Highlights of meiotic genes in Arabidopsis thaliana

Meiosis is a fascinating and complex phenomenon and, despite its central role in sexual plant reproduction, little is known on the molecular mechanisms involved in this process. We review the progress made in recent years using Arabidopsis thaliana mutants for isolating meiotic genes. In particular, emphasis is given on the description of mutants affecting either the regular commitment to meiosis, or the mechanisms of synapsis, recombination, and cytokinesis. We believe that the isolation of genes affecting some crucial meiotic events may represent the first step towards the practical use of meiotic genes in plant breeding. The introduction of deviations in the meiotic pathway into sexual crops will have important implications for the exploitation of apomixis and sexual polyploidization.


INTRODUCTION
Meiosis is a particularly significant process in sexual plant reproduction. The complexity of the events involved, such as DNA replication, chromosome pairing and synaptonemal complex formation, suggests that many tightly regulated genes are involved in this multistep process. Many meiotic genes have been reported in yeast, Drosophila, Caenorhabditis elegans, humans (Schwarzacher, 2003). As far as plants are concerned, species like maize, tobacco, and pea have provided good *Corresponding author. Tel: ++39-081-2539225. Fax: ++39-081-7753579. E-mail: carputo@unina.it.
Abbreviations. Ath: Arabidopsis thaliana. SC: synaptonemal complex. DSBs: Double Strand Breaks. models to understand the cytological events and the genetic control of meiosis. However, these species still represent difficult systems for the molecular characterization of the genes involved in meiosis, mainly for the difficulty of using functional genetic approaches.
Recently, Arabidopsis thaliana has become the model plant to study meiosis. The availability of a large number of T-DNA and transposon mutated lines represents a powerful tool to isolate meiotic mutants and to identify the corresponding mutated genes. The scientific approach used recentely in Arabidopsis, as in yeast and other eucaryotic organisms, is to dissect meiosis into single events through the isolation of mutants specific for each meiotic step.
This review focuses on the progress made in the recent years in the identification of premeiotic and meiotic mutants in A. thaliana (Ath) through innovative approaches. We believe that in the near future this will contribute to a full understanding of meiosis in plants and will open new perspectives for the biotechnological manipulation of plant reproduction.

PREMEIOTIC MUTANTS
The discovery of premeiotic mutants and the related genes could provide an insight into the mechanisms involved in the early step of sporogenesis for meiocyte differentiation and meiosis commitment. This is very important in that, despite decades of research in plant reproduction, there is only limited knowledge of how sporocytes differentiate and how meiosis is initiated and regulated in plants.
One premeiotic mutant defective in the specification of reproductive cell type has been discovered in Ath by Yang et al. (1999) and Schiefthaler et al. (1999). They identified sporocyteless/nozzle (spl/nzz) mutant characterized by archesporial cells unable to differentiate into both megasporocytes and microsporocytes. Molecular cloning of the SPL gene by both research teams provided evidence that it encodes a novel nuclear protein related to MADS box transcription factors. The sequence of SPL suggests that the gene product functions as a transcriptional regulator essential for sporocyte development.

MEIOTIC MUTANTS
Several Ath mutants defective in chromosome pairing and segregation, and in cytokinesis have been identified (Table 1). Their phenotypes are always associated with a specific stage of the meiotic cycle (Figure 1), and in most cases either microsporogenesis or macrosporogenesis are affected. Given the difference in form and function between male and female gametes, it is expected that the two pathways of reproductive developments are under the control of overlapping but distinct genetic programmes. From genetic evidence it is known that different genes are involved in male (Hulskamp et al., 1997;Yang et al., 1999) and female meiosis (Byzova et al., 1999;Siddiqi et al., 2000), and only few of them control both processes (Couteau et al., 1999;Grelon et al., 2001).

Prophase I
Most of Ath meiotic mutants recently isolated are defective in prophase I, a crucial meiotic stage. Indeed, synapsis, synaptonemal complex (SC), crossing-over, and chiasmata formation that lead to recombination events between homologous chromosomes and subsequently to a proper segregation, occur during this stage. At least two Ath mutants are affected in sister chromatid cohesion; switch1/dyad and syn1/dif1.
Three alleles have been isolated with mutation in the SWI1/DYAD locus: switch1-1 (Motamayor et al., 2000), switch1-2 (Mercier et al., 2001) and dyad (Agashe et al., 2002). All the three alleles cause similar defects in the female meiosis pathway. By contrast, only swi1-2 causes aberrations in microsporogenesis. The majority of megaspore mother cells (MMCs) perform a mitotic-like division instead of meiosis I, exiting meiosis with an unreduced chromosome number. However, the MMCs are not committed to a mitotic programme but to a meiotic cycle, and the two daughter cells in the dyad perform a second cell cycle, which is either a novel mitosis-like division or an aberrant one, with unequal chromosome segregation. At the end, degenerated cells are observed instead of the embryo sac. During microsporogenesis, swi1-2 mutant allele causes male sterility with severe defects in chromosome behaviour.
The SWI1 gene has been cloned and the putative protein does not show any similarity to known proteins. It is proposed that SWI1 protein promotes the establishment of sister chromatid cohesion, but it is not responsible for its manteinance. Mercier et al. (2003) reported that SWI1 is also required for the formation of SC, and for recombination. Syn1/dif1 causes both male and female sterility (Bhatt et al., 1999;Bai et al., 1999). During early leptotene and zygotene of microsporogenesis, it causes irregular chromosome condensation and extensive chromosome fragmentation. Cytological analysis showed that sister chromatid cohesion is lacking and homologous pairing does not occur (Cai et al., 2003). The SYN1/DIF1 gene has been cloned, and it shows homology to RAD21/REC8 cohesin family. Also the asynaptic Ath mutant (asy1) exhibits defects in homologous chromosome synapsis (Ross et al., 1997;Caryl et al., 2000). This mutation leads to reduced male and female fertility. ASY1 gene encodes a homologous of yeast HOP1 protein, which is essential for SC assembly and normal synapsis.
In plants, few molecular data are available on genes involved in meiotic recombination. Functional data come from the analysis of the Ath dmc1 and spo11 mutants, that show poor bivalent formation during male and female meiosis (Couteau et al., 1999;Grelon et al., 2001 (Sun et al., 1989;Cao et al., 1990) and follows the DSB repair model (Szostak et al., 1983;Sun et al., 1991), there is evidence that SPO11 protein is the endonuclease responsible for the induction of DSBs (Keeney et al., 1997). There is also indication that DMC1 is involved in the first step of DSB repair (Bishop et al., 1992). So far, the generation of DSBs in micro-and macrosporocytes has not been demonstrated in plant meiosis. However, the presence in Ath genome of yeast homologous SPO11 and DMC1 as well as of other genes involved in meiotic recombination (Klimyuk and Jones 1997;Doutriaux et al., 1998;Hartung and Puchta, 2000;Gallego and White, 2001) strongly suggests that the essential features of meiotic recombination in yeast could be conserved in plants. In addition, the Ath spo11 and dmc1 mutant phenotypes have contributed to better understand the link between synapsis and meiotic recombination in plants. Historically, it has been proposed that homologous chromosomes synapse before they recombine. However, recent data provided evidence that in yeast, synapsis is not required for meiotic recombination and that it may depend on DSB processing. In zip1 and red1 strains, for example, despite the absence of SC, recombination occurs (Rockmill and Roeder, 1990;Sym et al., 1993;1994). In addition, none of the many yeast mutants characterized so far has been found to make SC in the absence of DSBs. Plants may have a similar control mechanism, and thus synapsis is affected when SPO11 and DMC1 functions are absent. Recently, another ath mutant -solo dancers (sds) isolated by Azumi et al. (2002) -showed defects similar to spo11 and dmc1 in meiotic recombination, SC and bivalent formation. It could provide a new and important entry point to understand the regulation of homolog synapsis and recombination in plant meiosis.

Anaphase I
In other Ath mutants, the earlier meiotic events proceed normally without any visible anomaly in synapsis and recombination, but the subsequent stages are affected. Yang et al. (1999) reported on the male sterile ask1 mutant that exhibits a failure in separation of homologous chromosomes during anaphase I. The ASK1 gene has homology to the yeast SKP1, essential to regulate the mitotic cell cycle by targeting specific proteins for ubiquitin-mediated proteolysis. Until now, it was unknown that SKP1 is also required for meiosis. ASK1 may play a role in the control of homologous separation by degrading/removing a protein that is required for the homologue association in prophase I.

Tetrad stage
Mutations have been detected also in the last stage of meiosis. Ath pollenless3/tmd1 male-sterile mutant produces microspores in excess of four during the tetrad stage, usually 8 +/-1 or +/-2 (Ross et al., 1997;Sanders et al., 1999). Meiosis I and II proceed normally and the production of the 'octads', with unbalanced chromosome number, is caused by an additional division without DNA replication. It has been found that the POLLENLESS3 gene encodes a protein with limited homology to CDC23P and RAD3 proteins, involved in cell cycle of Saccharomyces cerevisiae and Schizosaccharomyces pombe, respectively.

Cytokinesis
Meiosis ends with cytokinesis, in which the four haploid nuclei formed after the two meiotic divisions are partitioned by a cell wall into four spores. Stud (std)/tetraspore (tes) mutant show normal meiotic divisions but it lacks male meiosis cytokinesis (Hulskamp et al., 1997;Spielman et al., 1997). This mutation leads to pollen of bigger size than wild type due to tetranucleate microspores. In std/tes mutant, the post-meiotic development of micropores proceed relatively normally. In fact, the nuclei are capable of independently undergoing the complete mitotic cell division leading to pollen grains with variable number of sperm nuclei. Some microspore nuclei can fuse before the first mitotic division, leading to polyploid sperm nuclei instead of the normal haploid nuclei. Defects in male cytokinesis were also observed in the tardy asynchronous meiosis (tam) mutant. However, in this case the abnormality seems a consequence of alteration in cell cycle progression during meiosis (Magnard et al., 2001). In tam, cytokinesis occurs at the end of the first nuclear division, giving a dyad with two daughter cells. In this mutant, Ath cytokinesis (that is normally simultaneous), mimics a successive type cytokinesis, typical of monocots. TAM gene maps to chromosome 1, but it has not been isolated yet. It is proposed that TAM protein positively regulates the cell cycle progression, perhaps promoting the G2/M transition, and that TAM has a role in coupling the normal pace of cell cycle progression with the synchrony of cell division during male meiosis.

CONCLUSIONS
Meiosis is a fascinating phenomenon with intriguing biotechnological potential. Given the high frequency with which new meiotic mutants are being isolated in Ath, this plant probably is the best model system for elucidating the molecular mechanisms of meiosis. Moreover, progress in meiotic studies in Ath will stimulate research on meiotic genes in cultivated species. The possibility to isolate genes altering the normal meiotic pathway and to engineer them in crops has important implications for plant biotechnology and plant breeding. Indeed, meiotic mutants can be used to exploit the advantages of apomixis in F 1 hybrid production, in sexual polyploidization crossing schemes to generate polyploid genotypes, and in promoting genetic recombination in hybrids between genetically distant parents.