|Home | About | Journals | Submit | Contact Us | Français|
Endoreduplication is the major source of endopolyploidy in higher plants. The process of endoreduplication results from the ability of cells to modify their classical cell cycle into a partial cell cycle where DNA synthesis occurs independently from mitosis. Despite the ubiquitous occurrence of the phenomenon in eukaryotic cells, the physiological meaning of endoreduplication remains vague,although several roles during plant development have been proposed, mostly related to cell differentiation and cell size determination.
Here recent advances in the knowledge of endoreduplication and fruit organogenesis are reviewed, focusing on tomato (Solanum lycopersicum) as a model, and the functional analyses of endoreduplication-associated regulatory genes in tomato fruit are described.
The cyclin-dependent kinase inhibitory kinase WEE1 and the anaphase promoting complex activator CCS52A both participate in the control of cell size and the endoreduplication process driving cell expansion during early fruit development in tomato. Moreover the fruit-specific functional analysis of the tomato CDK inhibitor KRP1 reveals that cell size and fruit size determination can be uncoupled from DNA ploidy levels, indicating that endoreduplication acts rather as a limiting factor for cell growth. The overall functional data contribute to unravelling the physiological role of endoreduplication in growth induction of fleshy fruits.
Plant growth is an important process, providing food, and natural resources for industry, chemicals, pharmaceuticals and renewable energy. In the present worldwide socio-economical context, plant biologists are facing challenges related to food and human nutrition supply, plant adaptation to occurring climate change, exhaustion of fossil resources, and are thus expected to produce fundamental knowledge to be transferred to the societal and economical demands.
Plant growth and development involve fundamental cellular processes such as cell division, cell expansion and cell differentiation, impacting plant yield and, consequently, the quality of plant products. Therefore the development of our knowledge on the plant cell cycle leading to mitotic cell division has been a major issue for the last two decades and has allowed the understanding of how the progression within the different stages of the cell cycle is regulated, through the characterization of many key regulators such as the CDKs (cyclin-dependent kinases), the cyclins and other regulatory elements (Inzé and De Veylder, 2006). During plant development the canonical cell cycle can be replaced by an altered cell cycle where mitosis is bypassed. This modified cell cycle called the endoreduplication cycle or endocycle consists of one or several rounds of DNA synthesis in the absence of mitosis, and participates in the control of the cell expansion process, thus contributing to plant organ growth and thus plant yield. Endoreduplication in plants can occur in many cell types, especially those undergoing differentiation, putatively related with peculiar metabolic properties (Joubès and Chevalier, 2000; Kondorosi and Kondorosi, 2004; Sabelli and Larkins, 2009). Although accumulating data have provided important insight into the molecular effectors triggering the exit of the classical cell cycle towards the onset of the endoreduplication cycle, it is still poorly understood in plants and its physiological role remains unclear. Hence understanding the functional role of endoreduplication during plant development has become a major field of interest in plant biology, as it potentially impacts plant yield.
The endoreduplication process in tomato (Solanum lycopersicum) is of considerable significance, not only because high levels of endopolyploidy occur in the course of fruit development (Bergervoet et al. 1996; Joubès et al., 1999; Cheniclet et al., 2005; Bertin et al., 2007) which attain unmatched values by other model plants such as Arabidopsis, maize or Medicago (Melaragno et al., 1993; Vilhar et al., 2002; Kondorosi et al., 2005), but most importantly because it does contribute to fruit growth in a developmentally and/or genetically regulated manner (Cheniclet et al., 2005; Chevalier, 2007).
As a major fruit produced in the world, tomato has been established as the model species for all fleshy fruits, displaying both agronomical and scientific advantages. Tomato displays a highly favourable biology with a short life cycle, high multiplication rate, easy crosses and self-pollination and a wide range of genetic resources (cultivars, mutants, segregating populations, reverse genetic tools based on stable or transient transformation protocols) for analysing the function of candidate genes. Furthermore it has now entered into the post-genomic era as the draft assembled genome sequence of Solanum lycopersicum (‘Heinz 1706’) was recently released (http://solgenomics.net/genomes/Solanum_lycopersicum/index.pl).
This review aims at describing the recent advances in the knowledge of endoreduplication and fruit organogenesis, focusing on tomato as a model.
The fruit is a specialized organ which results from the development of the ovary after successful flower pollination and fertilization, and provides a suitable environment for seed maturation and seed dispersal mechanisms.
Tomato fruit organogenesis results from the relationship between cell division and cell expansion which determines the cell number and the relative cell size inside fruit, respectively (Bohner and Bangerth, 1988). These two developmental phenomena which are under the control of complex interactions between internal signals (due to hormones) and external factors (carbon partitioning, environmental influences) represent crucial determinants of essential criteria for morphological fruit quality traits such as the final size, weight and shape of fruits (Tanksley, 2004). In addition, organoleptic and nutritional quality traits of tomato ripe fruit relevant to composition in primary and secondary metabolites are also determined early during fruit development.
The development of tomato fruit was classically described as proceeding in four distinct phases: fruit set (I), a phase of intense cell divisions (II) and a phase of cell expansion (III) both contributing to fruit growth, and finally ripening (IV) (Gillaspy et al., 1993). However, this classical developmental scheme for tomato fruit is mostly a short cut, since the distribution of mitotic activities (Joubès et al., 1999) and occurrence of cell expansion (Cheniclet et al., 2005) are spatially and temporally regulated according to the various fruit tissues.
Upon completion of pollination and fertilization, the presence of fertilized ovules triggers very rapidly (within 2 d post-anthesis, dpa) the development of the ovary into a fleshy pericarp, encompassing the placental tissue and the seeds. This gives rise to the tomato fruit berry during a phase of rapid growth due to high mitotic activity which lasts for about 7–10 d. Thereafter and until ripening, fruit growth to almost full size is principally obtained through cell expansion. Cell size in tomato fruit can reach spectacular levels, e.g. >0·5 mm in diameter inside the pericarp according to the variety (Cheniclet et al., 2005). These very large cells are found within the central part of the fleshy pericarp and the locular (jelly-like) tissue. Both tissues are characterized by the arrest of mitotic activity during the third phase of growth of tomato fruit development, namely the phase of cell expansion (Joubès et al., 1999). The loss of mitotic activity is concomitant with endopolyploidy, as a spectacular increase in DNA ploidy levels leading to DNA contents up to 256C or even 512C and a consequent hypertrophy of the nucleus are observed in tomato fruit tissues (Bergervoert et al., 1996; Joubès et al., 1999; Cheniclet et al., 2005; Bourdon et al., 2010).
Various mechanisms can account for the occurrence of endopolyploidy in plants: it may originate from the generation of multinucleate cells originating from acytokinetic mitosis, from nuclear fusion, from endomitosis or from endoreduplication. Evidence that endoreduplication per se is the main mode of cell endopolyploidization in tomato has been provided. Using a fluorescence in situ hybridization approach on pericarp nuclei sorted by flow cytometry according to their DNA content and chromosome-specific probes, it was demonstrated that endopolyploidization in tomato fruit tissues does not lead to a doubling of the chromosome number in the nucleus as expected for endomitosis, but to endoreduplication producing chromosomes with 2n chromatids without any change in chromosome number (Bourdon et al., 2010).
During tomato fruit development, clear positive correlations have been established between the mean cell size within the fruit pericarp and the mean ploidy level of various tomato genotypes, and between the mean ploidy level and final fruit size in genotypes displaying two to three carpellar locules (Cheniclet et al., 2005). Therefore endoreduplication is a major determinant for the final size of the cell, which can explain the observed gradation in cell size in tomato fruit but contributes also in part to the variation in fruit size.
Fruit size/weight in tomato is genetically under the control of nearly 30 quantitative trait loci (QTLs) (Grandillo et al., 1999). The QTL named fw2·2 (for fresh weight locus no. 2 on chromosome 2) encodes a single gene (FW2·2) that is to date the only fruit size-controlling locus to be cloned and characterized at the molecular level (Frary et al., 2000). Its quantitative effect on fruit size is outstanding as it can explain, on its own, as much as a 30 % difference in fruit fresh weight between the domesticated tomato and its wild relatives (Alpert et al., 1995). At the level of gene expression, the timing of FW2·2 transcription (heterochronic changes) and the overall quantity of transcripts account for the quantitative effect on fruit size between small and large fruits (Cong et al., 2002). FW2·2 is a plant-specific and fruit-specific protein that regulates negatively the mitotic activity of the developing fruit and thus results in a modulation of final fruit size (Frary et al., 2000; Cong et al., 2002). The nature of FW2·2 and its precise biological function in controlling cell division remain elusive. However, as a transmembrane protein, FW2·2 appears to be part of a cell-cycle control signal transduction pathway (Cong and Tanksley, 2006), operating early after fruit set and upstream of the endoreduplication onset. Whether it participates in the signal transduction chain triggering endoreduplication, as an inhibitor of cell division, is unknown
As it is widely occurring in angiosperms, the endoreduplication process is likely to have been selected during evolution to the benefit of plant and organ development. According to the different types of situation encountered in various plant species, in relation to plant-, organ- or cell physiology, several functional roles – not exclusive to each other – were reported to explain tentatively the relevance of endoreduplication.
Endoreduplication might contribute to the adaptation to adverse environmental factors, allowing the maintenance of growth under stress conditions. The extended amplification of nuclear DNA may provide a means to protect the genome from DNA-damaging conditions such as UV damage or prevent uneven chromosome segregation during mitosis. For instance, exposure to UV-B induces endoreduplication in epidermal cells surrounding trichomes in cucumber cotyledons (Yamasaki et al., 2010), and the UV-B-insensitive 4 mutation (uvi4) in arabidopsis stimulates endoreduplication during hypocotyl growth and leaf development (Hase et al., 2006). In addition, it was found that endoreduplication may confer an increased tolerance to UV-B exposure, since tetraploid arabidopsis was hyper-resistant to UV-B compared with diploid arabidopsis (Hase et al., 2006). Endoreduplication is also involved in the adaptation to high salt concentration (Ceccarelli et al., 2006), to water deficit (Cookson et al., 2006) and low temperatures (Barow, 2006), thus suggesting that an increase in DNA content can be of advantage to cope with an adverse environment.
Endoreduplication often occurs during the differentiation of cells that are highly specialized in their morphology and its blockage by mutation results in developmental abnormalities (Kondorosi et al., 2000, 2005; Edgar and Orr-Weaver, 2001). In arabidopsis, the influence of endoreduplication on cell growth was best characterized in epidermal cells of mature leaves (Melaragno et al., 1993), during hypocotyl development in which the ploidy levels vary according to light conditions (Gendreau et al., 1997), and in leaf single-celled trichomes according to a genetically regulated process (Hulskamp et al., 1999; Larkin et al., 2007). The growth of trichomes was shown indeed to be dependent on the succession of endocycles. The formation of a two-branched trichome cell requires three rounds of endocycle, leading to a 16C DNA ploidy level, and a supplementary endocycle may eventually occur to give rise to the formation of a third branch and 32C DNA content.
A new and original role for endoreduplication has been recently reported in arabidopsis trichomes. Bramsiepe et al. (2010) demonstrated very elegantly that endoreduplication is an important determinant for cell fate, as they managed to change trichome fate into an epidermal pavement cell fate even in already advanced stages of trichome differentiation by compromising endoreduplication. Conversely they could restore the trichome fate in a patterning mutant by promoting endoreduplication.
As illustrated for trichomes, endoreduplication often occurs during the differentiation of cells that are highly specialized in their morphology. The influence of endoreduplication on the differentiation of metabolically specialized cells was also reported. For instance the highly polyploid endosperm cells of maize kernels accumulate large amounts of starch and storage proteins (Kowles et al., 1990; Sabelli and Larkins, 2009). During the formation of nitrogen-fixing root nodules in legumes as a response to interaction with the symbiotic bacterium Sinorhizobium meliloti, the symbiotic nodule cells hosting the rhizobia and programmed to fix nitrogen develop into very large and highly endoreduplicated cells during their differentiation process (Cebolla et al., 1999; Vinardell et al., 2003; González-Sama et al., 2006). Interestingly this endoreduplication-associated differentiation is tightly linked to an important transcriptional activity that is remarkably specific to the nodule (Mergaert et al., 2003). As far as tomato fruit is concerned, only descriptive analyses for transcriptional and metabolic profiling in cells from various tomato fruit tissues were provided (Lemaire-Chamley et al., 2005; Schauer et al., 2006; Mounet et al., 2009; Steinhauser et al., 2010), without any direct link with the extent of endoreduplication.
Since a correlation can be found between endoreduplication and cell differentiation-specific metabolism, it is commonly stated that endoreduplication may contribute to modulate transcriptional activity by increasing the availability of DNA templates for gene expression as the gene copy number is obviously multiplied, and therefore may contribute to modulate subsequent translational and metabolic activities. Although this is true for animal cells (Hu and Cross, 2010) or yeasts (Galitski et al., 1999), the physiological role of endoreduplication in stimulating transcriptional activity has never been convincingly demonstrated or disproven in plant cells. For example, altering the DNA ploidy levels has no clear impact on the expression level of some endosperm-specific genes in maize (Leiva-Neto et al., 2004), thus making endoreduplication in maize endosperm a more likely mechanism to provide a store of nitrogen and nucleotides during embryogenesis and/or germination.
The most unanimously accepted functional role of endoreduplication relates to cell- and organ-size determination, since endoreduplication and cell size in many different plant species, organs and cell types are naturally and intimately positively correlated (Joubès and Chevalier, 2000; Sugimoto-Shirasu and Roberts, 2003; Kondorosi and Kondorosi, 2004). As a result of successive rounds of DNA synthesis during endoreduplication, nuclei become hypertrophied which in turn influences the final size of a cell which therefore adjusts its cytoplasmic volume with respect to the nuclear DNA content (according to the ‘karyoplasmic ratio’ theory; Sugimoto-Shirasu and Roberts, 2003). Therefore endoreduplication is a likely driver for cell expansion. Such a positive correlation between cell size and ploidy level in developing tomato fruit tissue was indeed demonstrated (Cheniclet et al., 2005). However, the ability to form large cells is not fully restricted to endoreduplicating cells. For instance, amongst the various species of fleshy fruits, cell sizes of approx. 200 µm, 220 µm and 350 µm in diameter can be observed in kiwi, persimmon and grape, respectively, while endoreduplication never occurs in these three species. These cell sizes are quite comparable to the sizes measured for the largest cells present in fruits which undergo high numbers of endocycle rounds (diameters of >200 µm in cucumber, 450 µm in melon, 600–1000 µm in tomato; Bourdon et al., 2010). Conversely smaller cell diameters can be associated to endoreduplicating cells of strawberry (50 µm), cherry (65 µm) or peach (120 µm). Nevertheless it appears as a rule of thumb that the largest cells are always present in fruits which undergo the highest number of endocycles, which suggests that polyploidy via endoreduplication might be necessary for plant cells to reach very large sizes.
Cheniclet et al. (2005) reported that the level of endoreduplication is tightly correlated with final fruit size in tomato, and therefore endoreduplication could participate in modulating the rate of organ growth and/or cell expansion.
In a recent analysis (Bourdon et al. 2010), it was reported that endoreduplication always occurs in fleshy fruits which develop rapidly (in <13 weeks) comprising three to eight rounds of endocycle, in particular in the Solanaceae and Cucurbitaceae species analysed so far. With the exception of some Rosaceae species (apricot, peach and plum), endoreduplication does not occur in most of the species where fruit development lasts for a very long period of time (over 14 weeks; Fig. 1). It was thus concluded that endoreduplication does indeed influence the fruit growth rate, most probably at the level of the cell expansion rate.
The endoreduplication cycle (endocycle) corresponds to a truncated variation of the canonical eukaryotic cell cycle where the mitosis phase is aborted thus accounting for the cessation of cell division. As a consequence the endocycle is only made of the succession of an undifferentiated G phase and the S phase for DNA synthesis resulting in an exponential increase in ploidy level (Joubès and Chevalier, 2000; Edgar and Orr-Weaver, 2001; Vlieghe et al., 2007). Therefore some of the existing molecular controls for the classical cell cycle regulation are conserved in the endocycle, especially at the G → S phase and S → G transitions.
The progression within the distinct phases of the plant canonical cell cycle and the endocycle requires the activity of a class of conserved heterodimeric protein complexes consisting of a catalytic subunit referred to as cyclin-dependent kinase (CDK) and a regulatory cyclin (CYC) subunit. Within the complex, there is an absolute need of the CYC moiety for the CDK activity, not only for the activation of the phosphorylation activity on protein targets, but also for the specification of the substrates, and the stability and subcellular localization of the complex itself (Inzé and De Veylder, 2006). At the boundary between the G1 and the S phases, the canonical A-type CDK (CDKA; Joubès et al., 2000a) harbouring the PSTAIRE hallmark in the cyclin-binding domain plays a pivotal role during the cell cycle in association with D-type cyclins allowing the commitment to DNA replication. In the course of the DNA replication process of the endocycle, it is highly probable that CDKA also participates in a different CDKA/CYC complex involving the presence of CYCA3;1, a typical S-phase cyclin, to trigger the specific phosphorylation of protein substrates (Joubès et al., 1999, 2000b; Mathieu-Rivet et al., 2010b).
Most of the reported literature related to the control of endoreduplication has addressed the exit from the mitotic cycle and commitment to the endocycle. Consequently the attention of plant cell cycle investigators was drawn to the lack of mitosis during endoreduplication, which was proposed to occur because of the impairment of a so-called mitosis inducing factor (MIF), normally governing the passage through the G2 → M transition (Inzé and De Veylder, 2006). Vlieghe et al. (2007) showed that the down-regulation of M-phase-associated CDK activity is sufficient to drive cells into the endoreduplication cycle. Since the plant M-phase-specific CDKB1;1 activity is required to prevent a premature entry into the endocycle (Boudolf et al., 2004), CDKB1;1 was proposed as the most likely candidate kinase to be part of the MIF. Convincing in planta functional analyses highlighted the A-type cyclin CYCA2;3 as an appropriate partner of CDKB1;1 within the MIF (Yu et al., 2003; Imai et al., 2006), and Boudolf et al. (2009) demonstrated recently that CDKB1;1 and CYCA2;3 do interact in vivo to form a functional complex that inhibits endoreduplication. This process is regulated at the level of the CDKB1;1 activity which relies on the stability of the CYCA2;3 moiety: the commitment to endoreduplication and consequent exit from mitosis is then achieved through the selective degradation of CYCA2;3 (Boudolf et al., 2009) which requires the activation of the anaphase-promoting complex/cyclosome (APC/C), an E3-ubiquitin ligase part of the 26S proteasome degradation pathway (Capron et al., 2003; Fig. 2).
Once the cell has committed to endoreduplication, the sole control of MIF activity and inhibition cannot account for the progression within the endocycle. Reduplicating the genomic DNA requires the fluctuation in the activity of S-phase CDK at the S–G boundary to allow re-licensing of the origins of replication which subsequently trigger the assembly of the pre-replication complex (pre-RC) harbouring several components such as the ORC, CDC6, CDT1 and MCMs proteins (Bryant and Francis, 2008). As a conserved molecular control between the endocycle and the canonical cell cycle, the control of the G → S transition is exerted through the retinoblastoma-related protein (RBR) pathway (Gutierrez et al., 2002). Following hyperphosphorylation via the activity of a CDKA/CYCD complex, RBR releases the trapped E2F-DP dimeric transcription factor. E2F-DP now activates the transcription of E2F-responsive genes required for the commitment to the S phase, such as those encoding the pre-RC components (Fig. 2). The deregulated expression of all these genes in transgenic plants affects endoreduplication; for instance, the ectopic expression of E2Fa with its dimerization partner DPa (De Veylder et al., 2002), of CDC6 (Castellano et al., 2001) and CDT1 (Castellano et al., 2004), or RBR suppression (Park et al., 2005) are sufficient to trigger extra endocycles, consistent with the involvement of the RBR pathway in the G → S transition.
The balance between mitosis and commitment to endoreduplication is thus exerted at the level of CDK activity control (De Veylder et al., 2007). Three distinct mechanisms are proposed to affect the components of the CDK/CYC complexes at the post-translational level (Inzé and De Veylder, 2006). The kinase activity of the CDK/CYC complexes is dependent on (a) the phosphorylation/dephosphorylation status of the kinase itself, (b) the availability and binding of the cyclin regulatory subunit to a CDK partner and (c) the binding of specific CDK inhibitors. Since endoreduplication occurs naturally and plays such an important part during fruit development (Chevalier, 2007), the relative contribution of these different control mechanisms on CDK activity has been addressed in tomato.
In fission yeast (Schizosaccaromyces pombe), the commitment to mitosis is regulated by the successive action of the SpWEE1 and SpCDC25 phosphoregulators of CDK activity. Following DNA replication in the S phase, the activity of the yeast CDK, CDC2, is inhibited by phosphorylation on the Y15 residue mediated by the activity of the WEE1 kinase during G2. The progression within the cell cycle is then stopped to ensure that DNA replication and repair of damaged DNA have been completed (Russell and Nurse, 1987). At the G2/M boundary, CDC2 is then dephosphorylated on Y15 within the CDC2/CYCB complex through the action of CDC25 to drive the cell through mitosis (Russell and Nurse, 1986; O'Farrell, 2001).
Plant homologues to SpWEE1 have now been identified in many different species, including maize (Sun et al., 1999), arabidopsis (Sorrell et al., 2002) and tomato (Gonzalez et al., 2004), and the over-expression of maize and arabidopsis WEE1 in S. pombe induced an expected long-cell phenotype when compared with fission yeast (Russell and Nurse, 1986; 1987), as the length of the G2 phase is affected. The existence of a true CDC25 homologue in higher plants is still questioned, although a putative candidate named Arath;CDC25 was identified in arabidopsis, and shown to be able to dephosphorylate plant CDKs and activate CDK activity (Landrieu et al., 2004) and to induce a short-cell phenotype when over-expressed in fission yeast (Sorrell et al., 2005). Hence the functionality of the WEE1/CDC25-like phosphoregulators of CDK activity seems to operate in plants.
Expression analyses revealed that transcripts for WEE1 accumulate in maize endosperm (Sun et al., 1999), as well as in developing fruit tissues in tomato (Gonzalez et al., 2004), thus suggesting that the inhibitory WEE1 kinase activity could contribute to the endoreduplication process. Contrary to this proposed role for WEE1 in the onset of endoreduplication, De Schutter et al. (2007) demonstrated that WEE1 knock-out mutants in arabidopsis grew normally under standard conditions, indicating that WEE1 is not rate-limiting for cell cycle progression. WEE1 was shown to be part of the DNA-damage signal transduction pathway operating in the G2 phase by arresting the cell cycle in response to genotoxic stresses. Although the work of De Schutter et al. (2007) was extremely convincing, we intended to provide a new functional analysis of WEE1 in a highly endoreduplicating cell context, i.e. during tomato fruit development.
Tomato transgenic plants were generated with the aim of down-regulating the expression of WEE1 (plants referred to as Pro35S:Slwee1AS; Gonzalez et al., 2007). These transformants displayed a gradation in mature fruit size which correlated nicely with the degree to which the endogenous WEE1 gene was down-regulated: the lower the expression of WEE1, the smaller the fruit size. In addition, a clear reduction in the level of endoreduplication was found in 25-dpa fruits when compared with wild type; the number of nuclei at the 4C and 8C DNA levels was increased, while that at 16C, 32C and 64C DNA levels was reduced (Gonzalez et al., 2007). The molecular characterization revealed that the CDK/CYC Histone H1 kinase activity measured in young leaves of the Pro35S:Slwee1AS plant was enhanced as a result of WEE1 down-regulation. Most interestingly, a spectacular decrease in the amount of Tyr15-phosphorylated CDKA was observed. Since Tyr15 is the natural target of the WEE1 kinase activity, thus leading to the inactivation of CDKA activity, the down-regulation of WEE1 in tomato transgenic plants resulted in an increased quantity of dephosphorylated CDKA, and thus increased CDKA activity. In all the vegetative plant organs that were tested (cotyledons, young leaves and stems) it has been possible to show that the effect of a WEE1 down-regulation resulted in a short-cell phenotype. The fruit size reduction originates in a reduction in pericarp width because of a significant reduction in cell size, especially within cell layers five to nine in the central pericarp which is characterized by the presence of the largest cells in wild-type fruits. New data relative to WEE1 were thus provided, showing that WEE1 is an important regulator of tomato fruit development as its activity, at the interplay between normal cell cycle events and endoreduplication, has an important impact on organ and cell size determination.
According to the authors' data, it is tempting to speculate that WEE1 is involved in the control of the endocycle G-phase length during endoreduplication to allow the sufficient cell growth in response to nuclear DNA amplification and nucleus size increase (Fig. 2). This function of WEE1 in triggering the cell expansion- and endoreduplication-dependent growth of the fruit organ is likely to be species-dependent, and could explain the active WEE1 transcription in endoreduplicating tissues such as tomato fruit pericarp and jelly or maize endosperm (Sun et al., 1999; Gonzalez et al., 2004) where spectacular levels of endoreduplication can be reached, far ahead of those encountered in arabidopsis. Nevertheless, this role for WEE1 in the control of cell size and endoreduplication does not exclude a putative function at the DNA integrity checkpoints, since endoreduplication as a DNA amplification process may require a sustained WEE1 activity to cope with the DNA damage potentially occurring during the successive rounds of DNA replication.
The commitment towards endoreduplication requires the loss of mitotic CDK/CYC complex activity which can occur upon the selective proteolytic destruction of the cyclin subunits via the ubiquitin proteasome pathway. The ubiquitin-dependent proteolysis of mitotic cyclins requires the involvement of a specific E3-type ubiquitin ligase named the anaphase-promoting complex/cyclosome (APC/C) (Capron et al., 2003). In plants the APC/C is activated by the CCS52A protein (Cebolla et al., 1999), homologous to the mammalian CDH1 and Drosophila FZR, which binds to cyclins to drive them towards the degradation process by the 26S proteasome (Fülop et al., 2005; Boudolf et al., 2009). Like its eukaryotic counterparts, CCS52A was found to promote the onset and progression of endoreduplication in different plant organs (Cebolla et al., 1999; Lammens et al., 2008; Boudolf et al., 2009; Larson-Rabin et al., 2009; Vanstraelen et al., 2009), so the functional role of CCS52 during tomato fruit development was investigated.
Tomato plants over-expressing SlCCS52A were also generated (plants referred to as Pro35S:SlCCS52A°E). In the most extreme case, leaf development and morphology of the phenotype of Pro35S:SlCCS52A°E plants were deeply affected (Mathieu-Rivet et al., 2010a). The plant phenotype was characterized by the appearance of under-developed (small and curly) leaves, resembling those of arabidopsis plants over-expressing CDK inhibitors (Wang et al., 2000; De Veylder et al., 2001), thus suggesting that cell division was deeply impaired in these plants. Interestingly, the ploidy level in these plants was increased towards high DNA levels. The Pro35S:Slccs52AAS plants displayed smaller fruits than wild-type plants. The ploidy level of the Pro35S:Slccs52AAS was reduced, in correlation with a reduced mean cell size and an increased cell number. Using protoplasts prepared from Pro35S:Slccs52AAS leaves and a luciferase gene reporter assay, it was possible to demonstrate that the phenotype observed in Pro35S:Slccs52AAS fruits originated from a decrease in the APC/CCCS52A activity targeting CYCA3;1 for subsequent proteasome destruction, thus resulting in a higher stabilization of cyclins.
Tomato plants over-expressing SlCCS52A were also generated (plants referred to as Pro35S:SlCCS52A°E) and analysed for fruit growth characteristics (Mathieu-Rivet et al., 2010b). Here similar fruit phenotypes to those seen in Pro35S:Slccs52AAS loss-of-function plants, i.e. significantly smaller fruits than wild type, were also obtained. The reduction in fruit size in Pro35S:Slccs52AAS does not originate from any harm to cell division activities during the very early fruit development, but from the reduction in cell size within the pericarp, due to a severe impairment in endoreduplication in accordance with the preferential role of APC/CCCS52A in endoreduplication-driven cell expansion. Similarly, a SlCCS52A over-expression does not impact the very early fruit development between anthesis and 8 dpa, when fruits remain mitotically active and display high cyclin transcription rates (Joubès et al., 2000b). Hence cyclin production rates are so high at this stage that cyclin abundance is insensitive to the counteracting action of the over-expressed CCS52A. Between 8 and 15 dpa, fruit growth is deeply affected in Pro35S:SlCCS52A°E plants, resulting in a reduced fruit size as a consequence of impairment in endoreduplication. Since APC requires a cyclic activation to target A- and B-type cyclins for cyclic destruction for the progression in the endocycle (Narbonne-Reveau et al., 2008; Boudolf et al., 2009), this reduction in endoreduplication rate may be caused by a lack of target oscillation in Pro35S:SlCCS52A°E lines. Later on, the fruit growth resumes between 15 and 25 dpa, concomitantly with an enhanced production of 32C and 64C nuclei, which results in a higher endoreduplication index and production of larger cells, both contributing to a recovery in fruit size compared with wild type. During this period, the cyclin production rates reduce triggering oscillations in CDK activity and the effects of increasing CCS52A levels may become apparent on endoreduplication levels influencing fruit growth via the generation of highly polyploid expanded cells in very late development. The increased regulation at the level of target protein stability mediated by the activity of APC/CCCS52A, in combination with reduced cyclin production rates, is likely to be crucial for the control of the spectacular extent of endoreduplication during fruit development.
The fine tuning of the ICK/KRP protein abundance was demonstrated to be a key feature of cell cycle control, and especially to trigger the onset of the endoreduplication cycle (Verkest et al., 2005b; Vlieghe et al., 2007). It has been possible to identify and isolate so far a total of four different KRP inhibitors in tomato (named KRP1 to KRP4) (Nafati et al., 2010). Using quantitative PCR following reverse transcription (RT-qPCR), we showed that the four tomato KRPs display unique expression profiles during tomato fruit development. Interestingly, KRP4 is mostly expressed during the very early development of fruit when cell divisions predominantly drive fruit growth, while the expression of KRP3 is increasing at the latest stages of fruit development when cell expansion predominantly accounts for fruit growth. These data suggest that these KRPs may display distinct physiological and functional roles during the cell cycle and eventually during the endoreduplication cycle. This assertion was strengthened by our recent phylogenetic, structural and functional analysis of tomato ICK/KRPs which aimed at assigning putative roles to the conserved motifs in specific protein–protein interactions, subcellular localization or proteolytic degradation (Nafati et al., 2010).
During tomato fruit development, the large and hyper-vacuolarized cells constituting the jelly-like (gel) locular tissue undergo multiple rounds of endoreduplication. Within this particular tissue, mitosis is arrested after 20 dpa and only endoreduplication occurs thereafter, concomitantly with a strong post-translational inhibitory regulation of the CDKA activity (Joubès et al., 1999). The origin of this post-translational regulatory mechanism resides in part in the accumulation of SlKRP1 which accounts for the inhibition of CDK/CYC kinase activity during the development of the gel tissue in tomato fruit (Bisbis et al., 2006).
In plants, no direct phenotypic effects induced by the reduced expression or mutation of ICK/KRP genes has been reported so far, most probably as a result of gene functional redundancy since ICK/KRP belongs to a multigene family. In contrast, numerous studies described the effects of ICK/KRP over-expression in arabidopsis (Wang et al., 2000; Schnittger et al., 2003; Verkest et al., 2005a; Weinl et al., 2005; Bemis et al., 2007), maize (Coelho et al., 2005), tobacco (Jasinski et al., 2002), rice (Barrôco et al., 2006) and brassica (Zhou et al., 2002). A low level of ICK/KRP over-expression appears to block cell division and induce nuclear endoreduplication, together with only a slight alteration in both cell size and a slight decrease in final plant size (Verkest et al., 2005a; Weinl et al., 2005). By contrast both cell division and endoreduplication are negatively affected by a high level of ICK/KRP over-expression, generating cells of bigger size, and leading to an overall plant dwarfism (De Veylder et al., 2001; Jasinski et al., 2002; Zhou et al., 2003).
A functional analysis of the tomato ICK/KRP gene SlKRP1 during fruit development was performed by over-expressing SlKRP1 under the control of the tomato PhosphoEnolPyruvate Carboxylase 2 (PEPC2) promoter (Fernandez et al., 2009), which is specifically and strongly expressed in the expanding cells of mesocarp during the early fruit development (plants referred to as ProPEPC2:SlKRP1°E) (Nafati et al., 2011). Using this promoter, the aim was thus to investigate the effects of endoreduplication induced by SlKRP1 on fruit growth independently from cell divisions.
While DNA ploidy levels were clearly lowered in tomato fruit mesocarp from ProPEPC2:SlKRP1°E plants, as a result of the effect of a high level of SlKRP1 over-expression on endoreduplication, it was not possible to detect any apparent phenotype at the morphological and cytological levels; the final fruit size in ProPEPC2:SlKRP1°E plants was indistinguishable from that of wild-type fruit, and the mean cell diameter within the mesocarp tissue and the number of cell layers across the pericarp were both unaffected (Fig. 3). Therefore the PEPC2-driven over-expression of SlKRP1 during the phase of cell expansion induced a complete disruption in the nuclear : cytoplasmic ratio, thus uncoupling endoreduplication from cell size, as the decrease in ploidy levels and consequent reduction in nuclear size in ProPEPC2:SlKRP1°E fruit occurred while cell size was unaffected. It was thus possible to reproduce the data from Leiva-Neto et al. (2004) who reported that cell size determination can be independent of endoreduplication level when a dominant negative form of CDKA was over-expressed in the endoreduplicating cells of the maize endosperm using the 27-kDa γ zein promoter.
The strong proportionality between ploidy level and cell size during normal plant development has been widely observed, and applies to cell populations within a tissue or an organ but rarely to individual cells because it is highly probable that, for a given cell size, several distinct ploidy levels may occur. Our functional analysis revealed that the population of cells in ProPEPC2:SlKRP1°E fruit mesocarp performed one endoreduplication round under wild-type cells when referring to the measured endoreduplication index. In ProPEPC2:SlKRP1°E fruits, DNA ploidy levels were indeed lowered. However, the levels of endoreduplication attained seemed to be still sufficient to support growth and an optimal cell size increase as compared with wild type, which suggests that endoreduplication is likely to support a range of cell size rather than a defined one. Thus endoreduplication does not exert a direct control on cell growth but would rather be a limiting factor for cell growth, in accordance with the proposed model from Schnittger et al. (2003).
The extent of endoreduplication in various plant organs and species is clearly correlated with cell growth, suggesting a causal relationship between these two cellular processes, and thus represents a major determinant for organ size. However, the role of endoreduplication in modulating the rate of organ growth and/or cell expansion has been recently questioned (John and Qi, 2008), because all the in planta functional analyses reported so far have impacted genes that regulate the progression into the cell cycle and therefore both affect cell division and cell expansion-associated endoreduplication. As a consequence, two opposing mechanistic views arise from the literature for describing organ growth. First, the final size of an organ can result from the combination of cell number and cell size which are respectively determined by cell division and cell expansion together with endoreduplication, according to probabilistic decisions of individual cells, as nicely demonstrated for sepal epidermal patterning in arabidopsis (Roeder et al., 2010). Opposed to this cell-based (autonomous) developmental process accounting for growth, some authors argued for a deterministic organismal level of regulation where growth itself, via a signalling pathway putatively involving hormones such as auxin, would primarily control cell proliferation, thus determining the final size of the cell and ultimately organ size (Mizukami, 2001; John and Qi, 2008). In light of the undisputed ‘karyoplasmic ratio’ theory (Sugimoto-Shirasu and Roberts 2003), what comes first – cell growth or endoreduplication? If growth occurs according to the organismal control, cell division and cell enlargement driven by endoreduplication are just secondary processes to maintain the karyoplasmic ratio (Cookson et al., 2006; John and Qi, 2008). However, as often in biological processes, things may not appear to be so dualistic since it is likely that a minimal cell size is required prior to commitment to DNA replication during the endocycle, and, alternatively, once DNA synthesis is completed, the doubling of the DNA quantity can promote cell growth, with respect to the karyoplasmic ratio. Since the endocycle is made successively of S phase and undifferentiated G phase, the latter corresponds to a fused G1 and G2 phase where the two naturally occurring periods of cell growth in the canonical cell cycle would appear at once.
In studies of tomato fruit development and the contribution of endoreduplication in fruit growth, it has been possible to demonstrate that the down-regulation of endoreduplication-promoting genes such as SlWEE1 and, more importantly, SlCCS52A clearly impacted negatively cell size as a result of lower levels in DNA ploidy, and affected the whole tomato fruit development and final fruit size (Gonzalez et al., 2007; Mathieu-Rivet et al., 2010b). When SlCCS52A was over-expressed, fruit growth was first impaired as endoreduplication was initially delayed, accounting for the altered final fruit size. However, endoreduplication then resumed and was even enhanced after 15 dpa leading to fruit growth recovery (Mathieu-Rivet et al., 2010b).
Surprisingly, the fruit-specific over-expression of the CDK inhibitor SlKRP1 under the control of the strong PEPC2 promoter induced a disruption in the karyoplasmic ratio, as it negatively impacted on endoreduplication within fruit pericarp without affecting cytological characteristics and final fruit size (Nafati et al., 2011). The time frame of expression of SlKRP1 under the control of this cell expansion-specific promoter in fruit allowed the onset of endoreduplication during early fruit growth, and subsequently impaired its progression. However, cell enlargement could happen just as in wild type, suggesting that endoreduplication operates as a trigger for cell and organ growth.
Dealing with a hypothesized role of endoreduplication during plant development, endoreduplication is unlikely to contribute to the regulation of transcriptional activity and subsequent metabolic activity by increasing the availability of DNA templates for gene expression. Indeed, we reported that the tomato fruits over-expressing SlCCS52A displayed metabolic modifications that were in fact attributable to alterations in fruit growth rather than endoreduplication itself (Mathieu-Rivet et al., 2010a). Moreover, comparing the metabolic profiling of SlKRP1 over-expressing fruits with that of wild-type fruits did not reveal any significant difference (Nafati et al., 2011), in accordance with the findings of Leiva-Neto et al. (2004) who demonstrated that inducing lower levels of endoreduplication in maize endosperm did not impact on starch or storage protein contents and the associated biosynthetic gene transcripts.
The data thus argue for a physiological role of endoreduplication as a facilitator of cell growth (Nafati et al., 2011), but more importantly as an accelerator for organ growth, such as in fruit (Gonzalez et al., 2007; Bourdon et al., 2010; Mathieu-Rivet et al., 2010b). As Barow and Meister (2003) proposed that endoreduplication is important for plant growth in response to environmental constraints, whether this process was selected during evolution prior to fleshy fruit domestication to cope with hostile growth conditions, is still an intriguing question.
I would like to express my deepest thanks to Aurélie Honoré and Patricia Ballias, and Sophie Salar, for their excellent technical work, especially for taking care of the plants in the greenhouse and for performing morphometrical and cytological analyses, respectively.