Backcross Breeding

Efforts to restore the American chestnut have been ongoing for over ninety years. Publically funded breeding programs, initiated in the 1920’s by the USDA and Brooklyn Botanical Garden, hybridized C. dentata with an Asian Castanea species resistant to chestnut blight. Hybrid progeny were not sufficiently competitive in the forested environment of C. dentata, and these early chestnut breeding programs were largely discontinued by the 1960’s.

The Backcross Method

In 1983, The American Chestnut Foundation (TACF) was founded as a non-profit to continue the species restoration effort. For three decades, TACF has pursued backcross breeding to generate hybrids that combine the pathogen resistance of Chinese chestnut (Castanea mollissima) and the timber-type growth form of American chestnut. The backcross method to introduce blight resistance from Chinese chestnut into American chestnut was first proposed by one of TACF’s founders, Dr. Charles Burnham, a renowned maize geneticist. Burnham’s rationale for backcrossing was based on the hypothesis that a few genes from Chinese chestnut are responsible for its blight resistance. Thus, it should be possible to dilute out most of the genes inherited from Chinese chestnut except for those involved in blight resistance and recover hybrids that are morphologically indistinguishable from American chestnut.

This has been accomplished through four successive crosses of a Chinese chestnut and its descendants to American chestnut (i.e., the F1, B1, B2, and B3). Each generation of backcrossing halves the proportion of the genome inherited from C. mollissima and the resulting third backcross (B3) hybrids are 15/16ths American chestnut in their genomic composition. The hybrid progeny of backcrosses to C. dentata inherit varying numbers of resistance genes from C. mollissima. In each generation the most blight-resistant individuals have been selected by artificially inoculating the hybrids with Cryphonectria parasitica (the fungus that causes chestnut blight) and culling out all trees except for those with with the smallest cankers. At the third backcross stage the small portion of the genome inherited from Chinese chestnut is expected to harbor genes or alleles that confer resistance to chestnut blight.

At locations in the chestnut genome where resistance genes reside, backcross hybrids selected for resistance are expected to be heterozygous, meaning they inherited one copy of a resistance allele (i.e., a gene variant) from Chinese chestnut and corresponding allele for susceptibility from American chestnut. Blight resistance is an incompletely dominant trait, meaning that individuals heterozygous for resistance alleles from Chinese chestnut will have blight resistance that is intermediate between Chinese chestnut and American chestnut.

To enhance blight-resistance to levels closer to Chinese chestnut, the third backcross trees selected for resistance have been intercrossed to produce the B3F2 generation. Presumably, a subset of the progeny from these intercrosses will be homozygous for resistance from Chinese chestnut, meaning they inherited two alleles for resistance from Chinese chestnut from each B3 parent at some or all of the locations of the genome where resistance alleles reside. To avoid the deleterious effects of inbreeding upon intercrossing the B3 trees, 20 + American chestnut backcross lines have been created for each source of resistance at each location where chestnuts are being bred. American chestnut backcross lines refer to non-overlapping sets of American chestnut individuals that are bred with backcross selections in each generation. Because blight-resistance is controlled by multiple genes and the inheritance of these genes is a random process, it is necessary to generate large populations of B3F2 trees to produce individuals that are homozygous for resistance at multiple loci. For single source of resistance with 20 American chestnut backcross lines, over 30,000 B3F2 trees have been planted in seed orchards at TACF’s Research Farms at Meadowview to obtain ~300 selections. Approximately 80% of the B3F2 trees are culled after artificial inoculation with a weak strain of the blight fungus. We are currently testing the resistance of the progeny of trees that remain to determine the genetic resistance of their parents. We are also developing genomic selection to predict resistance based on DNA sequence (see Genomic selection). Genomic selection holds promise to accelerate and increase the accuracy of making the final selections.