Prevalent role of homologous recombination in the repair of specific double-strand breaks in Rhizobium etli .

Double-strand breaks (DSBs) are the most dangerous injuries for a genome. When unrepaired, death quickly ensues. In most bacterial systems, DSBs are repaired through homologous recombination. Nearly one-quarter of bacterial species harbor a second system, allowing direct ligation of broken ends, known as Non-Homologous End Joining (NHEJ). The relative role of both systems in DSBs repair in bacteria has been explored only in a few cases. To evaluate this in the bacterium Rhizobium etli , we used a modified version of the symbiotic plasmid (264 kb), containing a single copy of the nifH gene. In this plasmid, we inserted an integrative plasmid harboring a modified nifH gene fragment containing an I-SceI site. DSBs were easily inflicted in vivo by conjugating a small, replicative plasmid that expresses the I-SceI nuclease into the appropriate strains. Repair of a DSB may be achieved through homologous recombination (either between adjacent or distant repeats) or NHEJ. Characterization of the derivatives that repaired DSB in different configurations, revealed that in most cases (74%), homologous recombination was the prevalent mechanism responsible for repair, with a relatively minor contribution of NHEJ (23%). Inactivation of the I-SceI gene was detected in 3% of the cases. Sequence analysis of repaired derivatives showed the operation of NHEJ. To enhance the number of derivatives repaired through NHEJ, we repeated these experiments in a recA mutant background. Derivatives showing NHEJ were readily obtained when the DSB occurred on a small, artificial plasmid in a recA mutant. However, attempts to deliver a DSB on the symbiotic plasmid in a recA background failed, due to the accumulation of mutations that inactivated the I-SceI gene. This result, coupled with the absence of derivatives that lost the nonessential symbiotic plasmid, may be due to an unusual stability of the symbiotic plasmid, possibly caused by the presence of multiple toxin-antitoxin modules.