Development of synthetic Brassica amphidiploids by reciprocal hybridization and comparison to natural amphidiploids.

In a previous study we proposed that cytoplasmic genomes have played an important role in the evolution of Brassica amphidiploid species. Based on this and other studies, we hypothesized that interactions between the maternal cytoplasmic genomes and the paternal nuclear genome may cause alterations in genome structure and/or gene expression of a newly synthesized amphidiploid, which may play an important role in the evolution of natural amphidiploid species. To test this hypothesis, a series of synthetic amphidiploids, including all three analogs of the natural amphidiploids B. napus, B. juncea, and B. Carinata and their reciprocal forms, were developed. These synthetic amphidiploids were characterized for morphological traits, chromosome number, and RFLPs revealed by chloroplast, mitochondrial, and nuclear DNA clones. The maternal transmission of chloroplast and mitochondrial genomes was observed in all of the F1 hybrids examined except one hybrid plant derived from the B. rapa x B. oleracea combination, which showed a biparental transmission of organelles. However, the paternal chloroplast and mitochondrial genomes were not observed in the F2 progeny. Nuclear genomes of synthetic amphidiploids had combined RFLP patterns of their parental species for all of the nuclear DNA clones examined. A variation in fertility was observed among self-pollinated progenies of single amphidiploids that had completely homozygous genome constitutions. Comparisons between natural and synthetic amphidiploids based on restriction fragment length polymorphism (RFLP) patterns indicated that natural amphidiploids are considerably more distant from the progenitor diploid species than the synthetic amphidiploids. The utility of these synthetic amphidiploids for investigating the evolution of amphidiploidy is discussed.