Genome stability in radioresistant bacterium Deinococcus radiodurans depends on RecA, the main bacterial recombinase. Without RecA, gross genome rearrangements occur during repair of DNA double-strand breaks. Long repeated (insertion) sequences have been identified as hot spots for ectopic recombination leading to genome rearrangements, and single-strand annealing (SSA) postulated to be the most likely mechanism involved in this process. Here, we have sequenced five isolates of D. radiodurans recA mutant carrying gross genome rearrangements to precisely characterize the rearrangements and to elucidate the underlying repair mechanism. The detected rearrangements consisted of large deletions in chromosome II in all the sequenced recA isolates. The mechanism behind these deletions clearly differs from the classical SSA; it utilized short (4–11 bp) repeats as opposed to insertion sequences or other long repeats. Moreover, it worked over larger linear DNA distances from those previously tested. Our data are most compatible with alternative end-joining, a recombination mechanism that operates in eukaryotes, but is also found in Escherichia coli. Additionally, despite the recA isolates being preselected for different rearrangement patterns, all identified deletions were found to overlap in a 35 kb genomic region. We weigh the evidence for mechanistic vs. adaptive reasons for this phenomenon.

Homologous recombination repairs potentially lethal DNA lesions such as double-strand DNA breaks (DSBs) and single-strand DNA gaps (SSGs). In Escherichia coli, DSB repair is initiated by the RecBCD enzyme that resects double-strand DNA ends and loads RecA recombinase to the emerging single-strand (ss) DNA tails. SSG repair is mediated by the RecFOR protein complex that loads RecA onto the ssDNA segment of gaped duplex. In both repair pathways, RecA catalyses reactions of homologous DNA pairing and strand exchange, while RuvABC complex and RecG helicase process recombination intermediates. In this work, we have characterised cytological changes in various recombination mutants of E. coli after three different DNA-damaging treatments: (i) expression of I-SceI endonuclease, (ii) γ-irradiation, and (iii) UV-irradiation. All three treatments caused severe chromosome segregation defects and DNA-less cell formation in the ruvABC, recG, and ruvABC recG mutants. After I-SceI expression and γ-irradiation, this phenotype was efficiently suppressed by the recB mutation, indicating that cytological defects result mostly from incomplete DSB repair. In UV-irradiated cells, the recB mutation abolished cytological defects of recG mutants and also partially suppressed the cytological defects of ruvABC recG mutants. However, neither recB nor recO mutation alone could suppress the cytological defects of UV-irradiated ruvABC mutants. The suppression was achieved only by simultaneous inactivation of the recB and recO genes. Cell survival and microscopic analysis suggest that chromosome segregation defects in UV-irradiated ruvABC mutants largely result from defective processing of stalled replication forks. The results of this study show that chromosome morphology is a valuable marker in genetic analyses of recombinational repair in E. coli.

The bacterium Deinococcus radiodurans is known to efficiently and accurately reassemble its genome after hundreds of DNA double-strand breaks (DSBs). Only at very large amounts of radiation-induced DSBs is this accuracy affected in the wild-type D. radiodurans, causing rearrangements in its genome structure. However, changes in its genome structure may also be possible during the propagation and storage of cell cultures. We investigate this possibility by listing structural differences between three completely sequenced genomes of D. radiodurans strains with a recent common ancestor—the type strain stored and sequenced in two different laboratories (of the ATCC 13939 lineage) and the first sequenced strain historically used as the reference (ATCC BAA-816). We detected a number of structural differences and found the most likely mechanisms behind them: (i) transposition/copy number change in mobile interspersed repeats—insertion sequences and small non-coding repeats, (ii) variable number of monomers within tandem repeats, (iii) deletions between long direct DNA repeats, and (iv) deletions between short (4–10 bp) direct DNA repeats. The most surprising finding was the deletions between short repeats because it indicates the utilization of a less accurate DSB repair mechanism in conditions in which a more accurate one should be both available and preferred. The detected structural differences, as well as SNPs and short indels, while being important footprints of deinococcal DNA metabolism and repair, are also a valuable resource for researchers using these D. radiodurans strains.