Rice Genetics IV - IRRI books - International Rice Research Institute

aposporous apomicts such as H. aurantiacum and H. piloselloides, both embryo andendosperm form spontaneously inside an unreduced embryo sac (Koltunow et al 1998).Both fis production (Peacock et al 1995) and fie production (Ohad et al 1996) havebeen achieved by mutagenesis in Arabidopsis. Should we attempt to produce someequivalent form of fie formation in a synthetic apomictic of rice? Or should we retainthe sexual process to preserve the triploidy of the endosperm? Savidan (2000) presentedan argument in favor of retaining triploidy. We examine both sides of the argumentbelow.Triploidy results when the maternal binucleate central cell of the embryo sac isfertilized by a paternal sperm nucleus. The genotype of the endosperm is representedas 2m:1p. In maize, any departure from this genotype has deleterious effects on thequality and quantity of the endosperm (Birchler 1993). One explanation for this effectinvokes imprinting, the phenomenon in which expression of alleles differs dependingon whether they originate from the male or female parent. Imprinting in the endospermof angiosperms has been shown to explain most failures of interploidy or interspecificcrosses in plants (Haig and Westoby 1991, Birchler 1993). Because of imprinting,seeds develop normally only if the 2m:1p dosage is represented in the endosperm.Not all species appear to be strongly imprinted for the 2m:1p genotype. InTripsacum, all polyploids reproduce through diplospory. Meiotic failure inmegasporocytes leads to the development of eight-nucleate unreduced megagametophytes.Microgametophytes remain unaffected. Grimanelli et al (1997) used flowcytometry to determine ploidy levels in the endosperm of both apomictic and sexualTripsacum accessions. In both cases, fertilization appeared to involve only one spermnucleus. Endosperm of apomictic Tripsacum developed normally even though theratio of genomic contributions deviated from the normal 2m:1p ratio. Ratios of 2:1,4:1, 4:2, 8:1, and 8:2 were observed, depending on both the ploidy level of the parentsand the mode of reproduction. Thus, specific dosage effects are seemingly not requiredfor endosperm development in Tripsacum. Grimanelli et al (1997) suggest thatevolution of diplosporous apomixis might have been restricted to species with few orno imprinting requirements.In work on the transfer of apomixis from Tripsacum to maize, Sokolov andKhatypova (1999) encountered problems with small seed size and high sterility (associatedwith ovary death in the early stages of development in intergeneric hybrids).The material used comprised F 1 and backcross hybrids from the crossing of tetraploidmaize with Tripsacum (2n = 56). Hybrids were obtained with different ratios of completeparental genomes or of complete maize genomes and Tripsacum subgenomes orchromosomes. Sokolov and Khatypova (1999) concluded that the availability ofTripsacum chromosomes in some way inhibited imprinting expression of the maizeparental genome in the hybrids. For total suppression, a critical number of nineTripsacum chromosomes was needed.Imprinting in Arabidopsis is less marked than in maize (Scott et al 1998). Thismay reflect the fact that Arabidopsis is inbreeding while maize is outcrossing. Theparental conflict model of imprinting (Haig and Westoby 1991) predicts that the parentinvesting the most in reproduction should have the major influence on the out-Molecular tools for achieving synthetic apomixis in hybrid rice 385