Efficient elimination of selectable marker genes from the plastid genome by the CRE-lox site-specific recombination system

Incorporation of a selectable marker gene during transformation is essential to obtain transformed plastids. However, once transformation is accomplished, having the marker gene becomes undesirable. Here we report on adapting the P1 bacteriophage CRE-Iox site-specific recombination system for the elimination of marker genes from the plastid genome. The system was tested by the elimination of a negative selectable marker, codA, which is flanked by two directly oriented fox sites (>codA>). Highly efficient elimination of >codA> was triggered by introduction of a nuclear-encoded plastid-targeted CRE by Agrobacterium transformation or via pollen. Excision of >codA> in tissue culture cells was frequently accompanied by a large deletion of a plastid genome segment which includes the tRNA-Val(UAC) gene. However, the large deletions were absent when cre was introduced by pollination. Thus pollination is our preferred protocol for the introduction of cre. Removal of the >codA> coding region occurred at a dramatic speed, in striking contrast to the slow and gradual build-up of transgenic copies during plastid transformation. The nuclear cre gene could subsequently be removed by segregation in the seed progeny. The modified CRE-Iox system described here will be a highly efficient tool to obtain marker-free transplastomic plants.