giving a value of 22 min for the mitotic period, after transfer to fresh medium at 17 C it increased

12 12 /. M. Zada-Hames and J. M. Ashzoorth <B 6 ; i E CD blin 3 o D :al 0 "o I 4 co 1 3 o o 2 2 o "o a? Culture doubling time, h Culture doubling time, h 16 Fig. 5. Change in cell cycle phase lengths with changing culture doubling time. A, change in G t ( ), S (O), and G± ( ) in terms of absolute time (h). B, change in mitosis in terms of absolute time (min). c, D, changes in the proportions of the total cell cycle time that the various phases occupy. Symbols as for A and B. to 3-2%. For an 8-h cell cycle this gives a value of 22-1 min for mitosis, i.e. identical to that seen in the pre-transfer culture. These results suggest that although the timing of S phase and mitosis is unaltered in fresh medium, G 2 can be shortened, and that in conditioned medium there may be certain factors which regulate the length of the G a period or the rate at which cells pass through it and into mitosis. Cell cycle arrest of stationary phase cultures When Dictyostelium amoebae deplete the medium of one or more nutrient components they enter a stationary phase in which no further increase in cell number occurs. The cells secrete a growth inhibitor which prevents stationary phase medium from supporting the growth of freshly inoculated cells even if it is supplemented

13 Cell cycle of D. discoideum 13 with all the components of HL-5 medium (Yarger, Stults & Soil, 1974). It has been suggested that cells in the stationary phase of growth accumulate at a stage late in the cell cycle (Soil, Yarger & Mirick, 1976). The following experiments were carried out to determine whether this is indeed the situation. If the cells are blocked in G ± phase of the cell cycle, one should see DNA synthesis before cell division in reinoculated cultures, but if the cells are blocked in G 2 phase of the cell cycle, one should see cell division before DNA synthesis CD o 1-6 x CD D 12 = o Time, h Fig. 6. Synchronous division of stationary phase cells after reinoculation into fresh medium (HL-5).» cells/ml, x icr*; O» % mitotic index. Cells which had been in stationary phase at a density of 3-6 x io 7 per ml at 22 C for h were washed twice with sterile distilled water and resuspended in fre3h HL-5 medium at a density of 1-7 x io 6 per ml at 22 C. Two parallel cultures were set up. Nuclear DNA synthesis was detected autoradiographically by means of both pulse and continuous labels of phjthymidine as described in Methods. Cell number and mean cell volume were also monitored. There was no increase in cell number during the first 5 h after reinoculation of cells into fresh medium but this was followed by a period of rapid cell division between 5 and 8 h during which cell number almost doubled before remaining constant for the next 3 h. At 11 h cell number began to increase again (Fig. 6). The 3 CEL 32

14 14 I. M. Zada-Hames andj. M. Askworth period of rapid cell division was accompanied by a rapid increase in mitotic index from a value of below 0-2 % at 4 h to a peak value of 2-2% at 6-7 h (Fig. 6). It should be mentioned that the rather extensive washing required to remove radioisotope prior to the preparation of autoradiographs resulted in the mitotic index being lower than would otherwise be expected; in cultures fixed immediately after sampling the peak mitotic index has been found to be as high as 5%. It is interesting to note that during the period of rapid cell division there was not a complete doubling in cell number, the increase being 84%. Cells were examined during the experiment to see whether the percentage of cells that were multinucleate was changing. Fig. 7 shows that from o to 4 h, 83 % of the cells were mononucleate but that during the period of rapid cell division the percentage decreased to %, i.e. a decrease of 14-15%. Conversely, the percentage of binucleate cells increased by 14-15% over this period. If one calculates the increase in the number of nuclei rather than cells over the period of cell division the result is that there is a 103% increase, i.e. a complete doubling in numbers of nuclei. The increased frequency of binucleate cells which accompanies the period of cell division suggests very strongly that in axenic cultures, multinucleate cells arise in the main from the failure of cytokinesis to follow karyokinesis. The presence of multinucleate mitotic figures suggests that cytokinesis is not just delayed in axenic cultures but fails to occur completely until the next mitosis. Since the proportion of cells that are multinucleate remains fairly constant in exponentially growing cultures there must be a continual break-up as well as formation of these cells. As mentioned above cytological evidence confirms this (Fig. 1K, L). After the period of rapid cell division the mean cell volume of the culture was halved, appearing bimodal while the cells were dividing (data not shown) in agreement with the work of Soil et al. (1976). Analysis of the DNA-synthetic activity showed that very few cells synthesized nuclear DNA during the first 3-4 h in fresh medium but that there was then a rapid rise in synthetic activity, reaching a maximum at 7 h (Fig. 8). At this time 50% of the total culture had begun or completed synthesizing DNA and 22% of the cells at that time were actually involved in DNA synthesis. The peak of DNAsynthetic activity coincided almost exactly with the peak of mitotic activity. However, the mitotic activity fell off before the DNA-synthetic activity, reaching a minimum value (zero activity) at 9 h, whereas the DNA-synthetic activity did not reach a minimum value until h. Indeed, in the presence of a continuous label of PHJthymidine 100% of the cells were not labelled until n-i2h. Furthermore, whereas the second increase in mitotic activity began at 9-10 h, the second wave of DNA synthesis did not begin until 11-12I1. Therefore DNA synthesis followed division in these cells. The first and second periods of cell division were only 6 h apart in contrast to 8 h, the known culture doubling time of cells growing under these conditions and at this temperature. The results above suggest that this was because the first G 1 period was very short or absent. The latter is probably correct because in the autoradiographs prepared from the pulse-labelled cultures, labelled telophase figures

15 Cell cycle of D. discoideum 15 could be seen during the first period of cell division, suggesting that DNA synthesis Could begin before the chromosomes had fully decondensed. By the second cell cycle the G x period was restored to that found in cells growing with a doubling time of 8 h and there was a lag of about 1-2 h between mitosis and the onset of nuclear DNA synthesis. 100 i a c oco 40 -! Time, h Fig. 7. Change in nuclearity of cells after reinoculation of stationary phase cells into fresh medium (HL-5). 0, mononucleate; Q, binucleate cells. Periods of rapid cell division occurred between 5 and 8 h, and between 11 and 13 h. Since in the second cycle DNA synthesis clearly followed cell division we can conclude that stationary phase cells must have been arrested in the G 2 phase of the cell cycle. Furthermore, examination of the autoradiographs from the continuously labelled cultures showed that the metaphase figures in the first wave of cell division were not labelled, confirming that these cells had not gone through S phase before dividing, i.e. the stationary phase cells could not have been arrested in G v In the experiments described above, cells were harvested after they had been in stationary phase for only h. Upon reinoculation into fresh medium, synchronous cell division and DNA synthesis occurred. However, if cells are left in stationary phase for 40 h or more the release from growth inhibition is less synchronous.

16 16 I.M.Zada-Hames andj.m.ashworth Thus, although the initial lag period where cell number remains constant is still about 5 h, the period of cell division may take 7-8 h, suggesting that the longer cells remain at the G z block the longer it takes them to reinitiate cell division. A similar phenomenon has been observed in stationary phase cultures of E. colt (Cutler & Evans, 1966). 24 r "5 o I Time, h Fig. 8. DNA synthesis in stationary phase cells washed free of stationary phase medium and reinoculated into fresh medium (HL-5). Cells were grown either in the continuous presence of [Me *H]thymidine (#) or given a 12-min pulse label ( ) prior to preparation of autoradiographs. Periods of rapid cell division occurred between 5 and 8 h, and between 11 and 13 h. DISCUSSION In this paper we have defined the cell cycle in exponential-phase axenic cells of D. discoideum. The use of asynchronously growing cells has enabled us to do this in the absence of any artifacts that could be introduced by attempts to synchronize

17 Cell cycle of D. discoideum 17 cells by induction. The cell cycle as defined here differs from that described by Katz & Bourguignon (1974), the greatest difference being in the estimate of the mitotic period. Using the time taken for doubling of cell number in synchronous cultures they estimated that mitosis occupied 1-2 h of an 8-10 h cell cycle. Since individual cells probably exhibit considerable variation in the time required to resume growth and cell division after induction of synchrony, this probably represents an overestimate. Our value of 15 min is in agreement with the recent work of Robson & Williams (1977) who found the mitotic index of cells growing axenically at 22 C to be 2%, and with earlier work using time-lapse photography on bacterially grown cells, where it was observed that the time taken for a single amoeba of D. discoideum to round up and subsequently divide into 2 daughter amoebae was between 3-10 min, representing 1-4% of a total cycle time of 4 h (Ross, i960; Bonner, i960). The G 2 period occupies the greatest part of the cell cycle under all the conditions tested. At 22 C it represents 55 % of the total cell cycle time and this is in contradiction to that predicted by Leach & Ashworth (1972). On the basis of DNA measurements of bacterially and axenically grown amoebae they concluded that in axenic medium amoebae must be largely in the G x phase of the cell cycle and that the faster doubling time seen in bacterially grown cultures was due to a contraction of the G x phase. Since we find that G x represents only 19% of the total cell cycle in axenically grown cells and there is a 2-fold difference in growth rate between axenically and bacterially grown cells, this cannot be the explanation for the differences in DNA contents that Leach & Ashworth (1972) found. It is not unprecedented that G x is so short in D. discoideum since in other lower eukaryotes G x can also be very short, e.g. in Schizosaccharomyces pombe (Mitchison & Creanor, 1971) or absent altogether, e.g. in Amoeba (Prescott & Goldstein, 1967), Physarum (Nygaard, Gflttes & Rusch, i960; Mohberg & Rusch, 1969) and the micronuclei of Euplotes and Tetrahymena (Flickinger, 1965). In addition to occupying the greatest part of the cell cycle, the G 2 phase is also the phase of the cell cycle that is the most sensitive to temperature change. G x and S are the least affected, changing very little even when culture doubling time doubles. There have been many reports on the response of cell cycle times to variations in temperature and although in higher animals and plants most workers have found that temperature affects all phases of the cell cycle, it is frequently G x that is the most sensitive (Sisken, Morasca & Kibby, 1965; Wimber, 1966; Watanabe & Okada, 1967). The results of studies on the lower eukaryote Tetrahymena are rather contradictory, one group of workers finding that all phases of the cell cycle occupy the same percentage of the cell cycle at different temperatures (Mackenzie, Stone & Prescott, 1966) and another that G x is the most sensitive to variations in temperature (Cameron & Nachtwey, 1967). Our results show that axenic cultures contain a significant proportion of multinucleate cells and we have presented evidence that the majority of these multinucleate cells probably arise through failure of cytokinesis to follow karyokinesis, but from genetic evidence it is clear that a very small proportion ( < io~*) must arise through cell and nuclear fusion (see Jacobson & Lodish, 1975). On rare occasions cell and

18 18 I. M. Zada-Hames and J. M. Ashtoorth nuclear fusions have been observed microscopically (Huffman, Kahn & Olive, 1962; Huffman & Olive, 1964). Since multinucleate cells represent a considerable proportion of the total cell population in axenic cultures their frequency should be monitored during studies involving cell counting and growth kinetics. Their presence, probably accounts for the great heterogeneity in cell size (larger than could be explained by volume changes during the cell cycle) that have been reported for axenic cultures (Soil et al. 1976). It should be noted that when nuclear constituents are measured on a per cell basis the value obtained will not represent the value for a single haploid nucleus but be an overestimate. Our results show that multinucleate cells may take slightly longer to traverse G 2 than mononucleate cells. Several workers using other cell systems, have found that multinucleate cells have slightly longer cycle times than mononucleate cells in the same population, (Gime'nez-Martfn, Gonzalez-Ferndndez & Lo"pez-Saez, 1966; Gonzalez-Fernandez, L6pez-Sdez & Gime"nez-Martm, 1966). However, since in the percentage labelled mitoses curve we have reported here labelled multinucleate cells are not late in appearing in the second wave of labelled mitoses it may be that their total cycle time is the same as that of mononucleate cells but that G 2 and G x are longer and shorter respectively. We have demonstrated that when amoebae of D. discoideum reach the stationary phase of growth they become arrested in the G 2 phase of the cell cycle. This is in agreement with the results of Soil et al. (1976), who on the basis of measurements of cell volume, dry weight, protein and DNA content, concluded that stationary phase cells accumulate at a stage late in the cell cycle. When one looks at the data of Soil et al. (1976) (their fig. 4) it is interesting to note that, as reported in this paper, not all synchronous divisions achieve a complete doubling in cell number. Although many cell types in vivo and vitro arrest in the G x phase of the cell cycle (Bender & Prescott, 1962; Nilausen & Green, 1965) others are known to arrest in G 2, for example the mouse ear epidermis (Gelfant, 1962). In higher plants dormant seeds and starved meristems can arrest in either G 1 or G 2 in proportions that differ between species and are characteristic for each species, opening up the possibility that cell populations within a species are genetically determined to arrest in certain cycle periods (Van't Hof, 1974). In D. discoideum the near or complete absence of a G ± period in released stationary phase cells suggests that preparations for DNA synthesis may be able to commence in the G 2 phase of the cell cycle and that certain important 'trigger' events are located within this phase. Likewise, the great sensitivity of G 2 to changes in growth rate indicates that rate-limiting steps important in the regulation of growth may be situated within this phase. We thank the S.R.C. forfinancialsupport and Dr B. D. Hames for critical reading of this manuscript.

19 Cell cycle of D. discoideum 19 REFERENCES BENDER, M. A. & PRESCOTT, D. M. (1962). DNA synthesis and mitosis in cultures of human peripheral leukocytes. Expl Cell Res. 27, BONNER, J. T. (i960). Development in the cellular slime molds: the role of cell division, cell size, and cell number. In Developing Cell Systems and Their Control, 18th Growth Symp. (ed. D. Rudnick), pp New York: Ronald Press. CAMERON, I. L. & NACHTWEY, D. S. (1967). DNA synthesis in relation to cell division in Tetrahymena pyriformis. Expl Cell Res. 46, CLEAVER, J. E. (1967). In Thymidine Metabolism and Cell Kinetics, pp Amsterdam: North-Holland Publishing. CUTLER, R. G. & EVANS, J. E. (1966). Synchronization of bacteria by a stationary phase method. J. Bad. 91, FLICKINGER, C. ]. (1965). The fine structure of the nuclei of Tetrahymena pyriformis throughout the cell cycle. J. Cell Biol. 27, GELFANT, S. (1962). Initiation of mitosis in relation to the cell division cycle. Expl Cell Res. 26, GIMENEZ-MART/N, G., GONZALEZ-FERNANDEZ, A. & LOPEZ-SAEZ, J. F. (1966). Duration of the division cycle in diploid, binucleate and tetraploid cells. Expl Cell Res. 43, GONZALEZ-FERNANDEZ, A., L6PEZ-SAEZ, J. F. & GiMENEz-MARTfN, G. (1966). Duration of the division cycle in binucleate and mononucleate cells. Expl Cell Res. 43, HUFFMAN, D. M., KAHN, A. J. & OLIVE, L. S. (1962). Anastomosis and cell fusions in Dictyostelium. Proc. natii. Acad. Sci. U.S.A. 48, HUFFMAN, D. M. & OLIVE, L. S. (1964). Engulfment and anastomosis in the cellular slime molds (Acrasiales). Am. J. Bot. 51, JACOBSON, A. & L0DI8H, H. F. (1975). Genetic control of development of the cellular slime mold, Dictyostelium discoideum. A. Rev. Genet. 9, KATZ, E. R. & BOURGUIGNON, L. Y. W. (1974). The cell cycle and its relationship to aggregation in the cellular slime mold, Dictyostelium discoideum. Devi Biol. 36, LEACH, C. K. & ASHWORTH, J. M. (1972). Characterization of DNA from the cellular slime mould Dictyostelium discoideum after growth of the amoebae in different media. J. molec. Biol. 68, LOOMIS, W. F. (1975). Dictyostelium discoideum: A Developmental System. New York and London: Academic Press. MACKENZIE, T. B., STONE, G. E. & PRESCOTT, D. M. (1966). The durations of G u S and G 3 at different temperatures in Tetrahymena pyriformis HSM. J. Cell Biol. 31, MITCHISON, J. M. (1971). In The Biology of the Cell Cycle, pp Cambridge University Press. MITCHISON, J. M. & CREANOR, J. (1971). Further measurements of DNA synthesis and enzyme potential during cell cycle of fission yeast Schizosaccharomyces pombe. Expl Cell Res. 69, MOENS, P. B. (1976). Spindle and kinetochore morphology of Dictyostelium discoideum. J. Cell Biol. 68, MOHBERG, J. & RUSCH, H. P. (1969). Growth of large plasmodia of the myxomycete Physarum polycephalum. J. Bact. 97, NILAUSEN, K. 8c GREEN, H. (1965). Reversible arrest of growth in G x of an established fibroblast line (3T3). Expl Cell Res. 40, NYGAARD, O. F., GUTTES, S. & RUSCH, H. P. (i960). Nucleic acid metabolism in a slime mold with synchronous mitosis. Biochim. biophys. Ada 38, PRESCOTT, D. M. & GOLDSTEIN, L. (1967). Nuclear-cytoplasmic interaction in DNA synthesis. Science, N.Y. 155, ROBSON, G. E. & WILLIAMS, K. L. (1977). The mitotic chromosomes of the cellular slime mould Dictyostelium discoideum: a karyotype based on giemsa banding. J. gen. Microbiol. 99, Ross, I. K. (i960). Studies on diploid strains of Dictyostelium discoideum. Am. J. Bot. 47,

20 20 I. M. Zada-Hames and J. M. Ashworth SISKEN, J. E. (1964). Methods for measuring the length of the mitotic cycle and the timing of DNA synthesis for mammalian cells in culture. In Methods in Cell Physiology, vol. 1 (ed. D. M. Prescott), pp New York: Academic Press. SISKEN, J. E., MORASCA, L. & KIBBY, S. (1965). Effects of temperature on the kinetics of the mitotic cycle of mammalian cells in culture. Expl Cell Res. 39, SOLL, D. R., YARGER, J. & MIRICK, M. (1976). Stationary phase and the cell cycle of Dictyostelium discoideum in liquid nutrient medium. J. Cell Sci. 20, STEEL, G. G. & HANES, S. (1971). The technique of labelled mitoses: analysis by automatic curve-fitting. Cell Tiss. Kinet. 4, SUSSMAN, M. (1966). Biochemical and genetic methods in the study of cellular slime mold development. In Methods in Cell Physiology, vol. 2 (ed. D. M. Prescott), pp New York and London: Academic Press. VAN'T HOF, J. (1974). Control of the cell cycle in higher plants. In Cell Cycle Controls (ed. G. M. Padilla, I.L.Cameron & A.Zimmerman), pp New York and London: Academic Press. WATANABE, I. & OKADA, S. (1967). Effects of temperature on growth rate of cultured mammalian cells (L5178Y). J. Cell Biol. 32, WATTS, D. J. & ASHWORTH, J. M. (1970). Growth of myxamoebae of the cellular slime mould Dictyostelium discoideum in axenic culture. Biochem. J. 119, WIMBER, D. E. (1966). Duration of the nuclear cycle in Tradescantia root tips at three temperatures as measured with H'-thymidine. Am. jf. Bot. 53, YARGER, J., STULTS, K. & SOLL, D. R. (1974). Observations on the growth of Dictyostelium discoideum in axenic medium: evidence for an extracellular inhibitor synthesized by stationary phase cells. J. Cell Sci. 14, ZADA-HAMES, I. M. (1977). Analysis of karyotype and ploidy of Dictyostelium discoideum using colchicine-induced metaphase arrest. J. gen. Microbiol. 99, ZADA-HAMES, I. M. & ASHWORTH, J. M. (1977). The cell cycle during the growth and development of Dictyostelium discoideum. In Development and Differentiation in the Cellular Slime Moulds, Developments in Cell Biology, vol. 1 (ed. P. Cappuccinelli & J. M. Ashworth), pp Amsterdam: Elsevier/North-Holland Publishing (Received 27 September 1977)