Effects of mevinolin on cell cycle progression and viability of tobacco BY-2 cells
Centre National de la Recherche Scientifique, Institut de Biologie Moléculaire des Plantes, Département d’Enzymologie Cellulaire et Moléculaire, Institut de Botanique, Université Louis Pasteur, 28 rue Goethe, F-67083 Strasbourg, France
Centre National de la Recherche Scientifique, Institut de Biologie Moléculaire des Plantes, Département d’Enzymologie Cellulaire et Moléculaire, Institut de Botanique, Université Louis Pasteur, 28 rue Goethe, F-67083 Strasbourg, France
This manuscript is dedicated to the late Dr Claude Gigot, our colleague and friend.
*For correspondence (fax +33 3 88 35 84 84;
Mevinolin, an inhibitor of 3-hydroxy-3-methylglutaryl-CoA reductase, was used to study the importance of mevalonic acid (MVA) for cell cycle progression of tobacco (Nicotiana tabacum L.) BY-2 cells. After treatment with 5 μM mevinolin, the cell cycle progression was completely blocked and two cell populations accumulated (80% in phase G0/G1 and 20% in G2/M). The arrest could be released by subsequent addition of MVA. Effects were compared to those caused by aphidicolin, an inhibitor of α-like DNA polymerases that blocks cell cycle at the entry of the S phase. The 80% proportion of mevinolin-treated TBY-2 cells was clearly arrested before the aphidicolin-inducible block. By the aid of a double-blocking technique, it was shown that the mevinolin-induced cell arrest of highly synchronized cells was due to interaction with a control point located at the mitotic telophase/entry G1 phase. Depending on the developmental stage, mevinolin induced rapid cell death in a considerable percentage of cells. Mevinolin treatment led to a partial synchronization, as shown by the increase in mitotic index. The following decrease was correlated with the above-mentioned induction of cell death.
However, for a close inspection of how mevinolin interacts with the cell cycle, efficient methods of synchronization must be applied. The tobacco BY-2 (TBY-2) cell line, which had originally been selected for its rapid division cycle of about 14 h, can be synchronized to a higher degree than any other plant cell line (Nagata et al. 1992). The aim of the study presented here was to determine the sensitivity of unsynchronized and synchronized populations of TBY-2 suspension-cultured cells to mevinolin during different stages of the cell cycle. Furthermore, we wanted to elucidate the reasons for the apparently phytotoxic effects of mevinolin.
HMGR activity in TBY-2 cells
HMGR activity was determined in microsomal preparations. The apparent specific activity was remarkably high compared to values found in literature, using other cell lines of tobacco as an enzyme source (Bach 1987;Chappell et al. 1995;Gondet et al. 1992;Schaller et al. 1995). Determination of HMGR activity during the growth cycle showed a correlation with the capacity of the cells to divide (Fig. 1a). At the beginning of the growth curve HMGR activity was very high, steadily decreasing to about one-seventh of the initial value until the cells reached stationary phase. This time course of apparent HMGR activity was not reflected exactly by the quantity of enzyme protein, as determined by Western blot analysis (Fig. 1b).
Inhibition of cell cycle progression by mevinolin
It has been shown previously that mevinolin is a potent inhibitor of proliferation of TBY-2 cells (Crowell and Salaz 1992;Morehead et al. 1995). In our system its efficiency was even higher: significant growth inhibition was already observed at 10 nm. At 5 μm the inhibition reached its maximum and thus this concentration was chosen for further experiments. However, this inhibition never exceeded 90% (data not shown).
The inhibition of cell proliferation was analysed with the aid of cytometric scanning. While untreated, non-synchronized cells exhibited cell cycle oscillations after transfer to new medium (Fig. 2a), mevinolin induced an arrest after 48 h (Fig. 2b). This explains why only a maximum of 90% inhibition was observed, since apparently a certain proportion of cells being distributed randomly in the cell cycle is not affected by the inhibitor. Essentially two cell populations appeared: about 80% in phase G0/G1 and 20% in phase G2/M. No mitotic figures appeared after 48 h of mevinolin treatment (for example Fig. 6a), which indicates that this second site of control is localized at G2. The inhibition of cell division was overcome by subsequent addition of MVA as the immediate product of the inhibited enzyme reaction, and the cells needed only a few hours to recover (Table 1). This is a strong argument in favour of the hypothesis that cells are not blocked in G0 but rather in G1, because it has been reported that for re-entering the cell cycle from G0 to G1 there is a time lag of about 17 h (Gould 1984).
Table 1. . Reversion of the mevinolin-induced block by addition of mevalonate
In order to localize the block more precisely in phase G1, we compared the effects of mevinolin with those of aphidicolin, which was used at a concentration of 20 μg ml–1 and which completely blocks DNA replication (Reichheld et al. 1995). The inhibition by aphidicolin was found at the early S phase (Fig. 3a). Clearly some cells in the rather broad peak (2c) were already in S phase. In contrast to this, the sharp peak seen with mevinolin-treated cells was exclusively confined to G1 (Figs 3a, 2c domain). These observations were confirmed by directly measuring DNA synthesis (Fig. 3b) and by fluorimetric determination of the DNA content in individual cells of a given population (Fig. 3c). Inhibition of thymidine incorporation in the presence of aphidicolin was less drastic than that induced by mevinolin, which again indicates that inhibition of MVA formation leads to an arrest before S phase. Within 1 h thymidine incorporation was resumed after washing the cells free of inhibitors (Fig. 3b). The apparent incorporation rates were practically identical for both mevinolin- and aphidicolin-treated cells, which confirms the rapid entry of G1-blocked cells into S phase. This series of experiments clearly demonstrates that the mevinolin-induced arrest within the cell cycle is localized before that of aphidicolin.
Mevinolin effects on synchronized TBY-2 cells
To determine more precisely the phase of sensitivity towards mevinolin, we had to use synchronized cells. Cells were synchronized by aphidicolin treatment (control) and then allowed to pass through the cell cycle (Fig. 4a). After washing, control cells arrived nearly completely in G1 11 h later and continued to enter S and G2, whereas mevinolin-treated cells were slightly slower in cell cycle progression and arrived in G1 after 14 h, where they were apparently completely arrested (Fig. 4b). This retardation became evident 9 h after washing out aphidicolin and the addition of mevinolin (Fig. 4c), when the cells were in mitosis (Fig. 4d). This was the first indication that the period of sensitivity towards mevinolin must have been within or close to mitosis, although the cells were finally blocked in G1.
To confirm this further, we designed a blocking experiment with double-synchronized cells (Fig. 5a), in principle an approach similar to that described by De La Torre et al. (1989). This approach should give evidence of the transition point within the cell cycle where the inhibitor is effective and after which sensitivity is lost. Mevinolin was added at time 0 after release from the propyzamide block (cells were in prometaphase of mitosis;Nagata et al. 1992) and consecutively every hour thereafter up to 8 h. As a consequence of the propyzamide treatment, during the first 3 h cells were still in mitosis (0 h: prometaphase; 3 h: telophase) and thereafter in G1 phase. The development of the mitotic index was followed hourly for 9 h after release from the propyzamide block (Fig. 5b). In parallel, the position of cells within the cell cycle was analysed by microscopic scanning of Feulgen-stained nuclei at the moment of addition of mevinolin (Fig. 5c). Twelve hours after release from propyzamide, cells from the nine individual synchronizations were harvested and nuclei were analysed in the same way (Fig. 5d). Clearly, mevinolin treatment during the first 3 h led to an arrest of cells within G1 (Fig. 5d). However, beginning with a time delay of 3 h between release from the propyzamide block and addition of mevinolin, cells started to pass the G1/S boundary (Fig. 5d). This proves that TBY-2 cells were sensitive to mevinolin in mitosis, but not after entry into G1, even though treatment led to an arrest in G1. Thus some signal, arising from mevinolin treatment, generated during mitotic telophase must have affected the cells such that the majority could not pass the control point that leads to initiation of DNA replication.
Induction of cell death by mevinolin
When unsynchronized cells were treated with mevinolin, the peak of the mitotic index (25%) was reached after 20 h (Fig. 6a). The following decrease in the mitotic index was paralleled by a steep increase in the number of dead cells (Fig. 6b), followed by a low increase of about 10–15% over a time period of 6 days. These cells exhibited the morphological aspect of mitotic telophase and were not organized in small clusters as necrotic cells usually are, but rather were individually distributed within living cells (Fig. 7a,b). Therefore it appears as if mevinolin-treated cells left the cell cycle in mitosis and were possibly partially destined to die. It has to be noted that these cells were labelled not only by propidium iodide, which penetrates dead cells exclusively and leads to a red–orange staining of nuclei, but also weakly by fluorescein diacetate, which specifically stains living cells and leads to a yellow–green staining of the cytoplasm (Fig. 7a). Furthermore, in some cases Hoechst 33258-stained nuclei showed the appearance of microstructures after mevinolin treatment (Fig. 7c,d). In accordance with these observations, a long-term effect of mevinolin was an increase in DNA degradation (data not shown).
The induction of cell death was dependent on the age of the cells. When TBY-2 cells were cultured for 4, 5, 6, 7 or 8 days after transfer to new MS medium, and then treated for a further 48 h with 5 μm mevinolin, we noted that young cells were more sensitive than older ones in terms of cell death induction (Fig. 8a). Cytometric analysis (Fig. 8b) confirmed this observation. For instance, in 4- or 5-day-old cells this treatment led to the appearance of a population of nuclei having less than a G1 typical DNA content, which is characteristic of DNA degradation. This experiment provided evidence of a high toxicity of 5 μm mevinolin to TBY-2 cells during exponential growth phase. Moreover, the typical cell cycle phase distribution, i.e. about 80% in G1 and 20% in G2/M, was only observed with cells approaching stationary phase.
Because HMGR represents the target enzyme of mevinolin, and as corresponding data are not available for TBY-2 cells, we have measured activity during a normal growth cycle of this cell line and also analysed protein quantity. Compared with other tobacco cell lines, even those that have a higher activity of HMGR, such as triazole-resistant cells overproducing sterol intermediates (Gondet et al. 1992), in TBY-2 cells apparent HMGR activity was higher by another factor of three. The steep increase in apparent enzyme activity following new subculture was not paralleled exactly by the mass of HMGR protein. Cells from the stationary phase (quiescent or undividing cells), which initiate cell cycle progression when transferred to fresh medium, apparently need a rather high MVA production in order to ensure the supply for biosynthetic sinks such as formation of sterols as membrane constituents, etc. It has to be noted, however, that the antibodies used would not distinguish between different isoforms of HMGR. Nevertheless, the differences found might be interpreted such that regulation of HMGR in plants (Newman and Chappell 1997) and in TBY-2 cells can occur at the translational and post-translational levels.
For mammalian cell lines it has been suggested that a substantial induction of HMGR activity may be necessary to build up the cellular MVA and isoprenoid pools in cells that must undergo the transition to the S phase starting from quiescence (Maltese and Sheridan 1988). Endogenous sterol synthesis most probably represents a major sink for MVA. The experiments of Ryder and Goad (1980), using cell cultures of Acer pseudoplantanus treated with compactin (an analogue of mevinolin), showed the presence of a pool of MVA or of its derivatives that was gradually depleted by the process of sterol biosynthesis. If cells are allowed to build up such a pool of MVA-derived precursors, it would explain why cells apparently become insensitive to low concentrations of mevinolin (Morehead et al. 1995).
When cells were transferred from near stationary phase, where more than 90% of the cells are in G0/G1, mevinolin could only partially block cell division, as expressed in packed cell volume. This was a first indication that cells in this particular phase were not sensitive and could thus proceed in the cell cycle. The increase over initial cell mass actually reflected the passage through a complete division cycle, up to the point where mevinolin finally arrested cell division. This explains why maximum inhibition, independent of inhibitor concentration, remained below 90%.
Initiation of cell division activity is dependent on cytokinins (Binns 1994) and a correlation has been found between the endogenous concentration of cytokinins and cell cycle progression (Nishinari and Syôno 1986). These observations were recently confirmed by using TBY-2 cells, where apparently two peaks of cytokinin synthesis/accumulation appear, one in S phase and a second in mitosis (Redig et al. 1996). The effect of low concentrations of mevinolin (1 μm) on suspension cells derived from TBY-2 calli was overcome by simultaneous addition of kinetin (Crowell and Salaz 1992), although the precise reason for this remains to be elucidated. The putative increase of conditioning factors during the cultivation period might also explain why these TBY-2 cells, subcultured by dilution of 3 ml of stationary phase cells into 30 ml of fresh medium, are only inhibited by 1 μm mevinolin within the first 2 days of culture (Morehead et al. 1995).
The above-mentioned observations led us to choose the concentration of 5 μm mevinolin for the series of experiments. Although the inhibitory effect of mevinolin on the proliferation of plant cells has been examined previously, the mechanism of interaction has not yet been determined precisely. Instead, the interpretation of results is based largely on experiments carried out with various animal and human cell lines. Depending on which cell lines are used the observations made, as well as their interpretation, vary greatly.
Therefore, for the precise determination of the phase where TBY-2 cells are arrested by mevinolin, we had to use some marker point. For this we used the diterpene aphidicolin, an inhibitor that blocks the α-like DNA polymerase complex in eukaryotic cells. This might explain why cells treated with aphidicolin were blocked in the early S phase (rather than at the G1/S transition, as frequently described). Having this time-point for comparison, we have clearly shown that the major part of mevinolin-treated TBY-2 cells is arrested earlier, i.e. in the late G1 phase. More important, while the point of interaction of aphidicolin is a direct consequence of its mode of action, that of mevinolin is apparently not. Instead, it affects cells in a different phase of the cell cycle, at the transition from mitotic telophase to G1. The absence of a MVA derivative and/or a loss in functionality of a prenylated molecule at the end of the mitotic phase might lead to incapability of synthesizing DNA later on. For mammalian cells it has been shown that the different cell cycle phases are linked such that erroneous or incomplete passage through a given phase leads to the inhibition of transmission signals that allow transition into following stages. At the level of so-called checkpoints the cell verifies that the division was correctly affected (cf. Jacobs 1995). For instance, cells with errors during DNA replication are not able to effect mitosis before repair mechanisms have worked (Hartwell and Weinert 1989). Here, the cells were arrested in transition G2/M, although the malfunction was localized in S phase. The dependence of S phase transition from preceding phases, like mitosis, is less well characterized. In fission yeast Schizosaccharomyces pombe, the dysfunction of a protein kinase (chk1) led to an arrest at the checkpoint in late G1 (Carr et al. 1995). Only if this control point is correctly passed can cells start DNA synthesis. It is reasonable to assume that in TBY-2 cells mevinolin affects the process of isoprenylation of a regulatory molecule, for instance a GTP-binding protein playing a role in signalling (Bowler and Chua 1994), and might consequently disturb signal cascades that normally allow the G1/S transition. The regulatory role of GTP binding proteins in plants has been discussed (Ma 1994;Terryn et al. 1993), and isoprenylation also occurs in plant cells (Randall and Crowell 1997).
We have shown that mevinolin selectively induces cell death within a TBY-2 population. In the presence of both propidium iodide and fluorescein diacetate, necrotic cells, always occurring as clusters, are exclusively propidium iodide-stained, while dying cells, usually randomly distributed between unaffected cells, are stained with both dyes. Cells destined to die might irreversibly leave the cell cycle close to M phase (Havel and Durzan 1996). A specific response of TBY-2 cells compared to animal cells is the biphasic increase in the number of dead cells. The initial rapid increase coincides with the decrease in the mitotic index. The slow increase in cell death thereafter might be explained by the fact that other cells are largely prevented from passing the G1/S boundary. This requires further studies. The time-dependence of toxic effects of mevinolin, as indicated by induction of cell death, can be interpreted to mean that cells in exponential phase (4–5 days of culture), representing an unsynchronized and rapidly proliferating population, with a statistical distribution of cells in different stages of the cell cycle, contain a subpopulation where cell death can be induced by MVA deprivation. We also noted that cells within this exponential phase could not be synchronized, which apparently needs a clearly defined distribution within the cell cycle. At the entry into stationary phase (8-day-old cells) we determined that more than 95% accumulated in G1.
In conclusion, the experiments presented here provide unequivocal evidence for the importance of MVA synthesis for cell cycle progression. Several MVA derivatives, or signals whose formation is dependent upon those, might be implicated in the regulation of the cell cycle. Apparently, the mevinolin-induced deficit in such entities cannot be compensated for in vivo. Use of branch-specific inhibitors is necessary to identify precisely the member(s) of isoprenoid families that are responsible for control of cell division and, if absent or functionally blocked, lead to induction of cell death.
A suspension cell culture of tobacco cells, originally derived from young plants of Nicotiana tabacum L. cv Bright Yellow-2 (TBY-2), was made available by the Tobacco Science Research Laboratory, Japan Tobacco Inc., Tokyo, Japan. Mevinolin was a kind gift of Drs Michael Greenspan and Alfred W. Alberts (Merck Sharp & Dohme Research Labs, Rahway, NJ). Before use, the lactone was converted to the open-acid form according to the protocol of Kita et al. (1980).
TBY-2 cells were cultured in a modified Murashige–Skoog (MS) medium as described by Reichheld et al. (1995). Cell cultures (80 ml in 250-ml Erlenmeyer flasks) were kept in the dark at 26°C and shaken with 174 r.p.m. Cells were subcultured weekly (1.5 ml to 80 ml of new medium, if not stated otherwise). For inhibition studies the inhibitor solutions were sterile-filtered.
Incorporation of [3H]thymidine
[3H]thymidine incorporation into DNA followed the protocol of Sala et al. (1980) with some modifications. In brief, 1 ml of suspension was incubated for 1 h under standard culture conditions in the presence of 5 μCi [methyl-3H]thymidine (Amersham; 3.03 TBq/mmol, 82 Ci mmol–1). After centrifugation at 13 000 g, the supernatant was decanted and the pellet was washed twice with PBS (Sigma). Cells were frozen in liquid N2 and homogenized in 10% (w/v) TCA containing 50 μm of unlabelled thymidine. The slurry was kept at 0°C for 1 h, followed by centrifugation (13 000 g). The pellet was washed with 5% TCA, then by ice-cold 70% (v/v) ethanol (EtOH), and resuspended in 5% SDS. Aliquots of the homogenate were deposited on glass fibre filters (Whatman type GF/C) and radioactivity was determined (Packard Minaxi Tri-Carb 4000) using a water-compatible scintillation cocktail (Ready Gel, Beckman).
Cell biology techniques
For fluorescence microscopy nuclear DNA was stained by Hoechst (H) 33258 (Sigma; final concentration of 0.1 μg ml–1 in 0.005% Triton X-100) according to the protocol of Latt and Stetten (1976). After incubation for 5 min at room temperature (RT), the cells were examined with a Zeiss epifluorescence microscope (HBO lamp, 50 W). Mitotic index was determined by counting the percentage of H33258-stained nuclei in all phases of mitosis, compared to the total number of cells (average 500 cells).
Viability of cells (Huang et al. 1986) was tested by addition of propidium iodide (Sigma; 600 μg ml–1), which only penetrates dead cells and leads to a red–orange staining of nuclei, and of fluorescein diacetate (Sigma; 100 μg ml–1), specific to living cells and which leads to a yellow–green fluorescence-staining of the cytoplasm. Cells were incubated for 10 min at RT, then examined by fluorescence microscopy.
Cytofluorometry was carried out using a microscope connected to an image analysing system and operated with the aid of the program designed by LAB-EL (Institut Universitaire de Technologie, Mulhouse, France). H33258-stained cells were prepared as described above. This method allowed for quantification of fluorescence arising from the fluorochrome bound to the DNA with a scale from 0 to 256 relative units.
For cytometry TBY-2 cells were fixed for 10 min in Bohm-Sprenger solution [37% formaldehyde, 15%; acetic acid, 5%; methanol, 80% (v/v)]. Cells were then hydrated by immersion in distilled water (10 min, RT), and the fixed material was hydrolysed in 6 n HCl for 1 h at RT, which liberates the purine bases of DNA. The cells were washed by passage through water (four times, 1 min each). Nuclei were stained with Schiff's reagent (BDH Laboratory Supplies, UK) and developed by fourfold incubation in 1 n sulphuric acid. Finally the cells were deposited on microscope slides and impregnated with 3-aminopropyl-triethoxysilane (Fluka). When fixed, cells were dehydrated by passage through water (3 min), 70% EtOH (3 min), 95% EtOH (3 min), 95% EtOH (5 min), 100% ETOH (2 min) and 100% EtOH (5 min). The nuclei were then mounted in neutral resin (EUPARAL) and were covered with a slip that was sealed with nail polish. After drying their DNA content was analysed (system SAMBATM 2005, program STAT 2005). This allowed for a statistical analysis of the DNA content of a given population of nuclei. The resulting histogram was calibrated by comparison to a sample whose DNA content was known. With this a graphical representation of the number of nuclei belonging to a certain type of the haploid chromosomes (2c, 4c, etc.) was established. On the statistical basis of 300 cells per sample three categories were calculated as a function of colour intensity: non-replicated DNA, 2c, typical of phases G0 and G1; replicated DNA, 4c, typical of cells in phase G2 or M; intermediate content of DNA, typical of cells in phase S of the cell cycle.
Synchronization of TBY-2 cells
This method was first described by Nagata et al. (1992). Twenty millilitres of 7-day-old TBY-2 cells were transferred to 80 ml of new MS medium to which 60 μl of aphidicolin (Sigma; stock solution 5 mg ml–1 in DMSO) were added. The cells were kept for 24 h under otherwise standard culture conditions. Then cells were filtered and washed with 3% (w/v) sucrose in order to remove the inhibitor. As after this treatment most cells are in phase S or G2, the second inhibitor propyzamide (3 μm; Sumitomo Chemical Co., Japan; dissolved in DMSO) was added. When the maximum of cells had reached pro-metaphase, they were washed with 3% sucrose and transferred to new culture medium. The efficiency of synchronization as determined by the mitotic index was usually 80% after washing out the propyzamide. With aphidicolin alone (‘simple synchronization’) the mitotic index was about 50%.
Preparation of microsomes, protein determination and HMGR assay
Cells were frozen and kept at –80°C, then ground in a mortar in the presence of liquid N2. The powder was suspended in 2 ml g–1 FW in a 4°C cold phosphate buffer system and microsomes were isolated essentially as described previously (Bach et al. 1986). Microsomal fractions were supplied with DTE to > 20 mm and stored at –80°C. Proteins were determined by a Lowry method (Bensadoun and Weinstein 1976). HMGR activity was determined as described elsewhere (Bach et al. 1986). For Western blot analysis of HMGR 40 μg of microsomal proteins per slot were separated in 12% polyacryamide gels containing 1% SDS (Laemmli 1970), using a Protein II slab cell (BioRad). Proteins were electroblotted onto nitrocellulose membranes (Amersham) as described by Towbin et al. (1979) and immunostained using polyclonal antibodies raised against the soluble (catalytic) domain of radish HMGR, as described elsewhere (Vollack et al. 1994).
We wish to thank Professor A.-M. Lambert and Dr A.-C. Schmitt (IBMP, Strasbourg) for introducing us to basic techniques of cell synchronization. We are grateful to Dr R. Bronner (IBMP) for introducing us to the microscopic methods. We also thank Drs J.-P. Ghnassia (Centre Paul Strauss, Strasbourg) and G. Jung (Hôpital Emile Muller, Mulhouse) for letting us use their instrumentation. We are indebted to Drs A. W. Alberts and M. D. Greenspan (Rahway, NJ) for the generous gift of mevinolin, and to Professor T. Nagata (Tokyo) for permission to use TBY-2 cells. We appreciate the stimulating discussions with Dr S. C. Brown (Gif-sur-Yvette) and Drs C. Gigot, N. Chaubet and M.-A. Hartmann (IBMP) and are indebted to Mr J. Ackerson for reading the English manuscript.This study was made possible by a PhD fellowship of the Ministère de l’Education Nationale, de l’Enseignement Supérieur et de la Recherche, to A. Hemmerlin
e-mail address: firstname.lastname@example.org).
Essential results of this study have been presented at the EMBO Workshop Control of Cell Division Cycle in Higher Plants, Szeged (Hungary), 5–7 October 1995, and at the Annual Meeting of the ASPP, San Antonio, 27–21 July 1996.