A bacterial DNA repair pathway specific to a natural antibiotic

Abstract

All organisms possess DNA repair pathways that are used to maintain the integrity of their genetic material. Although many DNA repair pathways are well understood, new pathways continue to be discovered. Here, we report an antibiotic specific DNA repair pathway in Bacillus subtilis that is composed of a previously uncharacterized helicase (mrfA) and exonuclease (mrfB). Deletion of mrfA and mrfB results in sensitivity to the DNA damaging agent mitomycin C, but not to any other type of DNA damage tested. We show that MrfAB function independent of canonical nucleotide excision repair, forming a novel excision repair pathway. We demonstrate that MrfB is a metal-dependent exonuclease and that the N-terminus of MrfB is required for interaction with MrfA. We determined that MrfAB failed to unhook interstrand cross-links in vivo, suggesting that MrfAB are specific to the monoadduct or the intrastrand cross-link. A phylogenetic analysis uncovered MrfAB homologs in diverse bacterial phyla, and cross-complementation indicates that MrfAB function is conserved in closely related species. B. subtilis is a soil dwelling organism and mitomycin C is a natural antibiotic produced by the soil bacterium Streptomyces lavendulae. The specificity of MrfAB suggests that these proteins are an adaptation to environments with mitomycin producing bacteria.

Figure 1.. DNA damage sensitivity of ΔmrfAB is specific to mitomycin C.

(A) Relative fitness plots for the indicated gene disruptions from Tn-seq experiments previously reported (Burby et al., 2018). The mean fitness is plotted as a bar graph and the error bars represent the 95% confidence interval. (B) Spot titer assay using strains with the indicated genotypes grown on LB with the indicated supplements. (C) Spot titer assay using strains with the indicated genotypes grown on LB media with the indicated treatments. For UV irradiation, cells were exposed to the indicated dose after serial dilutions were spotted on plates. For trioxsalen plates, 1 μg/mL was used and the UV wavelength for irradiation was 365 nm.

Figure 2.. MrfA and MrfB function in the same pathway.

(A) Spot titer assay using strains with the indicated genotypes grown on the indicated media. (B) Bacterial two-hybrid assay using the indicated T18 and T25 fusions. (C) MrfA constructs used in deletion analysis of MrfA-MrfB interaction (upper) and a bacterial two-hybrid assay using T25-MrfB and the indicated MrfA-T18 fusions (lower). (D) MrfB constructs used in deletion analysis of MrfA-MrfB interaction (upper) and a bacterial two-hybrid assay using MrfA-T18 and the indicated T25-MrfB fusions (lower).

Figure 3.. MrfB is a metal-dependent exonuclease.

(A) A schematic of MrfB depicting putative catalytic residues and C-terminal tetratrichopeptide repeat (TPR) domain. (B) Spot titer assay using strains with the indicated genotypes spotted on the indicated media. (C) 1 μg of purified MrfB stained with Coomassie brilliant blue. (D) Exonuclease assay using pUC19 linearized with BamHI (lanes 3–7). Reactions were incubated at 37°C for 15 minutes with or without MrfB, MgCl₂, or EDTA as indicated, and separated on an agarose gel stained with ethidium bromide. Lane 1 is a 1 kb plus molecular weight marker (M) and lane 2 is undigested pUC19 plasmid. (E) Exonuclease assay testing substrate preference. The indicated exonucleases were incubated with a closed covalent circular plasmid (CCC), a nicked plasmid (Nicked) or a linear plasmid (Linear) in the presence of Mg²⁺ at 37°C for 10 minutes. Reaction products were separated on an agarose gel stained with ethidium bromide. Lane 1 is a 1 kb plus molecular weight marker (M).

Figure 4.. MrfAB function independent of UvrABC dependent nucleotide excision repair.

(A & B) Spot titer assays using strains with the indicated genotypes grown on the indicated media. (C) Survival assay using strains with the indicated genotypes. The y-axis is the percent survival relative to the untreated (0 ng/mL) condition. The x-axis indicates the concentration of MMC used for a 30 minute acute exposure. The data points represent the mean of three independent experiments performed in triplicate (n=9) ± SEM.

Figure 5.. MrfAB are not required for unhooking inter-strand DNA crosslinks.

(A) DNA crosslinking repair assay. Chromosomal DNA from untreated samples (U), 1 μg/mL MMC treated samples (T), and recovery samples (45’ and 90’) were heat denatured and snap cooled (upper) or native chromosomal DNA (lower) was separated on an agarose gel stained with ethidium bromide. A 1 kb plus molecular weight marker is shown in the first lane. (B) A bar graph showing the mean percent of crosslinked DNA (see methods) from two independent experiments, and error bars represent the range of the two measurements.

Figure 6.. MrfAB and UvrABC are required for efficient RecA-GFP focus formation.

(A) Representative micrographs of strains containing RecA-GFP expressed from the native locus in addition to the indicated genotypes. Images were captured at the indicated times following MMC addition (5 ng/mL). RecA-GFP is shown in green and the merged images show RecA-GFP (green) and membranes stained with FM4–64 (red). The white bar indicates 5 μm (B) Percentage of cells with a RecA-GFP focus or foci over the indicated time course of MMC treatment (5 ng/mL). The error bars represent the 95% confidence interval.

Figure 7.. MrfAB are conserved in diverse bacterial phyla.

(A) A rooted phylogenetic tree constructed using 16s rRNA sequences (18s rRNA for S. cerevisiae), aligned with muscle (Edgar, 2004), using the neighbor joining method (Saitou & Nei, 1987), and the evolutionary distances were calculated using the p-distance method (Nei & Kumar, 2000). The percentage of replicate trees that resulted in the associated species clustering together in a bootstrap test (500 replicates) is indicated next to the branches (Felsenstein, 1985). Evolutionary analysis was performed in MEGA (Kumar, Stecher, & Tamura, 2016). *In this organism MrfA and MrfB homologs are fused into a single protein. (B) Spot titer assay using codon optimized versions of MrfA and MrfB from the indicated species to complement ΔmrfA (upper) or ΔmrfB (lower).

Figure 8.. A model for MrfAB mediated nucleotide excision repair.

We propose that either an unknown factor or MrfA recognizes an MMC adduct. MrfB is then recruited, and MrfA uses its helicase activity to separate the strand containing the MMC adduct, facilitating MrfB-dependent degradation of the adduct containing DNA. The source of the nick used to direct excision is unknown.