Skip to main content
NCBI home page
Genetics logoLink to Genetics
. 2014 Sep 10;198(3):1231–1235. doi: 10.1534/genetics.114.170258

Stress-Induced Mutation Rates Show a Sigmoidal and Saturable Increase Due to the RpoS Sigma Factor in Escherichia coli

Ram Maharjan 1, Thomas Ferenci 1,1
PMCID: PMC4224162  PMID: 25213168

Abstract

Stress-induced mutagenesis was investigated in the absence of selection for growth fitness by using synthetic biology to control perceived environmental stress in Escherichia coli. We find that controlled intracellular RpoS dosage is central to a sigmoidal, saturable three- to fourfold increase in mutation rates and associated changes in DNA repair proteins.

Keywords: stress, mutation rates, DNA repair, Escherichia coli, dinB, mutS

MUTATIONS provide a source of evolutionary innovation (Radman 2001) but also contribute to the development of antibiotic resistance (Gutierrez et al. 2013) and to diseases like cancer (Alexandrov and Stratton 2014). The rate at which mutations occur is not constant, even within a species (Bjedov et al. 2003). There is a tradeoff between balancing genetic integrity in good times with increased evolvability in suboptimal environments. A way of shifting this balance is through stress-induced mutagenesis (SIM) (Radman 2001). In the laboratory, mutation rate in bacteria like Escherichia coli is increased by suboptimal conditions and regulated by adverse environments (Bjedov et al. 2003; Foster 2007; Galhardo et al. 2007; MacLean et al. 2013). Mutation rates indicative of SIM have been studied (Foster 2007; Galhardo et al. 2007; Matic 2013) but aspects of the evidence for SIM have been disputed (Roth et al. 2006). Particularly, the role of selection in increasing the frequency of drug-resistant mutants is contentious in studies like Bjedov et al. (2003), which relied on aged, stressed colonies. This is because some mutants supposedly have a growth advantage in old colonies on agar plates (Katz and Hershberg 2013).

These uncertainties are best resolved if the impact of stress on genomes could be assessed in the absence of fitness selection in aging colonies over extended periods. A problem for assaying SIM is that the level of stress in cultures and thus selection is difficult to control in colonies. The stress levels, fitness, and mutation rate can change over time in stationary-phase liquid culture as well (Yeiser et al. 2002; Loewe et al. 2003). An approach to circumventing the problem of selection and varying stress level is to use identical growth conditions, but with bacteria in which the level of SIM is under genetic control. A way of doing this is to control a major input into SIM regulation; RpoS has been implicated as a controller of SIM in several laboratories (Galhardo et al. 2007; Storvik and Foster 2010; Gutierrez et al. 2013). RpoS (also known as σS) is an alternative RNA polymerase sigma factor and the master stress regulator in E. coli, which is highly variable and dependent on both environmental signals and growth phase (Battesti et al. 2011; Hengge 2011). We use here a surrogate means of fixing stress by artificially setting levels of RpoS in a collection of strains to allow highly replicated measurements of mutation rates.

We recently constructed a set of synthetic strains to achieve fixed, environment-independent RpoS levels (Maharjan et al. 2013). The strains in the set do not have transcriptional, translational, and post-translational control over RpoS levels; instead these strains contain distinct synthetic promoters to express fixed levels of rpoS. All the strains along with their RpoS level are listed in Table 1 and were standardized to the content of RpoD, which is constant in different environments (Ishihama 2000; Gutierrez et al. 2013; Maharjan et al. 2013). The strain set covered a 200-fold range of sigma factor ratios and expressed the full range of environmental stress resistance phenotypes controlled by the general stress response (Maharjan et al. 2013). The RpoS/RpoD ratios in the synthetic strains (BW5205–BW5222) remain relatively constant in different growth phases, unlike the wild-type MC4100, as shown in Table 1. To check whether different cultures of strains maintained similar RpoS levels, we tested iodine staining levels of cultures using the method that we used in our previous study; iodine staining is linearly dependent on the RpoS content of cells (Maharjan et al. 2013). There was no change in staining level in cultures suggesting that storage and repeat culture did not affect the RpoS levels.

Table 1. Properties of strains used in this study.

RpoS/RpoD ratio (m ± SD)a
Strain Relevant genotype (primer set used for construction) Exponential bacteriab Stationary phase bacteriab Source
MC4100 F- araD139 Δ(argF-lac)U169 rspL150 deoCl relA1 thiA ptsF25 flb5301 rbsR 0.00 ± 0.00 0.40 ± 0.06 Ferenci et al. (2009)
BW2952 malG::λplacMu55 ϕ(malG::lacZ)rssB::IS1 0.76 ± 0.04 1.00 ± 0.00 Ferenci et al. (2009)
BW3709 BW2952 rpoS::Tn10 0.00 ± 0.00 0.00 ± 0.00 Notley-Mcrobb et al. (2002)
BW5205 BW2952 rpoSleader::bla (rpoSF and rpoSR1) 0.30 ± 0.07 0.31 ± 0.02 Maharjan et al. (2013)
BW5206 BW2952 rpoSleader::bla (rpoSF and rpoSR2) 0.01 ± 0.01 0.00 ± 0.01 Maharjan et al. (2013)
BW5207 BW2952 rpoSleader::bla (rpoSF and rpoSR3) 0.13 ± 0.09 0.16 ± 0.10 Maharjan et al. (2013)
BW5208 BW2952 rpoSleader::bla (rpoSF and rpoSR4) 0.86 ± 0.04 0.72 ± 0.13 Maharjan et al. (2013)
BW5213 BW2952 rpoSleader::bla (rpoSF and rpoSR5) 0.28 ± 0.03 0.22 ± 0.05 Maharjan et al. (2013)
BW5214 BW2952 rpoSleader::bla (rpoSF and rpoSR6) 0.28 ± 0.01 0.27 ± 0.06 Maharjan et al. (2013)
BW5219 BW2952 rpoSleader::bla (rpoSF and rpoSR7) 0.26 ± 0.03 0.22 ± 0.01 Maharjan et al. (2013)
BW5220 BW2952 rpoSleader::bla (rpoSF and rpoSR8) 0.02 ± 0.00 0.03 ± 0.01 Maharjan et al. (2013)
BW5222 BW2952 rpoSleader::bla (rpoSF and rpoSR10) 0.31 ± 0.00 0.38 ± 0.03 Maharjan et al. (2013)
Open in a new tab
a

RpoS/RpoD ratio data for the two culture conditions are from graphs in Maharjan et al. (2013) based on Western blot analysis of RpoS and RpoD detected using anti-RpoS and anti-RpoD antibodies. The mean (m) and standard deviation (SD) were from at least three independent experiments.

b

The growth medium used was Luria-Bertani broth.

The RpoS strain set was used to investigate SIM by measuring mutation rates (Figure 1, A and B) in fluctuation test-based, >20-fold replicated assays of rifampicin resistance (RifR) (Bjedov et al. 2003) and cycloserine resistance (CycR) (Feher et al. 2006) in strains with 11 different set levels of RpoS but under identical growth conditions. The RifR change is limited to a number of point mutations in rpoB (Wolff et al. 2004), while CycR does not change growth fitness and is due to a wide spectrum of loss-of-function mutations including large indels and transpositions at many points in the cycA gene (Feher et al. 2006). The strain set was further used to provide a quantitative link between RpoS and the regulation of DNA repair systems involving dinB (error-prone polymerase) and mutS (mismatch repair) implicated in SIM (Al Mamun et al. 2012), as shown in Figure 1C.

Figure 1.

Figure 1

Open in a new tab

The relationship between intracellular RpoS dosage and mutation rates. (A and B) The mutation rate of E. coli with set levels of RpoS was estimated using two different assays, involving resistance to cycloserine (CycR, A) and rifampicin (RifR, B). The mutation rates were estimated by fluctuation analyses in engineered strains with fixed ratios of RpoS/RpoD (plotted on the x-axis based on the data in Maharjan et al. 2013). The engineered strains do not have transcriptional, translational, and post-translational control over RpoS levels and contained synthetic promoters to express a fixed level of rpoS (Maharjan et al. 2013). A single colony of each strain was inoculated into 5 ml Luria–Bertani broth (LB) and allowed to propagate overnight at 37° with shaking. The overnight culture was diluted in 5 ml fresh LB medium and allowed to grow to optical density of 0.6 at 600 nm. The exponentially growing cultures were further diluted 10,000-fold and 150 µl was distributed into each of 40 wells in 96-well plates and incubated at 37° with shaking at 200 rpm. Aliquots of each well were then plated on rifampicin and cycloserine plates to detect CycR and RifR mutant colonies. The mutation rates were estimated from the number of resistant colonies per culture and total cell count by using the fluctuation analysis calculator (FALCOR) web tool (Hall et al. 2009). Error bars in A and B are upper and lower limits with 95% CI and are based on at least 20 replicate cultures from fluctuation tests. The mutation rates were then estimated from the number of resistant colonies per culture and total cell count by using the FALCOR web tool (Hall et al. 2009). At each RpoS/RpoD ratio, two to three independent fluctuation analyses, each with >20 replicates, were carried out on the same strain. (C) The levels of DinB and MutS proteins in exponentially growing cultures of engineered strains were determined by using anti-DinB and anti-MutS rabbit antibodies. The RpoD, MutS, and DinB proteins were quantified by Western blotting. Error bars represent the standard deviations from two independent experiments. For estimation of DinB and MutS levels in the strain set with fixed level of RpoS/RpoD, a single colony of each strain was grown overnight in 5 ml LB in a MacCartney bottle with shaking (200 rpm) at 37°. The overnight cultures were diluted 500-fold in 5 ml of fresh LB and allowed to grow to optical density 0.4–0.6 at 600 nm. The cultures (1 ml) were harvested by centrifugation (10000 × g) for 1 min and 200 µl of 1× SDS PAGE sample buffer was added after removing supernatant. The tubes containing protein samples were then snap frozen using dry ice and stored until use. of Protein samples, 10 µl, were revolved on a 12% polyacrylamide gel (Bio-Rad, BioRad Australia, Sydney, N.S.W., Australia). Proteins were then transferred in Optitran BA-S85 blotting membrane (GE Whatman, GE Healthcare Australia Pty. Ltd, Sydney N.S.W., Australia). Proteins DinB and MutS were detected using anti-DinB (Santa Cruz Biotechnology, Santa Cruz Biotechnology, Dallas, Texas 75220 U.S.A.) and anti-MutS (US Biological, United States Biological, Massachusetts, MA 01907, U.S.A.) rabbit antibodies using the same protocol as that used for the detection of RpoS and RpoD and band intensities were also quantified by densitometry as previously described (Maharjan et al. 2013). DinB and MutS protein levels in the RpoS strain set were expressed as a ratio to the RpoD band as reference. (D) The physiological levels of RpoS/RpoD in wild-type E. coli K12 bacteria (strain MC4100; Maharjan et al. 2013) are shown in different stress situations. Unstressed bacteria were grown to exponential phase in LB medium (open arrow) or grown to exponential phase in LB with 0.7M NaCl added to initiate osmotic stress (shaded arrow) while the ratio found in 16-hr culture to stationary phase in LB medium is shown by the solid arrow.

The dose-dependent increase in mutation rates with RpoS shown in Figure 1, A and B, strongly implicates the importance of RpoS level in SIM. The mutation rate in the strain with the highest RpoS content was 6.6 × 10−8 per locus per generation for RifR and 2.08 × 10−7 per locus per generation for CycR. These were three- and fourfold higher than the mutation rates found in strain with the lowest RpoS (2.4 × 10−8 per locus per generation for RifR and 0.53 × 10−7 per locus per generation for CycR per locus per generation respectively (two-tailed P < 0.05 in both cases). The RpoS-dependent elevation of mutation rates for RifR was slightly lower than the 5.5-fold increase in RifR mutants stress-induced in aged colonies (Bjedov et al. 2003). Nevertheless, the RpoS dosage is responsible for much of the SIM under stress although stress inputs not sensed through RpoS may contribute to the overall mutation rate (Al Mamun et al. 2012).

To compare our RifR mutation rates with the previously estimated mutation rate per genome by Drake (2012) for E. coli, we converted our per-locus mutation rates into per genome per generation (Table 2). Given that 79-point mutations in rpoB can confer RifR (Garibyan et al. 2003) and a genome size of 4.6 Mb, we estimated our genomic mutation rate to be 0.0038 per genome per generation for the high RpoS strain and 0.0013 per genome per generation for the low RpoS strain. The low-RpoS mutation rates are similar to the 0.001 per genome per generation estimated by Lee et al. (2012) in mutation accumulation (MA) experiments with E. coli. On the other hand, the high perceived stress, high-RpoS mutation rate is within the range 0.0022–0.0043 per genome per generation in Drake’s studies (Drake et al. 1998; Drake 2012). This raises the interesting possibility that the MA data are characteristic of a low-stress environment, which is consistent with the frequent subculture in fresh media in MA experiments (Lee et al. 2012).

Table 2. Mutation rates in Escherichia coli K12.

Strain Assay Mutation rate per locus (×108) [95% CL] Mutation rate per bp (×1010) Mutation rate per genome References
BW2952 (high RpoS)a RifR fluctuation test 6.55 [4.9–7.9] 8.3 0.0039 This study
BW5206 (low RpoS)a RifR fluctuation test 2.2 [1.3–2.6] 2.8 0.0013 This study
BW2952 (high RpoS)a CycR fluctuation test 20.8 [15.7–26.7] NA NA This study
BW5206 (low RpoS)a CycR fluctuation test 5.3 [3.5–7.5] NA NA This study
E. coli K-12b CycR fluctuation test 6.5 NA NA Feher et al. (2006); Posfai et al. (2006)
E. coli K-12b Mutation Accumulation NA 2.2 0.0010 Lee et al. (2012)
E. coli K-12b lacI (Lac reversion) NA 4.1–9.3 0.0019–0.0043 Drake et al. (1998); Drake (2012)
Open in a new tab
a

RifR and CycR mutation rates and 95% confidence interval (CL) in the strains with high and low levels of RpoS were calculated as described in Figure 1. The RifR mutation rates per base pair (bp) per generation were estimated from mutation rates per locus by assuming RifR is conferred by 79 different point mutations in the rpoB gene (Garibyan et al. 2003).

b

Mutation rates were based on published studies.

For CycR, we were unable to convert mutation rates into per genome per generation due to the lack of information on the number of sites in the cycA gene that can give rise to the CycR phenotype, which includes point mutations, small and large indels, and transpositions in cycA (Feher et al. 2006). Nevertheless, the CycR per-locus mutation rate in the strain with low RpoS is very close to the previously estimated mutation rate using the identical method (Table 2;Feher et al. 2006; Posfai et al. 2006). So for both the CycR and RifR assays, the low-RpoS strain reflects previously obtained lab data, but the high-RpoS strain is closer to the SIM state.

Interestingly, both the CycR and RifR assays in Figure 1, A and B, showed a sigmoidal increase before saturation at high RpoS levels. The RpoS threshold needed to induce CycR is slightly lower compared to RifR, possibly because different mutations are involved. Nevertheless, as shown in Figure 1D, mutation rates with either RifR or CycR increased at RpoS levels higher than those found in rapidly growing bacteria. Even osmotic stress levels that elicit partial induction of the general stress response are not high enough to induce SIM. This stress threshold for inducing SIM makes ecological sense in balancing genetic integrity, so that low levels of environmental stress do not elevate mutation rates. These results are consistent with a stationary-phase increase in mutation rates, when RpoS levels are elevated to the position of the solid arrow shown in Figure 1D, where mutation rates are elevated. It should be noted though that the cellular levels of RpoS can differ even in closely related E. coli K-12 strains (Spira et al. 2008), which in turn can affect the mutation rate. The parentage strain used in this study as the wild type (MC4100) exhibits similar RpoS level to MG1655 (Spira et al. 2008), the strain used in other mutation rate measurements (Lee et al. 2012).

The effect of RpoS on SIM is thought to be through modulating both the mismatch repair system and error-prone polymerase DinB (Godoy et al. 2007; Al Mamun et al. 2012; Gutierrez et al. 2013). Our strain set allows testing of these notions. We find, as shown in Figure 1C, that RpoS level changes regulation of both repair systems simultaneously, so both together contribute to SIM, even though different networks regulate the two processes (Al Mamun et al. 2012; Gutierrez et al. 2013). For example, other influences besides RpoS (such as from RpoE;Gibson et al. 2010) may additionally affect the overall mutation rate. Nevertheless, the mutation rate threshold in Figure 1 is correlated with the regulation of mutS and dinB in the strains set.

In conclusion, stress sensed solely through RpoS resulted in a three- to fourfold increase in mutation rates in Figure 1. Thus even in the absence of selection for competitive fitness as may occur in aging colonies, SIM clearly has a component influenced by the general stress response. A central conclusion of this study is that SIM exhibits a nonlinear stress–dose response, with a clear threshold. The sigmoidal dose effects and saturability prevent a simple linear extrapolation between levels of stress, mutation rates, and evolvability. The final conclusion is that the mutation rate we observe for both CycR and RifR with high levels of perceived stress, but not low RpoS, is higher than that found in the oft-used MA laboratory experiments (Lee et al. 2012). Given the use of an unstressed environment in MA experiments, the threefold discrepancy in mutation rates between MA experiments and other estimates of genomic mutation rates (Drake et al. 1998) may well be due to different levels of environmental stress in different laboratory experiments and in nature.

Acknowledgments

We thank Edward C. Holmes and Laurence D. Hurst for constructive comments on the manuscript. The work was funded by a Discovery grant from the Australian Research Council.

Footnotes

Communicating editor: J. Lawrence

Literature cited

  1. Al Mamun A. A. M., Lombardo M.-J., Shee C., Lisewski A. M., Gonzalez C., et al. , 2012.  Identity and function of a large gene network underlying mutagenic repair of DNA breaks. Science 338: 1344–1348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alexandrov L. B., Stratton M. R., 2014.  Mutational signatures: the patterns of somatic mutations hidden in cancer genomes. Curr. Opin. Genet. Dev. 24: 52–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Battesti A., Majdalani N., Gottesman S., 2011.  The RpoS-mediated general stress response in Escherichia coli. Annu. Rev. Microbiol. 65: 189–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bjedov I., Tenaillon O., Gerard B., Souza V., Denamur E., et al. , 2003.  Stress-induced mutagenesis in bacteria. Science 300: 1404–1409. [DOI] [PubMed] [Google Scholar]
  5. Drake J. W., 2012.  Contrasting mutation rates from specific-locus and long-term mutation-accumulation procedures. G3 (Bethesda) 2: 483–485. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Drake J. W., Charlesworth B., Charlesworth D., Crow J. F., 1998.  Rates of spontaneous mutation. Genetics 148: 1667–1686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Feher T., Cseh B., Umenhoffer K., Karcagi D., Posfai G., 2006.  Characterization of cycA mutants of Escherichia coli: an assay for measuring in vivo mutation rates. Mutat. Res. Fundamental Mol. Mech. Mutagenesis 595: 184–190. [DOI] [PubMed] [Google Scholar]
  8. Ferenci T., Zhou Z. M., Betteridge T., Ren Y., Liu Y., et al. , 2009.  Genomic sequencing reveals regulatory mutations and recombinational events in the widely used MC4100 lineage of Escherichia coli K-12. J. Bacteriol. 191: 4025–4029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Foster P. L., 2007.  Stress-induced mutagenesis in bacteria. Crit. Rev. Biochem. Mol. Biol. 42: 373–397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Galhardo R. S., Hastings P. J., Rosenberg S. M., 2007.  Mutation as a stress response and the regulation of evolvability. Crit. Rev. Biochem. Mol. Biol. 42: 399–435. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Garibyan L., Huang T., Kim M., Wolff E., Nguyen A., et al. , 2003.  Use of the rpoB gene to determine the specificity of base substitution mutations on the Escherichia coli chromosome. DNA Repair 2: 593–608. [DOI] [PubMed] [Google Scholar]
  12. Gibson J. L., Lombardo M. J., Thornton P. C., Hu K. H., Galhardo R. S., et al. , 2010.  The sigma E stress response is required for stress-induced mutation and amplification in Escherichia coli. Mol. Microbiol. 77: 415–430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Godoy V. G., Jarosz D. F., Simon S. M., Abyzov A., Ilyin V., et al. , 2007.  UmuD and RecA directly modulate the mutagenic potential of the Y family DNA polymerase DinB. Mol. Cell 28: 1058–1070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gutierrez A., Laureti L., Crussard S., Abida H., Rodriguez-Rojas A., et al. , 2013.  beta-lactam antibiotics promote bacterial mutagenesis via an RpoS-mediated reduction in replication fidelity. Nat. Commun. 4: 1610. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hall B. M., Ma C.-X., Liang P., Singh K. K., 2009.  Fluctuation AnaLysis CalculatOR: a web tool for the determination of mutation rate using Luria–Delbrück fluctuation analysis. Bioinformatics 25: 1564–1565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hengge, R., 2011 The General Stress Response in Gram-Negative Bacteria. American Society of Microbiology, Washington, DC.
  17. Ishihama A., 2000.  Functional modulation of Escherichia coli RNA polymerase. Annu. Rev. Microbiol. 54: 499–518. [DOI] [PubMed] [Google Scholar]
  18. Katz S., Hershberg R., 2013.  Elevated mutagenesis does not explain the increased frequency of antibiotic resistant mutants in starved aging colonies. PLoS Genet. 9: 3968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lee H., Popodi E., Tang H. X., Foster P. L., 2012.  Rate and molecular spectrum of spontaneous mutations in the bacterium Escherichia coli as determined by whole-genome sequencing. Proc. Natl. Acad. Sci. USA 109: E2774–E2783. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Loewe L., Textor V., Scherer S., 2003.  High deleterious genomic mutation rate in stationary phase of Escherichia coli. Science 302: 1558–1560. [DOI] [PubMed] [Google Scholar]
  21. MacLean R. C., Torres-Barcelo C., Moxon R., 2013.  Evaluating evolutionary models of stress-induced mutagenesis in bacteria. Nat. Rev. Genet. 14: 221–227. [DOI] [PubMed] [Google Scholar]
  22. Maharjan R., Nilsson S., Sung J., Haynes K., Beardmore R. E., et al. , 2013.  The form of a trade-off determines the response to competition. Ecol. Lett. 16: 1267–1276. [DOI] [PubMed] [Google Scholar]
  23. Matic I., 2013.  Stress-induced mutagenesis in bacteria, pp. 1–19 in Stress-Induced Mutagenesis, edited by Mittelman D. Springer, New York. [Google Scholar]
  24. Notley-McRobb L., King T., Ferenci T., 2002.  rpoS mutations and loss of general stress resistance in Escherichia coli populations as a consequence of conflict between competing stress responses. J. Bacteriol. 184: 806–811. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Posfai G., Plunkett G., Feher T., Frisch D., Keil G. M., et al. , 2006.  Emergent properties of reduced-genome Escherichia coli. Science 312: 1044–1046. [DOI] [PubMed] [Google Scholar]
  26. Radman M., 2001.  Fidelity and infidelity. Nature 413: 115. [DOI] [PubMed] [Google Scholar]
  27. Roth J. R., Kugelberg E., Reams A. B., Kofoid E., Andersson D. I., 2006.  Origin of mutations under selection: the adaptive mutation controversy. Annu. Rev. Microbiol. 60: 477–501. [DOI] [PubMed] [Google Scholar]
  28. Spira B., Hu X., Ferenci T., 2008.  Strain variation in ppGpp concentration and RpoS levels in laboratory strains of Escherichia coli K-12. Microbiology 154: 2887–2895. [DOI] [PubMed] [Google Scholar]
  29. Storvik K. A. M., Foster P. L., 2010.  RpoS, the stress response sigma factor, plays a dual role in the regulation of Escherichia coli’s error-prone DNA polymerase IV. J. Bacteriol. 192: 3639–3644. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Wolff E., Kim M., Hu K. B., Yang H. J., Miller J. H., 2004.  Polymerases leave fingerprints: analysis of the mutational spectrum in Escherichia coli rpoB to assess the role of polymerase IV in spontaneous mutation. J. Bacteriol. 186: 2900–2905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Yeiser B., Pepper E. D., Goodman M. F., Finkel S. E., 2002.  SOS-induced DNA polymerases enhance long-term survival and evolutionary fitness. Proc. Natl. Acad. Sci. USA 99: 8737–8741. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Genetics are provided here courtesy of Oxford University Press

RESOURCES