Karl Haro von Mogel has a better way to revisit domestication:
I wanted to point out that for shattering, the mutant is likely to be recessive. If you stuck a mutant allele into a normal plant you would likely get a normal plant. So I think what would work better is to make an RNAi construct using the shattering gene as the template, that way it would silence the genes involved and be a dominant trait.