Résumé
Apomixis in plants covers a variety of cloning systems through seeds of great potential for plant breeding. Among long-standing approaches for crop improvement is the attempt to exploit wild relatives as natural, vast reservoirs for novel genetic variation. With regard to apomixis, maize possesses an apomictic wild relative, Tripsacum, which we used to produce advanced maize-Tripsacum hybrid generations. However, introgression of apomixis in maize has failed so far. In order to understand the hows and whys, we undertook characterization of seed development and inheritance studies in these materials. We show that apomictic seeds suffer from epigenetic loads. Both seed tissues, the endosperm and the embryo, displayed developmental defects resulting from imbalanced parental genomic contributions and aberrant methylation patterns, respectively. Progeny characterization of several maize-Tripsacum hybrid generations allowed significant progress toward the unraveling of the genetics of apomixis. First, chromosome deletion mapping showed that expression of apomixis requires one single Tripsacum chromosome. However, inheritance studies revealed that female gametes inheriting this segment were unequivalent carriers depending on their origin: unreduced gametes transmit a functional segment, whereas progeny derived from reduced ones reproduced sexually. Finally, chromosomal or genomic dosage variation barely affected the apomictic phenotype suggesting no dependency for ploidy in these materials. We conclude that epigenetic information imposes constraints for apomictic seed development and seems pivotal for transgenerational propagation of apomixis. The nature of the triggering mechanisms remains unknown as-yet, but it certainly explains the modest success relative to the development of apomictic maize thus far.