RecA balances genomic stability and evolution using many successive mismatch tolerant homology tests

A double-strand break (DSB) must usually be repaired with as little alteration to the genome as possible, though some rare alterations provide valuable genomic evolution. In E.coli, a DSB undergoes resection to give 3 ssDNA tails. These invading strand tails are loaded with RecA protein and then rapidly search the genome for the corresponding (allelic) partner. Thus, a searching ssDNA/RecA filament must almost never make stable non-allelic contact; therefore, it has been puzzling that RecA forms stable products that join partially homologous sequences. Homology testing by RecA family proteins begins with an 8-bp test, followed by successive homology tests of base pair triplets. Here we introduce a highly simplified homology recognition model to highlight how mismatch sensitivity could affect non-allelic pairing in bacterial genomes. The model predicts that even if each triplet test accepts 2 mismatches, RecA can have [~] 95% probability of establishing allelic pairing after a DSB in E. coli; however, that accuracy requires homology testing {gap}50 contiguous base pairs, consistent with the homology lengths probed in vivo. In contrast, if no mismatches are accepted testing 14 bp is sufficient, and testing more base pairs does not reduce non-allelic pairing because bacterial genomes contain long repeats.