Inheritance of Nucellar Embryony

The work of Parlevliet and Cameron (1959) suggests that nucellar embryony is con- trolled by a single major dominant gene that is heterozygous in trifoliate and absent in ‘Chandler’ pummelo. They also suggest minor genes may control the level of expression. Other work suggests that sev- eral genes control nucellar embryony and that polyembryony is an independent trait (Garcia et al., 1999; Asins et al., 2002). Garcia et al. (1999) evaluated the progeny of a cross between two parents known to have nucellar embryony, Citrus volkameriana (‘volkamer’ lemon) and Poncirus trifoliata var ‘Rubidoux’. The cross produced 50 fruit-yielding hybrid progeny from which seed samples were germinated. Twenty-fi ve random seedlings from each individual were genotyped with isozymes to deter- mine the seedlings’ origin as either nucellar or zygotic. Eight to ten seeds were scored for multiple embryos and the percentage of polyembryonic seed calculated for 38 of the

50 individuals in the test population. A variety of marker types were used in their

mapping analysis, with 73 polymorphic markers in P. trifoliata and 97 in C. volka- meriana. They propose a model with two quantitative trait loci (QTLs) in P. trifoliata and four QTLs in C. volkameriana control- ling apomixis, with individual QTLs con- tributing up to 24% of the total variation. Markers linked to polyembryony were found at different positions in each parent. TAA15 (a C. volkameriana marker for poly- embryony) was linked to Apo2, the QTL with the strongest effect on apomixis.

Kepiro (2003) studied inheritance of nucellar embryony in 88 progeny of a cross of Chandler pummelo ´ trifoliate orange. They scored the trait by counting the number of seedlings that germinated from an average of 283 seeds per progeny tree. Only 17 progeny produced more than 8.5% polyembryonic seeds. Mapping and QTL analysis identifi ed a major, dominant QTL that was heterozygous in trifoliate orange and which appears essential for production of more than about 2% polyembryonic seeds. DNA marker studies on a subset of seedlings supported a correlation between production of nucellar seedlings and pro- duction of a high proportion of polyembry- onic seeds. Some trees that produced 1–2% polyembryonic seed did produce nucellar embryos, but in others no apparently nucel- lar seedlings were found among seedlings from polyembyronic seeds. Therefore, the proportion of polyembryonic seeds could not be used to determine whether or not such trees have nucellar embryony. This major locus and markers associated with it showed signifi cant segregation distortion, with only about 32% of progeny producing at least 1.4% polyembryonic seed instead of the expected 50%. In those progeny having polyembryony, a second unlinked QTL accounted for much of the variation in the percentage of polyembyronic seeds. These QTLs were confi rmed in a population of open-pollinated (mostly selfed) progeny of trifoliate orange. No QTLs infl uencing poly- embryony were detected in the Chandler pummelo parent. Thus far, it has not been possible to compare the map locations of QTLs detected by Garcia et al. (1999) with

those studied by Kepiro because the maps developed do not share any common mark- ers. While the results of Kepiro suggest that inheritance of polyembryony from Poncirus involves only a few genes, it is quite possi- ble that additional polymorphic genes involved in nucellar embryony occur in Citrus or other Poncirus genotypes.

The relationship between the propor- tion of polyembryonic seeds and the pro- portion of nucellar seedlings has not been adequately tested in any segregating popu- lation. Garcia et al. (1999) tested this rela- tionship, but they examined only 25 seedling per progeny tree, too few to have a high probability of detecting polyembryony in genotypes that produce only 1–5% poly- embryonic seeds.