Research ArticleCELL BIOLOGY

Heterogeneity of spontaneous DNA replication errors in single isogenic Escherichia coli cells

See allHide authors and affiliations

Science Advances  20 Jun 2018:
Vol. 4, no. 6, eaat1608
DOI: 10.1126/sciadv.aat1608

Abstract

Despite extensive knowledge of the molecular mechanisms that control mutagenesis, it is not known how spontaneous mutations are produced in cells with fully operative mutation-prevention systems. By using a mutation assay that allows visualization of DNA replication errors and stress response transcriptional reporters, we examined populations of isogenic Escherichia coli cells growing under optimal conditions without exogenous stress. We found that spontaneous DNA replication errors in proliferating cells arose more frequently in subpopulations experiencing endogenous stresses, such as problems with proteostasis, genome maintenance, and reactive oxidative species production. The presence of these subpopulations of phenotypic mutators is not expected to affect the average mutation frequency or to reduce the mean population fitness in a stable environment. However, these subpopulations can contribute to overall population adaptability in fluctuating environments by serving as a reservoir of increased genetic variability.

INTRODUCTION

Mutations are the primary source of genetic variation, without which there is no evolution. Therefore, understanding how mutations arise is fundamental to understanding the evolution of living systems. Scientists broadly classify mutations into two categories: spontaneous mutations and mutations induced by environmental stress. While there is ample evidence about the mechanisms of environmental stress–increased mutation rates, it remains obscure how spontaneous mutations arise. Most of the knowledge about the molecular mechanisms controlling spontaneous mutagenesis comes from studies of mutator alleles, which identify pathways that cells use to prevent mutations (1). These include metabolic control over concentrations of endogenous and exogenous mutagens, DNA repair systems, deoxynucleotide triphosphate pool balance, the insertion accuracy of DNA polymerases, DNA polymerases’ 3′→ 5′ exonucleolytic proofreading, and several postreplication systems for repairing mismatches (2, 3). Despite extensive knowledge about these antimutator mechanisms, it is still impossible to answer the following question: Which molecular mechanism is responsible for mutant generation when a mutant arises in the population of bacteria whose mutation-prevention systems are operative?

The most reliable information about the actual sources and mechanisms of spontaneous mutation in cells has come from the studies of antimutator phenotypes. For example, screening for antimutator mutants that reduce spontaneous mutagenesis in carbon-starved, nongrowing, and stationary-phase Escherichia coli populations showed that activation of several stress responses is required for the mutagenic repair of DNA breaks (4). To date, the only well characterized antimutator alleles in growing E. coli cells are mutants of the α subunit of the replicative DNA polymerase III (5), suggesting that DNA replication errors are a major source of spontaneous mutagenesis in E. coli cells growing under optimal conditions, without exogenous stress. However, these DNA polymerase III antimutator alleles also reduce A:T-to-C:G transition mutations in the mutT mutant genetic background by reducing mispairing of 8-oxo-deoxyguanine with adenine (6). Thus, it is unclear whether spontaneous DNA replication errors arise in wild-type cells owing to limits of the intrinsic fidelity of DNA polymerase III, which is determined by base selection and proofreading activities, or in a subpopulation of cells with special phenotypic properties. For example, the cellular amounts of the mutagenic 8-oxo-deoxyguanine, which can increase in some cells, owing to increased reactive oxidative species production, can be incorporated into DNA by DNA polymerase III (7). Ninio (8) also proposed that mutations could arise at a higher-than-average mutation frequency in subpopulations of cells in which transcription or translation errors attenuate DNA replication fidelity transitorily. Although it was shown that the reduction of translation fidelity by mutations or by antibiotic treatment is mutagenic, to the best of our knowledge, this hypothesis was never tested using wild-type cells growing without exogenous stress (9).

This study aimed to determine the impact of phenotypic variability on spontaneous DNA replication errors in populations of isogenic E. coli cells growing under optimal conditions, without exogenous stress. It is not possible to answer this question by using standard mutation assays, which detect phenotypes resulting from mutations, or by whole genome sequencing (10). Both assays detect mutations, which are heritable, but do not provide information on the physiological state of a cell when mutation occurs. In addition, fitness effects of newly arisen mutations affect both assays. To solve these problems, we used a reporter system that allows visualization of the appearance of DNA replication errors in living individual cells (1114). This mutation assay is based on the activity of the E. coli mismatch repair system, which corrects 99.99% of all spontaneous DNA replication errors (15). E. coli mismatch repair involves three dedicated proteins: MutS, MutL, and MutH. The MutS protein detects and binds diverse base pair mismatches. The MutL protein interacts with the MutS-mismatch complex and triggers MutH endonuclease activity at a distal-strand discrimination DNA site. Finally, coordinated helicase/exonuclease activities remove the error from the newly synthesized strand. The mutation assay used in this study is based on the property of the MutL protein to polymerize on the sites of unrepaired mismatches (13). Tagging the MutL protein with a fluorescent protein allows monitoring DNA replication errors, which appear as fluorescent foci, in single living cells. MutL fluorescent foci allow detection of DNA replication errors independently of their future effect on phenotype: beneficial, neutral, deleterious, or even lethal. This mutation assay is also unaffected by inheritance of mutations by the offspring of mutant cells because MutL fluorescent focus disappears when the next DNA replication cycle separates the two DNA strands and mutations are fixed. We coupled the MutL-based mutation assay with different reporters for the growth phase and stress response induction. Results of our study indicate that most spontaneous DNA replication errors in proliferating cells arise in subpopulations of cells suffering from endogenous stresses.

RESULTS

Quantification of DNA replication errors in proliferating cells

We used previously constructed yellow fluorescent protein (YFP)–MutL reporter fusion, but to be able to combine MutL-based mutation reporter with green or yellow fluorescence–based gene expression reporters, we also constructed mCherry fusion to the wild-type mutL gene (table S1). We confirmed that the mCherry-labeled MutL protein retains its normal function, by measuring the frequency of rifampicin-resistant mutants in the strain whose chromosomal mutL gene was deleted and that carried only the labeled MutL derivative (table S2). We also verified that there was no statistically significant difference in the MutL foci frequency between YFP-MutL and mCherry-MutL reporter strains (fig. S1A and table S3). To detect all DNA replication errors, we introduced MutL reporter fusions into a mutH strain. In this strain, which is MutS- and MutL-proficient, mismatches can be detected but cannot be repaired (13). We first examined MutL foci frequency in populations of mutH cells during growth in rich Luria Broth (LB) medium using fluorescence microscopy (Fig. 1A). As expected, there were more DNA replication errors in exponentially growing cells than in cells approaching the stationary phase (Fig. 1, B and C).

Fig. 1 Visualization and quantification of DNA replication errors at a single-cell level.

(A) MutL foci in the mutH strain examined under a microscope during the exponential phase. Each fluorescent focus represents one unrepaired DNA replication error. (B) Growth kinetics of E. coli MG1655 cells. OD600 values of 0.2 (mid-exponential), 0.8 (late exponential), and 1.4 (early stationary) were selected for further analysis. (C) Quantification of DNA replication errors of mutH cells in different growth phases. Cells in the mid-exponential phase had statistically significantly more DNA replication errors than in two other growth phases (see table S3). (D) Quantification of DNA replication errors in different strains during the mid-exponential phase. All strains were statistically significant different from the mutH strain (see table S3).

We decided to follow DNA replication fidelity in cells during the exponential growth phase, that is, cell culture with the optical density at 600 nm (OD600) at 0.2. At this growth phase, cells are metabolically active and at maximum growth rate. First, we confirmed that MutL foci are tagging DNA replication errors by introducing the antimutator allele dnaE911 (5), which enhances the fidelity of the DNA polymerase III (table S2). The dnaE911 allele significantly decreased the frequency of the MutL foci (Fig. 1D and table S3). Second, we confirmed that the detected MutL foci were not artifactual aggregates of the MutL protein fusions independent of mismatches, by inactivating the mutS gene that codes for the mismatch binding protein (13). We did not observe any MutL focus in the mutS mutH mutant cells (Fig. 1D and table S3). Third, because the formation of MutL foci requires the MutS protein, some mismatches may not be tagged by MutL in the cells with an insufficient amount of the MutS protein (16, 17). To test this possibility, we overproduced the MutS protein and observed that the MutL foci frequency increased slightly but significantly (Fig. 1D and table S3). This finding indicated that the amount of MutS protein in some cells was insufficient for efficient mismatch repair detection, thus resulting in increased mutagenesis because of undetected and therefore unrepaired DNA replication errors. The cellular amount of the MutL protein is not limiting in this study because the mutL gene is driven by an inducible promoter (13).

Next, because cell size varies according to growth phase—that is, growing cells are bigger than stationary-phase cells—we examined the correlation between cell size and occurrence of MutL foci. We found that cells with ≥1 MutL foci were bigger than cells with no focus (fig. S1, B and C). For this reason, all data presented hereinafter were normalized to cell size. To further examine the link between growth phase and DNA replication fidelity, we coupled the MutL-based assay with fluorescent transcriptional reporters of the rrnB ribosomal RNA operon. We used fusion of the P1rrnB or P2rrnB operon promoters to the genes coding for green and yellow fluorescent proteins, respectively. P1 promoter activity increases with growth rate, while the P2 promoter is active during transitions, that is, at the point of entry and outgrowth from the stationary phase (18, 19). We found that cells having ≥1 MutL focus had significantly higher P1-driven reporter fluorescence than cells without foci (Table 1). There was no significant difference in P2-driven reporter fluorescence between cells having ≥1 and no MutL foci (Table 1 and fig. S2), whereas there was a positive correlation between the expression levels of the P1-driven reporter and MutL foci (Fig. 2A). Finally, there were fourfold more cells with MutL foci among the top 10% of the population showing strong P1-driven reporter fluorescence when compared to the 10% of the cells that showed weak reporter fluorescence (Fig. 2C and table S3). Together, these data indicated that DNA replication errors occurred more frequently in actively growing cells.

Table 1 Mean fluorescence of different transcriptional reporters among cells with ≥1 and no MutL focus.

We observed a statistically significant difference in the fluorescence between cells with ≥1 and no MutL focus among all reporters except for P2rrnB-YFP.

View this table:
Fig. 2 DNA replication errors in various subpopulations.

We coupled the MutL-based assay with the following fluorescent transcriptional promoter reporters: P1rrnB (A to C), PkatG (D to F), PrecA (G to I), and PibpA (J to L). We plotted the mean fluorescence detected per cell against the number of foci per cell (A, D, G, and J). Heterogeneity of fluorescence detected in isogenic cells (B, E, H, and K). We further examined the cells having the strongest (top 10% of the population) and the weakest (the bottom 10% of the population) fluorescence for the distribution of the number of foci per cell (C, F, I, and L). We observed a statistically significant difference between cells with ≥1 and no MutL foci among all reporters (see Table 1 and table S3). AU, arbitrary units.

DNA replication errors in subpopulations of cells suffering from endogenous stresses

There is growing awareness that there is important cell-to-cell phenotypic heterogeneity in isogenic bacterial populations exposed to the same environmental conditions (20, 21). To reveal subpopulations of cells suffering from endogenous stresses, we used transcriptional reporters for the induction of three stress response regulons. First, we used a recA gene promoter reporter, which is induced by genotoxic stresses as part of the LexA-regulated SOS response (fig. S3) (22). Second, we used the ibpA gene promoter as a reporter for the induction of RpoH-regulated heat shock response (fig. S3) (23). Third, we followed the induction of the katG gene promoter, which is induced as part of the OxyR regulon by hydrogen peroxide (fig. S3) (7). We observed cell-to-cell heterogeneity of the expression of all three reporters (Fig. 2).

Next, by using the MutL-based mutation assay, we examined whether DNA replication errors emerge more frequently in subpopulations of cells experiencing endogenous stresses. We found that cells having ≥1 MutL focus showed significantly higher mean fluorescence for all three transcriptional reporters than cells without MutL foci in the mutH background (Table 1). The number of cells with ≥1 MutL focus in the top 10% of the population showing strong reporter fluorescence was three-, four-, and twofold higher for the SOS, heat shock, and oxidative stress reporters, respectively, than in the 10% of the cells with weak reporter fluorescence (Fig. 2, F, I, and L, and table S3). Therefore, DNA replication errors occur at a higher frequency in subpopulations of endogenously stressed cells.

We observed the strongest increase in DNA replication errors in cells that exhibited induced SOS and heat shock responses. Pennington and Rosenberg (24) showed that SOS response can be induced in subpopulations of E. coli cells growing in stable environmental conditions. It is also well known that several LexA-controlled genes, particularly those coding for translesion synthesis DNA polymerases, increase mutagenesis (22). However, much less is known about DNA replication errors in cells exhibiting heat shock response induction. Heat shock response can be induced by temperature shift or by the accumulation of denatured proteins due to translation errors (25). In our experiments, we incubated cells at constant, optimal growth temperature. Thus, we tested whether an increase in spontaneous translation error rates might be associated with an increase in DNA replication errors.

The impact of the treatments with translation error–inducing antibiotics on DNA replication errors

We first tested whether increased translation errors increase MutL foci frequency, by treating cells with aminoglycoside antibiotics streptomycin and paromomycin (fig. S4). We have chosen aminoglycoside antibiotics because they are known to increase translation errors (26). In addition, Ren et al. (27) also show that streptomycin increases DNA replication errors. We found that treatment of E. coli cells with sub-minimum inhibitory concentrations of both antibiotics significantly increased the frequency of the MutL foci (fig. S4). We also found that the introduction of the dnaE911 antimutator allele abolished streptomycin-induced increase in MutL foci frequency (fig. S5). The fact that there were no MutL foci in mutS-deficient cells showed that streptomycin did not induce aggregation of the MutL protein fusions. Finally, the overproduction of MutS protein did not have any impact on the MutL foci frequency in streptomycin-treated cells. This result indicated that the increase in DNA replication errors was not due to the decreased amount of the MutS protein.

The incidence of the spontaneous translation and DNA replication errors

Therefore, it is clear that decrease in translation fidelity by exogenous stresses results in decrease of DNA replication fidelity. However, we wanted to test whether there are more DNA replication errors in cells having high spontaneous translation errors. Spontaneous translation error rates are at least an order of magnitude higher than transcription and DNA replication error rates (8). Drummond and Wilke estimated that more than 20% of individual proteins in E. coli contain at least one misincorporated amino acid in an average protein length of around 300 amino acid residues (28). These high translation error rates allow monitoring translation fidelity at the single-cell level (29). For this study, we constructed a translation error reporter based on the introduction of a stop codon in the gfp gene. Stop codons cause termination of translation and ribosome detachment from the messenger RNA. However, functional green fluorescent protein (GFP) can be produced not only by translation errors but also by transcription errors. To construct the reporter, which is not affected by transcription errors, we introduced two mutations in the gfp gene generating the UAG stop codon from the lysine AAA codon. Production of functional GFP due to translation errors therefore occurs at the transfer RNA level. Double mutation renders very unlikely that transcription errors produce functional mRNA. We then introduced the mutated GFP reporter in the chromosome with a Ptet promoter (Fig. 3). As a control for the activity of the Ptet promoter, the promoter was coupled with a functional mCherry gene in a nearby locus of the chromosome. We confirmed the validity of the translation error reporter by treating cells with sub-inhibitory concentrations of streptomycin, an antibiotic known to increase mistranslation, as well as by introducing an rpsD mutation, which is known to increase translation errors (30). We also observed cell-to-cell heterogeneity of the spontaneous translation errors (Fig. 3 and fig. S6).

Fig. 3 Translation error reporter and DNA replication errors in cells with different levels of translation errors.

(A) Construct of the chromosomal translation error reporter. We introduced two point mutations in the sfgfp gene, which is under control of the Ptet promoter, creating the UAG stop codon from the lysine AAA codon. As a control, a functional mCherry gene, which is also under control of the Ptet promoter, was added. (B to D) Validation of the translation error reporter. (B) Correlation between the fluorescence signals from mCherry and sfGFP. We found a significant but weak correlation, and we examined 334 cells. (C) The fluorescence signals increased in cells with an rpsD mutation, which is known to increase translation errors, and when cells were treated with sub-minimal inhibitory concentrations of streptomycin, an antibiotic known to increase mistranslation. (D) Representative flow cytometry histogram showing GFP intensity in cells after heat shock treatment from exponentially growing culture (OD600 = 0.2) to later growth phase (OD600 = 0.6). (E) The MutL-based assay was coupled with a translation error reporter in the mutH strain and the wild-type (WT) strain. The 5% of the cells with the strongest translation error reporter fluorescence (high translation error) had statistically significant more cells with ≥1 focus than the 5% of the cells with the weakest fluorescence (low translation error) in both strains (see table S3).

Next, we coupled the translation error reporter with the MutL-based assay in the mutH strain, and we found that the top 5% of the cells showing strong translation error reporter fluorescence had significantly more MutL foci than the 5% of the cells with the weakest translation error reporter fluorescence (Fig. 3D and table S3). We did not observe this when we compared the top 10% of the cells showing strong fluorescence versus the 10% with the weakest fluorescence (fig. S6C and table S3). Therefore, the association between spontaneous translation errors and DNA replication errors is observed only in small subpopulations, which corroborated a hypothesis based on theoretical calculations (8).

Finally, we verified whether an observed increase in the frequency of MutL foci between the top 5% of the cells showing the strongest and the top 5% of the cells showing the weakest translation error reporter fluorescence observed in the mutH background could also be observed in the mismatch repair–proficient background. We found that an increase in the frequency of MutL foci was even stronger, that is, 4.6-fold versus 2.1-fold (Fig. 3E). Overproduction of the MutS protein slightly (but not significantly) reduced this increase (fig. S7 and table S3), just like in the streptomycin-treated cells (fig. S5). This suggests that an increase in translation errors affects some proteins participating in the mismatch repair–mediated elimination of DNA replication errors other than MutS and MutL proteins, such as UvrD helicase and different exonucleases.

DISCUSSION

Here, we explored the origins of spontaneous mutations in rapidly growing E. coli cells using the MutL-based assay that allows detection of DNA replication errors occurring in single cells, while simultaneously assaying induction of stress responses and translation fidelity. Results of our study indicate that most spontaneous DNA replication errors in proliferating cells arise in subpopulations of cells suffering from endogenous stresses, that is, problems with proteostasis, genome maintenance, and reactive oxidative species production.

Endogenous stresses may be induced by toxic metabolic by-products, spontaneous chemical alteration of cellular components, and stochastic fluctuations in the abundance of the low-copy proteins, all of which can also affect genome integrity and fidelity of DNA replication (3133). For instance, Pennington and Rosenberg (24) observed that around 0.9% of the exponentially growing E. coli cells have spontaneously induced SOS response, which is mainly induced by double-strand DNA breaks. Cellular protection and repair systems fix most damages, but damages in some cells may escape repair and cause mutations. Jee et al. (34) observed that overproduction of catalase, which eliminates H2O2, decreases eightfold spontaneous mutation rates in populations of growing E. coli cells. This suggests that there is either high reactive oxidative species production or suboptimal capacity to deal with this stress in some cells. The latter possibility is supported by the observation that fluctuations in O6-alkylguanine transferase protein abundance can compromise repair of spontaneous mutagenic DNA alkylation damages in E. coli and create a subpopulation of cells with increased mutation rates (35). Here, we also observed that MutS protein amount might be limiting for the spontaneous DNA replication error detection in growing cells.

The quantity of proteins can vary not only because of the random distribution of the low-abundance protein during cell division but also because of high translation errors that may produce mutated and truncated proteins, which may have loss of function, partial function, and dominant negative properties. The resulting suboptimal quantity and/or quality of proteins implicated in DNA replication, DNA repair, and other cellular functions may affect spontaneous mutation rates. We found that there are significantly more DNA replication errors in subpopulations of cells having high translation errors in both mismatch repair–proficient and –deficient strains.

Spontaneous mutation rates are kept at a minimum level because the probability that one newly generated mutation is deleterious or lethal is much higher than the probability that it is beneficial (36). The presence of subpopulations of phenotypic mutators, similar to those we described in this study, is not expected to affect the average mutation frequency and therefore to reduce the population mean fitness in a stable environment. Therefore, there should be no selective pressure for the evolution of molecular mechanisms, which will further decrease subpopulations of phenotypic mutators. Nevertheless, these subpopulations can contribute to the population adaptability in fluctuating environments because they represent a reservoir of increased genetic variability (37). In particular, their presence could be important when a combination of several mutations is required for adaptation, such as to cross fitness valleys (38).

MATERIALS AND METHODS

Bacterial strains and growth conditions

Bacterial strains used in this study are described in table S1. All strains were derivatives of the E. coli K12 MG1655 strain. When needed, mutants were constructed by P1 transduction of alleles from the INSERM U1001 laboratory strain collection or the Keio collection (39).

Overnight liquid cultures were grown in LB medium (Difco) at 37°C with 150 rpm shaking. When needed, medium was supplemented with antibiotics (Sigma-Aldrich) at the following concentrations: chloramphenicol (30 μg/ml), ampicillin (100 μg/ml), kanamycin (50 μg/ml), norfloxacin (0.04 μg/ml), paromomycin (5 μg/ml), rifampicin (100 μg/ml), and streptomycin (1 or 5 μg/ml).

Construction of the MutL-based DNA replication error reporter

The construction of mCherry fusions to wild-type mutL expressed from the Plac promoter on the pET-32a–type plasmid is described below. The plasmid encoding mCherry-MutL was made by using primers 5′-GCAGATCTATTAAAGAGGAGAAAGGTACCATGGTGAGC-3′ and 5′-ATGAATTCGCCAGATCCTGATCCAGATCCTGAGCCGCTCTTGTACAGCTCGTCCATGCCG-3′ to amplify the fragment encoding mCherry from the template p37pRD131 (laboratory collection). These primer pairs contain a Bgl II and an Eco RI restriction site at the N terminus and the C terminus, respectively. The polymerase chain reaction (PCR)–amplified fragments were digested by Bgl II and Eco RI and cloned in the pYFP-MutLCmR (13), creating the plasmid pmCherry-MutLCmR. Then, using the gene replacement technique, the plasmid pmCherry-MutLCmR was used as a template to integrate the mCherry-mutL-cat under a lactose promoter into the chromosome. Primers that include sequences homologous to the chromosomal lactose region were used to amplify the mCherry-mutL-cat. The primer 5′-TGTGTGGAATTGTGAGCGGATAACAATTTCACACAGGAAACAGCTATGGTGAGCAAGGGCGA-3′ contains a 45-bp sequence homologous to the lactose promoter (from −1 to −45), and the primer 5′-CACCAGACCAACTGGTAATGGTAGCGACCGGCGCTCAGCTGTTAGCAGCCGGATCTCAGTG-3′ contains a 40-bp sequence homologous to the lacZ C terminus (from 3026 to 3065). The amplified PCR fragments were introduced to the MG1655 mutL::frt strain by transformation containing the plasmid pKD46. Clones with a chloramphenicol-resistant and nonfunctional lacZ gene were selected and named AW052.

Construction of the translation error reporter

The construction of the translation error reporter is described below. Primers 5′-CGCCATCATACCCAACGTAACAAAACTCAGGCACGTTCCCGATATCAAATTACGCCCCGC-3′ and 5′-GAAAGAATAATGCAAAACAGAAAATGGATTTTGACCTCGCGATGCCTCTAGATTTAAATG-3′ that contain a sequence homologous to yzgL were used to amplify the fragment encoding the cat-sfgfp from strain RB166. The PCR fragments were integrated into the chromosome of strain MG1655, resulting in strain TD049. Primers 5′-CGCCATCATACCCAACGTAACAAAACTCAGGCACGTTCCCGATATCAAATTACGCCCCGC-3′ and 5′-GAAAGAATAATGCAAAACAGAAAATGGATTTTGACCTCGCCATATGAATATCCTCCTTAG-3′ that contain a sequence homologous to yzgL were used to amplify the fragment encoding the cat-sfgfp-kan from strain RB166, and the PCR fragments were integrated into the chromosome of strain MG1655, resulting in strain TD050. Next, the primer 5′-ATGGCCGACTCTGGTGACTACCCTGACCTAGGGTGTTCAGTTAAGACCCACTTTCACATT-3′, which contains a mutated stop codon switched from AAA to UAG at the codon coding for the 53rd amino acid in the flanking homologous regions of the sfgfp gene, and the primer 5′-AATCATGCTGCTTCATGTGATCCGGGTAACGAGAAAAACACTAAGCACTTGTCTCCTG-3′ were used to amplify the tetracycline-resistant gene, tet, and introduced into TD049 by lambda Red recombination (40). By this, the downstream part of the sfgfp gene was replaced by the tet gene. The primer 5′-CTGACCCTGAAATTCATCTGCACTACCGGTTAGCTGCCGG-3′, which also carried a mutated stop codon switched from AAA to UAG in the flanking homologous regions of the sfgfp gene, and the primer 5′-GAAAGAATAATGCAAAACAGAAAATGGATTTTGACCTCGC-3′ were then used to amplify the downstream part of the sfgfp gene and kan from TD050 and introduce it into the cat-sfgfp-tet construct. By this, the tet gene was replaced with the downstream part of the sfgfp gene and kan, resulting in the kanamycin-resistant but tetracycline-sensitive strain TD299. Finally, primers 5′-CGCCATCATACCCAACGTAACAAAACTCAGGCACGTTCCCGATATCGACGTCTAAGAAAC-3′ and 5′-GAAAGAATAATGCAAAACAGAAAATGGATTTTGACCTCGCCATATGAATATCCTCCTTAG-3′ were used to transfer the stop codon-sfgfp-kan construct from TD299 into MG1655, resulting in strain TD339.

Determination of rifampicin-resistant mutant frequencies

LB medium was inoculated with fewer than 100 cells from overnight culture so that there were no preexisting rifampicin mutants in the starting inoculum. Cells were grown in LB medium overnight with shaking at 37°C. Cells with appropriate dilutions were plated on LB agar medium to determine the total number of colony-forming units and on a selective medium, LB agar medium supplemented with rifampicin (100 μg/ml), to detect rifampicin-resistant mutants. Colonies were counted after 24 hours of incubation at 37°C.

Microscopy and image analysis

Tester strains were grown in LB medium with 0.1 mM isopropyl-β-d-thiogalactopyranoside overnight at 37°C with shaking. Overnight culture was diluted 1:400 (v/v) and incubated at 37°C with shaking. Cells from exponentially growing culture (OD600 = 0.2) were washed with phosphate-buffered saline (PBS) and inoculated onto a solid matrix of minimal medium agarose in microscope cavity slides, as described previously (14). Cells were examined with a 100× objective on an Axiovert 200 inverted microscope (Carl Zeiss) equipped with a Photometrics CoolSNAP camera (Princeton Instruments). Images were recorded and analyzed by MetaMorph software and ImageJ software, respectively.

Flow cytometry and cell sorting

Bacterial cultures were grown in LB medium overnight with shaking at 37°C. Overnight cultures were diluted 1:400 (v/v) and incubated at 37°C with shaking. Different treatments were applied to cells from exponentially growing culture (OD600 = 0.2): streptomycin (1 μg/ml), norfloxacin (0.04 μg/ml), and 5 mM hydrogen peroxide. Cells were washed with PBS and analyzed with a cytometer (Beckman Coulter) when the cultures reached an OD600 of 0.6.

To test the impact of heat shock response on the reporters, cultures were grown overnight with shaking at 30°C. Overnight cultures were diluted 1:400 (v/v) and incubated at 30°C with shaking until the cultures reached an OD600 of 0.2. Then, cultures were grown at 42°C with shaking, and cells were analyzed with a cytometer when cultures reached an OD600 of 0.6. FlowJo (LLC) was used to analyze data.

Samples for cell sorting were incubated overnight at 37°C with shaking. Overnight cultures were diluted 1:400 (v/v) and grown at 37°C with shaking until cultures reached an OD600 of 0.2. Cultures were then washed and resuspended in PBS before sorting using a cell sorter (BD FACSAria). Cells were analyzed for GFP expression before cell sorting to determine the populations to be sorted. Gates were drawn to separate the 5% of the cells expressing lower levels of GFP and the 5% of the cells expressing higher levels of GFP. Cells (7 × 105) were collected for each population. Cells were concentrated and analyzed with a microscope after cell sorting.

Statistical analyses

All statistical analyses were performed using the software MATLAB.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/6/eaat1608/DC1

fig. S1. MutL reporters, cell area, and number of foci per cell.

fig. S2. DNA replication errors and transcriptional promoter reporter P2rrnB.

fig. S3. Induction of fluorescent transcriptional promoter reporters.

fig. S4. Determination of minimum inhibitory concentration (MIC) of aminoglycosides for the E. coli wild-type strain MG1655 and the impact of sub-MIC of aminoglycoside on DNA replication errors.

fig. S5. DNA replication errors of different strains with sub-MIC of streptomycin.

fig. S6. DNA replication errors and translation error reporter.

fig. S7. DNA replication errors and translation errors in wild-type cells with excess MutS proteins.

table S1. Bacterial strains and plasmids.

table S2. The frequency of appearance of rifampicin mutants in mismatch repair–proficient and –deficient strain.

table S3. Chi-square test results of MutL foci frequency in this study.

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: Funding: This work was supported by FRM Grant DBF20160635736. A.C.W. was supported by a postdoctoral fellowship from Labex “Who am I?” Idex ANR-11-IDEX-0005-01/ANR-11-LABX-0071, Idex-Sorbonne Paris Cité. Author contributions: A.C.W. and I.M. designed the study. A.C.W. and L.F. performed experiments and analysis. A.C.W., L.F., and T.D. constructed strains. A.C.W. and I.M. wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Navigate This Article