Maternal genotype influences pea seed size by controlling both mitotic activity during early embryogenesis and final endoreduplication level/cotyledon cell size in mature seed

Nathalie Munier-Jolain 0 1 Jean-Pierre Boutin 0 1 0 INRA, Station d'Agronomie, BV 1540, 21034 Dijon Cedex, France 1 INRA, Unit e de Recherche en Ge ne tique et Am elioration des Plantes , BV 1540, 21034 Dijon cedex, France 2,4 4 To whom correspondence should be addressed. Fax: +33 3 8063 3263. E-mail: - When reciprocal crosses are made between different pea genotypes, there is a strong maternal influence on mature seed size of the reciprocal hybrids, i.e. their dry weights are similar to that of seeds obtained from their maternal parents. Reciprocal crosses between pea varieties having very different mature seed sizes were used to investigate how the maternal genotype controls seed development and mature seed size. The differences in dry seed weight between genotypes and reciprocal hybrids reflected differences in both cotyledon cell number and mean cell volume, and the maternal control on the establishment of these two traits was investigated. Using flow cytometry, data relative to endoreduplication kinetics in cotyledons during the transition between the cell division phase and maturation were obtained. The appearance of nuclei having an 8C DNA content indicates the initiation of the endoreduplication phenomenon and thus the end of the cell division phase. It was shown that the duration of the cell division phase was the same in the reciprocal hybrids, its value being intermediate between those recorded for their maternal parents. This result indicates that the timing of development of the embryo is not under maternal control, but depends on its own genotype. Consequently, maternal genotype must influence the mitotic rate during the cell division phase to achieve differences in cell number found in the cotyledons of mature F1-reciprocal hybrids. The final level of endoreduplication in cotyledons of mature seeds was also investigated. This study showed that 2 there is a close relationship (r =0.919) between the endoreduplication level in mature cotyledons and seed dry weight or mean volume of cotyledon cells, suggesting that both maternal and non-maternal factors could control the number of endoreduplicating cycles in the cotyledons and, hypothetically, the cotyledon cell size. Pea seed development has already been described in many studies (Bain and Mercer, 1966; Smith, 1973; Hedley and Ambrose, 1980). It can be divided into three distinct phases. In the first, the cell division phase, the cotyledon cells actively divide. In the second phase, maturation, the cotyledon cells expand, and reserve compounds (starch and proteins) are stored. The third phase concerns seed desiccation. At the end of the initial phase of development, the number of cells in the cotyledons is established (Smith, 1973). In legume seeds, cotyledons represent the major storage organ, as the endosperm is restricted to a nutrientrich apoplastic liquid which is almost totally resorbed at the beginning of the maturation (Marinos, 1970). Studies in diVerent legume species such as faba bean (Davies, 1977), pea (Davies, 1975) and soybean ( Egli et al., 1981) have established a positive correlation between cotyledon cell number and mature seed size. The number of cells formed in the cotyledons determines the capacity of the storage organ to accumulate dry matter (Munier-Jolain and Ney, 1998 ). Some previous studies of parental eVects have shown that the maternal genotype influences mature seed mass in pea (Davies, 1975 ). However, the way by which the maternal genotype aVects seed development is not understood. The presence of cells of diVerent ploidy levels in somatic tissues is called endoreduplication or polyteny. In contrast to dividing cells, endoreduplicating cells are not believed to undergo mitosis, and in such cells, nuclear DNA content successively doubles from 2C to 4C to 8C to 16C etc., where C is the haploid DNA content per nucleus. Endoreduplication was first described during seed development. In maize developing endosperm, the average DNA content per nucleus increases sharply, as the mitotic index decreases. This increase can reach levels of 384C in some individual nuclei (Schweizer et al., 1995 ). During the early period of field pea (Pisum arvense) seed development, the DNA content of the cells remains at the diploid level until cell division is complete, after which it begins to increase (Smith, 1973). Using microdensitometry, a C-value of 64 has been measured in cotyledon cells in two pea genotypes (Davies and Brewster, 1975 ). In Arabidopsis thaliana, endoreduplication occurs in cells of the hypocotyl during the elongation of this organ, and reaches a 16C value in dark-grown seedlings (Gendreau et al., 1997). Although endoreduplication has been already described in several plant species, its significance still remains uncertain. It is often related to nuclear genome size and/or cellular dimension (Galbraith et al., 1991; Melaragno et al., 1993; Gendreau et al., 1998). This suggests that the endoreduplication phenomenon may be related to the mature seed size, and may be more important in large seeds than in small seeds. In order to investigate how the maternal genotype controls seed development and mature seed size, reciprocal crosses between four varieties of pea having diVerent seed sizes were used. The stage of appearance of the endoreduplication phenomenon in cotyledon cells was investigated during the transition period between cell division phase and maturation using flow cytometry. Next, the endoreduplication level in the mature seed of nine pea varieties and hybrids was described in order to study the relationship between this trait and cotyledon cell number and cotyledon cell volume. Materials and methods Plant material and growing conditions Garden pea (Pisum sativum L.) is a diploid (2n=14) and autogamous plant, so varieties used are considered as pure homozygous lines. Five varieties were used, which show large diVerences in their mature seed weights: cvs GSP6, Cation and Frisson have a low seed weight (respectively 58, 180 and 247 mg in non-limiting conditions) whereas cvs Solara and Imposant have a high seed weight (413 and 548 mg). Plants were grown in pots filled with expanded clay in the greenhouse during the spring of 1996 and 1997. They were supplied daily with a complete nutrient solution (Lesaint and Coc, 1983). Reciprocal crosses were made manually between cvs Frisson and Solara, and cvs Cation and Imposant, on floral buds before the natural self-pollination occurred. No more than two pods were grown on each plant, so that the supply of nutrients to the pod was never limiting for its growth. Temperature variations were measured from the beginning of flowering until sampling. Time was expressed in cumulative degree-days after pollination (C DAP), using 0 C as the base temperature ( Eteve and Derieux, 1982). Seeds were sampled first at 150 C DAP, and then 1, 2, 3, 4, and 5 d after this date. Seed coats and apoplastic liquid were removed, and embryos were (...truncated)