Sugar Control of the Plant Cell Cycle: Differential Regulation of Arabidopsis D-Type Cyclin Gene Expression

Abstract

In most plants, sucrose is the major transported carbon source. Carbon source availability in the form of sucrose is likely to be a major determinant of cell division, and mechanisms must exist for sensing sugar levels and mediating appropriate control of the cell cycle. We show that sugar availability plays a major role during the G₁ phase by controlling the expression of CycD cyclins in Arabidopsis. CycD2 mRNA levels increase within 30 min of the addition of sucrose; CycD3 is induced after 4 h. This corresponds to induction of CycD2 expression early in G₁ and CycD3 expression in late G₁ near the S-phase boundary. CycD2 and CycD3 induction is independent both of progression to a specific point in the cell cycle and of protein synthesis. Protein kinase activity of CycD2- and CycD3-containing cyclin-dependent kinases is consistent with the observed regulation of their mRNA levels. CycD2 and CycD3 therefore act as direct mediators of the presence of sugar in cell cycle commitment. CycD3, but not CycD2, expression responds to hormones, for which we show that the presence of sugars is required. Finally, protein phosphatases are shown to be involved in regulating CycD2 and CycD3 induction. We propose that control of CycD2 and CycD3 by sucrose forms part of cell cycle control in response to cellular carbohydrate status.

In multicellular organisms, cell division is normally coordinated with growth and cellular differentiation (38, 46). Plants respond flexibly to the environment, modulating both their growth rate and developmental pattern in response to external conditions (8, 13, 31). One possible mediator of these events is sucrose, the major transported product of photosynthesis, which is involved in the regulation of a large number of genes (25, 59). Sucrose is a good candidate for modulating cell division rates because its availability to proliferating cells located in meristems will be a reflection of overall photosynthetic capacity and hence the prevailing conditions (27).

Cell division control in plants shares some features with that in eukaryotic microorganisms and animals, including a major control point during the G₁ phase (12, 21, 39). Many plant cells arrest in G₁ in response to nutrient limitation or differentiation (12). G₁ is also the major extendable part of the cell cycle in response to alterations in conditions such as temperature (14), CO₂ levels (26), and light levels (42), and it might be expected that a primary control of cell cycle entry will therefore be imposed during the G₁ phase (41).

In both yeasts and mammals, such controls operate primarily through transcriptional control of the levels of G₁ cyclins, which are unstable regulatory subunits of cyclin-dependent kinases (Cdk) required for G₁ exit (45, 57). In mammals, the main response mechanism in G₁ is due to serum growth factors, whose presence leads to a rapid rise in transcript levels of D-type cyclins (35, 57). These associate with Cdk4 and Cdk6, and the resulting kinase activity is directed to the retinoblastoma protein (Rb). Rb keeps genes responsive to the E2F transcription factors inactive during G₁, so that the phosphorylation of Rb results in activation of genes under E2F control and hence in S-phase entry (10, 57, 58).

Unlike yeasts, plants have D-type cyclins, Rb homologs, and E2F-like proteins (7, 9, 16, 17, 47, 54, 62, 63). Plants therefore appear to be more similar to mammals than to yeasts in the operation of G₁-to-S-phase control (1, 20), and the plant homologs of D-type cyclins (CycD) are thus good candidates for influencing plant cell division in response to external conditions. The plant CycD cyclins have highest homology to mammalian D-type cyclins and have a highly conserved Rb-binding motif that is functional in binding plant and human Rb proteins (1, 20). Available evidence on expression of CycD genes in Arabidopsis is also consistent with their playing a role in response to external signals, with their major point of action occurring during G₁. Experiments with callus material deprived of sucrose and the hormones auxin and cytokinin for 48 h suggested that CycD2 mRNA levels were responsive to sucrose and CycD3 mRNA levels were responsive to cytokinin (62). Subsequent analysis has shown that CycD3 is a major target of cytokinin in Arabidopsis cell division control, since CycD3 is rapidly induced by cytokinin without a requirement for protein synthesis and CycD3 overexpression can replace the cytokinin requirement for callus induction from Arabidopsis leaf tissue (50).

The timing of expression of a number of Arabidopsis cyclins has also been examined using a suspension culture (36). Partial synchronization was achieved by blocking exponentially growing cells in G₁ with low concentrations of cycloheximide (15). Within the limitations of the synchrony achieved, it was concluded that CycD2 did not show cell cycle regulation and that CycD3 expression appeared to be induced at or, more probably, slightly before the S phase, after release of the blocked cells. No subsequent oscillations in CycD3 mRNA abundance were observed. However, that study did not address whether CycD2 and CycD3 expression depends on cell division or its possible response to external signals such as carbon source availability or hormones (62). In this latter case, expression would be expected to be constant in the presence of the external inducing signal, even if the cells were not undergoing active division.

A number of cell cycle genes in plants are regulated by hormones and/or sucrose availability (16, 24, 37, 50). It is often difficult to distinguish a direct response to the stimulus from indirect effects due to triggering of the process of cell division. To identify genes directly imparting external information to cell division processes, it is necessary to uncouple cell division from the stimulus (41).

Here we show that expression of both CycD2 and CycD3 is dependent on sucrose. From the relative timing of induction of CycD2 in early G₁ and CycD3 in late G₁, we propose a model for the control of G₁ exit in Arabidopsis.

MATERIALS AND METHODS

Arabidopsis cell suspension culture, seedlings, and induction experiments.

A suspension culture of Arabidopsis thaliana ecotype Landsberg erecta (15, 36) was grown in Murashige and Skoog (MS) medium (ICN Biomedicals Inc., Costa Mesa, Calif.) with 3% (wt/vol) sucrose, 0.5 mg of α-naphthaleneacetic acid (Sigma, St. Louis, Mo.) per ml, and 0.05 mg of kinetin (Sigma) per ml added; the pH was adjusted to 5.8 with 1 N KOH. Plant hormones (plant growth regulators; PGRs) were added after the mixture was autoclaved at 120°C for 20 min. Cells were grown at 23°C under continuous light conditions in a 500-ml flask shaken at 120 rpm and were diluted by adding 10 ml to 200 ml of fresh medium every 7 days (D7 cells). A. thaliana Landsberg erecta seedlings were grown in MS liquid medium (3% sucrose, no PGRs) under the same conditions.

Experiments involving carbon source induction were carried out as described previously (50). Early-stationary-phase cells 7 days after the previous subculture (D7 cells) were washed with 3 to 4 culture volumes of fresh MS medium (lacking sucrose but containing PGRs) by vacuum-assisted filtration, resuspended at the original cell density of the D7 culture, and incubated under culture conditions for 24 h in the same medium until time T0. The cells were concentrated by vacuum-assisted filtration to a total volume of about 100 ml, and 3 ml of concentrated cell suspension was added to each of 100-ml flasks containing 25 ml of MS medium with additions as described below. Individual flasks were processed as a single sample.

Similarly, for induction experiments examining cytokinin induction, D7 cells were washed with MS medium containing 3% sucrose (lacking PGRs) and cultivated for 24 h in this medium until T0. In certain experiments, cells were washed and cultivated for 24 h in MS medium lacking both sucrose and PGRs.

Inhibitors.

Stock solutions of drugs (Sigma) were prepared as follows. Cycloheximide (Chx) was dissolved in 100% ethanol at 10 mg/ml, okadaic acid (OA) was dissolved in 0.15% ethanol and prepared at 25 μg/ml, and tautomycin (Biomol Research Lab. Inc.) was dissolved in 100% ethanol at 100 μg/ml. Hydroxyurea was dissolved directly at the desired concentration in MS medium.

RNA extraction and Northern blot analysis.

RNA, extracted as described previously (69) except for the use of extraction buffer (61) was analyzed by RNA gel blotting (40). After transfer, membranes were baked for 2 h (80°C). Probes were labeled using a Rediprime kit (Amersham-Pharmacia Biotech, Little Chalfont, United Kingdom), purified using Nick columns (Amersham-Pharmacia Biotech), and denatured by being adjusted to 0.4 N NaOH. The CycD3 probe (62) used corresponded to the complete cDNA coding region (1.2 kb), the CycD2 probe (62) corresponded to a 0.7-kb EcoRI fragment, and the histone H4 probe (5) corresponded to a 0.2-kb BamHI-EcoRI fragment. Hybridized membranes were exposed to autoradiography film and quantitated using NIH Image 1.62. Quantitation is valid only within one experiment, and so comparisons between gels cannot be made. All experiments were independently repeated at least twice. RNA loading was controlled by methylene blue staining of the membranes. The membranes were soaked in 1× SSPE (53) containing 0.02% (wt/vol) methylene blue stock solution prepared in 0.5 M sodium acetate (pH 5.2) and were then rinsed in 20% ethanol.

[methyl-³H]thymidine incorporation.

For each measurement, 10 μCi of [methyl-³H]thymidine (65 to 80 Ci/mmol; Amersham-Pharmacia Biotech) was added to a 1-ml sample of cells, which were incubated under culture conditions for a further 30 min. The cells were centrifuged, washed three times in 15 ml of fresh medium, and immediately frozen. The cell pellet was resuspended in 1 ml of 7% (vol/vol) trichloroacetic acid (TCA) and disrupted using an Ultra-Turrax T8 hand-held cell disrupter (IKA-Labortechnik) for 10 s. After centrifugation (3,040 × g), the pellet was washed with a mixture of 80% (wt/vol) ethanol and 0.2 N perchloric acid and recentrifuged. The pellet was resuspended in hot 0.5 N perchloric acid and incubated at 65°C for 30 min. Radioactivity was measured in the supernatant using CytoScint (ICN). The DNA concentration was determined by measuring the optical density at 260 nm, and results were expressed as cpm per microgram of DNA. All samples were measured in duplicate.

Kinase assays.

Pelleted cell suspension (1 g) was ground in liquid nitrogen, resuspended in cell lysis buffer (1 ml of 50 mM Tris [pH 7.5], 75 mM NaCl, 15 mM EGTA, 15 mM MgCl₂, 1 mM dithiothreitol [DTT], 0.1% Tween 20, 1× complete Tm protease inhibitors [Roche], 1 mM NaF, 0.2 mM Na₃VO₄, 2 mM sodium pyrophosphate, 60 mM β-glycerophosphate), and homogenized for four 30-s bursts with 30 s on ice between homogenizations. Then 0.5 to 1 mg of extract was preincubated with 20 μl of protein A-Sepharose (50% suspension) for 30 min at 4°C. The supernatant was incubated with 1 μl of specific antiserum (2 h on ice), 20 μl of protein A-Sepharose was added, and the sample was rotated at 4°C for 1 h. Samples, washed four times with 1 ml 50 mM Tris (pH 7.5)–250 mM NaCl–5 mM EDTA–5 mM NaF–0.1% Tween 20–0.5 mM phenylmethylsulfonyl fluoride, were washed twice in kinase buffer (50 mM Tris [pH 7.5], 100 mM NaCl, 15 mM EGTA, 1 mM DTT), resuspended in 15 μl of assay buffer (50 mM Tris [pH 7.5], 100 mM NaCl, 5 mM EGTA, 10 mM MgCl₂, 1 mM DTT, 1 mM NaF, 0.2 mM sodium orthovanadate, 2 mM sodium pyrophosphate, 25 mM β-glycerophosphate, 0.5 mg of histone H1 per ml, 0.5 mM phenylmethylsulfonyl fluoride, 2 μCi of [γ-³²P]ATP [>5,000 Ci/mmol] per 15-μl reaction volume), and incubated at room temperature for 30 min. The reaction was stopped by adding gel loading buffer, and samples were analyzed by sodium dodecyl sulfate-polyacrylamide electrophoresis and quantitated using a PhosphorImager (Molecular Dynamics). Polyclonal rabbit antibodies were raised against full length Arabidopsis CycD2 expressed in Escherichia coli and against the Arabidopsis CycD3 C-terminal peptide MRGAEENEKKKPILHLPWAIVATP.

In vivo labeling of proteins.

For each measurement, 17 μCi of Redivue Pro-Mix L-[³⁵S] in vitro cell-labeling mix (>1,000 Ci/mmol; Amersham) was added to 10 ml of cell suspension 1 h after Chx addition, and incubation was continued for a further 6 h. Cells were washed three times with 15 ml of MS medium, cell pellets were ground in liquid nitrogen, and 150 mg was resuspended in cell lysis buffer. The suspension was homogenized four times for 10 s with 20 s on ice between homogenizations and centrifuged at 13,000 rpm (MSE Microcentaur; Sanyo) at 4°C. Then 30 μl of extracted protein solution was added to 100 μl of bovine serum albumin (1 mg/ml), 1 ml of ice-cold 10% (wt/vol) TCA solution was added, and the suspension was mixed vigorously and incubated for 30 min on ice. The suspension was filtered through a disposable GF/C filter funnel (Whatman, Maidstone, United Kingdom) under vacuum, and the filters were washed three times with 5 ml of 10% (wt/vol) TCA and twice with 5 ml of 100% ethanol, air dried, and counted in CytoScint. Counts were expressed as the ratio of the total activity present in 30 μl of protein extract.

RESULTS

CycD2 and CycD3 expression is not growth phase dependent, but is regulated by carbon source availability.

We examined whether CycD expression is growth phase related. Both CycD2 and CycD3 mRNA levels are constant through the growth cycle from day 1 to early stationary phase on day 7 (D7; Fig. 1A and B) for cells grown in 88 mM (3%) sucrose. In contrast, the expression of histone H4, which is expressed only in cells in S phase, is high in exponential cells but low by D7. This indicates that the expression of CycD2 and CycD3 is not dependent on active division or on cells being at a particular cell cycle stage.

FIG. 1.

CycD2 and CycD3 expression in A. thaliana cell suspension cultures is not growth phase regulated. (A) Early-stationary-phase (D7) cells were subcultured, and cell growth was monitored by determination of the number of cells per milliliter of culture medium each day after dilution (upper). Samples were taken for RNA gel blot analysis (lower). As loading control, the blot was reprobed with c7, a cDNA unregulated during cell cycle (R. Fuerst, unpublished data). (B) In a separate experiment, cells were treated as above and Arabidopsis histone H4 and CycD2 expression was determined by RNA gel blot analysis. Relative signal quantitation is shown under each gel band. Loading control was provided by methylene blue staining of the membrane after transfer.

To understand CycD responses to external cell signals, we examined the effect of sucrose removal on the continued expression of CycD2 and CycD3. Early-stationary-phase D7 cells were incubated for 24 h in medium lacking sucrose, after which the levels of CycD2 and CycD3 mRNA were both reduced approximately 2-fold (Fig. 2A and B, Table 1; compare early-stationary-phase [D7] cells before sucrose removal with T0 cells 24 h after sucrose removal). Within 4 h of readdition of sucrose, strong expression of both CycD2 and CycD3 was present (Fig. 2A, +S), corresponding to a fourfold increase over T0 levels for CycD2 mRNA and a ninefold increase for CycD3 mRNA. In both cases the induced level of mRNA was higher than that observed in D7 cells (Table 1). Equal effects were observed with 1 to 3% sucrose, and the use of mannitol as a substitute osmoticum showed that the decline in expression on sucrose removal and subsequent reinduction is not due to osmotic changes but to the loss of a readily metabolizable carbon source (results not shown). The decline in CycD2 and CycD3 mRNA levels on sucrose removal and the induction of their expression on its resupply were consistently seen in every repeat of this experiment carried out (n = 6 to 8; Table 1).

FIG. 2.

Regulation of CycD2 and CycD3 expression depends on the carbon source availability. (A) D7 cells were depleted of sucrose for 24 h (T0) in the presence of auxin and cytokinin. Then 3% sucrose was added for 4 h, and RNA was hybridized to CycD2 and CycD3 probes. (B) Effects of various carbon sources at a range of concentrations for 6 h on CycD mRNA levels in sucrose-depleted cells (T0). (C) CycD3 expression depends on sugar availability. Sucrose was replaced by glucose at 88 or 176 mM in cell suspensions. The glucose concentration (% Glucose) remaining in the culture supernatant was determined using test strips for blood glucose determination. Relative signal quantitation is shown under each gel band. Loading control was provided by methylene blue staining of the membrane.

TABLE 1.

Response of CycD2 and CycD3 mRNA levels to sucrose removal and resupply for 4 to 6 h^a

Gene	Fold decline after sucrose removalb			Fold induction after resupplyb
	Mean ± SD	Range	n	Mean ± SD	Range	n
CycD2	1.56 ± 0.37	1.05–1.96	7	3.09 ± 0.75	1.65–4.0	8
CycD3	3.18 ± 1.79	1.80–6.90	6	7.73 ± 2.37	4.75–12.5	7

^aThe mean value of the decline in mRNA level 24 h after removal of sucrose from D7 cells and the subsequent induction 4 h after sucrose readdition was measured in six to eight independent RNA gel blot experiments involving separate treatment of cells on different occasions and separate independent hybridization analyses. 

^bStandard deviations of measurements and minimum and maximum values obtained are shown. n is the number of independent experiments from which analysis carried out. 

CycD3 expression was found to be inducible by both glucose and sucrose within 4 h of addition to starved cells, and the increase in expression was observed at concentrations down to 1 mM sucrose or glucose, with maximal induction from 10 mM (Fig. 2B). The increases in CycD2 levels appeared to be somewhat more responsive to glucose than to sucrose, and induction reached its maximal level from 10 mM glucose. CycD2 and CycD3 expression therefore responds to low concentrations of sugars compared to that initially present in the growth medium (88 mM sucrose; 176 mM hexose equivalent) and within the physiological range of concentrations observed in meristems (D. Francis, personal communication). Since sucrose and glucose both induce CycD2 and CycD3 expression, we conclude that CycD2 and CycD3 induction is not linked to a specific sugar uptake pathway.

The flux of substrate through the hexokinase-catalyzed step has been proposed as a primary mechanism of sugar sensing in many organisms, including plants (22, 27). The differential effects of the glucose analogs 6-deoxyglucose (6DG), which is taken up by cells but is not a substrate for hexokinase, and 2-deoxyglucose (2DG), which is a substrate of hexokinase but is not readily further metabolized, have been widely used as evidence for the involvement of hexokinase as a sugar sensor. Neither CycD2 nor CycD3 were induced by 6DG, whereas low concentrations of 2DG induced CycD2 but not CycD3 expression (Fig. 2B). Furthermore, 10 mM mannose, a sugar that is also phosphorylated by hexokinase but is poorly metabolized subsequently, also induced CycD2 expression (results not shown). These data suggest that CycD2 induction by sucrose requires hexokinase whereas CycD3 induction probably does not.

A feature of the responses of CycD2 and CycD3 expression to inducing signals in these experiments is that removal of the stimulus followed by its reapplication results in a higher level of expression than if the stimulus remains constantly present. The “induced” level of CycD2 is 1.9 ± 0.43-fold and that of CycD3 is 3.0 ± 1.06-fold higher than the D7 level and hence than their normal levels in untreated cultures. It is also apparent that the dynamic range of CycD3 regulation by sucrose is greater than that of CycD2 (Fig. 2; Table 1).

To understand why early-stationary-phase cells continue to express CycD3, we measured residual hexose present in the culture supernatant of cells grown on 176 mM glucose, thereby providing the same molar hexose equivalent as the normal culture conditions of 88 mM (3%) sucrose. We found that at D5, the glucose concentration in the culture was 60 mM and by D10 it had only fallen to 30 mM. These cells were in stationary phase, since no further increase in cell number was observed (Fig. 1) and histone H4 mRNA was not detectable (data not shown), indicating that the cells were entering stationary phase, but this is not due to a lack of metabolizable sugars. However, when the cells were cultured in half the concentration of glucose (88 mM), no detectable glucose remained by D9. RNA prepared from these cultures showed that CycD3 expression is correlated with hexose availability, since at D9 it was still present in cells originally subcultured into 176 mM glucose in which more than 30 mM still remained but was absent from cells originally subcultured into 88 mM glucose, in whose medium no glucose remains (Fig. 2C). From the correlation of CycD levels with sugar availability and their induction on sugar resupply, we conclude that CycD2 and CycD3 respond sensitively to sugar availability.

CycD2 and CycD3 are sucrose regulated in Arabidopsis seedlings.

To confirm the in planta relevance of the analysis, we examined the response of CycD2 and CycD3 in intact seedlings. Seedlings grown in liquid MS medium without sucrose for 8 days were transferred for 24 h to medium containing sucrose. Expression of CycD2 and CycD3 was increased 5.8- and 3.6-fold, respectively, by the presence of sucrose (Fig. 3A, +S) compared to that in control seedlings left in medium lacking sucrose (Fig. 3A, −S). In a further experiment, 7-day-old seedlings were transferred from medium containing sucrose to medium lacking sucrose. At 48 h after the transfer (Fig. 3B, lane 4) the seedlings showed an eightfold reduction in both CycD2 and CycD3 mRNA levels compared to control seedlings maintained with sucrose (lane 3). When transferred back to medium containing 15 or 30 mM (0.5 or 1%) sucrose, CycD2 and CycD3 mRNA levels both increased 15-fold within 6 h (lanes 1 and 2). As previously observed with suspension-cultured cells, CycD2 and CycD3 were induced to a level approximately twofold higher than that seen in seedlings maintained continuously in the presence of sucrose.

FIG. 3.

Low concentrations of sucrose rapidly induce CycD2 and CycD3 expression in young seedlings. (A) Seedlings were grown under light for 8 days in MS liquid medium lacking sucrose. Half the sample was left in the same medium for a further 24 h (−S), and the other half was placed for 24 h in medium supplemented with 3% sucrose (+S). (B) Seedlings (7 days old) grown in MS liquid medium with 3% sucrose were depleted of sucrose for 48 h (0 mM; lane 4), whereas control seedlings were maintained in the presence of sucrose for 9 days (+S; lane 3). Seedlings were transferred from −S conditions to 15 mM (0.5%) or 30 mM (1%) sucrose for 6 h (lanes 1 and 2) before samples were taken for RNA gel blot analysis. Relative signal quantitation is shown under each gel band.

We conclude that CycD2 and CycD3 respond rapidly to the presence of sucrose at low concentrations in seedlings, consistent with the results obtained from suspension cultures. The cytokinin response of CycD3 in suspension cultures is also consistent with results obtained with whole plants (50).

The timing of CycD2 and CycD3 induction by sucrose corresponds to early and late G₁ phases, respectively.

We examined the timing of CycD2 and CycD3 mRNA accumulation in response to sucrose readdition to the culture medium. Sucrose was added to D7 cells depleted of sucrose for 24 h, and samples were taken for mRNA analysis from 30 min to 24 h and also assessed for [methyl-³H]thymidine incorporation into DNA. A partially synchronous entry into S phase from the quiescent state induced by sucrose removal was observed (50), characterized by S-phase onset after 6 h as detected by histone H4 mRNA accumulation and the incorporation of [methyl-³H]thymidine into cellular DNA (Fig. 4A).

FIG. 4.

The timing of sucrose induction of CycD2 and CycD3 corresponds to the early G₁ phase and the late G₁ phase, respectively, in cells reentering the cell cycle. D7 cells were depleted of sucrose for 24 h (T0), and a time course of induction was monitored and analyzed by thymidine incorporation (A) and RNA gel blot analysis (B). (A) Cells were incubated with [methyl-³H]thymidine for 30 min, and incorporation of thymidine into cellular DNA was measured (open squares). Background incorporation (solid squares) was similarly measured in the presence of 10 mM hydroxyurea, a ribonucleotide reductase inhibitor that blocks DNA synthesis in plant cells. Net incorporation (circles) was calculated by subtraction of the background incorporation measured and is presented as a percentage of the maximum observed. (B) RNA gel blot analysis of CycD2, CycD3, and histone H4 expression for the cells sampled in panel A. The relative signal quantitation is shown under each gel band. (C) Fold increase in CycD2 and CycD3 kinase activity at the times shown over background measured at the time of sucrose addition (T = 0). CycD2 (open squares)- and CycD3 (solid squares)-containing kinase complexes were immunoprecipitated using specific antisera, assayed on histone H1 as a kinase substrate, and quantified by PhosphorImager analysis.

A constant low level of CycD3 mRNA was observed until 3 h after sucrose addition, which increased more than fourfold between the 3- and 4-h samples, 2 h in advance of the main increase in histone H4 expression and before the onset of significant DNA synthesis (Fig. 4B). The induction of CycD3 before S phase was confirmed by the finding that CycD3 was also induced by sucrose in the presence of the ribonucleotide reductase inhibitor hydroxyurea, which blocks S-phase entry (results not shown). This refines earlier observations (15, 62) and confirms that CycD3 is induced late in G₁ but before S-phase onset in cells reentering the cycle.

In contrast, 50% of the total increase in the CycD2 mRNA level occurred within 30 min of sucrose readdition, and the CycD2 mRNA level then further increased to its maximal value of a 3.6-fold induction by the 6-h sample. We conclude that onset of CycD2 mRNA accumulation is a rapid response to sucrose readdition whereas CycD3 mRNA accumulated after approximately 4 h. In unperturbed cells, this corresponds to late G₁ phase.

In both yeasts and mammals, transcriptional control is an important mechanism of regulation of G₁ cyclin activity (35, 45, 57). To confirm that results obtained from RNA analysis of CycD2 and CycD3 expression reflect subsequent enzyme activity, kinase assays were carried out using immunoprecipitates of cell extracts with either a CycD2- or CycD3-specific antiserum. CycD2 kinase activity increased threefold above its basal level within 2 h of sucrose readdition, whereas CycD3 kinase activity showed no increase until 4 h after readdition and reached a threefold induction only after 8 h (Fig. 4C), consistent with the timing of induction of CycD2 and CycD3 mRNA.

CycD2 and CycD3 induction by sugars is independent of cell cycle progression.

Induction of CycD gene expression could be a direct response to sucrose or an indirect result of the ability of cells to grow and divide again. To test this, we used the protein synthesis inhibitor Chx. At concentrations of 150 nM or above, Chx blocks cells of this Arabidopsis suspension culture in G₁ (15). At higher concentrations (e.g., 100 μM), de novo protein synthesis is also inhibited (50).

Chx at 100 μM had no effect on the induction of CycD3 levels by sucrose after 4 h, demonstrating that neither progression through G₁ nor new-protein synthesis is necessary (Fig. 5A). The CycD2 mRNA level increased 3.6-fold in response to Chx alone, equal to the induction of expression by sucrose (Fig. 5A), but the combined presence of sucrose and Chx led to a further 2.5-fold increase in the CycD2 mRNA level, indicating that sucrose induction is still occurring. Chx concentrations of 50 μM or above reduced the incorporation of [³⁵S]methionine into proteins to 2.3% of the incorporation level in untreated cells (Table 2), confirming the effectiveness of Chx inhibition of protein synthesis in this experiment.

FIG. 5.

Different pathways are involved in CycD induction by sucrose, but none requires de novo protein synthesis. (A) Chx does not block CycD2 or CycD3 induction by sucrose. D7 cells deprived of sucrose were pretreated with 100 μM Chx for 1 h, and at T0 3% sucrose was added in the presence of Chx for a further 4 h (Chx + S). Control cells were induced with sucrose (S) or 100 μM Chx alone (Chx) for 5 h. (B) Cells starved of sucrose for 24 h were treated with OA at different concentrations for 1 h, and then 3% sucrose was added for 4 h in the continuing presence of OA. Control cells were treated with 3% sucrose (S) or 1 nM OA alone (OA) for 4 h. (C) Cells were treated as in panel B but with different concentrations of tautomycin (Tau). Relative signal quantitation is shown under each gel band. Loading control was provided by methylene blue staining of the membrane.

TABLE 2.

Chx inhibition of protein synthesis in the Arabidopsis cell suspension culture^a

Chx concn (μM)	% Incorporation of label ± SD in presence of:
	Chx alone	Chx + 3% sucrose
0		100 ± 8.9b
10	3.1 ± 0.29	3.3 ± 0.45
50	2.3 ± 0.10	2.3 ± 0.10
100	1.9 ± 0.16	2.3 ± 0.16

^aD7 cells were incubated for 24 h in medium without sucrose and then labeled with ³⁵S-labeled amino acids in the presence of Chx and/or sucrose for 6 h. The label incorporated into proteins was determined for triplicate samples by TCA precipitation. Control cells supplied with sucrose alone (0 μM Chx) for 6 h incorporated 48% ± 4.3% of the total intracellular ³⁵S label into protein. This was defined as maximal (100%) incorporation, and other results are presented as a percentage of this level. 

^bSucrose only. 

Analogous accumulation of certain transcripts in the presence of Chx has previously been reported (4, 28). We conclude that neither new-protein synthesis nor progression through G₁ is required for increases in CycD2 or CycD3 mRNA levels, consistent with their proposed roles as effectors of nutrient status information in cell cycle control. As expected, CycD2 and CycD3 were also induced by sucrose when the cell cycle was blocked with hydroxyurea at the G₁/S boundary (results not shown).

Different protein phosphatases (PP) are involved in CycD2 and CycD3 induction.

Chx inhibition experiments show that all the proteins necessary for CycD2 and CycD3 induction by sucrose are present in sucrose-starved stationary-phase cells, implicating posttranslational signaling processes in their induction. Protein phosphorylation plays important roles in the responses of plant cells to external signals (30, 33, 34, 52). We therefore investigated the effects of protein kinase and phosphatase inhibitors on induction of CycD2 and CycD3.

OA is a potent inhibitor of protein serine-threonine phosphatase type 2A (PP2A) (50% inhibitory concentration [IC₅₀] = 0.1 to 1 nM in Arabidopsis extracts) and PP1 (IC₅₀ = 50 nM) (52). PP2B is inhibited only at OA concentrations in excess of micromolar, and PP2C is unaffected by OA (3, 6). CycD3 induction by sucrose was found to be inhibited by OA (IC₅₀ = 10 nM) and abolished by 50 nM OA (Fig. 5B). In contrast, CycD2 induction was inhibited (IC₅₀ = 100 nM) and was therefore 10-fold less sensitive to OA than was CycD3 induction (Fig. 5B). These conclusions were extended by the use of tautomycin, also an inhibitor of PP1 and PP2A (32), and cyclosporin A, a specific inhibitor of the calcium-dependent phosphatase PP2B. CycD3 was sensitive to inhibition by tautomycin (IC₅₀ = 10 nM) (Fig. 5C) and cyclosporin A (IC₅₀ = 10 nM) (results not shown). CycD2 mRNA accumulation was unaffected by 100 nM tautomycin (Fig. 5C) or 50 nM cyclosporin A (not shown).

The differential sensitivity of CycD2 and CycD3 induction by sucrose to PP inhibitors suggests that different PP pathways may be involved in induction of the two genes.

Sucrose operates upstream of phytohormones in CycD3 regulation.

CycD3 expression is induced by cytokinins when sucrose remains present (50). Previous work has shown links between cytokinin and sucrose responses (11). To understand the relationship between cytokinin and sucrose in control of CycD3 expression, we examined CycD mRNA levels in response to hormone removal and readdition in the continuous presence of sucrose. CycD3 levels declined more than 4-fold after incubation of the cells without auxin or cytokinin for 24 h and were induced 10-fold by readdition of the cytokinin zeatin, as previously reported (50). In contrast, CycD2 levels were unaffected by hormone removal and zeatin addition (Fig. 6A).

FIG. 6.

Sucrose is dominant over cytokinin regulation of CycD3 induction, and CycD2 mRNA induction is unaffected by hormones. (A) D7 cells were depleted of auxin and cytokinin for 24 h (T0), and cytokinin (1 μM zeatin) was added for 4 h (+Zea). (B) Hormone- and sucrose-depleted cells (T0) were treated for 4 h by addition of 3% sucrose only (S), 1 μM zeatin only (Zea), or 3% sucrose plus 1 μM zeatin (S+Zea).

When D7 cells were incubated in the absence of both sucrose and hormones for 24 h, low levels of CycD3 and CycD2 mRNA were detected (Fig. 6B, T0). Readdition of cytokinin alone did not result in CycD2 or CycD3 mRNA induction above the T0 level (Fig. 6B, compare T0 and +Zea), even after 24 h of treatment (results not shown). In contrast, sucrose alone without exogenous hormones could induce the expression of both CycD2 and CycD3 within 4 h to a similar level to that seen when hormones are continuously present (Fig. 6B, S [CycD2 induction, 1.7-fold; CycD3 induction, 3.8-fold] [compare Table 1]). It should be noted that the pretreatment of cells for the experiment in Fig. 6A involved removal of hormones only and that for the experiment in Fig. 6B involved removal of both hormones and sucrose. In all other experiments reported here, only sucrose was removed.

We conclude that continued expression of CycD2 requires sucrose and is independent of the presence of hormone. Continued CycD3 expression requires both sucrose (Fig. 2A) and hormones (Fig. 6A). CycD2 and CycD3 expression in sucrose- and hormone-deprived cells is inducible by sucrose alone. Resupply of cytokinin to hormone-deprived cells results in CycD3 induction, provided that sucrose is present. Sucrose is therefore dominant and upstream of hormones in regulating CycD3 expression.

DISCUSSION

Sucrose plays a central role in regulating cellular metabolism and physiology in plants, functioning as the major transported sugar from photosynthetic tissues to sink organs and as a signaling molecule (27, 51). Since cell division in plants is responsive to energy availability, the presence of sucrose or its metabolites is a likely regulator of cell cycle progression. Control mechanisms that interface between the presence of sugar and the cell cycle are therefore probable, as previously reported for yeast (44, 48).

Here we confirm direct carbon source regulation of the plant cell cycle and show that sugar availability plays a major role during the G₁ phase of the cell cycle by controlling the expression of CycD cyclins. Sugar induces the expression of CycD2 within 30 min of application, whereas no increase in CycD3 levels is observed until 4 h after application, results as shown by measurement of CycD2- and CycD3-associated kinase activity. In cells allowed to progress through the cell cycle, this corresponds to induction of CycD2 very early in G₁ and of CycD3 in late G₁, close to the S phase boundary. However, the use of Chx at a concentration that inhibits both cell cycle progression and protein synthesis shows that neither CycD2 nor CycD3 induction requires progress to a specific point in the cell cycle. Moreover, all proteins required for their induction are already present in sucrose-starved cells, thereby implicating CycD2 and CycD3 as direct mediators of the presence of sucrose in cell cycle control.

We may therefore contrast the expression control of genes which are cell cycle regulated and activated as a consequence of cells reaching a specific cell cycle stage with that of genes whose expression conveys information on external conditions. Cell cycle-regulated genes include those encoding histones and A- and B-type cyclins (49). In contrast, induction of CycD2 and CycD3 expression is a direct response to the presence of sugars.

Sugar control of G₁ commitment and progression.

A common theme in eukaryotic cell division is that a major control point operates during G₁ which is under the control of multiple cyclins whose expression normally responds to different signals. After transit through this point, cells become committed to completing a full cell cycle. Dividing plant cells can experience wide fluctuations in carbon source availability due to alterations in environmental conditions, and cell cycle controls in plants in response to nutrient conditions would therefore be expected. Indeed, sucrose can act as a primary inducer of cell division in plant cells and tissues (2, 66, 68). The expression of CycD2 and CycD3 in G₁ in response to sucrose application is consistent with a role for these genes in the sucrose-responsive commitment of cells to division.

In budding yeast, nutritional status is an important factor in transit of G₁, and levels of the G₁ cyclin Cln3 reflect carbon source availability and allow cell cycle progression only under appropriate conditions (18, 44, 48). Indeed, the control of CLN3 expression has parallels with CycD regulation in Arabidopsis, since CLN3 expression is not controlled directly by growth but is induced by glucose and nitrate availability. As observed for CycD induction in Arabidopsis, CLN3 induction also does not require cell cycle progression (44), and, like CycD2, CLN3 mRNA accumulates on cycloheximide treatment (18).

In mammalian cells, the best-studied aspect of commitment to cell division is the transcriptional regulation of cyclin D genes by serum growth factors (35, 57), but the glucose-responsive hormone insulin (65) is among other factors that act on cyclin D activity (29).

In experiments reported here, early-stationary-phase Arabidopsis cells were rendered quiescent by sucrose starvation. On sucrose resupply, the cells left this resting state, reentered the cell cycle, and transited G₁ before reaching S phase. We found that CycD2 mRNA level shows a modest but particularly rapid response to the presence of sucrose within 0.5 h whereas the CycD3 mRNA level is induced after 4 h. This suggests that CycD2 induction may be the primary signal for sugar responses in cell cycle control early in G₁ phase and that CycD3 operates late in G₁ (50). We note that late G₁ induction of CycD3 has also been found in several systems, including tobacco BY-2 cells (63), and in Arabidopsis cultures using Chx or hydroxyurea as synchronizing agents (15, 62).

The presence of sugars is also a prerequisite for cytokinin regulation of CycD3. Like sucrose induction of CycD3, cytokinin induction is independent of protein synthesis (50). However, induction appears to involve a different signal transduction pathway, since the kinase inhibitor staurosporine can mimic cytokinin induction of CycD3 (50) but not sucrose induction of either CycD2 or CycD3 (results not shown).

In Fig. 7, a model is presented in which sugars acts early in G₁ to increase the CycD2 mRNA level and late in G₁ to induce CycD3, where it may integrate independent pathways responding to sucrose and cytokinin (50) to control a commitment point for S-phase entry by participating in the phosphorylation of plant Rb proteins, which are probable substrates for CycD kinases (1, 17, 20).

FIG. 7.

Conceptual scheme for control of CycD2 and CycD3 gene induction during the reentry of quiescent or nondividing cells into the cell cycle. Sugars rapidly induce CycD2 expression, corresponding to early G₁ phase. No protein synthesis is required, and the induction is inhibited by concentrations of OA that suggest the involvement of a single or multiple PP2A in the signal transduction from sugar to gene expression. The induction of CycD2 by 2DG suggests that hexokinase signaling could be involved. CycD3 mRNA starts to accumulate 4 h after sucrose addition, corresponding to the late G₁ phase. Interestingly, this induction also does not require protein synthesis and is also sensitive to OA at levels suggestive of PP2A involvement. The sensitivity of CycD3 mRNA accumulation to the PP2B inhibitor cyclosporin A suggests that a PP2B may also be involved. Previous analysis has shown that cytokinin also induces CycD3 mRNA accumulation provided that sucrose is present. Cytokinin induction is independent of protein synthesis, is stimulated by a PP2A, and is inhibited by a staurosporine-sensitive kinase (50). In each case, although the inhibition observed falls into the range previously defined for PP2A inhibition by OA (52), we cannot exclude the alternative involvement of a PP1.

In mammalian cells, genes whose promoters contain E2F binding sites are activated in late G₁ phase. Progressive phosphorylation causes Rb to dissociate from E2F, which is thereby converted from a transcriptional repressor to a transcriptional activator (10, 57, 58). In plants, no genes have yet been confirmed as being E2F regulated, but the timing of induction of CycD3 in late G₁ phase in Arabidopsis cells in response to both sucrose and cytokinin (50) suggests CycD3 as a possible candidate. We note that near-matches to mammalian E2F binding sites are present in the CycD3 promoter at positions (relative to ATG = +1) −945 (TTTGGCGT) and −461 (TTTGGCGA) and that the very high conservation of the E2F DNA-binding domain between mammals and plants (47, 54) suggests that binding to similar DNA sequences is likely. We propose that CycD3 may be E2F regulated to account for the timing of its activation.

Induction of CycD2 and CycD3 involves independent sugar response mechanisms.

A hexokinase-signaling pathway has been previously characterized for the repression of photosynthetic genes by sucrose (55), and the same pathway may also be involved in sugar induction of some genes, since Arabidopsis strains overexpressing hexokinase show higher levels of induction of the sucrose-induced nitrate reductase gene NR1 (22). In the hexokinase-mediated sugar response pathway, hexokinase activity itself is responsible for mediating the response, since hexokinase substrates such as 2DG and mannose (which are not readily further metabolized after hexokinase-mediated phosphorylation) are powerful gratuitous effectors whereas sugars such as 6DG, which are not hexokinase substrates but are nevertheless transported into cells, have no effect. CycD3 was not induced by any of these analogs, and indeed its basal level was reduced by 2DG; we therefore conclude that the hexokinase pathway is not involved in mediating the CycD3 response to carbon source. However, CycD2 was fully induced by 1 mM 2DG (Fig. 2B) or 10 mM mannose. Since CycD2 levels did not increase in response to 6DG, hexokinase may be involved in the CycD2 response to carbon source availability.

CycD gene expression responds within 0.5 to 4 h to sugars. In contrast, many sugar-modulated genes show changes in gene expression over extended periods (3 to 7 days), although some show altered transcript abundance within hours (reference 27 and references therein). We also observed a relatively rapid decline in CycD2 and CycD3 mRNA levels after sucrose removal from suspension-cultured cells, reaching basal levels within 12 h (results not shown). However, in intact seedlings, CycD2 and CycD3 levels took up to 48 h to decline (Fig. 3B and data not shown), presumably because of mobilization and gradual depletion of carbohydrate reserves. In agreement with this suggestion, seedlings kept in the dark underwent a more rapid loss of CycD2 and CycD3 transcript levels (results not shown).

We also observed that in common with most other sugar-regulated genes, removal and resupply of sucrose resulted in a greater CycD2 and CycD3 response than its continued presence. This suggests that flux rather than the steady-state sugar level is the primary signal for CycD2 and CycD3 regulation (27).

PP involvement in sucrose induction of CycD2 and CycD3.

PP have been implicated in many response pathways in plant cells (60). In particular, a PP1 has been implicated in light-inducible gene expression (56) and PP1 and PP2A have been implicated in sucrose regulation of gene expression in sweet potato (64).

The differential sensitivities of PP2A and PP1 to OA inhibition can be used to distinguish the class of PP involved in a particular signal transduction pathway. In Brassica napus seed extracts, PP2A is inhibited by OA (IC₅₀ = 0.1 nM) whereas PP1 activity is 100-fold less sensitive to OA (IC₅₀ = 10 nM) (60). However, extrapolation to intact cells is complicated by the uptake of the inhibitor and by reduced inhibition if high intracellular PP concentrations prevail. In intact Arabidopsis plantlets, an extracellular concentration of 1 μM was needed to achieve inhibition of PP2A equal to the effect of 1 nM OA in cell extracts (52). Complete inhibition of jasmonic acid-induced expression of JR1 gene expression was observed with 100 nM extracellular OA (IC₅₀ = 20 nM). The authors concluded from in vitro and in vivo experiments that inhibition in this range was due to loss of PP2A activity (53).

We report here the inhibition of CycD3 and CycD2 induction by extracellular OA concentrations of 10 and 100 nM, respectively, for intact Arabidopsis cells. Inhibition of both genes falls into the range of PP2A inhibition (52). Similar results were obtained with tautomycin, another inhibitor of both PP2A and PP1. CycD3 induction was inhibited by tautomycin (ID₅₀ = 10 nM) whereas CycD2 induction was unaffected by 100 nM tautomycin. This difference in inhibition of CycD2 and CycD3 induction by OA and tautomycin may be due either to the involvement of different PP2As or to different levels of inhibition of a single enzyme needed to block the two pathways. In addition, a PP2B is involved in CycD3 regulation, since induction is strongly inhibited by cyclosporin A. Combined with the evidence that protein synthesis is not required for CycD2 or CycD3 gene induction, we suggest that PP may play a direct role in producing the signal for the sucrose induction of these genes (Fig. 7).

Cell cycle control in mammals and yeast is regulated by phosphorylation and dephosphorylation (43). The only studies of the effects of PP inhibitors on cell cycle in plants have used concentrations 2 to 3 orders of magnitude higher than these reported here, and widespread and nonspecific effects are anticipated at such high (micromolar) concentrations. OA blocks the cell cycle during early mitosis in suspension cultures of Nicotiana plumbaginifolia when a concentration of 12 μM is applied (72) and of BY-2 cells when 1 μM OA is used (19). To our knowledge, no previous data show an effect of PP inhibitors on cell cycle regulation of plant cells during the G₁ phase or at concentrations below 1 μM.

Physiological significance.

Sugar availability in whole plants is controlled by photosynthesis and the use of stored carbohydrate reserves. Direct evidence for the role of sucrose in stimulating cell division comes from the response of plant cells in intact or excised plant tissues to sucrose (2, 66–68, 71). Withdrawal of sucrose from the medium of actively dividing cells causes the majority of cells to arrest in G₁, and resupply of sucrose leads to S-phase progression and a resumption of division (66, 68). Cell division within a tissue can also be synchronized by withholding and resupplying sugars (70).

Evidence also links endogenous sugar levels to the control of cell division in intact plants. At elevated CO₂ concentrations, a more rapid progression through the cell cycle is due to an increase in the number of rapidly cycling cells in the apical meristem and to a reduction in cell cycle length (23, 26). More rapid cell divisions due to a shorter G₁ phase were observed when Chrysanthemum plants were exposed to increase light levels (42). Both high light and elevated CO₂ levels are likely to increase the photosynthetic yield and sugar levels. In addition, cell divisions can be induced in latent buds of sunflower by sugar application (2).

All these examples have been shown to be, or can be predicted to be, linked to a reduced length of the G₁ phase of the cell cycle or an increased rate of commitment to cell division by otherwise resting cells. Both of these phenomena may therefore be linked to CycD2 or CycD3 by sugar induction, which, as we show here, occurs at low concentrations of applied sucrose in both Arabidopsis cells and intact seedlings.