DNA end resection controls the balance between homologous and illegitimate recombination in Escherichia coli.

Even a partial loss of function of human RecQ helicase analogs causes adverse effects such as a cancer-prone Werner, Bloom or Rothmund-Thompson syndrome, whereas a complete RecQ deficiency in Escherichia coli is not deleterious for a cell. We show that this puzzling difference is due to different mechanisms of DNA double strand break (DSB) resection in E. coli and humans. Coupled helicase and RecA loading activities of RecBCD enzyme, which is found exclusively in bacteria, are shown to be responsible for channeling recombinogenic 3' ending tails toward productive, homologous and away from nonproductive, aberrant recombination events. On the other hand, in recB1080/recB1067 mutants, lacking RecBCD's RecA loading activity while preserving its helicase activity, DSB resection is mechanistically more alike that in eukaryotes (by its uncoupling from a recombinase polymerization step), and remarkably, the role of RecQ also becomes akin of its eukaryotic counterparts in a way of promoting homologous and suppressing illegitimate recombination. The sickly phenotype of recB1080 recQ mutant was further exacerbated by inactivation of an exonuclease I, which degrades the unwound 3' tail. The respective recB1080 recQ xonA mutant showed poor viability, DNA repair and homologous recombination deficiency, and very increased illegitimate recombination. These findings demonstrate that the metabolism of the 3' ending overhang is a decisive factor in tuning the balance of homologous and illegitimate recombination in E. coli, thus highlighting the importance of regulating DSB resection for preserving genome integrity. recB mutants used in this study, showing pronounced RecQ helicase and exonuclease I dependence, make up a suitable model system for studying mechanisms of DSB resection in bacteria. Also, these mutants might be useful for investigating functions of the conserved RecQ helicase family members, and congruently serve as a simpler, more defined model system for human oncogenesis.