INTRODUCTION
Characterization of the molecular detail of the cell division cycle in eukaryotic cells, based on research carried out over the last few decades, has shown that a period of DNA synthesis (termed S-phase) alternates with division of the nucleus, cytoplasm and other organelles and cell cleavage (M-phase) (Baserga, 1999; Puri, 1999). There are several variations on this theme, mainly associated with the length of the cycle due to the presence or absence of gap phases (G1 and G2) separating the main S- and M-phases and of a quiescent (G0) state in which the cell cycle is arrested. Despite the differences in the process, the roles of regulatory molecules, principally the cyclin-dependent protein kinases, seems well conserved throughout eukaryotes (Golias et al. 2004). It is still unclear how this cell-cycle model applies to the various life-cycle stages of the malarial parasite Plasmodium: which gap phases are present, the duration of the different phases and how progression through nuclear replication and division is controlled (Arnot and Gull, 1998; Doerig and Chakrabarti, 2004; Doerig et al. 2000; Leete and Rubin, 1996). Despite the discovery in the major human malarial parasite Plasmodium falciparum of apparent homologues of known eukaryotic cell-cycle regulatory proteins (Doerig and Chakrabarti, 2004), the apparently asynchronous division of multiple nuclei in a single parasite in erythrocytic schizonts (see below) is one indication that the parasite does not follow the consensus model of cell-cycle progression or regulation.
The erythrocytic cycle of P. falciparum has been the best studied because of its relevance to disease and treatment and our ability to reproduce this stage in culture. The multiplicative, asexual cycle takes approximately 48 h in the body or in culture (Garnham, 1988). It begins with the invasion of the host erythrocyte by 1 or more haploid merozoites and ends with the release of up to approximately 32 new merozoites, that can go on to invade fresh erythrocytes. In this way the parasitaemia increases and, in general, higher parasitaemia is associated with poorer disease prognosis. It is obvious that inhibiting this multiplication would prevent the disease, and molecules forming part of the cell-cycle apparatus or contributing to its regulation are proposed to be promising drug targets (Hammarton et al. 2003). The first recognizable cell-cycle event during the asexual erythrocytic cycle is the initiation of DNA replication, i.e. the beginning of the first S-phase. Owing to the range of ages found in cultured parasites and the difficulty in reducing that age range without drastically reducing the parasitaemia, it is difficult to determine the time after invasion of the start of this first S-phase: values around 30 hours have been reported (Inselburg and Banyal, 1984). After this time, DNA synthesis in the parasite population appears to be more or less continuous until around 38–40 h post-invasion: this presumably represents parasites containing S-phase nuclei, non-S-phase nuclei or a mixture of the two, as the nuclear bodies contain no more than diploid DNA amounts, at least in P. berghei (Janse et al. 1986). The onset of the first M-phase is marked by the first nuclear division, i.e. the beginning of schizogony, which occurs around 36 h post-invasion (Leete and Rubin, 1996). Approximately 4 or 5 divisions, presumably interspersed by periods of further DNA replication, are completed in a common cytosol before the nuclei and other organelles are partitioned into nascent merozoites. The latter events (known as segmentation) occur rapidly in the last few hours of the cycle and their exact timings are not known. It will be apparent from the above summary that understanding of the cell-cycle aspects of the asexual erythrocytic cycle is still very fragmentary: we do not know (i) whether parasites in the merozoite, ring and trophozoite stages before the initiation of DNA synthesis are in G0, in G1, or start in G0 and progress to G1 at some point, (ii) whether there are G1 phases between M and subsequent S phases and/or G2 phases between S and M and (iii) how long the various phases last. To confuse matters further, the numbers of nuclei do not increase geometrically as expected for a syncitium (Rao and Johnson, 1970). Whether this is the result of genuinely asynchronous nuclear division in a common cytosol or the restraint of some nuclei but not others from further division is not known (Read et al. 1993).
The generation of populations of parasites synchronized to the same cell-cycle phase would aid the study of the cell-cycle in this organism, for example by allowing us to measure the lengths of the phases and the concentrations, locations and activities of different cell-cycle regulatory elements at different times. Previous methods said to give ‘synchronized’ cultures of erythrocytic P. falciparum (using density gradients (Jensen, 1978; Kutner et al. 1985), temperature shifts (Rojas and Wasserman, 1993) or differential osmotic lysis of parasitized erythrocytes (Lambros and Vanderberg, 1979)) actually select a particular stage-range of the asexual developmental cycle. Multiple applications of these techniques can produce a population covering a much narrower age range than that normally found in cultured P. falciparum: as low as 3–5 h (Hoppe et al. 1991; Rojas and Wasserman, 1993). For the purposes of investigating the regulation of processes involved in cell-cycle progression and the events occurring at each cell-cycle phase (see above), this age range is too large. We reasoned that ‘batch’ synchronization of parasite cultures using inhibitors of specific cell-cycle processes might allow a population of parasites truly synchronized to a specific cell-cycle phase to be isolated and studied. This approach is based on the assumption from cell-cycle studies in other organisms that a series of checkpoints exists in the cell, which control whether it progresses to the next phase of the cycle or not. For example, entry into or progression through S phase (when DNA synthesis occurs) can be prevented by DNA damage. Inhibiting the completion of processes involved in each phase is thought to prevent cells from progressing past these checkpoints, thus arresting them in a single phase of the cell cycle. Agents that inhibit DNA synthesis (e.g. aphidicolin), mitosis (e.g. nocodazole) or molecules involved in cell-cycle control such as cyclin-dependent kinases (e.g. flavopiridol), have been used on eukaryotic cells, apparently to bring about synchronization of populations to particular cell-cycle phases (Bishay et al. 2000; Davis et al. 2001; Gewirtz, 1993; Merrill, 1998). In this study we have attempted to apply this principle to erythrocytic P. falciparum.
MATERIALS AND METHODS
Cell culture
P. falciparum clone 3D7 (obtained from M. Grainger, National Institute of Medical Research, London, UK) was cultivated in human O+ erythrocytes as previously described (Fennell et al. 2006). Age-selection of parasites was carried out by two-step sorbitol treatment as previously described (Lambros and Vanderberg, 1979) with minor modifications (Nankya-Kitaka et al. 1998). Cultures were examined by Giemsa staining. Rings were defined as immature trophozoites with a ring-like appearance and lacking obvious pigment, trophozoites had even cytosolic staining and visible pigment with a single nucleus, and schizonts were parasites containing greater than one nucleus. Numbers of segmenters were very low and they were therefore included with the schizonts.
Reagents
All chemicals were obtained from Sigma Aldrich, Dublin, Ireland unless otherwise stated. Trifluralin, amiprophos-methyl (APM), roscovitine, olomucine, nocodazole and Taxol (paclitaxel; ICN Biomedicals Inc., Aurora, Ohio, USA) were dissolved in dimethylsulphoxide. Hoechst 33342, vinblastine, and aphidicolin were dissolved in dH2O and L-mimosine was dissolved in NH4OH, and all 3 solutions were filter-sterilized.
Determination of median inhibitory concentrations (IC50)
Inhibitor susceptibility was determined in 96-well dishes using the parasite lactate dehydrogenase assay as described previously (Makler et al. 1993) with slight modifications (Fennell et al. 2006).
Inhibition of cell-cycle progression
Inhibition of cell-cycle progression was determined using cultures at around 12–18 h post-invasion (assuming a 48 h life-cycle) i.e. before the first S phase. Cultures were exposed to between 2 and 10 times the IC50 value of each agent (0·2–0·3 μM of Hoechst 33342; 0·75–2·5 μM of vinblastine; 5–20 μM of aphidicolin; 10–30 μM of trifluralin, oryzalin, or APM; 50–100 μM of roscovitine, olomucine, or nocodazole). Parasitaemia on initial treatment was 2–3% and experiments were carried out 3 times in duplicate. The same starting culture was used for all treatment groups within individual experiments. At specific time-points samples were taken for (i) morphological analysis by Giemsa staining and light microscopy and (ii) DNA content determination by flow cytometry. Parasite DNA content was determined by measuring propidium iodide staining following RNAse A treatment as previously described (Fennell et al. 2006). The stained cells were analysed using a FACScan flow cytometer (Becton-Dickinson, CA, USA) as described previously (Pattanapanyasat et al. 1997).
Reversibility of inhibition
Reversibility of inhibition of schizogony by Hoechst 33342, roscovitine and L-mimosine was determined using cultures aged 12–18 h post-invasion exposed to the inhibitors for 18, 24 and 32 h. Cultures were re-incubated in inhibitor-free medium for 30 and 78 h, at which times microscopical examination was carried out.
RESULTS
Inhibition of cell-cycle progression
The ability of each agent to block cell-cycle progression at a particular phase, thus producing a synchronous population of parasites, was assessed. The parasites were selected to a narrow age range and allowed to progress to the early trophozoite stage before various agents believed or expected to block cell-cycle progression were added. The compounds used included DNA-binding agents/DNA synthesis inhibitors (Hoechst 33342, aphidicolin (Inselburg and Banyal, 1984)), microtubule inhibitors (vinblastine, Taxol, trifluralin, nocodazole: (Bell, 1998)) and inhibitors of CDKs (roscovitine and olomucine (Woodard et al. 2003)). L-mimosine inhibits specific protein hydroxylases involved in post-translational hydroxylation of eukaryotic translation initiation factor-5A (eIF-5A) (McCaffrey et al. 1995). All of these compounds had been previously reported to have antimalarial activity in culture, except L-mimosine and Hoechst 33342, which we found to have 72-h IC50 values of 76 μM and 0·021 μM respectively.
Hoechst 33342, roscovitine and L-mimosine all inhibited parasite progression into schizogony, i.e. both nuclear division (Fig. 1) and DNA synthesis (Fig. 2) were blocked. Following 24 or 32 h exposure to all agents a noticeable disruption of the clarity of the traces obtained on flow-cytometric analysis of the cultures could be observed (Fig. 2; results for 24 h exposure qualitatively similar to those for 32 h). This is probably due to damage to the parasites caused by prolonged exposure to these agents and subsequent loss of DNA from dead or damaged cells. Notably, the morphological appearances of these (roscovitine, Hoechst 33342 and L-mimosine)-blocked parasites under the microscope were not those of late trophozoites. The cells present were larger than expected, being similar in size to mid- to late schizonts. Both aphidicolin and olomucine reduced the proportion of cells progressing into schizogony, but didn't block it completely (Fig. 1, Fig. 2). Taxol and vinblastine prevented cytokinesis (and the production of new rings) but did not prevent DNA synthesis and/or low levels of nuclear division. This can be seen in Fig. 1 in parasites treated with vinblastine, in which there were no new rings produced but there appeared to be schizonts present that contained multiple nuclei. These data were confirmed by flow cytometry (Fig. 2C) although at the later time-point (Fig. 2F) there seemed to be loss of DNA from dead or damaged parasites. Results for Taxol were similar to those obtained for vinblastine (data not shown). This indicates that in agreement with previous findings (Sinou et al. 1998), inhibiting mitosis does not stop DNA synthesis in P. falciparum, so microtubule inhibitors do not appear to synchronize the cell cycle. Continuation of DNA synthesis after treatment with the microtubule inhibitor Taxotere has been noted previously (Sinou et al. 1998) and suggests the absence of a G1 checkpoint found in some other eukaryotic cells.
Fig. 1. Effects of inhibitors on progression into schizogony. Cells were selected to an age range of 12-18 h post-invasion and incubated with the inhibitors. Samples were taken 18 h (A) and 32 h (B) post-treatment for microscopical analysis of Giemsa-stained smears. Results for Hoechst 33342 and L-mimosine (not shown) were similar to those for roscovitine. Aphidicolin, olomucine and vinblastine produced similar results. Results for the other microtubule inhibitors were similar to those for vinblastine. The results shown are means for duplicate experiments and vertical bars indicate standard errors.
Fig. 2. Flow-cytometric analyses of the inhibitor-treated parasites. Cultures of parasites 12–18 h post-invasion were exposed to no inhibitor (A, D), 50 μM roscovitine (B, E) or 750 nM vinblastine (C, F). Samples were taken at intervals as in Fig. 1, fixed in 1% formaldehyde, stained with propidium iodide and fluorescence levels measured. Fifty thousand cells from each sample were counted and the experiment was performed 3 times in duplicate. The results from an untreated culture (A) and the corresponding cultures exposed to roscovitine (B) or vinblastine (C) for 18 h are shown. A second untreated culture (D) and the corresponding cultures exposed to roscovitine (E) or vinblastine (F) for 32 h are also shown. Results for Hoechst 33342- and L-mimosine-treated cultures were similar to those for roscovitine (data not shown), and results for aphidicolin, olomucine and nocodazole were similar to those for vinblastine. Results for 24 h exposure were similar to those for 32 h. The maximal displacement on the y-axis is 100 cells. HEn and HEp represent uninfected and infected erythrocytes, respectively.
Reversibility of cell-cycle block
One of the main methods used in studying the cell-cycle in other organisms is to synchronize the cells in culture to a particular cell-cycle phase, reverse the synchronizing block and observe various features of the emerging cells (Harper, 2005). An attempt was made to develop such a technique for P. falciparum. As Hoechst 33342, roscovitine and L-mimosine all successfully blocked both DNA synthesis and nuclear division and caused accumulation of parasites at the trophozoite–schizont boundary (presumably in the first S-phase, or G2 phase if it exists: Figs. 1 and 2), the ability to remove the block applied by these agents was assessed. Following 18, 24 or 32 h exposure to the inhibitor, samples were taken for microscopical analysis, the cells were washed extensively with culture medium to remove the inhibitor (or at least that proportion of the inhibitor that was not tightly bound to the cells) and then re-incubated at 37 °C for 30 h. Results are shown for cultures exposed to the inhibitors for 18 h (Fig. 3A); qualitatively similar findings were obtained at all 3 time-points. For both roscovitine- and L-mimosine-treated cultures, slowing down of development was apparent, with parasitaemias lower than the control apparent and schizonts with multiple nuclei being the most prominent stage present (Fig. 3B). Following a further 48 h incubation (Fig. 3C), the roscovitine- and L-mimosine-treated cultures had increased parasitaemias approaching that of the control culture, including new rings and trophozoites. This confirmed that the block was removed and that cytokinesis and re-invasion of new merozoites occurred. Somewhat disappointingly, the cultures did not remain completely synchronous following the removal of the block, progression through schizogony and re-invasion, although this might be improved by multiple staggered treatments. In the case of Hoechst 33342 (Fig. 3B), neither schizonts nor new rings were apparent after the inhibitor was removed, indicating that the block was irreversible. In agreement with this conclusion, the Hoechst 33342-treated cultures did not recover after subsequent incubation (Fig. 3C).
Fig. 3. Reversibility of inhibition by Hoechst 33342, roscovitine and L-mimosine. Cells were selected to an age range of 12–18 h post-invasion and incubated with the inhibitors. A sample was taken 18 h post-treatment for microscopical analysis, when the control parasites were aged about 30–36 h post-invasion (A). The cultures were washed and re-incubated in culture medium for a further 30 h, at which time a sample was again taken for microscopical analysis (B). The cultures was then diluted 1[ratio ]3 into fresh erythrocytes and incubated for a further 48 h, at which time samples were again taken for microscopical analysis (C). Results are means of triplicate experiments and vertical bars show standard errors.
DISCUSSION
The cell cycle of the malaria parasite P. falciparum is a complex process essential for parasite development and so potentially contains novel targets for new chemotherapeutic agents (Hammarton et al. 2003). The limited knowledge available about the processes involved in the progression of P. falciparum through intraerythrocytic schizogony is in part due to the difficulty in obtaining truly cell-cycle-synchronized populations of parasites. Non-inhibitor based methods used for synchronizing the cycle in mammalian cells, yeasts and other organisms on the basis of collection of new daughter cells cannot be applied to these parasites because of their intracellular location, so the development of an inhibitor-based method was attempted.
Inhibitor-based methods of ‘batch’ synchronization of eukaryotic cell populations to particular cell-cycle phases have been widely reported (Harper, 2005). These techniques are based on applying an inhibitor to prevent cell-cycle progression past a certain point, and then once ample time has been allowed for cells in the population to reach this point, the effects of the inhibitor are reversed and the cells allowed to proceed through the cell cycle uninhibited. By this definition, roscovitine and L-mimosine appear to achieve synchronization in P. falciparum, by inhibiting progression into schizogony. All other agents examined either failed to produce a successful block of schizogony, as seen with aphidicolin and the microtubule inhibiting agents Taxol and vinblastine, or, as was the case with Hoechst 33342, inhibition of schizogony was irreversible.
In recent years (and largely while this study was under way) the validity of using inhibitor-based methods (and other ‘batch’ methods such as nutrient starvation, double thymidine block and treatment with hormones such as Saccharomyces cerevisiae alpha-factor) for synchronizing the cell cycle of populations of eukaryotic cells has been questioned (Cooper et al. 2006; Shedden and Cooper, 2002). The objection stems from a belief that while a population of cells can be obtained which exhibit a common property, e.g. all cells have a similar DNA content, such an alignment does not mean that the cells are arrested at the same point in the cell cycle. In addition, one cannot be sure that the treated cells are unperturbed and truly representative of normal cycling cells. Cooper and colleagues (Cooper et al. 2006; Shedden and Cooper, 2002) have argued that the results of certain studies of cell cycle-dependent gene expression based on batch-synchronized human or yeast cells are largely artefactual (for a heated debate on these issues, see (Cooper, 2004a,b; Spellman and Sherlock, 2004a,b; Cooper et al. 2006). This viewpoint may still be a minority one, but there is an additional problem to be considered in the case of P. falciparum. As previously stated, the success of batch synchronization of cultures depends on the presence of checkpoints at the end of each cell-cycle phase preventing the cell from progressing past a certain point until the cell is ready to do so. Based on the results in this study with the microtubule inhibitors, as well as a previously published study using Taxotere (Sinou et al. 1998) which showed that DNA synthesis restarts despite lack of completion of mitosis, it is unlikely that a conventional checkpoint exists at the end of G1 phase. Also, while roscovitine has been shown to interact with a recombinant form of a putative P. falciparum cyclin-dependent kinase (Harmse et al. 2001), it may have more than one target protein, and the molecular target of L-mimosine in Plasmodium is undefined, undermining one's confidence that these agents have blocked the parasites at a specific phase of the cell cycle. This, in conjunction with the lack of knowledge of the control of the cell cycle makes it impossible to assume that all the parasites in a single population are in the same phase when treated with these inhibitors. Moreover, our data indicate that at the concentrations of roscovitine and L-mimosine required to achieve the (apparent) block, there is a degree of irreversible damage to the parasites.
In conclusion, we have found that some of the inhibitors previously reported or assumed to be specific for particular cell-cycle phases or to ‘synchronize’ erythrocytic P. falciparum cells do not block development at specific points as expected. Three agents were found to block development at the trophozoite–schizont boundary, presumably in S- or G2-phase. The effects of 2 of these (roscovitine and L-mimosine) were (at least partially) reversible, and these 2 agents might in principle have potential as P. falciparum cell-cycle synchronizing agents. However, on the basis of the properties of the recovering roscovitine- and L-mimosine-treated parasites described here and the arguments that such methods ‘cannot and do not’ synchronize other cell types (Cooper et al. 2006), we now feel that (contrary to our original idea) any results obtained using inhibitor-treated parasites are unlikely to be reflective of the normal cell cycle. While inhibitors may still be used to shed light on certain aspects such as the presence of checkpoints, study of the Plasmodium cell cycle will continue to be difficult and live-cell microscopy using fluorescently-tagged molecules would seem to offer the best hope of understanding this important and unusual process.
This work was supported by the Irish Higher Education Authority, Programme for Research in Third-Level Institutes (HEA-PRTLI)-funded Institute for Information Technology and Advanced Computing (IITAC) programme at Trinity College, Dublin. We thank Ralph Gräser for discussions that led to this study and Christian Doerig for helpful comments on the manuscript.
