Multicellular plants with a known genome sequence possess two or more copies of a DEFECTIVE EMBRYO AND MERISTEMS (DEM) gene. These genes are highly conserved between land plants, and distant and less-conserved homologs exist in other diverse eukaryota. The high conservation and retention of the DEM gene family in multicellular plants suggests that these genes have an important biological function in these organisms. In dicot species such as Arabidopsis and tomato, there are two DEM genes, namely DEM1 and DEM2. The DEM1 gene was first identified and cloned by transposon tagging with a Dissociation (Ds) element in tomato (Keddie et al., 1998). The insertion mutant in tomato, dem1Ds, lacks organised cell division in shoot and root apical meristems, and fails to grow beyond the seedling stage of development. Thus, the DEM1 gene is required for organised cell divisions during embryonic and post-embryonic growth in tomato (Keddie et al., 1998).
Prior to this study, Hawker (2004) characterised the dem1-1 and dem2-1 mutants of Arabidopsis ecotype Wassilewskija (Ws), and the results indicated that DEM1 and DEM2 of Arabidopsis are functionally redundant but required for cell divisions during male and female reproductive development. Here, we confirmed and extended the characterisation of these mutants and two additional mutants, dem1-2 and dem2-2, of ecotype Columbia (Col-0). Due to genetic redundancy, the Arabidopsis single mutants, dem1 and dem2, are viable during most stages of plant development. Plants harbouring both dem1 and dem2 alleles, however, exhibited meiotic and/or mitotic cell division defects during microgametophyte (pollen) and megagametophyte (female gametophyte) development.
The transmission of the dem1 dem2 genotype was almost abolished through the male germline of Arabidopsis, indicating the essential requirement for DEM expression during pollen development. The presence of DEM in both developing pollen and the sporophyte (adult plant tissue) was found to enhance pollen viability.
DEM was also required for development of megagametophytes in Arabidopsis. However, our results showed that many dem1 dem2 megagametophytes are viable, and that the major requirement for DEM in megagametophyte development is in the sporophyte. Fully homozygous adult dem1 dem2 plants were never recovered, and the dosage of DEM in the male parent contributed to the post-fertilization fate of dem megagametophytes.
In agreement with the requirement of DEM in tissues undergoing organised cell division, high-level expression of GFP-tagged DEM1 under the control of the DEM1 promoter was observed in developing pollen, sporophytic tissues of the developing ovule, embryos and the root apical meristems of Arabidopsis. The GFP-tagged DEM1 was found to localise to the nuclear envelope and/or the cytoplasm of microspores and cells in developing pollen and the root apical meristem. Expression of GFP-tagged DEM1 significantly suppressed pollen defects and restored viability to dem1 dem2 male gametes in dem plants. Thus, the expression pattern and subcellular localisation of GFP-tagged DEM1 was likely to be a reliable reporter for endogenous DEM1 in microspores and during microgametogenesis.
Phenotypic analysis of plants harbouring dem1 and dem2 mutants and GFP-tagged DEM1 localisation studies in Arabidopsis, suggested that DEM1 interacts with proteins required for cell division at the nuclear envelope or in the cytoplasm. Earlier, a two-hybrid screen identified Ras-related Nuclear protein-1 (RAN1) as a potential interacting partner of DEM (Matthew, 2003). RAN is a GTPase that is highly conserved between plants, fungi and animals. In animals and yeast, RAN has been shown to play key roles in cell cycle progression, reformation of the nuclear envelope following cell division, and nucleo-cytoplasmic transport. Here, we provide additional evidence that DEM1 interacts with RAN1 from pull-down binding assays, and that DEM1 and RAN1 fused to GFP and RFP, respectively, co-localize at the nuclear envelope of microspores, microgametophytes (pollen) and root tip cells. Furthermore, the sub-cellular localisation of DEM1 resembles other Ran binding-proteins (RanBPs) that typically localise to the nuclear envelope. Together, our results indicate that DEM mediates its effects on cell division by interacting with RAN at the nuclear envelope.
We also discovered a novel phenomenon in male gametophyte development of tomato. Three independent dem1+7 or dem1+4 frameshift alleles were generated by excision of the Ds transposon from dem1Ds, leaving a 7 or 4 bp frameshift mutation at codon 121 of the DEM1 coding sequence. Homozygous dem1 frameshift and dem1Ds mutants showed identical defects in embryo and post-embryonic development. Surprisingly however, in DEM1/dem1 plants, all three independent dem1 frameshift alleles but not dem1Ds were preferentially transmitted five to nine times greater through the male germline compared to the DEM1 wild-type allele. The preferential transmission through the male germline has been confirmed over two generations for one dem1 frameshift allele, demonstrating the phenotype is linked to the mutant dem1 allele. Thus, our results indicate that the preferential transmission of each of these three independent dem1 frameshift alleles through the male germline of DEM1/dem1 plants, is conferred by the dem1 allele itself, and not by mutations in other genes linked in coupling or repulsion on the same chromosome. The phenomenon of preferential transmission of a mutant allele through the germline is very rare, and to our knowledge, this is the first reported case of such an allele in plants. Clearly, this genetic material provides a unique opportunity to investigate novel aspects of microgametophyte function and male reproductive development in plants.