Published Online: 22 May, 2018 | Supp Info: http://doi.org/10.1083/jcb.201803020 Downloaded from jcb.rupress.org on July 4, 2019 REPORT Dynamics of RecA-mediated repair of replication- dependent DNA breaks Vincent Amarh, Martin A. White, and David R.F. Leach Chromosomal replication is the major source of spontaneous DNA double-strand breaks (DSBs) in living cells. Repair of these DSBs is essential for cell viability, and accuracy of repair is critical to avoid chromosomal rearrangements. Repair of replication-dependent DSBs occurs primarily by homologous recombination with a sister chromosome. However, this reaction has never been visualized at a defined chromosomal locus, so little is known about its spatial or temporal dynamics. Repair of a replication-independent DSB generated in Escherichia coli by a rare-cutting endonuclease leads to the formation of a bundle of RecA filaments. In this study, we show that in contrast, repair of a replication-dependent DSB involves a transient RecA focus localized in the central region of the cell in which the DNA is replicated. The recombining loci remain centrally located with restricted movement before segregating with little extension to the period of postreplicative sister- chromosome cohesion. The spatial and temporal efficiency of this reaction is remarkable. Introduction Duplicating the genome is a fundamental requirement of life. nonhomologous end joining. However, such DSBs are likely to be Although the DNA replication machinery is capable of doing so particularly well suited for repair by homologous recombination successfully, it occasionally encounters obstacles that lead to rep- because there is no requirement for an extensive DNA homology lication fork stalling (Cox et al., 2000). Stalled replication forks search as the site of the DSB is in close proximity to the unbroken are potential sources of double-strand breaks (DSBs; Michel et sister chromosome. al., 2004). These spontaneous DSBs can be generated as a conse- The dynamics of DSBR in bacteria have been studied previ- quence of replication forks colliding with DNA-bound proteins ously using live-cell fluorescence imaging. In these studies, DSBs such as transcription complexes (Marnef et al., 2017). Sponta- have been generated by the rare-cutting I-SceI endonuclease neous DSBs are also formed at DNA nicks and gaps encountered (Lesterlin et al., 2014; Badrinarayanan et al., 2015), DNA damage– by progressing replication forks (Kuzminov, 2001). Because one inducing drugs (Kidane and Graumann, 2005), or UV irradiation unrepaired DSB can be a lethal event, DNA DSB repair (DSBR) (Renzette et al., 2005; Centore and Sandler, 2007). Notably, RecA plays a critical role in underpinning chromosomal replication. bundles or thread-like structures were detected after DSB induc- This importance of DSBR during chromosomal replication is tion. These extended RecA structures were proposed to mediate predicted to increase in organisms with larger genomes because the extensive DNA homology search that was required for repair the probability of DSB formation is expected to increase pro- of I-SceI–induced DSBs (Lesterlin et al., 2014; Badrinarayanan et portionally with the length of replicated DNA. The numbers of al., 2015). We reasoned that RecA bundles might not be required DSBs detected in Escherichia coli and human cells confirm this during repair of a replication-dependent DSB if the repair was prediction. It has recently been estimated that the replication initiated during the period of postreplicative cohesion of sis- forks that duplicate the 4.6-Mbp genome of E. coli have an 18% ter chromosomes in which an extensive DNA homology search probability of breakage as estimated by the percentage of cells is not required. with broken forks detected in the absence of repair by homol- In this study, we investigated the spatial and temporal dynam- ogous recombination (Sinha et al., 2018). In human cells with ics of RecA during the repair of a replication-dependent DSB in a genome size of 3.2 Gbp, it is estimated that ∼50 spontaneous E. coli. We addressed this question by inducing a replication- DSBs are repaired per cell cycle during chromosomal replica- dependent break in the lacZ gene, which is located on the right tion (Vilenchik and Knudson, 2003). DSBs emanating from bro- arm of the chromosome, approximately halfway between the ken replication forks are one-ended and cannot be repaired by origin and the terminus. Our study addressed the following Institute of Cell Biology, School of Biological Sciences, University of Edinburgh, Edinburgh, Scotland, UK. Correspondence to David R.F. Leach: D.Leach@e d .ac. uk; V. Amarh’s present address is Dept. of Biochemistry, Cell and Molecular Biology, West African Centre for Cell Biology of Infectious Pathogens, University of Ghana, Legon-Accra, Ghana; M.A. White’s present address is Dept. of Molecular and Cellular Biology, Harvard University, Cambridge, MA.  © 2018 Amarh et al. This article is available under a Creative Commons License (Attribution 4.0 International, as described at https://c reativecommons .org/ licenses/ by/4 . 0/ ). Rockefeller University Press https://doi.org/10.1083/jcb.201803020 2299 J. Cell Biol. 2018 Vol. 217 No. 7 2299–2307 questions: How long is a fluorescent derivative of RecA visible as formation using our inducible system, confirming that the com- a focus at the site of the DSB? Does the RecA focus at the site of bination of RecA-mCherry and RecA conferred recombination the DSB mature to form a RecA bundle? Where is the RecA focus proficiency. For the slow growth condition (M9 salt medium located in the cell, and how does its position relate to the local- supplemented with glycerol) that was used in this study, cells ization of the DNA replication machinery? Does DSBR affect the had a single replicating chromosome in the absence of SbcCD duration of postreplicative cohesion at the lacZ locus? expression (Fig. 1 D). It was therefore expected that only one We show that RecA works fast and does so in the center of DSBR reaction would occur per cell cycle when expression of the cell, close to where the break was formed during replica- SbcCD was induced in the palindrome-containing strain. tion. No RecA bundle is observed, indicating that formation of this extended structure is not a necessary consequence of DNA RecA forms transient foci at the DNA DSB site breakage. The whole reaction (from breakage to separation of After induction of SbcCD expression in the palindrome-con- the recombining loci) is remarkably efficient as expected if the taining strain, RecA-mCherry proteins were assembled to form cell has evolved a mechanism to permit the repair of replication- a distinct focus, which colocalized with both the TetR-YPet and dependent breaks with minimal perturbation to DNA replication. LacI-Cerulean foci, the fluorescent markers for the DSB site (Fig. 2 A; upper limit for colocalization is 0.4 µm). Subsequently, the RecA-mCherry focus disassembled to background fluores- Results and discussion cence within the cell. Disassembly of the RecA-mCherry focus System for visualizing replication-dependent DSBR at the was followed by segregation of the sister lacZ loci into the oppo- E. coli lacZ locus site cell halves and, eventually, cell division (Fig. 2 B). The median Live-cell fluorescence imaging was used to investigate the duration between disassembly of the RecA-mCherry focus and dynamics of repair of a site-specific DSB whose formation was segregation of sister lacZ loci was 18 min (Fig. 2 C). dependent on chromosomal replication. The DSB was generated Previous studies have suggested that the RecA nucleoprotein at the lacZ locus of the E. coli chromosome by SbcCD-mediated filament is stabilized by the DinI protein and is destabilized by cleavage of a DNA hairpin structure formed by a 246-bp, inter- the RecX and UvrD proteins in vitro (Stohl et al., 2003; Drees et rupted palindrome on the lagging-strand DNA template during al., 2004; Lusetti et al., 2004b; Veaute et al., 2005; Cox, 2007). In replication (Fig. 1 A; Eykelenboom et al., 2008). Arrays of lac this study, we investigated the effect of ΔuvrD, ΔrecX, or ΔdinI and tet operator sequences were inserted on either side of the mutation on the duration of RecA-mCherry focus at the DSB site. DSB site (White et al., 2008) and were preceded by arrays of Our data revealed that the median duration of the RecA-mCherry three Chi sites (Fig. 1 B) to stimulate RecA loading by the Rec- focus was 1.5 min in WT cells, 1.3 min in the ΔdinI mutant, and BCD enzyme (Cockram et al., 2015). The tet and lac repressor 2.1 min in the ΔrecX mutant (Fig. 2 D). The effects of the ΔdinI genes were coupled to YPet and Cerulean genes, respectively, and ΔrecX mutations were small and not statistically significant. and inserted in tandem at the ykgC chromosomal locus, with However, they were as expected for antagonistic roles of the RecX these genes under the control of a strong constitutive synthetic and DinI proteins on the stability of the RecA nucleoprotein fil- promoter, Pmw1. The Pmw1 promoter was derived from the ftsk ament (Lusetti et al., 2004a). Deletion of the uvrD gene caused gene promoter (Wang et al., 2005) by mutating the −10 and a very minimal, statistically insignificant change to the median −35 elements to their respective consensus sequences using duration of the RecA-mCherry focus (median duration was 1.6 site-directed mutagenesis. Binding of TetR-YPet and LacI-Ce- min), although some longer-lasting (4–6 min) foci were observed rulean proteins to the operator arrays generated coincident (Fig. 2 D). The RecA-mCherry foci colocalizing with the TetR-YPet fluorescent foci, which marked the cellular location of the lacZ and LacI-Cerulean foci were inferred to define RecA participating locus of the chromosome. Visualization of RecA was achieved in DSBR at the site of the SbcCD-mediated cleavage of the DNA by inserting a codon-diversified recA-mCherry gene in tandem palindrome inserted in the lacZ gene. The transient duration with the endogenous recA gene, with both recA alleles under of these RecA-mCherry foci suggest that the homology search the control of the native promoter (Fig. 1 B). Codon diversifica- and strand exchange are rapid events during repair of a replica- tion was performed to limit runs of homology between the recA tion-dependent DSB. This compares with periods on the order of alleles to <14 bp to minimize the probability of recombination 1 h after I-SceI cleavage, where RecA bundle formation to focus between the two genes. Finally, the promoter of the sbcD and pairing takes a mean of 47 min, and bundles disassemble after sbcC genes was replaced by the arabinose-inducible promoter a further 17 min (Lesterlin et al., 2014). Bundles or thread-like (ParaBAD) to enable induction of expression of the SbcCD protein, structures of RecA-mCherry were not detected in cells undergo- which is responsible for DNA hairpin cleavage, to generate the ing DSBR at the lacZ locus. DSB at the lacZ locus (Eykelenboom et al., 2008). The tandem insertion of recA-mCherry and recA genes was constructed DSBR localizes the lacZ locus to the midcell region because the RecA-mCherry protein is only partially active, but The effect of DSBR on the spatial localization of the lacZ locus recA-mCherry is recessive, and the partial diploid strain con- was also investigated in strains containing, or not containing, the taining recA-mCherry and recA is fully recombination profi- interrupted palindrome at the lacZ locus. After SbcCD expression cient, despite the incorporation of RecA-mCherry proteins into in the strain that did not contain the interrupted palindrome, the mixed filaments. Fig. 1 C shows that the partial diploid strain lacZ locus exhibited dynamic movement, primarily on one side did not suffer any detectable loss of viability in response to DSB of the midcell, until segregation of the sister loci (Fig. 3 A). After Amarh et al. Journal of Cell Biology 2300 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020 Figure 1. System for visualizing DSBR at the lacZ locus. (A) Schematic representation of replication-dependent DSB formation at the E. coli lacZ locus. (B) Schematic representation of the construct used for visualizing the site of the DSB at the lacZ locus. The location of the palindrome at the lacZ locus is indicated as 0 kb. The arrays of three Chi sites are represented by three arrows. The arrays of 171 tetO and 143 lacO sites are shown in yellow and blue, respec- tively. The precise number of operator sites in these constructs was determined by DNA sequencing with the primers listed in Table S3. A codon-diversified recA-mCherry gene was inserted in tandem with the endogenous recA gene at the native chromosomal locus. The promoter for sbcC and sbcD genes was replaced by the arabinose-inducible promoter, ParaBAD. The tetR-YPet and lacI-Cerulean genes were expressed from a strong, constitutive, synthetic promoter at the chromosomal ykgC locus. (C) Spot-test assay showing that addition of the recA-mCherry gene does not affect cell viability after DSB induction at the lacZ locus. (D) DNA content of cells grown in M9–glycerol medium. The experiment was performed in triplicate. Representative data from one experiment are shown. segregation, one locus stayed close to the midcell, whereas the segregation of the sister lacZ loci, is distinct from the locations other migrated further, consistent with the left–right–left–right of lacZ before DSBR and during replicative cohesion in cells not symmetry of the chromosome arms (White et al., 2008). In the undergoing DSBR. palindrome-containing strain, the dynamic movement of the lacZ locus, which was also primarily on one side of the midcell DSBR occurs close to the site of lacZ replication became constrained in the vicinity of the midcell during the To investigate the relationship between DSBR and DNA rep- formation and after disassembly of the RecA-mCherry focus at lication, the dnaN gene encoding the β-sliding clamp of the the site of the DSB (Fig. 3 B). Distances of centroids of lacZ foci replisome was tagged with the fluorescent protein YPet. Pre- (before segregation) to an arbitrary cell pole are shown in Fig. 3 viously, fluorescent foci formed by YPet-DnaN proteins have (C and D) and Fig. S1, whereas distances to midcell are shown been shown to be indicative of the cellular location of the chro- in Fig. S2. Collectively, these data demonstrate that the central mosomal replisome (Wang et al., 2011). Time-lapse imaging location of lacZ, from RecA-mCherry focus formation until confirmed that molecules of YPet-DnaN were assembled from Amarh et al. Journal of Cell Biology 2301 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020 Figure 2. Formation of a reversible and transient RecA focus at the lacZ locus during DSBR. (A) Time-lapse fluorescence imaging of RecA-mCherry and the site of the replication-dependent DSB at the lacZ locus. The lacZ locus was visualized by TetR-YPet and LacI-Cerulean proteins bound to the tetO and lacO arrays, respectively. (i) Red, yellow, and blue spots represent RecA-mCherry, TetR-YPet, and LacI-Cerulean foci, respectively. (ii) Colocalization of RecA-mCherry focus with the LacI-Cerulean and TetR-YPet foci. Asterisks indicate the images used for the colocalization in ii. (B) Time-lapse fluorescence imaging of RecA- mCherry and the site of the DSB during repair, segregation, and cell division. Bars, 2 µm. (C) Duration of cohesion of sister lacZ loci after disassembly of the RecA-mCherry focus at the site of the DSB. n is the number of cells analyzed. (D) Duration of the RecA-mCherry focus at the site of the DSB WT cells and in recX, dinI, and uvrD mutants. Time-lapse imaging was performed at 1-min intervals for each strain after induction of SbcCD expression in the palindrome-containing strains. Dash lines represent the extrapolation of the median duration of the RecA-mCherry focus at the site of the DSB. Median durations for each of the mutants were not significantly different from the WT (P > 0.05) using the Mood’s median test. At least 80 cells were analyzed for the WT and each mutant strain. background fluorescence to form a distinct focus in newborn duplicate each arm of the circular E. coli chromosome usually cells, which was later disassembled before cell division (Fig. S3 coexist in the vicinity of the midcell (Mangiameli et al., 2017), A). Time-lapse imaging of the YPet-DnaN protein revealed that there is no requirement for them to do so to complete the cell cells had either one or two YPet-DnaN foci during chromosomal cycle (Reyes-Lamothe et al., 2008). replication. Interestingly, 28 out of 41 cells exhibited a single, The distribution of durations of the YPet-DnaN foci per rep- predominant YPet-DnaN focus during the replication cycle, and lication cycle was not greatly affected by repair of the DSB at the this focus localized at or near the midcell (Fig. S3, A and B). In lacZ locus (Fig. 4 A), indicating that repair of the palindrome-in- these cells, the single YPet-DnaN focus occasionally underwent duced DSB had minimal effect, if any, on the time required to rep- transient separation to generate two foci, which were very close licate the entire chromosome (mean of 69 ± 8 min in the absence to each other at the midcell (Fig. S3, A and B). In the remaining of DSBR and 68 ± 9 min in the presence of DSBR; Fig. 4 B). These 13 cells, the two YPet-DnaN foci that were generated by spatial observations are in accordance with our expectation that the separation of the sister replisomes were longer lived (Fig. S3 C). DSB is generated behind the progressing replication fork at the These observations suggest that, although the replisomes that lacZ locus (Eykelenboom et al., 2008). On the assumption that Amarh et al. Journal of Cell Biology 2302 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020 Figure 3. Effect of DSBR on localization of the lacZ locus. (A) Spatial dynamics of the lacZ locus in the absence of DSB induction at the lacZ locus. Dashed lines represent the midcell. Each kymograph shows a representative cell and was compiled from the phase-contrast and LacI-Cerulean fluorescence images of a cell acquired at 45-s intervals. (B) Spatial dynamics of the lacZ locus in the presence of DSB induction at the lacZ locus. Dash lines represent the midcell. Black arrows indicate the formation of the RecA-mCherry focus at the site of the DSB. Each kymograph shows a representative cell and was compiled from the phase-contrast and LacI-Cerulean fluorescence images of a cell acquired at 45-s intervals. (C) Localization of LacI-Cerulean foci, before foci splitting, in the absence of DSB induction at the lacZ locus. The data were separated into foci visible in the period before the last 18 min before foci splitting and foci visible during the last 18 min before foci splitting. The first period is expected to comprise mostly foci before DNA replication and replicative cohesion. The second period is expected to comprise mainly foci after DNA replication, during the period of postreplicative cohesion. n = 82 for the before-18-min mean cohesion period; n = 163 for the during-18-min mean cohesion period. (D) Localization of LacI-Cerulean foci, before foci splitting, in the presence of DSB induction at the lacZ locus. The formation of a RecA-mCherry focus at the DSB site was used as an indicator for repair (n = 144 before repair and n = 244 during repair). Heat maps are shown as colored bars in A and B, where 0 represents the background fluorescence within the cell, and 1 represents the maximum fluorescence of the LacI-Cerulean foci. replication occurs at an approximately equal rate around the Analysis of the data obtained by time-lapse microscopy also chromosome as has been demonstrated (Skovgaard et al., 2011), revealed the relationship between DNA replication and DSBR at the time taken to replicate the chromosome predicts that the the lacZ locus (Fig. 4 D). We determined that the formation of a lacZ locus is replicated at 32 min after initiation of replication. RecA-mCherry focus occurred at a median duration of 2.5 min The median time required for segregation of the sister lacZ loci after the colocalization of YPet-DnaN focus with the lacZ locus, after initiation of replication (formation of a YPet-DnaN focus which indicates the likely time of lacZ replication (Fig. 4, D and in newborn cells) was 50 min in the absence of DSB induction E). The YPet-DnaN focus did not always remain localized at the (Fig. 4 C), arguing that cohesion of sister chromosomes at the midcell after the formation of the RecA-mCherry focus at the lacZ locus lasts a median period of 18 min. The median period site of the DSB (second YPet-DnaN kymograph in Fig. 4 D). This between initiation of replication and the separation of lacZ loci observation demonstrates that the constraint on the dynamic was 56 min in cells undergoing DSBR (Fig. 4 C), suggesting that movement of the lacZ locus after the formation and disassembly DSBR extends sister chromosome cohesion by 6 min, to give a of the RecA-mCherry focus, before lacZ focus segregation, is inde- duration of 24 min. pendent of the spatial localization of the replisome and is likely Amarh et al. Journal of Cell Biology 2303 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020 Figure 4. Localization of the replisome and the lacZ locus during DSBR. (A) Distribution of the duration of YPet-DnaN foci per replication cycle under slow-growth conditions. (B) Mean duration of chromosomal replication under slow-growth conditions. Error bars represent SD. (C) Replication and segregation of the lacZ locus in the absence and presence of DSB induction at the lacZ locus. The dashed lines represent extrapolation of the median values for the “NO DSB” and “DSB” data. (D) Localization of the lacZ locus (LacI-Cerulean) and the replication site (YPet-DnaN) during formation and repair of the replication- dependent DSB. For all the kymographs, the dash lines represent the midcell. The black arrows represent the formation of the transient RecA-mCherry focus at the site of the DSB. The black circles represent colocalization of the LacI-Cerulean focus with the YPet-DnaN focus. Each black circle containing a black arrow represents the colocalization of the LacI-Cerulean focus with the YPet-DnaN focus until disassembly of the transient RecA-mCherry focus at the DSB site. Each kymograph shows a representative cell and was compiled from the phase-contrast and fluorescence (LacI-Cerulean or YPet-DnaN) images of a cell acquired at 1-min intervals. Heat maps are shown as colored bars, where 0 represents background fluorescence within the cell and 1 represents the maximum fluorescence of fluorescent foci. (E) Phase-contrast and fluorescence (LacI-Cerulean and YPet-DnaN) images acquired at 1-min intervals were used to estimate the duration between colocalization of a YPet-DnaN focus with a LacI-Cerulean focus and formation of the RecA-mCherry focus at the site of the DSB. For cells with multiple colocalizations (between LacI-Cerulean and YPet-DnaN foci) preceding the formation of the RecA-mCherry focus, the last colocalization event was chosen. to be local to the site of DSBR. The period of 2.5 min between DNA in the cells undergoing DSBR, consistent with the independent replication and RecA-mCherry focus formation, added to the 1.5 measure of 24 min of total cohesion (Fig. 4 C). min median duration of the RecA-mCherry focus (Fig. 2 D) and We have shown that RecA forms a distinct and transient the 18 min between disassembly of the RecA-mCherry focus and focus at the site of a replication-dependent DSB induced at the lacZ focus splitting (Fig. 2 C), sums to a total of 22 min of cohesion lacZ locus of the E. coli chromosome. The transient focus was Amarh et al. Journal of Cell Biology 2304 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020 disassembled before segregation of the sister loci. After disas- adjusted to 1.0 and serially diluted 10-fold. Aliquots of 4 µl of the sembly of RecA, the recombining loci remained centrally located serially diluted cultures were spotted onto LB agar plates con- in the cell and showed reduced mobility, consistent with local taining either 0.2% arabinose or 0.5% glucose. The LB agar plates constraints that might include the resolution of the DNA struc- were incubated overnight at 37°C. tures such as the Holliday junctions generated by the action of RecA. The RecA focus did not mature into an elongated or bundle Microscopy and image analysis structure as reported previously, when rare-cutting endonucle- Conventional widefield fluorescence microscopy was performed ase systems were used to generate a site-specific DSB (Lesterlin with a Zeiss Axiovert 200 fluorescence microscope equipped with et al., 2014; Badrinarayanan et al., 2015). RecA-GFP also formed a 100× 1.4 NA oil Plan Apochromat objective (phase or differential elongated structures during repair of DSBs generated using interference contrast [DIC]), dual OptoLED light source (Cairn mitomycin C (Kidane and Graumann, 2005) and UV irradiation Research), an MS-2000 Piezo Z-Stage (Applied Scientific Instru- (Renzette et al., 2005; Centore and Sandler, 2007), despite these mentation), and an Evolve 512 electron-multiplying charge-cou- treatments being predicted to produce replication-dependent pled device camera (Photometrics). With the exception of time- DSBs. These studies reported that the formation, maturation, and lapse imaging of the WT and ΔrecX, ΔuvrD, and ΔdinI mutants disassembly of the RecA filaments or bundles lasted for ≥45 min shown in Fig. 2 D, which were acquired with the DIC objective, during DSBR (Kidane and Graumann, 2005; Lesterlin et al., 2014; all the other images were acquired with the phase objective. The Badrinarayanan et al., 2015). In contrast, our data indicate that a microscope was enclosed in an incubation chamber to control the single, site-specific, replication-dependent DSB can be repaired temperature during live-cell fluorescence imaging. Before either much more efficiently. We conclude that a rapid and local mech- snapshot or time-lapse microscopy, the incubation chamber of anism has evolved to repair replication-dependent DSBs by the microscope was kept at 37°C, which was the temperature at homologous recombination, taking advantage of the proximity which the cells were grown in the liquid M9–glycerol medium. of the intact sister chromosome and the consequent facilitation Molten agarose (1.5%) was prepared in M9–glycerol medium of DNA homology searching. supplemented with arabinose and was mounted within a 1.5 × 1.6–cm Gene Frame (Thermo Fisher Scientific), which was sealed onto a microscope slide. A 6-µl aliquot of the cells growing in Materials and methods M9–glycerol medium was spread onto the solidified agarose and Bacterial strains and growth conditions sealed with a cover slide. The MetaMorph software (Molecu- All the E. coli strains used in this study are derivatives of lar Devices) was used for image acquisition. Phase-contrast or BW27784 (Khlebnikov et al., 2001) and are described in Table S1. DIC images were acquired concurrently with the fluorescence This background strain enables homogenous expression of the images. For snapshot microscopy, each fluorescence image con- SbcCD endonuclease from the arabinose-inducible promoter. sisted of 11 z sections with 200 nm z distance. The fluorescence Mutations were introduced by plasmid-mediated gene replace- images that were acquired during time-lapse microscopy con- ment (PMGR; Merlin et al., 2002) and confirmed by sequencing sisted of 6 z-sections with 350nm z-distance to minimize poten- and spot-test assays, where applicable. All plasmids used for tial photobleaching. The following settings were used during introducing the mutations are derivatives of pTOF24 (Table S2). image acquisition: an exposure time of 50 ms and a gain of one Sequences of the oligonucleotides that were used for construct- for phase and DIC; an exposure time of 50 ms and a gain of 150 for ing the pTOF24 derivatives and sequencing of the tet operator RecA-mCherry; and an exposure time of 100 ms and a gain of 250 (tetO) and lac operator (lacO) arrays are listed in Table S3. for LacI-Cerulean, TetR-Ypet, and DnaN-YPet images. For all microscopy experiments, the E. coli strain of interest The images obtained from microscopy were analyzed with was grown overnight at 37°C in M9-minimal medium supple- either ImageJ (National Institutes of Health) or MetaMorph mented with 0.2% of glycerol. The M9–glycerol medium was fur- software. During analysis, the z sections of each fluorescence ther supplemented with 100 ng/ml of anhydrotetracycline when image were converted to a single image that corresponded with strains were grown that contained the tetO array and expressed the sum of the stacked images. The sum image was deconvolved the TetR-YPet protein (Possoz et al., 2006). The overnight cul- with AutoQuant X software (Media Cybernetics). Deconvolution ture was diluted to an OD600 = 0.09 and grown for 3 h (OD600 = was performed to reduce noise and improve the resolution of 0.3–0.35). The bacterial culture was diluted again in the same the fluorescence image. Colocalization analysis was performed medium, and 0.2% of arabinose was added for induction of the with the MetaMorph software and ImageJ (0.4 µm was used as expression of the SbcCD endonuclease. The diluted culture was the upper limit for the distance between centroids of the two further grown for 2 h before microscopy. These conditions were foci under consideration). The OUFT I software (Paintdakhi et also used for growing E. coli cells during analysis of DNA con- al., 2016) was used for generating kymographs of the images tent by flow cytometry. For the spot-test assay, cells were grown that were obtained from time-lapse microscopy. OUFT I software overnight in Luria-Bertani (LB) medium before spotting of the was also used for measuring cell lengths and distances between cultures on LB agar plates. the centroid of the LacI-Cerulean foci and an arbitrary cell pole or the midcell (in AU). In all experiments the AU are the same. Spot-test assay The cftool function of MATL AB (MathWorks) was used for fitting A colony of the E. coli strain of interest was grown overnight in Gaussian curves on the data illustrating distances of LacI-Ceru- liquid LB medium at 37°C. The OD600 of the overnight culture was lean foci from the midcell. Amarh et al. Journal of Cell Biology 2305 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020 Quantification of cellular position of lacZ locus durations of RecA foci that were either above or below the cal- Time-lapse images were acquired for RecA-mCherry and LacI-Ce- culated median value (from Fig. 2 D) were counted and used for rulean at time intervals indicated. The time point at which two calculating the χ2 statistic. The χ2 statistic was calculated for the distinct LacI-Cerulean foci were formed during lacZ loci segre- WT strain and a mutant strain to ascertain whether the differ- gation was used as a reference point for determining the cellular ence between the median durations of RecA foci for these two position of the LacI-Cerulean foci during cohesion, which was 18 strains were statistically significant. This procedure for calcu- min before the occurrence of segregation (see Results and dis- lating the χ2 statistic was separately performed for each of the cussion). The time points preceding the 18-min cohesion were three mutants in comparison with the WT strain. defined as “before cohesion.” The procedure was repeated for predicted durations of cohesion (7, 9, 12, and 15 min, instead of Online supplemental material the 18 min that was determined in the absence of DSB induction. Figs. S1 and S2 confirm that DSBR and not postreplicative cohe- In the presence of DSB induction, appearance of the RecA- sion localizes the lacZ locus in the midcell region. Fig. S3 shows mCherry focus was used as the reference point for determin- the localization of the replisome in a cell during DNA replication. ing the cellular position of LacI-Cerulean foci before repair The E. coli strains, plasmids, and oligonucleotides used in this and during repair. During-repair data were collated from the study are provided in Tables S1, S2, and S3, respectively. appearance of the RecA-mCherry focus until splitting of the LacI- Cerulean focus. This approach was used because disassembly of the RecA-mCherry focus might not signify the end of the DSBR Acknowledgments because of the presence of unresolved joint molecules. We thank Ewa Okely, Mahedi Hasan, and Benura Azeroglu for technical assistance; Elise Darmon and Benura Azeroglu for crit- Analysis of DNA content by flow cytometry ical reading of the manuscript; and Elise Darmon for preparation Cells in exponential phase of growth (OD600 = 0.3–0.35) at 37°C in of the manuscript. M9–glycerol medium were treated with cephalexin and rifampi- This work was supported by the UK Medical Research Council cin. Cephalexin and rifampicin were added at a final concentra- (grant MR/M019160/1). tion of 10 and 150 µg/ml, respectively, and the cells were grown The authors declare no competing financial interests. for a further 3 h. The overnight, exponential phase and cepha- Author contributions: V. Amarh conceptualized the study, lexin/rifampicin-treated (runout) cultures were separately fixed provided the methodology, performed the investigation, pro- in 70% ethanol and stored overnight at 4°C. The cells in 70% eth- vided visualizations, and wrote, reviewed, and edited the orig- anol were harvested and washed in 1× PBS, and the DNA of these inal draft. M.A. White conceptualized the study, provided the cells was stained with 1× propidium iodide solution for 1 h in the methodology, performed the investigation, and reviewed and dark at room temperature. An A50 Micro Flow Cytometer (Apo- edited the original draft. D.R.F. Leach conceptualized the study, gee Flow Systems) was used for recording the fluorescence signal provided the methodology, provided resources and visualiza- generated by the stained cells after excitation with the blue laser tions, wrote, reviewed, and edited the original draft, provided (488 nm). The data obtained were analyzed with Apogee Histo- supervision, and acquired funding for the study. gram Software (version 3.1). Submitted: 6 March 2018 Sanger sequencing of DNA Revised: 27 April 2018 The complete DNA sequences of the tetO and lacO arrays were Accepted: 30 April 2018 determined via Sanger sequencing of PCR products amplified from the chromosomal loci bearing these operator arrays. PCR products were purified using the QIAquick PCR Purification kit (QIA GEN) before sequencing. The Sanger sequencing reaction References was set up using the BigDye Terminator v3.1 cycle-sequencing kit Badrinarayanan, A., T.B. Le, and M.T. Laub. 2015. Rapid pairing and resegre- (Applied Biosystems), following the manufacturer’s guidelines. gation of distant homologous loci enables double-strand break repair in bacteria. J. Cell Biol. 210:385–400. https://d oi. org/1 0. 1083/j cb. 201505019 The samples were sent to the Edinburgh Genomics Facility for Centore, R.C., and S.J. Sandler. 2007. UvrD limits the number and intensities analysis using the ABI 3730XL capillary-sequencing instrument. of RecA-green fluorescent protein structures in Escherichia coli K-12. J. Bacteriol. 189:2915–2920. https:// doi .org/ 10 .1128/ JB .01777 -06 Cockram, C.A., M. Filatenkova, V. Danos, M. El Karoui, and D.R. Leach. 2015. Statistical analysis Quantitative genomic analysis of RecA protein binding during DNA dou- The Mood’s median test was used for ascertaining whether there ble-strand break repair reveals RecBCD action in vivo. Proc. Natl. Acad. was a statistically significant difference between the median Sci. USA. 112:E4735–E4742. https:// doi. org/ 10 .1073/ pnas .1424269112Cox, M.M. 2007. Regulation of bacterial RecA protein function. Crit. Rev. Bio- duration of the RecA foci in WT cells compared with the cor- chem. Mol. Biol. 42:41–63. https:// doi. org/ 10 .1080/ 10409230701260258 responding duration in each deletion mutant that was studied Cox, M.M., M.F. Goodman, K.N. Kreuzer, D.J. Sherratt, S.J. Sandler, and K.J. (ΔuvrD, ΔrecX, or ΔdinI mutant). The Mood’s median test was Marians. 2000. The importance of repairing stalled replication forks. Nature. 404:37–41. https:// doi. org/ 10 .1038/ 35003501 chosen for this analysis because the durations of RecA foci were Drees, J.C., S.L. Lusetti, S. Chitteni-Pattu, R.B. Inman, and M.M. Cox. 2004. A determined independently for each bacterial strain and the dis- RecA filament capping mechanism for RecX protein. Mol. Cell. 15:789– tributions of these durations were similar (right-skewed) for all 798. https:// doi .org/ 10 .1016/ j .molcel .2004 .08 .026Eykelenboom, J.K., J.K. Blackwood, E. Okely, and D.R. Leach. 2008. SbcCD four strains. For each of these bacterial strains, the individual causes a double-strand break at a DNA palindrome in the Escherichia Amarh et al. Journal of Cell Biology 2306 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020 coli chromosome. Mol. Cell. 29:644–651. https://d oi. org/1 0. 1016/j . molcel DNA in vivo. EMBO J. 25:2596–2604. https:// doi .org/ 10 .1038/ sj .emboj .2007 .12 .020 .7601155 Khlebnikov, A., K.A. Datsenko, T. Skaug, B.L. Wanner, and J.D. Keasling. 2001. Renzette, N., N. Gumlaw, J.T. Nordman, M. Krieger, S.P. Yeh, E. Long, R. Homogeneous expression of the P(BAD) promoter in Escherichia coli Centore, R. Boonsombat, and S.J. Sandler. 2005. Localization of RecA by constitutive expression of the low-affinity high-capacity AraE trans- in Escherichia coli K-12 using RecA-GFP. Mol. Microbiol. 57:1074–1085. porter. Microbiology. 147:3241–3247. https:// doi .org/ 10 .1099/ 00221287 https:// doi .org/ 10 .1111/ j .1365 -2958 .2005 .04755 .x -147 -12 -3241 Reyes-Lamothe, R., C. Possoz, O. Danilova, and D.J. Sherratt. 2008. Indepen- Kidane, D., and P.L. Graumann. 2005. Dynamic formation of RecA filaments at dent positioning and action of Escherichia coli replisomes in live cells. DNA double strand break repair centers in live cells. J. Cell Biol. 170:357– Cell. 133:90–102. https:// doi. org/ 10 .1016/ j .cell .2008 .01. 04418394992 366. https:// doi. org/ 10 .1083/ jcb .200412090 Sinha, A.K., C. Possoz, A. Durand, J.M. Desfontaines, F.X. Barre, D.R.F. Kuzminov, A. 2001. Single-strand interruptions in replicating chromosomes Leach, and B. Michel. 2018. Broken replication forks trigger heritable cause double-strand breaks. Proc. Natl. Acad. Sci. USA. 98:8241–8246. DNA breaks in the terminus of a circular chromosome. PLoS Genet. https:// doi. org/ 10 .1073/ pnas .131009198 14:e1007256. https:// doi. org/ 10 .1371/ journal .pgen .1007256 Lesterlin, C., G. Ball, L. Schermelleh, and D.J. Sherratt. 2014. RecA bundles Skovgaard, O., M. Bak, A. Løbner-Olesen, and N. Tommerup. 2011. Genome- mediate homology pairing between distant sisters during DNA break wide detection of chromosomal rearrangements, indels, and mutations repair. Nature. 506:249–253. https:// doi .org/ 10 .1038/ nature12868 in circular chromosomes by short read sequencing. Genome Res. 21:1388– Lusetti, S.L., J.C. Drees, E.A. Stohl, H.S. Seifert, and M.M. Cox. 2004a. The 1393. https:// doi. org/ 10 .1101/ gr .117416 .110 DinI and RecX proteins are competing modulators of RecA function. J. Biol. Chem. 279:55073–55079. https:// doi .org/ 10 .1074/ jbc .M410371200 Stohl, E.A., J.P. Brockman, K.L. Burkle, K. Morimatsu, S.C. Kowalczykowski, Lusetti, S.L., O.N. Voloshin, R.B. Inman, R.D. Camerini-Otero, and M.M. Cox. and H.S. Seifert. 2003. Escherichia coli RecX inhibits RecA recombinase 2004b. The DinI protein stabilizes RecA protein filaments. J. Biol. Chem. and coprotease activities in vitro and in vivo. J. Biol. Chem. 278:2278– 279:30037–30046. https:// doi .org/ 10 .1074/ jbc .M403064200 2285. https:// doi .org/ 10 .1074/ jbc .M210496200 Mangiameli, S.M., B.T. Veit, H. Merrikh, and P.A. Wiggins. 2017. The repli- Veaute, X., S. Delmas, M. Selva, J. Jeusset, E. Le Cam, I. Matic, F. Fabre, and somes remain spatially proximal throughout the cell cycle in bacteria. M.A. Petit. 2005. UvrD helicase, unlike Rep helicase, dismantles RecA PLoS Genet. 13:e1006582. https:// doi .org/ 10 .1371/ journal .pgen .1006582 nucleoprotein filaments in Escherichia coli. EMBO J. 24:180–189. https:// Marnef, A., S. Cohen, and G. Legube. 2017. Transcription-Coupled DNA Dou- doi .org/ 10 .1038/ sj .emboj .7600485 ble-Strand Break Repair: Active Genes Need Special Care. J. Mol. Biol. Vilenchik, M.M., and A.G. Knudson. 2003. Endogenous DNA double-strand 429:1277–1288. https:// doi. org/ 10 .1016/ j .jmb .2017 .03 .02428363678 breaks: production, fidelity of repair, and induction of cancer. Proc. Merlin, C., S. McAteer, and M. Masters. 2002. Tools for characterization of Natl. Acad. Sci. USA. 100:12871–12876. https:// doi .org/ 10 .1073/ pnas Escherichia coli genes of unknown function. J. Bacteriol. 184:4573–4581. .2135498100 https:// doi .org/ 10 .1128/ JB .184 .16 .4573 -4581 .2002 Wang, X., C. Possoz, and D.J. Sherratt. 2005. Dancing around the divisome: Michel, B., G. Grompone, M.J. Florès, and V. Bidnenko. 2004. Multiple asymmetric chromosome segregation in Escherichia coli. Genes Dev. pathways process stalled replication forks. Proc. Natl. Acad. Sci. USA. 19:2367–2377. https:// doi. org/ 10 .1101/ gad .345305 101:12783–12788. https:// doi. org/ 10 .1073/ pnas .0401586101 Wang, X., C. Lesterlin, R. Reyes-Lamothe, G. Ball, and D.J. Sherratt. 2011. Rep- Paintdakhi, A., B. Parry, M. Campos, I. Irnov, J. Elf, I. Surovtsev, and C. lication and segregation of an Escherichia coli chromosome with two Jacobs-Wagner. 2016. Oufti: an integrated software package for high-ac- replication origins. Proc. Natl. Acad. Sci. USA. 108:E243–E250. https:// curacy, high-throughput quantitative microscopy analysis. Mol. Micro- doi .org/ 10 .1073/ pnas .1100874108 biol. 99:767–777. https:// doi .org/ 10 .1111/ mmi .13264 White, M.A., J.K. Eykelenboom, M.A. Lopez-Vernaza, E. Wilson, and D.R. Possoz, C., S.R. Filipe, I. Grainge, and D.J. Sherratt. 2006. Tracking of controlled Leach. 2008. Non-random segregation of sister chromosomes in Esche- Escherichia coli replication fork stalling and restart at repressor-bound richia coli. Nature. 455:1248–1250. https://d oi. org/1 0. 1038/n ature07282 Amarh et al. Journal of Cell Biology 2307 RecA dynamics at DNA double-strand breaks https://doi.org/10.1083/jcb.201803020