Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Genetic analysis of epistasis between nucleotide excision repair and homologous recombination in the recovery of persisters after fluoroquinolone treatment

  • Nashaly Soto-Echevarria,

    Roles Conceptualization, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing

    Affiliation Department of Molecular Biology, Princeton University, Princeton, New Jersey, United States of America

    ⨯
  • Annabel S. Lemma,

    Roles Conceptualization, Investigation, Methodology, Writing – original draft, Writing – review & editing

    Affiliation Department of Chemical and Biological Engineering, Princeton University, Princeton, New Jersey, United States of America

    ⨯
  • Mark P. Brynildsen

    Roles Conceptualization, Funding acquisition, Methodology, Writing – original draft, Writing – review & editing

    mbrynild@princeton.edu

    Affiliations Department of Molecular Biology, Princeton University, Princeton, New Jersey, United States of America, Department of Chemical and Biological Engineering, Princeton University, Princeton, New Jersey, United States of America, Omenn-Darling Bioengineering Institute, Princeton University, Princeton, New Jersey, United States of America

    ⨯

Abstract

Stationary-phase cultures contain high abundances of persisters, which are bacterial cells that are hyper-tolerant to antibiotics due to phenotypic reasons. Fluoroquinolones (FQs) are one of the most effective antibiotic classes for the treatment of non-growing bacteria, such as those found in stationary-phase cultures, and persisters in those populations have been found to survive FQ treatment by repairing the DNA damage caused by the antibiotic rather than by avoiding damage through inactivity. Interestingly, previous work demonstrated that transient growth inhibition after the conclusion of FQ treatment significantly increased persister levels from stationary-phase Escherichia coli populations if recA, a mediator of homologous recombination (HR), or uvrD, a component of nucleotide excision repair (NER), were present. Here, we sought to identify additional epistatic interaction partners in that persister recovery network by deleting DNA repair enzymes known to interact with RecA or UvrD. We found that uvrA, uvrB, and mfd also epistatically interact with recA in FQ persister recovery, whereas recB and recC are additional epistatic interaction partners of uvrD. While these results indicated that HR and NER can contribute to persister recovery, different combinations of genetic mutants suggested that the phenomenon is more nuanced than compensation for loss of one repair pathway by another. Specifically, loss of recA had farther reaching epistatic consequences than loss of other HR genes, and loss of uvrD had a greater impact on the network than any other NER gene. Delving deeper into the roles of RecA and UvrD, we found that the recombination function of RecA, via RecA(N304D), and helicase function of UvrD, via UvrD(R284A), were required for their participation in FQ persister recovery. Collectively, the data presented here deepen understanding of the roles of HR and NER machinery in the recovery FQ persisters and further emphasize the important roles of RecA and UvrD in this phenomenon.

Author summary

Persisters are rare phenotypic variants that survive antibiotic treatments that kill the vast majority of cells in cultures, and evidence has accumulated to support their role in chronic and recurrent infections. Persisters are abundant in populations where nutrients have been exhausted, such as stationary-phase cultures, and FQs are one of the few antibiotic classes that remain efficacious against such populations. Previous work found that transient growth inhibition after FQ treatment increased the abundance of persisters from stationary-phase E. coli populations by providing time for bacteria to repair FQ-induced DNA damage before resuming growth. Further, an epistasis between recA (a mediator of homologous recombination) and uvrD (a participant in nucleotide excision repair) was uncovered in the phenomenon. Here we adopted a genetic approach to identify additional members of this epistatic network by focusing on known interaction partners of RecA and UvrD. We found that uvrA, uvrB, and mfd also epistatically interact with recA, whereas recB and recC are additional epistatic interaction partners of uvrD. Additional analyses revealed the central roles of recA and uvrD in this epistatic network that has the capacity to reduce the efficacy of FQs by up to an order of magnitude.

Citation: Soto-Echevarria N, Lemma AS, Brynildsen MP (2026) Genetic analysis of epistasis between nucleotide excision repair and homologous recombination in the recovery of persisters after fluoroquinolone treatment. PLoS Genet 22(9): e1012305. https://doi.org/10.1371/journal.pgen.1012305

Editor: Morten Kjos, Norwegian University of Life Sciences: Norges miljo- og biovitenskapelige universitet, NORWAY

Received: January 31, 2026; Accepted: August 25, 2026; Published: September 21, 2026

Copyright: © 2026 Soto-Echevarria et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The raw data that support the findings of this study have been deposited in figshare and can be accessed at the following DOI: https://doi.org/10.6084/m9.figshare.33411667.

Funding: This work was supported by the National Institute of Allergy and Infectious Diseases (NIAID) of the National Institutes of Health (NIH) (https://www.niaid.nih.gov). Award number is R21AI182745 and it was awarded to M.P.B. This work was also supported by the National Institute of General Medical Sciences (NIGMS) of the National Institutes of Health (NIH) (https://www.nigms.nih.gov). The award number is T32GM007388 and it was awarded to N.S.E. The content is solely the responsibility of the authors, does not necessarily represent the official views of the funding agencies, and the funders had no role in the study design, analysis, data collection, manuscript preparation, or the decision to publish.

Competing interests: The authors have declared that no competing interests exist.

1. Introduction

Persisters survive antibiotics by occupying transient phenotypic states that die more slowly during antibiotic treatments than the majority of their clonal neighbors [1–3]. After antibiotic exposure concludes, persisters exit those tolerant phenotypic states and give rise to new populations that are as susceptible to antibiotics as the original cultures [1–4]. Persisters have been detected in a wide variety of bacterial species, are thought to contribute to infection relapse, and have been shown to facilitate the development of antibiotic-resistant mutants [4–11]. Overall, increased understanding of persister survival tactics has the potential to improve outcomes for relapsing infections [7,12].

Fluoroquinolones (FQs) are antibiotics that target type II topoisomerases (e.g., DNA gyrase, topoisomerase IV) that depending on the concentration used produce single and/or double strand breaks (SSBs and DSBs) in DNA [13–17]. Previous work has shown that persisters survive FQ treatment by repairing DNA damage that ensued from topoisomerase corruption, rather than by avoiding FQ-induced DNA damage through topoisomerase inactivity [18–21]. Two prominent DNA repair pathways used by FQ persisters are homologous recombination (HR) and nucleotide excision repair (NER) [10,19,21–26]. HR can repair DSBs via the RecBCD pathway and single strand lesions via the RecFOR pathway [27–29]. RecA is a key HR enzyme as both pathways load RecA onto nascent single-stranded DNA (ssDNA); for DSBs, RecBCD will help load RecA onto ssDNA generated after Chi site recognition [30], whereas for single strand gaps (SSGs), RecFOR binds to double-stranded DNA (dsDNA)-ssDNA junctions and facilitates RecA loading in a 5’ to 3’ direction [31]. Once loaded, ssDNA-bound RecA forms helical filaments that protect damaged DNA from further degradation, initiate the SOS response, and facilitate strand exchange [27–29]. NER can be divided into three sub-pathways which sense and fix DNA damage via the UvrABC machinery [32,33]. In global genomic repair (GGR), DNA damage is sensed through the UvrA2B2 complex by a change in helix rigidity; in Mfd-mediated transcription coupled repair (TCR), Mfd binds to a stalled RNA polymerase and displaces it by pushing it forward to expose the lesion for UvrA2B2 binding; and in UvrD-mediated TCR, UvrD binds to a stalled RNA polymerase but backtracks it to expose the damaged site for UvrA2B2 [32–37]. Once UvrA2B2 detects the lesion, UvrC is recruited to generate a dual incision 3’ and 5’ from the damaged nucleotides [32,33,35]. In addition to binding stalled RNA polymerase in UvrD-mediated TCR, UvrD plays a key role in all three sub-pathways by releasing UvrC and employing its helicase activity to remove the damaged oligonucleotides [32,33]. Interestingly, our recent work revealed an epistatic interaction between UvrD and RecA in the recovery of persisters from FQ treatment [23], which is the focus of this study.

A growing body of evidence has demonstrated the importance of post-treatment events on FQ persister survival [18,19,23,38]. For example, our group showed that recA and the SOS response were not needed by FQ persisters from stationary-phase populations until FQ treatment had ended and cultures were exposed to fresh nutrients [18]. We then found that the relative timing of DNA repair and growth resumption following FQ treatment was a critical parameter in defining persister levels of starving populations [19]. Later, we observed that inhibition of transcription or translation after FQ treatment had ended produced significant increases in persistence from stationary-phase populations [23]. That phenomenon, which we will call inhibition-assisted recovery, required recA or uvrD, because it was absent from the double deletion mutant, ΔuvrD ΔrecA, but present in the single deletion strains, ΔrecA and ΔuvrD [18]. Notably, the origin of the epistasis was not uncovered, which is a complex task given the large number of interaction partners both RecA and UvrD have [27,28,35,39–42].

In this study, we set out to identify RecA and UvrD interaction partners that participate in the epistasis in inhibition-assisted recovery of Escherichia coli FQ persisters. Our results revealed that uvrA, uvrB, mfd, recB, and recC are all members of the epistatic interaction network along with recA and uvrD, and that recA and uvrD exhibit the strongest degrees of epistasis with other genes. Further, analyses of loss of function and impaired function mutants of RecA and UvrD demonstrated that the recombination abilities of RecA and helicase activity of UvrD were critical to this epistatic phenomenon. Overall, the results presented elaborate on an epistasis observed between recA and uvrD to include additional genes involved in NER and HR DSB repair and provide deeper understanding of the functions of RecA and UvrD that contribute to inhibition-assisted recovery of FQ persisters.

2. Results

2.1. uvrA, uvrB, and mfd exhibit epistasis with recA in the recovery of FQ persisters

In previous work, translational inhibition with chloramphenicol (CAM) following FQ treatment increased survival in ΔrecA and ΔuvrD, but not in the ΔuvrD ΔrecA double mutant [23]. We sought to identify additional DNA repair genes that were part of this epistatic network. We began by screening DNA repair genes whose products were known to interact with UvrD for whether their knockouts when combined with ΔrecA would resemble ΔuvrD ΔrecA in FQ persister recovery assays. We considered genes involved in NER (uvrA, uvrB, uvrC, mfd) and mismatch repair (MMR) (mutS, mutH, mutL) due to the prominent role of UvrD in these repair pathways [32,33,39,43,44]. The screen suggested that none of the MMR genes could recapitulate the phenotype of uvrD in FQ persister recovery assays, whereas uvrA, uvrB, and mfd did (Fig 1). Specifically, the recoveries of ΔuvrA ΔrecA, ΔuvrB ΔrecA, and Δmfd ΔrecA resembled that of ΔuvrD ΔrecA, where survival with or without CAM were near equivalent (Fig 1). Interestingly, the phenotype of uvrC, the other NER gene that is involved in all three NER pathways (GGR, Mfd-mediated TCR, and UvrD-mediated TCR) [32,33], when knocked out in ΔrecA did not resemble that of ΔuvrD ΔrecA. E. coli encodes another excision nuclease named Cho that interacts with the NER machinery and has homology to UvrC [35]. We considered whether cho might be participating in FQ persister recovery and assayed Δcho ΔrecA and Δcho ΔuvrC ΔrecA. Neither Δcho ΔrecA nor Δcho ΔuvrC ΔrecA resembled ΔuvrD ΔrecA in FQ persister recovery assays; however, the triple mutant exhibited attenuated recovery (Fig 1).

thumbnail
Download:
Fig 1. Screen for epistatic interactions with recA in FQ persister recovery.

Stationary-phase cultures of strains harboring indicated mutations were treated with 5 µg/mL ofloxacin (OFL) and assayed for recovery by washing three times with PBS and plating on LB agar supplemented with 25 µg/mL CAM on top of 0.2 µm filters. After 12 h, filters were transferred to LB agar plates for 24 h. ΔuvrA ΔrecA, ΔuvrB ΔrecA, and Δmfd ΔrecA strains did not undergo inhibition-assisted recovery, whereas Δcho ΔuvrC ΔrecA exhibited an attenuated response. Strains in this screen contained kanamycin resistance cassettes from the knockout procedure (S1 Table). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. *: reflects statistically significant differences (p ≤ 0.05) between the indicated mutant and ΔrecA, but not ΔuvrD ΔrecA. †: reflects statistically significant differences (p ≤ 0.05) between the indicated mutant and both ΔrecA and ΔuvrD ΔrecA. #: reflects statistically significant difference between the indicated references strains.

https://doi.org/10.1371/journal.pgen.1012305.g001

With these screening results suggesting that loss of uvrA, uvrB, or mfd from ΔrecA eliminates its ability to undergo inhibition-assisted recovery, we advanced uvrA, uvrB, and mfd to further testing with time-course recovery assays. Importantly, time-course assays reveal the dynamics of recovery, which were different between ΔrecA and ΔuvrD (Fig 2, green-shaded background). For instance, defining the time to recover as a significant increase in persisters that exceeds 10-fold, it can be observed that the time to recover for wild-type (WT) was 2 h, for ΔuvrD was 4 h, and for ΔrecA was 8 h. Such data suggests that the timescale to restore culturability after FQ treatment depends to a greater extent on recA than uvrD under the conditions considered here.

thumbnail
Download:
Fig 2. Epistasis between recA and uvrD in FQ persister recovery occurs with post-treatment translational inhibition.

Stationary-phase cells were treated with 5 µg/mL OFL, CFU/mL were monitored, and survival fractions were calculated. At 5 h post-OFL treatment, cells were spotted on filters placed on top of LB agar supplemented with 25 µg/mL CAM. Samples were incubated at 37 °C for the times indicated before being transferred to LB agar without CAM for 24 h. Strains in these assays were cured of kanamycin resistance cassettes (S1 Table). WT, ΔrecA, and ΔuvrD exhibited significant increases in persister levels from post-FQ translational inhibition. However, the recovery period did not increase persistence in ΔuvrD ΔrecA. Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. *: indicates statistically significant (p ≤ 0.05) differences between time points and t = 0 h of the same samples.

https://doi.org/10.1371/journal.pgen.1012305.g002

In time-course recovery assays (Fig 3), ΔuvrA, ΔuvrB, and Δmfd showed significant increases in survival when cultures were exposed to CAM after FQ treatment compared to cultures that were not exposed to CAM (t = 0 h of recovery). That data showed that loss of uvrA, uvrB, or mfd on their own does not eliminate inhibition-assisted recovery. When ΔuvrA, ΔuvrB, or Δmfd were combined with ΔrecA, little difference in survival was observed for cultures with or without CAM, which demonstrated that loss of these NER genes in a strain devoid of recA eliminates inhibition-assisted recovery (Fig 3). These results recapitulate the phenotype of ΔuvrD where in isolation inhibition-assisted recovery of FQ persisters was observed, whereas in combination with ΔrecA recovery was not seen (Fig 2). Further, inspection of the dynamics revealed recovery that occurred within a 2 h timescale for ΔuvrA, ΔuvrB, or Δmfd, which resembled the timescales of WT and ΔuvrD. In addition, the lack of recovery in all three double deletion strains suggests that uvrA, uvrB, and mfd are involved in processes that enable recovery at times scales of 8 h and later, which is similar to uvrD.

thumbnail
Download:
Fig 3. ΔuvrA, ΔuvrB, and Δmfd recapitulate the phenotype of ΔuvrD in FQ persister recovery.

Stationary-phase cells were treated with 5 µg/mL OFL, CFU/mL were monitored, and survival fractions were calculated. At 5 h post-OFL treatment, cells were spotted on filters placed on top of LB agar supplemented with 25 µg/mL CAM. Samples were incubated at 37 °C for the times indicated before being transferred to LB agar without CAM for 24 h. Strains in these assays were cured of kanamycin resistance cassettes (S1 Table). ΔuvrA (A), ΔuvrB (B), Δmfd (C), and ΔrecA (A-C) exhibited significant increases in persister levels from post-FQ translational inhibition. However, the recovery period did not increase persistence in ΔuvrA ΔrecA (A), ΔuvrB ΔrecA (B), or Δmfd ΔrecA (C). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. *: indicates statistically significant (p ≤ 0.05) differences between time points and t = 0 h of the same samples.

https://doi.org/10.1371/journal.pgen.1012305.g003

To confirm that uvrA, uvrB, mfd, and uvrD were components of this epistatic persister recovery network, we performed complementation experiments with the double knockout mutants. Expression of recA or the missing NER gene from their native promoters on plasmids restored inhibition-assisted recovery in double knockout mutants, whereas empty vectors had little difference in survival (S1 and S2 Figs). Collectively, these data suggested that uvrA, uvrB, mfd, and uvrD exhibit epistasis with recA in the phenomenon of inhibition-assisted recovery of FQ persisters.

2.2. recB and recC exhibit epistasis with uvrD in the recovery of FQ persisters

To further identify DNA repair genes that are part of the epistatic network responsible for FQ persister recovery, we screened genes whose products were known to interact with RecA for epistasis with uvrD [27,45]. Specifically, we screened genes involved in recombinatorial repair (recB, recC, recD, recJ, recN, recF, recO, recR) with knockouts, and the SOS response with an uninducible allele of lexA (lexA3) [18,27–29,46]. Results suggested that ΔrecD, ΔrecJ, ΔrecF, ΔrecO, ΔrecR, and lexA3 cannot recapitulate the phenotype of ΔrecA in its epistasis with ΔuvrD in FQ persister recovery assays (Fig 4). However, both ΔuvrD ΔrecB and ΔuvrD ΔrecC failed to exhibit inhibition-assisted recovery and thus could recapitulate the phenotype of ΔuvrD ΔrecA, whereas ΔuvrD ΔrecN exhibited attenuated recovery (Fig 4). With these screening results suggesting that loss of recB or recC from ΔuvrD eliminates its ability to undergo inhibition-assisted recovery, we advanced recB and recC to further testing with time-course recovery assays.

thumbnail
Download:
Fig 4. Screen for epistatic interactions with uvrD in FQ persister recovery.

Stationary-phase cultures of strains harboring indicated mutations were treated with 5 µg/mL OFL and assayed for recovery by washing three times with PBS and plating on LB agar supplemented with 25 µg/mL CAM for 12 h on top of 0.2 µm filters. After 12 h, filters were transferred to LB agar plates for 24 h. ΔuvrD ΔrecB and ΔuvrD ΔrecC did not undergo inhibition-assisted recovery, whereas ΔuvrD ΔrecN exhibited an attenuated response. Strains in this screen contained kanamycin resistance cassettes from the knockout procedures (S1 Table). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. *: reflects statistically significant differences (p ≤ 0.05) between the indicated mutant and ΔuvrD, but not ΔuvrD ΔrecA. †: reflects statistically significant differences (p ≤ 0.05) between the indicated mutant and both ΔuvrD and ΔuvrD ΔrecA. #: reflects statistically significant difference between the indicated references strains.

https://doi.org/10.1371/journal.pgen.1012305.g004

In time course assays, both ΔrecB and ΔrecC displayed significant increases in survival when cultures were exposed to CAM after FQ treatment compared to cultures that were not exposed to CAM (Fig 5). That data showed that loss of recB or recC on their own does not eliminate inhibition-assisted recovery. However, when ΔrecB or ΔrecC were combined with ΔuvrD little difference in survival was observed for cultures with or without CAM (Fig 5). These results recapitulate the phenotype of ΔrecA where in isolation inhibition-assisted recovery of FQ persisters was observed, whereas in combination with ΔuvrD recovery was not seen (Fig 2). Analysis of the recovery dynamics revealed that the time to recover for ΔrecB and ΔrecC was 6 h (Fig 5), which more closely matched the dynamics of ΔrecA recovery compared to the dynamics of WT, ΔuvrA, ΔuvrB, ΔuvrD, and Δmfd. These results suggest that recB and recC are involved in processes that enable recovery at timescales of 4 h and earlier, which is similar to recA.

thumbnail
Download:
Fig 5. ΔrecB and ΔrecC recapitulate the phenotype of ΔrecA in FQ persister recovery.

Stationary-phase cells were treated with 5 µg/mL OFL, CFU/mL were monitored, and survival fractions were calculated. At 5 h post-OFL treatment, cells were spotted on filters placed on top of LB agar supplemented with 25 µg/mL CAM. Samples were incubated at 37 °C for the times indicated before being transferred to LB agar without CAM for 24 h. Strains in these assays were cured of kanamycin resistance cassettes (S1 Table). ΔrecB (A), ΔrecC (B), and ΔuvrD (A-B) exhibited significant increases in persister levels from post-FQ translational inhibition. However, the recovery period did not increase persistence in ΔuvrD ΔrecB (A) or ΔuvrD ΔrecC (B). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. *: indicates statistically significant (p ≤ 0.05) differences between time points and t = 0 h of the same samples.

https://doi.org/10.1371/journal.pgen.1012305.g005

To confirm that recB and recC were components of this epistatic persister recovery network, we performed complementation experiments with the double knockout mutants (S3 Fig). Expression of recB and recC from their native promoters on plasmids significantly restored inhibition-assisted recovery in double knockout mutants; however, the empty vector pUA66 and expression of uvrD from its native promoter on pUA66 presented issues with colony counts falling under the limit of detection (S3 Fig). To address this technical hurdle, we used Tn7-mediated site-specific transposition to insert deleted genes and their native promoters directly into the chromosome (S1 Table) [47]. As an empty control, we incorporated the associated multiple cloning site (MCS) at the same genomic position (S1 Table). Expression of uvrD from its native promoter located in the attTn7 site restored inhibition-assisted recovery in both ΔuvrD ΔrecB and ΔuvrD ΔrecC, whereas the empty control had little difference in survival from before and after 12 h of recovery in ΔuvrD ΔrecB (S3 Fig). For ΔuvrD ΔrecC, the empty control still resulted in colony counts below the limit of detection. Altogether, these data suggested that recB and recC, which with recA comprise major DSB repair machinery of HR [27,29], exhibit epistasis with uvrD in the phenomenon of inhibition-assisted recovery of FQ persisters.

2.3. recA and uvrD are more critical to FQ persister recovery than other NER and HR genes

From the above genetic analyses, we identified that inhibition-assisted recovery of FQ persisters required mfd, uvrA, uvrB, and uvrD in strains devoid of recA; and recA, recB, and recC were required in strains devoid of uvrD. To further delineate interactions in this epistatic network, we generated the remaining double knockout mutants (ΔuvrA ΔrecB, ΔuvrB ΔrecB, Δmfd ΔrecB, ΔuvrA ΔrecC, ΔuvrB ΔrecC, Δmfd ΔrecC) and assayed them for FQ persister recovery. For combinations of Δmfd with ΔrecB or ΔrecC, FQ persister recovery was still observed (Fig 6), which differed from results with Δmfd ΔrecA where FQ persister recovery was eliminated (Figs 1 and 3). Alternatively, combinations of ΔuvrA with ΔrecB or ΔrecC or ΔuvrB with ΔrecB or ΔrecC exhibited attenuation of inhibition-assisted recovery (Fig 6). These results suggested that epistasis in persister recovery is not simply explainable by epistasis between NER and HR DSB machinery, but rather recA and uvrD appear to be more critical for FQ persister recovery than recB, recC, uvrA, uvrB, and mfd.

thumbnail
Download:
Fig 6. Screen for epistatic interactions of gene combinations that recapitulated the phenotypes of either recA or uvrD in FQ persister recovery.

Stationary-phase cultures of strains harboring indicated mutations were treated with 5 µg/mL OFL and assayed for recovery by washing three times with PBS and plating on LB agar supplemented with 25 µg/mL CAM on top of 0.2 µm filters. After 12 h, filters were transferred to LB agar plates for 24 h. ΔuvrA ΔrecB, ΔuvrB ΔrecB (A), and ΔuvrA ΔrecC, ΔuvrB ΔrecC (B) strains exhibited an attenuated response to inhibition-assisted recovery, whereas Δmfd ΔrecB (A) and Δmfd ΔrecC (B) did not exhibit epistasis. Strains here contained kanamycin resistance cassettes from the knockout procedure (S1 Table). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. †: indicates statistical significance (p ≤ 0.05) when compared to single mutants (ΔrecB or ΔrecC) and double mutants (ΔuvrD ΔrecB or ΔuvrD ΔrecC). #: reflects statistically significant difference between the indicated references strains.

https://doi.org/10.1371/journal.pgen.1012305.g006

2.4. RecA recombination activity and UvrD helicase activity are needed for FQ persister recovery

Due to their importance to FQ persister recovery and the fact that both RecA and UvrD have multiple functions, we used a panel of characterized mutants to assess the role of different RecA and UvrD functions in persister recovery. For UvrD, we considered UvrD(R284A), which was shown to eliminate helicase activity while preserving DNA binding activity [48]; UvrD(R396E), which was shown to reduce dsDNA binding activity while preserving helicase activity and ssDNA binding [49]; and UvrD(ΔCTD), which corresponds to deletion of the C-terminal domain (amino acids 645–720), which was shown to reduce the binding of UvrD to RNA polymerase [50]. As depicted in Fig 7A, mutants with reductions in dsDNA binding or RNA polymerase binding were able to restore persister recovery of ΔuvrD ΔrecA to levels that were comparable and statistically indistinguishable from native UvrD. However, UvrD(R284A) failed to restore persister recovery to ΔuvrD ΔrecA and was statistically indistinguishable from the empty vector control. These data suggest that the helicase activity of UvrD was involved in the recovery of FQ persisters, whereas the dsDNA and RNA polymerase binding activities appear to be dispensable functions for the phenomenon. Notably, UvrD(ΔCTD) exhibits a 2-fold reduction in RNA polymerase binding affinity and a significant reduction in UvrD-dependent RNA polymerase backtracking [50,51], which are important interactions for UvrD-mediated TCR [50,52]. The observation that inhibition-assisted recovery with UvrD(ΔCTD) resembles that of native UvrD thereby suggests that UvrD-mediated TCR is not required for the recovery of FQ persisters.

thumbnail
Download:
Fig 7. Assessment of separation of function mutants of RecA and UvrD in FQ persister recovery.

Stationary-phase cultures of strains harboring plasmids with indicated mutant were treated with 5 µg/mL OFL and assayed for recovery by washing three times with PBS and plating on LB agar supplemented with 25 µg/mL CAM and 50 µg/mL kanamycin on top of 0.2 µm filters. After 12 h, filters were transferred to LB agar plates with kanamycin for 24 h. Mutants of UvrD along with empty vector and native UvrD controls (A), and mutants of RecA along with empty vector and native RecA controls (B). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. *: reflects statistically significant differences (p ≤ 0.05) between the indicated mutant and native UvrD or RecA, but not empty vector control. #: reflects statistically significant difference between the indicated references strains.

https://doi.org/10.1371/journal.pgen.1012305.g007

For RecA, we considered RecA(G204S), which is also known as RecA430 and has retained much of its recombination activity while exhibiting little ability to initiate the SOS response [53], and RecA(N304D), which is recombination deficient, but proficient at cleaving LexA and initiating the SOS response [53]. As illustrated in Fig 7B, RecA(N304D) cannot restore persister recovery in ΔuvrD ΔrecA, whereas RecA(G204S) restores persister recovery to levels equivalent to that of native RecA. These results suggest that the recombination functions of RecA provide the basis of the epistatic interaction between RecA and UvrD in the recovery of FQ persisters, whereas the ability of RecA to initiate the SOS response could be a dispensable function, which agrees with the results of lexA3 in Fig 4.

2.5. Loss of DSB and SSG repair machinery do not recapitulate loss of recA in inhibition-assisted recovery

To investigate the role of RecA recombination in persister recovery more deeply, we considered that RecA contributes to both DSB repair with recB and recC and SSG repair with the RecFOR pathway [27,45]. We reasoned that loss of DSB repair and SSG repair would more closely resemble loss of recombination in ΔrecA because both types of damage repair involve HR. When mutants that were defective in NER (ΔuvrA), DSB repair (ΔrecB), and SSG repair (ΔrecF, ΔrecO, or ΔrecR) were assayed, FQ persister recovery was found to be equivalent to that of ΔuvrA ΔrecB and significantly greater than that of ΔuvrA ΔrecA (Fig 8). These data suggest that loss of DSB repair via ΔrecB and SSG repair via ΔrecF, ΔrecO, or ΔrecR fails to recapitulate the phenotype of ΔrecA in FQ persister recovery.

thumbnail
Download:
Fig 8. Screen for epistatic interactions of strain combinations deficient in NER, DSB repair, and SSG repair.

Stationary-phase cultures of strains harboring indicated mutations were treated with 5 µg/mL OFL and assayed for recovery by washing three times with PBS and plating on LB agar supplemented with 25 µg/mL CAM on top of 0.2 µm filters. After 12 h, filters were transferred to LB agar plates for 24 h. Strains here contained kanamycin resistance cassettes from the knockout procedure (S1 Table). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey tests were conducted to assess statistical significance. None of the triple mutants were statistically different from ΔuvrA ΔrecB and equivalent to ΔuvrA ΔrecA, which would constitute an absence of recovery; and none of the triple mutants were statistically different from ΔuvrA ΔrecA and ΔuvrA ΔrecB, which would constitute an intermediate recovery between the responses of those strains. #: reflects statistically significant difference between the indicated references strains.

https://doi.org/10.1371/journal.pgen.1012305.g008

2.6. Rifampicin exhibits inhibition-assisted recovery that also depends on recA and uvrD

In previous work, we observed that CAM or rifampicin (RIF) could produce inhibition-assisted recovery of FQ persisters in WT populations [23]. Given the roles of uvrA, uvrB, uvrD, and mfd in TCR we sought to determine whether inhibition-assisted recovery had different genetic dependencies when growth inhibition was triggered at the level of transcription (RIF), rather than translation (CAM). Conceivably, if TCR was solely required for persister recovery in the absence of recA, the use of RIF would recapitulate the phenotype of ΔuvrA, ΔuvrB, ΔuvrD, or Δmfd when combined with ΔrecA. As depicted in Fig 9, inhibition-assisted recovery was observed with WT, ΔuvrD, and ΔrecA, but recovery was not observed with ΔuvrD ΔrecA. This genetic dependency mirrors that of recovery with CAM and suggests that TCR is not solely required for FQ persister recovery in ΔrecA. Rather, results suggest that under conditions with active transcription (CAM used for recovery) Mfd-dependent TCR is required for FQ persister recovery in ΔrecA (Fig 3), whereas under conditions without transcription (RIF used for recovery) GGR likely provides the analogous function for FQ persister recovery in ΔrecA (Fig 9).

thumbnail
Download:
Fig 9. FQ persister recovery with RIF exhibits epistatic interaction with recA and uvrD.

Stationary-phase cultures of strains harboring indicated mutations were treated with 5 µg/mL OFL and assayed for recovery by washing three times with PBS followed by resuspension in LB with 750 µg/mL RIF. Filters were not used due to carryover of RIF, and liquid equivalency was established in previous work (see Materials and Methods). After 12 h, samples were washed and plated on LB agar for 24 h. Strains in these assays were cured of kanamycin resistance cassettes (S1 Table). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means.

https://doi.org/10.1371/journal.pgen.1012305.g009

3. Discussion

While persistence to antibiotics had previously been attributed to a lack of antibiotic-induced damage due to cell dormancy, several studies have shown that persisters accumulate DNA damage from FQ treatment, which they have to repair to survive [10,18–26,54]. Our group found that both stationary-phase susceptible cells and OFL persisters experienced DNA damage, which persisters could repair after the conclusion of treatment using DNA repair systems involving RecA [18]. Further, we reported that surviving persisters activated the SOS response and underwent filamentation prior to resuming replication following FQ-induced damage, a response that was highly dependent on RecA [10]. Similarly, Goormaghtigh and Van Melderen showed that this response was conserved in exponential-phase cells, as OFL persisters also experienced DNA damage, and, after treatment, underwent filamentation and induced the SOS response [20]. Later, we discovered that persisters containing two chromosome copies were more likely to survive FQ treatment than one chromosome cells due to their ability to participate in HR DSB repair, as the persister levels in cells with one and two chromosomes were near equivalent in ΔrecA and ΔrecB strains [22]. Notably, one chromosome cells still formed persisters, though their growth dynamics differed from both two chromosome persisters and untreated one chromosome controls, suggesting the involvement of HR-independent pathways in the survival of one chromosome cells [22]. Indeed, our group found that RecA was particularly important for the survival of two chromosome FQ persisters, whereas UvrD was important for the survival of one chromosome cells [54]. Further, Wilmaerts and coworkers found that persisters accumulate oxidative damage in response to OFL treatment, and that this damage can be fixed via NER [25]. Combined, these studies show that FQ persisters actively repair antibiotic-induced DNA damage through pathways that include both HR and NER.

Previous studies have suggested that the period after FQ treatment is critical for stationary-phase persister survival [19,23,38]. For example, our group showed that starving E. coli cells after FQ treatment resulted in increased persister levels and that this increased survival depended on the presence of RecA [19]. A similar response has also been observed in Salmonella enterica cultures where survival increased ~5-fold when cells were exposed to a nutrient-poor environment after FQ treatment compared to immediate exposure to nutrient-rich conditions [38]. Further, this phenomenon is not exclusive to starvation, as delaying growth resumption after FQ treatment either by the accumulation of intracellular toxins, such as MazF and LdrD, or by exposing cells to translational or transcriptional inhibitors, such as RIF or CAM, also resulted in higher persister levels [23]. Specifically, we showed that inhibiting translation following FQ treatment resulted in increased survival in ΔrecA and ΔuvrD strains, but not in the double mutant (ΔuvrD ΔrecA), providing further evidence for the role of HR and NER in FQ persistence [23]. Here, we took inspiration from that previous study and further investigated the epistatic interactions between DNA repair machinery that underlie the ability of persisters to recover from FQs when translation is impaired following treatment.

Using a genetic approach, we found that uvrD, uvrA, uvrB, and mfd exhibit epistasis with recA during inhibition-assisted recovery (Figs 1–3). These genes, along with uvrC, are known to participate in NER, which can be divided into three sub-pathways (GGR, Mfd-mediated TCR, and UvrD-mediated TCR) [32,33,36,37]. Importantly, all NER pathways repair damage on one strand of DNA and are not associated with repair of DSBs, which are discussed below. In GGR, DNA lesions are recognized by changes in helix rigidity [32–34]. In Mfd-mediated TCR, Mfd pushes forward an RNA polymerase stalled at a damage site to expose the lesion, whereas in UvrD-mediated TCR, the lesion is exposed when UvrD promotes backtracking of the stalled RNA polymerase [32,33]. In all three pathways, DNA lesions are recognized by the UvrA2B2 complex, which scans the genome for helix distortions [26,27]. Specifically, UvrA initiates contact at the damaged site and facilitates loading of UvrB onto the lesion [26,27]. Upon lesion identification, UvrA dissociates, allowing UvrB to recruit UvrC [26,27]. UvrC introduces two incisions, the first on the 3′ side and the second on the 5′ side of the lesion [32,33,35]. UvrD then uses its helicase activity to help remove the damaged nucleotides, and the resulting SSG is filled by DNA polymerase I and sealed by ligase [32,33]. Interestingly, neither UvrC nor Cho, the nucleases associated with NER, were required for inhibition-assisted recovery, although removal of both in ΔrecA led to an attenuation in persister recovery (Fig 1) [35]. Notably, a prior study by Wilmaerts and colleagues found UvrA, UvrB, UvrD, and Mfd to be important for FQ persister survival, but not UvrC [25]. However, that study did not examine the role of Cho and was not investigating survival changes in populations exposed to a growth-inhibited environment immediately after treatment [25]. We postulate that another nuclease that has not been associated with canonical NER functions substitutes for UvrC and Cho in this process when they are absent. Identifying such a nuclease and how it interacts with canonical NER machinery represents an interesting area for future work. In addition, we observed that transcriptional inhibition after FQ exposure produced persister recovery with the same epistatic interactions between recA and uvrD that translational inhibition did (Fig 9). Such data suggested that transcription is not required to observe FQ persister recovery in ΔrecA, and that TCR is not solely responsible for persister recovery in the absence of HR.

To further identify genes in this epistatic network, we implemented a complementary approach to identify uvrD epistatic interaction partners. We found that recA, recB, and recC exhibited epistasis with uvrD in inhibition-assisted recovery (Figs 2, 4, and 5), which together with recD, repair DSBs through the RecBCD pathway of HR [27,29]. Briefly, the RecBCD helicase activity unwinds duplex DNA and degrades both strands of DNA with its nuclease activity until a Chi site is encountered, which triggers a shift in nuclease activity that leads to the production of a 3’ ssDNA tail [27,29]. The RecBCD complex promotes RecA loading onto that ssDNA, allowing the resulting RecA filament to promote the search for a homologous sequence and strand invasion [27,29]. The involvements of RecB and RecC, but not of RecD, in FQ persistence has been observed previously; however, those studies did not focus on the post-treatment recovery period [25,55]. Several works have reported a less prominent role of RecD in recombination [29,56–60]. Knockout mutants of recB, which has both the nuclease domain and 3’ to 5’ helicase activity, or recC, which is crucial for Chi site recognition, exhibit reduced viability and recombination deficiency, whereas a ∆recD mutation has no effect on viability and cells remain proficient in recombination [29,56–60]. It has been postulated that this may occur due to the ability of RecD to inhibit recombination until encountering a Chi site; thus, ΔrecD strains would constitutively mimic the post-Chi state, allowing RecBC to perform recombination and repair independently of RecD [29,56–60]. Overall, our results suggest that in the absence of UvrD, persisters can recover from FQ treatment by repairing DSBs through the RecBCD pathway of HR.

The epistatic interactions of uvrA, uvrB, and mfd with recA, suggest that persister levels can increase through the repair of ssDNA lesions with the use of NER. Such lesions could result from gyrase monomers or topoisomerase IV subunits that remain covalently bound to the phosphate backbone on one side of DNA, or from the accumulation of oxidative lesions caused by FQ treatment [13–16,25,61]. In contrast, the epistatic interactions of recB and recC with uvrD suggest that persisters survive by repairing DSBs caused by FQ treatment through the RecBCD pathway of HR. Previous work has reported epistatic interactions between HR and NER in the repair of DNA-protein crosslinks (DPCs) [62–64]. NER can repair DPCs smaller than 12–14 kDa [62,63], which could mimic lesions formed by the degradation of gyrA or parC monomers that had remained bound to DNA following FQ treatment. This restriction reflects the limited ability of UvrB to load onto larger DPCs sites, which is needed to recruit UvrC to excise the damage [62,63]. Conversely, HR can process DPCs across a wider size range, including both small and large lesions [62,63]. Given that HR represents a more versatile system for repairing DNA damage than NER, it could explain why mutants deficient in NER but with intact HR recover faster post-FQ treatment than those deficient in HR but with intact in NER (Figs 2–3 and 5). Importantly, while restoration of culturability was measured here for a variety of DNA repair mutants, measurement of DNA damage in persisters was not feasible. Specifically, persisters are minority populations that are difficult to differentiate from viable but non-culturable cells (VBNCs) prior to outgrowth [65], which makes measurement of DNA damage in them difficult. Conceivably, fluorescent reporters of DNA damage can be used [18]; however, those reporters often rely on the SOS response, which is inactive in most of the strains examined here due to the use of ΔrecA. Alternatively, population level metrics of DNA damage can be used to assess whether correlations exist between population-wide measurements and persister recovery, such as gyrase cleavage site sequencing (GCS-seq) [55] and bacterial rapid approach to DNA adduct recovery (RADAR) [66]. In the future, application of such techniques to the strains examined here could identify whether differential levels of DNA damage underlie any of the trends observed; however, given that the OFL concentration used here was ~ 80-fold the MIC of WT [67], it is likely that any difference in DNA damage detected would reflect differences in ability to repair the damage or differences in DNA gyrase or topoisomerase activity between strains.

To delineate additional dependencies in the epistatic interaction network underlying FQ persister recovery, we generated the remaining double knockout mutant combinations between the NER and HR machinery that contribute to the recovery of FQ persisters (ΔuvrA ΔrecB, ΔuvrB ΔrecB, Δmfd ΔrecB, ΔuvrA ΔrecC, ΔuvrB ΔrecC, Δmfd ΔrecC), all of which retained functional RecA and UvrD. Unexpectedly, none of those combinations completely recapitulated the absence of inhibition-assisted recovery observed in strains devoid of recA or uvrD (Fig 6). Given that HR can proceed through the RecFOR pathway to repair SSGs [27,45], we constructed triple knockout mutants defective in SSG repair, DSB repair, and NER (ΔrecF ΔuvrA ΔrecB, ΔrecO ΔuvrA ΔrecB, ΔrecR ΔuvrA ΔrecB) while retaining functional RecA and UvrD. Loss of both HR pathways through combined deletions of ΔrecB and ΔrecF, ΔrecO, or ΔrecR still failed to recapitulate the phenotype of ΔrecA (Fig 8). Likewise, loss of ΔuvrA, ΔuvrB, and Δmfd did not recapitulate loss of uvrD. Together, these findings suggest that inhibition-assisted recovery depends on RecA- and UvrD-mediated activities that go beyond simply a loss of HR or NER in their respective deletion mutants. To elucidate the functions of RecA and UvrD that were required for inhibition-assisted recovery we used a panel of mutants where selective functions of RecA and UvrD had been knocked down. Those experiments revealed that the recombination activity of RecA and helicase activity of UvrD were important to FQ persister recovery, whereas reductions in SOS induction by RecA and dsDNA and RNA polymerase binding by UvrD did not impact the phenomenon (Fig 7).

Beyond the prototypical role of UvrD in NER, studies have suggested a role for it outside of NER [63,68–70]. As ΔuvrD strains exhibit elevated levels of HR, it has been proposed that UvrD helps remove RecA filaments from stalled replication forks, which prevents unnecessary recombination that could otherwise be lethal to the cell [68–70]. Additionally, uvrD mutants showed modest sensitivity to azacytidine, which induces DPCs that exceed the size limit for NER and are therefore expected to be processed exclusively by HR [62,63]. Beyond the prototypical role of RecA in HR, Buljubašić and colleagues described a third HR pathway active in ΔrecBCD sbcB15 sbcC(D) ΔrecF(OR) mutant derivatives that enables efficient DNA repair independent of RecBCD and RecFOR, but dependent on RecA and the SbcB15 allele [71]. In that pathway, the RecQ helicase unwinds DNA and the RecJ exonuclease produces a 3’ ssDNA overhang where RecA can bind [71]. Exploring the additional interactions of uvrD and recA that give them higher degrees of epistasis in this phenomenon are interesting areas of future study.

Collectively, the results presented here suggest that HR machinery facilitated FQ persister recovery that was observable in the immediate hours after FQ treatment, whereas the contributions of NER machinery became apparent at longer timescales in the absence of RecA. Further, combinations of NER and HR DSB knockout mutants failed to fully recapitulate the phenotypes of uvrD or recA, which suggested the existence of an epistatic network where RecA and UvrD had stronger connections to other nodes than other members (Fig 10). In addition, we determined that the recombination activity of RecA and helicase activity of UvrD were essential to their epistatic interaction in this network. Overall, this study elaborates on an epistatic interaction network involving HR and NER machinery in the recovery persisters from FQ treatment and further elaborates on the critical roles that RecA and UvrD in this phenomenon.

thumbnail
Download:
Fig 10. Diagram of the epistatic interaction network driving FQ persister recovery.

Solid lines indicate epistatic interactions of uvrA, uvrB, and mfd with recA, and epistatic interactions of recB and recC with uvrD. Dashed lines indicate attenuated recovery levels for the ΔuvrD ΔrecN and Δcho ΔuvrC ΔrecA mutants, as well as when uvrA or uvrB were deleted alongside either recB or recC.

https://doi.org/10.1371/journal.pgen.1012305.g010

4. Materials and Methods

4.1. Bacterial strains and plasmids

All strains used in this study were derived from E. coli MG1655, and details about their construction are provided in S1 Table. Genetic knockouts were transduced from the respective Keio collection mutant via P1 phage transduction [72,73]. When indicated, kanamycin resistance markers were removed using FLP recombinase expressed from pCP20 or pGL04 [74,75]. Knockouts were confirmed via PCR with primers that anneal to the gene of interest (internal primer pair), the upstream and downstream sequences of the gene of interest (external primer pair), and upstream to the gene and internal to the kanamycin resistance cassette. All primer sequences are provided in S2 Table. The uncleavable lexA mutant, lexA3, was constructed in a previous study [24], and the mutation was confirmed by whole genome sequencing (Plasmidsaurus Inc.).

For complementation of recA, uvrA, uvrB, mfd, uvrD, recB, and recC, pUA66 plasmids containing either PrecA-recA, PuvrA-uvrA, PuvrB-uvrB, Pmfd-mfd, PuvrD-uvrD, PrecB-recB, or PrecC-recC were used. These plasmids were constructed using Gibson Assembly [76], verified by whole plasmid sequencing (Plasmidsaurus Inc.), and transformed into strains of interest by electroporation (see below). A pUA66 plasmid containing a promoterless gfpmut2 was transformed into our strains of interest and used as an empty vector control. For chromosomal complementation at the Tn7 attachment site (attTn7), transpositions were conducted using pGRG25 as described previously [47]. A fragment containing PuvrD-uvrD was introduced into the MCS of pGRG25 via Gibson assembly [76] and confirmed by whole plasmid sequencing (Plasmidsaurus Inc.). The non-coding MCS between the left and right transposition arms of pGRG25 was used as an empty control. Plasmids were transformed into ΔuvrD ΔrecB and ΔuvrD ΔrecC by electroporation. Transpositions of interest were confirmed via PCR with primers that flank the attTn7 site listed in S2 Table, and loss of pGRG25 was confirmed by the absence of growth in ampicillin at 30–32 ℃.

Separation-of-function mutations were introduced into plasmids harboring the WT sequence, pUA66 PrecA-recA or pUA66 PuvrD-uvrD, using the Q5 Site-Directed Mutagenesis Kit (New England Biolabs). Briefly, plasmid backbones were amplified with mutagenic primers using the provided Q5 Hot Start High-Fidelity 2X Master Mix. Following amplification, products were treated with the provided Kinase-Ligase-DpnI (KLD) enzyme mix at room temperature for 5 min. All mutations were confirmed via whole plasmid sequencing (Plasmidsaurus Inc.), and plasmids were transformed into ΔuvrD ΔrecA by electroporation. All plasmids used in this study are listed in S3 Table, and the primers used for plasmid construction can be found in S4 Table.

To generate electrocompetent cells, strains of interest were first inoculated from -80 °C, 25% glycerol stocks into 3 mL of LB media and incubated at 37 °C with shaking at 250 r.p.m for at least 16 h. Overnight cultures were then diluted 100-fold in 3 mL of SOB media. When cultures reached mid-exponential phase (OD600 = ~0.4-0.6), 1 mL samples were added to microcentrifuge tubes and chilled on ice for 10 min. Chilled cultures were washed three times. In each wash, cells were collected by centrifugation at 5,000 r.p.m for 5 min at 4 ºC, followed by removal of supernatants. In the first wash, cell pellets were resuspended in 1 mL of ice-chilled 1% glycerol. In the second and third washes, cell pellets were resuspended in 1 mL of ice-chilled 10% glycerol. After the third wash, cell pellets were resuspended in 0.1 mL of ice-chilled 10% glycerol to produce electrocompetent cells. For electroporation, electrocompetent cells were mixed with ~100 ng of purified plasmid and chilled on ice for 30 min. The mixture was then added to an electroporation cuvette and electroporated. Afterward, 900 μL of SOC media was added to cuvettes and mixtures were transferred to microcentrifuge tubes. Cultures were then incubated at 37 ºC (for pUA66 or pGL04 transformations) or 30–32 ºC (for pGRG25 or pCP20 transformations) for 1 h, after which they were centrifuged at 15,000 r.p.m for 3 min. After incubation, 900 μL of supernatants were removed and the remaining culture volume was plated on LB agar containing the appropriate selection marker.

4.2. Chemicals and media

All chemical components were obtained from Fisher Scientific or Sigma Aldrich, unless otherwise specified. All media were prepared in distilled water purified using a Millipore Milli-Q lab water system (Burlington, MA) to a resistivity of 18.2 MΩ.cm. LB media was made by dissolving 10 g/L tryptone, 5 g/L yeast extract, and 10 g/L NaCl in MilliQ water, which was then autoclaved. LB agar plates were made with 25 g/L pre-mixed LB Miller broth and 15 g/L agar dissolved in MilliQ water, which was autoclaved. For P1 transduction, special LB medium was made by adding 1.75 mL of 1 M CaCl2 and 3 mL of 1 M MgSO4 250 mL of LB media, and for mutant selection sodium-citrate plates were made by supplementing LB agar plates with 25 mM of sterile-filtered sodium citrate and kanamycin. For bacterial transformations, SOB media was prepared by autoclaving 20 g/L tryptone, 5 g/L yeast extract, 5 mM NaCl (5 M filter sterilized stock solution), and 2.5 mM KCl (1 M filter sterilized stock solution) in MilliQ water. After, a 1 M MgSO4 stock solution was prepared in MilliQ water, filter-sterilized, and added to a final concentration of 10 mM. SOC media was prepared by supplementing SOB media with 20 mM of filter-sterilized 1 M glucose. To generate electrocompetent cells and cells stocks, 1%, 10%, and 50% glycerol solutions were prepared by diluting the appropriate amount of glycerol in autoclaved water and sterilizing with 0.22 μM bottle top filters (Merck Millipore Ltd, Burlington, MA). For curing of pGL04, no salt LB plates were prepared by dissolving 1 g tryptone, 0.5 g yeast extract, and 1.5 g agar in 80 mL MilliQ water and autoclaving. Once that mixture was cool to touch, 20 mL of 1.5 M filter-sterilized sucrose was added. To induce the expression of tnsABCD from pGRG25, 1 M stocks of L-arabinose were prepared in autoclaved water and filter sterilized. Gutnick media used for persister and recovery assays was made by mixing 10 mM glucose (filter sterilized), 10 mM NH4Cl (filter sterilized), and 1X Gutnick salts in autoclaved MilliQ water, and sterilizing with 0.22 μM bottle top filters. Autoclaved 10-fold concentrated (10X) Gutnick salts solution contained 47 g/L KH2PO4, 135 g/L K2HPO4, 10 g/L K2SO4, and 1 g/L MgSO4·7H2O in MilliQ water. For all wash steps, sterile-filtered 1X phosphate-buffered saline (PBS) solution prepared from a 10X stock was used. The 10X stock was prepared by mixing 98.9 g of powdered PBS (81% NaCl, 14% Na2HPO4, 3% KH2PO4, and 2% KCl by weight) in 1 L of Milli-Q water followed by autoclaving.

For mutant selection and plasmid retention, a final concentration of 100 μg/mL ampicillin or 50 μg/mL kanamycin were used in media and plates. For FQ persister assays, 5 μg/mL OFL was used. For translational inhibition, a final concentration of 25 μg/mL CAM was added to autoclaved LB agar. Stock solutions for all antibiotics and chemicals, except CAM, RIF, and OFL, were prepared in autoclaved MilliQ water and filter sterilized before use. CAM was dissolved in ethanol, RIF was dissolved in DMSO, and 1M NaOH was added to OFL stock solution in autoclaved MilliQ water until it was fully dissolved and then filter sterilized.

4.3. Persistence assays

Cultures were first inoculated from -80 °C, 25% glycerol stocks into 3 mL of LB media and incubated at 37 °C with shaking at 250 r.p.m. After 4 h, those cultures were diluted 100-fold in 3 mL of Gutnick media with 10 mM glucose as the sole carbon source and incubated at 37 °C with shaking at 250 r.p.m. for 20 h. After 20 h of incubation, 500 µL samples were removed for t = 0 h measurements prior to antibiotic treatment. Then OFL was added to a final concentration of 5 µg/mL, cultures were incubated at 37 °C with shaking at 250 r.p.m. and additional 500 µL samples were taken after 1, 3, and 5 h. For all time points, samples were washed three times by centrifugation at 15,000 r.p.m for 3 min, removal of 450 µL of supernatant, and resuspension of cell pellets in 450 µL of sterile PBS. After the three washes, samples were spun once more and concentrated 5-fold by removing 400 µL of supernatant and resuspending the cell pellet in the remaining 100 µL of PBS. Ten µL of the concentrated samples were used for 10-fold serial dilutions in 90 µL of PBS. For each sample, 10 µL per dilution were spotted onto LB agar plates that were incubated at 37 °C for 24 h, after which CFUs were enumerated for dilutions that were countable and contained at least 5 colonies.

4.4. Recovery assays

After 5 h of OFL treatment, 500 µL samples were taken and washed three times with sterile PBS. Each wash step consisted of a centrifugation step at 15,000 r.p.m for 3 min, removal of 450 µL of supernatant, and resuspending pellets in 450 µL of PBS. After the three washes, samples were spun once more and concentrated 5-fold by removing 400 µL of supernatant and resuspending cells in the remaining 100 µL of PBS. Ten-fold serial dilutions in PBS were made for each sample. For t = 0 h, a 10 µL spot of each dilution was plated directly onto an LB agar plate. For all other timepoints, 10 µL of each dilution were spotted onto Supor 200 polyether sulfone membranes with 0.2 µm pores (Pall Corporation) that were placed atop of LB agar plates supplemented with 25 µg/mL CAM and incubated at 37 °C. As shown previously, membranes were used to immobilize cells to ensure that CFU counts were not influenced by any cell divisions that occurred during the recovery period [19,23]. After 2, 4, 6, 8, 10, or 12 h, membranes were transferred to LB agar plates and incubated at 37 °C for 24 h, after which CFUs were counted for dilutions that were countable and contained at least 5 colonies. For initial mutant screening, membranes were transferred to LB agar plates only at t = 12 h. Agar thickness was doubled for LB plates receiving membranes from CAM-supplemented plates to support bacterial growth, as it allowed any residual antibiotic from the membrane to disperse.

For RIF recovery assays, use of filters was not possible due to the affinity of the filters for RIF [23]. Therefore, liquid based assays were used, which were shown previously to be indistinguishable from filter-based assays using CAM [23]. In brief, after 5 h of OFL treatment, 500 µL samples were taken and washed three times with sterile PBS. Each wash step consisted of a centrifugation step at 15,000 r.p.m for 3 min, removal of 450 µL of supernatant, and resuspension of pellets in 450 µL of PBS. After three washes, samples were spun once more and concentrated 5-fold by removal of 400 µL of supernatant and resuspension of cells in the remaining 100 µL of PBS. Cells were then resuspended in liquid LB supplemented with 750 µg/mL RIF, a concentration previously observed to halt transcription and TCR in E. coli [37,77]. Specifically, 30 µL aliquots of the previously mentioned 100 µL cell suspensions were inoculated into 3 mL of liquid LB supplemented with RIF. These cultures were incubated at 37 °C with shaking at 250 r.p.m for 12 h. A 500 µL sample was taken before and after the 12 h incubation. For each time point, samples were washed three times, concentrated 5-fold, and 10-fold serially diluted in PBS. Ten µL per dilution were plated onto LB agar plates and incubated at 37 °C for 24 h. Further, an additional 75 µL of undiluted samples were also plated in the case of low culturability samples. CFUs were enumerated for dilutions that were countable and contained at least 5 colonies.

To confirm that 750 µg/mL of RIF was sufficient to halt bacterial growth under our experimental conditions, control experiments were performed using untreated (no FQ) cultures (S4 Fig). Cells were incubated for a total of 25 h, to account for the standard 20 h overnight growth period and a 5 h mock-treatment period. Following incubation, cultures were processed as described above, with a single modification to account for the absence of FQ-induced cell death. Instead of inoculating a 30 µL aliquot of the 100 µL concentrate, the untreated cultures were diluted to approximately the same survival levels observed in their respective OFL-treated counterparts. Specifically, the WT strain was diluted an additional 20-fold, whereas the mutant strains ΔuvrD, ΔrecA, and ΔuvrD ΔrecA were diluted an additional 5,000-fold, 50,000-fold, and 125,000-fold, respectively.

4.5. Statistical analysis

Data points indicate the mean values of at least three biological replicates and error bars represent the standard errors of the means. Where indicated, one way ANOVA with post-hoc Tukey tests were performed to assess statistical significance among the different timepoints of the same condition or between the indicated mutants. Asterisks (*), obelisks (†), and hash marks (#) denote p-values ≤ 0.05.

Supporting information

S1 Fig. Complementation of RecA or UvrD restores inhibition-assisted recovery.

ΔuvrD ΔrecA containing pUA66 plasmids with PuvrD-uvrD or PrecA-recA, or without an expression cassette (empty control) were treated with 5 µg/mL OFL for 5 h. All samples exhibited biphasic killing during OFL treatment, indicating the presence of persisters. For recovery assays, cells were washed with PBS after OFL treatment and plated on filters on top of LB agar with 25 µg/mL of CAM for the designated time before being transferred to LB agar for 24 h. Complementation of uvrD or recA in the double deletion mutant restored inhibition-assisted recovery, whereas the empty vector control showed no difference in survival. Data points indicate the means of at least three biological replicates with error bars indicating standard errors of the means. One-way ANOVA with post-hoc Tukey test were conducted to assess significance. *: indicates statistically significant (p < 0.05) differences between time points and t = 0 h of the same samples.

https://doi.org/10.1371/journal.pgen.1012305.s001

(TIF)

S2 Fig. Complementation of ΔuvrA, ΔuvrB, Δmfd, and ΔrecA restores inhibition-assisted recovery.

ΔuvrA ΔrecA containing pUA66 plasmids with PuvrA-uvrA, PrecA-recA, or without an expression cassette (empty control) (A), ΔuvrB ΔrecA containing pUA66 plasmids with PuvrB-uvrB, PrecA-recA, or without an expression cassette (empty control) (B), and Δmfd ΔrecA containing pUA66 plasmids with Pmfd-mfd, PrecA-recA, or without an expression cassette (empty control) (C) were treated with 5 µg/mL OFL for 5 h. All samples exhibited biphasic killing during OFL treatment, indicating the presence of persisters. For recovery assays, cells were washed with PBS after OFL treatment and plated on filters on top of LB agar with 25 µg/mL of CAM for the designated time before being transferred to LB agar for 24 h. Complementation of uvrA, uvrB, mfd, or recA in the double deletion mutants restored inhibition-assisted recovery. Empty vector controls showed only small changes in survival. Note that the survival change for the t = 12 h CAM exposure for ΔuvrA ΔrecA with the empty vector was significant but only ~3-fold higher than t = 0 h, whereas the recA- and uvrA-complemented strains were significant and ~20- and ~40-fold higher, respectively. Also, the timepoints for ΔuvrB ΔrecA with the empty vector that were significantly different than t = 0 h were all lower survival than t = 0 h. Data points indicate the means of at least three biological replicates with error bars indicating standard errors of the means. One-way ANOVA with post-hoc Tukey test were conducted to assess significance. *: indicates statistically significant (p < 0.05) differences between time points and t = 0 h of the same samples. P-values for ΔuvrB ΔrecA containing pUA66 plasmids with PuvrB-uvrB at t = 8 h and t = 10 h were ~0.064 and ~0.054, respectively.

https://doi.org/10.1371/journal.pgen.1012305.s002

(TIF)

S3 Fig. Complementation of ΔuvrD ΔrecB and ΔuvrD ΔrecC restores inhibition-assisted recovery.

ΔuvrD ΔrecB with pUA66 plasmid containing PrecB-recB, PuvrD-uvrD or without an expression cassette (empty control) and ΔuvrD ΔrecB with a chromosomal insertion at the Tn7 site containing an empty MCS control or PuvrD-uvrD (A), and ΔuvrD ΔrecC with pUA66 plasmid containing PrecC-recC, PuvrD-uvrD or without an expression cassette (empty control) and ΔuvrD ΔrecC with a chromosomal insertion at the Tn7 site containing an empty MCS control or PuvrD-uvrD (B) were treated with 5 µg/mL OFL for 5 h. All samples, except ΔuvrD ΔrecC containing an empty MCS control at the Tn7 site, exhibited biphasic killing during OFL treatment, indicating the presence of persisters. For recovery assays, cells were washed with PBS after OFL treatment and plated on filters on top of LB agar containing 25 µg/mL of CAM for the designated time before being transferred to LB agar for a 24 h incubation at 37 ℃. Plasmid-mediated complementation of recB or recC in the double deletion mutants and chromosomal complementation of uvrD at the attTn7 site for both double knockout mutants restored inhibition-assisted recovery. Plasmid-mediated complementation of uvrD in ΔuvrD ΔrecC showed no significant increases in survival. Since plasmid-mediated empty vector control for ΔuvrD ΔrecB resulted in colony counts that were below the limit of detection, we inserted a MCS at the chromosomal attTn7 site to use as our empty control, which showed no increase in survival. Both plasmid and chromosomal empty vector controls for ΔuvrD ΔrecC resulted in colony counts that were below the limit of detection either after 1 h of treatment or during the recovery period. Data points indicate the means of at least two biological replicates with error bars indicating the standard errors of the means. One-way ANOVA with post-hoc Tukey test were conducted to assess significance. *: indicates statistically significant (p < 0.05) differences between time points and t = 0 h of the same samples.

https://doi.org/10.1371/journal.pgen.1012305.s003

(TIF)

S4 Fig. RIF treatment inhibits bacterial growth.

Stationary-phase cultures of WT or the indicated mutant strains were washed three times with PBS followed by resuspension in LB or in LB supplemented with 750 µg/mL RIF for 12 h. Filters were not used due to carryover of RIF, and liquid equivalency was established in previous work (see Materials and Methods). Before and after the 12 h incubation, samples were washed and plated on LB agar for 24 h. Strains in these assays were cured of kanamycin resistance cassettes (S1 Table). Data points reflect the mean values of at least three biological replicates with error bars indicating the standard errors of the means. *: reflects statistically significant differences (p < 0.05) between the 0 and 750 µg/mL RIF concentrations for the indicated strains.

https://doi.org/10.1371/journal.pgen.1012305.s004

(TIF)

S1 Table. List of strains used in the manuscript.

https://doi.org/10.1371/journal.pgen.1012305.s005

(DOCX)

S2 Table. List of primers (5’ → 3’) used to verify genetic modifications.

https://doi.org/10.1371/journal.pgen.1012305.s006

(DOCX)

S3 Table. List of plasmids used in this study.

https://doi.org/10.1371/journal.pgen.1012305.s007

(DOCX)

S4 Table. List of primers used for plasmid construction.

https://doi.org/10.1371/journal.pgen.1012305.s008

(DOCX)

Acknowledgments

We thank the National BioResource Project (NIG, Japan) for distribution of the Keio collection.

References

  1. 1. Balaban NQ, Merrin J, Chait R, Kowalik L, Leibler S. Bacterial persistence as a phenotypic switch. Science. 2004;305(5690):1622–5. pmid:15308767
  2. 2. Balaban NQ, Helaine S, Lewis K, Ackermann M, Aldridge B, Andersson DI, et al. Definitions and guidelines for research on antibiotic persistence. Nat Rev Microbiol. 2019;17(7):441–8. pmid:30980069
  3. 3. Van den Bergh B, Fauvart M, Michiels J. Formation, physiology, ecology, evolution and clinical importance of bacterial persisters. FEMS Microbiol Rev. 2017;41(3):219–51. pmid:28333307
  4. 4. Levin-Reisman I, Ronin I, Gefen O, Braniss I, Shoresh N, Balaban NQ. Antibiotic tolerance facilitates the evolution of resistance. Science. 2017;355(6327):826–30. pmid:28183996
  5. 5. Spoering AL, Lewis K. Biofilms and planktonic cells of Pseudomonas aeruginosa have similar resistance to killing by antimicrobials. J Bacteriol. 2001;183(23):6746–51. pmid:11698361
  6. 6. Lafleur MD, Qi Q, Lewis K. Patients with long-term oral carriage harbor high-persister mutants of Candida albicans. Antimicrob Agents Chemother. 2010;54(1):39–44. pmid:19841146
  7. 7. Fauvart M, De Groote VN, Michiels J. Role of persister cells in chronic infections: clinical relevance and perspectives on anti-persister therapies. J Med Microbiol. 2011;60(Pt 6):699–709. pmid:21459912
  8. 8. Windels EM, Michiels JE, Fauvart M, Wenseleers T, Van den Bergh B, Michiels J. Bacterial persistence promotes the evolution of antibiotic resistance by increasing survival and mutation rates. ISME J. 2019;13(5):1239–51. pmid:30647458
  9. 9. Liu J, Gefen O, Ronin I, Bar-Meir M, Balaban NQ. Effect of tolerance on the evolution of antibiotic resistance under drug combinations. Science. 2020;367(6474):200–4. pmid:31919223
  10. 10. Barrett TC, Mok WWK, Murawski AM, Brynildsen MP. Enhanced antibiotic resistance development from fluoroquinolone persisters after a single exposure to antibiotic. Nat Commun. 2019;10(1):1177. pmid:30862812
  11. 11. Bartell JA, Cameron DR, Mojsoska B, Haagensen JAJ, Pressler T, Sommer LM, et al. Bacterial persisters in long-term infection: Emergence and fitness in a complex host environment. PLoS Pathog. 2020;16(12):e1009112. pmid:33315938
  12. 12. Fisher RA, Gollan B, Helaine S. Persistent bacterial infections and persister cells. Nat Rev Microbiol. 2017;15(8):453–64. pmid:28529326
  13. 13. Pohlhaus JR, Kreuzer KN. Norfloxacin-induced DNA gyrase cleavage complexes block Escherichia coli replication forks, causing double-stranded breaks in vivo. Mol Microbiol. 2005;56(6):1416–29. pmid:15916595
  14. 14. Tamayo M, Santiso R, Gosalvez J, Bou G, Fernández JL. Rapid assessment of the effect of ciprofloxacin on chromosomal DNA from Escherichia coli using an in situ DNA fragmentation assay. BMC Microbiol. 2009;9:69. pmid:19364397
  15. 15. Drlica K, Zhao X. DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiology and Molecular Biology Reviews. 1997;61(3):377–92.
  16. 16. Drlica K, Malik M, Kerns RJ, Zhao X. Quinolone-mediated bacterial death. Antimicrob Agents Chemother. 2008;52(2):385–92. pmid:17724149
  17. 17. Snyder M, Drlica K. DNA gyrase on the bacterial chromosome: DNA cleavage induced by oxolinic acid. J Mol Biol. 1979;131(2):287–302. pmid:226717
  18. 18. Völzing KG, Brynildsen MP. Stationary-phase persisters to ofloxacin sustain DNA damage and require repair systems only during recovery. mBio. 2015;6(5):e00731-15.
  19. 19. Mok WWK, Brynildsen MP. Timing of DNA damage responses impacts persistence to fluoroquinolones. Proc Natl Acad Sci U S A. 2018;115(27):E6301–9. pmid:29915065
  20. 20. Goormaghtigh F, Van Melderen L. Single-cell imaging and characterization of Escherichia coli persister cells to ofloxacin in exponential cultures. Sci Adv. 2019;5(6):eaav9462. pmid:31223653
  21. 21. Theodore A, Lewis K, Vulic M. Tolerance of Escherichia coli to fluoroquinolone antibiotics depends on specific components of the SOS response pathway. Genetics. 2013;195(4):1265–76. pmid:24077306
  22. 22. Murawski AM, Brynildsen MP. Ploidy is an important determinant of fluoroquinolone persister survival. Curr Biol. 2021;31(10):2039-2050.e7. pmid:33711253
  23. 23. Lemma AS, Brynildsen MP. Toxin Induction or Inhibition of Transcription or Translation Posttreatment Increases Persistence to Fluoroquinolones. mBio. 2021;12(4):e0198321. pmid:34399616
  24. 24. Lemma AS, Soto-Echevarria N, Brynildsen MP. Fluoroquinolone Persistence in Escherichia coli Requires DNA Repair despite Differing between Starving Populations. Microorganisms. 2022;10(2):286. pmid:35208744
  25. 25. Wilmaerts D, Focant C, Matthay P, Michiels J. Transcription-coupled DNA repair underlies variation in persister awakening and the emergence of resistance. Cell Rep. 2022;38(9):110427. pmid:35235801
  26. 26. Dörr T, Lewis K, Vulić M. SOS response induces persistence to fluoroquinolones in Escherichia coli. PLoS Genet. 2009;5(12):e1000760. pmid:20011100
  27. 27. Dillingham MS, Kowalczykowski SC. RecBCD enzyme and the repair of double-stranded DNA breaks. Microbiol Mol Biol Rev. 2008;72(4):642–71, Table of Contents. pmid:19052323
  28. 28. Bianco PR. The Role of the RecFOR Complex in Genome Stability. Multidisciplinary Digital Publishing Institute. 2025;26(12).
  29. 29. Amundsen SK, Taylor AF, Smith GR. The RecD subunit of the Escherichia coli RecBCD enzyme inhibits RecA loading, homologous recombination, and DNA repair. Proc Natl Acad Sci U S A. 2000;97(13):7399–404. pmid:10840065
  30. 30. Anderson DG, Kowalczykowski SC. The Translocating RecBCD Enzyme Stimulates Recombination by Directing RecA Protein onto ssDNA in a x-Regulated Manner.
    • 31. Morimatsu K, Kowalczykowski SC. RecFOR proteins load RecA protein onto gapped DNA to accelerate DNA strand exchange: a universal step of recombinational repair. Mol Cell. 2003;11(5):1337–47. pmid:12769856
    • 32. Monnet J, Grange W, Strick TR, Joly N. Mfd as a central partner of transcription coupled repair. Transcription. 2013;4(3):109–13. pmid:23756341
    • 33. Kraithong T, Hartley S, Jeruzalmi D, Pakotiprapha D. A Peek Inside the Machines of Bacterial Nucleotide Excision Repair. Int J Mol Sci. 2021;22(2):952. pmid:33477956
    • 34. Jaciuk M, Nowak E, Skowronek K, Tańska A, Nowotny M. Structure of UvrA nucleotide excision repair protein in complex with modified DNA. Nat Struct Mol Biol. 2011;18(2):191–7. pmid:21240268
    • 35. Moolenaar GF, van Rossum-Fikkert S, van Kesteren M, Goosen N. Cho, a second endonuclease involved in Escherichia coli nucleotide excision repair. Proc Natl Acad Sci U S A. 2002;99(3):1467–72. pmid:11818552
    • 36. Kamarthapu V, Nudler E. Rethinking transcription coupled DNA repair. Current Opinion in Microbiology. 2015;24:15–20.
    • 37. Bharati BK, Gowder M, Zheng F, Alzoubi K, Svetlov V, Kamarthapu V, et al. Crucial role and mechanism of transcription-coupled DNA repair in bacteria. Nature. 2022;604(7904):152–9. pmid:35355008
    • 38. Fanous J, Claudi B, Tripathi V, Li J, Goormaghtigh F, Bumann D. Limited impact of Salmonella stress and persisters on antibiotic clearance. Nature. 2025;639(8053):181–9. pmid:39910302
    • 39. Matson SW, Robertson AB. The UvrD helicase and its modulation by the mismatch repair protein MutL. Nucleic Acids Res. 2006;34(15):4089–97. pmid:16935885
    • 40. Nudler E. Transcription-coupled global genomic repair in E. coli. Trends Biochem Sci. 2023;48(10):873–82. pmid:37558547
    • 41. Simmons LA, Foti JJ, Cohen SE, Walker GC. The SOS regulatory network. Cellular and Molecular Biology of E coli, Salmonella, and the Enterobacteriaceae. 2008;3(1).
    • 42. Cox MM, Goodman MF, Keck JL, van Oijen A, Lovett ST, Robinson A. Generation and Repair of Postreplication Gaps in Escherichia coli. Microbiol Mol Biol Rev. 2023;87(2):e0007822. pmid:37212693
    • 43. Hickson ID, Arthur HM, Bramhill D, Emmerson PT. The E. coli uvrD gene product is DNA helicase II. Mol Gen Genet. 1983;190(2):265–70. pmid:6135974
    • 44. Caron PR, Kushner SR, Grossman L. Involvement of helicase II (uvrD gene product) and DNA polymerase I in excision mediated by the uvrABC protein complex. Proc Natl Acad Sci U S A. 1985;82(15):4925–9. pmid:3161077
    • 45. Sakai A, Cox MM. RecFOR and RecOR as distinct RecA loading pathways. J Biol Chem. 2009;284(5):3264–72. pmid:18986990
    • 46. Markham BE, Little JW, Mount DW. Nucleotide sequence of the lexA gene of Escherichia coli K-12. Nucleic Acids Res. 1981;9(16):4149–61. pmid:6272195
    • 47. McKenzie GJ, Craig NL. Fast, easy and efficient: site-specific insertion of transgenes into enterobacterial chromosomes using Tn7 without need for selection of the insertion event. BMC Microbiol. 2006;6:39. pmid:16646962
    • 48. Hall MC, Matson SW. Mutation of a highly conserved arginine in motif IV of Escherichia coli DNA helicase II results in an ATP-binding defect. J Biol Chem. 1997;272(30):18614–20. pmid:9228029
    • 49. Lee JY, Yang W. UvrD helicase unwinds DNA one base pair at a time by a two-part power stroke. Cell. 2006;127(7):1349–60.
    • 50. Kawale AA, Burmann BM. UvrD helicase-RNA polymerase interactions are governed by UvrD’s carboxy-terminal Tudor domain. Commun Biol. 2020;3(1):607. pmid:33097771
    • 51. Sanders K, Lin C-L, Smith AJ, Cronin N, Fisher G, Eftychidis V, et al. The structure and function of an RNA polymerase interaction domain in the PcrA/UvrD helicase. Nucleic Acids Res. 2017;45(7):3875–87. pmid:28160601
    • 52. Epshtein V, Kamarthapu V, McGary K, Svetlov V, Ueberheide B, Proshkin S, et al. UvrD facilitates DNA repair by pulling RNA polymerase backwards. Nature. 2014;505(7483):372–7. pmid:24402227
    • 53. Adikesavan AK, Katsonis P, Marciano DC, Lua R, Herman C, Lichtarge O. Separation of recombination and SOS response in Escherichia coli RecA suggests LexA interaction sites. PLoS Genet. 2011;7(9):e1002244. pmid:21912525
    • 54. Tang J, Herzfeld AM, Leon G, Brynildsen MP. Differential impacts of DNA repair machinery on fluoroquinolone persisters with different chromosome abundances. mBio. 2024;15(5):e0037424. pmid:38564687
    • 55. Tang J, Brynildsen MP. Genome-wide mapping of fluoroquinolone-stabilized DNA gyrase cleavage sites displays drug specific effects that correlate with bacterial persistence. Nucleic Acids Res. 2023;51(3):1208–28. pmid:36631985
    • 56. Chaudhury AM, Smith GR. A new class of Escherichia coli recBC mutants: implications for the role of RecBC enzyme in homologous recombination. Proc Natl Acad Sci U S A. 1984;81(24):7850–4. pmid:6393130
    • 57. Amundsen SK, Taylor AF, Chaudhury AM, Smith GR. recD: the gene for an essential third subunit of exonuclease V. Proc Natl Acad Sci U S A. 1986;83(15):5558–62. pmid:3526335
    • 58. Myers RS, Kuzminov A, Stahl FW. The recombination hot spot chi activates RecBCD recombination by converting Escherichia coli to a recD mutant phenocopy. Proc Natl Acad Sci U S A. 1995;92(14):6244–8. pmid:7603978
    • 59. Jockovich ME, Myers RS. Nuclease activity is essential for RecBCD recombination in Escherichia coli. Mol Microbiol. 2001;41(4):949–62. pmid:11532156
    • 60. Amundsen SK, Smith GR. RecBCD enzyme: mechanistic insights from mutants of a complex helicase-nuclease. Microbiol Mol Biol Rev. 2023;87(4):e0004123. pmid:38047637
    • 61. Huang S-YN, Michaels SA, Mitchell BB, Majdalani N, Vanden Broeck A, Canela A, et al. Exonuclease VII repairs quinolone-induced damage by resolving DNA gyrase cleavage complexes. Sci Adv. 2021;7(10):eabe0384. pmid:33658195
    • 62. Nakano T, Morishita S, Katafuchi A, Matsubara M, Horikawa Y, Terato H, et al. Nucleotide excision repair and homologous recombination systems commit differentially to the repair of DNA-protein crosslinks. Mol Cell. 2007;28(1):147–58. pmid:17936711
    • 63. Salem AMH, Nakano T, Takuwa M, Matoba N, Tsuboi T, Terato H, et al. Genetic analysis of repair and damage tolerance mechanisms for DNA-protein cross-links in Escherichia coli. J Bacteriol. 2009;191(18):5657–68. pmid:19617358
    • 64. Fang Q. DNA-protein crosslinks processed by nucleotide excision repair and homologous recombination with base and strand preference in E. coli model system. Mutat Res. 2013;741–742:1–10. pmid:23500083
    • 65. Orman MA, Brynildsen MP. Establishment of a method to rapidly assay bacterial persister metabolism. Antimicrob Agents Chemother. 2013;57(9):4398–409. pmid:23817376
    • 66. Aldred KJ. Bacterial RADAR Assay: Isolation and Detection of Intracellular Covalent Complexes. Methods Mol Biol. 2025;2928:17–23. pmid:40372633
    • 67. Aedo SJ, Orman MA, Brynildsen MP. Stationary phase persister formation in Escherichia coli can be suppressed by piperacillin and PBP3 inhibition. BMC Microbiol. 2019;19(1):140. pmid:31234796
    • 68. Morel P, Hejna JA, Ehrlich SD, Cassuto E. Antipairing and strand transferase activities of E. coli helicase II (UvrD). Nucleic Acids Res. 1993;21(14):3205–9. pmid:8341594
    • 69. Veaute X, Delmas S, Selva M, Jeusset J, Le Cam E, Matic I, et al. UvrD helicase, unlike Rep helicase, dismantles RecA nucleoprotein filaments in Escherichia coli. EMBO J. 2005;24(1):180–9. pmid:15565170
    • 70. Lestini R, Michel B. UvrD controls the access of recombination proteins to blocked replication forks. EMBO J. 2007;26(16):3804–14. pmid:17641684
    • 71. Buljubašić M, Hlevnjak A, Repar J, Đermić D, Filić V, Weber I, et al. RecBCD- RecFOR-independent pathway of homologous recombination in Escherichia coli. DNA Repair (Amst). 2019;83:102670. pmid:31378505
    • 72. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, et al. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2006;2:2006.0008. pmid:16738554
    • 73. Thomason LC, Costantino N, Court DL. E. coli genome manipulation by P1 transduction. Curr Protoc Mol Biol. 2007;Chapter 1:1.17.1-1.17.8. pmid:18265391
    • 74. Cherepanov PP, Wackernagel W. Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene. 1995;158(1):9–14. pmid:7789817
    • 75. Leon G, Brynildsen MP. Conservation and divergence of ciprofloxacin persister survival mechanisms between Pseudomonas aeruginosa and Escherichia coli. PLoS Genet. 2025;21(9):e1011840. pmid:40892829
    • 76. Gibson DG, Young L, Chuang R-Y, Venter JC, Hutchison CA 3rd, Smith HO. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat Methods. 2009;6(5):343–5. pmid:19363495
    • 77. Martinez B, Bharati BK, Epshtein V, Nudler E. Pervasive Transcription-coupled DNA repair in E. coli. Nat Commun. 2022;13(1):1702. pmid:35354807