Management of E. coli sister chromatid cohesion in response to genotoxic stress

Abstract

Aberrant DNA replication is a major source of the mutations and chromosomal rearrangements associated with pathological disorders. In bacteria, several different DNA lesions are repaired by homologous recombination, a process that involves sister chromatid pairing. Previous work in Escherichia coli has demonstrated that sister chromatid interactions (SCIs) mediated by topological links termed precatenanes, are controlled by topoisomerase IV. In the present work, we demonstrate that during the repair of mitomycin C-induced lesions, topological links are rapidly substituted by an SOS-induced sister chromatid cohesion process involving the RecN protein. The loss of SCIs and viability defects observed in the absence of RecN were compensated by alterations in topoisomerase IV, suggesting that the main role of RecN during DNA repair is to promote contacts between sister chromatids. RecN also modulates whole chromosome organization and RecA dynamics suggesting that SCIs significantly contribute to the repair of DNA double-strand breaks (DSBs).

Figure 1. SCIs are preserved in the presence of MMC via a RecN-dependent pathway.

(a) The number of SCIs is under the control of replication that produces cohesive sister chromatids and segregation that separates them. (b) MMC causes rapid replication arrest. An MG1655 wild-type strain was used to monitor EdU incorporation in the presence and absence of MMC (10 μg ml⁻¹) for 10 min. (c) Quantification of a 10 min EdU incorporation pulse in WT cells treated or not with MMC. (d) SCIs were estimated at different time points after replication arrest in WT and dnaC^ts cells treated or not with MMC. SCIs were measured in WT and dnaC^ts strains by adding arabinose during 20 min pulses after MMC addition. In the dnaC^ts strain initiation of replication was arrested at 40 °C. (e) Upon MMC treatment, SCIs are dependent on RecA, LexA and RecN. SCIs were estimated using the frequency of loxP/Cre recombination at the ori-3 locus. The results at 40 min are presented; the results at 10, 20 and 30 min are presented in Supplementary Fig. 1. The results are expressed as the relative loxP recombination normalized to the untreated WT strain. Statistical comparisons of histological data were performed using Student's t-test. P values are considered to be significant for α=0.05. *P<0.05, **P<0.01, ***P<0.001. (f) SCIs are not maintained in a recN mutant. MMC was added and SCIs were measured as described in d. (g) MMC prevents sister chromatid segregation in synchronized cells. dnaC^ts and dnaC^ts recN cells tagged with a parS^pMT1/ParB-GFP system at ori-3 were synchronized for 90 min at 40 °C, and replication was re-initiated at 30 °C for 10 min. Cells were then placed on an agar pad, with or without MMC, and a time course analysis was performed. The images represent kymographs of a single-cell (fire colour map) and an overlay of about 20 cells (ice colour map). (h) Quantification of segregation events in dnaC^ts WT and recN strains (100 cells were analysed for each strain). Cells were treated as described in g. Error bars are s.d. of four experiments.

Figure 2. RecN-mediated SCIs are specifically established in response to DSBs.

(a) Measurements of SCIs following Topo IV alteration in the presence of MMC. LoxP assays were performed at 10, 20, 30 and 40 min after the addition of MMC. The results at 30 min are presented; the results at 10, 20 and 40 min are presented in Supplementary Fig. 3. The results are expressed as the relative loxP recombination, with MMC normalized to untreated WT. Cells were incubated for 25 min at 42 °C before the addition of 0.1% arabinose and MMC. (b) Influence of Topo IV alteration on WT, recN and recA mutant viability in the presence of MMC. The cell viability assay was performed at non-permissive temperature of 42 °C. (c) Influence of RecN overexpression on SCIs. The plasmid pZA31 carries recN under the control of a leaky promoter. The results are expressed as the relative loxP recombination of the MMC-treated sample normalized to the untreated WT. (d) Viability of WT, recN and recA mutants in the presence of SS gaps formed by AZT. (e) Measurement of SCIs in the presence of SS gaps. LoxP assays were performed as described in A. Error bars are s.d. of four experiments.

Figure 3. RecN participates in sister locus re-pairing and nucleoid rearrangement in response to DSBs.

(a) Representative kymographs of sister focus dynamics in the absence or presence of MMC. Kymographs were constructed along the long axis of the cell. Time-lapse imaging starts 5 min after the initial contact with MMC. Images were acquired every 3 min for 2 h. The fire lookup kymographs represent a single cell. The Ice lookup kymographs are an overlay of kymographs. (b) Frequency of the different types of sister focus dynamics was measured in the presence of MMC. Results are expressed as a percentage. About 100 cells were observed. (c) Representative kymographs of nucleoid (HU-mCherry) dynamics in the absence or presence of MMC. (d) The frequency of the different types of nucleoid dynamics was measured in the presence of MMC for the WT strain and the recN mutant. Results are expressed as a percentage. About 100 cells were observed. (e) Dynamics of sister foci and nucleoids in the presence of MMC. A time course was performed in a strain tagged with a parS^pMT1/ParB tag at ori-3 and labelled with HU-mCherry. (f) The distance between nucleoid edges and the distance between sister foci were recorded at each time point of the experiment presented on e. Distances were normalized to 1 for the time +5 min after MMC addition and are an average of 50 cells. Error bars are standard deviations of 50 cells. Scale bars are 1 μm.

Figure 4. RecN influences SCC but not DNA condensation.

(a) The number of foci per cell at the yajR-yajQ::parS^pMT1 site was counted in cells treated with MMC for 0, 15 or 45 min. (b) The number of foci per cell at the crl::parS^P1 site was counted in cells treated with MMC for 0, 15 or 45 min. (c) Same as a, but performed in the recN mutant. (d) Same as b, but performed in the recN mutant. In a–d, error bars represent standard deviations of 300 cells. (e) The distance between two loci on the same replichore of the chromosome, spaced by 188 kb and tagged with a parS^P1 or a parS^pMT1 site (crl::parS^P1 and yajQ-yajR::parS^pMT1, respectively), was measured after treatment with 10 μg ml⁻¹ MMC for 0, 15 or 45 min. The results are shown as a box plot representing the median, first and forth quartiles (N=300). (f) Same as e, but performed in the recN mutant. (g) The distance between two loci on the same replichore of the chromosome, spaced by 188 kb and tagged with a parS^P1 or a parS^pMT1 site (crl::parS^P1 and yajQ-yajR::parS^pMT1, respectively), was measured after treatment with 30 μg ml⁻¹ of chloramphénicol for 0, 15 or 45 min. (t-test ***P<10⁻³⁰, NS (not significant) P>10⁻⁵).

Figure 5. recN mutant has altered homology search and delayed cell cycle restart.

(a) Cell cycle restart pattern after a brief MMC treatment using a microfluidic platform. WT cells were introduced into the chambers, and fresh minimal medium A was perfused for 20 min; 10 μg ml⁻¹ MMC was then perfused for 10 min and immediately washed with clean medium, and incubation and imaging were continued for over 3.5 h. Cell lineage was measured for 100 cells; each colour corresponds to a given state of the cell. (b) The same experiment as described in a was performed in the recN mutant, which exhibits delayed cell cycle restart. (c) Representative time-lapse microscopy of RecA-mCherry focus dynamics in the presence of MMC in the WT strain. Time-lapse imaging starts at 5 min after initial contact with MMC. Pictures were acquired every 3 min for 2 h on an agarose pad with MMC. (d) RecA-mCherry focus dynamics in the presence of MMC in the recN mutant. Experiments were performed as described for c. (e) Analysis of RecA focus dynamics. Kymograph and time series of cell slices for RecA-mCherry WT. The experiment was performed as described in c. (f) Analysis of RecA foci dynamics in the recN mutant. The experiment was performed as described in d. (g) The frequency of bundles was estimated as a function of the shape of the RecA mCherry signal in WT and the recN mutant. The WT and recN distribution are significantly different (t-test P=10⁻³⁰). Scale bar is 1 μm.

Figure 6. Roles of RecN during repair of an induced DSB.

Our observations suggest that when a replicative DSB occurs, RecA (Green) is responsible for RecN (orange) expression (through the SOS response) and RecN loading onto the sister chromatids. RecN loading prevents the complete removal of SCIs by Topo IV (blue) and may participate in DNA end joining. In a second step, RecN may propagate on the newly replicated chromatids to mediate regression of the segregated sister chromatids and re-mixing of brother nucleoids.