Doubled-haploid broccoli production using anther culture: effect of anther source and seed set characteristics of derived lines.

摘要

Using anther culture to generate doubled-haploid (DH) homozygous lines for use as parents in F1 hybrid crosses is common practice in breeding broccoli (Brassica oleracea var. italica). During anther culture and subsequent embryogenesis and plant regeneration, polyploidization of microspore-derived embryos may not occur or it may occur accompanied by a doubling, tripling, quadrupling, octupling or irregular polyploidization of the genome. Thus regenerants from the process can be haploids, diploids,triploids, tetraploids, octaploids or aneuploids. Repeat cycles of broccoli anther culture was conducted using a group of F1 hybrids as anther donors to develop populations of regenerants. The resulting populations were analysed using DNA flow cytometry,and the influence of F1 source on frequency of different ploidy levels among regenerants was determined. Seed set in broccoli inbreds developed in a traditional selfing programme was compared to that in DH broccoli derived from anther culture. In 2 cycles (1994 and 1995) of anther culture, anther-derived populations of regenerants were developed using the F1 hybrids Marathon, Everest, High Sierra and Futura as sources of anthers. In 1994, Everest, High Sierra and Futura yielded populations that included 2-7% haploids, 53-56% diploids, 32-38% tetraploids and 5-6% others. Marathon-derived regenerants were 5% haploid, 78% diploid, 15% tetraploid and 2% other, showing significantly more diploids. In 1995, Marathon regenerants included significantly more diploids and fewer tetraploids than those derived from other F1 sources, suggesting that the genotype of the anther source affects the frequency of a particular ploidy level among regenerants derived from culture. In manual self-pollinations of the 1994 regenerants, only diploids and rare tetraploids set seed. When plants that set no seed were discounted, seed production following manual self pollinations of 1995 regenerants was not significantly different from that of traditional inbreds derived from the same F1 sources.