Relationship between the timing of dna replication and the developmental competence in Acanthamoeba castellanii

The unicellular eukaryote Acanthamoeba castellanii has the option of two pathways of cell-type generation. Trophozoites are generated in the mitotic division cycle, whereas cysts are generated in the differentiation pathway. The relationship between particular cell cycle phases and differentiation into cysts has been investigated recently (Stbhr et al. 1987) and may be summarized as follows. In trophozoites flow fluorometric measurements of nuclear DNA content and determination of bivariate DNA/BrdUrd distribution indicate a lack of Gi phase nuclei and a cell cycle start with a short S phase (about 0-5 h) followed by a long G₂ phase (about 7h), mitosis and cytokinesis. When cells of the growth stage are transferred into non-nutrient medium, only 7% of cells encyst, whereas 50 to 70% of cells of a stationary stage culture develop into cysts. Synchronization of cells by release from the stationary stage reveals that encystation is initiated at a particular time in late G₂ phase. This result suggests that in contrast to growing cells, the high encystation competence of stationary stage cells is due to arrest of cells at this particular point in G₂ phase. These findings were obtained with cells cultivated in a nondefined medium containing proteose peptone and yeast extract (ND cells). In contrast, cells that have been adapted to a chemically defined medium containing 11 amino acids (D cells) exhibit a low encystation competence. On the basis of the results obtained with ND cells, it was suggested that the low encystation competence of D cells is due to a loss of accumulation of stationary stage D cells at the particular position of G₂ phase. This suggestion is supported by the finding that the nuclear DNA content of stationary stage D cells is about 15% lower than in G₂ phase cells. Furthermore, growing D cells indicate altered nuclear states, since in 10-20% of cells, an increase of up to 40% of the G₂ phase DNA content can be observed.

In an effort to study a possible relationship between the timing of DNA replication and developmental competence, this study examines more closely the dynamics of nuclear DNA synthesis in Acanthamoeba cultivated in a chemically defined medium. The results show that in D cells DNA synthesis takes place not only immediately after completion of mitosis as in ND cells, but also before mitosis. The occurrence of premature DNA replication is related to the inconstancy of nuclear DNA content as well as to the low competence of development in these cells.

Cultivation of Acanthamoeba castellanii (Neff strain) in a nondefined yeast extract/proteose peptone medium (ND cells) or in a chemically defined medium (D cells), and encystation in non-nutrient medium, were carried out as described (Jantzen, 1974; Jantzen & Schulze, 1987). In order to obtain synchronously growing ND cultures, standard stationary stage cells were taken 3 days after the termination of growth and diluted out into fresh non-defined nutrient medium (Stôhr et al. 1987).

Labelling of cells by concomitant addition of BrdUrd and fluorodeoxyuridine (FdUrd), preparation of cells for flow cytometric analysis by staining of fixed cells with fluorescein (FITC)-conjugated anti-BrdUrd and propidium iodide (PI), and the flow cytometric measurements of bivariate DNA/BrdUrd distribution were carried out according to Dolbeare et al. (1983) as described (Stôhr et al. 1987). Analysis of stained cells by fluorescence microscopy revealed that the FITC label was located exclusively and the PI label mainly within the nuclei. Cells not pulsed with BrdUrd/FdUrd, but stained with anti-BrdUrd/PI showed no anti-BrdUrd staining.

Nuclei were prepared by the Triton lysis method as described by Stôhr et al. (1987). For conventional DNA preparation, 5 ×10⁷ nuclei were digested with 100 tig proteinase K in 1ml 45 mM-EDTA, 80 mM-NaCl, 0 ·5% SDS (pH 7 ·8) for 2h at 37°C. After extraction with phenol and chloroform/isoamyl alcohol and DNA precipitation, the DNA was incubated in the presence of deoxyribonuclease-free ribonuclease A (100 μgml⁻¹) in 80mM-NaCl, 5 mM-EDTA (pH7-5) for 0-5h at 37°C, and thereafter in the presence of proteinase K (SOjUgml⁻¹) for 0-5 h. Following extraction with phenol/chloroform/isoamyl alcohol, the DNA was precipitated in the presence of 0 ·2M-sodium acetate with 2 ·5 vol. ethanol.

DNA was also prepared by embedding the cells or nuclei in agarose blocks and then incubating the blocks with proteinase K (Schwarz & Cantor, 1984). Cells or nuclei were suspended in PBS (0 ·15M-NaCl, 0 ·05M-Na₃PO₄(pH 7·2)) and mixed with an equal volume of a 1% solution of low melting agarose in PBS at 42°C. The mixture was immediately poured into a slot former so that an 80 id agarose block contained 2 ×10⁶ cells or nuclei. Blocks were incubated for 48 h at 55 °C in 1% N-lauryl sarcosine, 0 ·5M-EDTA (pH8) with proteinase K (2mgml⁻¹), and then stored in 0 ·5 M-EDTA (pH 8). For enzyme digestion, blocks were rinsed in TE (10mM-Tns-HCl, 1 mM-EDTA (pH 8)) and incubated twice in TE containing 0-04 mg ml⁻¹ phenylmethylsulphonyl fluoride for 30 min at 55 °C. For cleavage of DNA with restriction endonucleases, blocks were washed in TE and a 40 μl block was incubated with 17 units of enzyme for 4 h in 200 μ1 restriction enzyme buffer as recommended by the suppliers. A 20 μl block containing 0 ·5 ×10⁶ cells or nuclei was placed into a mixture of 0’5 × TBE gel running buffer (45mM-Tris base, 45mM-boric acid, 1 ·25 mM-EDTA) and 0-1 vol. of 10 × dye solution (50% (w/v) Ficoll, 0-2% Bromphenol Blue, 400 mM-EDTA) and then loaded into a slot of a 0 ·3% agarose gel. Electrophoresis was performed at 25 V for 14 h and 40 V for 8h. Following electrophoresis, gels were stained for 30 mm in 0 ·2 μg ml⁻¹ ethidium bromide and photographed. DNA samples prepared either by the lysis or by the agarose block method contained about 0 ·5 μg DNA (the nuclear DNA content of Acanthamoeba is 1 ·2pg cell⁻¹).

Growing D and ND cells were labelled with BrdUrd, fixed and subsequently stained with FITC-conjugated anti-BrdUrd to show nuclear DNA synthesis and with PI to show nuclear DNA content. Analysis by flow cytometry yielded bivariate DNA/BrdUrd distributions. When cells are labelled for 0 ·5 h, contour plots indicate a small FITC/PI-labelled cohort that exhibits a lower DNA content than the major FITC-negative, Pl-positive cell population (Fig. 1). From similar results it has been concluded that the S phase in both D and ND cells takes place immediately and exclusively after completion of mitosis (Stôhr et al. 1987). However, BrdUrd/FdUrd labelling of growing cells for extended time periods reveals obvious differences between bivariate DNA/BrdUrd distributions in D and ND cells. These differences may be described qualitatively as follows: First, the rate of increase of FITC-labelled cohort per labelling time is lower in D cells than in ND cells. Second, in D cells, especially after a prolonged period of BrdUrd/FdUrd labelling, FITC fluorescence is found in a first small cohort, in a second major cohort and in a third smaller population. In ND cells, however, the third FITC-positive cohort seems to be either absent or very small. The first small FITC-positive cohort represents postmitotic, DNA-synthesizing cells and the second represents cells that have finished their S phase. The third cohort, clearly seen in D cells, represents premitotic cells in which DNA synthesis is, nevertheless, taking place. This interpretation of bivariate DNA/BrdUrd distributions is supported by the following observations. Out of the three FITC-positive cohorts in D cells and the two in ND cells, only the second FITC-positive cohort increases during the labelling period, whereas the FITC-negative cohort concomitantly decreases. Hence, this cohort should represent cells that have finished their S phase. In this case, the distribution of the second FITC-positive cohort (representing G₂ phase cells that have progressed through the S phase) should coincide with the distribution of the FITC-negative cohort (also representing G₂ phase cells that, however, have not entered S phase). This is obviously not the case in D cells. However, when DNA synthesis in growing BrdUrd-labelled D cells is inhibited by deficiency of essential amino acids, the distributions of the second FITC-positive and the FITC-negative cohort coincide (Fig. 2). Thus, since the first FITC-positive cohort exhibits a somewhat lower DNA content than the second, this first cohort represents post-mitotic S phase cells. From the observation that the third FITC-positive cohort, clearly observed in D cells, exhibits a higher DNA content than the second FITC-positive cohort, it follows that this cohort represents premitotic, DNA replicating cells. Thus, these results support the recent finding that ND cells progress through a cell cycle in which DNA replication takes place immediately and exclusively after completion of mitosis. In D cells, however, DNA synthesis takes place not only after nuclear division but also before division.

Despite the occurrence of premitotic DNA synthesis in D cells, the following results indicate that DNA synthesis in these cells also correlates with cell proliferation. When a cell population is in balanced exponential growth, the slope of the regression line obtained by plotting the percentage of BrdUrd-labelled nuclei (labelling index) against labelling time roughly indicates the rate of entry of percentage of cells into S phase perh. The duration of the S phase may be estimated from the negative intercept of the line with the horizontal axis, and the percentage of cells of a growing culture being in S phase by extrapolation of the line to zero time (Sasaki et al. 1987). Fig. 3 shows that in both cultures, the labelling index increases linearly with the labelling time up to an increase in the relative cell number of about 1-7-fold. The duration of the phase of DNA synthesis in D cells seems to be slightly higher (0 ·55 ± 0 ·05 h, 3 experiments) than in ND cells (0 ·3 ± 0 ·05 h, 2 experiments). Relative to ND cells, the rate of entry of D cells into S phase is 1-7-fold lower, corresponding approximately with the slower growth rate of these cells. However, Fig. 3 also indicates an obvious difference between the growth characteristics of D and ND cells. Unlike ND cells, cytokinesis in D cells seems to take place in several crises of short duration.

On the basis of the finding of a correlation between BrdUrd incorporation and cell growth, the following

Cell number of control cultures that were not labelled with BrdUrd/FdUrd. suggestion may be made with regard to the appearance of premitotic DNA synthesis associated with an atypical nuclear DNA content in D cells. If premitotic DNA synthesis indicates endoreduplication of part of the genome twice instead of once or a severalfold amplification of particular DNA sequences, then total nuclear DNA content should increase with increasing generations, unless either DNA is degraded in subsequent cycles or excess DNA synthesis before mitosis consists only of duplication of a section of the genome, and is followed by non-replication of that section in the subsequent generation.

A rough estimation of DNA amplification can be made by restriction fragment analysis of DNA. The Acanthamoeba genome exhibits a relatively low genetic complexity of 3×10⁷ base-pairs (bp) (Jantzen, 1973) and therefore a 10-fold increase in copy number of particular DNA sequences may be detected by direct visualization of ethidium bromide-stained gels of separated restriction fragments (Coderre et al. 1983). By conventional preparation of DNA from D cell nuclei, we were unable to obtain high molecular weight DNA, whereas with ND nuclei a considerably lower degree of degradation occurs. However, by embedding the cells or nuclei in agarose blocks and subsequently digesting protein with high concentrations of proteinase K, non-degraded DNA can be obtained (Fig. 4). Fig. 4 shows restriction fragment analysis of DNA prepared from cells and nuclei. In the restriction pattern of total cellular DNA, several distinct fragments are visible, however, these are due to the presence of mitochondrial DNA. The characteristic restriction fragment pattern of the Acanthamoeba Neff strain mitochondrial DNA and the genome size of 4 × 10⁴bp that is obtained are in agreement with recent results (Bohnert & Gordon, 1980; Bogler et al. 1983). Since, in relation to the nuclear DNA complexity of 3 × 10⁷bp (Jantzen, 1973), the mitochondrial DNA complexity of 4×10⁴bp is 750-fold lower and the DNA content about fivefold lower (Byers, 1979), a copy number of about 150 of mitochondrial DNA per cell may be assumed. Thus, if the excess nuclear DNA observed in about 20% of growing D cells is due to amplification of particular DNA sequences, a copy number of about 100 (i.e. 20-fold amplification per total cells) should be detectable. The restriction pattern of nuclear DNA shows that this is not the case (Fig. 4). These results, however, do not exclude the occurrence of DNA amplification of low frequency. The alternative interpretation is that the cell cycle is altered in such a way that DNA replication is begun before mitosis and completed after mitosis. Several observations could support this suggestion: (1) postmitotic cells of D cultures (the first FITC-positive cohort in DNA/BrdUrd distributions) already exhibit an enhanced FITC/PI fluorescence (Fig. 1); (2) it has been shown (Pritchard & Lark, 1964) that if a culture is deprived of required amino acids, cells in the process of a cycle of DNA replication complete that cycle but are unable to initiate a new one. This also seems to be the case in D cells. Fig. 3 shows that about 5% of cells of a growing D cell culture are engaged in DNA synthesis.

Accordingly, during an 18 h BrdUrd/FdUrd labelling period of an arginine-deficient log stage culture, 4-8% of nuclei become FITC-positive. Arginine is one of the five essential amino acids in Acanthamoeba (Dolphin, 1976). Thus, in response to amino acid deficiency, the bivariate DNA/BrdUrd distribution should lack the first FITC-positive cohort (postmitotic) and the third cohort (premitotic), whereas the percentage of cells of the second FITC-positive cohort (G₂ phase cells that have finished their DNA replication) should be increased by about 5%. In fact, this is the observed behaviour (Fig. 2). (3) Recent experiments revealed that stationary stage ND cells are arrested at late G₂ phase of the cell cycle (Stôhr et al. 1987). Thus, if D cells are also stalled at a late cell cycle phase and replication takes place both before and after mitosis, stationary stage D cells should be stalled within their replication phase. To address this question, the distribution of bivariate DNA/BrdUrd and growth were examined in stationary stage D cells transferred into fresh D medium. Figs 5F and 6B show that during the lag of cytokinesis, the FITC and the PI labels in the cells increase slightly, indicating the arrest of cells within the replication phase.

Taken together, ND cells progress through a cell cycle in which a short replication phase (about 4% of the cell cycle) occurs immediately and exclusively after completion of mitosis. In D cells the duration of the replication phase is similar; however, DNA replication is observed in postmitotic as well as in premitotic cells, indicating an alteration in the timing of DNA replication.

Recent results indicate that stationary stage ND cells are committed to encystation whereas stationary stage D cells are not (Stôhr et al. 1987). In an effort to test the hypothesis that the incapacity of D cells to become committed to development is due to the altered timing of DNA replication, D cells were cultivated in ND medium. The occurrence of premitotic DNA synthesis in log stage cells was related to the encystation competence of stationary stage cells. Fig. 7B shows that immediately after transfer of log stage D cells into ND medium, premitotic DNA synthesis still takes place. When these cells reach the stationary stage for the first time, encystation can be initiated in only 15-30% of cells. However, after further cultivation of D cells in ND medium, distributions of bivariate DNA/BrdUrd are indistinguishable from those obtained with ND cells (Fig. 1), and stationary stage cells show a high level of encystation. Thus, the timing of DNA replication depends on the growth condition of cells and the encystation competence of cells depends on the timing of DNA replication.

As shown in the preceding section, the consequence of an altered timing of DNA replication is that stationary stage D cells are arrested within their replication phase, whereas stationary stage ND cells are stalled at late G₂-Since stationary stage D cells show a low competence of encystation, whereas stationary stage ND cells show a high competence, it may be suggested that these nuclear states are antagonistically mutually exclusive with regard to commitment. It is not, however, the status of stalled replication that is antagonistic to the status of commitment, since log stage G2 phase ND cells are also not committed for encystation. Thus, it may be suggested that commitment is inhibited in cells that express irregular (i.e. non-postmitotic) replication (D cells) or have the capacity for expression. In an effort to test this possibility, the capacity for non-regular DNA synthesis in log stage ND cells that do not encyst was compared with that in stationary stage ND cells that do encyst. When log stage ND cells are pelleted by low-speed centrifugation (50g; 1 · 5 min, r_av 12cm) and resuspended in fresh ND medium, a small percentage of cells divide but the majority of these cells do not incorporate BrdUrd and do not divide for about 3 h. Thereafter, coordinated DNA synthesis and cytokinesis takes place (Fig. 5A). A control culture that was centrifuged in the same manner and resuspended in its supernatant did not show this lag in growth, indicating that cell growth is not perturbed by this mode of centrifugation. In contrast, when pelleted log stage ND cells are resuspended in D medium, DNA synthesis is initiated immediately, and during the 3 h lag of cytokinesis 15% of nuclei become FITC-positive (Fig. 5B). Since these cells contain only one FITC-positive nucleus, this indicates the induction of non-postmitotic DNA synthesis. The distribution of bivariate DNA/BrdUrd shows that incorporation of BrdUrd wasfound both in cells that had a PI fluorescence similar to BrdUrd-negative cells and in cells that had a higher PI fluorescence than BrdUrd-negative cells. This indicates the occurrence of concomitant DNA degradation (Fig. 7A). The induction of premitotic DNA synthesis is inhibited by the absence of any of the five essential amino acids (Fig. 5C). When growing D cells are transferred into fresh medium, only a short lag of DNA synthesis and cytokinesis occurs (Fig. 5D). However, in stationary stage ND cells, neither BrdUrd incorporation nor cell growth can be induced by transfer of cells into D medium (Figs 5E and 6A). At 24 h after transfer, about 40% of cells die, and the remaining cells seem to have encysted incompletely.

The low capacity of stationary stage ND cells for initiation of premitotic DNA synthesis and proliferation after transfer into D medium may be due either to the stationary stage of cells or related to the specific cell cycle position at which cells are stalled. Recent results show that transfer of stationary stage ND cells into fresh ND medium leads to synchronous growth and concomitant

DNA synthesis after a lag of 4h (Stôhr et al. 1987). Thus, the low capacity of stationary stage ND cells for growth in D medium is not due to a general reduction in the capacity for growth and regular postmitotic DNA synthesis. Results, to be shown elsewhere, indicate a decrease in the rate of pinocytosis (examined by the transport of FITC-conjugated dextran into cells) in stationary stage cells in comparison to log stage cells. Since postmitotic and premitotic DNA synthesis in D cells as well as in log stage ND cells transferred into D medium is dependent on amino acid availability (Figs 2 and 5C), the rate of FITC-dextran transport was measured in log stage and stationary stage ND cells transferred into fresh D medium. However, no differences were found. Since pinocytosis seems to be the major mechanism for uptake of dissolved substances (Byers, 1979), this indicates that the low capacity of stationary stage ND cells for initiation of premitotic DNA synthesis and growth in D medium is not caused by nutrient limitation. Thus, the reduced capacity of stationary stage ND cells for growth in D medium may berelated to the particular cell cycle position at which these cells are arrested. To follow the growth of ND cells in D medium during the cell cycle, cells were synchronized by release from the stationary stage and transferred into D medium at various times. Fig. 8 shows that the proliferation property of late G₂ phase cells is low in comparison with cells of other cell cycle phases.

Taken together, the results reveal a relationship between the timing of DNA replication and developmental competence of cells. The only cells that become arrested at late G₂ stage and are committed to development are those cells in which DNA replication is exclusively postmitotic. Since in committed cells premitotic DNA synthesis is not inducible, it may be suggested that a reduction in the capacity for non-regular DNA synthesis is a prerequisite for development.

The cell cycle timing of ND cells is characterized by a short S phase (about 0-3 h), which takes place immediately and exclusively after completion of mitosis, and a long G₂ phase (about 7h). Accordingly, as measured by flow cytometry, a unimodal distribution of nuclear DNA content has been obtained. In growing D cells, however, about 20% of nuclei can show an increase in DNA content of up to 40% (Stôhr et al. 1987). The appearance of a subset of cells with increased DNA content per cell corresponds with the finding that DNA synthesis in D cells takes place before as well as immediately after nuclear division. The second difference between the nuclear states of stationary stage D and ND cells also seems to be related to the alternate timing of DNA replication in D cells. ND cells that replicate exclusively and early during the cell cycle, are stalled at late G₂ phase, whereas D cells, which also replicate at a late phase of the cell cycle, are arrested within their replication phase.

The suggestion that premitotic incorporation of BrdUrd in D cells indicates partial premature S phase DNA replication seem to be supported by several findings. (1) Distributions of bivariate DNA/BrdUrd indicate that premitotic incorporation of BrdUrd is associated with an increase in DNA content and that, in comparison to ND cells, postmitotic D cells already exhibit an elevated FITC/PI content (Fig. 1). (2) After inhibition of initiation of DNA synthesis by essential amino acid deficiency, neither postmitotically nor premi-totically DNA synthesizing cell populations are present in bivariate DNA/BrdUrd distributions. The percentage of FITC-positive G₂ phase cells increases to equal the proportion of DNA synthesizing cells seen before the onset of amino acid deficiency (Fig. 2). (3) The results indicate that premitotically synthesized DNA does not consist of a high copy number of particular DNA sequences (Fig. 4). However, DNA amplification of low frequency cannot be excluded. On the other hand, the finding that nuclease activity in D nuclei is higher than in ND nuclei (Fig. 4), and that transfer of log stage ND cells into D medium induces premitotic DNA synthesis and presumably also DNA degradation (Fig. 7A), may indicate the possibility that at least part of the excess DNA is degraded.

In another amoeba, Amoeba proteus, cytofluorometric determination of the nuclear DNA content also suggests that part of the genome is replicated in interphase more than once (Makhlin et al. 1979), and pulse-labelling of this organism with [³H] thymidine showed two waves of nuclear DNA synthesis in different parts of the cell cycle (Ord, 1968). However, it may be suggested that this atypical mode of DNA replication reflects the perturbation of growth conditions, and may play a role in adaptation to a changing environment, since in Acanthamoeba it was only found in cells cultivated in the defined medium and in A. proteus, Goldstein & Prescott (1967) found that all [³H]thymidine labelling took place once within a cell cycle. In lower eukaryotic cells, aberrant replication of DNA is induced when cell division is blocked and cell mass increases or, alternatively, the gene concentration may become reduced when the cell mass decreases (Berger, 1984). Thus, the abnormal DNA replication in D cells may be related to the imbalanced cell growth observed in these cultures. Estimates of cell cycle timing in Acanthamoeba differ considerably (Byers, 1979, 1986). These varying results may be partly due to differences in the methodology; however, in the light of the present findings, they may also result from different culture conditions.

The evidence found to date suggests that, in higher eukaryotic cells, an important component in the alteration of cells that generate an overt cancer and/or in its subsequent malignant progression is the relaxation or loss of control of the number and timing of initiations of DNA replication (Stark & Wahl, 1984; Stark, 1986; Schimkeei al. 1986). It is interesting to note that in Acanthamoeba an alteration in the timing of DNA replication is associated with an interference in normal developmental processes. ND cells that replicate only after the completion of the cell cycle show a high encystation competence, whereas D cells that also replicate before mitosis are constitutively committed to generation of trophozoites. However, when D cells are cultivated in ND medium the timing of DNA replication as well as the encystation competence changes so that the degree of premitotic DNA synthesis successively decreases and the percentage of cells that become committed during the stationary stage increases. This result indicates an incompatibility between a cell cycle progression that includes premitotic DNA replication and the developmental competence of cells. This suggestion is consistent with the recent finding that in Acanthamoeba, as in other unicellular eukaryotic organisms (Weijer et al. 1984; Sharpe & Watts, 1985; Nurse, 1985), the differentiation of cells is related to a particular cell cycle phase. Synchronization experiments revealed that this developmental decision point is in late G₂ phase of the Acanthamoeba cell cycle (Stôhr et al. 1987). Thus, these results indicate that, owing to an improper timing of DNA synthesis, D cells do not progress into late G₂ phase of the cell cycle and hence do not have the competence to become committed to development.

The molecular mechanisms involved in developmental commitment are not known in Acanthamoeba. We found that those cells that express irregular DNA synthesis (D cells or log stage ND cells in D medium) are in the noncommitted state, whereas premitotic DNA synthesis cannot be induced in committed cells. This suggests that the reduction in the capacity for irregular premitotic DNA synthesis may be a prerequisite for development.

This work was supported by DFG grant II Bl-Ja 217/8-1 to H.J.