A nucleoid-associated protein is involved in the emergence of antibiotic resistance by promoting the frequent exchange of the replicative DNA polymerase in Mycobacterium smegmatis

ABSTRACT Antibiotic resistance in Mycobacterium tuberculosis exclusively originates from chromosomal mutations, either during normal DNA replication or under stress, when the expression of error-prone DNA polymerases increases to repair damaged DNA. To bypass DNA lesions and catalyze error-prone DNA synthesis, translesion polymerases must be able to access the DNA, temporarily replacing the high-fidelity replicative polymerase. The mechanisms that govern polymerase exchange are not well understood, especially in mycobacteria. Here, using a suite of quantitative fluorescence imaging techniques, we discover that in Mycobacterium smegmatis, as in other bacterial species, the replicative polymerase, DnaE1, exchanges at a timescale much faster than that of DNA replication. Interestingly, this fast exchange rate depends on an actinobacteria-specific nucleoid-associated protein (NAP), Lsr2. In cells missing lsr2, DnaE1 exchanges less frequently, and the chromosome is replicated more faithfully. Additionally, in conditions that damage DNA, cells lacking lsr2 load the complex needed to bypass DNA lesions less effectively and, consistently, replicate with higher fidelity but exhibit growth defects. Together, our results show that Lsr2 promotes dynamic flexibility of the mycobacterial replisome, which is critical for robust cell growth and lesion repair in conditions that damage DNA. IMPORTANCE Unlike many other pathogens, Mycobacterium tuberculosis has limited ability for horizontal gene transfer, a major mechanism for developing antibiotic resistance. Thus, the mechanisms that facilitate chromosomal mutagenesis are of particular importance in mycobacteria. Here, we show that Lsr2, a nucleoid-associated protein, has a novel role in DNA replication and mutagenesis in the model mycobacterium Mycobacterium smegmatis. We find that Lsr2 promotes the fast exchange rate of the replicative DNA polymerase, DnaE1, at the replication fork and is important for the effective loading of the DnaE2-ImuA′-ImuB translesion complex. Without lsr2, M. smegmatis replicates its chromosome more faithfully and acquires resistance to rifampin at a lower rate, but at the cost of impaired survival to DNA damaging agents. Together, our work establishes Lsr2 as a potential factor in the emergence of mycobacterial antibiotic resistance.

pathogen Mycobacterium tuberculosis is exclusively driven by chromosomal mutagene sis, as M. tuberculosis has no ability for horizontal gene transfer (2,3).Thus, the mechanisms regulating polymerase expression and activity are of particular importance for mycobacterial evolution.
In mycobacteria, the fidelity of chromosome replication is established by the major replicative polymerase DnaE1, which possesses endogenous proofreading activity (4).In addition to DnaE1, mycobacteria encode a suite of other polymerases with specialized repair functions that are only expressed under certain conditions, such as oxidative or DNA-damaging stress (5)(6)(7).For example, the mycobacterial specialist polymerases DinB1, DinB2, DinB3, and DnaE2 are induced under different conditions and have varying repair activities and mutational signatures (5)(6)(7)(8)(9)(10).Of particular importance is the error-prone polymerase DnaE2.It and its accessory proteins ImuA' and ImuB-collec tively referred to as the "mutasome"-are highly expressed in conditions that damage DNA and are the primary drivers of stress-induced mutagenesis in vitro and potentially in vivo (5,6,11).However, the mechanisms that facilitate switching between polymerases at the site of repair are not well understood, especially in mycobacteria (12).
It is known that in non-DNA damaging conditions, in Escherichia coli, components within the large multi-protein complex that replicates DNA (the replisome) exchange rapidly, likely when encountering DNA-protein obstacles.More specifically, fluorescence recovery after photobleaching (FRAP) and single molecule studies have revealed that, despite being highly processive in vitro, replisome components are highly dynamic within the cell (13).For example, at room temperature, the alpha subunit of Pol III in E. coli exchanges with the cytoplasmic pool at a timescale of approximately 4 seconds, suggesting that subunits are constantly being replaced during the 150 minutes it takes to replicate the chromosome at this temperature (13).These data underlie the notion that the replisome is dynamic and flexible, allowing new subunits the opportunity to exchange frequently and suggest that protein obstacles influence the exchange rate of replisome components.
Mycobacteria differ from model bacteria like E. coli in several ways.Most notable is their slow growth and long doubling times.Pathogenic mycobacteria, like M. tuberculosis and M. leprae, are notorious for their long doubling times, which can vary from 18 hours to several weeks.Even the so-called fast growers like Mycobacterium smegmatis double on the hour-long timescale compared with the minute-long timescale of commonly studied organisms like E. coli.Intriguingly, the duration of chromosomal replication at least partially scales with the growth rate: M. smegmatis exhibits a longer duration of chromosome replication than E. coli (~100-150 minutes vs ~41 minutes, respectively) despite the ability of DnaE1 to polymerize DNA at a faster rate than E. coli Pol IIIa in vitro (4,14,15).
These observations led us to wonder if components within the replisome exchange at a different timescale in mycobacteria than in faster-growing bacteria.Using various fluorescence imaging techniques, we find that, as in E. coli, the replicative polymer ase is exchanged rapidly in M. smegmatis.Surprisingly, this fast exchange rate is, in part, mediated by an actinobacteria-specific DNA-binding protein Lsr2, as DnaE1 is exchanged more slowly in cells missing lsr2.Interestingly, when the DNA is damaged, ImuB, a key scaffolding protein that recruits the error-prone polymerase into an active complex, is exchanged more rapidly, suggesting ineffective loading of the mutasome.Consistently, loss of lsr2 results in growth defects in conditions that damage DNA and reduced mutagenesis.Overall, our data deepen our understanding of DNA metabolism in mycobacteria and provide another unexpected layer to the mechanisms mycobacteria use to mutate and acquire drug resistance.

The mycobacterial replisome is highly dynamic
To investigate the dynamics of DnaE1, we created a strain expressing DnaE1-mScarlet from its native promoter.A similar fusion encodes functional DnaE1 (16).To verify that DnaE1-mScarlet reports on the localization of the replisome, we performed time-lapse microscopy and analyzed fluorescence distributions as a function of both cell length and cell cycle (Fig. 1A and B).Consistent with the reported placement of the replisome in mycobacteria, DnaE1 localizes at the approximately mid-cell before disappearing, presumably after DNA replication has been completed (16).Toward the end of the cell cycle, DnaE1 reappears at one-and three-quarter positions to initiate replication in newly synthesized sister chromosomes (Fig. 1A and B).These data show that DnaE1-mScarlet is reporting on the localization of the replisome in M. smegmatis.
To test the exchange rate of DnaE1, we took the same approach as Beattie et al. and used FRAP to measure the dynamics of DnaE1 recovery in cells (13).Fast recovery of fluorophores into a bleached spot indicates fast dissociation of bleached molecules.Conversely, slow or incomplete recovery of non-bleached fluorophores indicates slow dissociation of bleached molecules and a stably associated complex.To account for differences in the degree of bleaching per cell, we normalized the fluorescence intensity after bleaching to zero and the final intensity to the maximum recovery we expec ted based on the percentage of fluorophores bleached.We fit the data to a diffusion exchange reaction and calculated both the timescale of recovery, reflecting the dissociation time of bound subunits, and the maximum recovery, reflecting the mobile fraction of proteins.Imaging was done in the same conditions as above, in which we measured the duration of chromosomal replication to be approximately 100 minutes based on the time between the appearance and disappearance of DnaE1 (Fig. 1B; Fig. S1).We found that, as in E. coli, DnaE1 exchanged quickly and nearly completely with a timescale of approximately 8 ± 3 seconds (mean ± SD) (Fig. 2A through C).Prior studies investigating the exchange of proteins at the replication fork have found that protein concentration positively correlates with its exchange rate (inverse of bound time) (17).Thus, as we employed a strain that expresses a second copy of DnaE1, the values we report here may represent lower bounds.Nevertheless, these data are consistent with the model that has emerged in E. coli, which suggests that the replisome constantly encounters obstacles that promote the exchange of individual subunits with the cytoplasmic pool.

Lsr2 promotes the frequent exchange of DnaE1
In bacteria, nucleoid-associated proteins (NAPs) are important for bacterial chromosome organization and influence replication, transcription, and cell cycle regulation either directly or indirectly.Lsr2 is one of the most well-studied NAPs in mycobacteria.Lsr2 is conserved across mycobacteria and related actinomycetes (18)(19)(20)(21)(22). Like H-NS in E. coli, Lsr2 primarily binds AT-rich regions of DNA and functions as a global transcriptional repressor (19).Notably, in vitro studies show that Lsr2 binding interferes with transcrip tion, inhibits topoisomerase I-dependent supercoil relaxation, and protects DNA from DNase I digestion and H 2 O 2 -mediated degradation (18).Loss of lsr2 is associated with a wide range of different phenotypes in mycobacteria, including altered cell cycle timing and growth defects upon entering/exiting hypoxic conditions and when exposed to oxidative stress and various antibiotics (18,(23)(24)(25)(26). Lsr2 is also involved in resistance against phage infection and protection against reactive oxygen intermediates during macrophage infection, and cells missing lsr2 show reduced virulence in various hosts, including mice (21,23,24,27).
ChIP-seq data show an abundance of Lsr2 binding sites across the chromosome, and consistent with this, Lsr2 decorates the entire nucleoid in discrete and highly dynamic puncta (25,26).Interestingly, the major Lsr2 foci are localized near, but not co-localized with, the replisome (26).We verified this localization pattern by creating strains that express Lsr2-mScarlet and DnaE1-mGFPmut3.Expression of Lsr2-mScarlet in Δlsr2 restores normal cell morphology, showing that it encodes a functional protein (Fig. S2A).Consistent with prior reports, Lsr2 and the replisome exhibit distinct localizations, and although Lsr2 localizes nearby DnaE1 at certain times in the cell cycle, Lsr2 does not co-localize with DnaE1 despite Lsr2's abundance throughout nucleoid (Fig. S2B and  C) (26).
Given the importance of Lsr2 to mycobacterial physiology and its localization, we wondered if deletion of lsr2 would influence replisome exchange dynamics.While Lsr2 is not known to affect the expression of DNA metabolism genes, we first profiled the transcriptome of both wild type and Δlsr2 cells in our conditions.Consistent with prior studies, we could not detect changes in the abundance of transcripts that encode proteins known to be at the replication fork (Table S1-S11).Next, to investigate the effect of Lsr2 on polymerase exchange, we deleted lsr2 in our strain expressing DnaE1-mScarlet.Consistent with prior findings using a DnaN reporter (26), we find that DnaE1 fluorescence disappears from the mid-cell region faster in cells missing lsr2, suggesting that DNA replication is faster by approximately 20% (Fig. S1A and B).Although by RNA-sequencing (RNA-seq) we did not observe differences in the expression of DnaE1 or any other DNA metabolism genes, we observed a small (~50%) increase in DnaE1-mScar let fluorescence in Δlsr2 (Fig. S1C, Table S1-11).To understand how or if this contributed to DnaE1 dynamics at the replication fork, as before, we measured the rate of DnaE1 recovery in Δlsr2 cells using FRAP (Fig. 2A).If changes to DnaE1 exchange were due to the slight increase in DnaE1 concentration, we expected to observe faster exchange in Δlsr2.Instead, we observed a longer recovery time (38 ± 10 seconds) compared with wild type cells (Fig. 2B and C).These data suggest that DnaE1 remains associated with the replisome for longer in cells missing lsr2.Alternatively, a longer recovery time in the lsr2 mutant could be due to reduced DnaE1 access to the replisome (e.g., due to increased replisome access by other polymerases).However, reduced access by DnaE1, an essential polymerase in M. smegmatis, would likely result in a growth rate difference in the mutant, which we and others do not observe (Fig. S1D) (26).Likewise, reduced access by DnaE1 would also increase basal mutation rates, as this polymerase is responsible for faithful chromosome replication in standard laboratory medium (4).Instead, we observe that Δlsr2 cells mutate at decreased frequency as measured by spontaneous resistance to rifampin (Fig. 2D).The magnitude of the decrease in mutation was similar to the increase in DnaE1-bound time (four-to five-fold), further supporting the notion of increased stability of DnaE1 in Δlsr2.Taken together, these data show that Lsr2 promotes the rapid exchange of DnaE1.In its absence, DnaE1 is more stably associated, and cells replicate their chromosome more quickly and more faithfully.

Lrs2 is important for the acquisition of mutations by promoting the stable formation of the mutasome
We next hypothesized that loss of lsr2 would affect polymerase exchange, mutagenesis, and bacterial fitness in conditions that damage DNA and increase the expression of the error-prone polymerase DnaE2 and its associated complex.As Lsr2 transcriptionally silences the expression of many genes in M. smegmatis, we first asked if Δlsr2 cells have an intact DNA damage response by transcriptional profiling.To test this, we treated cells with mitomycin C (MMC) at 1, 3, 6, and 24 hours, and performed RNA-seq.Across three biological replicates, we could not detect changes in the expression of any gene known to be involved in the DNA damage response at any of the time points (Fig. 3; Tables S1 to  S11).Thus, we conclude that cells missing lsr2 have an intact DNA damage response.
Next, to test polymerase exchange, we created strains carrying fluorescent protein fusions to ImuA' , ImuB, and DnaE2 integrated either at a phage site (imuA' and imuB in ΔimuA'B) or at the chromosomal locus (dnae2).When exposed to UV light, these strains acquired resistance to rifampin at a similar rate as the wild type, showing that the fusion proteins retain the ability to repair damaged DNA (Fig. S3) (28).ImuB recruits ImuA' and DnaE2 into an active complex by binding to the β-clamp subunit of the replisome.ImuA' and DnaE2 bind ImuB but not the β-clamp and therefore require ImuB to associate with the replisome and catalyze mutations (8,11,28,29).Despite being co-transcribed, ImuA'-mScarlet and mScarlet-ImuB had dramatically different fluorescence intensities: ImuA'-mScarlet was barely detectable above background fluorescence levels, while mScarlet-ImuB was brightly fluorescent (Fig. S4A and B).To confirm that this was not due to cleavage of the fluorescent protein, we conducted a Western blot analysis against the C-terminal FLAG epitope we inserted at the 5′ end of mScarlet (Fig. S5).Consistent with the low fluorescence we observed by microscopy, we could barely detect ImuA'-mScarlet.Interestingly, we observed two bands-one at the predicted full-length size of ImuA'-mScarlet (51 kDa) and one approximately 5 kDa smaller.Importantly, we could not detect any bands corresponding to the size of free mScarlet, suggesting that cleavage of the fluorescent protein was not the reason for the dim fluorescence we observed by microscopy.Thus, cellular levels of ImuA' are likely much lower than ImuB, implying that additional post-transcriptional mechanisms exist (e.g., translational regulation or proteolysis) to limit ImuA' abundance.Consistent with the low abundance of ImuA' , we also observed very low fluorescence of DnaE2-mGFPmut3 (Fig. S4C).Others have reported similar results using N-terminal fusions (28).These data suggest that as in E. coli, the translesion polymerase abundance may be tightly regulated post-transcriptionally (30).Further, these data suggest that one limiting factor in forming an active mutasome complex may be the low cellular concentrations of ImuA' and/or DnaE2.
Due to the dim fluorescence of ImuA' and DnaE2, it was difficult to draw conclusions about their localization or dynamics in cells with or without lsr2.Thus, we decided to determine the exchange rate of ImuB.We treated cells with a low concentration of mitomycin C to induce expression of ImuB and, as before, performed FRAP.In wild type cells, ImuB displayed a much longer timescale of exchange (74 ± 14 seconds) compared with DnaE1, suggesting ImuB remains associated with the β-clamp for a longer time (Fig. 4A and B).The long bound time of ImuB is consistent with the need to recruit low-abundance proteins ImuA' and/or DnaE2.In cells missing lsr2, ImuB exchanged more quickly (46 ± 6 seconds) than in the wild type, at a rate similar to DnaE1-mScarlet (Fig. 2C).Thus, as ImuB and DnaE1 both bind the β-clamp, these data suggest that the mutasome complex is less stably associated in cells missing lsr2, possibly due to the increased stability of DnaE1.
We reasoned that increased stability of DnaE1 would again manifest in the ability of Δlsr2 cells to acquire mutations and that the mutasome would be less effective in these cells.Thus, we assayed the ability of Δlsr2 to acquire resistance to rifampin when exposed to UV light in the presence and absence of imuA'B.We chose a dose of UV that resulted in little survival difference between wild type and Δlsr2 (Fig. S6A).Consistent with our model, cells deleted for lsr2 were approximately 4-5 times less likely to acquire resistance to rifampin when exposed to UV (Fig. S6B) compared with either wild type or comple mented cells, similar to the decreased ability of these cells to mutate in basal conditions (Fig. 2D).Additionally, deleting imuA'B had a smaller effect on mutation frequency in Δlsr2 (17-fold less likely to acquire mutations) than in the wild type (60-fold less likely to acquire mutations) (Fig. 5A).Together, these results suggest that cells missing lsr2 have an intact DNA damage response but that its effect is partly blunted by the increased DNA replication fidelity due to a stably associated DnaE1.In other words, a maximal DNA damage response as mediated by the ImuA'B-associated complex requires Lsr2.
As DnaE1 is a faithful enzyme but unable to repair lesions, we hypothesized that cells missing lsr2 would have a growth defect when exposed to conditions that damage DNA.To test this, we grew cells on agar plates with different concentrations of mitomycin C. Consistent with our hypothesis, deletion of lsr2 results in a severe growth defect in the presence of MMC, which could be complemented by the expression of lsr2 from M. smegmatis or M. tuberculosis (Fig. 5B; Fig. S7A).While deletion of imuA'B results in a small survival defect on MMC plates, the effect of deleting lsr2 was much greater, suggesting that increased residence of the replicative polymerase is highly detrimental in conditions that result in DNA lesions.Deletion of imuA'B in Δlsr2 appeared to have little to no effect, suggesting that ImuA'B's role in survival to MMC depends on Lsr2 (Fig. 5C; Fig. S7B).Taken together, our data show that Lsr2 is involved in the exchange of polymerases at the replication fork in both standard culture medium and conditions that damage DNA and upregulate repair pathways.This function is needed for M. smegmatis to thrive in conditions that damage DNA and to effectively acquire mutants that result in drug resistance.

DISCUSSION
Lsr2 acts as a transcriptional repressor for many genes on the M. smegmatis and M. tuberculosis chromosomes (18,(23)(24)(25).Additionally, multiple lines of evidence suggest that Lsr2 has other roles aside from its function as a transcription factor (23,26,27).Our results implicate Lsr2 as a potential factor in mycobacterial mutagenesis and antibiotic resistance (Fig. 6).
While Lsr2 binds to and/or alters the transcription of many genes in both M. smegmatis and M. tuberculosis, very few phenotypes associated with the deletion mutant have been connected to altered expression of target genes (18,23,25).One of the few exceptions is the change in colony morphology, which can be attributed to the increased expression of a putative polyketide synthase (MSMEG_4727) in M. smegmatis (25,31).Why, then, is Lsr2 associated with so many bacterial behaviors, including bacterial growth, the cell cycle, robust infection, survival to hydrogen peroxide and antibiotics, survival in low oxygen, and resistance to phage?In conditions that do not induce DNA damage, these polymerases may include DinB1, which contributes to spontaneous mutagenesis (5).In conditions that damage DNA, the mutasome, which is comprised of DnaE2 and its accessory proteins (ImuA' and ImuB), is induced and needs to exchange with the replicative polymerase to repair lesions (6,11,12).Deletion of Lsr2 reduces the flexibility of the replisome needed for sufficient exchange, thereby resulting in cells that are more reliant on DnaE1, a higher fidelity polymerase that is unable to bypass lesions.Thus, cells missing lsr2 are impaired in their ability to grow in conditions that damage DNA and in their ability to acquire mutations.
Our results indicate that one of Lsr2's functions in the cell is to promote the flexibility of the mycobacterial replisome, enabling specialist polymerases like the DnaE2 translesion complex (ImuB/DnaE2/ImuA') to access the replication fork, leading to efficient, but error-prone, DNA replication and cell growth in conditions that damage DNA.Importantly, we cannot rule out the possibility that Lsr2's effect on the repli some is mediated by another protein whose expression is controlled by Lsr2.While our transcriptional profiling did not detect changes in expression for genes involved in DNA metabolism or repair, Lsr2 could be controlling the expression of a gene of unknown function that is important for the phenotypes we observe here.Nevertheless, our findings support a model whereby in non-DNA-damaging conditions, Lsr2 (or a protein whose expression is dependent on Lsr2) serves as an obstacle that promotes the frequent exchange of DnaE1, allowing other polymerases to exchange into the replisome (Fig. 6).This model is consistent with Δlsr2 having a reduced basal mutation rate, as Lsr2 could be allowing exchange of polymerases besides DnaE1 at a low frequency.Likewise, in DNA-damaging conditions, our data suggest a model in which Lsr2-medi ated exchange of DnaE1 promotes effective loading of ImuB and thus DnaE2 and ImuA' .DnaE1 is a more faithful enzyme but is unable to bypass lesions; thus, cells missing lsr2 grow more slowly under conditions that damage DNA but replicate with higher fidelity.
Numerous studies in eukaryotes and other bacteria show that the rate of replication fork progression is actively regulated, in some cases by machinery that slow down DNA replication to ensure coordination among DNA replication, DNA repair, and transcription (32)(33)(34)(35).We hypothesize that Lsr2 may regulate replication fork progression in myco bacteria and that this function may provide a unifying explanation for many of the phenotypes associated with loss of lsr2, including slower DNA replication in the host and other environmental conditions likely to assault DNA.Likewise, in conditions that do not significantly damage DNA, like those found in laboratory growth media, loss of lsr2 leads to faster and higher fidelity chromosomal replication, possibly explaining the emergence of spontaneous lsr2 mutants in several laboratories (31,36,37).Additionally, loss of lsr2 is associated with increased resistance against infection by certain phages and fewer and smaller zones of phage DNA replication (27).We speculate that efficient turnover of the bacterial host DNA replication machinery mediated by Lsr2 could be required for robust replication of phage DNA and productive infection.
Intriguingly, Lsr2 is phosphorylated by PknB, a plasma membrane-bound eukaryotic serine/threonine kinase involved in cell growth and division (38).Given the effect of Lsr2 phosphorylation on DNA binding (38) and our results showing Lsr2's role in mediating the dynamics of the replisome over the cell cycle , we hypothesize that such regulation could be used to coordinate cell wall synthesis and DNA synthesis during environmental changes or even during a typical cell cycle.
Lsr2 may be a promising drug target for preventing the emergence of antibiotic resistance in mycobacteria.Our work adds to the growing body of evidence that Lsr2 is at the nexus of several evolutionary tradeoffs, including fast chromosome replication, mutagenesis, replisome flexibility, and phage resistance.Indeed, loss of Lsr2 results in severe growth defects in M. tuberculosis and Mycobacterium abscessus in vivo (21,24).Thus, targeting Lsr2 therapeutically may help treat drug-susceptible infections, while preventing the emergence of drug resistance.
Genetic modifications to the native locus were generated as previously described using the mycobacterial Che9c phage RecET recombination system (39).To construct a knockout, we designed a dsDNA substrate encoding a zeocin resistance cassette flanked by loxP sites and ~500 bp of homology on either end of the locus of interest.The dsDNA substrate was then electroporated into M. smegmatis strains carrying a plasmid encoding isovaleronitrile-inducible expression of RecET.Colonies were selected by zeocin resistance and screened by PCR using primers annealing to regions flanking dsDNA insertion.In some cases, the zeocin resistance cassette was removed by Cre-Lox recombination.Strains and primers are listed in Tables S12 and S13, respectively.

Fluorescence microscopy and analysis
All microscopy experiments were performed on a Nikon TI-E inverted wide-field microscope equipped with the Nikon Perfect Focus System, an ORCA-Flash4.0Digital CMOS Camera (Hamamatsu) and an environmental control chamber maintained at 37°C.

Time-lapse and kymograph analysis
Time-lapse imaging was done using a 60× Oil 1.45NA Plan Apochromat phase-contrast objective.Bacterial samples were loaded and grown in a B04A microfluidic plate (CellAsic ONIX, B04A-03-5PK).Phase and fluorescence images were acquired every 30 minutes using NIS Elements software.Fluorescence images were taken sequentially using a 470-nm excitation laser and 515/30 emission filter for msfGFP and mGFPmut3 and a 555-nm excitation laser and a 595/40 emission filter for mScarletI.
Kymograph analysis was performed using a custom pipeline for segmenting and tracking cells and extracting fluorescence profiles.Cell segmentation was performed on the phase image using a U-Net network trained for segmenting mycobacterial cells (40).Semi-automated cell tracking and extraction of fluorescence profiles from backgroundsubtracted images were done using a custom MATLAB program and open-source image analysis software Fiji (41).Average kymographs were generated by 2D interpolation of at least 20 individual kymographs.

FRAP
FRAP experiments were done using a 60× oil DIC objective.Bacterial cells were imaged under 7H9 agar pads [1% agar, 0.2% (vol/vol) glycerol] in 50-mm glass-bottomed dishes.Fluorescent spots were manually selected for localized bleaching with a focused 405-nm laser.Two pre-bleach fluorescence images were acquired: one to identify and center a target focus and a second for quantification of total pre-bleach fluorescence.Fluores cence recovery was monitored by acquisition at pre-defined intervals (2 to 10 seconds) following photobleaching using an exposure time of 200 ms.
Cell outlines for determination of total pre-bleach fluorescence were manually drawn using the brightfield image.Fluorescence images were background subtracted and corrected for photobleaching.To quantify the fluorescence of bleached foci over time, a circular region-of-interest (ROI) was manually drawn around the bleached foci and intensity profiles over time were obtained using Fiji (41).Fluorescence recovery curves were generated by first normalizing fluorescence intensity profiles to pre-bleach focus fluorescence.The post-bleach value was then subtracted from all time points.Maximum recovery was estimated by calculating the total amount of fluorescence in the cell after photobleaching outside the focus.Finally, all time points were divided by the estimated maximum recovery value, such that a value of 0 and 1 corresponds to no recovery and maximum recovery, respectively.

Snapshot imaging
Cells were immobilized under 7H9 agar pads containing 1% UltraPure Agarose (Thermo Fisher, #16500500).Phase and fluorescence images were generated using NIS Elements software.Fluorescence images were taken sequentially using a 470-nm excitation laser and 515/30 emission filter for msfGFP and mGFPmut3 and a 555-nm excitation laser and a 595/40 emission filter for mScarletI.Fluorescence images were background subtracted.Cell lengths and fluorescence intensities were obtained using MicrobeJ (42).

RNA-seq of WT M. smegmatis and Δlsr2
WT M. smegmatis and Δlsr2 were grown in triplicate at 37°C to optical density at 600 nm (OD600) ~0.7-1.0.For mitomycin C treatment, 1 mL of cells was added to 4 mL of 7H9 containing 100 ng/mL mitomycin C for a final concentration of 80 ng/mL mitomycin C. Approximately 2-4 mL of cells was harvested at each timepoint and incubated with two volumes of RNAprotect (Qiagen) at room temperature for 5 minutes prior to centrifuga tion for 5 minutes at 4,000 × g at 4°C.The supernatant was discarded, and the cells were then immediately used for RNA extraction or stored at −80°C.
RNA extraction was performed using the Qiagen RNeasy Kit (74104).Cells were resuspended in 500 μL of RLT buffer (Qiagen) with 2-mercaptoethanol and homogenized by bead beating for 3-minute pulses at maximum speed, three times, with 5 minutes of rest on ice in between.The lysate was then centrifuged at maximum speed for 1 minute, and the supernatant was transferred to a new tube.The remainder of the RNA extraction was performed using the Qiagen RNeasy Kit according to manufacturer instructions.RNA quality was verified using a bioanalyzer (Agilent Technologies TapeStation 2200).DNase treatment and ribosomal RNA depletion were performed at the Microbial Genome Sequencing Center before library preparation and sequencing.

Mutation frequency analysis
Ten milliliters of bacterial cells was grown to OD600 ~0.7-1.0.Five milliliters of cells was UV irradiated (20 mJ/cm2 UV; Stratalinker UV crosslinker), while the remaining 5 mL of cells was used for the untreated control.Treated and untreated samples were then added to an equal volume of fresh 7H9 media and outgrown for 3 hours at 37°C.One hundred microliters of liquid culture was used for CFU determination, and the remainder of the culture was plated onto LB agar plates containing rifampin (rif ) (200 μg/mL) for mutant detection.All mutation frequency measurements were performed using eight biological replicates.Rif-resistant colonies were enumerated after 5-7 days of incubation at 37°C.CFU plates were read after about 2-3 days of growth on non-selective media.Mutation frequencies were calculated by the number of Rif-resistant colonies per mL of culture by the CFU/mL.

Spot assay
Bacterial cells were grown to OD600 ~0.7-1.0,normalized to the same OD, and then 10-fold serially diluted.Five microliters of culture was spotted on LB agar containing varying concentrations of antibiotic.Plates were imaged after incubation at 37°C for 2-3 days.All assays were done in biological triplicate.

Western blot analysis
Twenty milliliters of M. smegmatis culture was grown to OD600 ~0.7-1.0,pelleted by centrifugation at 4,000 × g for 10 minutes, and re-suspended in 500 μL of TBS + Roche cOmplete Protease Inhibitor Cocktail.Cells were lysed by bead beating for 45-second pulses, four times, with 2 minutes of rest on ice between pulses.Following lysis, cells were centrifuged at max speed for 5 minutes at 4°C.Supernatants were recovered, and protein concentrations were determined by measuring absorbance at A280.Protein samples were diluted to the same concentration and then labeled using the Amersham QuickStain Protein Labeling Kit according to the manufacturer's instructions.Samples were then heated in Laemmli buffer at 95°C for 10 minutes and resolved on 4%-12% Bis-Tris SDS-PAGE gels by running at 140V for 90 minutes.After electrophoresis, proteins were transferred to a PVDF membrane at 15V for 45 minutes using a semi-dry system.The membrane was blocked for 30 minutes at room temperature in 5% dry milk in Tris-buffered saline containing Tween 20 (TBST; 50 mM Tris, 0.5 M NaCl, and 0.02% Tween 20) and then incubated overnight at 4°C in primary antibody.Anti-FLAG antibody produced in mouse (Sigma F1804-50UG) was used at a 1:1,000 dilution in TBST with 5% milk.Following incubation in primary antibody, the membrane was then washed twice with TBST and then incubated at room temperature for 1 hour in secondary antibody [anti-mouse HRP produced in goat from Thermo Fisher Scientific (A28177)].Finally, the membrane was washed five times in TBST before chemilumines cent development.

FIG 1
FIG 1 DnaE1 localization over time in M. smegmatis.(A) A representative cell expressing DnaE1-mScarlet from its native promoter is imaged over time with both phase and fluorescence microscopy (WN1384).Fluorescent spots appear at mid-cell, disappear, and then reappear at roughly one-and three-quarter positions at the end of one cell cycle/beginning of another.(B) To quantify fluorescence probability distributions over time and space, we constructed kymographs by averaging individual distributions that were normalized to cell cycle time and length (N = 43).

FIG 2
FIG 2 The replicative polymerase is highly dynamic, which is in part mediated by Lsr2.(A) Representative time sequence of FRAP of DnaE1-mScarlet in wild type (WT) (top) and Δlsr2 (bottom) imaged at 2and 4-second intervals, respectively (WN1384, WN1385).Blue arrow indicates the location of the focus targeted for bleaching.(B) Fluorescence recovery curves for DnaE1-mScarlet (N = 21 for WT, 12 for Δlsr2).Solid lines show a one-phase exponential association fit to the data.Individual points show the mean ± SEM of fluorescence intensity after bleach correction.Dashed line represents the estimated maximum possible fluorescence recovery.(C) DnaE1-mScarlet bound times (mean ± SD) obtained from one-phase exponential association fits in panel B for both WT and Δlsr2.****P < 0.0001 by unpaired t-test.(D) Frequency of spontaneous rifampin-resistant mutants per 10 8 CFU in WT, lsr2 knockout, and complement (WN116, WN788, WN906; eight biological replicates).lsr2 complement was expressed from the native lsr2 promoter on a phage integrative plasmid (WN906).Middle horizontal line represents the median.LOD refers to the limit of detection.P values were obtained by one-way analysis of variance (ANOVA).*P < 0.0021 and **P < 0.0002.

FIG 3 FIG 4
FIG 3 Δlsr2 cells have an intact DNA damage response.RNA from three biological triplicates of WT and Δlsr2 was extracted at 0, 1, 3, 6, and 24 hours post-treatment with 80 ng/mL mitomycin C and quantified by RNA sequencing (WN116, WN788).Normalized counts per million (CPM) (mean ± SD) for different DNA damage-inducible genes over time in either the WT (green) and Δlsr2 (blue) are shown.In the imuA'B plot, imuA' is represented by solid lines and imuB by dashed lines.None of the genes shown are differentially expressed in Δlsr2 compared with WT (|log2FC| > 1, P < 0.05).

FIG 6
FIG6 Model of Lsr2 regulation of DNA replication and mutagenesis.Our data suggest that loss of lsr2 results in a decrease in mutagenesis due to a novel role for Lsr2 in modulating replisome dynamics.Specifically, our data are consistent with a model in which Lsr2 acts as a roadblock on the DNA to promote the exchange of the major replicative polymerase DnaE1 with others in the cytoplasmic pool.