INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Basic features of cell cycle control are remarkably conserved in all eukaryotes and principal control points at the G1/S boundary before entry into S phase, and at the G2/M boundary before mitosis have been identified in yeast, animals, and plants (Pines, 1995; Huntley and Murray, 1999). Transit through these control points requires activated kinase complexes consisting of a cyclin-dependent Ser/Thr kinase (CDK) bound to a cyclin. CDK activity is dependent on the cyclin, which also determines the substrate specificity and the subcellular localization of the CDK complex (Pines, 1995). The cyclin is therefore regarded as the regulatory component of the complex, a role reflected in its highly regulated pattern of transcription and degradation. In contrast to cyclins, there is little evidence (in yeast and mammals) for the specific regulation of CDK expression with transcript and protein levels generally observed at a constant level throughout the cell cycle, suggesting that the activity of the complex is not regulated by changes in the abundance of the CDK subunit.

In yeasts a single CDK (encoded by cdc2⁺ in the fission yeast Schizosaccharomyces pombe) in association with different stage-specific cyclins regulates the progression through all phases of the cell cycle, whereas in animals several distinct CDKs have been isolated and shown to function at different stages in the cell cycle (Morgan, 1997). These CDKs are characterized by distinct sequences within the cyclin binding motif of the CDK, namely PSTAIRE in the case of the yeast CDK and the principal mitotic CDK of animals.

Plants also contain multiple CDKs, including those of the PSTAIRE type known as cdc2a or CDK-a, and a novel type of plant-specific CDK characterized by the variant sequences PPTALRE or PPTTLRE, known as CDK-b (for review, see Segers et al., 1998; Huntley and Murray, 1999; Mironov et al., 1999). The CDK-b proteins appear to fall into two subgroups on the basis of sequence relationships (Huntley and Murray, 1999; Umeda et al., 1999; Joubès et al., 2000). One group contains Arabidopsis CDC2b, snapdragon (Antirrhinum majus) Cdc2c, and alfalfa (Medicago sativa) cdc2MsD, which all contain the sequence PPTALRE and for which the name CDK-b1 subgroup has been proposed (Hirayama et al., 1991; Imajuku et al., 1992; Hirt et al., 1993; Fobert et al., 1996; Segers et al., 1996; Magyar et al., 1997; Huntley and Murray, 1999). The other subgroup named CDK-b2 contains snapdragon Cdc2d, alfalfa cdc2MsF, rice (Oryza sativa) Cdc2Os3, and Arabidopsis Cdc2dAt (Hirt et al., 1993; Kidou et al., 1994; Fobert et al., 1996; Magyar et al., 1997; Umeda et al., 1999; Huntley and Murray, 1999). The CDK-b2 group CDKs have the sequence P(S/P)TTLRE with the exception of Cdc2Os3, which has PPTALRE. (Fig. 1; see legend for nomenclature of CDK genes and proteins used here.)

View larger version (16K): [in this window] [in a new window]   Figure 1.   A, Tobacco CDKB1 structural features. Numbers indicate boundaries of conserved features; T14, Y15, and T170 correspond to conserved residues involved in CDK phosphoregulation in yeast and mammals. For details, see text. B, The relationship between the protein sequences of tobacco CDKs including that of CDKB1;1 and CDKA;4 (isolated in this study) and other plant a- and b-type CDKs. The CDKs have been named (both in this figure and throughout the text) according to original designations; the species abbreviation has been appended when not evident from the original name. Am, A. majus (snapdragon); At, Arabidopsis; Cr, Chenopodium rubrum (red goosefoot); Ms, M. sativa (alfalfa); Nicta, N. tabacum (tobacco); Os, O. sativa (rice); Zm, Zea mays (maize). The genes isolated in this study are named according to C.P.E.N. nomenclature (Price et al., 1996). Nicta; CDKA;4 is the fourth tobacco a-type CDK described; cdc2Nt1 was previously presented by Setiady et al. (1996) and further unpublished sequences in Joubès et al. (2000). Here we designate the CDK groups by CDK-a, CDK-b1, or CDK-b2, the protein products as CDKA/CDKB1, and the genes as CdkA/CdkB1. 1, Colasanti et al., 1991; 2, Hirayama et al., 1991; 3, Hashimoto et al., 1992; 4, Hirt et al., 1993; 5, Kidou et al., 1994; 6, Fobert et al., 1996; 7, Renz et al., 1997; 8, accession no. AAD30597.1 cited in Huntley and Murray, 1999; 9, accession no. AJ297936; 10, accession no. AF289465; and 11, accession no. AF289467.

Uniquely, plant CDK-b-type CDKs are strongly transcriptionally regulated during the cell cycle. Transcripts of Arabidopsis CDC2b, snapdragon cdc2c, and alfalfa cdc2MsD are reported to be present during S-G2-M phases (Fobert et al., 1996; Segers et al., 1996; Magyar et al., 1997), whereas transcripts of snapdragon cdc2d, alfalfa cdc2MsF, and rice cdc2Os3 are only detected during G2-M (Fobert et al., 1996; Magyar et al., 1997; Umeda et al., 1999). However the abundance of the protein products of CDK-b genes has only been examined for alfalfa cdc2MsD and cdc2MsF and for rice cdc2Os3, and only in the case of the alfalfa proteins was this in a synchronized cell culture system (Magyar et al., 1997; Umeda et al., 1999). These results suggest protein levels broadly follow transcript abundance. Only the associated kinase activity of CDK-b2 protein cdc2MsF has been studied and was found to be only present in samples co-incident with the mitotic index peak of the cells (Magyar et al., 1997). In no case has kinase activity of a CDK-b1 been reported.

The tobacco (Nicotiana tabacum L. cv Bright Yellow- 2 [BY-2]) cell line (Nagata et al., 1992; Nagata and Kumagai, 1999) is the most highly synchronizable plant cell system and is thus ideal for studies of the plant cell cycle. Previously, a cDNA of a PSTAIRE (CDK-a) gene cdc2Nt1 (renamed Nicta;CdkA;3 by Joubès et al., 2000) has been cloned from tobacco, and RNA gel-blot analysis showed this gene to be preferentially expressed in dividing BY-2 cells but not to show significant cell cycle regulation of transcript abundance (Setiady et al., 1996). A later study of CDC2a (CDK-a) protein levels and histone H1 kinase activities in propyzamide synchronized cells showed that the protein levels remained at a constant level throughout the cell cycle but that kinase activity was cell cycle regulated (Reichheld et al., 1999).

Here we report the isolation from a BY-2 cell cDNA library of a PSTAIRE CDK-a (Nicta;CdkA;4 [accession no. AF289467]) highly related to cdc2Nt1 (Setiady et al., 1996) and two closely related novel tobacco CDKs (Nicta;CdkB1;1 [accession no. AF289465] and Nicta;CdkB1;2 [accession no. AF289466]) containing the PPTALRE sequence and belonging to the CDK-b1 subgroup, the first tobacco non-PSTAIRE CDKs to be identified. We have carried out the first analysis of tobacco CDK-b expression, protein abundance, and kinase activity during the BY-2 cell cycle and show that CDKB1 is likely to have a role in the regulation of mitosis. We confirm the potential role of CDKA kinase activity in G2/M suggested by the results of Reichheld et al. (1999), and we propose that the functions of CDKA and CDKB1 however may be distinct due to the different timing of their activity. We also show that CDKA but not CDKB1 shows activity at the G1/S boundary.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Isolation of a CDK-b1-Type Tobacco CDK

A cDNA library from exponentially growing BY-2 cells was screened at low stringency with the snapdragon Amcdc2c cDNA, which encodes a CDK-b1-type CDK carrying a PPTALRE sequence (Fobert et al., 1996; see "Materials and Methods"). Twenty-four positive clones were initially obtained, and after further screening rounds two non-PSTAIRE clones were sequenced completely; one (named Nicta;CdkB1;1) was 1,173 bp in length and the other (Nicta;CdkB1;2) was 1,334 bp. Both encode open reading frames (ORF) of 303 amino acids, which are identical except for a single amino acid difference (N60 replaced by H). However, the clones show significant differences in their nucleotide sequences with 100 nucleotide differences in their 5'- and 3'-untranslated regions and 25 nucleotide differences in the region encoding the ORF. This results in an overall 88% DNA sequence identity (97.25% within the ORF-encoding region) and suggests that these clones arise from separate genes. To confirm this result, PCR amplification of the genomic sequences corresponding to the two cDNAs was carried out using a common 5' primer and 3' primers specific to Nicta;CdkB1;1 and Nicta;CdkB1;2 (details of primers available from the authors). Amplification of a BY-2 genomic DNA template produced a 3-kb product using the primer combination specific for Nicta;CdkB1;1 and a 4.5-kb product using the primer combination specific Nicta;CdkB1;2. The cDNA clones produced products the expected size of approximately 1 kb. This indicates that additional intron sequences are present in the CdkB1;2 genomic DNA that are not present in the CdkB1;1 genomic sequence and demonstrates that the cDNA clones correspond to separate genes. These might arise from the two ancestral genomes that comprise tobacco.

Six of the remaining clones did not hybridize to a CdkB1;1 probe at high stringency but were found to produce a PCR product with primers designed against cdc2Nt1 (Setiady et al., 1996), a PSTAIRE CDK (data not shown). One of these clones (named Nicta;CdkA;4) was subcloned and sequenced and compared with the previously isolated tobacco PSTAIRE CDK, cdc2Nt1 (Setiady et al., 1996). CdkA;4 was found to have 11 nucleotide differences compared with cdc2Nt1 within the ORF-encoding region, resulting in a single amino acid difference in the encoded proteins (E99 replaced by K) and an identity 98.75%. CdkA;4 and cdc2Nt1 could be alleles of the same gene and therefore reflect genetic differences between cultivars (cdc2Nt1 was isolated from tobacco cv Samsun), or they could be separate genes with origins in the different parental genomes of tobacco. Neither the nucleic acid probes nor antisera used in the work here can distinguish the differences between these clones because of their close similarity, so we refer simply to CdkB1 and CdkA.

CDK-B1 Sequence Features and Relationship to Other CDKs

The CdkB1 clones encode non-PSTAIRE plant CDKs, containing a PPTALRE motif in the region of the CDK predicted to bind to a cyclin partner (Fig. 1A; Jeffrey et al., 1995). CDKB1 also contains all the other conserved regions of a functional CDK including the ATP-binding region, catalytic domain, and T-loop (De Bondt et al., 1993; Jeffrey et al., 1995). In addition the T14, Y15, and T160 (at position 170 in CDKB1) residues involved in the phosphoregulation of yeast and vertebrate CDKs are conserved (Lew and Kornbluth, 1996).

Comparison of CDKB1 with sequence databases using BLAST confirmed that it is most similar to plant and animal CDKs (data not shown). A comparison of CDKB1 with the polypeptide sequences of other plant CDKs using CLUSTAL X is shown graphically in Figure 1B. The relationships between sequences obtained in this analysis are consistent with those obtained using other algorithms (Doonan and Fobert, 1997; Segers et al., 1998; Umeda et al., 1999). CDKB1 falls into the b1 group of plant CDKs (Segers et al., 1998; Huntley and Murray, 1999), showing closest similarity to snapdragon Cdc2c (Fobert et al., 1996), alfalfa cdc2MsD (Magyar et al., 1997), and Arabidopsis CDC2b (Hirayama et al., 1991; Imajuku et al., 1992; Segers et al., 1996).

To assess the similarity within and between the CDK groups, consensus sequences for each group were generated from a CLUSTAL X alignment of the plant a- and b-type CDK groups with human PSTAIRE-containing CDK1 and used to calculate the homology within and between groups. The percentage identity between the a- and b-type CDK groups is 42% to 50%, whereas within each CDK group it is 70% to 77%. As previously reported, the a-type CDKs are more similar to human CDK1 (58% identity) than they are to the b-type CDKs (42%-46% identity) (Doonan and Fobert, 1997; Segers et al., 1998).

CdkB1 RNA Is Expressed in a Growth Phase-Dependent Manner

The RNA expression of CdkA and CdkB1 was compared during the complete BY-2 cell growth cycle. Changes in cell number and mRNA transcript levels during the growth cycle are shown in Figure 2. CdkB1 was highly expressed at d 1 and 3 (i.e. within the exponential growth phase, which lasts until d 4). By 5 d, CdkB1 transcript levels had declined substantially as cells exited the cell cycle and entered the stationary phase. At 7 d, CdkB1 transcripts were not detected. CdkA expression showed a similar expression pattern, except that the transcripts stabilized at a higher level in stationary phase cells (Fig. 2; Setiady et al., 1996). The expression of histone H4 (data not shown) and CycD3;2, both markers for the exponential phase of growth in BY-2 cells, were also included for comparison. We conclude that CdkB1 RNA is expressed in a growth phase-dependent manner in BY-2 cells.

View larger version (25K): [in this window] [in a new window]   Figure 2.   Growth phase-dependent RNA expression of tobacco CDKs. A, Growth curve of tobacco BY-2 cells in batch suspension culture. Stationary phase cells (7 d after previous subculture) were subcultured into fresh medium and incubated for 7 d. The mean cell number was determined daily; the error bars represent the SE determined from three samples. B, Total RNA isolated from samples taken at the indicated time points was blotted and hybridized with the probes indicated. CycD3;2 was included for comparison as a gene expressed during the exponential phase of growth (Sorrell et al., 1999). The loading control was by methylene blue staining of the membrane (Riou-Khamlichi et al., 2000).

CdkB1 RNA Expression Is Induced from Early S Phase As Stationary Phase BY-2 Cells Re-Enter the Cell Cycle

The timing of CdkB1 RNA expression, protein expression, and kinase activity was investigated as partially synchronized cells exit from stationary phase and re-enter the cell cycle (Sorrell et al., 1999). Stationary BY-2 cells were subcultured into fresh medium and samples taken at time points over 10 h for RNA and protein-blot analyses and protein kinase activity assays. Histone H4 was included in the RNA expression experiments to act as marker for the start of S phase and started to accumulate rapidly between 6 and 7 h (Fig. 3A), reaching a high level at 10 h. This indicated that the majority of cells started to enter S phase in a partially synchronous manner at approximately 7 h as previously described (Sorrell et al., 1999). CdkB1 transcript levels also started to accumulate rapidly between 6 and 7 h, suggesting that their expression was induced as cells entered S phase. In contrast, CdkA transcripts showed only a slight increase on cell cycle re-entry. Although the CdkA transcripts were only marginally induced above the level seen in stationary phase cells, the signal was readily detected (2- to 3-h exposure time), indicating that they were moderately abundant throughout the time course of the experiment. An independent experiment confirmed that the increase in CdkB1 transcripts occurred at the same time or slightly before histone H4 transcript accumulation (data not shown).

View larger version (20K): [in this window] [in a new window]   Figure 3.   RNA levels, protein abundance, and histone H1 kinase activity of tobacco CDKs in stationary BY-2 cells re-entering the cell cycle. A, Stationary BY-2 cells were incubated in fresh medium for 10 h, total RNA was extracted from samples at the indicated time points and was blotted and hybridized with the cDNA probes indicated on the left. CycD3;2 was included for comparison as a gene, which is induced in mid-G1 in cells re-entering the cell cycle (Sorrell et al., 1999). Histone H4 expression was included to monitor the entry into S phase (G0, G0 phase; G1, G1 phase; S, S phase). To control for loading, the membrane was stained with methylene blue. B, Samples from the same time points as A were used for western (protein gel) blots and immunoprecipitated histone H1 kinase assays. The western blot of the time point samples probed with CDKA antibody is shown above the quantitation of the corresponding kinase activities (measured in arbitrary units), measured after immunoprecipitation of CDKA-containing protein complexes. To control for protein concentrations, a duplicate gel was stained with Coomassie Blue. CDKB1 protein and kinase activity were not detected in these samples, although CDKB1 could be readily detected in a positive control extract (data not shown).

The protein levels of CDKA were similar to its RNA abundance, except that low levels present in stationary phase cells (0 h) increased in the first 2 h after subculturing of the cells, and then remained relatively constant as cells moved through G1 into S phase (Fig. 3B). In contrast to the RNA and protein expression, the immunoprecipitated histone H1 kinase activity of CDKA increased 3-fold between 4 and 6 h after re-entry into the cell cycle at or shortly before the G1/S boundary (Fig. 3B). In contrast, CDKB1 protein and kinase activity were undetectable in the 10 h after cell cycle re-entry (data not shown).

CdkB1 RNA and Protein Levels, and the Associated Kinase Activity Are Dependent on Cell Cycle Phase

The induction of CdkB1 at the G1/S boundary could be characteristic only of quiescent cells re-entering the cell cycle as found for cyclin CycD3;2 (Sorrell et al., 1999) or could occur in every cell cycle in rapidly dividing cells. We therefore investigated the expression of CdkB1 during the cell cycle in synchronously cycling BY-2 cells. The cells were synchronized by blocking the cell cycle in early S phase with aphidicolin, an inhibitor of DNA polymerase  activity (Nagata et al., 1992). After the aphidicolin block was released, progress of cells through the cell cycle was followed by monitoring changes in mitotic index and the proportion of cells in S phase using flow cytometry (Fig. 4A). Changes in the abundance of CdkB1 mRNA levels were compared at different times during the cell cycle. As previously described for the closely related cdc2Nt1, CdkA transcripts remained at an approximately constant level during the cell cycle (Setiady et al., 1996). In contrast, CdkB1 transcript levels varied markedly with the lowest levels in G1 and the highest levels in S, G2 (Fig. 4B), and M phases (data not shown). The same results were obtained with both a full length probe and a 3'-end probe specific for the CdkB1;1 cDNA (data not shown), the latter being used to confirm that the expression pattern observed was not due to cross-hybridization with other as-yet-unidentified tobacco CDK transcripts. This result also suggests that Cdkb1;1 and Cdkb1;2 are similarly regulated.

View larger version (20K): [in this window] [in a new window]   Figure 4.   RNA expression, protein expression, and kinase activity of tobacco CDKs during the cell cycle in synchronous BY-2 cells. A, Cells were synchronized in early S phase with aphidicolin as described in Sorrell et al. (1999). Progress of the cells through the cell cycle was monitored by following changes in mitotic index and by determining the proportion of S phase cells using flow cytometry. B, RNA was extracted from the cell samples at time points indicated and probed with cDNA probes as indicated on the left (M, mitosis; G1, G1 phase; S, S phase; G2, G2 phase). Loading control was by methylene blue staining of the membrane. C, Protein expression and histone H1 kinase activity of the CDKs was assayed in equivalent synchronization samples. The western blots of the time point samples were probed with antibodies as shown on the left. The corresponding histone H1 kinase activity present in CDKA and CDKB1 immunoprecipitates is quantitated in arbitrary units and shown below the corresponding western blot. CDKA kinase levels were determined independently between three and six times at each time point. The average and SE of these measurements are shown.

Corresponding protein and histone H1 kinase activities of the CDKs were monitored in similar synchronization experiments (Fig. 4C). CDKA protein levels, like the RNA levels, were relatively constant throughout the cell cycle except for a slight decline during G1 phase, but the immunoprecipitated kinase activity was higher during the S and G2 phases of the cell cycle. CDKA kinase activity remained high until the G2/M transition, and then declined between 6 and 8 h after aphidicolin release.

CDKB1 protein levels showed a gradual accumulation from S phase until mitosis, followed by a gradual decline more marked than the regulation of its mRNA. CDKB1-associated histone H1 kinase levels showed a sharp peak at the G2/M boundary, followed by an abrupt decline at the same time as CDKA histone H1 kinase activity was also lost.

To confirm that RNA transcript levels of the CdkA and CdkB1 genes show little variation between the S and M phases, their expression was examined in cells blocked in S phase with aphidicolin or in mitosis using the anti-tubulin drugs oryzalin or propyzamide. Stationary cells were diluted in fresh medium containing aphidicolin or oryzalin and cultured for 24 h. For the propyzamide treatment, G2/M cells obtained 4 to 5 h after the release from an aphidicolin block were treated with the drug for 5 h (Nagata et al., 1992). Figure 5 shows the transcript abundance in blocked, stationary phase and exponentially growing (3 d after subculture) cells. The histone H4 transcript levels and the mitotic index show that aphidicolin blocked the cells in S phase, and oryzalin or propyzamide blocked cells in mitosis. The mitotic block of oryzalin was less efficient than that of propyzamide as seen by the higher histone H4 transcript level and lower mitotic index. However, this lower mitotic index (28%) is consistent with other studies (Shaul et al., 1996) and indicates that a substantial number of cells are in mitosis compared with stationary, exponentially growing, or S phase-blocked cells. As expected, the two genes showed very little variation in transcript abundance between S and M phase blocked cells, consistent with the results of the aphidicolin experiment (Fig. 4).

View larger version (47K): [in this window] [in a new window]   Figure 5.   RNA expression of tobacco CDKs in cells treated with cell cycle inhibitors. Cells were arrested in S phase with aphidicolin or in mitosis with oryzalin or propyzamide (Sorrell et al., 1999) and harvested for RNA analysis. Cell cycle arrest was confirmed by determining the mitotic index and level of histone H4 transcripts. Samples from stationary and exponentially dividing (3 d after subculture) cells are included for comparison. CycD3;2 transcripts are also included for comparison of transcripts that remain constant throughout the cell cycle. The loading control was by methylene blue staining of the membrane.

Further confirmation that CdkB1 transcript levels are lower in G1 than in other phases of the cell cycle was shown by examining its expression in synchronous cells progressing through G1 and into S phase after release from a sequential aphidicolin-propyzamide mitotic block (Fig. 6). After release from the block, the progress of the synchronous cells was followed by monitoring changes in mitotic index (Fig. 6A) and the expression of histone H4 (Fig. 6B). As expected, when cells left mitosis and entered G1 the expression of CdkB1 declined, CdkB1 transcript levels started to accumulate again in late G1 phase 6 h after release and at the same time or shortly before histone H4 induction. These data are consistent with the late G1 induction of CdkB1 RNA seen as stationary phase BY-2 cells re-enter the cell cycle (Fig. 3A). As expected, CdkA transcripts remained at a constant level.

View larger version (27K): [in this window] [in a new window]   Figure 6.   RNA expression of tobacco CDKs in propyzamide-synchronized cells. Cells were synchronized in mitosis with sequential aphidicolin-propyzamide treatment (Nagata et al., 1992; Sorrell et al., 1999). After release from the propyzamide block, progress through the cell cycle was monitored by following changes in the mitotic index (A). B, Cells were harvested for RNA analysis at the indicated time points. The approximate position of the cells in the cell cycle at each time point is shown (M, mitosis; G1, G1 phase; S, S phase). Histone H4 expression was included as a marker for S phase, and CycD3;2 was included as a comparison of transcripts that remain constant throughout the cell cycle (Sorrell et al., 1999). The loading control was by methylene blue staining of the membrane.

In summary, during the cell cycle CdkB1 transcript and protein levels fluctuate, being highest in S, G2, and M phase, whereas its associated histone H1 kinase activity has a narrow period of activity at the G2/M boundary. In contrast CdkA RNA transcripts remain constant throughout the cell cycle, consistent with previous observations (Setiady et al., 1996). CDKA protein expression also remains relatively constant throughout the cell cycle and its kinase activity is high for approximately 5 h during S and G2 phases, consistent with the results of Reichheld et al. (1999) and declines along with CDKB1 histone H1 kinase activity after the G2/M boundary.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Cyclin-dependent kinases play key roles in controlling cell cycle progression in all eukaryotes. Their activity is regulated through multiple mechanisms of which the binding of cyclins and activating and inhibitory phosphorylations are of particular importance (Lew and Kornbluth, 1996). The archetypal CDK of yeast is involved in both the G1/S and G2/M phase transitions and contains the sequence PSTAIRE in the 1 helix of its cyclin binding domain (Jeffrey et al., 1995) and binds different cyclins at different cell cycle stages. This basic model is elaborated in higher organisms by the presence of multiple CDKs, some of which possess variant PSTAIRE motifs. In mammals, for example, the PSTAIRE-containing CDK1 most closely related to the yeast CDK is involved only in mitosis, and different proteins including CDK2 and the non-PSTAIRE CDK4 and CDK6 are involved in G1/S and S phase control (Morgan, 1997). However, in general the cyclin partner is preserved as the component whose abundance is regulated. Cell cycle regulation of CDK gene expression has not been observed in yeast, although in mammals there is limited cell cycle transcriptional regulation of CDK kinases (McGowan et al., 1990; Dalton, 1992).

The plant cell cycle also uses the same basic mechanisms as other eukaryotes, but significant differences have evolved in the classes of molecules and modes of regulation (Mironov et al., 1999). In particular, plants contain a novel group of CDKs, which not only possess a unique cyclin-binding motif in the 1-helix characterized by the sequence PPTA/TLRE, but also exhibit strong transcriptional regulation during the cell cycle (Hirayama et al., 1991; Fobert et al., 1996; Segers et al., 1996; Magyar et al., 1997; Segers et al., 1998; Huntley and Murray, 1999; Mironov et al., 1999; Umeda et al., 1999). The CDK-b group has recently been divided into two distinct subgroups (Huntley and Murray, 1999; Umeda et al., 1999), CDK-b1, whose members are expressed from S-G2-M phase, and CDK-b2, which are expressed in a narrower window from G2-M (Fobert et al., 1996; Magyar et al., 1997). In contrast, genes of the CDK-a (cdc2a) group have been found to show little cell cycle regulation and in general are rather widely expressed not only in actively dividing cells but also in cells that retain the potential to resume division once given the appropriate signal (Martinez et al., 1992; Hemerly et al., 1993; for review, see Mironov et al., 1999).

Despite a number of studies that have examined the transcriptional regulation of CDK genes in plants by in situ hybridization or RNA gel blotting, there has been little detailed analysis of protein abundance or protein kinase activities during cell cycle progression. In alfalfa, protein levels have been analyzed in synchronized cultures for CDK-a (cdc2MsA/B), CDK-b1 (cdc2MsD), and CDK-b2 (cdc2MsF; Magyar et al., 1997). These authors also presented histone H1 protein kinase activities for CDK-a and CDK-b2-type proteins only during the cell cycle. The results showed that CDK-a histone H1 kinase activity was high at the G1/S boundary and early S phase and declined after the G2/M boundary. In contrast, the CDK-b2 cdc2MsF-associated histone H1 kinase activity was restricted to samples with a high proportion of mitotic cells. A limited analysis of tobacco CDKA protein levels and histone kinase activities in propyzamide synchronized cells has been described (Reichheld et al., 1999). The results showed that CDKA protein levels remained at a constant level throughout the cell cycle, whereas CDKA kinase activity was low in early G1, increased at the G1/S phase transition, peaked during S to mid-G2 phase, and declined in early mitosis. Here we present the cloning of cDNAs corresponding to tobacco CDK-b1 genes, and examine the expression, protein abundance, and histone H1 kinase activity of tobacco CDKA and CDKB1. This represents the first analysis of the histone H1 kinase activity of a CDK-b1 and the first study of CDK activity of cells transiting from stationary phase, across G1 into S phase.

Our results show that CDKA protein levels increase in early G1 phase within 2 h of supplying fresh medium but subsequently remain relatively constant although with a slight decline after cells exit mitosis and enter a second G1 phase. CDKA histone H1 kinase activity remains constant until mid-G1 phase and then increases approximately 3-fold in late G1 before S phase is initiated. Kinase activity increases further during S phase. These results clearly demonstrate that CDKA activity increases before the G1/S transition in tobacco BY-2 cells re-entering the cell cycle. Consistent with the association of CDK-a with D-type (CycD) cyclins in plants and the presence of CycD-associated histone H1 kinase activity in late G1 phase (Nakagami et al., 1999; Cockcroft et al., 2000; Riou-Khamlichi et al., 2000), the CDKA activity we observe may be due to its association with CycD proteins and could therefore be participating in phosphorylation of the tobacco retinoblastoma (Rb) protein (Huntley et al., 1998; Nakagami et al., 1999). We conclude therefore that the protein kinase activity of the CDK-a (cdc2a) group is involved not only in mitosis, but also in the G1/S transition in plants.

During G1 phase, CdkB1 mRNA, protein, and kinase activity are only detectable at low levels, but using aphidicolin to synchronize cells in early S phase, it is possible to examine the behavior of CDKA and CDKB1 later in the cell cycle. Consistent with the results of Reichheld et al. (1999) CDKA histone H1 kinase activity was observed to be high in S phase and remain high until the G2/M boundary or early mitosis, when it shows a significant decline. CDKB1 histone H1 kinase activity, in contrast, does not increase until G2 phase and shows a narrow peak of activity corresponding to the G2/M transition. Thus CDKA kinase activity appears to increase from mid-G1 phase and remain high throughout S phase and decrease at G2/M, whereas CDKB1 kinase activity appears only during G2 but disappears at the same time as CDKA activity. It may be noted that the CDKB1 activity declines before the mitotic index peak, whereas the kinase activity of the CDK-b2 protein Cdc2MsF in synchronized alfalfa cells is coincident with the mitotic index peak (Magyar et al., 1997). Taken together, these results suggest that CDK-b1 kinase activity is needed for the G2/M boundary together with CDK-a activity, whereas CDK-b2 kinase activity is involved in a later mitotic control. This suggestion is also supported by the later transcript accumulation of CDK-b2 genes (Fobert et al., 1996; Magyar et al., 1997).

The data presented here confirm the value of tobacco BY-2 cells for analysis of the different kinase activities involved in plant cell cycle progression and show novel aspects of the plant cell cycle in the regulation of CDK-b kinases and the activity of PSTAIRE-type CDK-a kinases during G1 phase.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Isolation of Tobacco (Nicotiana tabacum) CDKs

Randomly primed [-³²P]-labeled snapdragon (Antirrhinum majus) Amcdc2c cDNA probe (Fobert et al., 1996) was used to screen approximately 5 × 10⁵ clones from a cDNA library constructed in a Lambda Zap Express vector with poly(A⁺) RNA isolated from exponentially growing BY-2 tobacco cells (Sorrell et al., 1999). Hybridizations were carried out at 50°C, and the membranes were washed twice for 10 min in 2× SSC/0.1% (w/v) SDS at room temperature and once for 10 min in 0.1× SSC/0.1% (w/v) SDS at 40°C. Selected positive clones were purified by further screening rounds and the cDNA inserts were excised in vivo according to the manufacturers instructions and sequenced on both strands. The remaining clones were analyzed without further rounds of purification, by rehybridizing with a CdkB1;1 probe at high stringency, and those that did not hybridize were subject to a PCR with primers designed against cdc2Nt1 (Setiady et al., 1996). One of the clones that gave a product with these primers was subcloned into a modified pBluescript vector (Borovkov and Rivkin, 1997) and sequenced. Sequence analysis was carried out using the Genetics Computer Group package (Madison, WI) and the Sequencher 3.0 software program (Gene Codes Corporation, Ann Arbor, MI). The relationships between the tobacco and other plant CDKs were determined using BLAST (Altschul et al., 1990) and CLUSTAL X (Thompson et al., 1997).

Culture of Tobacco BY-2 Cells, Experimental Treatments, and RNA Analysis

Tobacco BY-2 cells were maintained as previously described (Nagata et al., 1992). Procedures for cell cycle re-entry analysis and cell synchronization have been described previously (Sorrell et al., 1999). DNA content was determined by flow cytometry using a Partec PAS-III (Partec GmbH, Münster, Germany) flow cytometer and Multicycle for Windows software (Phoenix Flow Systems, San Diego). Total RNA was extracted, separated on formaldehyde-agarose gels, blotted onto nylon membranes, and hybridized using standard procedures (Sorrell et al., 1999). RNA loading was controlled by methylene blue staining of transfer membranes before hybridization (Riou-Khamlichi et al., 2000).

Protein Methods

Procedures for protein extraction, SDS-PAGE, and western-blot analysis are described in Cockcroft et al. (2000). Antisera was used at 1/1,000 to 1/2,000 dilution and incubated with western blots overnight at room temperature. Polyclonal rabbit antisera were raised against the C-terminal peptides ARNALEHEYFKDIGYVP for CDKA and ALDHPYFDSLDKSQF for CDKB1. An additional antiserum raised against full length CDKB1 expressed in insect cells was also used. Specificity of antisera was confirmed by peptide competition or by using recombinant proteins. Immunoprecipitations and histone H1 protein kinase assays were as described in Cockcroft et al. (2000) using 2 to 3 µL of antiserum.