Pre-meiotic endomitosis in the cytokinesis-defective tomato mutant pmcd1 generates tetraploid meiocytes and diploid gametes

Journal of Experimental Botany, May 2013

Sexual polyploidization through the formation and functioning of 2n gametes is considered a major route for plant speciation and diversification. The cellular mechanism underlying 2n gamete formation mostly involves a restitution of the meiotic cell cycle, generating dyads and triads instead of tetrad meiotic end-products. As an alternative mechanism, the tomato mutant pmcd1 (for pre-meiotic cytokinesis defect 1), which generates diploid gametes through the ectopic induction of pre-meiotic endomitosis, is presented here. Using cytological approaches, it is demonstrated that male pmcd1 meiocyte initials exhibit clear alterations in cell cycle progression and cell plate formation, and consequently form syncytial cells that display different grades of cellular and/or nuclear fusion. In addition, it was found that other somatic tissue types (e.g. cotyledons and petals) also display occasional defects in cell wall formation and exhibit alterations in callose deposition, indicating that pmcd1 has a general defect in cell plate formation, most probably caused by alterations in callose biosynthesis. In a broader perspective, these findings demonstrate that defects in cytokinesis and cell plate formation may constitute a putative route for diplogamete formation and sexual polyploidization in plants.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

Pre-meiotic endomitosis in the cytokinesis-defective tomato mutant pmcd1 generates tetraploid meiocytes and diploid gametes

NicoDe Storme 0 DannyGeelen 0 0 In vitro Biology and Horticulture, Department of Plant Production, University of Ghent , Coupure Links 653, 9000 Ghent, Belgium Sexual polyploidization through the formation and functioning of 2n gametes is considered a major route for plant speciation and diversification. The cellular mechanism underlying 2n gamete formation mostly involves a restitution of the meiotic cell cycle, generating dyads and triads instead of tetrad meiotic end-products. As an alternative mechanism, the tomato mutant pmcd1 (for pre-meiotic cytokinesis defect 1), which generates diploid gametes through the ectopic induction of pre-meiotic endomitosis, is presented here. Using cytological approaches, it is demonstrated that male pmcd1 meiocyte initials exhibit clear alterations in cell cycle progression and cell plate formation, and consequently form syncytial cells that display different grades of cellular and/or nuclear fusion. In addition, it was found that other somatic tissue types (e.g. cotyledons and petals) also display occasional defects in cell wall formation and exhibit alterations in callose deposition, indicating that pmcd1 has a general defect in cell plate formation, most probably caused by alterations in callose biosynthesis. In a broader perspective, these findings demonstrate that defects in cytokinesis and cell plate formation may constitute a putative route for diplogamete formation and sexual polyploidization in plants. - In plant evolution, the induction of polyploidy (i.e. the condition of having three or more copies of the basic set of chromosomes) is generally believed to be one of the major factors driving speciation and diversification (Adams and Wendel, 2005). Based on the redundant genomic material and increased genomic variability, polyploids have a high level of genotypic plasticity that provides them with an evolutionary advantage and enhanced competitiveness compared with their diploid equivalents (Leitch and Leitch, 2008). The supremacy of polyploidy is reflected by the ubiquitous presence of polyploids in the plant kingdom. In general, it is suggested that 3570% of angiosperm species are polyploid or show remnants of an ancient polyploidization event. Although the exact mode of polyploidization during plant evolution is unknown, the overall presence of 2n gametes suggests that whole-genome duplication through sexual means is the predominant mode of polyploidization (Bretagnolle and Thompson, 1995). Different cytological mechanisms are known to generate diploid gametes. In most cases, diploid spores originate from cellular defects that convert the meiotic cell cycle into a mitotic-like division, generating dyads and/or triads that contain diploid instead of haploid spores. This process is termed meiotic restitution or meiotic non-reduction. Cellular defects leading to meiotic restitution are characterized either by: (i) aberrations in spindle biogenesis and organization; (ii) the complete omission of meiosis I(MI) or MII; or (iii) defects in meiotic cell plate formation. Moreover, based on the genotypic outcome of the resulting 2n gametes, restitution mechanisms can be classified as FDR (first division restitution) or SDR (second division restitution). In FDR, sister chromatids The Author [2013]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved. For permissions, please email: are disjoined and the resulting 2n gametes largely maintain parental heterozygosity and epistasis, particularly at the centromeric regions. In SDR mechanisms, on the other hand, sister chromatids are physically not separated, generating 2n gametes that have lost most parental heterozygosity and epistasis, especially at the centromeric regions (Brownfield and Kohler, 2011). Apart from meiotic non-reduction, diploid gametes can also be generated by (pre-)meiotic genome doubling. In this process, tetraploid meiocytes are ectopically generated and eventually lead to the formation of diploid gametes. Meiotic genome doubling may result from two different cytological alterations, namely cytomixis or syncyte formation. In these processes, respectively, migration of chromosomes through cytoplasmic channels and fusion of one or more nuclei through defects in cell plate establishment just before or during meiosis generate tetraploid meiocytes, which eventually yield diploid gametes (Falistocco et al., 1995). Pre-meiotic genome doubling has repeatedly been observed in animal species, particularly as a mechanism to alleviate hybrid-associated sterility or to generate clonal gametes in parthenogenetic species (Itono et al., 2006; Lutes et al., 2010). In contrast, in plants, the formation of 2n gametes through pre-meiotic genome duplication has only been reported for a small set of isolated species. For example, in Dactylis glomerata L., diploid pollen was found to originate from cytomictic genome duplication events, ectopically occurring in developing pollen mother cells (PMCs; Falistocco etal., 1995). Similarly, in several other species that show cytomixis or cell fusion in male sporogenesis (e.g. spearmint, Lindelofia, etc.), the production of 2n spores through the ectopic formation of tetraploid meiocytes has also been suggested (Tyagi, 2003; Singhal et al., 2011). In contrast, in many species, cytomixis generates multinuclear PMCs without any fusion of chromatin or polyploid meiocyte formation. In polyploid accessions of Brachiaria, for instance, fusion between two or more meiotic cells occurs throughout the whole process of microsporogenesis, but subsequent nuclear fusion and the associated formation of polyploid meiocytes has never been observed (Mendes-Bonato etal., 2003; Boldrini etal., 2005). In this report, the tomato mutant pmcd1 (pre-meiotic cytokinetic defect 1) that produces diploid pollen through the ectopic formation of tetraploid meiocytes is presented. Based on cytological observations, it was found that tetraploid meiocytes in pmcd1 originate from nuclear fusion events in PMC initials, ectopically generated by alterations in cell wall formation. Similar defects in cell wall establishment and the associated formation of endomitotic nuclei were also occasionally observed in petal and cotyledon tissues, indicating that the pmcd1 allele causes a weak, non-lethal defect in somatic cell plate formation, specifically leading to tetraploid meiocytes in male sporogenesis. Materials and methods Plant materials lycopersicum TR 5306) kindly provided by Rob Dirks (Rijk Zwaan Breeding B.V.). Tetraploid and diploid (M82) controls were obtained from Rijk Zwaan. Tomato plants were propagated through seed or cuttings, and plants were grown to maturity in a climate chamber under a 16 h/ 8h light/dark rhythm and a temperature of 21C. Meiocyte and pollen cytology For visualization of male gametophytic nuclei, flowers were put into 1 ml of distilled H2O and intensively vortexed for 5 min. After removal of the flowers, the solution was centrifuged for 5min at 10 000 g and the resulting pollen pellet was dissolved in a 100 mM citric/phosphate buffer (pH 4)containing 1g ml1 4,6-diamidino2-phenylindole (DAPI) and 1% TritonX-100. Fluorescein diacetate (FDA) staining was performed according to Heslop-Harrison and Heslop-Harrison (1970). Astock solution of 2 mg ml1 FDA was prepared in acetone and diluted to 100 l ml1 with 7% sucrose in water (w/v). Mature tomato flowers were immersed in 350 l of staining solution and vortexed for 5 min. Upon flower removal and centrifugation (5 min at 10 000 g), the stained pollen pellet was washed with 7% sucrose in water and transferred to a microscopyslide. Cellular and nuclear morphology of tomato PMCs was monitored using orcein (Sigma-Aldrich). Based on the close correlation between bud size and meiotic developmental stage, buds of the appropriate size (meiosis: 712 mm) were collected in a humidity box. After dissection of the anthers, PMCs of one anther were squashed in 45% (v/v) lactopropionic orcein in distilled water [LPO stock: 5% orcein (w/v) in lactopropionic acid] on a microscopyslide. For the analysis of meiotic chromosome behaviour, a modified PMC spreading protocol was performed (Ross etal., 1996). Briefly, inflorescence buds of the appropriate stage were fixed in Carnoys fixative (3:1 ethanol:acetic acid) for at least 48h. Fixed buds were then rinsed twice in distilled water and twice in 10 mM citrate buffer (pH 4,5) and subsequently incubated in a 0.3% (w/v) digestion mix [cellulase RS and pectolyase Y23 (Sigma) in 10mM citrate buffer] for 3h at 37C. After digestion, buds were washed twice in distilled water and squashed on a microscopy slide. To counter cell clustering, 10l of 60% acetic acid was added to the cell suspension and slides were heated for 30 s at 45C on a hot plate. Next, cells were fixed on the slide by pipetting Carnoys fixative around the cleared cell suspension. After air drying, fixed cells were stained and mounted with a 2g ml1 DAPI solution in Vectashield antifade mounting medium (Vector Laboratories). Other cytology To monitor the nuclear morphology of meiocyte precursor cells, pre-meiotic anthers were dissected and locules were dyed in 2 g ml1DAPI. For histochemical callose staining, isolated material was infiltrated with 0.1% (w/v) aniline blue in 0.033% (w/v) K3PO4 for 10 min, washed in distilled water, and observed using fluorescence microscopy. Visual assessment of epidermal cell walls and quantitative analysis of the stomatal clustering were performed according to Guseman et al. (2010). Cotyledon-stage seedlings were fixed in ethanol:acetic acid (3:1) for at least 1 h, washed in 70% ethanol, and used for brightfield microscopic analysis. Both brightfield and fluorescence observations were made using an Olympus IX81 microscope equipped with an X-Cite Series 120Q UV lamp. Photographs were taken with an Olympus XM10 camera driven by Olympus Cell M software. Sterol measurements The pmcd1 mutant was isolated from an M2 population of ethyl methanesulphonate (EMS)-treated tomato seeds (Solanum Quantification of sterols was performed using the Q-TOF method, previously described by Wewer etal. (2011). The tomato mutant pmcd1 produces a subset of enlarged polyploid pollen grains and enlargedseed In the search for mutants producing diploid pollen, a population of 6000 EMS-mutagenized M2 tomato (S.lycopersicum) plants was screened for larger pollen grains. In tomato, as in several other plant species, there is a clear correlation between pollen size and the gametophytic ploidy level. Pollen size, and more particularly the presence of larger pollen, can therefore be used as a reliable parameter to detect higher ploidy male gamete formation. Using this methodology, we isolated the mutant pmcd1, which showed a distinct bimodal pollen size distribution with a large subset of normally sized pollen grains (2428m) and a second subgroup of larger pollen grains (14.8 3.2%, 2834m), closely corresponding to the size of diploid pollen (Fig. 1A). As such, the normally sized spores most probably correspond to haploid pollen, whereas the larger ones suggest di- or polyploid spore formation. Microscopic analysis confirmed that pmcd1 produces a subpopulation of larger pollen grains at a frequency of 7.0 2.0% (Fig.1B). To verify whether the large pollen in pmcd1 have an increased ploidy level, nuclear DNA staining was performed. In wild-type diploid plants, mature spores always show the characteristic binuclear configuration, with one highly condensed, rod-shaped generative nucleus and one less condensed vegetative nucleus (Fig. 1C). In all wild-type spores, the size and fluorescence intensity of both nuclei types appeared highly uniform, reflecting the uniformity in haploid DNA content. Regular-sized pmcd1 and enlarged pmcd1 pollen also display this binuclear spore configuration, but in the enlarged spores both nuclei types appear substantially larger compared with their diploid counterparts, clearly indicating a higher gametophytic DNA content (Fig.1C). Similarly, in an earlier stage of microspore development, at the uninuclear stage, pmcd1 anthers not only contain wild-type-like spores, but additionally display a minor subset of larger spores with an enlarged nucleus (Fig.1D). This observation suggests that the causative genome duplication event in pmcd1 sporogenesis occurs during or before the uninuclear microsporestage. The viability of both pmcd1 pollen types (normal sized and larger) was tested using FDA staining. In this assay, fluorescent spores are considered metabolically active and viable, whereas non-fluorescent spores are non-viable. In the wild-type tomato, most pollen exhibited a clear green fluorescence, strongly reflecting the high male gametophytic fertility of these plants (Fig. 1E). Similarly, in the pmcd1 mutant, both large and normally sized pollen grains showed a bright fluorescent staining, indicating that both types of spores are viable. Enlarged polyploid pollen were also found to form a pollen tube in a liquid germination medium, suggesting that they have the capacity to engage in fertilization (data not shown). Strikingly, despite the supportive evidence for viability and germination capacity of polyploid pollen, tri- or polyploid pmcd1 progeny plants (n >50) were never retrieved. Indeed, although pmcd1 seeds are consistently larger compared with the wild type Fig.1. pmcd1 produces a subset of viable, higher ploidy male gametes. Volume-based diameter distribution (A) and microscopic analysis (B) of pollen at anthesis. Scale bar, 20m. Nuclear DNA staining (DAPI) of mature binuclear (C) and immature uninuclear spores (D) and fluorescein diacetate (FDA) viability staining (E) of mature pollen grains. Scale bar, 10m. (Supplementary Fig. S1 available at JXB online), a phenotype that could indicate polyploidy, DNA flow cytometry of pmcd1 selfed progeny plants (somatic leaf material) always revealed a diploid ploidy level (Supplementary Fig. S2), suggesting either that the frequency of pmcd1 polyploid gamete formation was too low to compete with the haploid pollen abundance, or that the formation of tri- and polyploid seeds was impaired by imbalances in the endosperm balance number (EBN; triploid block). Moreover, since large pmcd1 seeds consistently yielded diploid plants (n >50), it is concluded that the larger seed phenotype is not caused by an increased ploidy, but rather by an additional effect of the pmcd1 mutation. To determine the genetic background of the pmcd1 larger pollen grain mutation, pmcd1 was crossed into an M82 diploid control line and the phenotypic segregation of the selfed F2 progeny population was analysed. As F1 plants (n=2) did not show any phenotypic aberration and the F2 population exhibited a 1:3 Mendelian segregation of the larger pollen grain phenotype (n=22; five plants showed the phenotype), it is concluded that pmcd1 is sporophytic recessive and genetically constituted by a single mutation or by two or more closely linked mutations. pmcd1 male sporogenesis generates twin tetrads and enlarged tetrads and polyads In higher plants, polyploid gametes are generally formed by a defect in the meiotic cell division programme. To monitor pmcd1 meiotic cell division, meiotic end-products were examined using orcein-stained flower bud squashes. In wild-type male sporogenesis, meiosis always produces tetrads containing four equally sized spores (100%; n >250), reflecting the double reductional cell division (Fig. 2A, a). In the pmcd1 mutant, on the other hand, besides normal tetrads (68.8 6.2%), a variety of altered microspore configurations were observed (Fig. 2A, bf, B), predominantly including enlarged tetrads (9.0 2.0%) and enlarged polyads (11.5 4.2%), and additionally displaying twin meiocytes (6.3 2.1%), partial twins (3.2 0.4%), and large octads (1.4 0.7%). Twin meiocytes contain two normal tetrads physically attached to each other and generate normal haploid spores, albeit in a double tetrad configuration. Based on their fused phenotype, twin meiocytes point towards defects in meiocyte differentiation or pre-meiotic cytokinesis (Fig.2A, b). Enlarged tetrads generally contain four equally enlarged spores, similar to tetrads observed in a tetraploid background (Fig. 2A, c). As such, these meiocytes indicate a genome doubling event occurring either before or during meiosis, typically generating diploid instead of haploid spores. In some cases, however, the four spores in the enlarged tetrad are highly variable in size, suggesting imbalances in meiotic chromosome segregation (Fig. 2A, e). Polyads, by definition, contain more than four spores (Fig. 2A, d) and also indicate defects in meiotic chromosome segregation (e.g. recombination defects). Polyads typically generate aneuploid microspores and confer a high level of gametophytic sterility. Octads, on the other hand, contain eight equally sized spores and originate from a binuclear meiocyte initial, most probably generated by defects in pre-meiotic cell wall formation (Fig.2A, f). Since octads, partial twins, and twin meiocytes are expected to yield normal sized haploid spores, enlarged pmcd1 pollen was thought to originate from the enlarged tetrads and polyads, most probably originating from a genome doubling event occurring during or before the meiotic cell division. Tetraploid meiocytes in pmcd1 lead to the formation of diploid and aneuploid male gametes To check whether the variety of meiotic products in pmcd1 male sporogenesis originate from defects in meiotic cell division, meiotic chromosome behaviour was cytologically assessed. In wild-type tomato meiosis, chromosomes condense (Fig. 3a) and subsequently synapse to form homologous chromosome pairs at pachytene (Fig. 3d). Further condensation of the homologue duplexes generates 12 highly condensed bivalent chromosome structures (Fig. 3g, j) that subsequently align (Fig. 3m) and segregate to generate two polar sets of 12 chromosomes at telophase I. Following decondensation at meiotic interphase, the two chromosome sets again condense (Fig. 3p) and align according to two physically separated metaphase II plates. Since both division planes are perpendicularly orientated, subsequent division of chromatids generates four tetrahedrally arranged chromosome sets (12 chromosomes, e.g. haploid) at the end of MII (Fig.3s). In male sporogenesis of pmcd1, most PMCs show a normal chromosome behaviour and progression of the meiotic cell cycle. However, besides the normal diploid meiocytes, two types of meiotic aberrations were occasionally observed at all stages of meiosis; namely twin cells and enlarged meiocytes (Fig.3). In the twin meiocytes, which basically consist of two diploid meiocytes closely attached to each other, progression of the meiotic cell cycle in each individual cell appears normal and synchronous with the other one (Fig.3b, e, h, k, n, q), leading to the formation of twin tetrads that typically contain eight instead of four haploid microspores (Fig. 3t). In the enlarged pmcd1 meiocytes, meiosis always exhibits 48 chromosomes (e.g. tetraploid), suggesting that these PMCs resulted from an ectopic genome duplication event (Fig.3l). Since PMCs with a doubled chromosome number were consistently observed in all stages of pmcd1 male meiosis (Fig.3c, f, i, l, o, r), the ectopic genome duplication was found to occur before meiosis, in the pre-meiotic germline initials. Tetraploid pmcd1 meiocytes display a normal progression through the meiotic cell cycle and mostly generate balanced tetrads containing diploid (2x) spores (Fig. 3u). In some cases, however, unbalanced chromosome sets were observed at the end of MII, typically resulting in the formation of imbalanced tetrads that contain aneuploid gametes. Similar imbalances in tetrad-stage chromosome segregation have already been observed in tetraploid PMCs of other species and have been attributed to an unequal segregation of quadrivalent chromosome structures at MI. In line with this, tetraploid pmcd1 male meiocytes predominantly display quadrivalent chromosome structures at diakinesis, together with a minor subset of bivalent homologue duplexes (Supplementary Fig. S3). Although accurate quantification of bi- and quadrivalents cannot be performed based on chromosome spreads, pmcd1 tetraploid male diakinesis was found to generate a minimum number of four bivalents, whereas the other chromosome structures can be quadrivalents or superimposed bivalents. Asimilar level of bivalent formation was observed in tetraploid control lines (Supplementary Fig. S3). Based on the Fig.2. pmcd1 male meiosis generates altered meiotic products. (A) Orcein staining of male meiocytes at the end of meiosis II (tetrad stage) shows normal tetrads in the wild type (a) and a variety of altered meiotic products in pmcd1, including twin tetrads (b), enlarged tetrads (c), enlarged polyads (d and e), and octads (f). Scale bar, 10m. (B) Relative frequency of meiotic tetrad configurations at the end of pmcd1 male meiosis. (partial) formation of quadrivalents, it was concluded that tetrad-stage aneuploidy in pmcd1 tetraploid meiosis most probably originates from defects in MI chromosome segregation. In line with this, 47.4% of tetraploid pmcd1 PMCs (n=19) showed an unbalanced chromosome segregation at the start of MII. It was also noted that pmcd1 tetraploid meiocytes occasionally exhibit one or more lagging chromosomes, yielding polyads that typically contain 24 aneuploid spores with one or two additional micronuclei. Although this observation suggests that pmcd1 polyads originate from cellular alterations associated with chromosome duplication, the possibility that they result from other defects in meiosis, such as partial PMC fusion or chromosome migration, cannot be excluded. Based on the consistent observation of tetraploid meiocytes as well as twin meiocytes, throughout the whole process of Fig.3. All stages of pmcd1 male meiosis exhibit diploid, tetraploid, and twin meiocytes. DAPI-stained chromosome spreads of male meiocytes displays twin (middle column) and tetraploid (right column) meiocytes in all stages of pmcd1 meiosis. Scale bars, 10m. male sporogenesis, it is concluded that pmcd1 generates diploid (and a subset of aneuploidy) pollen through the ectopic induction of pre-meiotic endomitosis, most probably resulting from a defect in cytokinesis or cell wall establishment. In order to check whether similar meiotic alterations also occur in pmcd1 female sporogenesis, female meiosis was cytologically monitored. Although technically challenging, it was possible to visualize a small number of female pmcd1 prophase I meiocytes (e.g. through orcein staining) and it was found that they do not show any defect in nuclear and cellular morphology and/or size (n=14; Supplementary Fig. S4). These findings suggest that female pmcd1 meiosis is not affected or shows a lower level of cellular alterations compared with the male reproductive side. Pre-meiotic genome doubling in pmcd1 is caused by defects in cell plate formation In order to reveal the cellular mechanism underlying the formation of both tetraploid and twin meiocytes in pmcd1, orcein staining was performed at different stages of male sporogenesis, from the pre-meiotic up till the meiotic tetrad stage. Typically, in all stages of meiosis, besides single diploid and tetraploid (Fig. 4e, j, o, t) PMCs, pmcd1 also displayed a substantial number of doubled meiocytes containing two individual PMCs that are physically separated by either a full (Fig.4b, g, l, q), partial (Fig.4c, h, m, r), or incomplete cell wall (Fig. 4d, i, n, s). In the majority of these meiotic doublets, individual meiocytes remained physically isolated, without displaying any nuclear fusion or chromosome migration, and displayed synchronous meiotic cell cycle progression (Fig.4bd, gi, ln), typically generating octads instead of the normal tetrads at the end of MII (Fig.4qs). In some twin meiocytes, however, partial or complete fusion of the two nuclei was observed, leading to polyads and enlarged tetrads, respectively, at the end of MII. Hence, based on the observation that (partial) cell plate formation in pmcd1 twin meiocytes is retrieved in all stages of male meiosis together with the consistent presence of tetraploid PMCs, it is presumed that both types of PMCs originate from defects in pre-meiotic cell division or cytokinesis. To check whether pmcd1 meiotic precursor cells have any defect in mitotic progression or cell wall formation, pre-meiotic anther locules were cytologically assessed. In wild-type tomato, meiotic precursor cells always exhibit uniformly sized nuclei (Fig.5A), clearly reflecting the synchronous development and uniform ploidy of pre-meiotic germ cells. In pmcd1, on the other hand, clear alterations in pre-meiotic nuclear morphology were observed. Indeed, besides the normally sized, diploid nuclei, a significant subset of enlarged (fused) and paired (fusing) nuclei were also registered, indicating the occurrence of pre-meiotic nuclear fusion (Fig. 5B, arrows). In addition, examination of DAPI-stained pmcd1 PMC initials showed an asynchronous cell proliferation pattern, with some cells exhibiting condensed chromosomes (e.g. M-phase) and others displaying a non-condensed nuclear configuration (Fig.5A, asterisks indicate cells undergoing division). As such, it is concluded that the ectopic induction of endomitotic polyploidy in pmcd1 PMC initials is caused by alterations in pre-meiotic cell cycle progression and associated defects in cell plate formation. Moreover, the relative abundance of condensed chromosomes (rarely seen in the wild Fig.4. pmcd1 twin meiocytes display a variable gradient of cell wall formation. Orcein staining of wild-type and pmcd1 male meiocytes at different stages in meiotic cell division. Scale bars, 10m. type) suggests that pmcd1 PMC initials suffer from a (partial) arrest in mitotic metaphase, possibly forming the basis for subsequent defects in somatic cell plate formation. As endomitosis typically causes a doubling of the absolute number of chromosomes, whereas other genome duplication processes (such as endoreduplication) do not alter the absolute chromosome number, somatic chromosome counts were next performed in pmcd1 pre-meiotic flower buds to determine the mode of polyploidization. In contrast to the diploid wild-type control, in which pre-meiotic flower bud cells typically contain 24 chromosomes, pmcd1 pre-meiotic buds additionally displayed a subset of cells with an increased chromosome number (Fig.5C), supporting the earlier notion that tetraploid pmcd1 meiocytes originate from a pre-meiotic genome doubling event. Moreover, since chromosomes in these polyploid cells are often in a condensed state, pre-meiotic polyploidy in pmcd1 is caused by the ectopic induction of endomitotic genome duplication. The combined presence of physically non-separated nuclei and asynchronously dividing polyploid nuclei in pmcd1 meiocyte initials indicates a deregulation of cell cycle progression and demonstrates that meiotic aberrations in pmcd1 originate from defects in pre-meiotic cell division. Moreover, based on the (pre-)meiotic alterations, it is hypothesized that the altered cell cycle control in pmcd1 meiocyte initials causes an extra, often incomplete round of mitotic cell division (putatively blocked at metaphase) and consequently generates twin (nuclear division with cell wall formation), tetraploid (nuclear division without cell wall formation), and partially separated (partial cell plate formation) meiocytes. Defects in pmcd1 somatic cell proliferation are not confined to male meiocyte initials To monitor whether ectopic endomitosis in pmcd1 also appears in other tissues, the somatic endoploidy pattern in leaves and flower organs was analysed. DNA flow cytometry demonstrated that pmcd1 leaves have an endopolyploidy pattern that is highly similar to that of the wild type (Fig.6A). Statistical analysis, however, revealed that the 8x and the 16x nuclear population in young and old leaves, respectively, was significantly higher in pmcd1 compared with the wild type. Similarly, in pre-meiotic flower buds, DNA flow cytometry revealed significant endoploidy differences between the wild type and pmcd1 (Fig. 6B). Indeed, in pmcd1 pre-meiotic flower buds, the frequency of diploid nuclei was significantly lower and the Fig.5. pmcd1 meiocyte initials have an increased genomic DNA content and show defects in cell wall establishment. In vivo DAPI staining (A) and orcein labelling (B) of pre-meiotic male archesporal cells. Asynchronously dividing, enlarged cells are indicated with an asterisk and defects in cell plate formation are highlighted by arrows. (C) Chromosome spreading of cells isolated from pre-meiotic flower buds. Scale bars, 20m. level of 4x and 8x nuclei was significantly higher compared with the wild type, indicating a general increase in endopolyploidy. Although this is at least partially caused by ectopic genome duplications in meiocyte initials, the strong endoploidy increase in pmcd1 flower buds led to the hypothesis that other flower tissues also show ectopic endopolyploidization. Strikingly, in both floral organ types (e.g. sepals and petals), no major differences in endoploidy were observed and pmcd1 sepals even showed a significant reduction in the 16x nuclei population compared with the wildtype. To investigate the somatic endoploidy level and cell wall morphology, cytological observations were additionally Fig.6. pmcd1 pre-meiotic flower buds show a significant increase in endoploidy. Comparative endoploidy distribution in wild-type and pmcd1 leaves (A) and flower organs (B), determined by DNA flow cytometry. Significant differences between pmcd1 and the wild type were determined by Students t-test analysis (*P<0.05). performed in somatic tissue types. In wild-type petals, adaxial epidermal pavement cells showed a typical continuous jigsawshaped wall that fully enclosed each cell from the adjacent cells (Fig. 7A). Although epidermal patterning in pmcd1 petals was highly similar to that of the wild type, occasional defects in cell plate formation and nuclear configuration were observed (<5%). Indeed, pmcd1 petal cells occasionally exhibited incomplete cell walls and cell wall stubs (Fig. 7A, arrows), indicating defects in somatic cell proliferation or cytokinesis. In addition, pmcd1 petals also displayed a small number of cells containing two normally sized nuclei (syncytes) or one larger nucleus (polyploid), often coinciding with defects in cell plate formation (Fig.7A, asterisks). These observations confirm the earlier hypothesis that defects in cell plate formation yield syncytial cells that may perform nuclear fusion to generate endomitotic polyploidcells. To monitor whether other pmcd1 somatic organ types also exhibit defects in cell wall establishment, the epidermal cell layer was next assessed in cotyledons of 7-day-old seedlings. In contrast to wild-type cotyledons, which do not show any defect in cell wall formation (Fig. 7B, a), pmcd1 seedlings occasionally showed similar alterations to those observed in pmcd1 petal tissue (Fig.7B, b, c; see arrows). In both cotyledons and petals, however, the frequency of cells containing incomplete cell walls was fairly low (<5%), indicating that pmcd1 is a weak cytokinesis-defective mutant. Stomatal clustering in pmcd1 points to alterations in plasmodesmatal callose deposition all stomata appear clustered, and additionally revealed that these clusters consist of no more than two adjacent stomata (Supplementary Table S1). In contrast, pmcd1 cotyledons exhibit a significant number of stomatal cell clusters and occasionally display more than two stomata in one cluster (Fig.7B, e, f). Quantitative analysis revealed that, although the majority of pmcd1 stomata appear as single units, a significant percentage groups in clusters of two (4.8 1.5%) or three stomata (0.5 0.7%) (Supplementary Table S1). Although alterations in stomatal patterning can be caused by a defect in one of the molecular factors governing stomatal cell proliferation (Nadeau, 2009), the concomitant presence of altered cell walls rather suggests that pmcd1 has a defect in the formation and/or deposition of callose, both at the newly formed cell plate and at the plasmodesmatal (PD) channels (Guseman etal., 2010). To monitor deposition of callose at PD channels, aniline blue staining of 7-day-old seedlings was performed. In wild-type cotyledons, the aniline blue staining revealed an abundance of PD callosic plugs uniformly spread along the epidermal cell walls (Fig.7C, a, b), indicating a proper deposition of callose at the plasmodesmata. In pmcd1 cotyledons, epidermal cells also showed the presence of some PD callosic plugs; however, the general frequency and intensity of aniline blue fluorescent dots were substantially lower compared with the wild type (Fig.7C, c, d). We therefore conclude that pmcd1 is impaired in plasmodesmatal callose deposition. pmcd1 is not affected in phytosterol biosynthesis Besides alterations in cell wall formation, it was also observed that pmcd1 cotyledons display distinct alterations in stomatal patterning. In wild-type cotyledons, stomata always appear as single units randomly dispersed along the adaxial leaf side (Fig. 7B, d). Quantitative analysis of wild-type cotyledons demonstrated that only 0.2 0.2% of Defects in somatic cell wall establishment and the associated formation of tetraploid PMCs through pre-meiotic endomitosis have also been observed in the Arabidopsis mutants frill1 and cvp1-3 (De Storme etal., 2013). Both lines have a mutation in STEROL METHYL TRANSFERASE 2 (SMT2) (Carland et al., 2002; Hase et al., 2005), encoding a major sterol biosynthesis enzyme which operates at the branching point between phytosterol and brassinosteroid biosynthesis. In both frill1 and cvp1-3, loss of SMT2 function reduces the relative content of phytosterols (e.g. stigmasterol and sitosterol) and leads to an increased level of brassinosteroids (e.g. brassinolide). To check whether defects in cell wall establishment and the associated formation of polyploid meiocytes in pmcd1 are caused by alterations in sterol biosynthesis, the relative content of the most abundant sterols (e.g. sitosterol, stigmasterol, and campesterol) was quantified in pmcd1 leaves. For all three sterols analysed, the relative content per mg fresh weight (FW) in the pmcd1 background appeared nonsignificantly different from the wild type (Supplementary Fig. S3a), indicating that pmcd1 has a normal level of sterols in its somatic tissues. This was phenotypically confirmed by the absence of morphological defects in cotyledonal vein patterning in pmcd1 seedlings (Supplementary Fig. S3a) (Carland etal., 2002). Pre-meiotic doubling as a mechanism of 2n gamete formation inplants In this study, the tomato mutant pmcd1 which generates diploid as well as tetraploid meiocytes in male sporogenesis is presented. Tetraploid pmcd1 meiocytes are the result of premeiotic nuclear fusion that ectopically occurs in flower tissue as well as in vegetative somatic tissue. It was additionally found that pmcd1 tetraploid PMCs predominantly generate diploid pollen together with a minor subset of aneuploid spores. However, despite proven metabolic activity and in vitro germination capacity, diploid pmcd1 pollen never yielded polyploid offspring. The absence of triploid progeny is most probably caused by the strong triploid block in tomato (Nilsson, 1950), which means that triploid seeds are not able to develop due to an imbalanced maternal/paternal genome dosage in the endosperm (EBN; endosperm balance number) (Scott etal., 1998). Indeed, although in some cases triploid tomato plants have been recovered through interploidy crosses (Kaganzur etal., 1991), these events are rare and thus confirm the existence of a strong triploid block in tomato, as has also been reported for other Solanaceae species (Ehlenfeldt and Ortiz, 1995; Ramsey and Schemske, 1998). In plants, formation of diploid gametes generally occurs via a cytological alteration of the reproductive programme. As such, diploid gametes can be generated post-meiotically by an ectopic genome duplication event occurring during micro- or megaspore development (Bastiaanssen etal., 1998). In most cases, however, diploid gametes are formed by a restitution of the meiotic cell cycle, which then generates dyads and/or triads instead of the normal tetrads (Brownfield and Kohler, 2011). In addition to meiotic restitution and post-meiotic genome doubling, pre-meiotic endomitosis is here presented as an alternative mechanism for diploid and polyploid spore formation in plants. In contrast to meiotic cytomixis, which constitutes the migration and occasional fusion of chromosomes through cytoplasmic channels during the early steps of meiosis (Mursalimov and Deineko, 2011), pre-meiotic endomitosis involves a complete fusion of nuclei before the initiation of meiosis, in the archesporal germ cell lineage. As a consequence, pre-meiotic endomitosis results in the stochastic formation of tetraploid PMC initials, which then proceed through meiosis as a normal tetraploid meiocyte. In animals, pre-meiotic genome doubling has repeatedly been observed in unisexual and parthenogenetically reproducing species, more particularly as a method to alleviate hybrid-associated reproductive sterility or to generate clonal gametes. For example, in female salamanders of the genus Ambystoma, pre-meiotic endomitosis doubles the absolute number of chromosomes in oogonial cells so that, after meiosis, the mature eggs have the somatic ploidy level (Bi and Bogart, 2010). Similarly, in other unisexual vertebrates, such as loach, hybrid medakas, hybrid frogs (e.g. Rana), and parthenogenetically reproducing lizards, pre-meiotic endomitosis generates meiocytes with a doubled chromosome number, yielding gametes that have the somatic DNA content (Cuellar, 1971; Heppich et al., 1982; Shimizu et al., 2000; Itono et al., 2006; Yoshikawa etal., 2009). In contrast, in the plant kingdom, the spontaneous formation of polyploid meiocytes has only been documented for a small number of species. Moreover, in all these cases, the cellular mechanism underlying PMC genome duplication has never been documented. In Brassica, Mason etal. (2011) presumed that the giant tetrads, occasionally observed in some specific hybrids, were polyploid and resulted from a pre-meiotic genome duplication event. Similarly, in tetraploid Turnera hybrids, PMCs with a doubled amount of chromosomes were also thought to result from a chromosome doubling event occurring before initiation of meiosis (Fernandez and Neffa, 2004). In oat, Katsiotis and Forsberg (1995) reported that a specific line produces PMCs with double the amount of chromosomes and suggested that these PMCs result from a failure of cell wall formation during the last mitotic division that precedes meiosis. Based on the latter finding, it is concluded that the process of pre-meiotic endomitosis, as observed in pmcd1, is a naturally occurring phenomenon and constitutes an alternative cellular basis to generate 2n gametes. Similarly, fusion of syncytial nuclei and the associated formation of tetraploid PMCs has also been observed in diploid Chrysanthemum (Kim et al., 2009). However, in contrast to pmcd1, which shows ectopic endomitosis prior to meiosis initiation, nuclear fusion in Chrysanthemum occurs in the early phase of meiosis, more specifically in prophase I.It is therefore concluded that the cellular mechanism underlying 2n gamete formation in pmcd1 provides new insights into the developmental relevance of cytokinesis and cell cycle control in plant sporogenesis and putatively constitutes an alternative route for 2n gamete formation and sexual polyploidization in natural plant populations. pmcd1 suggests a molecular control of pre-meiotic ploidy stability The mutation-based induction of pre-meiotic endomitosis in pmcd1 indicates the existence of one or more molecular factors that regulate the ploidy level of the meiocyte-producing cell lineage. Genetic studies in Arabidopsis have already revealed some proteins that play a role in the regulation of the pre-meiotic genome and ploidy stability. Henriques et al. (2010) reported that the Arabidopsis 40S ribosomal protein S6 kinases S6K1 and S6K2 play an important role in the ploidy maintenance of meiocyte initials. Both a hemizygous s6k1s6k2/++ plant and S6K1 RNAi (RNA interference) lines generate tetraploid PMCs together with the ectopic formation of endomitotic nuclei in somatic cells, indicating that loss of S6K function induces endomitosis, in both somatic and reproductive cell lineages. S6K1 associates with the RBR1-E2FB complex and thereby regulates nuclear localization of RBR1 and E2F-dependent expression of cell cycle genes (Henriques etal., 2010). As such, pre-meiotic endomitosis in S6K loss-of-function lines was suggested to originate from defects in cell cycle progression or cell proliferation. Although pmcd1 showed a similar phenotype, s6k-specific somatic polyploidy and aneuploidy and a high level of pollen abortion were not observed, indicating that the induction of pre-meiotic endomitosis in pmcd1 is most probably not caused by a loss of S6K function. Recently, two other Arabidopsis proteins were found to play a role in the ploidy maintenance of PMC initials, namely glucan synthase-like 8 (GSL8) and sterol methyl transferase 2 (SMT2) (De Storme etal., 2013). GSL8 is a callose synthase that is required for the production and deposition of callose at newly formed cell plates and plasmodesmata (Chen etal., 2009; Guseman etal., 2010). SMT2, on the other hand, is a sterol biosynthesis enzyme which operates at the branching point of phytosterol and brassinosteroid biosynthesis. Loss of SMT2 function and a specific mutant allele of GSL8 both induce the ectopic formation of tetraploid PMCs, typically yielding viable diploid spores. Moreover, both mutants show non-lethal defects in somatic cell plate formation and the occasional fusion of syncytial nuclei, indicating that polyploid smt2 and gsl8 meiocytes originate from pre-meiotic endomitosis, ectopically induced by defects in cell wall formation (De Storme etal., 2013). In pmcd1, several types of somatic tissue (e.g. cotyledons and petals) display similar defects in cell wall formation. In addition, plasmodesmatal deposition of callose appears strongly affected, without exhibiting significant changes in the sterol profile. Since these aberrations are highly similar to defects in gsl8, it is presumed that pmcd1 defects in cell wall formation are caused by an altered deposition of callose at the developing cell plate, most probably by an alteration of callose synthase functionality. PMCD1 is therefore not directly involved in the molecular regulation of premeiotic endoploidy stability, but rather constitutes a general factor involved in plant cytokinesis, supporting the notion that proper cell wall formation is essential for the genomic stability and ploidy maintenance of the reproductive lineage. Based on the minor alterations in cell plate formation in several organ types, pmcd1 can be considered as a weak, non-specific cytokinesis-defective mutant. Interestingly, the frequency of cytokinesis defects and the associated formation of polyploid cells appeared much higher in the pre-meiotic germ lineage (~20%) compared with levels observed in other tissues. Asimilar discrepancy has also been reported for the mild cytokinesis-defective Arabidopsis smt2 and et2 mutants (De Storme etal., 2013). Based on these findings, it is postulated that the pre-meiotic germ lineage is extremely susceptible to defects in cell plate formation and thereby constitutes a putative mechanism for the developmental-specific induction of polyploidy in the germ lineage and the associated induction of sexual polyploidization. Besides alterations in the reproductive organs, pmcd1 also shows phenotypic aberrations in its progeny seeds. Indeed, despite having the basic diploid ploidy level, pmcd1 seeds appeared substantially larger compared with the wild-type control. Although the larger pmcd1 seed size could be attributed to a mutation in another gene, genetically linked to PMCD1, the possibility that it is a pleiotropic effect of the diploid pollen-causing pmcd1 mutation cannot be excluded. Moreover, as altered cell cycle regulation and defective cytokinesis have already been found to yield larger seeds in several Arabidopsis mutants (Liu and Meinke, 1998; Li etal., 2008), it is assumed that the mutation underlying the cytokinesis aberrations and ectopic endomitosis in pmcd1 also forms the basis for enlarged seed formation. However, to determine the cellular basis of pmcd1 enlarged seed size and to assess the genetic determinant(s) underlying this aberrant phenotype, more extensive cytological and genetic (e.g. mapping) studies are needed. Tetraploid meiosis in a diploid plant: a tool to study polyploid meiosis The newly isolated tomato mutant pmcd1 generates a substantial number of tetraploid meiocytes in a diploid background, without affecting the general growth and development of the PMC-producing mother plant. Based on this unique characteristic, pmcd1 offers new perspectives for the genetic and molecular analysis of autopolyploid meiosis in plants. Indeed, as the elucidation of molecular factors involved in the regulation of autopolyploid meiosis is hampered by technical difficulties related to the increased chromosome number (e.g. low frequency of recessive mutants, complex mapping), pmcd1 and other mutants that produce tetraploid meiocytes in a diploid background circumvent these ploidy-related issues and thus enable the study of polyploid meiosis using well-established diploid-based (cyto)genetic and molecular approaches. Supplementarydata Supplementary data are available at JXB online. Figure S1. pmcd1 produces largerseeds. Figure S2. DNA flow cytometry of leaf material isolated from wild-type and pmcd1 F1 progeny. Figure S3. pmcd1 tetraploid meiocytes generate both biand tetravalents at meiosisI. Figure S4. Wild-type (a) and pmcd1 (b) female meiocytes at prophaseI. Figure S5. pmcd1 is not affected in sterol biosynthesis. Table S1. Frequency of stomatal clustering in cotyledons of 7-day-old seedlings. Acknowledgements The authors wish to thank Vera Wewer and Peter Drmann (University of Bonn) for performing sterol measurements, and Rob Dirks and Eveline van der Zeeuw (Rijk Zwaan Breeding B.V.) for providing plant material. This research is supported by an aspirant fellowship to NDS and research grant G006709N offered by the Flemish Funding Agency for Scientific Research (FWO). Collaborations and travel were supported by the COST action FA0903. mutation in Arabidopsis CHORUS (GLUCAN SYNTHASE-LIKE 8). Development 137, 17311741.

This is a preview of a remote PDF:

Nico De Storme, Danny Geelen. Pre-meiotic endomitosis in the cytokinesis-defective tomato mutant pmcd1 generates tetraploid meiocytes and diploid gametes, Journal of Experimental Botany, 2013, 2345-2358, DOI: 10.1093/jxb/ert091