AtSGP1, AtSGP2 and MAP4Kα are nucleolar plant proteins that can complement fission yeast mutants lacking a functional SIN pathway

Cytokinesis is the final event of the cell cycle that divides the mother cell into two daughter cells, after the completion of mitosis. In contrast to animal cytokinesis, where a contractile ring tightens the cytoplasm, plant somatic cytokinesis is characterised by the formation of a transient membrane compartment called the cell plate, which matures into a new cell wall flanked by cytoplasmic membranes. The cell plate is built de novo by material mobilised from Golgi-derived vesicles and delivered by a dynamic cytokinesis-specific cytoskeletal array, the phragmoplast. The cell plate is generally initiated in the centre of the cell division plane, defined by a preprophase band (PPB) of microtubules, and fuses with the parental plasma membrane (Staehelin and Hepler, 1996).

Several studies demonstrate that mitogen-activated protein kinases (MAPK) are involved in the formation of the cell plate. Two orthologous MAPK from tobacco and alfalfa (NtNTF6 and MsMMK3), as well as the tobacco MAP kinase kinase kinase (MAP3K) NtNPK1 have been shown to be activated in late mitosis. Furthermore, these proteins were shown to localise to the midplane during late anaphase and to the phragmoplast (Bögre et al., 1999; Calderini et al., 1998; Nishihama et al., 2001). Overexpression of a dominant negative form of NtNPK1 in tobacco cells inhibits expansion of both the phragmoplast and the cell plate, resulting in multinucleate cells with incomplete cell plates (Nishihama et al., 2001). Krysan et al. have shown that the three A. thaliana orthologues of NtNPK1 (ANP1, ANP2, ANP3) are also involved in the formation of the cell plate (Krysan et al., 2002).

Despite recent advances in our understanding of the formation of the cell plate in plant cells, little is known about the mechanism(s) responsible for initiating cytokinesis and coordinating its timing with mitosis. The phenotype of titan and pilz mutants, affected in genes encoding tubulin-folding cofactors (Liu and Meinke, 1998; Steinborn et al., 2002), illustrate that spatial and temporal coordination between mitosis and cytokinesis is mediated in part by cytoskeletal components. The co-localisation of the cdc2 kinase and three types of microtubular structures, namely the PPB, the spindle, and the cell plate, suggests that the latter kinase also plays an important role in this coordination (Weingartner et al., 2001).

The need shared by all eucaryotes to coordinate mitosis and cytokinesis, in order to ensure an even distribution of chromosomes over the daughter cells, suggests that mechanisms promoting communication between the regulation of the cell cycle and the organisation of the cytoskeleton probably involve some common features across kingdoms. The mechanisms that regulate the onset of mitosis are relatively well understood in contrast to the mechanisms ensuring the control of signal transduction during late mitosis and cytokinesis (Heese et al., 1998). Mitotic progression relies on the activity of the cyclinB-CDK complex during M phase. Before anaphase, the latter complex activates the anaphase promoting complex/cyclosome (APC/C), and promotes exit from mitosis by inactivating the cyclinB-cdc2 complex itself. Several studies in Saccharomyces cerevisiae and Schizosaccharomyces pombe support the idea that the latter mechanism is not sufficient for the initiation of cytokinesis, but rather, that an additional protein kinase cascade triggering a cytokinetic pathway is required (Nigg, 2001). Indeed, genetic and molecular analyses in S. pombe provided evidence for the role of the septum initiation network (SIN) involving numerous protein kinases (Bardin and Amon, 2001; McCollum and Gould, 2001) (see Fig. 1). Localisation studies have revealed that the SIN cascade functions by transducing septum-promoting signal(s) from a single spindle pole body (SPB) to the division site. Upon entry into mitosis, a yeast monomeric G protein, named spg1p, initially activated on both SPB, is inactivated on one SPB at anaphase onset, by the byr4p and cdc16p proteins, which form a two-component GTPase activating protein (GAP). This allows spg1p to recruit cdc7p, a MAP3K-related protein, to a single SPB. Cdc7p determines the localisation of two other SIN components to the SPB, i.e. sid1p, a MAP kinase kinase kinase kinase-related (MAP4K) protein, and cdc14p. The sid1p-cdc14p complex then promotes activation of the SPB-localised sid2p-mob1p protein kinase complex, which translocates to the cell-division site to trigger initiation of septum formation. Coordination between mitotic exit and septum assembly occurs through the action of the APC, which degrades B-type cyclins and possibly other proteins, resulting in cdc2p inactivation, necessary for proper SPB localisation of SIN proteins (Bardin and Amon, 2001).

Jouannic et al. have previously reported cdc7p-related MAP3Kϵ1 proteins from Brassica napus and Arabidopsis thaliana which can complement an S. pombe cdc7 mutant, suggesting their involvement in cell division in plants (Jouannic et al., 2001). We report on the identification and characterisation of new SIN-related components in plants (Fig. 1). The AtSGP1/2 and BnMAP4Kα2 plant proteins are able to complement cytokinesis deficiency in fission yeast lacking spg1p or sid1p, respectively. Moreover, in the yeast two-hybrid system, interaction was detected between the AtSGP1/2 proteins and the fission yeast SIN pathway proteins cdc7p and byr4p, and between the plant proteins AtSGP1 and BnMAP3Kϵ1. Localisation studies using green fluorescent protein (GFP) fusion proteins suggest that the plant SIN-related proteins are nucleolar in higher plants. Using real-time PCR and a synchronised cell suspension cultures of Arabidopsis, we demonstrated that the plant SIN-related genes are co-expressed in a cell-cycle dependent manner, with a higher accumulation at G₂-M. The data are compatible with a role of the plant SGP1/2, MAP3Kϵ1/2 and MAP4Kα2 proteins in a SIN-related network regulating cell division in plants.

AtSGP1, AtSGP2, AtMAP4Kα1 and AtMAP4Kα2 coding regions were amplified from an A. thaliana root cDNA library by PCR using gene-specific primers flanking the corresponding sequences according to genomic DNA annotations: AtSGP1 5′-TCTCTGTTTCATCAATGGCCG-3′, 5′-AATTAGGAATCGTAGTCG-3′; AtSGP2 5′-TTCTCAATGGCTCAATCGTG-3′, 5′-CATATTCTAGAAGTCGATGA-3′; AtMAP4Kα1 5′-GTTGTGGAAATGGATG-3′, 5′-TCTGTGAGTTTCTCAAC-3′; AtMAP4Kα2 5′-TAGTGGTTGGAGCGGGAGAGAAGGTGG-3′, 5′-ATTCCATCTCTGTAGGTTGTCTCAGGG-3′. These PCR fragments were cloned into the pBluescript SK vector by blunt-end ligation, and their identities were confirmed by DNA sequencing.

Arabidopsis thaliana ecotype Columbia seeds were sown in soil in a growth chamber (17°C, 65% humidity, light 250 νE/m²/second) with a 16 hour photoperiod. Plantlets were harvested 17 days after sowing, at the four leaf stage. Roots, rosette leaves, cauline leaves, inflorescence stems, flower buds, mature flowers and 0.5 cm long green siliques were harvested from a pool of 50 plants, 60 days after sowing. While harvesting the primary inflorescence stems, stem nodes were discarded. The A. thaliana cell suspension culture was synchronised using aphidicoline, according to the method of Doutriaux et al. (Doutriaux et al., 1998). The T0 reference point corresponds to a sample collected immediately after aphidicoline release. After removal of the drug, an aliquot of the cell suspension culture was collected every 2 hours and frozen in liquid nitrogen.

RNA extractions, cDNA preparations and real-time quantitative RTPCR analysis were performed according to the method of Charrier et al. (Charrier et al., 2002). PCR amplification was performed using specific oligonucleotides (Table 1). The oligonucleotides used for the amplification of AtMAP3Kϵ1, AtMAP3Kϵ2, AtMAP4Kα1, AtMAP4Kα2 and ACTIN 2/8 have been described previously (Charrier et al., 2002).

DNA fragments corresponding to the full length AtSGP1, AtSGP2 and BnMAP4Kα2 coding regions were generated by digestion and cloned into the S. pombe expression vector pREP3ΔATG. Standard techniques for transformation and genetic manipulation of S. pombe were used as described previously (Okasaki et al., 1990). S. pombe spg1-B8- and sid1-239-defective strains were used for complementation tests. Transformed colonies were replicated on EMM2-agar plates containing thiamine (non-inducing medium) or without thiamine (inducing medium). Nuclei and septa were labelled using DAPI and calcofluor, respectively, according to Fankhauser and Simanis (Fankhauser and Simanis, 1994).

The coding sequences of plant cDNAs were fused in frame to the GAL4 activation domain (AD) or the GAL4 binding domain (BD) by cloning into the pGAD424 or pGBT9 vectors, respectively. The coding regions were amplified by PCR using specific oligonucleotides: AtSGP1 5′-CTCGAATTCTCTGTTTCATCAATGGCC-3′, 5′-CTAGAATTCAATTAGGAATCGTAGTC-3′; AtSGP2, 5′-GTACGAATTCATGGCTCAATC-3′, 5′-ATAGGATCCCTAGAAGTCGATG-3′; modified with restriction sites (underlined sequences). DNA fragments corresponding to the full length BnMAP3Kϵ1, BnMAP3Kα1 and BnMAP3Kβ1 coding regions were generated by digestion and cloned into pGAD424 or pGBT9.

BD-cdc7, BD-spg1, BD-byr4 and AD-cdc14, AD-cdc7, AD-spg1, AD-byr4 constructs were a gift from V. Simanis. The BD and AD constructs were used to transform the yeast strain Y190 (Harper et al., 1993), and protein interactions were assayed using lacZ-filter lifts and by monitoring growth on SD-LTH medium containing 20 mM 3-amino-1,2,4-triazole (3-AT).

The coding sequences of plant genes were fused in frame with the smGFP encoding gene in the CD3-326 vector (Davis and Vierstra, 1998). The coding regions were amplified by PCR using specific oligonucleotides: BnMAP3Kϵ1 5′-AGGAATTCGAGCTCGGGATCCATGACG-3′, 5′-TTGGATCCACAAGACGGTGTTGATGTG-3′; AtSGP1 5′-GAATTCTCTGGATCCTCAATGGC-3′, 5′-GAATGGATCCAGGAATCGTAGTC-3′; AtSGP2 5′-GTACGGATCCATGGCTCAATC-3′, 5′-CAGGATCCAGAAGTCGATG-3′; AtMAP4Kα1 5′-GTTGGATCCATGGATGATGTTGCTGG-3′, 5′-TCTGTGAGGGATCCAACTTTGGTTAAGATCACGC-3′; AtMAP4Kα2 5′-TAGTGGATCCAGCGGGAGAGAAGGTGG-3′, 5′-GGGGATCCAAAGATCACGTGATGATTG-3′; modified with restriction sites (underlined sequences).

The cell suspension of the tobacco Bright-Yellow 2 (BY2) mutant was maintained according to standard protocols (Murashige and Skoog, 1962) and transformed with the plasmids mentioned above. 50 ml of cell suspension were harvested before it reached the stationary growth phase and subjected to cell wall digestion using the following mixture: 12% sorbitol, 0.1% pectolyase, 1% cellulase pH 5.5 at 30°C under gentle constant swirling. After completion (60-90 minutes), the protoplasts were washed twice with protoplast medium (0.10 M glucose, 0.25 M mannitol pH 5.8 in Murashige and Skoog medium). Transformations were carried out with 10 μg of plasmid per 10⁶ protoplasts in the protoplast medium supplemented with 12.5% polyethylene glycol 4000. After dilution to a 5 ml volume with protoplast medium, protoplasts were cultured for 24-36 hours in the dark at 25°C prior to examination.

Cells were observed under the epifluorescence microscope after staining the nuclei with 4′-6 diamidino-2-phenyl indole (DAPI, 10 μg/ml) and the cell walls with Calcofluor (Tinopal LPW™, 100 μg/ml). GFP fluorescence was imaged with an epifluorescence confocal Sanastro 2000 microscope (Moleculars Dynamics). Excitation of in vivo GFP fluorescence was achieved by the Ar laser line at 488 nm, while the fluorescence emission was collected by a broad band pass filter (500-550 nm). The resulting signal was amplified, digitalized and the consistent picture reconstituted by Leica software. As classically used for animal cells, nucleolar localisation was ascertained by analysing the plant cells by phase contrast microscopy, DAPI staining and GFP observation (Sirri et al., 2002).

To identify plant genes that encode putative SIN components, fission yeast spg1 and sid1 sequences were used to screen the GenBank database using the Basic Local Alignment Search Tool. The Arabidopsis genome contains two homologues of each of the fission yeast genes, namely AtSGP1, AtSGP2, AtMAP4Kα1 and AtMAP4Kα2, which were amplified by PCR from an A. thaliana root cDNA library. The genes AtSGP1 and AtSGP2 (for Arabidopsis thalianaspg1-like Gprotein) are highly similar to the gene encoding the yeast GTPase spg1p (Schmidt et al., 1997). AtSGP1 and AtSGP2 encode polypeptides of 290 and 292 amino acids, respectively, sharing 52% overall sequence identity. Spg1p from S. pombe showed sequence identities of 52% and 46% with AtSGP1 and AtSGP2, respectively (Fig. 2A). However, the plant and fungal proteins differ in their organisation in that the two plant proteins possess an amino terminal extension, consisting of 92 residues in AtSGP1 and 94 residues in AtSGP2, sharing 31% identity (Fig. 2A). The neighbour-joining dendogram shown in Fig. 2B indicates that spg1p and Tem1p from S. cerevisiae, GiTem1p from Giardia intestinalis and AtSGP1 and AtSGP2 constitute a new monomeric G-protein subfamily. None of these proteins possess the carboxyl-terminal CAAX prenylation motif, required for the association with cell membranes by lipid anchoring.

Analysis of the AtMAP4Kα1 and AtMAP4Kα2 genes and their deduced protein sequences revealed that they are closely related to the sid1 gene (Guertin et al., 2000) (Fig. 2C). AtMAP4Kα1 and AtMAP4Kα2 encode protein kinases of 708 and 711 residues, respectively, belonging to the SOK group of MAP4Ks as illustrated in Fig. 2D. They show an overall identity of 78% and are highly similar to Brassica napus BnMAP4Kα1 and BnMAP4Kα2, previously reported by Leprince et al. (Leprince et al., 1999). In vitro kinase assays demonstrated that the BnMAP4Kα2 cDNA encodes a functional serine/threonine protein kinase with auto-phosphorylation activity, which is also able to phosphorylate myelin basic protein (data not shown). Fig. 2C shows that sid1p and the two AtMAP4Kα proteins from Arabidopsis share 50% identity over the catalytic domain, whereas they share 30% similarity over the short carboxyl-terminal domain. Moreover, the organisation of the plant proteins differs from that of sid1p by a non-catalytic carboxyl-terminal extension, consisting of 178 and 179 amino acids in AtMAP4Kα1 and AtMAP4Kα2, respectively (Fig. 2C).

No significant similarities were observed between AtSGP1/2 or AtMAP4Kα1/2 and proteins from the fully sequenced genomes of Caenorhabditis elegans, Drosophila melanogaster and Homo sapiens. Hence, the genes and their corresponding proteins described in this study are designated as `plant-specific SIN-related components'.

In fission yeast, the SIN pathway couples the end of anaphase with the initiation of cytokinesis (McCollum and Gould, 2001). The bioinformatical studies described here show that the plant SIN-related proteins are highly similar. Therefore, it was tempting to speculate that these plant proteins can complement the spg1 or sid1 yeast mutants by reconstituting a functional SIN pathway. To test this hypothesis, yeast complementation experiments, overexpression and two-hybrid assays were performed. The AtSGP1, AtSGP2 or BnMAP4Kα2 coding regions were linked to the thiamine-repressible nmt1 promoter in the vector pREP3ΔATG, and introduced into the spg1-B8 and sid1-239 mutant S. pombe strains. A comparison of the mutant phenotypes under permissive (Fig. 3A1,B1) and restrictive conditions (i.e. 36°C) showed that cells at 36°C failed to form septa (Fig. 3A2,B2). These cells go through additional nuclear division cycles without septation, leading to the production of elongated and multinucleated cells (Schmidt et al., 1997; Guertin et al., 2000). By contrast, 4 hours after the shift to 36°C, the yeast cells expressing AtSGP1, AtSGP2 or BnMAP4Kα2 produced normal septa (Fig. 3A3,B3). In addition, about 10% of the yeast cells expressing AtSGP1 produced additional septum material at the cell tips (Fig. 3A3). A very similar phenotype was previously observed after complementation of the yeast cdc7-24 mutant by the plant BnMAP3Kϵ1 protein kinase (Jouannic et al., 2001), which suggests that the two proteins transduce the same signal in S. pombe. No noticeable abnormalities were observed in cells containing either the constructs AtSGP1, AtSGP2 or BnMAP4Kα2 under non-inducing conditions (data not shown), or the empty pREP3ΔATG vector under inducing conditions (Fig. 3A2,B2).

In fission yeast, overexpression of spg1 and cdc7, using the thiamine-repressible expression vector pREP3ΔATG, induces multiple rounds of septum formation without cell cleavage (Fankhauser and Simanis, 1994; Schmidt et al., 1997). In contrast, overexpression of sid1 does not compromise the signalling of cytokinesis (Guertin et al., 2000). In order to investigate further the role of the plant homologues of cdc7 (Jouannic et al., 2001), spg1 and sid1 genes, we overexpressed the A. thaliana and B. napus proteins in the wild-type S. pombe background. Upon withdrawal of thiamine, growth of the yeast cells carrying the empty pREP3ΔATG plasmid was not affected (Fig. 4A1). In contrast overexpression of either AtSGP1 or BnMAP3Kϵ1 (i.e. homologues of cdc7) disturbed the formation of the septum in part of the cells, resulting in multiseptate cells (Fig. 4A2,A4). Furthermore, 50% of the septate cells overexpressing AtSGP2 showed abnormal cytokinesis, forming cells that were either multiseptate or uninucleate with one or more septa (Fig. 4A3). No abnormal phenotype was observed among the cells overexpressing BnMAP4Kα2 (Fig. 4A5). Cell populations overexpressing either AtSGP1 or BnMAP3Kϵ1 formed twice as many septa as wild-type cell populations (Fig. 4B). Conversely, cells carrying the BnMAP4Kα2 gene did not differ from the wild type with regard to the percentage of septum-forming cells. These results show that overexpression of either the plant BnMAP4Kα2 (this study) or sid1p (Guertin and McCollum, 2001) does not affect the cytokinesis process, and that overexpression of the plant homologues of spg1p or cdc7p is able to control, at least partially, cytokinesis in fission yeast.

Cdc7p and spg1p interact both genetically and physically (Schmidt et al., 1997). Byr4p binds to cdc16p and spg1p (Furge et al., 1998). In order to investigate a possible interaction between the plant SIN-related homologues and S. pombe SIN proteins, we performed yeast two-hybrid experiments. The assays based both on growth on medium lacking His and on β-galactosidase activities (Table 2), indicate an interaction between AtSGP1 or AtSGP2, and byr4p. Furthermore, we also detected positive interactions of cdc7p with either AtSGP1 or AtSGP2. Conversely, no interaction could be detected between BnMAP3Kϵ1 and spg1p. To test the ability of BnMAP3Kϵ1, AtSGP1 and AtSGP2 to interact, two-hybrid experiments were conducted. These experiments showed that BnMAP3Kϵ1, but not BnMAP3Kα1 or BnMAP3Kβ1, could interact with AtSGP1 in yeast (Table 2). No interaction was detected between AtSGP2 and BnMAP3Kϵ1 (Table 2), suggesting a functional difference between the two AtSGP proteins.

Thus the complementation of the yeast sin mutants by AtSGP1, AtSGP2 or BnMAP4Kα2 and the effects of the overexpression of these proteins and BnMAP3Kϵ1 on the division of yeast as well as their interactions in the yeast two-hybrid system with one another and with the yeast SIN proteins, provide complementary evidence that the plant SIN-related proteins are, in fission yeast cells, functional homologues of spg1p, cdc7p and sid1p.

We measured the relative transcript level of AtSGP1/2, AtMAP3Kϵ1/2 and AtMAP4Kα1/2 genes in various Arabidopsis organs, using real-time RT-PCR (Bustin, 2000). The actin genes ACT2 and ACT8, amplified simultaneously, were used as a standard because of their `housekeeping' functions, and demonstrated stable expression profiles (An et al., 1996). The AtSGP1, AtMAP3Kϵ1, AtMAP3Kϵ2, AtMAP4Kα1 and AtMAP4Kα2 genes were ubiquitously expressed, with slightly higher transcript levels in flower buds and mature flowers (Fig. 5A). In contrast, the AtSGP2 gene displayed a significant differential expression, with transcript levels 20 times higher in flower buds and 15 times higher in mature flowers than in other organs (Fig. 5A).

Jouannic et al. previously showed that the AtMAP3Kϵ1 transcript levels are regulated during the cell cycle, showing a peak of expression at the G₂-M phase transition (Jouannic et al., 2001). In order to get an overview of the expression of each A. thaliana SIN-homologue during the cell cycle, we performed expression analyses of the AtMAP3Kϵ1, AtMAP3Kϵ2, AtMAP4Kα1, AtMAP4Kα2, AtSGP1 and AtSGP2 genes in aphidicoline-synchronised cell suspension cultures of Arabidopsis using real-time RT-PCR (Fig. 5B). As reference genes we monitored the expression of the cell cycle phase-specific markers histone AtH4 (Reichheld et al., 1995) and cyclin CycB1;1At (Shaul et al., 1996), marking the S phase and the G₂-M transition, respectively. Examination of the transcript levels of these eight genes during the cell cycle revealed two types of expression patterns. The first expression profile is similar to the profile of CycB1;1At, and comprises transcripts profiles of AtSGP1, AtMAP3Kϵ1, AtMAP3Kϵ2 and AtMAP4Kα2 genes, which are all co-expressed at a higher level at the G₂-M transition. The second type consists of the transcript expression patterns of AtMAP4Kα1 and AtSGP2. AtSGP2 transcripts were markedly induced during S phase and the steady state level remained stable during G₂. In contrast to the other genes, AtMAP4Kα1 displayed a quasi-constitutive expression profile during the course of the cell cycle. Hence, expression of several of the plant SIN-related genes is tightly regulated during the cell cycle with respect to their transcription and/or transcript stability, with the highest mRNA accumulation at the G₂-M phases.

In S. pombe a major part of the regulation of the SIN pathway proteins resides in a strict control of their subcellular localisation during the cell cycle (Simanis, 2003). To investigate whether a similar regulation occurs in plant cells, the localisation of the plant SIN-related proteins was determined using transiently expressed GFP fusions. Because of their amenability to microscopic analysis, protoplasts of tobacco BY2 cells were used for these studies. Twenty-four hours after protoplast transformation, all GFP fusions with the different plant SIN-related homologues were localised to the nucleus of interphase cells (Fig. 6). The data shown in Fig. 6 illustrate that each of the GFP-tagged proteins appeared concentrated in the non-DAPI staining region of the nucleus, i.e. the nucleolus. These interphase cells were characterised by a large nucleolus. A faint, diffuse signal was also detectable in the nucleoplasm, all around the nucleus and in the cytoplasm (Fig. 6B2,B4-B6). In control cells transformed with GFP alone, fluorescence was localised both to the nucleus and to the cytoplasm, including cytoplasmic strands (Fig. 6B1). GFP alone is smaller than the accepted exclusion size limit of nuclear pores, in agreement with its distribution both in the cytoplasm and the nucleus (Chytilova et al., 2000). However, the size of the plant SIN-like component GFP fusion proteins, are above the exclusion size limit of nuclear pores indicating that the plant GFP-tagged proteins are translocated to the nucleus by an active process. We were unable to identify potential nuclear localisation signals similar to the ones identified by Kalderon et al. and Robbins et al. from database searches (Kalderon et al., 1984; Robbins et al., 1991). In conclusion, these results show that, in plant cells, all of the plant SIN-related components AtSGP1, AtSGP2, BnMAP3Kϵ1, AtMAP4Kα1 and AtMAP4Kα2 localise to the nucleolus.

A recent fundamental advance in understanding the mechanisms that coordinate mitosis and cytokinesis is the discovery of the SIN pathway. We present evidence that plants have SIN-related proteins that are related to the spg1p, cdc7p and sid1p proteins of fission yeast. Specifically, we established that (i) the Arabidopsis genome encodes homologues of the yeast spg1p and sid1p proteins; (ii) Arabidopsis AtSGP1, AtSGP2 and the Brassica napus BnMAP4Kα2 can complement the fission yeast septation initiation mutants spg1-B8 and sid1-239; (iii) overexpression of the plant homologues of spg1p and cdc7p interferes with cytokinesis in fission yeast, and (iv) the Arabidopsis AtSGP1 and AtSGP2 proteins interact with relevant partners of the yeast SIN pathway (i.e. cdc7p and byr4p) in the yeast two-hybrid system. These results suggest that AtSGP1, AtSGP2 and BnMAP4Kα2 function in yeast cells as orthologues of S. pombe SIN pathways proteins. The molecular similarities observed between the plant genes and their yeast counterparts, known to be involved in cytokinesis, indicate that the plant and the S. pombe sequences are probably derived from common ancestral genes. However, rescue of the mutant phenotype and dominant-negative effects, show a partial functional divergence between the plant and the yeast proteins. The amino acid sequences of the plant SIN-related components show some differences with the S. pombe proteins spg1p, cdc7p and sid1p, such as plant-specific amino- or carboxyl-terminal domains. This suggests that the plant SIN-related proteins perform novel functions compared to the S. pombe couterparts, by using these additional domains for distinct regulations or interactions.

We provided independent sets of data, which are compatible with the hypothesis that plant SIN-related elements act in a same signalling pathway. We have shown that the plant proteins SPG, MAP3Kϵ and MAP4Kα can functionally replace spg1p, cdc7p and sid1p, which are involved in the SIN signal transduction network in fission yeast. AtSGP1 and BnMAP3Kϵ1 interact as observed for spg1p and cdc7p in the yeast two-hybrid system. Schmidt et al. showed that one possible role for spg1p in fission yeast is to direct cdc7p to its correct subcellular location in order to activate substrates such as sid1p (Schmidt et al., 1997). It is noteworthy that in the same system, however, the protein BnMAP3Kϵ1 failed to interact with the plant paralogous sgp1p-like protein AtSGP2. One interpretation of this finding is that there is a functional difference between the two AtSGP proteins. Alternatively, the lack of interaction could be a yeast two-hybrid artefact or a difference between Brassica napus and Arabidopsis proteins. Additional experiments are required to resolve these issues.

SIN-related gene expression is not regulated in an organ-specific fashion. Similarly, BnMAP3Kϵ1 gene expression is not organ dependent, however, using in situ hybridisation signal quantification, BnMAP3Kϵ1 gene expression is higher in dividing cells than non-dividing ones (Jouannic et al., 2001). The expression of AtSGP1, AtMAP3Kϵ1, AtMAP3Kϵ2 and AtMAP4Kα2 genes increased coordinately during late G₂ phase of the cell cycle in Arabidopsis suspension cells, in contrast to AtSGP2, which is induced during S phase. The timing of mRNA accumulations was almost identical to CycB1;1At cyclin mRNA accumulation marking the G₂-M transition (Shaul et al., 1996). Similarly, we observed an accumulation of BnMAP4Kα2 transcripts in freshly isolated microspores (data not shown), whose nuclei are in late G₂ (Binarova et al., 1993). The expression profile of these elements in plant cells might be compatible with a role in control of cell division. AtMAP4Kα1 gene expression is not regulated in a cell cycle-dependent manner, suggesting that the two AtMAP4Kα genes possess different functions. In fact, in a previous study we showed that AtMAP4Kα1 positively responded at the transcriptional level to both saline and PEG treatments by contrast to AtMAP4Kα2, suggesting that AtMAP4Kα1 is involved in stress response (Charrier et al., 2002).

Huh et al. suggest that in yeast, the combination of transcriptional co-expression and protein co-localisation can be used to corroborate or predict the function of proteins (Huh et al., 2003). Our results indicated that all the proteins AtSGP1, AtSGP2, BnMAP3Kϵ1, AtMAP4Kα1 and AtMAP4Kα2 localised to the nucleolus during interphase. Such a compartmentalization of the plant elements should be compared to the co-localisation of various S. pombe SIN pathway proteins to the spindle pole body. One of the most striking examples is provided by the yeast protein phosphatase clp1p. This protein, acting downstream of the SIN pathway, appears in the nucleolus during interphase, but later during mitosis it exits from the nucleolus and interacts with the mitotic spindle and the actomyosin ring in a SIN-dependent signalling mechanism (Trautmann et al., 2001). Recent analysis of the human nucleolar proteome (Andersen et al., 2002) reinforces the hypothesis of a multi-protein complex of clustered network components. Andersen and collegues report a proteomic analysis using a combination of mass spectrometry and sequence database searches and highlight the dynamic nature of the nucleolar proteome (Andersen et al., 2002). Newly proposed roles for the nucleolus may include its function in the sequestration and the release of proteins involved in cell division, gene silencing and senescence.

In order to identify other components of the putative plant SIN-related pathway, such as regulators and targets of the cascade, an extensive search in the A. thaliana genomic sequence database was carried out. Although we have shown that AtSGP1 and 2 interact with the S. pombe cdc16p-associated protein byr4p, no putative homologues of byr4p have been identified in the available plant genome sequence databases. Furthermore, the Arabidopsis genome appears to lack genes homologous to the polo-like kinase plo1p and the protein phosphatase clp1p, acting upstream and downstream, respectively, of the SIN pathway. It has been shown (Sparks et al., 1999) that the sid2p and mob1p proteins from S. pombe form a complex, which seems to act downstream sid1p. Interestingly, mob1p-like protein-encoding genes are present in the A. thaliana genome (i.e. At5g45550 and At4g19050), but at present it is unclear whether plant genomes encode homologues of sid2p and cdc14p (i.e. sid1p-interacting proteins). Our yeast two-hybrid experiments did not show interaction between the sid1p-like protein BnMAP4Kα2 and cdc14p (unpublished data). These results indicate that some of the plant SIN pathway regulators and possible targets may have diverged during evolution to such an extent that interaction between the yeast and plant proteins is no longer possible. Moreover, sequence database analyses reveal that the SIN pathway components are differentially conserved between the three eukaryotic kingdoms as suggested by the fully sequenced genomes from S. cerevisiae, A. thaliana, Drosophila melanogaster, Anopheles gambiae, Caenorhabditis elegans and Homo sapiens. Some of them seem to be conserved in the three kingdoms such as mob1p- and sid2p-related proteins (Xu et al., 1995; Tao et al., 1999). Some elements are conserved in fungi and metazoan but not in plants, such as the polo kinases and the Cdc14p-clp1p protein phosphatases (Donaldson et al., 2001; Gruneberg et al., 2002). A few elements seem to be specific to fungi such as the cdc16-byr4/Bub2p-Bfa1p two-component GTPase-activating protein complexes and the Lte1p guanine exchange factor. Conversely, the core elements spg1p-, cdc7p- and sid1p-related proteins are conserved only in the fungal and plant kingdoms.

These observations raise the possibility that functional counterparts of several SIN regulators exist in higher eukaryotes.

The role of plant SIN-related proteins in plant cytokinesis remains to be functionally demonstrated. Nevertheless, based upon the homology with the SIN pathway from S. pombe and our data, we can propose the following model. AtMAP3Kϵ1, recruited by AtSGP1, is able to activate AtMAP4Kα2 resulting in the activation of downstream processes (Fig. 1). Several questions still have to be addressed, in particular the effect on plant development of either the overexpression or knock-down of the SIN-related gene and the localisation of these plant proteins during mitosis, in order both to clarify their roles in plant cell division and to understand the molecular network including AtSGP, AtMAP3Kϵ and AtMAP4Kα proteins.

We are grateful to C. Bergounioux and colleagues at the IBP for providing help with the cell cycle analysis. We would like to thank S. Brown from the ISV for many fruitful discussions and his expertise with confocal microscopy. We thank A. Mingam and A. Lecharny for their helpful discussions concerning the real-time RT-PCR technology. We gratefully acknowledge the assistance of I. Gy, R. Boyer and J.-P. Bares for DNA sequencing, for technical help with photographic art work and for access to plant growth facilities, respectively.