Synchronous Arabidopsis suspension cultures for analysis of cell-cycle gene activity


*For correspondence (fax: +44 1223 334162; email:


Synchronized suspension cultures are powerful tools in plant cell-cycle studies. However, few Arabidopsis cell cultures are available, and synchrony extending over several sequential phases of the cell cycle has not been reported. Here we describe the first useful synchrony in Arabidopsis, achieved by selecting the rapidly dividing Arabidopsis cell suspensions MM1 and MM2d. Synchrony may be achieved either by removing and re-supplying sucrose to the growth media or by applying an aphidicolin block/release. Synchronization with aphidicolin produced up to 80% S-phase cells and up to 92% G2 cells, together with clear separation of different cell-cycle phases. These synchronization procedures can be used for analysis of gene expression and protein activity. We show that representatives of three CDK gene classes of Arabidopsis (CDKA, CDKB1 and CDKB2) show differential expression timing, and that three CDK inhibitor genes show strikingly different expression patterns during cell-cycle re-entry. We propose that ICK2 (KRP2) may have a specific role in this process.


Dispersed plant cell suspension cultures allow the study of cell division in the absence of developmental processes, by providing a homogeneous population of near-identical cells (Gould, 1984). Synchronization of such cultures can provide material representative of specific cell-cycle phases, and this may be achieved either by removal and subsequent re-supply of a compound required for growth, such as phosphate, nitrate, hormones or sucrose (Amino et al., 1983; King et al., 1973; Kodama et al., 1991a; Nishida et al., 1992; Riou-Khamlichi et al., 1999; Riou-Khamlichi et al., 2000), or by applying reversible blocks at different stages of the cell cycle using specific inhibitors (Binarova et al., 1998; Fukuda et al., 1994; Glab et al., 1994; Gould, 1984; Ito et al., 1997; Lucretti and Dolezel, 1995; Magyar et al., 1993; Nagata et al., 1992; Perennes et al., 1993; Planchais et al., 1997; Planchais et al., 2000; Roudier et al., 2000).

However, rather few plant cell suspensions can be synchronized to a high degree, a notable exception which is widely used in cell-cycle studies being the tobacco BY-2 cell line (Nagata et al., 1992; Samuels et al., 1998). Using the drug aphidicolin to inhibit DNA polymerase α and δ (Sala et al., 1980), cells can be arrested in late G1/early S phase, resulting in an S-phase level of 90% cells (Sorrell et al., 2001) and a mitotic index of 45–50%. A wide variety of cell-cycle processes, including cell-cycle phase-specific gene expression (Combettes et al., 1999; Reichheld et al., 1995; Sorrell et al., 1999), cyclin-dependent kinase (CDK) activities (Porceddu et al., 2001; Sorrell et al., 2001) and microtubule rearrangements (Hasezawa and Nagata, 1991) have been studied. Further synchronizable systems include cultures of Acer (King et al., 1974), soybean (Mader and Hanke, 1996), Catharanthus roseus (Amino et al., 1983; Nishida et al., 1992) and alfalfa, in which expression of four CDK homologues has been studied (Magyar et al., 1997).

The resources associated with the completed Arabidopsis genome sequence offer the opportunity for genome-wide studies of cell-cycle regulation. However, few Arabidopsis cell cultures are available, and it is reported to be difficult to achieve synchrony in Arabidopsis (Breyne and Zabeau, 2001; Callard and Mazzolini, 1997; Richmond and Somerville, 2000; Stals et al., 2000). Fuerst et al. (1996) arrested Arabidopsis cells in G1 phase with a low concentration of cycloheximide and obtained a reported mitotic index maximum of 9%, but this was broadly spread over 12 h and separation of S-phase and M-phase cells was not achieved. A combined phosphate starvation followed by aphidicolin block and release produced 50–60% of cells in S phase and G2 phase, but synchrony was lost before mitosis, so that synchrony only persisted for a single phase of the cell cycle (G2) after the blocking of cells in S phase (Callard and Mazzolini, 1997). Riou-Khamlichi et al. (1999, 2000) achieved partial synchrony from G1 into S phase using sucrose removal and re-supply but no further synchrony was observed.

Here we describe synchronous Arabidopsis cultures suitable for following both the re-entry of cells into the cell cycle and progression from S phase following an aphidicolin block/release. In the latter case, a maximum of 80% of cells in S phase and 92% in G2 phase were measured, with observable re-entry into a second S phase. We report on the analysis of the expression and activity of cell-cycle-associated genes, including CDKs, D-type cyclins (CycD) and CDK inhibitors.


Selection of fast growing cell lines

A cell suspension culture of Arabidopsis thaliana ecotype Landsberg erecta with a reported 4C DNA content (Fuerst et al., 1996; May and Leaver, 1993; Riou-Khamlichi et al., 1999; Riou-Khamlichi et al., 2000) was used for the derivation of fast growing cell lines MM1 and MM2d. One derivative of the procedure carried out under continuous illumination was cell line MM1, which had developed a ploidy level of 6C as confirmed by flow analysis (data not shown). MM1 is a well-dispersed culture of green cells (Figure 1b) growing as very small clumps (80% of cells present in clumps of 30 cells or less). MM2d was derived from MM1 by being continuously subcultured for more than two years in darkness. Culture growth was optimized for MM2d by altering temperature, medium composition and hormone concentration. MM2d consists of similarly small clumps of creamy-coloured cells and shows faster growth and reaches a higher cell density than MM1 (Figure 1a). MM2d retains the ability to produce chlorophyll when transferred to cultivation under light. This difference in growth rate is also apparent at the cellular level, as 5 days after subculture, MM2d cells have already initiated the expansion and elongation characteristic of early stationary phase, whereas MM1 cells have not (Figure 1b).

Figure 1.

Growth characteristics of Arabidopsis cell lines MM1 and MM2d.

(a) Early stationary phase (D7) cells were subcultured, and cell growth monitored of cell lines MM1 (□) and MM2d (◆) by determination of the number of cells per ml culture medium each day after dilution. (b) Cell morphology of MM1 (upper panels) and MM2d (lower panels) at day 2 (-2d), day 5 (-5d) and day 7 (-7d) after subculture.

Sucrose starvation-induced synchrony

A low level of synchrony lasting from G1 into S phase may be achieved by sucrose starvation of Arabidopsis cell suspensions (Riou-Khamlichi et al., 1999; Riou-Khamlichi et al., 2000). Removal of sucrose, which provides the carbon source for growth, from an early stationary phase cell suspension (7 days after previous subculture) induces or enhances a reversible arrest of cells primarily in G1 or G0 (quiescent stage). Significantly increased synchrony was achieved by using the selected line MM1, coupled with the introduction of a new dilution step (see Experimental procedures), resulting in partial synchrony from G0/G1 until S phase, with some synchrony persisting until mitosis (Figure 2a). Synchrony was monitored by identifying S-phase cells by the incorporation of bromodeoxyuridine (BrdU) into newly synthesized DNA and its subsequent detection using indirect immunofluorescence with anti-BrdU antibodies (see Experimental procedures). The proportion of BrdU-positive cells was termed the labelling index (LI). The presence of cells in mitosis was detected by scoring the proportion of cells containing highly condensed metaphase and anaphase figures (M/A index), because of the greater certainty in scoring cells in these phases.

Figure 2.

Sucrose starvation-induced synchrony of MM1.

Partial synchronization can be achieved by sucrose starvation of early stationary (D7) Arabidopsis cell line MM1. (a) Labelling index (LI) and M/A index determination to detect S-phase (LI, □) and metaphase/anaphase cells (M/A index, ◆) after sucrose is re-added (t = 0). The M/A index was determined as the proportion of cells with DAPI-stained mitotic figures. (b) Northern blot of RNA samples prepared at 2 h intervals from MM1 cells synchronously re-entering the cell cycle from a G0 state (t = 0), probed as indicated.

This experiment was primarily designed to reveal gene expression as quiescent cells traverse G1 and S phase. However, despite overlap between the S-phase peak (indicated by LI) and the mitotic peak (M/A index), the expression of various marker genes could readily be followed, including those associated with mitosis (Figure 2b). The expression of histone H4 increased strongly between 6 and 10 h, consistent with the proportion of cells entering S phase (Figure 2a,b). Transcripts of the B-type cyclin CYCB1;1 were detected from 10 h and rose strongly between 12 and 14 h, and the peak of cells in metaphase/anaphase was observed after 18 h. A weak CYCB1;1 signal was also observable at 2 h, which may be due to a small proportion of cells arrested in G2 phase after sucrose starvation (data not shown). As previously reported (Riou-Khamlichi et al., 1999; Riou-Khamlichi et al., 2000), mRNA levels corresponding to the D-type cyclins CYCD2;1 and CYCD3;1 increased during traversal of G1 phase. Following sucrose starvation, CYCD2;1 mRNA levels increased within 2 h, whereas those of CYCD3;1 increased only after 6 h at the G1/S boundary (Figure 2a,b).

Representative members of the three main classes of CDK genes known to be involved in cell-cycle regulation were found to exhibit differential expression. The CDKA;1 gene (originally cdc2a;At; Hirayama et al., 1991), which encodes an archetypal CDK containing the PSTAIRE sequence and which is the CDK partner of D-type cyclins (Healy et al., 2001), is expressed in sucrose-starved cells, and its mRNA increases within 2 h of sucrose re-supply. In contrast, expression of CDKB1;1 (originally cdc2b;At; Hirayama et al., 1991) with the variant PPTALRE motif, increases at the onset of S phase (6–8 h) and declines after 16–18 h when the M/A index peak is reached. Expression of CDKB2;2, with the PPTTLRE motif (Huntley and Murray, 1999), increases only after 12–14 h.

The expression of genes encoding three CDK inhibitors (de Veylder et al., 2001; Lui et al., 2000; Wang et al., 1998) was also examined (Figure 2b). ICK2 (KRP2) is strongly expressed 2 h after cells re-enter the cell cycle, after which expression rapidly decreases, which could indicate a possible role during reactivation of cell division in early G1 phase. In contrast, expression of the CDK inhibitor KRP3 increases from 2 h and reaches a maximum level by 12 h, corresponding to predominantly G2 cells (Figure 2a,b), consistent with a regulatory role at the second principal control point between G2 and mitosis. ICK1 (KRP1) showed high expression after the cell suspension was starved for 24 h (0 h). The expression decreased within 2 h of sucrose being re-supplied, and expression then increased again, reaching a peak before mitosis (Figure 2a,b), indicating that ICK1 may have roles both in non-dividing cells and in the G2/M transition.

Sucrose starvation of MM2d was further optimized by using mid-exponential phase cells, and a higher partial synchronization was achieved. The LI and M/A index show separation of S and M phases (Figure 3a), which was confirmed using histone H4, CYCB1;1, CDKA;1, CDKB1;1 and CDKB2;2 probes (Figure 3b). CYCD2;1 and CYCD3;1 mRNA levels were also examined in this experiment, and we observed that CYCD3;1 mRNA increased earlier than observed in Figure 2, at 2–4 h, clearly before histone H4 expression was activated (4–6 h). CYCD2;1 mRNA accumulation was initiated at the same time (2–6 h) as that of CYCD3;1 (Figure 3b). This is in contrast to the results shown in Figure 2(b) using MM1 cells starved in the early stationary phase, where CYCD2;1 was induced earlier than CYCD3;1. This difference in the relative timing of CYCD2;1 and CYCD3;1 expression might reflect a difference between the cell lines, or be due to the application of the treatment to mid-exponential rather than early stationary phase cells. To resolve this discrepancy, early stationary phase MM2d cells were synchronized by sucrose removal. In this case, CYCD2;1 mRNA accumulated before CYCD3;1 (data not shown), as had previously been observed for MM1 cells synchronized using an early stationary phase culture. We conclude that CYCD2;1 and CYCD3;1 are regulated in a similar way in both cell lines, and that CYCD2;1 mRNA induction precedes CYCD3;1 in sucrose-starved stationary phase cells, but not in sucrose-starved mid-exponential phase cells. This difference in timing may indicate a difference in the re-entry of stationary phase cells into the cell cycle, consistent with the shorter time required for mid-exponential phase cells to reach the peak of S phase (10 h compared to 15 h; compare Figures 2a and 3a).

Figure 3.

Sucrose starvation-induced synchrony of MM2d.

Exponential-phase Arabidopsis MM2d cells were subjected to sucrose removal and re-supply. (a) Labelling index (LI) and M/A index determination to detect S-phase (LI, □) and metaphase/anaphase cells (M/A index, ◆) after sucrose was re-added at time t = 0, determined as in Figure 2. (b) Northern blot of RNA samples prepared at 2–3 h intervals from MM2d cells synchronously re-entering the cell cycle from a G1 state (t = 0), probed as indicated. (c) Protein kinase activity (arbitrary units) of CDK bound to Suc1 beads assayed using histone H1 as a substrate at the times indicated. (d) CDKA;1-immunoprecipitated kinase activity (arbitrary units) assayed using histone H1 as a substrate at the times indicated.

The yeast CDK-binding protein Suc1 may be used as an affinity reagent to bind CDKs, whose activity can then be assayed using histone H1 as a substrate. An estimate of total CDK-associated kinase activity was therefore determined by assaying Suc1-bound histone H1 kinase activity (Figure 3c). CDK activity was found to be very low until late G1 phase, and then increased rapidly during late G1 and into S phase. Kinase activity remained relatively constant during S phase, but doubled at around the G2/M boundary at 16 h shortly before the peak of cells in metaphase/anaphase, followed by a decline. A similar result was obtained by determination of immunoprecipitation of CDKA;1-containing kinase activity using a C-terminal anti-peptide antibody and histone H1 as substrate (Figure 3d), suggesting that the Suc1-binding kinase activity parallels the activity of CDKA;1. This is consistent with the report of de Veylder et al. (1997) that Suc1 has a higher affinity for CDKA than other plant CDKs.

Aphidicolin block/release gives improved synchrony

Further enhancement of synchrony in the S–G2–M phases was achieved with reversible arrest cells in late G1/early S phase using aphidicolin block and release (Nagata et al., 1992), with the modifications described in Experimental procedures. Cell-cycle progression after release of the drug was followed by flow cytometry, which allows G1, S and G2 phase cells to be distinguished (Figure 4a). After treatment with aphidicolin, the majority of cells are arrested in G1/early S phase (G1/S). Within 1 h of removal of aphidicolin, an additional peak appeared corresponding to S phase cells, and the DNA content of this S-phase peak increased before reaching the G2-phase DNA content after 5–6 h. The appearance and shift of this additional peak clearly indicates that the majority of cells proceed synchronously through S phase (Figure 4a). Peak analysis of the flow cytometry data shows that 1 h after aphidicolin release, 79% of the cell population is in S phase (Figure 4b). After 9 h, a peak of cells in G2 is reached (G2, 92.5%; S, 0%; G1, 7.5%). Independent LI determination of the same samples confirmed these data (Figure 4c).

Figure 4.

Aphidicolin block/release of Arabidopsis cell lines MM1 and MM2d.

(a) Flow cytometry analysis of MM2d cells after release of aphidicolin block in late G1/early S phase showing the coherent population of cells progressing synchronously through S phase. Each block represents a sample taken at 1 h intervals from t = 0 (top left) to t = 20 (bottom right). (b) DNA histogram of flow analysis results in (a). (c)LI and M/A index determination of S-phase (LI: (□) MM1, (◆) MM2d) and metaphase/anaphase cells (M/A index: (○) MM1, (▴) MM2d) after aphidicolin block is released (t = 0), determined as in Figure 2.

At 10 h, a sharp increase in number of cells in metaphase/anaphase is observed, and the M/A index reaches a peak value of around 12%, 13 h after release of the block. From 10 h, a continuing increase in the proportion of G1 cells is observed (Figure 4b), and after 16–17 h, a second S phase is detectable. Similar results were obtained by monitoring cell-cycle progression after aphidicolin treatment in both light-cultivated Arabidopsis cell suspension MM1 and dark-cultivated MM2d (Figure 4c). We conclude that a high synchronization level and good separation of S and M phases can be achieved using the Arabidopsis cell suspensions MM1 and MM2d and the procedures described.

RNA gel blot analysis of samples from a similar experiment using MM2d showed that cell-cycle regulation of the histone H4, CDKA, CDKB1 and CDKB2 could be readily observed (Figure 5), and the later activation of CDKB2 expression compared with that of CDKB1 was confirmed in this experiment. We note that transcripts of both CDKB1 and CYCB1;1 are present in the blocked cells and during S phase. This result is consistent with the onset of the accumulation of these RNAs from the start of S phase observed in the synchrony obtained by sucrose starvation (Figure 3), and demonstrates that expression of both CDKB1 and CYCB1;1 begins during S phase, although the peak of their expression is in G2/M. We presume that the prolonged retention of Arabidopsis cells in S phase during the aphidicolin block allows these mRNAs, normally expressed at low levels during S phase, to accumulate.

Figure 5.

Expression and CDK activity analysis of aphidicolin-synchronized MM2d cells.

(a)Northern blot analysis of RNA samples prepared at 2–3 h intervals from MM2d cells after removal of aphidicolin block at time t = 0, probed as indicated. (b)Protein kinase activity (arbitrary units) of CDK bound to Suc1 beads assayed using histone H1 as a substrate at the times indicated.

Levels of CYCD2;1 and CYCD3;1 mRNA, particularly the latter, were low after aphidicolin treatment (Figure 5a, 0 h), and then increased within 6 h. Quantification of the RNA gel blots showed that CYCD3;1 mRNA levels continued to gradually increase until mitosis and those of CYCD2;1 increased until the subsequent G1 phase (Figure 5a). Total Suc1-binding CDK activity assayed on histone H1 (Figure 5b) showed high kinase activity in S phase cells, including aphidicolin-blocked cells (t = 0). Interestingly, a twofold drop in kinase activity was observed as cells exited S phase and during G2 (t = 6–8 h). Kinase activity then sharply increased at 10–12 h corresponding to the G2/M transition and early mitosis. Kinase activity then declined and rose again corresponding to the G1/S transition of the second cycle. These data illustrate clearly the low Suc1-binding CDK activity present in G2 cells and in early G1-phase cells. They also agree with the results presented in Figure 3(c), except that the higher level of synchrony from S into G2 using an aphidicolin block makes clear the drop in CDK activity after the end of S phase.


Few Arabidopsis cell cultures are available, and synchrony in these cultures has been difficult to achieve (Breyne and Zabeau, 2001; Callard and Mazzolini, 1997; Fuerst et al., 1996; Richmond and Somerville, 2000; Riou-Khamlichi et al., 2000; Stals et al., 2000). To date, no methods have been described that yield a complete partially synchronous cell cycle, and only partial synchrony for one additional cell-cycle phase beyond the point at which the cells are blocked (i.e. for cells blocked in S phase, synchrony is observed only in G2 and lost by M phase; Callard and Mazzolini, 1997). Here we demonstrate that the Arabidopsis cell suspensions MM1 and MM2d can be synchronized to a relatively high degree resulting in clear separation of different cell-cycle phases, and that such cultures can be used for the analysis of cell-cycle-regulated gene expression and kinase activities. These cultures were derived after extended culturing of an original line described by May and Leaver, (1993). We found that MM1 and its derivative MM2d have a hexaploid (6C) DNA content, compared to a reported tetraploid content in the original cell line (Fuerst et al., 1996). This suggests that the event(s) giving rise to MM1 were not simple genome doubling events, and may indicate possible genome re-organization as is common in animal cell lines.

Synchrony during cell-cycle re-entry may be achieved by starving cells of a nutrient source (Amino et al., 1983; King et al., 1973; Kodama et al., 1991a; Kodama et al., 1991b; Nishida et al., 1992; Riou-Khamlichi et al., 1999; Riou-Khamlichi et al., 2000). Starvation of MM1 and MM2d for sucrose followed by subsequent re-addition resulted in partial synchrony, and, in the case of MM2d, the first clear separation of S- and M-phase cells was achieved. However, the particular value of this synchrony method is that it allows cells to be observed synchronously moving from a non-dividing (quiescent) state, through G1 into S phase. The timing of cell-cycle gene activation during G1 progression and the G1/S transition can readily be followed. In addition, genes only expressed during the first G1 phase following cell-cycle re-entry can be studied. Analysis of expression of three genes encoding CDK inhibitors showed that ICK2 (KRP2) was only expressed in early G1 phase in cells re-entering the cycle after sucrose starvation. ICK2 expression could not be detected in cells synchronized with aphidicolin, and therefore passing through G1 from the preceding mitosis, suggesting that ICK2 is specifically involved in re-entry of cells into division and not in G1 transit per se.

A higher degree of S–G2–M phase synchrony was achieved by blocking cells in late G1/early S phase with the toxin aphidicolin (Callard and Mazzolini, 1997; Ito et al., 1998; Magyar et al., 1997; Nagata et al., 1992; Qin et al., 1996; Reichheld et al., 1995; Samuels et al., 1998). Plant cells treated with aphidicolin are generally arrested subsequent to the G1/S transition and the initiation of DNA replication. It is likely that the step affected is elongation of nascent DNA strands during S phase, as aphidicolin reversibly inhibits the mammalian counterpart of DNA polymerase α and δ in plants (Reichheld et al., 1998; Sala et al., 1980). After washing to release the aphidicolin block, we were able to demonstrate synchronous transition of a cell population of up to 80% through S phase, and 92.5% of cells were found to be in G2 phase only a few hours later, before the first mitotic cell divisions occurred. Entry into G1 and a second S phase could be observed. The greater proportion of cells detected in G2 phase is presumably due to the greater length of this phase relative to S phase, and possibly in part to a small proportion of cells (less than 5%) that could be permanently stuck in G2.

BY-2 cell cultures provide one of the highest synchrony levels of any higher eukaryotic cell type, and, by using aphidicolin to synchronize cultures, typical values of 90% S-phase cells and a mitotic index (MI) of 45–50% are observed (Nagata et al., 1992; Sorrell et al., 2001). In MM2d Arabidopsis cultures using aphidicolin synchrony, we observed 80% S phase cells. The peak value of cells observable in mitosis was about 13%, but it should be noted that only cells in metaphase and anaphase were scored for the M/A index. The small genome size and late condensation of Arabidopsis chromosomes in prophase (Heslop-Harrison and Maluszynska, 1994) make routine scoring of prophase and telophase more difficult, unlike the situation in tobacco BY-2 cells where all stages of mitosis are scored to give an overall MI. We note that, in cells of the Arabidopsis root meristem, the phases of metaphase and anaphase corresponding to those scored in our experiments comprise 33–40% of the total length of mitosis of about 45 min (Boisnard-Lorig et al., 2001). Assuming the relative lengths of mitotic phases to be comparable in a suspension culture, this would suggests that the M/A index we report could be up to threefold lower than the corresponding MI. In contrast, the complete scorable length of mitosis of BY-2 cells is around 1.5 h (Francis et al., 1995), and even in BY-2 cells subject to propyzamide block in metaphase and subsequent release resulting in a MI of 61%, the proportion of anaphase cells does not exceed 14% (Samuels et al., 1998).

The plant cell cycle is regulated by two key checkpoints, of which the first is in late G1, and the second in late G2 phase (Huntley and Murray, 1999; Mironov et al., 1999). Progression through these cell-cycle boundaries is dependent upon specific protein kinase complexes, consisting of the catalytic CDK subunit and a regulatory cyclin subunit. Using Suc1 beads or anti-CDKA;1 antibodies to bind CDKs (de Veylder et al., 1997), we observed peaks of kinase activity (Figures 3c and 5b). High CDK activity was observed in S-phase cells, and in highly synchronized cells produced by aphidicolin treatment a clear reduction in CDK activity was observed in G2, although the level of synchrony using sucrose starvation was not sufficient to observe this. In early mitosis, Suc1-binding CDK activity increased, and then dropped as cells passed from M phase into the subsequent G1 phase. A further rise in kinase activity was observed corresponding to re-entry into the second S phase. The results obtained with the sucrose-starvation-induced synchrony suggest that total Suc1-binding kinase activity increases in late G1 phase and continues increasing during S phase.

The expression of three CDK genes was observed in the synchronized cultures. CDKA is constantly expressed and shows activity at both transition points (Mironov et al., 1999). In contrast, CDKB1;1 expression is activated at the start of S phase and its mRNA level reaches a peak during late G2/early mitosis. The presence of CDKB1;1 mRNA in aphidicolin-blocked cells is consistent with its activation in early S phase, and in this regard we note that the CDKB1;1 gene contains an E2F-binding site in its promoter, as does D3;1 and other genes activated at the G1/S boundary (de Jager et al., 2001). Moreover, this result is also consistent with the activation of the tobacco CDKB1;1 gene from the onset of S phase in BY-2 cells (Sorrell et al., 2001). CDKB2;2 is a previously uncharacterized CDK of the PPTTLRE type (Huntley and Murray, 1999). Here we show that CDKB2;2 expression is clearly distinct in its timing from the CDKB1;1 class and appears to be limited to cells in mitosis. A similar conclusion was reached for expression of the Antirrhinum cdc2d gene (Fobert et al., 1996) and the alfalfa cdc2MsF gene (Magyar et al., 1997). These results clearly indicate that CDKB1 may play a role in the G2/M transition, whereas CDKB2 is probably involved in the control of mitotic processes.

Strikingly different cell-cycle regulation was observed for the three different CDK inhibitor genes examined (de Veylder et al., 2001; Lui et al., 2000; Wang et al., 1998). ICK1 (KRP1) mRNA levels were high in sucrose-starved cells, declined and then increased in mitotic cells. ICK2 (KRP2) expression was limited to cells transiting the first G1 after sucrose starvation and re-supply, whereas KRP3 expression was absent in starved cells and increased until G2/M phase. These results demonstrate that synchronized Arabidopsis cell suspensions can be used to identify the expression timing and activity of candidate cell-cycle-related genes, and thereby to infer potential roles in cell-cycle progression.

Experimental procedures

Selection of fast-growing Arabidopsis cell suspension cultures MM1 and MM2d

Arabidopsis thaliana cell suspension cultures (ecotype Landsberg erecta), originally described by May and Leaver, 1993) were obtained from Professor K. Lindsey (University of Durham) and grown in 1 × Murashige and Skoog medium (ICN Biomedicals, Aurora, Ohio, USA) supplemented with 3% w/v sucrose (Sigma), 0.5 mg l−1 NAA (Sigma), 0.05 mg l−1 kinetin (Sigma), with pH adjusted to 5.8 using 1 N KOH (MSS). Faster-growing derivatives were selected solely by rigorous attention to subculture every 7 days as follows. MM1 cells were maintained by weekly subculturing of 7.5 ml saturated culture into 200 ml of fresh MSS in 500 ml narrow-necked Erlenmeyer flasks, covered with two layers of domestic heavy-duty aluminium foil then loosely covered with PVC plastic film. MM1 was grown under continuous light (average 1300 lux) from Philips TLD HF 50 W/840 lamps, and rotated at 120 rev min−1 on a New Brunswick (Edison, New Jersey, USA) G10 Gyrator shaker with 25 mm orbit at a temperature of 24 ± 1°C. MM2d was derived by transferring a sample of MM1 into dark-growing conditions. Various temperatures of growth and medium compositions were tested, and optimum growth was obtained by weekly subculturing of 3.5 ml saturated culture into 100 ml fresh MSS in 300 ml narrow-necked Erlenmeyer flasks and growth in continuous darkness with rotation at 130 rev min−1 on a New Brunswick Innova Model 4230 incubator shaker with 19 mm orbit at a temperature of 27°C. Again, no further selection was applied other than rigorous attention to routine subculturing.

For growth analysis, 5 ml of an MM1 and an MM2d early stationary phase culture were inoculated into 100 ml fresh MSS and incubated. Samples were taken at approximately 24 h intervals and cells were microscopically observed and counted using a counting chamber (0.2 mm depth).

Sucrose starvation

To achieve synchronization, 200 ml of an early stationary phase cell suspension of MM1 (7 days after previous subculture) was washed by vacuum-assisted filtration (Whatman Membrane filter systems, glass holders; catalogue number 1960-004) with 1 litre of MS medium (MS salt, 0.5 mg l−1 NAA, 0.05 mg l−1 kinetin, lacking sucrose), left in a minimal volume and resuspended in 200 ml MS medium. Aliquots (40 ml) of this washed and resuspended cell suspension were diluted into 210 ml fresh MS medium to achieve a dilution factor of approximately 1:5, and incubated at 23°C, 120 rev min1 on a New Brunswick G10 gyrator shaker with 25 mm orbit, in the light for 24 h. This dilution step is important for full synchrony. After 24 h, sucrose was added to the medium to a final concentration of 3%, the cultures were incubated under cultivation conditions as above, and samples were removed hourly. Similarly, 600 ml of mid-exponential phase cell suspension MM2d (4 days after previous subculture), consisting of six individual flasks with 100 ml culture in each, were washed twice by repeated centrifugation (688 g/2 min/without brake) in MS medium (MS salt, 0.5 mg l−1 NAA, 0.05 mg l−1 kinetin, lacking sucrose) and resuspended and pooled in 600 ml MS medium. Aliquots (20 ml) of this pooled resuspended cell suspension were transferred into 100 ml fresh MS medium to achieve a diluting effect of approximately 1:5, and incubated at 27°C, 130 rev min1 in a New Brunswick Innova model 4230 incubator shaker with a 19 mm orbit, in darkness for 24 h before sucrose was added to the medium to a final concentration of 3%. The cultures were incubated under conditions as above and samples removed hourly.

Aphidicolin block/release

MM1 cells were reversibly blocked in late G1/early S phase with aphidicolin according to Nagata et al. (1992) with modifications as follows. A 40 ml aliquot of early stationary phase cell suspension culture (7 days after previous subculture) was subcultured into 200 ml fresh MSS, containing 4 µg ml−1 aphidicolin (Sigma) and incubated at 23°C, 120 rev min1 in a New Brunswick G10 gyrator shaker with 25 mm orbit, for 21.5 h. Cells were vigorously washed with 1 litre of MSS through a nylon net (mesh size 47 µm), followed by centrifugation (387 g/1 min/without brake) to remove aphidicolin. The cell pellet was resuspended in 250 ml MSS, incubated under cultivation conditions as above, and samples taken hourly after release of the drug for validation procedures. Synchronization of MM2d was performed as described above, with the exception that the culture conditions throughout the experiment were as described for MM2d.

Labelling index (LI) and metaphase/anaphase index (M/A index) determination

Cells actively replicating DNA were labelled using bromo-deoxyuridine incorporation according to the method described by Miyake et al. (1997) with modifications as follows. After fixation with 3.7% formaldehyde in PMEG buffer (50 mm PIPES (Sigma), 2 mm MgSO4 (Sigma), 5 mm EGTA (Sigma), 2% glycerol v/v, pH 6.8), aliquots of cells were washed twice with PBS buffer (2000 rev min1/1 min in an Eppendorf centrifuge 5417c), digested for 25 min at room temperature with 1% cellulase-R10 (Yakult Honsha Co. Ltd, Tokyo, Japan), 0.1% pectolyase Y-23 (Kikkoman, Tokyo, Japan) and 0.4 m mannitol (Sigma), followed by permeabilization for 25 min at room temperature with 0.1% Igepal CA-630 (Sigma) in PBS, washed three times with PBS buffer, transferred to a slide and left to settle overnight in a humidified chamber. After blocking the slide for 10 min (blocking solution: 0.1 mol l−1 glycine, 1% bovine serum albumin, 0.05% Triton X-100 in PBS containing 20 mmol l−1 phosphate and 150 mmol l−1 NaCl, pH 7.0), anti-5-bromo-2′-deoxyuridine/nuclease (RPN20, Amersham Pharmacia Biotech, Little Chalfont, UK) was added as a primary antibody and incubated for 60–90 min at 30°C in a humidified chamber. The attached cells were carefully washed with PBS. Texas red-conjugated donkey anti-mouse (Jackson ImmunoResearch, West Grove, PA, USA, catalogue number 715-076-150) was added as secondary antibody, and incubated as for the primary antibody. After washing with PBS, cells were mounted in Vectashield containing DAPI (Vector Laboratories, Burlingame, California, USA), and examined using fluorescent excitation at 575 nm on a Nikon Optiphot 2 microscope. To determine the labelling index, antibody-labelled and total cells were manually counted, using DAPI as nuclear stain. To determine the metaphase/anaphase index (M/A index), the proportion of cells with DAPI-stained metaphase and anaphase figures were counted in the same fields.

Flow cytometric analysis

A sample of frozen cell pellet was treated with the High Resolution Kit for plant ploidy level analysis (type P) (Partec GmbH, Münster, Germany) to determine the DNA content. To release cell nuclei, the cells were carefully chopped with a sharp razor blade in solution A and filtered prior to addition of solution B. On average, 8200 particles were counted with a flow cytometer (PASIII; Partec GmbH), using HBO lamp excitation (Partec GmbH, mercury lamp HBO 100 long-life, 100 W, excitation filters KG1, BG38, UG1) and detection of emission using a blue fluorescence emission filter GG435 (long pass colour glass). Cell-cycle phases were analysed using Multicycle for Windows (Phoenix Flow Systems, San Diego, California, USA).

RNA extraction and Northern blot analysis

Total RNA was extracted as described previously (Riou-Khamlichi et al., 2000; Verwoerd et al., 1989) except for the extraction buffer (Soni and Murray, 1994), and analysed by RNA gel blotting using probes labelled as described by Riou-Khamlichi et al. (2000). The following probes were used: histone H4 (Chaubet et al., 1996), 0.2 kb BamHI–EcoRI fragment; CYCB1;1 (Hemerly et al., 1992), 1.6 kb EcoRI fragment; CYCD2;1 (Soni et al., 1995), 0.7 kb EcoRI fragment; CYCD3;1 (Soni et al., 1995), complete cDNA coding region (1.2 kb); CDKA;1 and CDKB1;1 (gift of Dr P. Doerner) full-length coding region (1.3 kb); CDKB2;2, 0.6 kb fragment amplified from mRNA by RT–PCR using primers TTAAACCCGCTGTTTCCGCC and CCAACAGACCACATATCCACTCC; ICK2(KRP2) (de Veylder et al., 2001; Lui et al., 2000), 0.63 kb fragment amplified by PCR; KRP3 (de Veylder et al., 2001), 0.67 kb PCR fragment; KRP1 (ICK1) (de Veylder et al., 2001; Wang et al., 1998), 0.6 kb PCR-fragment. Hybridized membranes were exposed to autoradiography film, scanned and quantified using NIH Image 1.62 ( Equal loading was controlled by methylene blue staining of the membranes, the membranes were soaked in 0.02% w/v methylene blue in 0.5 m sodium acetate (pH 5.2), and then rinsed in 1 × SSPE (Sambrook et al., 1989).

Protein analysis and kinase activity

Procedures for protein extraction, immunoprecipitation and kinase assays have been previously described (Cockcroft et al., 2000; Riou-Khamlichi et al., 2000). For Suc1 pulldown, 15 µl Suc1 beads were incubated with 250 µg protein extract for 2 h, washed as described for immunoprecipitation, and a kinase assay performed.


The authors are very grateful to Masami Sekine for the gift of the KRP inhibitor clones. We thank Roderic Fuerst and Keith Lindsey for the original gift of the cell suspension from which the lines described here were derived, and Catherine Riou-Khamlichi for the initial development of sucrose starvation techniques. This work was financially supported by grants from the UK Biotechnology and Biological Sciences Research Council (BBSRC) and Aventis CropScience to J.A.H.M., and a BBSRC studentship to M.M. The cell lines MM1 and MM2d are freely available from the authors to interested researchers. A material transfer agreement with the commercial sponsors of the work is involved.