Tobacco retinoblastoma-related protein phosphorylated by a distinct cyclin-dependent kinase complex with Cdc2/cyclin D in vitro

Authors


*For correspondence (fax +81 0743 72 5469;
e-mail shinmyou@bs.aist-nara.ac.jp).

Summary

The retinoblastoma (Rb) protein was originally identified as a product of a tumour suppressor gene that plays a pivotal role in regulating both the cell cycle and differentiation in mammals. The growth-suppressive activity of Rb is regulated by phosphorylation with cyclin-dependent kinase (CDK), and inactivation of the Rb function is one of the critical steps for transition from the G1 to the S phase. We report here the cloning of a cDNA (NtRb1) from Nicotiana tabacum which encodes a Rb-related protein, and show that this gene is expressed in all the organs examined at the mRNA level. We have demonstrated that NtRb1 interacts with tobacco cyclin D by using yeast two-hybrid and in vitro binding assays. In mammals, cyclin D can assemble with CDK4 and CDK6, but not with Cdc2, to form active complexes. Surprisingly, tobacco cyclin D and Cdc2 proteins can form a complex in insect cells, which is able to phosphorylate tobacco Rb-related protein in vitro. Using immunoprecipitation with the anti-cyclin D antibody, cyclin D can be found in a complex with Cdc2 in suspension-cultured tobacco BY-2 cells. These results suggest that the cdc2 gene modulates the cell cycle through the phosphorylation of Rb-related protein by forming an active complex with cyclin D in plants.

Introduction

Progression through the eukaryotic cell cycle is regulated by distinct families of cyclin-dependent kinases (CDKs) whose activities are determined by the coordinated binding of different types of cyclins at each phase of the cell cycle. In mammals, the commitment to enter the cell cycle is governed during the G1 phase at a point called the ‘restriction (R) point’. The mammalian G1 cyclins consist of cyclin D, which associates with CDK4 and CDK6, and cyclin E, which associates with CDK2. Both types of cyclins sequentially phosphorylate the retinoblastoma (Rb) protein which inactivates its growth-suppressive activity ( Sherr 1994; Weinberg 1995). The growth-suppressive function of Rb is excerted by its binding to a variety of cellular proteins involved in DNA replication and control of the cell cycle ( Taya 1997). The ability of Rb to control the G1/S transition is mediated largely through its interactions with the E2F/DP transcription factor family. Hypophosphorylated Rb binds to the activation domain of E2F and actively represses its transactivation activity ( Hieber et al. 1992 ; Weintraub et al. 1992 ). At mid- to late G1, Rb becomes hyperphosphorylated by G1 cyclin-dependent kinases and is released from the promoter-bound E2F, allowing transcription of E2F-regulated genes.

Although substantial progress has been made in understanding the mechanisms that control the cell cycle in yeast and mammals, far less is known about cell cycle regulation in plants. Recently, cDNA clones for Rb-related protein were isolated from maize ( Grafi et al. 1996 ; Xie et al. 1996 ). Ach et al. (1997a) also reported that maize has two genes, RRB1 and RRB2, which encode Rb-related proteins. Although the maize Rb-related protein was shown to be phosphorylated during the course of endoreduplication in maize endosperm ( Grafi & Larkins 1995), it is not known which cyclin-dependent kinase can phosphorylate Rb-related protein in plants. The existence of cyclin D ( Dahl et al. 1995 ; Soni et al. 1995 ) and Rb-related genes ( Ach et al. 1997a ; Grafi et al. 1996 ; Xie et al. 1996 ) in plants suggests that the mechanisms of G1/S control in plants are more similar to those in mammals than those in yeast.

Genetic studies with yeasts have revealed a single CDK gene, cdc2 in Schizosaccharomyces pombe and cdc28 in Saccharomyces cerevisiae, required for both the G1/S and G2/M transitions ( Nasmyth 1993; Norbury & Nurse 1992). By contrast, the cell cycle is controlled by a family of CDKs in mammals ( Pines 1995). Plants contain a number of cdc2-related genes and several lines of evidence suggest that different cdc2-related genes are involved in different phases of the plant cell cycle as in mammals ( Burssens et al. 1998 ). However, which combination of Cdc2-related proteins forms active complexes with cyclins has not been elucidated ( Burssens et al. 1998 ; Magyar et al. 1993 ; Magyar & Meszaros 1997).

We have isolated a Rb gene, NtRb1, from tobacco. We have demonstrated that tobacco Cdc2 can form a complex with tobacco cyclin D in insect cells as well as in suspension cultured tobacco BY-2 cells, and that the complex expressed in insect cells can phosphorylate tobacco Rb-related protein in vitro. To our knowledge, this is the first evidence that a complex of Cdc2 with cyclin D can phosphorylate a Rb-related protein.

Results

Isolation of a tobacco cDNA encoding Rb-related protein

A partial cDNA of approximately 2.0 kb encoding a Rb-related protein was isolated by screening a cultured tobacco BY-2 cells cDNA library with the cDNA containing almost a full-length of maize Rb-related cDNA, ZmRb1 ( Xie et al. 1996 ). By screening a tobacco SR-1 shoot apex cDNA library with a partial cDNA, several clones were isolated, and one of the clones (NtRb1) contained an approximately 3.3 kb insert encoding a 961 amino acid peptide with a predicted molecular weight of approximately 107 kDa ( Fig. 1). Several in-frame stop codons were found at the 5′ non-translated region of the NtRb1 cDNA, indicating that this clone contains a full-length region.

Figure 1.

Nucleotide sequences and deduced amino acid sequences of the NtRb1 cDNA.

The A and B domains of the pocket region are boxed. The N-terminal leucine-rich domain conserved in the Rb family members from plants to mammalian cells is in italic and underlined. Potential CDK-phosphorylation sites (S/TP) are underlined.

Analysis of the deduced amino acid sequence of NtRb1 revealed that it contains the essential domain of homology with the A and B domains of the pocket region that is conserved among all the mammalian Rb family proteins ( Figs 1 and 2). The A domain of NtRb1 protein is approximately 31% identical and 66% identical in sequence to the A domains of human Rb protein and maize RRB1 protein, respectively. The B domain of NtRb1 protein is approximately 26% identical and 44% identical in sequence to the human Rb protein and maize RRB1 protein, respectively ( Fig. 2) ( Ach et al. 1997a ). The tobacco NtRb1 has the shortest B domain region among all the Rb family proteins ( Fig. 2). The critical cysteine residue at position 706 of human Rb ( Kaye et al. 1990 ) was present at comparable positions in both the tobacco and maize Rb-related proteins ( Fig. 2). NtRb1 also contains a leucine-rich domain in the N-terminal region which is conserved both in mammals ( Kaye et al. 1990 ) and in the maize RRB1 protein ( Fig. 2) ( Ach et al. 1997a ). There are 13 potential CDK phosphorylation sites in the tobacco NtRb1 which are highly clustered in the C-terminal region ( Fig. 1).

Figure 2.

Sequence alignment of domains A and B, and the N-terminal leucine-rich domain of NtRb1 with mammalian Rb family.

Black boxes indicate conserved amino acid residues, and dashes indicate gaps in the sequences. The bars in the A and B domains indicate the α-helix and β-sheet structures deduced from crystallization analysis ( Lee et al. 1998 ). The arrow in B domain denotes a conserved cysteine residue known to be critical for Rb function in mammalian cells ( Kaye et al. 1990 ).

NtRb1 is expressed in tobacco plants and cultured tobacco cells

The maize RRB genes are expressed in all the tissues, but the highest level of expression was seen in the shoot apex ( Ach et al. 1997a ). To examine NtRb1 expression in tobacco, poly(A) RNA from various organs was probed with a part of NtRb1. In all the organs examined, a transcript of approximately 3.3 kb was detected. High levels of NtRb1 expression were detected in stems, leaves, and roots, which consist predominantly of differentiated cells. Low levels of expression were found in flowers and in undifferentiated BY-2 cells in suspension culture ( Fig. 3). This result indicates that NtRb1 is definitely expressed in all the organs tested.

Figure 3.

Expression of NtRb1 in various tissues of tobacco plant.

Poly(A) RNAs (1 μg) from various tissues of tobacco plant were electrophoresed on an agarose–formaldehyde gel, and blotted onto a nylon membrane. The membrane was probed with the 32P-labelled NtRb1 cDNA. The same membrane was probed with actin cDNA to ensure equal loading of each sample.

NtRb1 can bind to tobacco cyclin D

The ability of the mammalian Rb protein family to bind a variety of viral and cellular proteins is conferred exclusively by the pocket region ( Weinberg 1991). The conservation of the pocket region in the NtRb1 protein suggests that it can bind these Rb-binding proteins. To test this possibility we used a yeast two-hybrid system to detect binding of NtRb1 to the tobacco cyclin A, Ntcyc27 ( Setiady et al. 1995 ), and the tobacco cyclin D, designated NtcycD3–1. A cDNA clone representing NtcycD3–1 was isolated from a cDNA library of suspension cultured tobacco BY-2 cells and was classified as the CycD3 class of cyclin D, based on its high sequence identity to Arabidopsis cyclinδ-3 ( Soni et al. 1995 ) and the CycD3 class of tobacco ( Sorrell et al. 1999 ; Sekine et al. unpublished results).

For two-hybrid assays, a plasmid that contains the part of the NtRb1 encoding amino acids 374–961 of NtRb1 was constructed with the Gal4 transactivation domain. Yeast strain HF7c was transformed with the plasmid together with the plasmids expressing tobacco cyclin D or tobacco cyclin A fused with the Gal4 DNA-binding domains. Although NtRb1 bound to tobacco cyclin D, tobacco cyclin A interacted very weakly with NtRb1 in this assay ( Fig. 4).

Figure 4.

Interaction of NtRb1 with NtcycD3–1 in yeast by the two hybrid system.

Plasmid that expresses the NtRb1 protein as a GAL4 DNA-binding domain fusion was transformed pairwise into HF7c yeast cells along with plasmids that express the GAL4 activation domain fusions of NtcycD3–1 and tobacco cyclin A, Ntcyc27. The NtRb1 fused with a Gal4 DNA-binding domain was also co-transformed into HF7c with pGBT9 vector. These transformants were streaked on glucose-containing plates with or without histidine.

We confirmed the ability of NtRb1 to bind to tobacco cyclin D in vitro using experiments with tagged proteins ( Fig. 5a). NtcycD3–1 was produced in baculovirus-infected insect cells and purified by an anti-FLAG M2 affinity gel. SDS–PAGE and staining with the anti-FLAG M2 monoclonal antibody showed that the product consisted of a major polypeptide of the size expected for FLAG–NtcycD3–1. Lysates containing equal amounts of GST–NtRb1 were mixed with lysates of vector-control infected cells and with lysates of cells infected with the FLAG–NtcycD3–1 vector. Proteins bound to an anti-FLAG M2 affinity gel were eluted, separated by SDS–PAGE, and analysed by immunoblotting with the anti-NtRb1 antibody (codons 654–671). NtRb1 immuno-reacted peptides of the size expected for GST–NtRb1 were only detected in the eluates obtained with lysates containing FLAG–NtcycD3–1 ( Fig. 5a, compare lanes 2–4). A reciprocal binding assay revealed that protein of the size expected for FLAG–NtcycD3–1 was only detected in lysates containing GST–NtRb1 (data not shown). Taken together, these results provide further support for the physical interaction between NtRb1 and a tobacco cyclin D.

Figure 5.

A complex of Cdc2/cyclin D can phosphorylate Rb-related protein in vitro.

(a) NtRb1 can bind to tobacco cyclin D. Lysates of insect Sf9 cells infected with a wild-type baculovirus (Ve) (lanes 1 and 2) and a vector encoding FLAG–NtcycD3–1 (D) (lanes 3 and 4) were incubated with GST–NtRb1 (Rb) for 1 h at 4°C on a rotating stirrer. After incubation, mixtures were bound to an anti-FLAG M2 affinity gel. Elute fractions (lane E) and crude extracts (lane C) were separated on a 10% polyacrylamide gel containing SDS. GST–NtRb1 was detected by immunoblotting with anti-NtRb1 antibody (upper column) and FLAG–NtcycD3–1 was detected by immunoblotting with the monoclonal antibody of anti-FLAG M2 (lower column).

(b) Tobacco Cdc2 can form a complex with cyclin D. Lysates of insect Sf9 cells co-infected with baculovirus vectors that express His–Cdc2Nt1 (2) (lanes 1, 2, 5 and 6) or His–Cdc2Nt1 and FLAG–NtcycD3–1 (D) (lanes 3, 4, 7 and 8) were purified by an anti-FLAG M2 affinity gel. Elute fractions (lane E) and crude extracts (lane C) were separated on 10% polyacrylamide gels containing SDS. FLAG–NtcycD3–1 was detected by immunoblotting with the monoclonal antibody of anti-FLAG M2 (lanes 1–4), and His–Cdc2Nt1 was detected by immunoblotting with the antibody for His tag (lanes 5–8). The positions of pre-stained molecular mass protein standards are indicated on the left.

(c) Rb kinase activity of Cdc2/cyclin D complex. Purified fractions, prepared from the lysates of insect cells which express the indicated proteins (Ve, vector; 2, His–Cdc2Nt1; D, FLAG–NtcycD3–1) with the anti-FLAG M2 affinity gel (F) or the TALON metal affinity resin (T), were incubated with a bacterial GST–NtRb1 for 10 min at 30°C in a kinase buffer containing [γ-32P]ATP.

(d) Association of NtcycD3–1 with Cdc2 in tobacco BY-2 cells. Total proteins extracted from tobacco BY-2 cells were treated with the indicated immunoaffinity beads (D, NtcycD3–1; Rb, NtRb1; 25, Ntcyc25; IgG, normal rabbit IgG; P, PSTAIRE). Proteins bound to the beads were eluted and separated on a 10% polyacrylamide gel containing SDS. Cdc2 was detected by immunoblotting with the anti-PSTAIRE antibody. Dots on the right indicate the locations of the three polypeptides recognized by the anti-PSTAIRE antibody.

Complex of Cdc2/cyclin D can phosphorylate the Rb-related protein in vitro

To determine whether Cdc2 and cyclin D can form a complex capable of phosphorylating NtRb1, we co-expressed a His-tagged tobacco Cdc2, cdc2Nt1 ( Setiady et al. 1996 ), together with FLAG–NtcycD3–1 using a baculovirus system. Proteins bound to an anti-FLAG M2 affinity gel were eluted, separated by SDS–PAGE, and analysed by immunoblotting with antisera directed against His tag ( Fig. 5b, lanes 5–8). Although His–Cdc2Nt1 was equally expressed in insect cells that were or were not expressing FLAG–NtcycD3–1, protein of the size expected for His–Cdc2Nt1 was only detected in the eluates obtained with lysates containing both His–Cdc2Nt1 and FLAG–NtcycD3–1 ( Fig. 5b, compare lanes 6–8). These results provide strong evidence that the tobacco cyclin D can specifically interact with the tobacco Cdc2 in vitro.

We tested whether the tobacco Cdc2/cyclin D complex exhibits Rb kinase activity. Phosphorylation of the GST–NtRb1 was only detected with lysate prepared from insect cells expressing both tobacco Cdc2 and cyclin D, and lysate prepared from insect cells expressing tobacco Cdc2 and cyclin D, and a vector alone had no kinase activity (data not shown). To exclude the possibility that tobacco Cdc2 and/or cyclin D would activate the endogenous CDK/cyclin complex from insect cells, the tobacco Cdc2/cyclin D complex was purified with the anti-FLAG M2 affinity gel or the TALON metal affinity resin. Phosphorylation of the GST–NtRb1 was only detected with the purified tobacco Cdc2/cyclin D complex in both purification procedures and the purified Cdc2 and cyclin D alone exhibited no kinase activity to the GST–NtRb1 ( Fig. 5c).

Cyclin D can be found in a complex with Cdc2 in tobacco BY-2 cells

To confirm that cyclin D binds with Cdc2 in vivo, we immunoprecipitated extracts from tobacco BY-2 cells with anti-cyclin D, anti-cyclin A, anti-NtRb1 and control antibodies. Ntcyc25 ( Setiady et al. 1995 ) was used for the tobacco cyclin A, and Cdc2 was detected by the anti-PSTAIRE antibody which recognizes the conserved ‘PSTAIRE’ motif of Cdc2 protein. The eluates from the beads bound with the anti-NtcycD3–1 antibody, anti-Ntcyc25 antibody, anti-NtRb1 antibody, normal rabbit IgG and anti-PSTAIRE antibody were immunoblotted with the anti-PSTAIRE antibody ( Fig. 5d). In the eluate from the beads bound with anti-PSTAIRE antibody, the anti-PSTAIRE antibody recognized three polypeptides in which the major band coincided with a molecular mass of approximately 34 kDa and two bands migrated more slowly. In the eluates from the beads bound with both anti-cyclin D and anti-cyclin A antibodies, the anti-PSTAIRE antibody cross-reacted with a polypeptide which has a molecular mass of approximately 34 kDa. However, a polypeptide cross-reacted with the anti-PSTAIRE antibody was not detected in the eluates from the beads bound with anti-NtRb1 antibody and normal rabbit IgG used as controls. This result indicated that Cdc2 can be found in a complex with both cyclin D and cyclin A in tobacco BY-2 cells.

Discussion

The first genes encoding plant Rb-related proteins were reported for a monocotyledonous plant, maize ( Ach et al. 1997a ; Grafi et al. 1996 ; Xie et al. 1996 ). We have cloned the first Rb gene for a dicotyledonous species, tobacco, and show that the Rb-related protein encoded by this gene can bind tobacco cyclin D and is a target for phosphorylation in vitro by a tobacco Cdc2/cyclin D complex.

The conserved A and B pocket domains and the leucine-rich domain of the N-terminal region present in the mammalian Rb protein family are also present in the maize and tobacco Rb-related proteins, suggesting that the plant and mammalian Rb proteins might have similar functions ( Fig. 2) ( Ach et al. 1997a ). Viral oncoproteins have been identified that bind the pocket region of Rb family proteins, which disrupts the interaction of these oncoproteins with Rb ( Chellappan et al. 1992 ; Zamanian & LaThangue 1992). These oncoproteins bind to Rb via a conserved motif, LXCXE, which is also found in cyclin D. Therefore, cyclin D can physically interact with Rb via the LXCXE motif ( Dowdy et al. 1993 ; Kato et al. 1993 ). We have demonstrated that the tobacco Rb-related protein binds a tobacco cyclin D by using a yeast two-hybrid system ( Fig. 4) and in vitro binding assay ( Fig. 5). This conserved physical interaction between tobacco Rb-related protein and cyclin D suggests that the pocket region of tobacco Rb-related protein may be involved in binding to a variety of viral and cellular proteins in plants. In fact, several viral proteins and Msi1-like proteins containing WD-40 repeats have been shown to bind maize Rb-related proteins ( Ach et al. 1997a ; Ach et al. 1997b ; Grafi et al. 1996 ; Xie et al. 1996 ).

Within the B pocket domain, the C706 in human Rb has been shown to be critical for Rb protein function ( Kaye et al. 1990 ). Although the A and B pocket domains are also conserved in the tobacco NtRb1 protein with the conserved C residue in the B domain, the length of the B domain is the shortest among all the identified Rb protein family ( Fig. 2). The crystal structure of the Rb pocket bound to the human papillomavirus HPV-16 E7 peptide containing the LXCXE motif illustrated both that the LXCXE peptide binds a highly conserved groove on the B pocket domain and that the A pocket domain is required for stable folding of the B pocket domain ( Lee et al. 1998 ). The A and B pocket domains each contain the cyclin-fold, a five-helix structural motif which has been found recently in the structures of cyclins and the basal transcription factor TFIIB ( Lee et al. 1998 ; Noble et al. 1997 ). However, we have demonstrated that NtRb1, which does not contain the part of the cyclin-fold structure in the B pocket domain corresponding to the α15 to β1 region in human Rb ( Fig. 2; Lee et al. 1998 ), can still bind cyclin D ( Figs 4 and 5). This result suggests that not all of the five-helix structure motifs in the B pocket domain are essential for the binding of Rb to proteins containing the LXCXE motif.

The Rb gene is expressed in all the tissues examined in mammals ( Bernards & Schackleford 1989). The maize RRB genes are also expressed in all the tissues, but the highest level of expression was detected in the shoot apex ( Ach et al. 1997a ). We showed that NtRb1 is expressed at the mRNA level in all the organs tested ( Fig. 3). Although high levels of NtRb1 expression were detected in stems, leaves and roots, which consist predominantly of differentiated cells, low levels of expression were found in flowers and in undifferentiated BY-2 cells ( Fig. 3). It is not known why NtRb1 mRNA levels are relatively high in these organs, but it is possible that NtRb1 could be involved in inducing and/or maintaining a differentiation state in plants.

In plants, a number of cdc2-related genes have been identified from several species such as maize ( Colasanti et al. 1993 ), rice ( Hashimoto et al. 1992 ), alfalfa ( Magyar et al. 1993, 1997 ), soybean ( Miao et al. 1993 ), Arabidopsis ( Hirayama et al. 1991 ) and pea ( Feiler & Jacobs 1990). The plant cdc2-related genes are categorized according to the similarity of the conserved ‘PSTAIRE’ motif (indicated by the single letter amino acid in the central region of the motif). Some of the plant cdc2-related genes contain an altered PSTAIRE motif, and several show different expression during cell cycle ( Burssens et al. 1998 ; Fobert et al. 1996 ; Hirt et al. 1993 ; Magyar et al. 1993 ; Magyar et al. 1997 ; Segers et al. 1996 ). We have previously isolated a cDNA clone (cdc2Nt1) that encodes a PSTAIRE-type of cdc2 gene from tobacco ( Setiady et al. 1996 ). We have demonstrated that the cdc2Nt1 gene was able to complement at both the G1/S (cdc28–4 and cdc28–13) and G2/M (cdc28–1N) transitions of yeast cdc28 mutants. The cdc2Nt1 gene is expressed constitutively throughout the cell cycle ( Setiady et al. 1996 ).

In mammals, cyclin D can associate with CDK4 and CDK6, but not with Cdc2, to form active complexes ( Dowdy et al. 1993 ; Kato et al. 1993 ). To assay whether tobacco Cdc2 is bound to tobacco cyclin D, we expressed Cdc2Nt1 and NtcycD3–1 in insect cells ( Fig. 5b). Surprisingly, Cdc2Nt1 can bind with NtcycD3–1 in vitro ( Fig. 5b). Sorrell et al. (1999) recently isolated three tobacco cyclin D clones and found that two belong to the CycD3 class and the third to the CycD2 class based on sequence criteria. NtcycD3–1 belongs to the CycD3 class but is not identical with their clone. This suggests that there are at least three distinct CycD3 cDNAs in tobacco.

We have tested the possibility whether or not Cdc2Nt1/NtcycD3–1 complex exhibits Rb kinase activity. The tobacco Cdc2/cyclin D complex purified from insect Sf9 cells co-infected with baculovirus vectors encoding cyclin D and Cdc2 indeed exhibited high level of a protein kinase activity that phosphorylated a GST–NtRb1 fusion protein in vitro. However, the purified cyclin D and Cdc2 alone did not exhibit Rb kinase activity at all ( Fig. 5c). This indicates that the phosphorylation of tobacco Rb protein by the tobacco Cdc2/cyclin D complex is not due to activation of endogenous proteins such as CDK and/or cyclins from insect cells. We also have evidence that lysate of the mammalian cyclin D/CDK4 complex produced by a baculovirus system did not exhibit the phosphorylation of the tobacco Rb protein, while the mammalian cyclin D/CDK4 complex can phosphorylate mammalian Rb protein under our assay conditions (data not shown). These results indicate that the tobacco Cdc2/cyclin D complex can truly phosphorylate the tobacco Rb-related protein in vitro.

To confirm that the tobacco cyclin D can bind with Cdc2 in vivo, immunoprecipitations were performed with cyclin D, cyclin A, NtRb1 and control antibodies. As shown in Fig. 5(d), a polypeptide cross-reacted with the anti-PSTAIRE antibody in the eluates from the beads bound with the anti-cyclin D and anti-cyclin A antibodies. However, the cross-reacted polypeptide could not be observed in the eluates from the beads bound with the anti-NtRb1 antibody and normal rabbit IgG. This indicates that Cdc2 can associate physically with cyclin D and cyclin A in tobacco BY-2 cells. We found that a polypeptide, cross-reacted with the anti-PSTAIRE antibody, coinciding with a molecular mass of approximately 34 kDa was detected in the immunoprecipitate with the anti-cyclin D antibody, while the anti-PSTAIRE antibody cross-reacted mainly with three polypeptides in tobacco BY-2 cells. It has been reported that the anti-PSTAIRE antibody recognizes at least two polypeptides in plants ( Feiler & Jacobs 1990; Magyar et al. 1993 ). Two slower-migrating bands were reduced when crude extracts were prepared from tobacco BY-2 cells without phosphatase inhibitors (data not shown). This suggests that these bands may represent distinct phosphrylation states of the polypeptide with a molecular mass of approximately 34 kDa. However, it remains to be concluded whether phosphorylation states of Cdc2 protein affect the binding with cyclin D. Cdc2Nt1 expressed in insect cells can be cross-reacted with the anti-PSTAIRE antibody (data not shown). We concluded that a single band cross-reacted with the anti-PSTAIRE antibody includes Cdc2Nt1, while we could not rule out the possibility that the band contains other Cdc2-related proteins. We observed here that the anti-tobacco cyclin A antibody also cross-reacted with Cdc2 which can be recognized with the anti-PSTAIRE antibody. This observation was supported by the finding that immunoprecipitation with anti-human cyclin A antibodies allowed the detection of Cdc2-related proteins cross-reacted with the anti-PSTAIRE antibody ( Magyar et al. 1993 ). Unfortunately, phosphorylation of the tobacco Rb-related protein has not been detected in the immunoprecipitated proteins with the anti-cyclin D antibody. It is possible that the tobacco Rb-related protein may not be phosphorylated by the Cdc2/cyclin D complex in vivo. Rather than considering this possibility, we believe that the activity of the Cdc2/cyclin D complex in vivo could not detected due to a very small amount of immunoprecititated proteins with the anti-cyclin D antibody. Further study will reveal whether the Cdc2/cyclin D complex can phosphorylate the tobacco Rb-related protein in vivo.

Our most important finding was that tobacco Cdc2 and cyclin D can form a complex both in vivo and in vitro, and the complex expressed in vitro phosphorylates Rb-related protein. Further study is required to determine whether the partner of the cyclin D is only Cdc2 in plants. Our result strongly suggests that the cdc2 gene modulates the cell cycle through the phosphorylation of Rb-related protein by forming an active complex with cyclin D in plants. Future work should elucidate other combinations of Cdc2-related proteins forming active complexes with various types of cyclins, especially in vivo, and the target proteins in which the Cdc2/cyclin complexes might participate.

Experimental procedures

Plant materials

Nicotiana tabacum L.cv.SR-1 was grown in a greenhouse. Tobacco BY-2 cells (N.tabacum L.cv. Bright Yellow-2) were cultured at 27°C in a modified Linsmaier and Skoog medium as described previously ( Setiady et al. 1995 ).

cDNA library construction and isolation of Rb cDNAs

A tobacco BY-2 cDNA library was constructed with a Uni-ZAP XR kit (Strategene) from RNA prepared from exponentially growing cells in suspension culture. A tobacco SR-1 shoot apex cDNA library was constructed with λ ZipLox kit (Gibco BRL) from RNA prepared from the shoot tip of mature plants. pGAD424Rb1 (kindly provided by Crisanto Gutierrez, CSIC-UAM, Spain) which contains ZmRb1 ( Xie et al. 1996 ) was digested with EcoRI and BamHI. The resulting 2.0 kb fragment was gel-purified by a Prep A gene kit (BioRad) and labelled with α-32P-dCTP by random priming (BcaBest labelling kit, Takara). This probe was used to screen a cultured tobacco BY-2 cell cDNA library as described previously ( Setiady et al. 1995 ). One positive clone was plaque-purified, and phage DNA was excised in vivo to recover pBluescript SK(–) plasmid according to Strategene’s protocol. The tobacco SR-1 shoot apex cDNA library was subsequently screened by a 32P-labelled tobacco cDNA clone. Positive clones were recovered by in vivo excision according to the manufacturer’s manual.

RNA blot analysis

Total RNA was isolated from various tobacco SR-1 organs as previously described ( Setiady et al. 1995 ) and poly(A) RNAs were purified by Oligotex-dT30 kit (Takara) according to the instruction manual. Samples of 1 μg were electrophoresed on an 1.0% agarose–formaldehyde gel, blotted on a nylon membrane (Hybond-N+, Amersham), and hybridized with a 32P-labelled 2.0 kbp NtRb1 probe (1288–3296 bp in Fig. 1) containing the A and B pocket regions. After autoradiography, the membrane was stripped and reprobed with a rice actin gene probe ( Setiady et al. 1995 ).

Yeast two-hybrid assays

The cDNA fragments encoding cyclin D (NtcycD3–1) and cyclin A (Ntcyc27) ( Setiady et al. 1995 ) were subcloned into the two-hybrid vector pGBT9, which produced fusion proteins with the Gal4 DNA-binding domain. A fragment encoding amino acids 374–961 of NtRb1 was fused with the Gal4 transactivation domain to construct a fusion protein in the plasmid pGAD424 as follows. An in-frame BamHI site was introduced upstream of the coding sequence by PCR with the forward primer 5′-GGATCCTTGCAATGGCTTCCCCAGC-3′ and the reverse primer 5′-GTCGACTAAGACTCAGGCTGCTCAG T-3′ to introduce a SalI site just after the stop codon of NtRb1. The PCR product was digested with BamHI and SalI and ligated to pGAD424. Yeast transformations were performed with S. cerevisiae strain HF7c as described previously ( Setiady et al. 1995 ).

Preparation of GST-NtRb1 fusion proteins

The pGEX-5X-2 plasmid was constructed to express a fusion protein consisting of amino acids 374–961 encoded by NtRb1 fused to glutathione S-transferase (GST). The recombinant plasmid pGAD424 (NtRb1) constructed for yeast two-hybrid assays was digested with BamHI and SalI and the insert fragment was subcloned into the BamHI and SalI sites of pGEX-5X-2. The recombinant plasmid pGEX-5X-2 (NtRb1) was introduced into the Escherichia coli strain BL21 (DE3) pLysS, and the transformant was grown to an OD600 of 0.4 in LB broth containing ampicillin and chloramphenicol at 37°C. Expression of GST–NtRb1 fusion proteins was induced with 0.25 m m isopropyl-β- d-thiogalactopyranoside (IPTG) at 18°C for 16 h. Cells from a 100 ml culture were lysed by sonication on ice in 5 ml of kinase buffer [50 m m Tris–HCl at pH 7.5, 10 m m MgCl2, 1 m m EGTA, 1 m m phenylmethylsulphonyl fluoride (PMSF), 1 m m dithiothreitol (DTT), 10 m m NaF, 25 m mβ-glycerophosphate, 2 m m sodium orthovanadate] and then cleared by centrifugation. Glutathion–Sepharose 4B beads (Pharmacia) were washed three times with kinase buffer and suspended in an equal volume of the same buffer. Washed beads (200 μl) were added to 1 ml bacterial supernatant and incubated for 1 h at 4°C on a rotating stirrer. The beads were then washed three times with kinase buffer and suspended in an equal buffer volume.

Preparation of antibody

Polyclonal antibodies raised against the N-terminal MVELNNCSNSEENGC peptide of NtRb1, the NLAPNGQIGDIRSPKKVC peptide corresponding to codons 654–671 of NtRb1, as well as against the N-terminal MGIQHNEHNQDQT peptide of NtcycD3–1 and the N-terminal MATTQNRRSSVSSA peptide of Ntcyc25 ( Setiady et al. 1995 ) were immunized in rabbits by using synthetic peptides coupled to keyhole limpet haemocyanin through an additional cysteine residue at their C-termini. Antibodies (MVELNNCSNSEENGC, MGIQHNEHNQDQT and MATTQNRRSSVSSA) were purified by affinity chromatography in which each peptide was coupled to 2-fluoro-1-methyl-pyridinium toluene-4-sulphonate (FMP)-activated cellulofine (Seikagaku Kougyou Co.).

Insect cell culture and baculovirus infection

Spodoptera frugiperda (Sf9) cells were maintained at 27°C in Grace’s insect medium containing 10% fetal bovine serum (FBS) and gentamicin in 100 ml spinner bottles. Virus infection was performed in 60 mm diameter dishes. The FLAG tag was fused to the N-terminal end of the entire coding region of NtcycD3–1 and inserted into the transfer vector pVL1392 (Pharmingen) as follows. To attach a FLAG sequence, an in-frame HindIII site was introduced upstream of the coding sequence by PCR with the forward primer 5′-AAGCTTATGGGAATACAACACAATGA-3′ and the reverse primer 5′-TCTAGATTAGCGAGGGCTGCCAAC-3′ to introduce the XbaI site just after the stop codon of NtcycD3–1 and inserted into the HindIII and XbaI sites of pFLAG-1 (Eastman Kodak). To subclone the FLAG-tagged fragment of NtcycD3–1 into pVL1392, an in-frame BglII site was introduced upstream of the coding sequence of pFLAG-1 (NtcycD3–1) by PCR with the forward primer 5′-AGATCTATGGACTACAAGGATGACGATG-3′ and the reverse primer as used above for introducing the XbaI site of pFLAG-1 (NtcycD3–1). pVL1392 (FLAG–NtcycD3–1) was co-transfected into Sf9 cells with linearized BaculoGoldTM DNA (Pharmingen) using a liposome-mediated transfection kit (Gibco BRL).

The cDNA fragment containing the entire coding region of cdc2Nt1 ( Setiady et al. 1996 ) was inserted into the pFastBac HTa plasmid (Gibco BRL) to construct a pFastBac HTa (His–cdc2Nt1) consisting of a His tag sequence fused to the N-terminal end of the Cdc2Nt1. An in-frame BglII site was introduced upstream of the coding sequence of cdc2Nt1 by PCR with the forward primer 5′-AGATCTGGATGGACCAGTATGAAAAAGT-3′ and the reverse primer 5′-GTCGACTCACGGAACATACCCAAT-3′ to introduce a SalI site. The resulting fragment was inserted into the BamHI and SalI sites of pFastBac HTa. This plasmid was transformed into E. coli strain DH5aBac (Gibco BRL) for transposition into the bacmid. The recombinant bacmid was isolated and transfected into Sf9 cells using a liposome-mediated transfection kit (Gibco BRL). Recombinant viruses were assayed for expression of their encoded proteins by immunoblotting. All sequences generated by PCR were verified by sequencing.

In vitro binding assay

Lysates of E. coli expressing GST–NtRb1 and GST were incubated with lysates of insect cells which were or were not expressing FLAG–NtcycD3–1 at 4°C for 1 h on a rotating stirrer. After incubation, the mixtures were purified by the anti-FLAG M2 affinity gel (Eastman Kodak), denatured in gel sample buffer, and separated on 10% polyacrylamide gels containing SDS. NtRb1 was then detected by immunoblotting, using the antibody against NtRb1 (codons 654–671) with alkaline phosphatase for detection.

Detection of NtcycD3–1/Cdc2Nt1 complex formation

FLAG–NtcycD3–1 and His–Cdc2Nt1 were co-expressed in Sf9 cells, and the lysate was purified by the anti-FLAG M2 affinity gel. Elute fraction was denatured in the gel sample buffer, separated on 10% polyacrylamide gels containing SDS, and then Cdc2Nt1 was detected by immunoblotting using the antibody for His tag (Santa Cruz).

In vitro kinase assay

At 72 h post-infection, infected Sf9 cells were lysed, and then FLAG-NtcycD3–1 and His–Cdc2Nt1 were purified by the anti-FLAG M2 affinity gel and the TALON metal affinity resin (Clontech), respectively. His–Cdc2Nt1/FLAG–NtcycD3–1 complex was purified from lysates of insect cells expressing both proteins using the anti-FLAG M2 affinity gel or the TALON metal affinity resin. Elute fractions which were prepared in the same way from the lysates of insect cells infected with a wild-type baculovirus were used as negative controls. Purified fractions were mixed with a bacterial GST–NtRb1 which were immobilized on glutathione–Sepharose 4B beads. Kinase reactions were initiated at 30°C by adding 10 μCi of [γ-32P]ATP (4500 Ci mmol–1, ICN) adjusted with unlabeled ATP to a final concentration of 50 μm. After incubation for 10 min, the beads were washed twice with cold kinase buffer, and the GST–NtRb1 eluted and resolved on denaturing polyacrylamide gels. The phosphorylated GST–NtRb1 was detected by autoradiography.

Protein extraction and immunoprecipitation

Tobacco BY-2 cells (4 days after subculture) were lysed by sonication on ice in extraction buffer [25 m m Tris–HCl at pH 7.6, 75 m m NaCl, 15 m m MgCl2, 15 m m EGTA, 0.1% NP-40, 1 m m phenylmethylsulphonyl fluoride (PMSF), 10 μg ml–1 leupeptin, 50 μg ml–1N-tosyl- l-phenylalanine chloromethyl ketone (TPCK), 5 μg ml–1 pepstatin A, 10 μg ml–1 aprotinin, 5 μg ml–1 antipain, 10 μg ml–1 trypsin inhibitor from soybean, 0.1 m m benzamidine, 10 m m NaF, 25 m mβ-glycerophosphate, 2 m m sodium orthovanadate] and then cleared by centrifugation. Protein concentrations were determined by use of the Protein Assay CBB solution (Nacalai Tesque Inc.) with bovine serum albumin as the standard.

The N-terminal anti-NtRb1 antibody, anti-NtcycD3–1 antibody, anti-Ntcyc25 anitibody, anti-PSTAIRE antibody (Santa Cruz) and normal rabbit IgG (Santa Cruz) were cross-linked to protein A Sepharose 4FF beads (Pharmacia) in 0.2 m H3BO3 buffer (pH 9.0) with 20 m m dimethylpimelimidate (DMP). Total protein (500 μg) in extraction buffer was pre-cleared with 10 μl of protein A Sepharose 4FF beads for 30 min at 4°C. After centrifugation (15 000×g,30 min, 4°C), supernatants were transferred into the microtubes containing 10 μl each of the immunoaffinity beads and incubated for 2 h at 4°C. Beads were washed three times with extraction buffer and once with 150 m m NaCl. The proteins were eluted from the beads with 10 μl of 0.1 m glycine–HCl (pH 2.5) and immediately neutralized with 0.5 μl of 1 m Tris. Eluants were denatured in the gel sample buffer, separated on 10% polyacrylamide gels containing SDS, and then Cdc2 was detected by immunoblotting with the anti-PSTAIRE antibody.

Acknowledgements

The authors wish to thank Drs Ko Kato, Jun-ya Kato, Hiroshi Kouchi (National Institute of Agrobiological Resources) and Kazuya Yoshida for helpful discussions and suggestions throughout this work. We are grateful to Dr Frederik Meins Jr (Friedrich Miescher Institut) for criticaly reading the manuscript and Dr Dee Worman for editing the manuscript. This research was supported by a Grant-in Aid for Scientific Research (Nos 09640773 and 10182217) from the Ministry of Education, Science and Culture, Japan.

GenBank database accession numbers AB015221 (NtRb1) and AB015222 (NtcycD3–1).

Ancillary