Regulatory dephosphorylation of CDK at G₂/M in plants: yeast mitotic phosphatase cdc25 induces cytokinin-like effects in transgenic tobacco morphogenesis

Abstract

Background

During the last three decades, the cell cycle and its control by cyclin-dependent kinases (CDKs) have been extensively studied in eukaryotes. This endeavour has produced an overall picture that basic mechanisms seem to be largely conserved among all eukaryotes. The intricate regulation of CDK activities includes, among others, CDK activation by CDC25 phosphatase at G₂/M. In plants, however, studies of this regulation have lagged behind as a plant Cdc25 homologue or other unrelated phosphatase active at G₂/M have not yet been identified.

Scope

Failure to identify a plant mitotic CDK activatory phosphatase led to characterization of the effects of alien cdc25 gene expression in plants. Tobacco, expressing the Schizosaccharomyces pombe mitotic activator gene, Spcdc25, exhibited morphological, developmental and biochemical changes when compared with wild type (WT) and, importantly, increased CDK dephosphorylation at G₂/M. Besides changes in leaf shape, internode length and root development, in day-neutral tobacco there was dramatically earlier onset of flowering with a disturbed acropetal floral capacity gradient typical of WT. In vitro, de novo organ formation revealed substantially earlier and more abundant formation of shoot primordia on Spcdc25 tobacco stem segments grown on shoot-inducing media when compared with WT. Moreover, in contrast to WT, stem segments from transgenic plants formed shoots even without application of exogenous growth regulator. Spcdc25-expressing BY-2 cells exhibited a reduced mitotic cell size due to a shortening of the G₂ phase together with high activity of cyclin-dependent kinase, NtCDKB1, in early S-phase, S/G₂ and early M-phase. Spcdc25-expressing tobacco (‘Samsun’) cell suspension cultures showed a clustered, more circular, cell phenotype compared with chains of elongated WT cells, and increased content of starch and soluble sugars. Taken together, Spcdc25 expression had cytokinin-like effects on the characteristics studied, although determination of endogenous cytokinin levels revealed a dramatic decrease in Spcdc25 transgenics.

Conclusions

The data gained using the plants expressing yeast mitotic activator, Spcdc25, clearly argue for the existence and importance of activatory dephosphorylation at G₂/M transition and its interaction with cytokinin signalling in plants. The observed cytokinin-like effects of Spcdc25 expression are consistent with the concept of interaction between cell cycle regulators and phytohormones during plant development. The G₂/M control of the plant cell cycle, however, remains an elusive issue as doubts persist about the mode of activatory dephosphorylation, which in other eukaryotes is provided by Cdc25 phosphatase serving as a final all-or-nothing mitosis regulator.

INTRODUCTION

Temporal and spatial regulation of the cell cycle is essential for establishment and maintenance of meristems as well as for control of organogenesis and growth, and represents one of the most intensively studied processes in contemporary plant biology. The basic mechanisms of cell cycle are conserved among all eukaryotes (Nurse, 1990) and much knowledge has been achieved with yeast and animal models. However, studies on the plant cell cycle have so far failed to lead to a comparably clear picture, partly because of the existence of unique plant regulatory pathways not observed in other eukaryotes (e.g. Joubès et al., 2000; Vandepoele et al., 2002; Bryja et al., 2008; Wang et al., 2008). These distinctive plant features reflect the particularities of plant development, including post-embryonic development, unlimited growth provided by life-long meristem activity enabling continuous organ formation or totipotency of somatic plant cells allowing regeneration of organs/tissues lost to disease, environmental stress and predators (e.g. Steeves and Sussex, 1989; Burssens et al., 2000; Fletcher, 2002; de Jager et al., 2005). Together these features create enormous plasticity that plant life, being sessile, requires.

MECHANISMS OF CELL CYCLE REGULATION IN EUKARYOTES

Proper maintenance of multicellular integrity in metazoa or plants requires accurate regulation of cell division, which is based on the ability of the cells to sense precisely when they are required to divide or when promotion of cell proliferation must stop or be modified (Joubès and Chevalier, 2000; Edgar and Orr-Weaver, 2001; Nurse, 2002; Tyson and Novak, 2008). Evidently, the decision to start cell division depends on signals resulting from the perception and integration of the information about nutrient availability, environmental conditions and spatio-temporal context of requirements within a body (Meijer and Murray, 2000). The signals find their targets in the proteins, namely the cyclin-dependent protein kinases (CDKs) and their associated subunits, which govern two major cell-cycle transitions: G₁/S and G₂/M (Hartwell et al., 1970; Van't Hof et al., 1978; Van't Hof and Bjerknes, 1981; Van't Hof, 1985; Murray, 1992; Norbury and Nurse, 1992; Guertin et al., 2002; Nurse, 2002; Francis, 2007; Oliferenko, 2009). The relative importance of each of these transitions is development- and species-dependent (O'Farrell, 2001; Doonan and Kitsios, 2009; Harashima and Schnittger, 2010). It became apparent (recently reviewed by Sullivan and Morgan, 2007; Bollen et al., 2009) that the control of these events is primarily ensured by reversible phosphorylation of key regulatory proteins. Among eukaryotes, there is a large diversity in protein kinases and phosphatases (Wang, 1994; Simanis, 1995; Goff et al., 1999; Bollen et al., 2009); especially, the precise timing of CDK dephosphorylated status at a specific site allows transition through mitotic events (O'Farrell, 2001; Sullivan and Morgan, 2007; Bollen et al., 2009; Carmera et al., 2009).

Two species of unicellular yeasts, the budding yeast, Saccharomyces cerevisiae, and the fission yeast, Schizosaccharomyces pombe, have been used extensively as models for eukaryote cell cycle studies. In S. cerevisiae, the cell cycle progression is controlled solely at G₁/S transition, which is known as START, and is governed by CDC28 (Hartwell, 1974; Nasmyth et al., 1991). In Schizosaccharomyces pombe, on the other hand, the cell cycle is regulated at both the G₁/S and G₂/M transitions by the cdc2 gene, which is functionally homologous to budding yeast's CDC28 (Nurse and Bissett, 1981; Nasmyth et al., 1991). Both of these transitions are highly conserved among eukaryotes, with regulation secured by cdc2 orthologues known as cyclin-dependent kinases (CDKs) (Nurse, 2002; Dewitte and Murray, 2003; Cools and De Veylder, 2009).

These main regulatory proteins of the cell-cycle progression, the cyclin-dependent serine/threonine protein kinases (CDKs), are, in turn, regulated by varying abundance of their non-catalytic partners – cyclins (Evans et al, 1983; Lees and Harlow, 1993; Pines, 1993; Vandepoele et al., 2002). Cyclin levels are tightly regulated through transcription or specific protein degradation (Lees and Harlow, 1993). The controlled destruction of cyclins within mitosis stops CDK activity and guarantees that cytokinesis follows after the chromosomes are separated (Guertin et al., 2002). The cyclins also regulate the substrate specificity of the CDK–cyclin complexes by directing the CDK subunits to appropriate subcellular spaces (Ubersax et al., 2003; Wittenberg, 2005).

Large superfamilies of CDKs and cyclins enable flexibility in eukaryotic cell-cycle regulation through the high possibility of combining to form made-to-measure CDK–cyclin complexes (Lees and Harlow, 1993; Menges et al., 2005; Nieuwland et al., 2009). The substrates of CDK–cyclin complexes include transcriptional regulators, cytoskeleton, nuclear matrix, nuclear membrane proteins, chromatin-associated proteins as well as other cell-cycle proteins (Norbury and Nurse, 1992; Koch and Nasmyth, 1994; Joubès et al., 2000).

The activity of CDK/CYC complexes is further influenced by direct association with other regulatory proteins, the CDK subunits (CKS) and CDK inhibitory (CKI) proteins. The CKS proteins function as docking factors for substrates of the CDK/CYC complex (Patra and Dunphy, 1998; Patra et al., 1999), whereas the CKI proteins inhibit CDK/CYC complexes in response to environmental and developmental signals (LaBaer et al., 1997; Cheng et al., 1999).

Furthermore, as mentioned above, the CDK/CYC complex activity is controlled through modification of its phosphorylation status, which is important for the spatio-temporal control of cell-cycle events. This control is partly provided by CDK-activating kinases and interplay between WEE1 inhibitory kinase and the activating CDC25 phosphatase (Murray, 1992; McGowan and Russell, 1995; Nurse, 2002; Kiyokawa and Ray, 2008; Tyson and Novak, 2008). CDC25 and WEE1 act as respective ‘on’ and ‘off’ switches on CDK activity (Cools and De Veylder, 2009). At G₂/M, dephosphorylation of the CDK occurs at tyrosine 15 in fission yeast and also at threonine 14 in animals. CDC25 phosphatase dephosphorylates these residues which allows the CDK to become catalytic and, thus, responsible for the final ‘all-or-nothing’ signal that drives the cell into mitosis (O'Farrell, 2001). Hence, a key feature of cell-cycle control is negative regulation that prevents the cell from dividing until conditions (or demands of development) are ideal.

Doonan and Kitsios (2009) discussed the roles of eukaryotic CDKs and pointed to similarities and differences in protein family organization and biological function between plants and animals. They propose that the prototypic kinases, evolved to control cell division in complex eukaryotic cells, are generalist kinases having a truly global effect on cellular structure, gene expression and metabolism. However, eukaryotic evolution has brought about family diversification resulting in the appearance of the more specialized family members, fulfilling more specialist roles and regulating particular aspects of gene expression.

MECHANISMS OF PLANT CELL-CYCLE REGULATION

As mentioned above, although basic mechanisms of cell cycle progression are in principle very similar in all eukaryotes, some differences in cell-cycle regulation machinery can be found in plants (larger variability in regulation, greater amounts of alternative proteins with redundant function, etc.; see below) reflecting the specificity of plant-life strategy.

Like other eukaryotic CDKs, plant CDKs are serine-threonine-specific protein kinases usually divided into seven classes according to the conserved cyclin-binding sequence motif (Vandepoele et al., 2002; Doonan and Kitsios, 2009). The kinases of the CDKA group are found in all tested organisms across the eukaryotic kingdom, and are thought to be pivotal for both the G₁/S and G₂/M transitions (reviewed, for example, by Porceddu et al., 2001; Vandepoele et al., 2002; Doonan and Kitsios, 2009). They share a conserved PSTAIRE amino acid sequence in a non-catalytic domain that is responsible for binding with a cyclin partner (Joubès et al., 2000). CDKB kinases constitute a plant-specific group of CDKs not yet identified in any other organism (e.g. Joubès et al., 2000; Inzé and DeVeylde; 2006; Boruc et al., 2010). In arabidopsis, CDKA is encoded by a single gene (Iwakawa et al., 2006; Nowack et al., 2006) but CDKB is represented by a multigene family, whose members may have more specialized roles in particular cell types (e.g. Boudolf et al., 2004). CDKA generally complements cdc2/cdc28 mutants in yeasts, whereas CDKB genes do not, although lower-plant CDKB genes seem to be an exception (Corellou et al., 2005). For the CDKB class, a cyclin-binding PPTA(T)LRE motif exists, and its time window of activity is almost exclusively at the G₂/M transition (Joubès et al., 2000; Inzé and De Veylder, 2006).

Members of the plant CDKD and CDKF kinases are sequentially and, as shown by yeast complementarity analyses, also functionally related to other eukaryotic CDK-activating kinases (CAK) or CAK-activating kinases, respectively (Yamaguchi et al., 1998, 2000; Vandepoele et al., 2002; Shimotohno et al., 2004; Takatsuka et al., 2009). CDKE kinase probably forms complexes with D-type cyclins, suggesting interaction of developmental and environmental signals with cell cycle progression (Boruc et al., 2010). Thus far, CDK classes C and G have not been linked with any cell-cycle regulatory functions (Vandepoele et al., 2002; Doonan and Kitsios, 2009).

In contrast to CDK expression profiles, which include non-dividing tissues with division potential (e.g. the pericycle, dormant buds), the cyclin subunits of the functional heterodimers are expressed almost exclusively in dividing cells and their protein levels are usually oscillating during the cell cycle (e.g. Morgan, 1997). Cyclin proteins also have a so-called cyclin box enabling binding with their CDK partners (e.g. Renaudin et al., 1996; Wang et al., 2004). Cyclins determine not only conformational changes in the catalytic site of CDK but are also responsible for substrate selection, subcellular localization and stability of the CDK–cyclin complexes (for review, see Wang et al., 2004).

In the plant kingdom, ten cyclin classes (A–D, H, L, T, U, SDS and CycJ18) have been established so far, according to the animal cyclin nomenclature, with five classes exclusively found in plants that comprise many more members than in animals (Wang et al., 2004). For example, 40 distinct cyclins have been found in the arabidopsis genome (Wang et al., 2004), which is far more than in any other complex eukaryote. A- and B-type cyclins are classified as mitotic cyclins with expression peaks in S-G₂-M and G₂/M phases, respectively. In promoters of plant B-type cyclins, a nine-nucleotide sequence (M-specific activator – MSA) homologous to animal Myb-binding domain has been recognized, indicating a G₂/M-specific transcription activation (Ito et al., 1998, 2001; Ito, 2000). Both A and B cyclins contain a conserved N-terminal sequence called the destruction box (D-box) to which unbiquitin residues bind in early anaphase, which is then followed by ubiquitin-dependent proteolysis (e.g. Renaudin et al., 1996; Vandepoele et al., 2002; Wang et al., 2004). Both A and B cyclins have CDKA as a binding partner but it seems highly probable that they can interact with CDKB partner(s) as well (Boruc et al., 2010).

Another important class of cyclins are the D-type. This group of proteins was originally called G₁ cyclins; however some members show constitutive expression throughout the cell cycle. Nevertheless, the main cell-cycle regulatory function of D cyclins is still mostly considered at the G₁/S transition, as the majority of D cyclins exhibit complementarity to yeast G₁ cyclins (for a review, see Renaudin et al., 1996). All D-type cyclins, have an N-terminal LxCx(D/E) retinoblastoma-binding motif and a PEST sequence responsible for targeting ubiquitin/proteasome-mediated proteolysis by the so-called SCF complex (Renaudin et al., 1996; Vandepoele et al., 2002; Planchais et al., 2004; Yanagawa and Kimura, 2005). The most frequent partners of D cyclins are CDKA and CDKB (for a review, see Vandepoele et al., 2002; Menges et al., 2005; Boruc et al., 2010).

H-type cyclins play a cell-cycle regulatory role as partners of CAK (CDK-activating kinase) which at G₀/G₁/S catalyses the phosphorylation of a tyrosine residue (Tyr-160/167) enabling CycD–CDKA binding (H cyclin peaks at the G₁/S transition; Yamaguchi et al., 2000; Shimotohno et al., 2004). For successive progression through the cell cycle the main CDK–cyclin complex has to be active. First, the activity is influenced by presence or absence of CDK inhibitor(s) (CKI). CDK inhibitors seem to be active in both G₁/S and G₂/M transitions; however, most data regarding their function come from regulation of entry into S phase. In arabidopsis, seven classes of CDK inhibitors (KRP1-7 from Kip-related proteins, originally designated as ICKs) have been identified to date (Wang et al., 1997, 2008; De Veylder et al., 2001). KRP proteins are able to inhibit complexes with D cyclins and, as it was found recently, also with A- and B-type cyclins bound to CDKA (Nakai et al., 2006; Wang et al., 2008; Boruc et al., 2010).

Beside the inhibition of CDK–cyclin complexes by CKI, the important roles of reversible (de)phosphorylation on specific CDK amino acid residues have to be mentioned. As described above, CAK, CDK-activating kinases, are responsible for specific activation of CDK by phosphorylation on conservative threonine 160 (or to a similar homologue-dependent positioned threonine residue) (e.g. Joubès et al., 2000). The main role of CAK is activation of CDK–cyclin at the G₁/S transition, while the precise activity at G₂/M remains to be established (De Veylder et al., 2003; Inzé and De Veylder, 2006). (De)phosphorylation on other amino acid residues controlling CDK–cyclin-complex activity at the G₂/M transition will be discussed in the next section.

Regulation of G₂/M transition

Although the data on the transition of a plant cell through G₁/S conforms well to a generic eukaryote model (for a review see, for example, Vandepoele et al., 2002; De Veylder et al., 2003; Dewitte and Murray, 2003; del Pozo et al., 2002, 2006; Inzé and De Veylder, 2006), information about the regulation of G₂/M transition is rather sketchy, and it is still not clearly understood whether plants, at least higher plants, have evolved an alternative pathway controlling optimal initiation of the mitotic process.

As reviewed, for example, by Inzé and De Veylder (2006), after successful accomplishment of DNA replication, the CDKA/B kinases form complexes with A, B and D cyclins. The function of CDKA complexes with A- and B-type cyclins is generally accepted (e.g. Joubès et al., 2000; De Veylder et al., 2003), the role of a CDK/D cyclin complex remains to be elucidated (Kono et al., 2003; Boruc et al., 2010). These heterodimers are rapidly phosphorylated, thus inactivated, by WEE1 kinase on the conserved threonine 14 and tyrosine 15 of CDKA/B (e.g. Sun et al., 1999; Sorrell et al., 2002; Gonzalez et al., 2004). WEE1 inactivation led to a shortening of cell size (Gonzalez et al., 2007) but its overexpression seemingly started the endoreduplication process (e.g. Boudolf et al., 2004). Further, the activatory phosphorylation on threonine 160–167 governed by CDK-activating kinase (CDKD; as mentioned above) is an essential post-translational modulation of the complex. For the cell to enter mitosis, the inhibitory phosphates have to be removed as they block the substrate and ATP-binding sites (e.g. De Veylder et al., 2003). Although much effort has been put into finding the plant phosphatase responsible for this activatory dephosphorylation, no full-length plant homologue of animal and yeast CDC25 protein has been revealed so far (Doonan and Kitsios, 2009; Boruc et al., 2010). The clear exception is the unicellular alga Ostreococcus tauri, where a full-length CDC25 was identified which is able to complement fission yeast cdc25⁻ mutants (Khadaroo et al., 2004). In the arabidopsis and rice genomes, genes/proteins have been identified containing solely the catalytic phosphatase domains that were able to activate a corresponding kinase in vitro, but they failed to complement cdc25⁻ mutants, presumably because they lack regulatory domains (Landrieu et al., 2004; Sorrell et al., 2005). Moreover, the arabidopsis CDC25 will act as an arsenate reductase if challenged with exogenous arsenate (Bleeker et al., 2006; Dhanker et al., 2006; Ellis et al., 2006). However, and notably, human CDC25 B and C will also show arsenate reductase activity in the same way (Bhattacharjee et al., 2010).

The role of signal molecules (phytohormones and saccharides) in regulation of G₂/M transition

For precise co-ordination and timing of cell division within a plant body, a highly important role has been attributed to external and internal signalling molecules. Among the generally accepted internal regulatory molecules, phytohormones are mentioned preferentially; nevertheless during the last decade saccharides have often been described as potent signalling agents.

Although all groups of phytohormones probably participate in regulation of cell-cycle gene expression, auxins and cytokinins are doubtless the most important of them. Auxins and cytokinins often act synergistically on expression of many cell-cycle genes (e.g. most of CDKA, CDKB, some of A-, B- and D-type cyclins and KRP) (e.g. Richard et al., 2002; Dewitte and Murray, 2003; del Pozo et al., 2005; Hartig and Beck, 2006; Cho et al., 2010); cytokinins also co-operate with saccharides in gene expression activation, e.g. D2 and D3 cyclins in arabidopsis (Riou-Khamlichi et al., 1999; 2000; Richard et al., 2002; Hartig and Beck, 2006).

At G₂/M a clear role of cytokinins has been documented many times (Redig et al., 1996; Zhang et al., 1996; Laureys et al., 1998). Cytokinins are supposed to affect cell-cycle progression mainly through CDK–cyclin-complex activation by inducing a signal transduction chain that ends in dephosphorylation of Tyr15 and Thr14 (Zhang et al., 1996, 2005). However, experiments with cytokinin analogues (roscovitine and olomoucine) proved that these molecules occupy the ATP-binding site (reversibly phosphorylated – inactivated by WEE1) thus disabling proper CDK function (e.g. Planchais et al., 1997; Vermeulen et al., 2003), which indicates strongly, a more direct involvement of a cytokinin molecule. More direct evidence of the cytokinin interaction with activatory dephosphorylation and the detailed characteristics of this important control mechanism, if it really exists (Boudolf et al., 2006), are out of reach as all attempts to identify the responsible phosphatase have been fruitless up to now.

CHARACTERISTICS OF PLANTS EXPRESSING FISSION YEAST CDC25

In the absence of an identified plant homologue to cdc25, Bell et al. (1993) raised questions about the regulation of the cell cycle and development by expressing the fission yeast (Schizosaccharomyces pombe) cdc25 (Spcdc25) in tobacco (Nicotiana tabacum ‘Samsun’) under the control of the 35S CaMV promoter. Since then many studies have been devoted to investigating the effects of Spcdc25 expression on plant growth and development (Bell et al., 1993; McKibbin et al., 1998; Wyrzykowska, 2002; Chrimes et al., 2004; Suchomelová et al., 2004; Orchard et al., 2005; Zhang et al., 2005; Teichmanová et al., 2007; Suchomelová-Mašková et al., 2008). The results of these studies have shown that Spcdc25 expression leads to marked changes in plant morphology and development and will be discussed in detail in the following sections. The plants expressing fission yeast cdc25 are referred to, hereafter, as Spcdc25 plants.

Spcdc25 expression influences flowering of day-neutral tobacco

Flowering is the most dramatic developmental phase in the life cycle of an angiosperm. In spite of tremendous progress in understanding the genetic control of flowering, major questions regarding flowering regulation have not been answered satisfactorily. This is especially true for day-neutral plants. Classic grafting experiments established that leaves produce a graft-transmissible floral stimulus (florigen) that is transported to the shoot apical meristem (SAM) which then switches from vegetative to floral morphogenesis (Chailakhyan, 1936; Habermann and Wallace, 1958; Zeevaart, 1962, 1976; Lang, 1965, 1989; Chailakhyan and Khazakhyan, 1974a, b). In addition, a floral inhibitor is supposed to be produced by the root system which is able to prevent the SAM from making an inflorescence until enough leaves are formed to displace the SAM from the zone of root inhibition (Gebhardt and McDaniel, 1991; McDaniel et al., 1992; McDaniel, 1996). Thus, the general picture in day-neutral plants is of a florigen able to prevail over root-produced inhibitor(s) and induce flowering in developmentally competent SAMs (e.g. Tran Thanh Van, 1973; Tran Thanh Van et al., 1974; McDaniel et al., 1996).

FT protein (FLOWERING LOCUS T) was recently demonstrated in the facultative long-day plant Arabidopsis thaliana, moving from the leaf to the SAM (An et al., 2004; Corbesier et al., 2007), and is proposed to be a component of florigen (reviewed by Lagercrantz, 2009). The genes that afterwards transform the SAM to a floral one have been well characterized in arabidopsis (Samach et al., 2000; Boss et al., 2004; Jack, 2004; Abe et al., 2005; Yoo et al., 2005; Aksenova et al., 2006; Blazquez et al., 2006; Liu et al., 2009).

In contrast to knowledge gained from studying arabidopsis, very little is known about molecular mechanisms that control flowering in day-neutral plants. However, in day-neutral tomato, an FT homologue was cloned (Lifschitz et al., 2006) that can also induce flowering in Arabidopsis thaliana and photoperiodically sensitive as well as day-neutral tobacco plants (Lifschitz et al., 2006). However, in tomato it was demonstrated that, in addition to functioning as a floral modulator, the FT homologue serves as general growth hormone with a much broader spectrum of targets (Shalit et al., 2009).

The main question remains whether physiological readiness to export sufficient florigen (Lang, 1965), or developmental competence of the SAM to respond to florigen (Robinson and Wareing, 1969; Bernier et al., 1981; Francis, 1987; McDaniel, 1992; Bernier et al., 1993), is critical for flowering in day-neutral plants. Interestingly, the study of flowering onset in Spcdc25 tobacco proved to be useful for addressing this problem.

Spcdc25 activatory phosphatase expression results in precocious flowering in day-neutral tobacco

Bell et al. (1993), in glasshouse experiments with Spcdc25 tobacco ‘Samsun’ (a day-neutral plant; gene expression verified), showed an earlier onset and enhanced intensity of flowering. The plants flowered after formation of less than half of the leaves characteristic for a flowering wild type (WT; Fig. 1; P. Vojvodová et al., unpubl. res.). Clearly, the premature flowering in Spcdc25 plants could be because of strengthened floral stimulus or earlier SAM floral competence, or both.

Fig. 1.

Precocious flowering in tobacco expressing Spcdc25 mitotic activator. WT, Flowering ‘Samsun’ tobacco plants; Spcdc25 plants, flowering Spcdc25-expressing ‘Samsun’ tobacco plants. Note that Spcdc25 plants also formed inflorescences from axillary buds and formed substantially fewer leaves until onset of flowering (P. Vojvodová et al., unpubl. res.). For details on tobacco plant transformation with the S. pombe cdc25 gene and other plant morphology characteristics see Bell et al. (1993).

Grafting experiments revealed primary responsibility for the observed early flowering onset resided in the apex (P. Vojvodová et al., unpubl. res.). Scions (apical buds) from Spcdc25-expressing plants were grafted onto WT root stocks and the floral response determined and compared with responses in reciprocal grafts. SAMs from Spcdc25 tobacco could flower early when grafted onto WT root stock after formation of a remarkably similar number of leaves to intact Spcdc25 plants. However, the converse graft, WT SAM to Spcdc25 root stock, did not flower early but instead retained WT floral kinetics. In other words, Spcdc25 stock did not speed up flowering in the WT scion, and thus earlier flowering is not caused by earlier acquisition of floral stimulus threshold level or lower production of inhibitory signal in Spcdc25 roots (P. Vojvodová et al., unpubl. res.).

Singer et al. (1992) demonstrated by grafting experiments between day-neutral and long-day tobacco plants that ontogenetically older day-neutral apical buds/meristems have a greater competence to respond to the floral stimulus than younger ones. They suggested that an increasing apical meristem competence during plant ageing is an important regulatory mechanism of the length of vegetative growth.

We propose that mitotic tyrosine phosphatase Spcdc25 expression makes the SAM florally competent much earlier and thus enables it to respond earlier to the floral stimulus than WT.

Spcdc25 activatory phosphatase expression alters the SAM structure

Nougarede (1967) studied SAM cell organization in several taxonomically unrelated angiosperms at vegetative and prefloral phases and revealed marked differences. The prefloral SAMs exhibited pronounced bulging structures with isodiametric cells forming the tunicas. A more convex shape of the SAM outer layers was also observed in a short-day tobacco after exposure to a series of inductive short days (Hopkinson and Hannam, 1969). Francis (1998) found frequent anticlinal divisions in the outer layers of prefloral SAMs, with no significant difference in mitotic cell size between vegetative and prefloral tunicas. Similarly, a predominance of anticlinal cell divisions in the tunicas of prefloral meristems was documented recently by Kwiatowska (2008) in a range of unrelated angiosperms.

The increase in meristem size immediately before floral realization was proposed to be the result of shorter cell cycles (e.g. Miller and Lyndon, 1976; Nougarede et al., 1991). In accordance, more cells in mitosis in the SAMs were reported to be connected with the earliest events of floral evocation, preceding by many hours any growth changes leading to formation of the first reproductive structures (Bernier et al., 1967; Bernier, 1988; Jacqmard et al., 2003). In the long-day plant Silene coeli-rosa, the transient increase in mitoses occurred several days before the start of floral realization (Francis and Lyndon, 1978; Ormrod and Francis, 1986). It is suggested that developmental control genes regulate cell division in the prefloral SAM (Vincent et al., 1995; Doonan, 1998; Bernier and Perilleux, 2005).

In Antirrhinum majus, determination of the spatial expression of CDKs in inflorescences revealed much stronger expression of CDKs in the outer layers of the SAMs (Doonan, 1998). In Spcdc25 plants, mounding up of cells in the outer layers of SAMs of very young vegetative plants was also observed (Fig. 2; P. Vojvodová et al., unpubl. res.). This is consistent with a propensity to periclinal cell division that is reflected in a reduced ratio of the number of cells in the outermost layer to the total number of meristematic cells in the SAM (P. Vojvodová et al., unpubl. res.). We propose that activatory phosphatase Spcdc25 expression induces this early change in SAM structure, enabling these plants to respond early to the florigen. In this context, it might be useful to remind ourselves that Gutierrez (2005) proposed that cell cycle regulators may serve as targets for integration of cell proliferation with differentiation, morphogenesis and growth. Further, Andersen et al. (2008) suggest that CDKB2 may be such a potential target within the SAM; the suggestion was based on the results showing that disruption of CDKB2 function resulted in failure to translate/interpret stimuli, e.g. hormone balance into a proper developmental output. Thus, the results obtained with Spcdc25 plants are in harmony with the concept of a connection between cell-cycle regulation and the ability to interpret signal(s) imported into the SAM.

Fig. 2.

Meristem structure as influenced by Spcdc25 expression, visualized by two-step staining with alcian blue and nuclear fast red. WT, ‘Samsun’ tobacco plants; line A and line C, lines A and C of ‘Samsun’ tobacco plants expressing the Spcdc25 gene (P. Vojvodová et al., unpubl. res.). For plant material characterization, see Bell et al. (1993) or Teichmanová et al. (2007).

Interestingly, Spcdc25 expression caused a change in the plane of cell division, resulting in the formation of clusters of nearly isodiametric cells, in both tobacco BY-2 cells expressing Spcdc25 and suspension cultures established from Spcdc25 tobacco ‘Samsun’ (Orchard et al., 2005; Suchomelová-Mašková et al., 2008). Note that in WT tobacco cells in culture, transverse division of elongated cells prevails. Moreover, the Spcdc25 phenotype could be simulated in cells of WT tobacco ‘Samsun’ with exogenous cytokinin application (Suchomelová-Mašková et al., 2008).

Cytokinins are reported to be important regulators of floral competence in the SAM (e.g. Bernier and Perilleux, 2005). Notably, Spcdc25 expression can replace cytokinins (discussed in more detail below): to induce de novo vegetative shoot formation (Suchomelová et al., 2004); to dephosphorylate plant CDK (Zhang et al., 2005); and to drive cells into premature mitosis (Orchard et al., 2005). Collectively, these observations point to a critical interface between cytokinins required for SAM development and a cytokinin requirement for activating more cells into mitosis. The presented data are consistent with Spcdc25 expression, at least partly, replacing a critical cytokinin requirement to establish florally competent SAMs.

Spcdc25 expression and sucrose act synergically in speeding up flowering onset

In photoperiodically sensitive plants, carbohydrates have an important function in floral transition (Bernier et al., 1993; Corbesier et al., 1998; Roldan et al., 1999; Ohto et al., 2001). However, reports dealing with the influence of carbohydrates on flowering in day-neutral plants are rather scarce, e.g. in tobacco (Konstantinova et al., 1972, 1976) and in tomato (Dielen et al., 2004). In day-neutral tobacco it is supposed that the floral stimulus is multicomponent and that sugars are indispensable sub-components (Bernier, 1988). In inducible plants, an inductive photoperiod can cause a rapid increase in sucrose levels in leaf exudates (Houssa et al., 1991; Lejeune et al., 1991) and rapid and, in some cases, transient sucrose accumulation in the SAM (Bodson and Outlaw, 1985; Bernier et al., 1993; Corbesier et al., 1998). Interestingly, the addition of sucrose can rescue the late-flowering phenotype of several arabidopsis mutants (Roldan et al., 1999).

The determination of endogenous carbohydrate levels in leaves revealed that Spcdc25 plants have a higher content of glucose and fructose compared with WT, and, moreover, sucrose found in Spcdc25 plants was barely detectable in WT (Teichmanová et al., 2007). Therefore, it is possible to consider a contribution of increased endogenous sugar levels to flowering promotion in Spcdc25 plants. Several authors reported considerable increases in the sucrose amounts reaching the apex after induction in photoperiodic plants. As this change long preceded any morphological events, they suggested a message-like role for sucrose in flowering initiation (Lejeune et al., 1993; Corbesier et al., 1998; Dielen et al., 2001). Essentially this was also a conclusion drawn from flowering responses of Pharbitis nil in culture (Durdan et al., 2000; Parfitt et al., 2004). Whether the role of sugars in flower induction in day-neutral plants is signalling or just a nutritive one is an on-going debate (Francis and Halford, 2006).

Roldán et al. (1999) proposed that sucrose could promote flowering by regulating the expression of the flowering repressor FLC, a key regulator of the autonomous regulatory pathway (Putterill et al., 2004) and, probably, also of ones controlled by day length (Noh et al., 2004). Teichmanová et al. (2007) followed the flowering onset in Spcdc25 tobacco plants derived from repeatedly sub-cultivated apical or basal-node segments under in vitro conditions, which allowed manipulating exogenous carbohydrate supply. Low-sucrose treatment (3 %) enabled flowering exclusively in Spcdc25 plants derived from the cultured nodes, while comparable WT plants did not flower even after a much higher number of sub-cultivations. When those WT plants were transferred to ex vitro, they formed a comparable number of leaves to seed-derived plants grown in soil. In the 5 % sucrose treatment, both the Spcdc25 and WT plants flowered, although Spcdc25 plants required a substantially lower number of sub-cultivations and formed considerably fewer leaves until flowering. Increasing sucrose supply to 7 % resulted in a time-to-flowering that was even shorter in Spcdc25 plants compared with Spcdc25 plants at both 3 % and 5 %. In all cases, flowering was significantly earlier in Spcdc25 plants in comparison to WT plants in corresponding sucrose treatments.

Indeed we concluded that, in day-neutral tobacco, Spcdc25 expression and increasing exogenous sucrose had a synergistic effect in shortening time-to-flower as well as reducing the number of leaves formed before flowering (Teichmanova et al., 2007). This is consistent with the picture of the SAM becoming prematurely florally competent due to yeast mitotic phosphatase Spcdc25 expression that reacts to floral stimulus strengthened by an enhanced level of one component: sucrose.

Acropetal floral gradient is disturbed in Spcdc25 plants

In accordance with published data (e.g. McDaniel and Hartnett, 1993), in WT tobacco an acropetal floral capacity gradient was observed when apical as opposed to basal nodes were cultured in vitro (Teichmanová et al., 2007). Surprisingly, in Spcdc25 plants with verified gene expression there was no significant difference between numbers of leaves formed and days-to-flowering in basal compared with apical cultured nodes. Hence a positive acropetal effect of node position on time-to-flowering observed in WT was abolished in the Spcdc25 plants. This is consistent with the observation that in intact Spcdc25 tobacco inflorescences are also formed from axillary buds (P. Vojvodová et al., unpubl. res.). Note that for WT tobacco only a terminal inflorescence is typical.

A consensus view is that the acropetal floral gradient exhibited by successive nodes in culture results from integrated signals of numerous chemical components. Using a late-flowering mutant uniflora, Dielen et al. (2001) identified sucrose, cytokinins and nitrogenous nutrients that promoted the floral transition in day-neutral tomato. Moreover, Corbesier et al. (2002) reported that the C : N ratio in phloem sap increased during inductive treatments and the inequality in the C : N supply may be important at the floral transition in Sinapis alba and arabidopsis.

However, grafting experiments, as mentioned above, proved that acquisition of premature floral competence by the SAM was responsible for earlier flowering in intact Spcdc25 plants, thus excluding marked changes in the levels of flowering-promoting compounds caused by Spcdc25 expression (P. Vojvodová et al., unpubl. res.). Taken together, the data support the idea of Spcdc25 expression inducing early floral competence even in axillary buds.

Spcdc25 expression affects de novo organogenesis

Primordium initiation could be driven by a local increase in the rate of cell division or by a reorientation of the plane of cell division, followed by expansion and differentiation processes in the proliferating tissue (Meyerowitz, 1997; Vernoux et al., 2000). Alternatively, bulging could be driven by local expansion of primordium initials (Fleming et al., 1997; Reinhardt et al., 1998) followed by division of expanded cells. Hence, any interference with cell-cycle regulation might result in changes to organogenic processes.

De novo shoot formation is more abundant in Spcdc25-expressing tobacco stem segments

Given the proposed links between cytokinins, Spcdc25 and the cell cycle (Zhang et al., 1996, 2005), and cytokinins and organ morphogenesis it can be expected that Spcdc25 expression could influence de novo organogenesis normally controlled in vitro by hormone balance. Suchomelová et al. (2004) showed earlier and more frequent shoot primordium formation on Spcdc25 internode segments cultivated in vitro on shoot-inducing media compared with WT. Spcdc25 expression could also induce de novo shoot morphogenesis in internodes of transformed tobacco plants on hormone-free medium, i.e. in the absence of exogenous cytokinins (Fig. 3). Moreover, the shoot development (together with scarce root formation) was observed in Spcdc25 segments even on a medium supplemented with a high auxin : cytokinin ratio where solely root morphogenesis is normally induced in WT (Suchomelová et al., 2004). Clearly, the presence of Spcdc25 in tobacco results in a response, which resembles a reaction to the exogenous hormone balance shifted towards cytokinins, thus creating a cytokinin-like effect. Riou-Khamlichi et al. (1999) reported cytokinin-like effects of constitutive expression of G₁/S cyclin (CycD3) in cell cultures of arabidopsis. In CycD3 transgenics, healthy calli can be induced and maintained for extended periods in the absence of exogenous cytokinin. These calli, however, were unable to regenerate shoots. As mentioned previously, both Spcdc25 expression and cytokinin treatment resulted in dephosphorylation of plant CDK in suspension cultures of Nicotiana plumbaginifolia (Zhang et al., 1996). This led to the hypothesis that ‘plant-like’ Cdc25 may be regulated by a cytokinin-mediated signalling pathway (John, 1996). Notably, when zeatin synthesis is partially inhibited by lovastatin (an inhibitor of mevalonic acid synthesis) in tobacco BY-2 cultures, cells are prevented from undergoing mitosis (Laureys et al., 1998), but tobacco BY-2 cells expressing Spcdc25 can overcome this block (Orchard et al., 2005). Thus, it is possible to suppose, that Spcdc25 compensates for the absence of a cytokinin-mediated dephosphorylation of a CDK at G₂/M, which then creates the mitotic clusters necessary for the formation of shoots. However, there was no apparent effect of Spcdc25 on cell size in meristematic cells in the culture explants (P. Mašková et al., unpubl. res.). This is similar to the effect of ectopic expression of Spcdc25 in leaves of tobacco; there were alterations in the number of cell layers in the leaf margin but there was no pronounced effect on cell size (Wyrzykowska et al., 2002). Nevertheless, mitotic cell size was smaller in root meristems of Spcdc25-expressing tobacco plants (Bell et al., 1993; McKibbin et al., 1998), and tobacco BY-2 cells (Orchard et al., 2005). Hence, we propose that fission yeast mitotic activatory phosphatase Spcdc25 can activate plant cells into mitosis at a reduced cell size, and thus substitute for the cytokinin treatment that would normally promote cell-cycle progression (e.g. Zhang et al., 1996) and subsequently influence cytokinin-dependent organogenesis.

Fig. 3.

Spcdc25-expressing stem segments form shoots on hormone-free medium (Murashige and Skoog medium without growth regulator). WT, stem segments of ‘Samsun’ tobacco; Spcdc25 plants, stem segments of tobacco plants expressing the Spcdc25 gene. For details of other characteristics of plant material, including determination of Spcdc25 expression by RT-PCR, see Suchomelová et al. (2004).

Spcdc25 expression restricts root development

In experiments studying de novo organ formation on Spcdc25 tobacco stem segments (with verified gene expression), Suchomelová et al. (2004) observed not only an enhanced tendency to shoot formation, but also substantially reduced root organogenesis. On the medium with a high auxin : cytokinin ratio, where plentiful root formation was induced in WT, the Spcdc25 segments scarcely formed any roots, and even some shoot formation was observed. In the field experiments designed to follow flowering onset in Spcdc25 plants, a restriction on the extent of root formation was found in Spcdc25 transgenics as well (P. Vojvodová et al., unpubl. res.), similar to Spcdc25 plants derived from nodal stem segments under in vitro conditions (Teichmanová et al., 2007). In previous sections, we have pointed out that Spcdc25 expression consistently resulted in cytokinin-like effects on the phenotypes studied. Kyozuka (2007) stated that, in contrast to a positive role of cytokinins on proliferation of the SAM cells, an opposite effect has been observed on root growth. In arabidopsis and tobacco with reduced cytokinin levels, Werner et al. (2003) found a promotion of root development. The study, using intact seed-derived Spcdc25 plants grown in Petri dishes, aimed to determine a component of the root development primarily responsible for Spcdc25-induced root restriction, revealed that the lengths of both the main and lateral roots were significantly reduced, together with a reduction in the total number of lateral roots formed (K. Čiháková and H. Lipavská, unpubl. res.). Root tips exhibited a shortened meristem region and it was also observed that the cells belonging to already differentiated tissues were shortened (K. Čiháková and H. Lipavská, unpubl. res.). Interestingly, Bell et al. (1993) showed, in root systems of tobacco, that Spcdc25 induces a smaller mitotic cell size. Dello Ioio et al. (2007) followed the changes resulting from reduced cytokinin levels in an arabidopsis ipt3/ipt5/ipt7 triple mutant and reported an increase in root meristem size. Based on a detailed study of the phenomenon, they proposed that the rate of differentiation of meristematic cells was reduced in cytokinin-deficient roots and that cytokinins operate at the transition zone – the border between the root meristem and the elongation zone. Taken together, we can propose that Spcdc25 expression causes changes in CDK activities that might be interpreted by the root cells as increased cytokinin availability and thus result in root system characteristics opposite to those found in cytokinin-deficient plant material (Werner et al., 2003; Dello Ioio et al., 2007).

McKibbin et al. (1998) found an increase in the frequency of lateral root per primary root length in tobacco with Spcdc25 under an inducible promoter. Lateral root primordia are initiated from the pericycle close to the primary apical meristem and immediately adjacent to protoxylem poles (Dubrovsky et al., 2001). Formative divisions are transverse as large founder cells divide into smaller descendants. In the next division, the plane of cell division changes from transverse to longitudinal so that these youngest primordia appear as a double filament of near-isodiametric cells (Dubrovsky et al., 2001). In appearance, they resemble the double filaments observed in the Spcdc25-expressing BY-2 lines (Orchard et al., 2005) or ‘Samsun’ tobacco cell cultures (Suchomelová-Mašková et al., 2008).

In Spcdc25 tobacco, a shift in carbohydrate allocation in favour of shoots was observed repeatedly (M. Teichmanová et al., Department of Experimental Plant Biology, Faculty of Science, Charles University, Czech Republic, unpubl. res.; K. Čiháková and H. Lipavská, unpubl. res.). Therefore, the possibility was examined that root reduction in intact plants is preferentially caused by a shortage of carbohydrate supply or other changes taking place in shoots. Examination of root cultures in vitro, however, showed similar characteristics of isolated roots to those obtained with intact plants, thus enabling us to exclude a substantial influence of the shoot on root development (K. Čiháková and H. Lipavská, unpubl. res.).

Collectively, the results of characterization of Spcdc25 tobacco root-system development show once more the cytokinin-like effect of Spcdc25 expression, indicating that cytokinin function in root meristem control might, at least partly, be performed via an effect on cell-cycle control at the G₂/M transition.

Spcdc25 expression and cell suspension characteristics

Changes in cell characteristics resulting from Spcdc25 expression were investigated in the studies using cell suspension cultures: Spcdc25-expressing BY-2 (referred hereafter as Spcdc25 BY-2 cells) and cell cultures derived from Spcdc25-transformed tobacco ‘Samsun’ (referred hereafter as Samsun Spcdc25 cells) were used as models.

Spcdc25 expression shortens G₂ in BY-2 cells

Tobacco BY-2 cells (Nagata et al., 1992) expressing Spcdc25 were used as a tool to investigate the effect of yeast mitotic activator on the length of cell cycle phases (Orchard et al., 2005). The results indicated that Spcdc25 expression shortens G₂, induces a smaller cell size but did not affect mitotic timing. In fission yeast, overexpression of Spcdc25 induced a reduction in mitotic cell size. However, it is generally over-looked that total cell cycle duration was very similar to that in WT as prolongation of G₁ compensated for the shorter G₂ (Russell and Nurse, 1986). In the BY-2 cell line, Spcdc25 induces small mitotic cells owing to a 50 % reduction in G₂ phase (Orchard et al., 2005). As in yeast, there was only a slight reduction in the duration of the cell cycle thanks to G₁ compensatory lengthening. This mechanism was proposed in previous work (McKibbin et al., 1998) where the authors demonstrated increased numbers of lateral root primordia composed of smaller cells in Spcdc25 tobacco plants.

Premature entry into mitosis is preceded by earlier CDK activitiy in Spcdc25 BY-2 cells

In the Spcdc25 BY-2 cells, constantly high CDKB1 kinase activity was detected in early S-phase, S/G₂ and early M (Orchard et al., 2005), while CDKA activity was relatively stable before peaking in M + G₁. The B-type CDK is unique to plants, while CDKA, as it was shown in arabidopsis, has the strongly conserved PSTAIRE motif also found in yeast and animal CDKAs (Joubès et al., 2000). In WT BY-2 cells, unlike CDKA, the highest CDKB1 activity is exclusively at the G₂/M transition (Porccedu et al., 2001; Sorrell et al., 2001). Thus, premature cell division at a small mitotic size in Spcdc25 BY-2 mentioned above is most likely the result of the premature and sustained CDKB1 activity.

Spcdc25 BY-2 cells overcome a cytokinin requirement at G₂/M

Wild-type BY-2 cells are dependent on zeatin and zeatin riboside synthesis at the G₂/M transition. This was proved by suppression of the isoprenoid pathway of cytokinin biosynthesis by lovastatin (Crowell and Salaz, 1992) that resulted in blocking the G₂/M transition (Redig et al., 1996; Laureys et al., 1998). However, the Spcdc25 BY-2 cell suspension cultures were able to continue cell division even in the presence of lovastatin (Orchard et al., 2005), overcoming or perhaps compensating for depleted cytokinin at the G₂/M transition. There are several pieces of evidence for a cytokinin-regulated G₂/M transition in the plant cell cycle: lovastatin blocks cytokinin biosynthesis and prevents cells from proceeding from G₂ to M (Redig et al., 1996; Laureys et al., 1998); cytokinin treatment activates the G₂/M transition (Zhang et al., 1996); and either Spcdc25 or a cytokinin treatment can dephosphorylate plant CDK (Zhang et al., 1996, 2005). In our opinion, the data showing that Spcdc25-expressing BY-2 cells can traverse from G₂ to M, regardless of a lovastatin block, represent convincing evidence for cytokinin-regulated CDK-activated dephosphorylation at G₂/M transition in plants.

Spcdc25-expressing suspension cultures consist of isodiametric cells with less-regular organization

Given that Spcdc25 expression leads to G₂ phase shortening, the premature entry into mitosis could cause a chaotic organization of cell division. This assumption was proved in the study of Samsun Spcdc25 cell suspensions (Suchomelová-Mašková et al., 2008), where determination of cell shape by measurement of circularity or elongation parameters (lower circularity resp. higher elongation of Spcdc25 cells) revealed significant alteration, arguing for changed mitotic cell planes. A similar phenotypic change was observed in fission yeast wee1 − mutants that formed very short, almost round cells (Sipiczki et al., 2000). Moreover, the Samsun Spcdc25 cells were arranged in clusters or, less frequently, chains of cell doublets (Fig. 4; Suchomelová-Mašková et al., 2008). The findings are fully in accordance with the results obtained by Orchard et al. (2005) in Spcdc25 BY-2 cells where sister filaments were documented. Note that WT cells of both BY-2 and Samsun cell cultures typically grow in long single filaments of oblong cells arising by transverse divisions.

Fig. 4.

Disorganization of cells in Spcdc25-expressing cell suspension culture shown in microphotographs under UV light in combination with Nomarski differential contrast. WT, cell suspension culture derived from ‘Samsun’ tobacco; Spcdc25 Samsun, cell suspension culture derived from ‘Samsun’ tobacco expressing the Spcdc25 gene. For details of cell-suspension culture preparation and other characteristics of the plant material e.g. determination of cell shape, see Suchomelová-Mašková et al. (2008).

In contrast to Spcdc25 BY-2 (Orchard et al., 2005), however, the Samsun Spcdc25 cells did not show smaller cells compared with WT (Suchomelová-Mašková et al., 2008). One possible explanation for this disparity could be that G₁-phase compensatory prolongation is deepened in Samsun Spcdc25 cells and therefore the overall cell cycle duration is similar to WT, as in the case of fission yeast overexpressing cdc25 (Russell and Nurse, 1986). Interestingly, Sorrell et al. (1999) proposed that different results for the D-cyclin expression profile in BY-2 cells compared with other plant species could be connected to specific properties of this culture, caused by its longevity and ‘immortality’ related to changed phenotype/genotype. Thus, the results achieved with the model BY-2 system, could in some aspects differ from those obtained with other tobacco cell cultures/lines. Nevertheless, the common features of both types of culture were more circular shapes of cells and formation of double chains or cell clusters, indicating the disturbance of the planes of cell divisions induced by yeast mitotic phosphatase Spcdc25 expression.

Cytokinin-treated Samsun WT cells exhibit a ‘Spcdc25-like’ phenotype

The cell morphology resembling that found in Samsun Spcdc25 cultures is inducible in WT with cytokinin application. The Samsun WT cells (note that WT cell cultures are exogenous-cytokinin independent) treated with BAP (6-benzylaminopurine) failed to form chains and were instead arranged in clusters or chains of cell doublets (Suchomelová-Mašková et al., 2008). Petrášek (1995) found out that BAP application to cytokinin-independent tobacco culture VBI-0 changed the filamentous phenotype to a spherical one. The effect has been ascribed to an influence of BAP on the cytoskeleton, especially microtubules (Shibaoka, 1994). In the context of contemporary models of cytokinin-induced regulation of G₂/M transition, other explanations for the observed phenomena in VBI-0 cells is at hand. BAP-treated tobacco cells could be driven faster into mitosis, which could result in ill-established cell division planes. Moreover, numerous data show a tight relationship of cell cycle machinery with the cytoskeleton (e.g. Hush et al., 1996; Weingartner et al., 2001, 2003, 2004) thus pointing to the possibility that cytoskeletal changes can be induced by cytokinins via their effects on cell-cycle regulation.

Most importantly, the low cytokinin treatments (0·1–1 mg BAP L⁻¹), which stimulated growth in WT cultures, were sensed by the Samsun Spcdc25 cells as inhibitory (presumably like a ‘cytokinin overdose’). Under those conditions, biomass accumulation of Samsun Spcdc25 cell cultures was lowered and giant clusters of cells with markedly reduced viability were formed resembling WT cultures under high (5 mg BAP L⁻¹) cytokinin treatments (Suchomelová-Mašková et al., 2008). This again shows that Spcdc25 expression in cultured tobacco cells and cytokinin application to WT cells are of a similar and additive nature, which points once more to a cytokinin-like effect of Spcdc25 expression.

Carbohydrate status of Spcdc25 Samsun cell cultures is inducible in WT with cytokinins

In addition to previously discussed features, the Samsun Spcdc25 cells exhibited different carbohydrate contents and spectra in comparison to WT (Suchomelová-Mašková et al., 2008). While a wealth of starch bodies was observed in transgenic cells, virtually no starch was detected in WT (Fig. 5). As carbohydrate metabolism functions as a network of negative- and positive-feedback signalling pathways, precisely regulating in particular the balance between soluble carbohydrates and starch levels (e.g. Munoz et al., 2006), it is not surprising that the levels of soluble carbohydrates differed in Samsun Spcdc25 and WT cells as well, exhibiting a higher total sugar content with a significant increase in the ratio of sucrose : hexoses in transgenics.

Fig. 5.

Starch (visualized by Lugol staining in native samples) deposition in tobacco cells induced by Spcdc25 expression. WT, cell suspension culture derived from ‘Samsun’ tobacco; Spcdc25 Samsun, cell suspension culture derived from ‘Samsun’ tobacco expressing the Spcdc25 gene. For details of other characteristics, e.g. soluble carbohydrate levels and spectra, see Suchomelová-Mašková et al. (2008).

It is well known that there exists an interrelation between cytokinin levels and carbohydrate status of a plant (e.g. Gibson, 2004; Brenner et al., 2005; Franco-Zorrilla et al., 2005). In cytokinin-independent BY-2 cell cultures, Miyazawa et al. (1999, 2002) observed that cytokinin application as well as auxin removal from the medium resulted in amyloplast accumulation. Cytokinin treatment enhanced the expression of plasma-membrane hexose-uptake carriers in Chenopodium rubrum cell suspension cultures and consequently improved carbohydrate availability (Balibrea et al., 2004). Clearly, in WT, the increased starch deposition as well as higher soluble carbohydrate levels induced by cytokinin application can be easily explained, based on published data. In contrast, an important question arises in trying to explain the similar changes induced by Spcdc25 transformation.

Spcdc25 expression results in a marked reduction in endogenous cytokinin levels

Naturally, one can speculate that Spcdc25 expression might interact with the tobacco genome in an unspecified way that indirectly enhances endogenous cytokinin levels that are finally responsible for cytokinin-like effects observed in Spcdc25 transgenics. Suchomelová et al. (2004) proposed that Spcdc25 acts as mitotic inducer simulating cytokinin activation of plant ‘cdc25-like’ phosphatase, and thus having an impact on cell division-dependent processes resembling cytokinin application. Further, providing that Spcdc25 phosphatase operates downstream of the cytokinin action (Zhang et al., 2005), constitutive Spcdc25 expression could be sensed by the plant as high cytokinin availability and as a consequence a compensatory decline in cytokinin levels might be expected (Suchomelová et al., 2004). The hypothesis was supported by the results obtained with Spcdc25-transformed BY-2 cells (Orchard et al., 2005) and Samsun Spcdc25 cell cultures (Suchomelová-Mašková et al., 2008). In both cases, measurements of endogenous cytokinin contents revealed a marked reduction in total cytokinins resulting from Spcdc25 expression. The Samsun Spcdc25 cells had significantly lower levels of cis-zeatin derivatives. Minor elevation of dihydrozeatin-type cytokinins in the transformant was fully compensated by elevated content of isopentenyladenine-type cytokinins in WT. Thus, the results verified the original hypothesis and correspond well with the cell cycle regulation model featuring cytokinins as active upstream of the plant ‘cdc25-like’ phosphatase in CDK dephosphorylation (Zhang et al., 2005). If this model is plausible, it is easy to understand that Spcdc25 expression-induced changes in cell division are manifested by the altered morphological characteristics presented in previous sections. In contrast, it is no help in elucidating the cytokinin-like nature of carbohydrate changes induced by Spcdc25 expression.

How can cytokinin-like changes in carbohydrate metabolism in low-cytokinin-containing Spcdc25 transgenics be explained?

In the light of the results in determining cytokinin levels, we have to abandon the explanation based on direct cytokinin impact on carbohydrate contents and spectra in Spcdc25 transgenics, which forces us to search for other explanations. It is now well documented that sugars contribute to the decision of the cell to continue in cell-cycle progression (e.g. via positive modulation of D-cyclin levels, and the negative regulation of inhibitor levels; e.g. Riou-Khamlichi et al., 2000; Healy et al., 2001; Richard et al., 2002). In connection with the results on carbohydrate status in Samsun Spcdc25 cells (Suchomelová-Mašková et al., 2008), we propose that it is not only a one-way relationship, i.e. sugar level is a component of the system controlling cell-cycle progression, but instead it also operates in an opposite direction. That also means that the decision to continue in cell-cycle progression, and especially to enter mitosis, produces a signal(s) aimed at preparing the cell for subsequent high carbon- and energy-demanding mitosis and cell division. The resulting metabolic changes might also include, besides others, changes in carbohydrate status. In green algae, like Chlorella vulgaris and Chlamydomonas reinhardtii, starch is synthesized through the cell cycle, and these reserves are then exhausted to meet the demands of nuclear and cellular divisions. Starch mobilization happens even in continuous light when enough soluble carbohydrates are available (V. Zachleder et al., Institute of Microbiology, Academy of Sciences, Czech Republic, unpubl. res.).

Hence, we propose that a higher proportion of active CDK in plants that express mitotic activatory Spcdc25 phosphatase in the original organism, fission yeast, functioning as an all-or-nothing signal driving cells into mitosis, resulted in changed signals to metabolic pathways, including carbohydrate metabolism. To support this hypothesis, a detailed study of CDK phosphorylation targets would be necessary. However, the question arises whether the enzymes/proteins/compounds involved in regulation of carbohydrate balance are the direct substrates of CDK or whether the control is more indirect. Some encouragement for the concept of CDK control of carbohydrate metabolism comes from an analysis of cell-cycle-related-protein complexes in A. thaliana (Van Leene et al., 2007) showing, besides others, interactions between CDKA;1 and components of carbohydrate metabolism (Rubisco subunit binding-protein and, especially, with cytoplasmic phosphoglucomutase). In addition, Geelen et al. (2007) found an interaction between trehalose-6-P-synthase protein and CDKA. In both cases, the data are interpreted as the mechanisms through which carbohydrate status modifies cell-cycle progression. We offer an alternative interpretation: the CDKA interaction with the above-mentioned proteins modulates carbohydrate metabolism in response to cell-cycle progression. Most interestingly, the unique study of CDK substrates in yeast identified proteins involved in saccharide biosynthesis and transport that are regulated by CDK activity (Ubersax et al., 2003). So, it is tempting to speculate that some of the published data on the influence of cytokinins on carbohydrate metabolism might in fact refer to carbohydrate-status modulations induced by cytokinins via cell-cycle control machinery. Clearly, more effort is required to resolve this issue.

CONCLUSIONS

Although a lot of work is needed to answer all the questions raised in this review, we think that summarizing the data achieved up to now, including, especially, those obtained with the mitotic-inducer Spcdc25-expressing tobacco, allows us to come to the following conclusion. CDKs in higher plants are most probably, as in other eukaryotes, negatively controlled by phosphorylation at G₂/M; as the presence of the highly conserved phosphoregulatory residues T14 and Y15 in plant CDKs was verified, their phosphorylation under cytokinin deprivation and stress conditions has been repeatedly detected and plant inhibitory Wee1 kinase orthologues have been identified. In addition, the alien Spcdc25 phosphatase recognizes plant CDK(s) as its substrate and dephosphorylates it, which leads to the enhanced activity of mitotic CDK compared with WT. Consequently, signalling pathways downstream of CDKs are also more active, drive cells prematurely into mitosis and, finally, influence the processes directly dependent on cell division in a cytokinin-mimicking way. This status might be sensed by the cell or whole plant as increased cytokinin availability, which afterwards results in a compensatory reduction in cytokinin levels. Furthermore, we propose that enhanced activity of CDK(s) at G₂/M might generate a signal modifying carbohydrate metabolism so that it can meet the requirements of the forthcoming cell division. Clearly, the results discussed bring more insight to the plant cell-cycle regulation, but they open new questions about complexity and interconnection of regulatory networks, e.g. between the cell cycle and carbohydrate metabolism in plants, that need further investigation.

For plants, the existence of regulatory networks with weaker hierarchy, compared with other eukaryotes, scarceness of control points of a true all-or-nothing nature, operations of parallel metabolic pathways or alternative solutions to most life situations are characteristic. In other words, ‘plants hardly ever hang anything on one nail’. Therefore, we propose that, even in the regulation of the cell cycle progression, the strength of controlling mechanisms, including activatory dephosphorylation at G₂/M, might be weakened in favour of greater flexibility and adaptability. Nevertheless, based on the results discussed above, we suggest that, although activatory dephosphorylation might lack the ability to govern G₂/M transition with the strictness typical of other eukaryotes, and so it might be ‘only’ one of the instruments contributing to the decision, it still represents an important part of the control mechanisms of G₂/M transition in plants.