Specific checkpoints regulate plant cell cycle progression in response to oxidative stress

Authors

  • Jean-Philippe Reichheld,

    1. 1Laboratorium voor Genetica, Departement Genetica, Vlaams Interuniversitair Instituut voor Biotechnologie, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium,2Department of Medicine, Laboratorium voor Hematologie, Universitaire Instelling Antwerpen (UIA), B-2610 Wilrijk, Belgium, and3Laboratoire Associé de l’Institut National de la Recherche Agronomique (France), Universiteit Gent,B-9000 Gent, Belgium
    Search for more papers by this author
  • 1 Teva Vernoux,

    1. 1Laboratorium voor Genetica, Departement Genetica, Vlaams Interuniversitair Instituut voor Biotechnologie, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium,2Department of Medicine, Laboratorium voor Hematologie, Universitaire Instelling Antwerpen (UIA), B-2610 Wilrijk, Belgium, and3Laboratoire Associé de l’Institut National de la Recherche Agronomique (France), Universiteit Gent,B-9000 Gent, Belgium
    Search for more papers by this author
  • 1 Filip Lardon,

    1. 1Laboratorium voor Genetica, Departement Genetica, Vlaams Interuniversitair Instituut voor Biotechnologie, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium,2Department of Medicine, Laboratorium voor Hematologie, Universitaire Instelling Antwerpen (UIA), B-2610 Wilrijk, Belgium, and3Laboratoire Associé de l’Institut National de la Recherche Agronomique (France), Universiteit Gent,B-9000 Gent, Belgium
    Search for more papers by this author
  • 2 Marc Van Montagu,

    1. 1Laboratorium voor Genetica, Departement Genetica, Vlaams Interuniversitair Instituut voor Biotechnologie, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium,2Department of Medicine, Laboratorium voor Hematologie, Universitaire Instelling Antwerpen (UIA), B-2610 Wilrijk, Belgium, and3Laboratoire Associé de l’Institut National de la Recherche Agronomique (France), Universiteit Gent,B-9000 Gent, Belgium
    Search for more papers by this author
  • and 1 , Dirk Inzé 1 ,3

    1. 1Laboratorium voor Genetica, Departement Genetica, Vlaams Interuniversitair Instituut voor Biotechnologie, Universiteit Gent, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium,2Department of Medicine, Laboratorium voor Hematologie, Universitaire Instelling Antwerpen (UIA), B-2610 Wilrijk, Belgium, and3Laboratoire Associé de l’Institut National de la Recherche Agronomique (France), Universiteit Gent,B-9000 Gent, Belgium
    Search for more papers by this author

*For correspondence (fax +32 9264 5349;
e-mail mamon@gengenp.rug.ac.be).

Summary

The effects of oxidative stress on plant cell cycle progression were studied both in cell suspensions andin planta. Oxidative stress of variable severity was imposed by the addition of different concentrations of the methyl-quinone, menadione, into the growth media. In cell suspensions, flow cytometry analyses demonstrated that low concentrations (20–50 μM) of menadione impaired the G1/S transition, slowed DNA replication, and delayed the entry into mitosis. Furthermore, cells in G1 were more sensitive to menadione-mediated oxidative stress than cells in S phase. Cell cycle arrest was associated with an inhibition of the activity of cyclin-dependent kinases, cell cycle gene expression, and a concomitant activation of stress genes. Menadione-mediated oxidative stress was shown to have very similar effects on tobacco plants, suggesting that a general regulation mechanism takes place in plants. These results define an oxidative stress checkpoint pathway that modulates both the expression of the core cell cycle genes and oxidative defence genes. Redox sensing could be of key importance in controlling cell cycle progression in environmental stress conditions.

Introduction

Aerobic organisms are continuously exposed to oxygen which renders them prone to damage generated by oxygen-derived free radicals. Oxidative stress is largely mediated by the active oxygen species (AOS), including the superoxide anion (O2·–), hydrogen peroxide (H2O2), and the hydroxyl radical (OH·) ( Foyer & Mullineaux 1994). AOS cause severe damage to critical cellular macromolecules including nucleic acids, proteins and lipids ( Halliwell & Gutteridge 1989; Imlay & Linn 1988). To cope with the toxic effects of AOS, cells use a number of defence mechanisms to sense and respond appropriately to oxidative stress. These defence mechanisms involve both low molecular weight antioxidants (glutathione, ascorbic acid, α-tocopherol) and antioxidant enzymes (superoxide dismutases, catalases, reductases, peroxidases) (for a review, see Inzé & Van Montagu 1995).

Another important response of cells exposed to oxidative stress is their capacity to arrest cell division. In animal and yeast cells, checkpoints have been identified that lead to cell cycle arrest after DNA damage (for reviews, see Jackson 1996; Lydall & Weinert 1996; Murray 1992; Paulovich et al. 1997) or inhibition of DNA replication (for a review, see Stewart & Enoch 1996). Indeed, the mechanisms triggering the oxidative stress-induced cell cycle arrests were shown to be similar to those induced by DNA-damaging agents, delaying transitions from G1 to S, G2 to M or inhibiting DNA replication ( Clopton & Saltman 1995; Corroyer et al. 1996; Russo et al. 1995). The delays in the progression of the cell cycle presumably allow time for the repair processes to take place or for completion of a critical cell cycle event. In animal and yeast cells, the DNA damage-induced checkpoints are controlled by specific genes, and mutations in checkpoint genes often result in increased sensitivity to damaging agents. Moreover, the checkpoint genes are commonly mutated in different cancers, emphasizing their importance in the maintenance of the cell cycle (for a review, see Hartwell & Kastan 1994). One of these genes, coding for the p53 protein, harbours mutations in more than half of human cancers ( Kastan 1996). This protein takes part in the G1 arrest in response to DNA damage. The DNA damage-induced arrests in the G1 and S phases may partly involve inhibition of the activity of G1 cyclin-dependent kinases (CDKs) and compounds of the DNA replication machinery by specific CDK inhibitors ( Xiong et al. 1993). Furthermore, the mechanism underlying the G2 arrest was shown to involve a specific inhibitory phosphorylation of the mitotic kinase, CDK1, in human cells ( Jin et al. 1996; O’Connor et al. 1993).

In plants, very few data are available on the possible link between the stress response and cell cycle progression. Logemann et al. (1995) showed a tight correlation between down-regulation of cell cycle genes and up-regulation of defense genes following UV irradiation or elicitor treatment in parsley cells. To gain a better understanding of how oxidative stress influences the plant cell cycle, it is important to determine the effects of stress on cell cycle progression and the relative sensitivity of different cell cycle phases to stress. We therefore studied the effects of oxidative stress on the plant cell cycle by using two different systems. Tobacco BY-2 cell suspensions were used because of their highly synchronisable character, while tobacco plants were used as a simple in planta model. To generate oxidative stress, we used 2-methyl-1,4-naphthoquinone (menadione, MD). Quinones that undergo redox cycling have been widely used to investigate oxidant-induced stress in cells. Two major mechanisms of quinone-induced cytotoxicity have been identified. First, quinones are reduced to hydroquinones or semiquinone radicals by cellular reductases at the expense of NADPH. Both hydroquinones and semiquinone radicals undergo rapid autoxidation with the regeneration of the parent quinones and formation of the AOS, superoxide, and H2O2. Second, quinones react with cellular nucleophiles, such as thiols or amines, which may impair cellular functions. Menadione, which has a –OCH3 group at its C2 position, cannot conjugate with glutathione or other compounds, and its toxicity is apparently mediated solely via the generation of AOS ( Halliwell & Gutteridge 1989; Shi et al. 1994).

Both in BY-2 cell suspension and tobacco plants, cells have a dose-dependent sensitivity to oxidative stress. At a low concentration of menadione, the cell cycle is transiently blocked. At higher concentrations, cell death is induced. Taking advantage of the highly synchronisable character of the BY-2 suspensions ( Nagata et al. 1992), we could detect specific cell cycle blocks at the G1/S and G2/M transitions and a reduction in DNA replication. The cell cycle arrest coincided with the down-regulation of cell cycle gene expression and with the inhibition of CDK activities. This down-regulation was concomitant with the induction of expression of stress genes. Furthermore, cells in the G1 phase were more sensitive to oxidative stress than cells in the S phase. Finally, we observed a similar effect of oxidative stress in tobacco plants, demonstrating that specific G1 and G2 phase arrests is a general feature that regulates the cell cycle progression upon oxidative stress. The implications of these different findings for the modulation of the cell cycle in response to environmental stimuli are discussed.

Results

Mild oxidative stress causes a transient arrest of cell growth

To have a first insight into the effect of oxidative stress on BY-2 cells, exponentially growing (3 days after refreshing) BY-2 cells were incubated with different concentrations of menadione ( Fig. 1). After addition of menadione, the level of DNA synthesis ( Fig. 1a), of cell suspension growth ( Fig. 1b), and of cell death ( Fig. 1c) were followed in parallel. The untreated control cells grew exponentially and showed a nearly constant level of DNA synthesis ( Fig. 1b). Incubation with 50 or 100 μm menadione led to a rapid decrease of the DNA replication that occurred within the first hour of the treatment ( Fig. 1a), showing a strong effect of oxidative stress on DNA replication. At the same time, a complete inhibition of the growth of the cell suspension was also observed ( Fig. 1b). Both concentrations of menadione caused cell death as revealed by the uptake of Evans Blue after 6 h of treatment, and this increased further after 24 h ( Fig. 1c). Microscopic observations of the Evans Blue-stained cells confirmed that the dose-dependent toxic effect was a direct effect of the sensitivity of cells to menadione treatment (data not shown). In contrast, cells exposed to 20 μm of menadione showed a progressive inhibition of DNA synthesis during the first 12 h of the treatment. This inhibition was transient as DNA replication was restored to a level comparable to control cells after 24 h ( Fig. 1a). Furthermore, the delay in cell division was confirmed by the 24 h lag observed in the growth of the treated cell suspension ( Fig. 1b) and by a transient decrease of the mitotic index (data not shown). As confirmed by cell viability measurements, exposure to 20 μm menadione did not have a cytotoxic effect ( Fig. 1c). Exposure of cells to lower menadione concentrations (0.5–10 μm) caused no visible effects on cell growth and did not modify DNA synthesis (data not shown).

Figure 1.

Dose-dependent effects of menadione on BY-2 cells.

Exponentially growing BY-2 cells (3-day-old) were mock-incubated (○) or exposed to different concentrations of menadione (MD 20 μm, • MD 50 μm, □ MD 100 μm, ▪). (a) DNA synthesis level, (b) packed volume of cell suspension and (c) cell death were determined at the indicated time points. Values are mean of three independent experiments.

Mild oxidative stress impairs the G1/S transition

To define in more detail the effect of oxidative stress on cell cycle progression, menadione was applied to BY-2 cells synchronized in G1 phase by an aphidicolin/propyzamide treatment, as described in Experimental procedures. Two hours after release from the propyzamide treatment, menadione was added at the previously described concentrations. Cell cycle progression was followed by flow-cytometrical analysis ( Fig. 2) and the level of DNA synthesis and the mitotic index were estimated ( Fig. 3a). Control cells progressed from the G1 phase to the S phase and subsequently to the G2 and M phases ( Fig. 2c), while cells treated with 20 μm menadione showed a delay in the progression of cells into the S phase. These data demonstrate a transient arrest of the cell cycle at the G1/S transition ( Fig. 2; MD 20 μm). According to the shift that was observed in the peak of DNA synthesis and in the onset of mitosis, the delay was estimated to be approximately 3 h ( Fig. 3a).

Figure 2.

Cell cycle progression of G1 phase cells exposed to menadione.

BY-2 cells synchronized in G1 phase (2 h after release from the propyzamide block) were mock-incubated (C) or exposed to different concentrations of menadione (MD). At the indicated times of incubation, cells were prepared and analyzed by flow cytometry.

Figure 3.

Effects of menadione on DNA synthesis, mitotic index and gene expression in G1 cells.

BY-2 cells were incubated under the same conditions as described in Fig. 2.

(a) The DNA synthesis rate (full lines) and mitotic index (dashed lines) were determined after propyzamide release (T0) in control cells (○) and in cells treated with menadione (MD 20 μm, • MD 50 μm, □ MD 100 μm, ▪) at the time indicated (arrow). Values are the mean of three independent experiments.

(b) Total RNAs (15 μg) isolated at the indicated times were blot-hybridized with different DNA probes, as described in Experimental procedures.

To study the effect of menadione on cell cycle gene expression, we followed the steady-state level of different cell cycle markers. Histone H4 is an S phase-specific gene. The CycA3;2 gene was shown to be expressed during a broad window going from the G1/S to the G2/M transition whereas CycA1;1 is specifically expressed from the S/G2 to the end of mitosis ( Reichheld et al. 1996 ). As shown in Fig. 3(b), the induction of the three genes was delayed in cells treated with 20 μm of menadione as compared to the control cells. In particular, the lag of induction of the CycA3;2 gene, which is assumed to play a function at the entry of the S phase ( Reichheld et al. 1996 ), strongly supports a transient cell cycle arrest occurring at the G1/S transition ( Fig. 3b). Interestingly, the cell cycle arrest was concomitant with the induction of defence genes, such as glutathione peroxidase (GPX) or PRB-1b genes. However, the cytosolic Cu/Zn superoxide dismutase (Cu/ZnSOD) is hardly induced ( Fig. 3b).

To test the effect of oxidative stress on CDK/cyclin activities, protein extracts were prepared from both control and stressed cells. Kinase activities of p9CKShs1 affinity-purified CDKs were determined in vitro by measuring the level of histone H1 phosphorylation ( Fig. 4, H1 p9CKS). In plant cells, p9CKShs1 binds at least two classes of CDKs (A-type and B-type) ( De Veylder et al. 1997 ; G. Segers, unpublished results). In control cells, the level of phosphorylated H1 had a basal level in G1 phase (0 h, C), increased at the G1/S transition (1–3 h, C), and peaked at the G2/M transition (9 h, C). We also measured the kinase activity of the specific CDC2aNt kinase purified by immunoprecipitation ( Fig. 4, H1 CDC2a). A similar kinetic of kinase activity was observed for this specific CDK ( Fig. 4, H1 CDC2a). Compared to the p9CKShs1-associated kinase activities, CDC2a activity peaked slightly earlier in S to the mid-G2 phase (6–9 h). CDC2a also showed a reduced kinase activity during the early M phase. This difference is probably due to additional kinases associated with p9CKShs1 (i.e. B-type CDK). These data correlated with previously described kinase activities of the alfalfa CDC2a homologues, CDC2A and CDC2B, which were proposed to play a dual role in the onset of S phase and in the entry into mitosis ( Magyar et al. 1997 ). In cells treated with 20 μm of menadione, the induction of kinase activities of both p9CKShs1-associated kinases and CDC2aNt kinase were delayed by approximately 3 h ( Fig. 4, MD20), confirming the shift in the cell cycle progression. Looking at the potential role of CDC2a in the onset of the S phase, the impairment of stressed cells to initiate DNA replication may be due to a specific inactivation of the CDC2a kinase. As expected, an antibody raised against the CDC2aNt protein recognized a 34 kDa protein in the p9CKShs1 affinity-purified protein extracts ( Fig. 4, CDC2a). Interestingly, the induction of the p9CKShs1-associated kinase activities was not accompanied by variations of the steady-state level of the CDC2a protein. These data confirm the constitutive expression of A-type CDKs during the cell cycle ( Magyar et al. 1997 ). The level of CDC2a was not altered under stress conditions ( Fig. 4, MD20) indicating that the effects of oxidative stress were mediated primarily through modifications in the specific activity of this protein.

Figure 4.

Effects of menadione on CDK kinase activities in G1 cells.

BY-2 cells were incubated under the same conditions described in Fig. 2. Total proteins were extracted at the times indicated after the addition of menadione. p9CKShs1-associated and CDC2aAt-specific kinase activities (H1 p9CKShs1 and H1 CDC2a) were determined by their ability to phosphorylate histone H1. Immunoblotting of p9CKShs1-bound proteins was performed with anti-CDC2aNt antibodies (CDC2a).

Exposure of cells to 50 or 100 μm menadione led to extensive cell death. The disappearance of the flow-cytometrical signals reflected genomic DNA degradation ( Fig. 2, MD50 and MD100). This observation was confirmed by loading genomic DNA of these cells on agarose gels. Unspecific DNA degradation revealed by a smear appeared after 2–4 h in cells treated with either 50 and 100 μm menadione (data not shown). Treatment of cells with 50 μm ( Fig. 3b; 50) and 100 μm (data not shown) menadione led to a complete disappearance of all mRNAs of cell cycle and stress defence genes as well as a complete inhibition of the CDK kinase activities associated with disappearance of the CDC2a protein ( Fig. 4, MD50).

Oxidative stress slows the S phase and delays entry into mitosis

To confirm the existence of a specific checkpoint occurring at the onset of the S phase in response to oxidative stress, we analyzed the effects of menadione on cells in early S phase (i.e. after the G1/S checkpoint). In this case, menadione was applied 7 h after release from propyzamide ( Fig. 5a). In contrast to cells stressed in the G1 phase, 20 μm menadione had a minor effect when added to cells during the progression through the S phase ( Fig. 5a), which strongly suggested that most of the cells have overrun the cell cycle arrest. This was also confirmed by flow cytometry that showed only slight differences between control cells and 20 μm menadione-treated cells ( Fig. 6). Cells treated with 50 or 100 μm menadione, however, were highly perturbed in their DNA replication, but the kinetics of inhibition were very different between both treatments ( Fig. 5a). The DNA synthesis level of cells treated with 50 μm menadione decreased slowly, whereas an exposure to 100 μm menadione led to a rapid drop (within 1 h) in the DNA replication. This latter observation suggested that a dose-dependent inhibition of DNA synthesis occurs with the two menadione concentrations. As indicated by flow-cytometrical analysis, the rapid drop in DNA synthesis in cells treated with 100 μm menadione reflected cell death, whereas the slow decrease in DNA synthesis in cells exposed to 50 μm menadione was due to a partial cell cycle arrest ( Fig. 6). By analyzing in more detail the flow-cytometrical data, two cell populations could be detected. One population represents cells that were capable of completing DNA replication and reached G2 phase after 7 h of treatment. The second population is comprised of cells that remain arrested ( Figs 6, 7h, MD50, G1 peak). Furthermore, no increase in the mitotic index was detected after 7 h of treatment with 50 μm menadione, excluding the possibility that the G1 peak would represent new post-mitotic cells ( Fig. 5a). Presumably, the first population represents cells that have escaped the effect of menadione treatment because they have already passed the checkpoint and are able to progress through the S phase. In the second population, cells are slightly upstream from the checkpoint and are still sensitive to G1/S arrest. Interestingly, 50 μm menadione had no toxic effect on both populations of cells ( Fig. 7; S) indicating that, unlike G1 phase cells ( Fig. 7; G1), both G1/S and S cells were resistant to this concentration of menadione.

Figure 5.

Effects of menadione on DNA synthesis, mitotic index, and gene expression in early S cells.

Cells were incubated under the same conditions as described in Fig. 3.

(a) DNA synthesis rate (full lines) and mitotic index (dashed lines) in control cells (○) and in cells treated with menadione (MD 20 μm, • MD 50 μm, □ MD 100 μm, ▪) 7 h after the release from propyzamide (arrow). Values are the mean of three independent experiments.

(b) Total RNA blot hybridizations were performed as described in Fig. 3.

Figure 6.

Cell cycle progression of early S phase cells exposed to menadione.

BY-2 cells synchronized in the early S phase (7 h after a release from propyzamide) were mock-incubated (C) or exposed to different concentrations of menadione (MD). At indicated times of incubation, cells were prepared and analyzed by flow cytometry.

Figure 7.

Menadione-induced cell death in G1 and early S phase BY-2 cells.

BY-2 cells synchronized in G1 (G1) or in early S phase (S) were mock-incubated (C, ○) or exposed to different concentrations of menadione (MD 20 μm, • MD 50 μm, □ MD 100 μm, ▪). Cell death assayed by Evans Blue staining were quantified by spectrophotometry at indicated times. Values are the mean of three independent experiments.

Expression of all cell cycle genes studied was affected by exposure to 50 μm menadione, confirming the modification of cell cycle progression in these cells ( Fig. 5b). As described previously, the inhibition of the CycA3;2 gene was indicative of a G1/S arrest. Moreover, the inhibition of expression of the CycA1;1 gene, which may function at the S/G2 and G2/M transitions ( Reichheld et al. 1996 ), suggests an additional block of these latter checkpoints ( Fig. 5b). Interestingly, minor modifications in the expression of cyclin genes were also detected in cells treated with 20 μm menadione, hinting that some cells were also affected by this weaker treatment, possibly the population of cells that was blocked by 50 μm menadione. Here again, cell cycle arrest correlated with an induction of GPX and PRB-1b gene expression and, later on, of Cu/ZnSOD.

We determined whether S phase progression and the onset of mitosis were further affected by a menadione treatment of cells in the mid-S phase ( Fig. 8). Menadione (50 μm) was applied to cells in the mid-S phase and cell cycle progression was measured by flow-cytometrical analysis ( Fig. 8). Compared to control cells, in which DNA replication was completed after 3 h of treatment ( Fig. 8,C), DNA synthesis was strongly reduced in stressed cells ( Fig. 8, MD 50 μm), reaching the G2 phase after 9 h of treatment. Furthermore, although DNA replication was completed, entry into mitosis was further delayed as only 5% of mitosis was detected after 12 h of treatment. Interestingly, no cell death was detected in these cells indicating that, unlike G1 phase cells, S and G2 phase cells are resistant to the toxic effect of 50 μm menadione (data not shown).

Figure 8.

Cell cycle progression of mid-S phase cells exposed to menadione.

Cells were released from an aphidicolin treatment. S phase was allowed to proceed without perturbation in control cells (C) or treated in mid-S phase (1 h after the aphidicolin release) by 50 μm menadione (MD 50 μm). At indicated times after the treatment, cells were prepared and analyzed by flow cytometry. Data from a representative experiment are shown.

Effects of oxidative stress on tobacco meristematic activity

To extend the observations made on cell suspensions, we investigated the effects of menadione on meristematic activities in tobacco plantlets. Five-day-old tobacco seedlings were exposed to menadione. Root growth was measured during the 5 days following the application of the drug ( Table 1). The primary roots from control plants increased in length by 4 mm per day. Exposure to 20 μm of menadione inhibited root growth by 46% after 1 day. Cell viability measured by fluoresceine diacetate (FDA) staining revealed no toxic effects of this treatment, excluding the possibility that the apparent inhibition of root growth would be due to cell ablation (data not shown). Moreover, flow-cytometrical analyses revealed modifications of the nuclear DNA content in the root tip cells upon oxidative stress. The percentage of cells in the S phase in the root tip of stressed plants was 39% of the control and cells accumulated in G2. These data strongly suggest that oxidative stress inhibits cell division in the root meristem via specific cell cycle arrests at G1 and G2 phases, as previously observed in tobacco cell suspensions. The same observation was made in the shoot apical meristem (SAM) where the percentage of S phase cells was 31% in stressed plants, compared to control plants. The growth was progressively resumed to 70% (data not shown) and 86% of the control after 2 and 5 days, respectively. This observation suggests an adaptive behaviour of plants to oxidative stress. After 5 days, the percentage of cells in the S phase in the root tip of stressed plants recovered to 75% of the control plants, confirming this rescue effect ( Table 1). The same phenomenon was observed in the distribution of the nuclear DNA content in the shoot apical meristem, suggesting that the adaptation occurred in all meristematic zones of the plant ( Table 1). Exposure to 40 μm menadione led to total inhibition of the root elongation. Cell viability measured by FDA staining revealed extensive cell death under these conditions (data not shown).

Table 1.  Effect of menadione on tobacco plantlets
DNA content
Root tipSAM
DayTreatmentsRoot growth (mm day–1) % G1% S% G2% G1% S% G2
  1. Average root growth (n = 20) and DNA content (pools of eight plants) were measured at different times after transfer to a medium supplemented with menadione (MD) or not (Control). SAM, shoot apical meristem. Data from a representative experiment are given.

1Control4.29 ± 0.7855.819.923.773.715.012.7
 20 μm MD2.30 ± 0.6251.77.740.380.34.715.7
5Control3.70 ± 0.4241.314.745.057.319.723.3
 20 μm MD3.19 ± 0.7242.311.047.363.619.917.0

Discussion

In the present study, we examined the impact of oxidative stress on the cell cycle progression using the redox-cycling drug menadione. We showed that mild oxidative stress leads to inhibition of cell division both in cell suspension and plants. Exposure to menadione of cells in G1 induces a transient G1 block, as revealed by the flow-cytometrical analyses, and inhibition of the induction of DNA synthesis ( Figs 2 and 3). Moreover, mild oxidative stress during the S phase slows down progression into the S phase and delays entry into mitosis, transiently blocking cells in G2 ( Figs 5, 6 and 8). These results reveal the existence of specific mechanisms responsible for transient cell cycle blocking at different checkpoints in response to sublethal menadione-induced oxidative stress. In planta, the growth arrest was accompanied by modifications of the nuclear DNA content in meristematic zones, showing accumulation of cells in the G1 and G2 phases ( Table 1). These observations may define a general oxidative stress checkpoint in plants to modulate the meristematic activity under oxidative stress conditions. Such a pathway closely resembles those existing in animal cells. Indeed, similar results were obtained in animal cells where G1 and G2 arrests, as well as inhibition of DNA replication, occur in response to oxidative stress ( Clopton & Saltman 1995; Russo et al. 1995) due to activation of specific signalling pathways ( Russo et al. 1995). Although these pathways are still largely unexplored in plants, they could be very similar to those occurring after DNA damage in animal and yeast cells. The human ATM gene, as well as its yeast homologue MEC1, were shown to trigger different pathways leading to G1/S, S and G2/M arrests in response to DNA damage (for a review, see Elledge 1996). Possibly, analogous proteins that co-ordinate stress-induced cell cycle arrest exist in plants.

To determine possible effectors of the oxidative stress-induced cell cycle block, we studied the transcription of several cell cycle genes in oxidative stress conditions ( Figs 3b and 5b). Concomitant with the arrest of the cell cycle, we observed in all treatments a general inhibition of the expression of genes involved in cell cycle, such as histone H4 and A-type cyclins (CycA3;2 and CycA1;1). Cyclins are crucial components of the cell cycle machinery as they bind and activate CDKs. Their synthesis and degradation were shown to be highly regulated during the cell cycle ( Evans et al. 1983). Therefore, the inhibition of expression of A-type cyclins that we observed under oxidative stress conditions may reveal a transcriptional mechanism that regulates cell cycle progression. The two A-type cyclins we used were assumed to play distinct roles during the cell cycle ( Reichheld et al. 1996). Inhibition of their synthesis could therefore account for the distinct cell cycle arrest we observed (i.e. G1 and G2 arrests). Nevertheless, the role of these A-type cyclins is still speculative and was mostly deduced by studying their respective expression pattern during the cell cycle ( Reichheld et al. 1996). Therefore, we do not know if the inhibition of their expression actually takes part in the mechanism triggering the oxidative stress-induced cell cycle arrest or if it is a consequence of this arrest.

We also analyzed whether modification of CDK activities is related to cell cycle arrest and we demonstrated inhibition of the p9CKShs1-associated kinase activities under mild stress conditions ( Fig. 3b). Planchais et al. (1997) reported significant p9CKShs1-associated kinase activities in BY-2 cells in late-G1 phase. Therefore, we can postulate that the inhibition of p9CKShs1-associated kinase activities might prevent stressed cells entering the S phase. We showed that p9CKShs1 binds the CDC2a kinase. In alfalfa and Arabidopsis, A-type CDKs were suggested to play a role in the G1/S transition ( Hemerly et al. 1995; Magyar et al. 1997) and inhibition of their activity might therefore be a way to prevent entry into the S phase under oxidative stress conditions. Such a statement would fit with data from animal cells showing that oxidative stress inhibits CycA/CDK2 kinase activities through a p21-dependent pathway ( Corroyer et al. 1996). Alternatively, the G1/S arrest could be under the control of another G1-specific CDK yet to be identified in plants, and the inhibition of the CDC2a and p9-associated kinase activities would be the result of this arrest. It is an open question whether other p9CKShs1-associated CDKs (i.e. B-type CDKs) are inhibited by oxidative stress. The constant level of CDC2a protein that we observed in the p9CKShs1-purified fraction, in contrast to the reduced H1 kinase activity, is strongly indicative of a post-transcriptional mechanism regulating the specific activity of this CDK under mild oxidative stress conditions ( Fig. 4). This regulation is possibly due to a lack of expression of the associated cyclin but additional mechanisms can be postulated. Among these are the induction of CKIs or involvement of CDK-inhibitory phosphorylations.

Interestingly, in animal cells, oxidative stress has previously been shown to induce the expression of the p21WAF1/CIP1 protein ( Corroyer et al. 1996; Russo et al. 1995). p21 is a potent inhibitor of the G1/S CDKs ( Harper et al. 1993) and can inhibit DNA synthesis by interfering with the proliferating cell nuclear antigen (PCNA), a protein implicated in the DNA replication process ( Waga et al. 1994). The potential implication of such a p21-homologue may explain both the cell cycle arrest in G1/S and the inhibition of DNA replication within the S phase. Recently, the first CKI homologue was isolated from Arabidopsis ( Wang et al. 1997), opening up new prospects in the study of cell cycle checkpoints in plants. In the same way, G2/M arrest following DNA damage has been linked to a failure to remove inhibitory phosphorylations from the ATP-binding domain of the CDC2 kinase ( O’Connor et al. 1993). Plant CDKs harbour potent inhibitory phosphorylation residues in these sequences and were shown to be controlled by such mechanisms ( Zhang et al. 1996). Therefore, it would be interesting to study whether such regulation mechanisms are acting under oxidative stress.

We found that the down-regulation of cell cycle genes in mild oxidative stress conditions correlates with an activation of defence genes such as GPX, PRB-1b and the Cu/ZnSOD ( Figs 3b and 5b). This response is probably indicative of the adaptive behaviour of cells to prolonged oxidative stress that consists of an increase in the defence capacities. In the same way, exposure of mammalian or plant cells to a prolonged sublethal oxidative stress leads to a sustained elevation of the intracellular glutathione pool ( Shi et al. 1994; J.P. Reichheld and T. Vernoux, unpublished results). Thus, redox homeostasis could possibly be a major signalling mechanism for cell division. Very similar results in parsley cell suspensions exposed to UV light or fungal elicitors were obtained by Logemann et al. (1995), who postulated that cell division inhibition and defence gene induction could be triggered by a common regulatory pathway. Such pathways that inhibit cell cycle progression and activate DNA repair system in response to DNA damage were largely studied in yeast and animal cells ( Hartwell & Kastan 1994; Navas et al. 1996). Adaptive responses can probably account for the transient arrest of tobacco root in response to mild oxidative stress. Such an adaptive response was reported in faba beans and sunflower roots submitted to water stress and meristematic activity recovered after a few days ( Robertson et al. 1990; Yee & Rost 1982). Here, we provide the first evidence that such adaptive responses also occur under oxidative stress conditions.

Measurements of cell viability revealed that cells in G1 are more sensitive than cells in the S phase to oxidative stress ( Fig. 7). The G1 phase is known to be particularly sensitive to environmental conditions. At the restriction point of mammalian cells (or START in yeast), cells switch on specific programmes that induce cell division, differentiation or cell death, depending on external signals (for a review, see Elledge 1996). In animals, oncogenes, such as ras, c-jun, c-fos, p53 and other members of signalling pathways, such as MAP kinases, are acting during the G1 phase to push cells through the restriction point. Furthermore, proteins such as c-jun, c-fos and p53 were shown to be sensitive to redox regulation of their activity. Although none of these proteins have yet been isolated in plants, an interesting hypothesis is that pathways which regulate progression through the cell cycle during the G1 phase are highly sensitive to a perturbation of the redox homeostasis of the cell. The fact that glutathione depletion impairs the G1/S transition and blocks cells into G1 phase strongly supports this idea (T. Vernoux, unpublished results). The importance of glutathione during the response to an environmental stress suggests the existence of such transduction pathways sensitive to the intracellular concentration of glutathione (M. May, unpublished results). Such pathways may have strong implications for environmental sensing and could at least partially account for our observations showing that glutathione concentration is modulated in response to oxidative stress.

In conclusion, we demonstrate that plants respond to oxidative stress by modulating the growth rates and inducing specific defence mechanisms. Possibly, this dual response to oxidative stress could mirror an evolutionarily conserved response to environmental stresses and thereby provide a good model to study the molecular events induced by specific environmental stresses that influence cell division.

Experimental procedures

Plant material, synchronisation and treatments

The tobacco BY-2 (Nicotiana tabacum L. cv. Bright Yellow 2) suspension was maintained by weekly dilution (1.8/100) of cells in fresh Murashige and Skoog (MS) medium modified according to Nagata et al. (1992 ) and cultured at 28°C and 0.6 g in the dark.

Synchronisation of cells was performed as described previously ( Reichheld et al. 1995 ). A stationary culture was diluted 1/8 in fresh medium supplemented with 4 μg ml–1 aphidicolin (Sigma, St. Louis, MO, USA). After 24 h of culture, the drug was removed by extensive washes and the cells were resuspended into fresh medium. When analyses of post-mitotic stages were needed, 0.8 μg ml–1 of propyzamide (Sumitomo Chemical Co., Tokyo, Japan) was added before the pre-prophase stage, approximately 4 h after aphidicolin release, maintained for 4 h and removed by extensive wash.

DNA synthesis was determined by pulse-labelling with [3H]-thymidine as described by Reichheld et al. (1995 ). Mitotic index was determined by UV light microscopic analysis of 500 cells stained with 0.1 μg ml–1 4′,6-diamino-2-phenylindole (Sigma) in the presence of 0.2% Triton X-100. Cell viability was measured as described previously ( Levine et al. 1994 ). One millilitre of cell suspensions were incubated with 0.05% Evans Blue (Sigma) for 15 min and then extensively washed to remove unbound dye. Dye bound to dead cells was solubilized in 50% methanol with 1% SDS for 30 min at 50°C and quantified by absorbance at 600 nm. The selective staining of dead cells with Evans Blue depends upon exclusion of this pigment from living cells by the intact plasmalemma, whereas it passes through the damaged membrane of dead cells and accumulates as a blue protoplasmic stain ( Turner & Novacky 1974).

Menadione (2-methyl-1,4-naphthoquinone; Sigma) treatments were performed by adding aliquots of a freshly prepared 50 m m stock solution.

Seeds of Nicotiana tabacum SR1 were germinated on vertical 0.6% (wt/vol) agarose plates ( Estelle & Somerville 1987) for 5 days at 24°C under a 16 h light/8 h dark regime. For treatment, seedlings were carefully transferred to plates containing the above medium supplemented with menadione and grown for a further 5 days.

Root length measurements and fluorescence microscopy

Upon transfer to the plates containing menadione, the position of the root tip was marked on the underside of the plate. Every day after the transfer, root growth was measured on 20 roots per treatment using a binocular microscope (Zeiss, Oberkochen, Germany) ( × 6 magnification) with the aid of an eyepiece graticule. For vitality staining, plants were incubated for 2 min in a 0.01% (wt/vol) fluoresceine diacetate (FDA) solution, rinsed in H2O, mounted in H2O and observed by UV light (λex = 365 nm, λem = 420 nm).

Nuclei purification and flow-cytometrical analysis

Tobacco BY-2 protoplasts were obtained from 5 × 104 cells by incubation with 1 ml of an enzyme solution (2% cellulase Onozuka R10, 0.1% pectolyase (Kikkoman Co., Tokyo, Japan), 0.66 m sorbitol) for 1 h at 37°C. After treatment, the enzyme solution was removed by centrifugation (5 min, 1000 g). Protoplasts were washed once with MS medium supplemented with 4.5% mannitol and nuclei were released from the protoplast pellet in Galbraith’s buffer ( Galbraith et al. 1983 ). After addition of 1% formaldehyde, nuclei were stored at 4°C until further analysis by flow cytometry. Before analysis, nuclei were filtered through a 25 μm nylon filter, treated with RNase A, and stained with propidium iodide (50 μg ml–1). Cytometrical analyses were performed on 104 nuclei with a FACS scan flow cytometer (Bio-Rad, Hercules, CA, USA).

For flow cytometry analysis of nuclear DNA content in tobacco plantlets, eight plants were pooled. Root tips (approximately 2 mm) and shoot apical meristem were chopped with a razor blade in Galbraith’s buffer ( Galbraith et al. 1983 ) and analyzed as described previously ( Gendreau et al. 1998 ).

RNA extraction and Northern analysis

Total mRNAs were extracted from frozen cell suspension pellets using the TRIzol Reagent (Gibco BRL, Gaithersburg, MD, USA) according to the manufacturer’s protocol. Total RNA (15 μg) was electrophoresed on formaldehyde/agarose gels and transferred to Hybond-N membranes (Amersham, Aylesbury, UK). Hybridization was performed at 65°C in a phosphate buffer (500 m m Na2PO4, pH 7.2, 1 m m EDTA, 1% BSA, 7% SDS) to random primed [32P] probes, corresponding to the coding region of H4 A748 gene from Arabidopsis ( Chaboutéet al. 1987 ), the 3′ end of the PRB-1b gene from tobacco ( Eyal et al. 1992 ) and the CycA1;1, CycA3;2 ( Reichheld et al. 1996 ), GPX ( Criqui et al. 1992 ) and Cu/ZnSOD ( Hérouart et al. 1993 ) cDNAs from tobacco. The hybridization signal was visualized through Phosphoimage scanning (Phosphorimager 445SI; Molecular Dynamics, Sunnyvale, CA, USA).

Protein extraction, immunoblotting and histone H1 kinase assays

Protein extracts were prepared by grinding cells with sea sand in homogenization buffer as described by Magyar et al. (1993 ). Protein concentrations were determined using the Protein Assay kit (Bio-Rad). SDS-PAGE gel electrophoresis and Western blots were performed according to standard procedures with primary antibody raised against the carboxyl-terminal peptide of the tobacco CDC2a protein (a gift from Dr P. John) diluted 1/500 and a secondary peroxidase-conjugated antibody (Amersham) diluted 1/5000.

For histone H1 kinase assays, 100 μg of protein extract were incubated with p9CKShs1-Sepharose beads (125 μg) on a rotating platform for 2 h at 4°C. Beads were then washed three times with homogenization buffer and once with kinase buffer ( Magyar et al. 1993 ). For immunoprecipitation, 100 μg protein extracts were pre-incubated for 1 h at 4°C on a rotating platform with 25% (v/v) protein A-Sepharose (Pharmacia, Uppsala, Sweden). After centrifugation, equal amounts of the supernatant proteins were incubated for 4 h with purified antibodies (1/50 diluted anti-CDC2aNt) and subsequently for 1 h with 25% (v/v) protein A-Sepharose beads. Beads were washed as described above. The histone H1 kinase assay was carried out by incubating 25 μl of the packed beads with 0.5 μCi [γ-32P]ATP in the presence of 17.5 μg histone H1 (Sigma) for 20 min at 30°C. Samples were analyzed on 12% SDS-PAGE, stained by Coomassie Blue, and autoradiographed.

Acknowledgements

The authors wish to thank the Tobacco Science Research Laboratory (Japan Tobacco Inc.) for the use of the BY-2 cell suspensions; Drs Marc Davey, Vladimir Mironov, Wim Van Camp, Mike May and Lieven De Veylder for critical reading of the manuscript; Martine De Cock for help preparing it; and Rebecca Verbanck and Karel Spruyt for illustrations. We would also like to thank the laboratory of Dr P. John (Canberra, Australia) for the gift of the tobacco CDC2a antibody. This work was supported by grants from the Interuniversity Poles of Attraction Programme (Belgian State, Prime Minister’s Office – Federal Office for Scientific, Technical and Cultural Affairs; no. 38 and P4/15), the International Human Frontier Science Program (IHFSP RG-434/94M), and the Fund for Scientific Research (Flanders) (G.0121.96). J.P.R. and T.V. are indebted to the European Union for a grant from the Research Training Project (ERBFMBI-CT96-1274) and Rhône-Poulenc Agrochimie for a postdoctoral and a predoctoral fellowship, respectively. D.I. is a Research Director of the Institut National de la Recherche Agronomique (France).

Ancillary