Dysregulated DnaB unwinding induces replisome decoupling and daughter strand gaps that are countered by RecA polymerization

Abstract

The replicative helicase, DnaB, is a central component of the replisome and unwinds duplex DNA coupled with immediate template-dependent DNA synthesis by the polymerase, Pol III. The rate of helicase unwinding is dynamically regulated through structural transitions in the DnaB hexamer between dilated and constricted states. Site-specific mutations in DnaB enforce a faster more constricted conformation that dysregulates unwinding dynamics, causing replisome decoupling that generates excess ssDNA and induces severe cellular stress. This surplus ssDNA can stimulate RecA recruitment to initiate recombinational repair, restart, or activation of the transcriptional SOS response. To better understand the consequences of dysregulated unwinding, we combined targeted genomic dnaB mutations with an inducible RecA filament inhibition strategy to examine the dependencies on RecA in mitigating replisome decoupling phenotypes. Without RecA filamentation, dnaB:mut strains had reduced growth rates, decreased mutagenesis, but a greater burden from endogenous damage. Interestingly, disruption of RecA filamentation in these dnaB:mut strains also reduced cellular filamentation but increased markers of double strand breaks and ssDNA gaps as detected by in situ fluorescence microscopy and FACS assays, TUNEL and PLUG, respectively. Overall, RecA plays a critical role in strain survival by protecting and processing ssDNA gaps caused by dysregulated helicase activity in vivo.

Graphical Abstract

Graphical Abstract.

Introduction

Replicative hexameric helicase enzymes are at the forefront of DNA replication, a highly conserved process fundamental to all life. Yet, the mechanisms for how these enzymes regulate their activities and function in concert with other replisome proteins are not fully understood at a mechanical or biological level. While the individual roles of replication proteins, including helicases, are well known, the mechanisms by which they regulate interactions with DNA and other proteins in a cellular environment remain to be fully elucidated. The bacterial hexameric DNA helicase, DnaB, is responsible for exposing single stranded DNA (ssDNA) strands as it creates the ‘fork’ around which replisome proteins orient (1,2). DnaB is a RecA-like superfamily 4 (SF4) helicase, translocating and unwinding in the 5′-3′ direction on the lagging strand with a C-terminal domain (CTD) first directionality (3,4). Not only are SF4 helicases well characterized, but Escherichia coli is a well-tested model organism, making this an ideal system for investigating replisome mechanics in vivo.

DnaB acts as a stable replisome hub that assists in organization and coordination of replication fork activities (5–8). The N-terminal domain (NTD) of DnaB interacts with the DnaG primase to control priming, and the Tau (τ) subunit of the clamp loader complex (CLC) physically and functionally connects the leading and lagging strand polymerases to DnaB. In addition to replisome proteins, DnaB interacts with DnaA and DnaC for replication initiation and loading and is involved in the resolution or bypass of certain replication blocks at the fork (2,9–13). Therefore, DnaB’s unwinding activity is likely regulated through multidimensional interactions with several replication enzymes for various replisome related purposes. Indeed, distinct modes of unwinding have been observed. Coupled unwinding and synthesis in an active replisome proceeds rapidly at ∼1000 nt/s (8,14), but during replication initiation in a complex called the preprimosome and when decoupled from the polymerase during replication, the helicase unwinds at a much slower rate of ∼35–100 nt/s (8,15–17). The mechanism for this observed kinetic regulation is still being understood.

Biochemical and structural investigations of DnaB have revealed two distinct structural conformers: a constricted state with a narrow central pore for tight interactions with ssDNA and fast unwinding and a slower, dilated state that allows for translocation over duplex DNA (Figure 1) (18–20). The dilated conformer is used for interactions with DnaC for loading at origins, and with DnaG when synthesizing RNA primers (18–20). Investigations into sterically restricted, dilated, or constricted mutants revealed a direct link between conformation, rate of unwinding, and protein–protein interactions in vitro. Studies have also shown that both DNA strands contribute to unwinding, and a mechanism involving steric exclusion and wrapping (SEW) of the excluded ssDNA strand contributes to helicase regulation in vivo and in vitro (21–23).

Figure 1.

DnaB structural conformations and nomenclature. The hexameric DnaB helicase can interconvert between a dilated state (D) (PDBID: 2R6A) in the Apo form or upon interaction with the DnaG primase with an interior diameter of 47 Å or a constricted state (C) induced upon ssDNA (pink), DnaC, or τ-CLC binding with a diameter of 22 Å. Point mutations in the DnaB protein (K180A, green or R328A/R329A, blue) also shift the equilibria towards a constricted state (C). Analogous E. coli chromosomal mutations in the dnaB gene have been made to create strains, MSB4 and MSB5, respectively, from the parent, HME63 strain.

Previous investigations of targeted mutations that disrupted helicase regulation led to proposal of a dual mechanism, wherein conformational changes in the helicase structure stimulated by protein–protein and protein–DNA interactions act as a molecular switch to conceal or expose the external residues for engagement with the excluded DNA strand (21). Upon loading onto ssDNA by DnaC or interaction with τ-CLC, the central channel of the DnaB hexamer constricts, making tighter and more productive contacts with the encircled strand for rapid unwinding (7,10,24,25) (Figure 1). In vivo, these targeted helicase regulation-deficient mutations led to substantial amounts of genomic stress and reduced overall strain growth and efficacy (21). While helicase regulation is important for genomic stability and damage mitigation, it is not known how dysregulated helicase activity alters the chromosome architecture nor what downstream DNA repair pathways are utilized to maintain a functional genome despite more rapid and unregulated unwinding.

RecA binds and cooperatively polymerizes on ssDNA to form filaments for the purposes of homologous recombination (HR) repair of double strand breaks (DSBs), marking lesion containing gaps for the RecFOR restart pathway, and induction of the bacterial SOS response upon sufficient RecA filamentation through allosterically inducing autolytic cleavage of the SOS repressor, LexA (26–32). While there are rare instances of RecA-independent recombinational repair in bacteria, E. coli lack the non-homologous end joining (NHEJ) pathway as an HR alternative. Sufficient RecA filamentation on ssDNA is needed to trigger SOS induction, and RecA-dependent recombination is the only efficient way for E. coli to repair DSBs (26,29,33). While DNA damage repair proteins and the translesion synthesis (TLS) polymerase IV (Pol IV) have low-level constitutive expression, activation of the SOS response dramatically upregulates multiple proteins important for rapid DNA repair in the face of severe stress, including a host of DNA damage tolerant (DDT) proteins (34). For one, the TLS polymerase with the least fidelity, Pol V, is induced and activated as part of the late-stage SOS response (35–37). While it is unclear at what threshold length RecA filaments trigger SOS induction, functional RecA binding and polymerization is necessary for an efficient response to high levels of genomic stress that include DNA damage and ssDNA gaps.

Exterior SEW mutations in DnaB previously shown to disrupt interactions with the excluded strand and enforce a fast constricted unwinding state (22) are now shown to decouple the replisome and limit leading strand synthesis. These same targeted genomic dnaB mutations (Figure 1) previously shown to have dysregulated unwinding leading to cellular stress and genomic instability (21) are now combined with an inducible plasmid-based switch that inhibits RecA filamentation at ssDNA gaps. We find that RecA filamentation activity is important for maintaining genomic integrity when helicase regulation is impaired. The high mutational frequency in these dnaB:mut strains is dependent on mutagenic repair or restart from proper RecA filamentation. Addition of exogenous DNA damage agents to these strains generally sensitized growth further when RecA filamentation was inhibited. A novel Pol I dUTP Gap filling (PLUG) assay was used to label single strand gaps in situ and found that dysregulated unwinding leads to increased ssDNA gaps in vivo, consistent with a model where helicase regulation encourages functional replisome coupling. In the absence of effective RecA filamentation, these ssDNA gaps are effectively converted to double strand breaks (DSBs) as detected by in situ fluorescence microscopy and fluorescence activated cell sorting (FACS) and Terminal dUTP Nick End Labeling (TUNEL) assays. This work advances our understanding of the impact of helicase regulation on efficient replisome progression and the dependence on RecA to mitigate consequences from replisome decoupling for cell survival.

Materials and methods

Materials

ATP was from Invitrogen (Carlsbad, CA). α-³²P- dATP was from Perkin Elmer (Waltham, MA). DNA substrates were HPLC or gel purified from Sigma-Aldrich (St. Louis, MO) or IDT (Coralville, IA). Mitomycin C (MMC) (Fisher Bioreagents, Waltham, MA) and Hydroxyurea (HU) (Acros Organics, Waltham, MS) were dissolved in DMSO; anhydrotetracycline (aTc) (Thermo Fisher Scientific, Waltham, MA) was dissolved in ethanol; Methyl methanesulfonate (MMS) (Alpha Aesar, Haverhill, MA) was dissolved in ultrapure water prior to use. Rifampicin (rif) (Fisher Bioreagents, Waltham, MA) was dissolved in DMSO. Sytox Green (SG) was from Invitrogen (Carlsbad, CA). All other materials were from commercial sources and were analytical grade or better.

Overexpressed and purified proteins

The E. coli DnaB helicase wild-type and mutants (R74A, R164A, K180A, RR328/329AA) (22), DnaC loader (21), Pol III core (38), ß-clamp (39), τ₃ clamp-loader-complex (τ₃-CLC) (40) and SSB (41) were expressed and purified essentially as previously described.

The plasmid pET28a-^HisDnaB (gift from James Keck) (42), was utilized for site-directed mutagenesis to create the K180A and RR328/9AA DnaB mutants using Platinum SuperFi II DNA Polymerase and partially overlapping primers containing novel restrictions sites, NaeI and EagI, respectively (Supplementary Table S3). Plasmids were isolated and confirmed by restriction digest and plasmid sequencing (Plasmidsaurus, Eugene, OR). His-tagged DnaB (WT, K180A, and R328A/R329A) was overexpressed in Rosetta 2 E. coli cells upon induction with 1 mM IPTG at OD₆₀₀ = 0.3 and then harvested by centrifugation after 3 hours of additional growth at 37°C. The pellet was resuspended in His A buffer (500 mM NaCl, 50 mM imidazole, 20 mM Tris [pH 7.5], 10% glycerol) supplemented with 2 mM PMSF and 0.4 mg/ml of lysozyme and allowed to rock at 4°C for 1 h. Cells were additionally lysed by sonication and then clarified by centrifugation at 21 000 rpm, 30 min, 4°C. The supernatant was filtered through a 0.45 micron PES syringe filter before loading onto a HisTrap column (Cytiva, Marlborough, MA), washing with His buffer B (His Buffer A with 50 mM NaCl), and eluting with a linear gradient with His buffer C (His buffer B with 500 mM imidazole). DnaB fractions were pooled and loaded onto a HiTrap QFF column, washed with QFF buffer A (20 mM Tris [pH 7.5], 10 mM MgCl₂, 0.1 M NaCl, 10% glycerol, 0.01 mM ATP, 1 mM BME) and eluted by QFF buffer B (QFF buffer A + 0.75 M NaCl) over a long 14CV gradient. Fractions containing DnaB were pooled, concentrated, and loaded onto a Superdex S-200 column (Cytiva, Marlborough, MA) equilibrated in GF Buffer (20 mm Tris–HCl [pH 8.5], 0.8 M NaCl, 10% glycerol, 5 mM MgCl₂, 0.1 mM ATP). Fractions containing pure DnaB were combined and dialyzed against storage buffer (20 mm Tris–HCl [pH 8.5], 0.5 M NaCl, 5 mM DTT, 50% glycerol) and stored at −80°C.

DNA unwinding and translocation

The unwinding and translocation assays were performed essentially as described previously (18,21). The DNA substrates (Supplementary Table S3) were HPLC purified from Integrated DNA technologies (IDT, Coralville, IA). For unwinding, Cy3-labeled DNA165 was mixed with BHQ-DNA181 in a 1:1.1 ratio in 1× annealing buffer [20 mM Tris–HCl, 4% glycerol, 0.1 mM EDTA, 40 μg/ml BSA, 10 mM DTT, 10 mM Mg(OAc)₂ (pH 8)] and incubated at 37°C overnight. For translocation, oligos Cy3 labeled-MAT165, BHQ-MAT181 and unlabeled MAT180 was mixed in a 1:5:1 ratio and annealed similarly. The reactions contained 1x reaction buffer [10 mM HEPES–KOH, 5 mM Mg(OAc)₂, 50 mM potassium glutamate, 5% glycerol, 0.2 mg/ml BSA, 4 mM DTT, pH 7.5], 50 nM of respective DnaB₆ protein, followed by addition of 50 nM of CLC or 300 nM of DnaC and incubated on ice for 10 minutes. 25 nM of annealed substrate was added to the mix and incubated for afurther 10 minutes on ice. The reactions were transferred to a 384-well black plate (Corning) and initiated simultaneously with 1 mM of ATP and 100 nM of trap DNA (unlabeled MAT165). The fluorescence (ex 510 nm/em 580 nm) corresponding to unwinding was measured at 30-second intervals on a Tecan Spark microplate reader (Männedorf, Switzerland).

DnaB–τ pull-down assay

Pull-down assays were performed by adding 50 μl of nickel resin (HisPur™ Ni-NTA Resin, Thermo scientific) to a 1.5 ml Eppendorf tube and spun down at 700 g for 2 min. The supernatant was removed, and the pellet was resuspended in 50 μl of binding buffer [20 mM sodium phosphate, 0.125 M NaCl, 40 mM imidazole, 1 mM ATP, pH 7.4], spun down again, and the supernatant discarded. A 1:2 ratio of DnaB WT to τ₃-CLC are mixed with the resin and allowed to incubate at 37°C for 5 min before spinning down and transferring the flow through (FT) solution to a new tube. The pellet was resuspended in 50 μl of binding buffer, spun down, and transferred to a new tube as the first wash, taking care not to disturb the resin pellet. This wash step is repeated 12 more times (13 total) to ensure that all unbound protein is removed before resuspending the resin in 50 μl of elution buffer [20 mM sodium phosphate, 0.125 M NaCl, 500 mM imidazole, pH 7.4], spinning down, and transferring the elution solution to a new tube. Finally, the FT, washes, and elution are electrophoresed on a 10% SDS-page gel and a western blot was performed using custom primary antibodies to detect DnaB and τ (7) and the secondary antibody, Goat anti-rabbit Cy5 (#ab6564, Abcam, Cambridge, UK), and visualized on the Typhoon RGB gel imaging scanner (Cytiva, Marlborough, MA).

TFII substrate preparation

The tailed form II (TFII) substrate was prepared using the pSCW01 plasmid (2 kb) as described previously (43). Briefly, pSCW01 was isolated from cell pellets using ZymoPURE plasmid midiprep kit (Irvine, CA). 100 μg of pSCW01 plasmid was treated with 1.5 units/μg of site-specific endonuclease, Nt.BstNBI (New England Biolabs, Ipswitch, MA), and 100× molar excess of displacer oligonucleotides (DNA197–199, sequences complementary to the single stranded DNA fragments created by Nt.BstNBI enzyme, Supplementary Table S3) in 1× NEB 3.1 buffer at 55°C for 4 h to create a gapped plasmid. The Nt.BstNBI enzyme was heat inactivated at 80°C for 20 min according to the manufacturer's instructions. In the same reaction tube, the nicked DNA fragments were annealed with the displacer oligos in a thermal cycler at a cooling rate of 1°C/min until the reaction reached 12°C. To extract the gapped plasmid, the displacer oligos were removed using PEG purification, where an equal volume of 2× PEG solution (26% [w/v] PEG-8000, 20 mM MgCl₂) was added to the reaction mixture, centrifuged at 21 000 × g for 1 h at 6°C and the supernatant was discarded. The pellet containing the gapped plasmid was washed with 70% (v/v) ice cold ethanol and centrifuged at 21 000 × g for 30 min at 6°C. The supernatant was discarded, air dried, and the resulting pellet was resuspended in 80 μl of Milli-Q water prewarmed to 65°C. The gapped plasmid was then annealed to 3× molar excess of fork oligonucleotide, DNA200 in thermal cycler in 1× Cutsmart buffer at 50°C for 10 min, followed by cooling to 16°C at a 1°C/min cooling rate. Next, ligation was performed by the addition of 62.5 units of T4 DNA ligase per μg of DNA substrate supplemented with 8 mM ATP and 1 mM DTT and incubated at 16°C for 18 hrs. The ligase was heat inactivated at 65°C for 20 min according to the manufacturer's instructions. The TFII substrate was purified as before by the addition of an equal volume of 2× PEG solution, followed by washing with 70% ethanol. The resulting pellet was resuspended in 1× TE buffer.

TFII rolling circle assay

The TFII substrate (5 nM) was incubated with 60 nM DnaB₆, 360 nM DnaC, 30 nM τ₃χφδδ’(τ₃-CLC), 90 nM Pol III core (αϵθ), 200 nM β₂ in 1X replication buffer at 37°C for 10 min. The replication reaction was initiated by adding a mixture of 50 nM SSB, 1.25 mM ATP, 125 μM dNTPs, α-³²P-dATP in 1× replication buffer (50 mM HEPES [pH 7.9], 12 mM Mg(OAc)₂, 0.1 mg/ml BSA, 10 mM DTT) prewarmed to 37°C. The replication reaction was stopped by mixing with an equal volume of quench buffer (200 mM EDTA, 300 mM NaOH, 18% Ficoll [w/v], 0.15% [w/v] Bromocresol green, 0.25% [w/v] xylene cyanol) at different time points. The reaction products were electrophoresed on a 1% alkaline agarose gel at 35 V until the dye front ran two thirds of the gel. The gels were dried, exposed overnight, and imaged on a Typhoon FLA 9000 phosphorimager (Cytiva, Marlborough, MA). Alternatively, alkaline agarose gels were neutralized in 2× TAE for 2 h on a rocker, stained with 1x SybrGold (Invitrogen) in 2× TAE for 2 h, and imaged similarly. Extended ³²P products above the 2 kb substrate were quantified as a function of time using ImageQuant software (Cytiva, v.10.1) and the data was plotted using Prism 10.1 (GraphPad, San Diego CA).

Linear TFII unwinding assay

The circular 5′Cy5 flap labeled TFII substrate was digested XhoI/AseI in NEB buffer 3.1 at 37°C for 2 h before inactivation at 65°C for 20 min, cooled at a rate of 1°C/min until the reaction reached 4°C to generate the linear 1219 bp substrate. 5 nM of the linear digested TFII was incubated with 60 nM DnaB₆, 360 nM DnaC in 1× reaction buffer at 37°C for 5 min. The reaction was initiated by adding a mixture of 50 nM SSB and 1.25 mM ATP prewarmed to 37°C. The unwinding reaction was quenched by mixing an equal volume of 2× LES buffer (Orange G, 200 mM EDTA, 2% SDS) at different time points and resolved on 1% alkaline agarose gels.

RecA56 vector engineering

To create a plasmid-based RecA56 expression system, the plasmid pEM-Cas9HF1-indRecA56 (Addgene ID: 102294) designed for the co-inducible expression of Cas9 and RecA56 (44) was modified using primers Cas9HF1delta F and R (Supplementary Table S3) to create pEM-indRecA56 (pEM-RecA56). These primers remove Cas9, keeping the tetracycline response element (TRE) and recA56 genes intact, allowing for inducible expression solely of the recA56 gene. Top10 cells were transformed with this modified plasmid and screened by restriction digest and DNA gel electrophoresis, before being electroporated into test strains (45).

Growth curves

Growth curves were recorded by diluting overnight clonal cultures 1:100 (OD₆₀₀ ∼ 0.01) in Miller LB (10 g typtone (Research Products International, molecular biology grade), 10 g NaCl, 5 g yeast extract/l, pH 7.0) and aliquoting 250 μl into a round bottom clear 96-well plate (Corning) with or without the addition of 100 ng/ml anhydrotetracycline (aTc). The cultures were incubated at 32°C (to prevent λ-Red induction in these strains) with aeration at 225 RPM and the OD₆₀₀ was recorded at 30-min intervals using a Tecan Spark microplate reader equipped with Spark Control Magellan software (Tecan, Männedorf, Switzerland). The specific growth rate (μ) was calculated from a rolling 1-h average of the slopes in the growth curves and the peak maximal specific growth rates () were determined and plotted using Prism 10.1 (GraphPad, San Diego CA).

Strain mutagenesis assay

To determine the mutation frequency of each strain, a rifampicin resistant (rif^R) assay was performed as previously described (46), with the modification of using Miller LB media instead (21). To test the mutagenesis rate with and without RecA activity impairment, fresh overnight cultures were diluted 1:100 in Miller LB media with or without the addition of 100 ng/ml aTc (Thermo Fisher Scientific, Waltham, MA). Cultures were grown at 32°C with aeration until OD₆₀₀ ∼ 0.4, then again diluted 1:100 in fresh Miller LB (± aTc) and allowed to grow for 24 h. 100 μl aliquots were spread onto plates containing 50 μg/ml rifampicin (ThermoFisher, Waltham, MA) (rif⁺). Identical aliquots were diluted and plated onto regular Miller LB (rif⁻) plates to calculate colony-forming units (CFUs). Plates were incubated at 32°C until colonies appeared or 48 h was reached. Mutation frequency was calculated as the ratio of mutants to total CFUs as follows:

(1)

where A is the number of mutant CFUs (colonies on the rif⁺ plate), B is the number of total CFUs (colonies on the rif⁻ plate), and 10⁶ is the dilution factor for B.

Minimum inhibitory concentration (MIC) assay

For MIC assays, fresh clonal cultures from glycerol stocks containing pEM-RecA56 were grown overnight at 32°C in Miller LB medium containing appropriate antibiotics and 100 ng/ml aTc for induction of RecA56. Clear, round bottom, 96-well culture plates were prepared by aliquoting 210 μl of Miller LB medium without antibiotics, 100 ng/ml aTc, the indicated concentration of genotoxin (MMC or MMS), and 2 μl of culture into each well. 96-well culture plates were transferred to a 32°C shaker and removed briefly for imaging and data collection at 24 and 48 h. 96-well culture plate images were taken on the benchtop using a handheld camera positioned underneath the plate, and OD₆₀₀ readings were collected with a Tecan Spark microplate reader (Tecan, Männedorf, Switzerland). Data was exported, plotted using Prism 10.1 (GraphPad, San Diego CA), and fit to the following equations to determine MIC (47),

(2)

(3)

where A is the lower asymptote of y, B is the slope, C is the amplitude and M is the log concentration of the inflexion point.

Genotoxin survival assays

For cell survival assays, fresh clonal cultures from glycerol stocks ± pEM-RecA56 were grown overnight at 32°C in Miller LB medium with 100 ng/ml aTc for induction of RecA56. Cultures were normalized by OD₆₀₀, then serially diluted and spotted in 2 μl aliquots onto Miller LB-agar plates containing the indicated concentration of genotoxin (MMC, HU, MMS) and 100 ng/ml aTc. In order to plate stains in the presence and absence of pEM-RecA56, spotting was performed on plates lacking chloramphenicol and so it is possible that pEM-RecA56 may have been lost in some fraction of the cells. Plates were incubated at 23°C for approximately 1 h, then transferred to a 32°C incubator for 24–48 h before imaging and data collection. All stock solutions were made at 1000× prior to mixing with cooled Miller LB agar. Plate images were taken on the benchtop using a handheld camera; data was processed using Excel and plotted using Prism 10.1 (GraphPad, San Diego CA).

Fluorescent detection of DNA nicks and gaps in situ

Exponential growth cultures were obtained by diluting overnight cultures 1:100 in Miller LB ± 100 ng/ml aTc for RecA56 induction and grown overnight at 32°C. Overnight cultures were again diluted 1:100 in 50 ml Miller LB ± 100 ng/ml aTc and grown with aeration at 32°C until OD ∼0.5. Cells were harvested by pelleting and washing with PBS, then fixed in 1 ml of ice cold formaldehyde solution (4% paraformaldehyde in 1× PBS) for 20 min at room temperature, pelleted, and washed with PBS as described (48). For internal controls, exponential growth cultures of the parental strain HME63, the K12 strain MG1655, and the Keio collection ΔrecA strain JW2669 (49,50) were exposed to 1 μg/ml MMC for 60 min prior to harvesting. After fixation, cells were permeabilized by rapid resuspension in 77% ice cold ethanol, then pelleted, washed, resuspended in a PBS storage solution containing 0.1% PFA, and kept at 4°C.

To visualize DNA breaks, Terminal dUTP Nick End Labeling (TUNEL) was utilized to label free 3′ DNA ends (51). For direct comparison of results, half of each fixed sample was processed for TUNEL analysis, and the remaining half was reserved for single strand gap detection (below). dUTP was added to DNA ends by pelleting stored cells and resuspending in 100 μl of elongation buffer (1× terminal deoxytransferase [TdT] buffer, 2.5 mM CoCl₂, 0.3 mM 5-bromo-2′-deoxyuridine [BrdU] [Invitrogen, Waltham, MA]), then adding 5 U of TdT (Thermo-Fisher, Waltham, MA), and incubating at 37°C for 60 min. After elongation, cells were pelleted, washed with PBS, and then resuspended in blocking solution (4% BSA in 1× TBST) for 30 min at room temperature. To fluorescently label BrdU labelled ends, blocked cells were pelleted and resuspended in 100 μl of primary antibody solution (1:100 mouse-α-BrdU [BD Bioscience, Franklin Lakes, NJ] in TBST with 2% BSA) for 60 min at room temperature. Afterwards, cells were pelleted and washed three times with blocking solution, then resuspended in 100 μl of secondary antibody (1:500 α-mouse IgG-Alexa647 [ThermoFisher, Waltham, MA] in TBST with 2% BSA), and incubated in the dark for 60 min at room temperature. Cells were pelleted again, washed with TBST, washed twice with PBS, then resuspended in the PBS storage solution and kept at 4°C for analysis.

To visualize single stranded DNA, an exonuclease deficient Klenow variant of Pol I (New England Biolabs, Ipswich, MA) was used to extend DNA ends across daughter strand gaps (52,53) in a Pol I dUTP Gap-filling (PLUG) assay. The reserved half of the stored fixed cells were pelleted and resuspended in 100 μl Klenow extension buffer (1× NEB buffer 2, 10 μM each of dATP, dCTP, dGTP, and BrdU [ThermoFisher]). To initiate the elongation, 5 U of Klenow was added, and the reaction was incubated at room temperature for 40 min. After elongation, cells were pelleted, washed thoroughly with PBS, then resuspended in blocking solution (4% BSA in 1× TBST) for 30 min at room temperature. Incorporated BrdU nucleotides were labeled as described above for the TUNEL assay. After secondary antibody incubation, cells were washed with TBST, then resuspended in PBS storage solution and kept at 4°C for analysis by epifluorescence microscopy and flow cytometry.

Microscopy

All microscopy images were obtained using an Olympus Brightfield Microscope IX-81 (Olympus Corp., Center Valley, PA) using a 60× objective with oil immersion. 2 μl of fixed samples from TUNEL and Klenow extension assays were spotted onto a microscope slide and allowed to dry. DAPI (ThermoFisher, Waltham, MA) was added to mounting media (2.5% DABCO, 90% glycerol, 7.5% PBS) to create a dual staining and mounting solution. 3 μl of this solution was added to cover the fixed cells, then immediately topped with a coverslip and sealed with clear polish. Slides were stored at 4°C in the dark overnight prior to imaging.

The DAPI stained images were used to set masks for individual cells with manual adjustments made for overlapping cells or those not identified using ImageJ (v.153) (54). The area embedded by the masks was utilized to give cell area. Alexa647 foci from TUNEL or PLUG assays were quantified from pixel maxima within the mask to give the foci per cell. Foci per cell was divided by cell area to obtain foci density. This data was analyzed by excel and plotted using Prism 9.5 (GraphPad, San Diego CA).

Flow cytometry

To quantify cell size and BrdU intensity, TUNEL and PLUG treated and Alexa647 antibody-stained cultures were pelleted, resuspended in 1 ml sheath fluid (sterile 1× PBS) with 1.5 μM Sytox Green, and incubated in the dark for 30 min. Samples were diluted with additional sheath fluid based on cell density and instrument parameters before being analyzed by FACSverse (BD Biosciences). HME63, MG1655 and JW2669 exposed to 1 μg/ml MMC prior to harvesting were used as positive controls; untreated cells stained with Sytox Green and unstained fluorescently labeled BrdU cells were used as signal controls (51).

Using FloJo software (v10.8.1, BD biosciences), cells were gated for single cells based on a forward scatter height vs. area scatter plot and universal gates for ±BrdU and ±SG subpopulations were set based on allophycocyanin (APC) and fluorescein isothiocyanate (FITC) controls, respectively. To calculate BrdU intensity for TUNEL and PLUG, FACS histograms of each sample were exported from FloJo as a series of individual values. Each sample's total fluorescence of +BrdU and +SG gates were independently calculated using the following equation:

(4)

where i is the bin number, X is the AFU of that bin, Y is the cell count of that bin, and n is the total number of bins associated with the sample. Total fluorescence values were calculated in Excel and data was plotted using Prism 10.1 (GraphPad, San Diego CA).

Results

Constricted DnaB mutants induce replisome decoupling

Previously, we have shown that several surface mutations on the DnaB helicase can induce a change in the hexameric structure to a more constricted conformation and dramatically increase unwinding activity (Figure 1) (21,22). The increase in DNA unwinding activity by these DnaB mutants suggested two possibilities when incorporated in the context of a replisome: (i) either faster unwinding leads to faster synthesis or (ii) faster unwinding decouples synthesis and unwinding activities resulting in slower replisome progression. To determine which of these possibilities occur, DnaB mutants were incorporated into a complete in vitro bacterial replisome system using a circular TFII DNA substrate (Supplementary Figure S1) to monitor coupled unwinding and leading strand synthesis (Figure 2A) (43). 50 nM of SSB was added in the initiation mix to prevent premature melting of the duplex region. A range of SSB concentrations have been used in these TFII assays by different groups (from 0 to 2000 nM) (43,55–59); however, at concentrations >50 nM, we noticed an inhibition in leading strand synthesis likely from premature duplex melting or equilibria binding competition (Supplementary Figure S2A). Additionally, when γ₃-CLC is used instead of τ₃-CLC to remove the contact with DnaB needed for coupled unwinding and synthesis, there is an almost total inhibition of product formation (Supplementary Figure S2B).

Figure 2.

Constricted DnaB mutants decouple the replisome. (A) Reaction scheme depicting the replisome components used including TFII substrate (5 nM), 60 nM DnaB₆, 360 nM DnaC, 30 nM τ₃χφδδ’ (CLC), 90 nM Pol III core (αϵθ), 200 nM β₂ in 1× replication buffer at 37°C for 10 min. The replication reaction was initiated by adding a mixture of 50 nM SSB, 1.25 mM ATP, 125 μM dNTPs, [α-³²P] dATP. (B) ³²P incorporated into all extended products with WT or mutant DnaB separated on a 1% alkaline agarose gel and (C) quantified at 0.5, 1, 2, 3 min to calculate the linear rate of total product synthesis. (D) Quantification of the average slope of the total extended product from three independent experiments. Error bars represent the standard error of the mean. Black bars indicate statistically significant differences, where P-values are indicated and represented by *P < 0.05, **P < 0.01 and ***P < 0.001, from an unpaired two-sided t-test. Reaction schematic and fluorescent plots of the (E) DNA unwinding or (F) duplex translocation assays sensitizing Cy3 fluorescence upon separation of the BHQ oligo for WT (grey), K180A (green), and RR328/9AA (blue) alone (○) at 50 nM hexamer, +300 nM DnaC (•), or + 50 nM τ-CLC (▴).

Leading strand synthesis was measured using α-³²P-dATP incorporation over time. Interestingly, leading strand synthesis, measured by total product abundance (or full-length product only, Supplementary Figure S3), is significantly reduced when fully constricted DnaB mutants (K180A and RR328/9AA) are utilized compared to WT (Figure 2B–D). WT DnaB had a total DNA synthesis rate of 3.6 ± 0.2 × 10⁶ min⁻¹, while moderately constricted mutants, R74A and R164A, had similar rates of 1.9 ± 0.5 × 10⁶ and 2.6 ± 0.5 × 10⁶ min⁻¹, respectively, with only a ∼1.9–1.4-fold reduction in total product (Figure 2D), but was insignificant when quantifying full-length product (Supplementary Figure S3). Alternatively, fully constricted mutants, K180A and RR328/9AA, had significantly reduced synthesis rates of 0.6 ± 0.3 × 10⁶ min⁻¹ (∼6-fold) and 0.02 ± 0.02 × 10⁶ min⁻¹ (∼180-fold), respectively, compared to WT. Therefore, mutations in DnaB that enforce a constricted hexamer state (i.e. K180A or RR328/9AA) contribute to decoupled regulation of unwinding and synthesis and create inefficient replisomes.

While the apparent leading strand replication rates were significantly affected with the constricted DnaB mutants, K180A and RR328/9AA, we could not completely disregard that decreased loading efficiencies, stabilities, or perturbed interactions with accessory proteins played a role in these reduced leading strand synthesis rates. His-tagged DnaB pull down assays showed that these mutants still interacted with τ-CLC (Supplementary Figure S4), but to verify that DnaC or τ-CLC are able to stimulate unwinding of K180A and RR328/9AA, we performed a fluorescent DNA unwinding assay (18,21). Both K180A and RR328/9AA were stimulated by DnaC or τ-CLC to a similar magnitude as that of WT DnaB (Figure 2E); however, we did notice a 5–10 min lag in unwinding for those mutants. To confirm that DnaC or τ-CLC stimulated K180A unwinding similar to WT, we also performed this assay at higher enzyme concentrations (Supplementary Figure S5), which confirmed a similar lag phase and significant stimulation of unwinding. Therefore, the TFII leading strand reactions utilized DnaC and extended preincubation times to allow for maximal loading of all DnaBs prior to initiation (Figure 2A). Interestingly, neither DnaC nor τ-CLC were able to induce significant hexamer dilation of K180A or RR328/9AA to traverse over duplex DNA in a fluorescent translocation assay (Figure. 2F). Therefore, both DnaB mutants, K180A or RR328/9AA, can be loaded onto DNA, have stimulated unwinding through interacting proteins but have lost the ability to effectively transition to a dilated state.

RecA filamentation is important for fast, efficient growth when helicase regulation is impaired

As DnaB mutations, K180A and RR328/9AA, decoupled unwinding and synthesis in an in vitro assay, and the identical genomic mutations (Figure 1) decreased cellular fitness and caused gross chromosomal instabilities in vivo (21), we sought to better understand how these dnaB:mut strains survive in spite of impaired helicase regulation and decoupling by investigating the ssDNA processor, RecA. The two constricted DnaB mutant strains MSB4 (dnaB:K180A) and MSB5 (dnaB:RR328/9AA), along with the parental strain HME63 (dnaB:WT), were transformed with a plasmid containing an inducible expression system for the dominant mutant allele RecA56 (pEM-RecA56) (44) (see Supplementary Tables S1 and S2 for a full list of strains and plasmids). RecA56 readily incorporates itself into nascent RecA presynaptic filaments on ssDNA without affecting wild-type ATP hydrolysis (60). However, RecA56 expressing strains do disrupt stable RecA filaments needed for homologous recombination, reduce LexA cleavage to limit SOS induction, and are effectively outcompeted by SSB for binding for ssDNA (61,62).

Mass doubling times for all strains were recorded over the course of 20 h by measuring the OD₆₀₀ in Miller LB media (Supplementary Figure S6) as performed previously (21). The tetracycline analog anhydrotetracycline (aTc) was used to induce RecA56 expression in the test strains at functional doses that had little to no effect when added to control strains, HME63 and JW2669 (ΔrecA) (49), that lacked the RecA56 plasmid (Figure 3A, lanes 1 & 2). Adding pEM-RecA56 to HME63 also showed no significant change in average maximal specific growth rate (), 0.28 ± 0.01 h⁻¹ and 0.29 ± 0.01 h⁻¹ without and with aTc, respectively (Figure 3A, lane 3). However, RecA56 expression significantly decreased of both dnaB:K180A and dnaB:RR328/9AA (Figure 3A, lanes 4 and 5). dnaB:K180A had the lowest of all the strains, increasing in density at a lower rate than even the RecA knockout, JW2669 (Supplemental Figure S6F). During log phase, dnaB:K180A grew with a of 0.25 ± 0.01 hr⁻¹ in the absence of aTc and was reduced by almost 16% to 0.21 ± 0.01 h⁻¹, a decrease of 0.04 ± 0.00 h⁻¹ when RecA56 was expressed with addition of aTc. Similarly, despite dnaB:RR328/9AA demonstrating more rapid growth with a of 0.48 ± 0.01 h⁻¹ under normal conditions, expression of RecA56 reduced the growth by ∼13% to 0.42 ± 0.01 h⁻¹, a decrease of 0.06 ± 0.00 h⁻¹.

Figure 3.

Constricted dnaB mutants have a significant reduction in growth rate and mutagenesis when RecA filamentation is disrupted. (A) The difference in the mean of the maximal specific growth rates () from RecA56 expression was measured for each strain and condition by taking the difference of the calculated maximum growth rates ± RecA56 expression from Supplemental Figure S6. Data is from a minimum of three trials, and individual data is presented with closed circles. (B) The mutagenicity of each strain (± pEM-RecA56 and ± aTc, as indicated) was monitored using a rifampicin resistant assay to quantify number of rif^R colonies after 24-hour growth in Miller LB. Average number of resistant colonies per 1 million CFU’s are shown. Data is from at least two trials of three technical replicates each (n ≥ 6) with mean values listed and is presented with closed or open circles indicating functional RecA filamentation or not, respectively. Hashed bars represent addition of aTc to induce RecA56 and inactivate RecA filaments. Error bars represent ± SD; Black bars indicate statistically significant differences, where P-values are indicated and represented by **P < 0.01 or ****P < 0.0001, from an unpaired two-sided t-test. n.s. is not significant.

The high mutagenesis of dnaB mutant strains is RecA-dependent

Previously, we showed that interfering with helicase regulation led to poor competitiveness in mixed cultures and a dramatic increase in mutational frequencies (21). From these initial investigations, it was unclear whether the increased mutagenesis was from SOS induction that amplified the concentration of low fidelity TLS polymerases, or some other aspect of helicase dysregulation, perhaps altering the fidelity of Pol III or inducing RecA-independent recombination from excessive ssDNA. To determine if the high level of mutagenesis seen in dnaB:K180A and dnaB:RR328/9AA was mediated by RecA filamentation, we tested the strains’ abilities to develop resistance to rifampicin with and without expression of RecA56 to disrupt RecA-induced SOS that may occur from decoupling (Figure 3B). Briefly, cells grown in aTc for RecA56 induction were exposed to rifampicin, and the number of resistant colonies (rif^R) that arose determined the mutational frequencies (63–65). dnaB:RR328/9AA had the greatest number of mutation events under normal conditions, with 8.4 ± 0.5 per million (10⁶) colonies but then drastically dropped to 0.4 ± 0.2 when RecA filamentation was impaired by RecA56 expression (Figure 3B, blue, solid versus hashed). dnaB:K180A showed a similar dramatic decrease in mutation frequency when aTc was added, from 4.1 ± 0.3 mutation events per 10⁶ viable cells to 0.7 ± 0.2, indicating a clear reliance on RecA-dependent pathways for mutability (Figure 3B, green, solid vs. hashed). While the control strain HME63 showed no difference in mutation rate without the RecA56 expression system (Figure 3B, black, solid vs. hashed), inhibition of RecA by the addition of aTc (purple, solid versus hashed) caused a reduction in mutagenesis from 0.9 ± 0.4 to 0.3 ± 0.1. Even in the absence of endogenous stress, RecA allows low level mutagenesis important for environmental adaptation.

Interestingly, dnaB:RR328/9AA not only had the largest drop in mutational frequency with RecA56 expression but was also reduced by the greatest percentage, dropping by 8.0 events per 10⁶ (95%), followed by dnaB:K180A which dropped by 3.4 (82%), and the parental strain dropped by 0.6 mutation events (69%) (Figure 3B). Direct comparison of the strains in the absence of aTc showed that the presence of the RecA56 expression system alone did not affect mutation rates, as HME63 ± pEM-RecA56 did not significantly change in mutagenesis frequency (Supplementary Figure S7A, lanes 3 versus 4, solid grey or purple), and the average number of mutations for dnaB:K180A and dnaB:RR328/9AA are consistent with previously reported values (21). Induction of RecA56 to disrupt RecA filamentation resulted in all strains having mutation frequencies less than 1 event per 10⁶ cells (Supplementary Figure S7B) showing a clear dependence on RecA-associated pathways for mutation.

RecA56 expression sensitizes decoupled replisomes to exogenous damage

Because we determined that the dysregulated helicase strains (dnaB:K180A and dnaB:RR328/9AA) have high levels of mutagenesis mediated by RecA, we wanted to further explore how they responded to different types of exogenous damage, and whether RecA plays a role in genotoxin evasion and survival. We selected two genotoxins to test in minimum inhibitory concentration (MIC) growth assays that elicit different DDT mechanisms: mitomycin C (MMC), which causes large DNA adducts and interstrand crosslinks and methyl methanesulfonate (MMS), which methylates purine bases to cause mispairing. Although we also tried hydroxyurea (HU), which inhibits ribonucleotide reductase causing transient polymerase stalling from depleted nucleotide pools and also induces direct DNA damage through toxic breakdown products (66,67), the strain growth measurements were inconsistent in this assay as caused by delayed cell death that resulted in cell aggregates, impairing absorbance readings.

All cultures were grown overnight in the presence of aTc to ensure strong RecA56 expression in plasmid carrying strains prior to genotoxin exposure. Overnight cultures were normalized for equal cell counts across strains and conditions and diluted ∼100-fold into individual wells of a 96-well plate containing increasing concentrations of genotoxin and grown in an incubator/shaker at 32°C for at least 24 h prior to taking OD₆₀₀ readings (Figure 4A, B). Interestingly, the dnaB:K180A strain alone had a significantly higher MIC value (about 2-fold) compared to the parental strain, HME63, or dnaB:RR328/9AA in the absence of RecA56 expression (Figure 4C, D, and Supplementary Figure S8A–D). The MIC values for HME63 or dnaB:K180A for MMC or MMS also did not significantly decrease upon expression of RecA56 (Figure 4G, H, lanes 1 versus 2 and lanes 3 versus 4). Conversely, dnaB:RR328/9AA showed a significant reduction in the MIC values (about 2-fold) for both MMC and MMS (Figure 4E,F & Supplementary Figure S8C, D) upon induction of RecA56 (Figure 4G, H, lanes 5 versus 6), suggesting the RecA filamentation is particularly important in mitigating survival in this strain. The ΔrecA knockout strain, JW2669, also showed some sensitivity to MMC compared to the parental MG1655 strain (Figure 4G, lanes 7 versus 8) but still had a significantly greater MIC values than dnaB:RR328/9AA+ RecA56 (Figure 4G, H, lanes 6 versus 7) for both MMC and MMS, indicating that decoupling had a more significant effect on survival for dnaB:RR328/9AA. Overall, the sensitivities of the dnaB:mut strains on their own or with induction of RecA56 are significant but also somewhat modest in magnitude compared to other restart or repair gene deficiencies (68–70).

Figure 4.

Functional RecA activity is important for genotoxin evasion and survival when replisomes are decoupled. Representative strain growth in 96 well plates shaking at 32°C after 24 hours of incubation in increasing concentrations of (A) MMC or (B) MMS. The growth of (C, D) dnaB:K180A (green) or (E, F) dnaB:RR328/9AA (blue) strains minus (•, solid line fit) and plus (○, dashed line fit) pEM-RecA56 was plotted and fit to Equations (2) and (3) to determine MIC for MMC or MMS, respectively. Quantification of the MIC values for all strains exposed to (G) MMC or (H) MMS. N ≥ 4 trials with individual data points represented as closed (•, minus pEM-RecA56) or open (○, plus pEM-RecA56) circles. Error bars represent ± SE. Black bars indicate statistically significant differences with P-values indicated and represented by *P < 0.05 or **P < 0.01 from an unpaired two-sided t-test. n.s. is not significant.

To better visualize the growth at genotoxic concentrations that were just above the relative IC₅₀ values, we serial diluted and spotted (left to right) strains onto Miller LB-agar plates containing both aTc and the indicated genotoxin and imaged after 24 hours (Supplementary Figure S8E and F). In particular, the log-fold survival of dnaB:RR328/9AA with exposure to 1.2 μg/ml MMC was significantly impaired upon expression of RecA56 (Supplementary Figure S8E, rows 5 versus 6). Similar sensitivities for dnaB:RR328/9AA+ RecA56 at 2 mM MMS were also observed (Supplementary Figure S8F, rows 5 versus 6). Interestingly, dnaB:K180A also appeared to show some sensitivity to 2 mM MMS when RecA56 was induced (Supplementary Figure S8E-F, rows 3 versus 4) that was not quantified in the MIC plate reader assay (Figure 4H). Therefore, 2 mM MMS appears to allow for differences in log growth to be visualized, unlike higher concentrations that more rapidly kill all strains.

RecA activity is critical for preventing DNA breaks caused by dysregulated unwinding

Because RecA activity was influential for growth, mutagenesis, and genotoxin survival in the dnaB mutant strains, we investigated the importance of functional RecA filamentation activity on the frequency of DSBs when DNA unwinding regulation is impaired. dnaB:K180A, dnaB:RR328/9AA, and control strains (HME63 and MG1655), were grown in Miller LB ± aTc until mid-exponential phase, fixed, and permeabilized. Terminal BrdU Nick End Labeling (TUNEL) was used to label free 3′-OH ends for detection of both ss- and dsDNA breaks (Supplementary Figure S9A). Cellular DNA was stained with 4′,6-diamidino-2-phenylindole (DAPI, microscopy) or Sytox Green (SG, fluorescence activated cell sorting [FACS]). As positive controls, HME63 and MG1655 were exposed to MMC for 45 min prior to harvesting (Supplementary Figure S10). Fixed BrdU-labeled cultures were imaged for TUNEL DNA breaks density by fluorescence microscopy and FACS (Figure 5).

Figure 5.

RecA filamentation prevents DSBs from decoupling in dnaB mutant strains. (A) Log phase cells grown in ±aTc for RecA56 induction were stained with DAPI (blue) and BrdU (pink) for in situ detection of DNA nicks and breaks by TUNEL. Labeled cells containing the pEM-RecA56 plasmid were imaged by epifluorescence microscopy and images shown are representative of the observed population. Images of control strains exposed to the genotoxin MMC can be found in Supplementary Figure S10. The cell area (px²) was quantified from (B) microscopy for HME63 (grey) and dnaB:RR328/9AA (blue) for more than 50 cells and presented in violin plots, where the solid black bar represents the median and the dashed lines are quartiles in the absence (solid fill) and presence (open fill) of aTc. (C) FACS quantification of cellular DNA content using Sytox Green autofluorescence units (AFU) in the absence (solid fill) and presence (hashed fill) of aTc from n = 10 000 events, where error bars represent ± SE. Representative histograms and gating controls can be found in Supplementary Figure S11. The intensity of TUNEL BrdU labeling was measured and quantified by (D) microscopy as BrdU foci/cell density to adjust for cell size or by (E) FACS for the gated population from n = 10000 events reported in AFU from the same exponential growth cultures. Error bars represent ± SE. Black bars indicate statistically significant differences with P-values indicated and represented by *P < 0.05, ***P < 0.001 and ****P < 0.0001 from a Mann–Whitney two-sided U-test for the microscopy violin plots and an unpaired two-sided t-test for the FACS data. n.s. is not significant.

HME63 showed mostly normal, short (∼2 μm) cells (Figure 5A, left) with the occasional filament, that was unchanged with addition of aTc (Figure 5B, grey) and contained minimal BrdU foci (Figure 5D, grey). However, both dnaB:K180A and dnaB:RR328/9AA had a high density of long filamented cells under normal growth conditions (Figure 5A, middle and right), as reported previously (21), but these mutant strains appeared to shorten and show more BrdU foci when aTc was added. Interestingly, quantification of dnaB:RR328/9AA microscopy images showed a significant ∼1.7-fold reduction in cell area when aTc was added (Figure 5B). To measure cell size more efficiently for all strains and conditions, the DNA content of fixed and permeabilized cells was stained with Sytox green, quantified, and reported in auto fluorescence units (AFU) by FACS (Figure 5C and Supplementary Figure S11). Again, for the control strain, HME63, addition of aTc (hashed bars) in the absence of the RecA56 plasmid showed no difference in DNA content per cell (Figure 5C, lanes 1 and 2); however, addition of MMC (pink, lane 3) significantly increased the Sytox Green AFU signal ∼2-fold. MMC is known to induce DSBs, which could increase the DNA content by generating cellular stress from complicated DNA replication processes and inhibited cell division by the SOS protein, SulA, or asymmetrical segregation of Holliday junction entangled chromosomes (71–75). Interestingly, the trends observed by microscopy quantification for HME63 and dnaB:RR328/9AA cell area were confirmed by FACS and showed a significant 2-fold (purple, lanes 4 versus 5), 1.6-fold (green, lanes 6 versus 7) and 1.8-fold (blue, lanes 8 versus 9) reduction in Sytox green AFU upon aTc addition for HME63, dnaB:K180A and dnaB:RR328/9AA, respectively.

As a decrease in cellular size was readily observed with addition of aTc, the BrdU TUNEL foci density was quantified for HME63 and dnaB:RR328/9AA from microscopy images (Figure 5D). First, in the absence of aTc, there was a significant increase in BrdU density per cell for dnaB:RR328/9AA compared to the control strain, HME63, indicating that DNA breaks are more prevalent in this strain (grey versus blue, solid fill). The BrdU TUNEL foci density measured by microscopy was significantly increased with addition of aTc for dnaB:RR328/9AA, indicating that a reduction in RecA filamentation affords more DNA breaks (blue, solid versus open fill). To better quantify the BrdU TUNEL signal in all strains and conditions, FACS was once again utilized (Figure 5E). Consistent with the microscopy quantification, the addition of aTc (hashed bars) had no effect on BrdU AFU for the control strain regardless of RecA56 expression, HME63 (grey, lanes 1 versus 2; or purple, lanes 4 versus 5). However, addition of MMC as a positive control for DNA breaks increased the signal > 2.7-fold to 5.6 × 10⁷ AFU (pink, lanes 1 versus 3). dnaB:RR328/9AA had a small but significant increase in TUNEL BrdU AFU (1.4-fold) in the absence of aTc compared to HME63 (grey versus blue, lanes 1 versus 8) consistent with the microscopy quantification (Figure 5D). Interestingly, the AFU signal increased dramatically (∼2-fold for both dnaB:K180A and dnaB:RR328/9AA) with addition of aTc to 3.6 × 10⁷ (green, lanes 6 versus 7) and 3.9 × 10⁷ (blue, lanes 8 versus 9), respectively, near that of the + MMC positive control (pink, lane 3).

dnaB:K180A and dnaB:RR328/9AA induce abundant daughter strand gaps

To determine if dysregulated helicase regulation led to the production of excess ssDNA gaps in vivo stimulating the protective activities of RecA, we utilized Klenow, an exonuclease deficient variant of Pol I, in a Pol I dUTP Gap filling (PLUG) assay (Supplementary Figure S9B) (51–53). Briefly, the same cells that were probed with TUNEL (Figure 5) were split and incubated at room temperature for 30 min with Klenow and a nucleotide mix containing BrdU to label any ssDNA gaps extended from a 3′-OH primer (Figure 6). To limit variables, PLUG-treated cells were analyzed by microscopy and FACS identically to TUNEL samples. The amount of PLUG signal detected in the control strains, HME63, MG1655, and JW21669, was minimal and did not appear to change with addition of aTc (Figure 6A and Supplementary Figure S10). However, the addition of MMC to HME63 caused a visual increase in PLUG signal, indicating a substantial amount of genomic stress (Supplementary Figure S10). Qualitatively, PLUG foci were generally brighter than TUNEL, likely due to multiple BrdU nucleotides being incorporated into a single ssDNA gap. Similar to TUNEL, PLUG foci and cell area were measured and used to calculate and compare PLUG density for HME63 and dnaB:RR328/9AA (Figure 6B). The PLUG foci density significantly increased for dnaB:RR328/9AA compared to HME63 in the absence of aTc (open fill, blue or grey). However, the PLUG foci density did not significantly change for HME63 or dnaB:RR328/9AA when aTc was added, even though the median increased (solid versus open fill, blue or grey), possibly from the limited number of quantified cells.

Figure 6.

Dysregulated helicase activity induces significant ssDNA gaps. (A) Log phase cells grown in ±aTc for RecA56 induction were stained with DAPI (blue) and BrdU (pink) for in situ detection of DNA nicks and breaks by PLUG. Labeled cells containing the pEM-RecA56 plasmid were imaged by epifluorescence microscopy and images shown are representative of the observed population. Images of control strains exposed to the genotoxin MMC can be found in Supplementary Figure S10. The intensity of PLUG BrdU labeling was measured and quantified by (B) microscopy as BrdU density to adjust for cell size, where the black bar represents the median in the absence (solid fill) and presence (open fill) of aTc or by (C) flow cytometry (FACS) for the gated population from n = 10 000 events reported in auto fluorescence units (AFU) from the same exponential growth cultures. Error bars represent ± SE. Black bars indicate statistically significant differences with P-values indicated and represented by *P < 0.05, ***P < 0.001 and ****P < 0.0001 from a Mann–Whitney two-sided U-test for the microscopy violin plots and an unpaired two-sided t-test for the FACS data. n.s. is not significant.

To better quantify PLUG signal intensity across multiple strains and conditions for larger sample sizes, the BrdU PLUG AFU signal intensity was quantified per cell by FACS (Figure 6C and Supplementary Figure S11A–E). First, the PLUG intensity measured for dnaB:K180A (4.5 × 10⁷ AFU) and dnaB:RR328/9AA (6.2 × 10⁷ AFU) was 1.9- and 2.6-fold greater than for HME63 (2.4 × 10⁷ AFU) respectively, indicating that an excessive number of single strand gaps are created when the helicase is dysregulated. The single strand gaps measured by PLUG AFU (Figure 6C) are ∼2-fold more intense in all cases compared to DNA breaks measured by TUNEL AFU (Figure 5E) when DnaB is dysregulated. Although there was no difference in HME63 ± aTc (Figure 6C, grey, lanes 1 versus 2; or purple, lanes 4 versus 5 and Supplementary Figure S11F–H), addition of MMC as a positive control for DNA breaks increased the signal >2.2-fold to 5.2 × 10⁷ AFU (Figure 6C, pink, lanes 3 versus 1), similar to that seen for TUNEL (Figure 5E, lane 3). The addition of aTc to dnaB:K180A or dnaB:RR328/9AA significantly increased the PLUG AFU further 1.3-fold for both (5.7 and 7.8 × 10⁷ AFU, respectively), (Figure 6C, green, lanes 6 versus 7; blue lanes 8 versus 9 and Supplementary Figure S11I-J). This result combined with the decrease in cellular area with RecA filament disruption (Figure 5C) allows us to conclude that dysregulation of replisome coupling produces ssDNA gaps that are protected and mediated by RecA. However, when RecA filamentation is impaired, these ssDNA gaps are converted to DNA breaks that are likely processed through a separate DDT pathway.

Discussion

In a previous study, we described the first in vivo investigation of genomic dnaB SEW-disrupting mutations, reporting that MSB4 (dnaB:K180A) and MSB5 (dnaB:RR328/9AA) stabilized a constricted DnaB conformer and had significant negative impact on replication fidelity and cellular efficacy (21). Both strains demonstrated a filamented cellular stress phenotype and had high incidence of mutations, implying a dependence on RecA for survival. Here, these constricted and fast unwinding mutants of DnaB (K180A and RR328/9AA) both drastically reduced productive leading strand synthesis products, implying that coupling between unwinding and synthesis must be maintained for a productive replisome. To better understand helicase regulation in vivo and the consequences of dysregulation, we investigated the role of RecA filamentation on strain survival and genomic stability using a RecA56 inducible system to inhibit stable RecA presynaptic filamentation. The high mutagenicity of dnaB mutant strains (21) was found to be dependent on effective RecA polymerization and is therefore the result of increased DDT, likely from SOS protein induction, including mutagenic Pol V. High levels of endogenous damage in dnaB mutant strains that presented as DNA nicks and gaps were tolerated by RecA filamentation for growth and efficacy, albeit with cellular stress and filamentation phenotypes consistent with SOS. When exposed to exogenous damage, dysregulated unwinding in these strains typically increased genotoxin sensitivity and RecA filamentation was moderately important in mediating survival. Interestingly, dysregulated helicase activity on its own produced excess ssDNA in vivo as detected by PLUG, and RecA-mediated HR is partially responsible for the cellular filamentation and asymmetric chromosome segregation observed in these strains previously (21). Importantly, RecA filamentation is utilized to modulate ssDNA gaps created by dysregulated helicase activity, and inhibition of RecA filamentation in these strains converts ssDNA gaps to breaks allowing for alternative DDT mechanisms and cellular division progression (Figure 7).

Figure 7.

Helicase-polymerase uncoupling induces ssDNA gaps that stimulate RecA filamentation to protect and repair the genome. Uncoupling of unwinding and synthesis from helicase dysregulation or other persistent blocks to replication lead to an increased frequency of ssDNA gaps that are modulated by RecA filamentation and downstream SOS and DDT processes. However, when RecA56 is induced, ssDNA gaps are unprotected, become labile, and are converted to more single and double strand breaks.

Although dnaB:RR328/9AA had a significantly higher mutation frequency than dnaB:K180A, both exhibited increased RecA-dependent mutagenesis relative to the parental strain, HME63. When RecA56 was induced to impair RecA filament formation, both dnaB mutants demonstrated fewer than 1 in 10⁶ mutation events, equivalent to parental strains, revealing that the increased mutation frequency of dnaB mutant strains can be attributed to productive RecA filamentation on these ssDNA gaps inducing SOS. Polymerase exchange at the fork when higher concentrations of mutagenic TLS Pols are induced by SOS likely allows for frequent and continuous replication by Pol V within large stretches of ssDNA gaps generated by dnaB:K180A and dnaB:RR328/9AA (76–79). Since dysregulated unwinding stimulates RecA polymerization on ssDNA gaps, we were interested in whether the growth of dnaB mutant strains during RecA inhibition impacted strain growth. RecA56 expression significantly reduced the growth rate of dnaB mutant strains. Exposure with targeted exogenous damage showed that dnaB:RR328/9AA with its greater SOS induction (21) relied more on RecA filamentation for survival and that expression of RecA56 reduced the MIC values for both MMC and MMS. For unknown reasons, dnaB:K180A had a greater resistance to both MMC and MMS and disruption of RecA filamentation only appeared to have a moderate effect on the measured MIC value. However, when dnaB:K180A was serial diluted on plates containing genotoxins near the IC₅₀ values, the sensitization in growth with RecA56 expression was more apparent. Therefore, dysregulated dnaB mutants depend on RecA activity to mitigate the impact of rapid unwinding in vivo that produce significant ssDNA stretches that are labile and may be processed into DSBs.

Previously we reported that dnaB:K180A and dnaB:RR328/9AA had increased chromosome complexities that included odd numbers and broad distributions and complementary increases in the ori:ter ratios, suggesting that fork stalling from decoupling was impacting chromosome segregation (21). Quantification of the large increase in DNA breaks and total DNA intensity in dnaB mutant strains relative to the parental shows that helicase-affected DNA damage and fork stalling forces cell filamentation and asymmetrical separation of chromosomes. With wild-type RecA ability, the majority of measured TUNEL foci in the dnaB:mut strains correlate directly with DSBs and are therefore substrates for HR repair, which are known to create temporarily entangled chromosomes through RecA-mediated strand invasion (74,80). While inhibition of RecA polymerization through expression of RecA56 reduced the DNA content and cellular filamentation of dnaB mutant cells, they still maintained higher DNA content than the control strains, indicating that not all chromosome entanglement and asymmetric segregation is from active Holliday junctions or that not all Holliday junctions are fully resolved. Significant stretches of ssDNA produced from leading strand decoupling or from extended gaps on the lagging strand resulting from an inability of DnaG to interact with the constricted DnaB state (18) were effectively measured by PLUG and are significantly increased as well as converted to DSBs in the absence of productive RecA filamentation. ssDNA secondary structures, DNA adducts, and other lesions are likely also involved, delaying fork progression to the terminus and sister chromosome separation.

This dependence on RecA activity to maintain fast and efficient growth when helicase activity is dysregulated or impeded indicates that measurably more endogenous damage and replication impediments are constantly challenging faithful and efficient replisome progression. Both DnaB K180A and RR328/9AA helicase mutants failed to produce long leading strand products in vitro, suggesting that even though these mutants have faster unwinding rates, they cause significant decoupling, compromising the replisome. This observed decoupling in vitro is further supported by excessive DNA breaks and even more ssDNA gaps detected endogenously by TUNEL and PLUG when helicase regulation is disrupted in vivo (i.e. dnaB:K180A and dnaB:RR328/9AA). Direct comparison of TUNEL and PLUG between dnaB:K180A and dnaB:RR328/9AA shows that despite the high incidence of DNA breaks, DNA gaps are the primary product of dysregulated unwinding. When faced with rapid ssDNA generation, RecA-associated repair is critical for cell survival, efficient replication, and damage mitigation in E. coli (81). It is also likely that other restart pathways (RecFOR, PriAB, PriABC, or RecBCD) (82) and both TLS Pols, IV and V (83) differentially contribute to cell survival when the replisome is decoupled and abundant ssDNA gaps are present. Even though complex and diverse fork interactions maintain a dynamic replisome, helicase regulation is necessary for optimal genome duplication to suppress DNA damage and genomic mutations.

Hereinto, targeted regulation-deficient helicase mutants induce replisome decoupling in vitro and in vivo, but the biological consequences are diminished by RecA stabilization and recombination. In an increasingly evident dynamic replisome, RecA-mediated DNA processing in E. coli compensates for non-ideal DNA biproducts of replication, including those caused by uncoupled replisomes. However, the fitness cost of managing dysregulated unwinding demonstrates the critical nature of helicase regulation and replisome coupling for efficacious and stable growth.

Notes

Present address: Megan S. Behrmann, National Cancer Institute, Bethesda, MD 20892, USA.

Present address: Himasha M. Perera, Novartis Institutes for Biomedical Research, Cambridge, MA 02139, USA.

Data availability

The data underlying this article are available in the article and in its online Supplementary material.

Supplemental data

Supplementary Data are available at NAR Online.

Funding

NSF MCB [NSF 2105167 to M.A.T.]; Baylor University. Funding for open access charge: NSF.

Conflicts of interest statement

None declared.