Alkylating agents induce UVM, a recA-independent inducible mutagenic phenomenon in Escherichia coli

Expression of mutant alanine tRNAs increases spontaneous mutagenesis in Escherichia coli

The expression of mutA, an allele of the glycine tRNA gene glyV, can confer a novel mutator phenotype that correlates with its ability to promote Asp-->Gly mistranslation. Both activities are mediated by a single base change within the anticodon such that the mutant tRNA can decode aspartate codons (GAC/U) instead of the normal glycine codons (GCC/U). Here, we investigate whether specific Asp-->Gly mistranslation is required for the unexpected mutator phenotype. To address this question, we created and expressed 18 individual alleles of alaV, the gene encoding an alanine tRNA, in which the alanine anticodon was replaced with those specifying other amino acids such that the mutant (alaVX) tRNAs are expected to potentiate X-->Ala mistranslation, where X is one of the other amino acids. Almost all alaVX alleles proved to be mutators in an assay that measured the frequency of rifampicin-resistant mutants, with one allele (alaVGlu) being a stronger mutator than mutA. The alaVGlu mutator phenotype resembles that of mutA in mutational specificity (predominantly transversions), as well as SOS independence, but in a puzzling twist differs from mutA in that it does not require a functional recA gene. Our results suggest that general mistranslation (as opposed to Asp-->Gly alone) can induce a mutator phenotype. Furthermore, these findings predict that a large number of conditions that increase translational errors, such as genetic defects in the translational apparatus, as well as environmental and physiological stimuli (such as amino acid starvation or exposure to antibiotics) are likely to activate a mutator response. Thus, both genetic and epigenetic mechanisms can accelerate the acquisition of mutations.

The Escherichia coli UVM response is accompanied by an SOS-independent error-prone DNA replication activity demonstrable in vitro

UVM is an SOS-independent inducible response characterized by elevated mutagenesis at a site-specific 3, N4-ethenocytosine (epsilonC) residue borne on M13 single-stranded DNA transfected into Escherichia coli cells pretreated with DNA-damaging agents. By constructing and using E. coli strain AM124 (polA polB umuDC dinB lexA1[Ind-]), we show here that the UVM response is manifested in cells deficient for SOS induction, as well as for all four of the 'non-replicative' DNA polymerases, namely DNA polymerase I (polA), II (polB), IV (dinB) and V (umuDC). These results confirm that UVM represents a novel, previously unidentified cellular response to DNA-damaging agents. To address the question as to whether the UVM response is accompanied by an error-prone DNA replication activity, we applied a newly developed in vitro replication assay coupled to an in vitro mutation analysis system. In the assay, circular M13 single-stranded DNA bearing a site-specific lesion is converted to circular double-stranded replicative-form DNA in the presence of cell extracts and nucleotide precursors under conditions that closely mimic M13 replication in vivo. The newly synthesized (minus) DNA strand is selectively amplified by ligation-mediated polymerase chain reaction (LM-PCR), followed by a multiplex sequence analysis to determine the frequency and specificity of mutations. Replication of DNA bearing a site-specific epsilonC lesion by cell extracts from uninduced E. coli AM124 cells results in a mutation frequency of about 13%. Mutation frequency is elevated fivefold (to 58%) in cell extracts from UVM-induced AM124 cells, with C --> A mutations predominating over C --> T mutations, a specificity similar to that observed in vivo. These results, together with previously reported data, suggest that the UVM response is mediated through the induction of a transient error-prone DNA replication activity and that a modification of DNA polymerase III or the expression of a previously unidentified DNA polymerase may account for the UVM phenotype.

Escherichia coli cells defective for the recN gene display constitutive elevation of mutagenesis at 3,N(4)-ethenocytosine via an SOS-induced mechanism

The Escherichia coli UVM (UV Modulation of mutagenesis) response is a DNA damage-inducible mutagenic pathway detected as significantly increased mutagenesis at 3,N4-ethenocytosine (epsilon C) lesions borne on transfected single-stranded M13 vector DNA. All major classes of DNA-damaging agents can induce UVM, and the phenomenon is independent of previously characterized mutagenic responses in E. coli. To understand this phenomenon further, we set out to identify and characterize mutants in the UVM response. Screening a mutant bank of cells defective for 1-methyl-3-nitro-1-nitrosoguanidine-inducible genes revealed that defects in the recN gene cause a constitutive elevation of mutagenesis at epsilon C residues. In contrast to normal cells that show approximately 6% mutagenesis at epsilon C lesions, but approximately 60% upon UVM induction, recN-defective strains display approximately 50% mutagenesis at epsilon C lesion sites in untreated cells. However, the recN-mediated mutagenesis response was found to require the recA gene and the umuDC genes, and could be suppressed in the presence of a plasmid harbouring the SOS transcriptional repressor LexA. These results imply that recN cells are constitutively active for SOS mutagenesis functions. The observation that epsilonC mutagenesis is enhanced in recN cells confirms previous findings that mutagenesis at epsilonC can also be independently elevated by the SOS pathway.

Enzymatic repair of 5-formyluracil. I. Excision of 5-formyluracil site-specifically incorporated into oligonucleotide substrates by alka protein (Escherichia coli 3-methyladenine DNA glycosylase II)

5-Formyluracil (fU) is a major thymine lesion produced by reactive oxygen radicals and photosensitized oxidation. We have previously shown that fU is a potentially mutagenic lesion due to its elevated frequency to mispair with guanine. Therefore, fU can exist in DNA as a correctly paired fU:A form or an incorrectly paired fU:G form. In this work, fU was site-specifically incorporated opposite A in oligonucleotide substrates to delineate the cellular repair mechanism of fU paired with A. The repair activity for fU was induced in Escherichia coli upon exposure to N-methyl-N'-nitro-N-nitrosoguanidine, and the induction was dependent on the alkA gene, suggesting that AlkA (3-methyladenine DNA glycosylase II) was responsible for the observed activity. Activity assay and determination of kinetic parameters using purified AlkA and defined oligonucleotide substrates containing fU, 5-hydroxymethyluracil (hU), or 7-methylguanine (7mG) revealed that fU was recognized by AlkA with an efficiency comparable to that of 7mG, a good substrate for AlkA, whereas hU, another major thymine methyl oxidation products, was not a substrate. (1)H and (13)C NMR chemical shifts of 5-formyl-2'-deoxyuridine indicated that the 5-formyl group caused base C-6 and sugar C-1' to be electron deficient, which was shown to result in destabilization of the N-glycosidic bond. These features are common in other good substrates for AlkA and are suggested to play key roles in the differential recognition of fU, hU, and intact thymine. Three mammalian repair enzymes for alkylated and oxidized bases cloned so far (MPG, Nth1, and OGG1) did not recognize fU, implying that the mammalian repair activity for fU resided on a yet unidentified protein. In the accompanying paper (Terato, H., Masaoka, A., Kobayashi, M., Fukushima, S., Ohyama, Y., Yoshida, M., and Ide, H., J. Biol. Chem. 274, 25144-25150), possible repair mechanisms for fU mispaired with G are reported.

Multiplex sequence analysis demonstrates the competitive growth advantage of the A-to-G mutants of clarithromycin-resistant Helicobacter pylori

Clarithromycin resistance in Helicobacter pylori is due to point mutation within the 23S rRNA. We examined the growth rates of different types of site-directed mutants and demonstrated quantitatively the competitive growth advantage of A-to-G mutants over other types of mutants by a multiplex sequencing assay. The results provide a rational explanation of why A-to-G mutants are predominantly observed among clarithromycin-resistant clinical isolates.

SOS and UVM pathways have lesion-specific additive and competing effects on mutation fixation at replication-blocking DNA lesions

Escherichia coli cells have multiple mutagenic pathways that are induced in response to environmental and physiological stimuli. Unlike the well-investigated classical SOS response, little is known about newly recognized pathways such as the UVM (UV modulation of mutagenesis) response. In this study, we compared the contributions of the SOS and UVM pathways on mutation fixation at two representative noninstructive DNA lesions: 3,N4-ethenocytosine (epsilonC) and abasic (AP) sites. Because both SOS and UVM responses are induced by DNA damage, and defined UVM-defective E. coli strains are not yet available, we first constructed strains in which expression of the SOS mutagenesis proteins UmuD' and UmuC (and also RecA in some cases) is uncoupled from DNA damage by being placed under the control of a heterologous lac-derived promoter. M13 single-stranded viral DNA bearing site-specific lesions was transfected into cells induced for the SOS or UVM pathway. Survival effects were determined from transfection efficiency, and mutation fixation at the lesion was analyzed by a quantitative multiplex sequence analysis procedure. Our results suggest that induction of the SOS pathway can independently elevate mutagenesis at both lesions, whereas the UVM pathway significantly elevates mutagenesis at epsilonC in an SOS-independent fashion and at AP sites in an SOS-dependent fashion. Although mutagenesis at epsilonC appears to be elevated by the induction of either the SOS or the UVM pathway, the mutational specificity profiles for epsilonC under SOS and UVM pathways are distinct. Interestingly, when both pathways are active, the UVM effect appears to predominate over the SOS effect on mutagenesis at epsilonC, but the total mutation frequency is significantly increased over that observed when each pathway is individually induced. These observations suggest that the UVM response affects mutagenesis not only at class 2 noninstructive lesions (epsilonC) but also at classical SOS-dependent (class 1) lesions such as AP sites. Our results add new layers of complexity to inducible mutagenic phenomena: DNA damage activates multiple pathways that have lesion-specific additive as well as suppressive effects on mutation fixation, and some of these pathways are not directly regulated by the SOS genetic network.

Escherichia coli cells expressing a mutant glyV (glycine tRNA) gene have a UVM-constitutive phenotype: implications for mechanisms underlying the mutA or mutC mutator effect

Transfection of M13 single-stranded viral DNA bearing a 3,N4-ethenocytosine lesion into Escherichia coli cells pretreated with UV results in a significant elevation of mutagenesis at the lesion site compared to that observed in untreated cells. This response, termed UVM, for UV modulation of mutagenesis, is induced by a variety of DNA-damaging agents and is distinct from known cellular responses to DNA damage, including the SOS response. This report describes our observation, as a part of our investigation of the UVM phenomenon, that E. coli cells bearing a mutA or mutC allele display a UVM-constitutive phenotype. These mutator alleles were recently mapped (M. M. Slupska, C. Baikalov, R. Lloyd, and J. H. Miller, Proc. Natl. Acad. Sci. USA 93:4380-4385, 1996) to the glyV (mutA) and glyW (mutC) tRNA genes. Each mutant allele was shown to arise by an identical mutation in the anticodon sequence such that the mutant tRNAs could, in principle, mistranslate aspartate codons in mRNA as glycine at a low level. Because a UVM-constitutive phenotype resulting from a mutation in a tRNA gene was unexpected, we undertook a series of experiments designed to test whether the phenotype was indeed mediated by the expression of mutant glycine tRNAs. We placed either a wild-type or a mutant glyV gene under the control of a heterologous inducible promoter on a plasmid vector. E. coli cells expressing the mutant glyV gene displayed all three of the following phenotypes: (i) missense suppression of a test allele, (ii) a mutator phenotype measured by mutation to rifampin resistance, and (iii) a UVM-constitutive phenotype. These phenotypes were not associated with cells expressing the wild-type glyV gene or with cells in which the mutant allele was present but was not transcriptionally induced. These observations provide strong support for the idea that expression of mutant tRNA can confer a mutator phenotype, including the UVM-constitutive phenotype observed in mutA and mutC cells. However, our data imply that low-level mistranslation of the epsilon subunit of polymerase III probably does not account for the observed UVM-constitutive phenotype. Our results also indicate that mutA deltarecA double mutants display a normal UVM phenotype, suggesting that the mutA effect is recA dependent. The observations reported here raise a number of intriguing questions and raise the possibility that the UVM response is mediated through transient alteration of the replication environment.

Nebularine (9-2'-deoxy-beta-D-ribofuranosylpurine) has the template characteristics of adenine in vivo and in vitro

Nebularine (9-beta-D-ribofuranosylpurine; Nb) is a naturally occurring nucleoside with structural features suggestive of a universal base. However, previous observations based on thermal melting characteristics of oligonucleotides suggested that Nb formed stable pairs only with thymine. To determine the template characteristics of Nb, we constructed M13 viral single-stranded DNA (ssDNA) molecules bearing a single site-specific deoxynebularine (9-2'-deoxy-beta-D-ribofuranosylpurine) residue. The ssDNA constructs were transfected into Escherichia coli cells to determine the specificity of base insertion opposite Nb, as well as to determine the effect of Nb on the replicability of the transfected DNA. Base insertion opposite Nb, analyzed by a multiplex sequencing technology, suggests that Nb has the template characteristics of adenine. Analysis of DNA replicability, measured as transfection efficiency, indicates that Nb does not block DNA replication. UV irradiation of host cells before transfection did not significantly affect survival or base insertion specificity within the limits of multiplex sequencing technology employed, suggesting that inducible mutagenic phenomena appear to have only minor effects on translesion synthesis across Nb. In addition, in vitro DNA elongation experiments on oligonucleotide templates using E. coli DNA polymerase I (Klenow fragment) as the model polymerase showed that Nb templates for T, but not for other bases under the tested conditions. The data reported in this communication underscore the importance of base-pair geometry as a specificity-determinant during base insertion by replicative polymerases.

Replication of M13 single-stranded DNA bearing a site-specific ethenocytosine lesion by Escherichia coil cell extracts

Previous investigation on the mutagenic effects of 3, N4-Ethenocytosine (epsilon C), a nonpairing DNA lesion, revealed the existence of a novel SOS-independent inducible mutagenic mechanism in E. coli termed UVM for UV modulation of mutagenesis. To investigate whether UVM is mediated by an alteration of DNA replication, we have set up an in vitro replication system in which phage M13 viral single-stranded DNA bearing a single site-specific (epsilon C) residue is replicated by soluble protein extracts from E. coli cells. Replication products were analyzed by agarose gel electrophoresis and the frequency of translesion synthesis was determined by restriction endonuclease analyses. Our data indicate that DNA replication is strongly inhibited by epsilon C, but that translesion DNA synthesis does occur in about 14% of the replicated DNA molecules. These results are very similar to those observed previously in vivo, and suggest that this experimental system may be suitable for evaluating alterations in DNA replication in UVM-induced cells.

Role of mismatch repair in the Escherichia coli UVM response

Mutagenesis at 3,N4-ethenocytosine (epsilonC), a nonpairing mutagenic lesion, is significantly enhanced in Escherichia coli cells pretreated with UV, alkylating agents, or H2O2. This effect, termed UVM (for UV modulation of mutagenesis), is distinct from known DNA damage-inducible responses, such as the SOS response, the adaptive response to alkylating agents, or the oxyR-mediated response to oxidative agents. Here, we have addressed the hypothesis that UVM results from transient depletion of a mismatch repair activity that normally acts to reduce mutagenesis. To test whether the loss of mismatch repair activities results in the predicted constitutive UVM phenotype, E. coli cells defective for methyl-directed mismatch repair, for very-short-patch repair, or for the N-glycosylase activities MutY and MutM were treated with the UVM-inducing agent 1-methyl-3-nitro-1-nitrosoguanidine, with subsequent transfection of M13 viral single-stranded DNA bearing a site-specific epsilonC lesion. Survival of the M13 DNA was measured as transfection efficiency, and mutation fixation at the lesion was characterized by multiplex sequencing technology. The results showed normal UVM induction patterns in all the repair-defective strains tested. In addition, normal UVM induction was observed in cells overexpressing MutH, MutL, or MutS. All strains displayed UVM reactivation, the term used to describe the increased survival of epsilonC-containing DNA in UVM-induced cells. Taken together, these results indicate that the UVM response is independent of known mismatch repair systems in E. coli and may thus represent a previously unrecognized misrepair or misreplication pathway.

Induction of the Escherichia coli UVM response by oxidative stress

UVM (ultraviolet modulation of mutagenesis) is a recently described recA-independent, inducible mutagenic phenomenon in which prior UV irradiation of Escherichia coli cells strongly enhances mutation fixation at a site-specific 3-N4-ethenocytosine (epsilon C) lesion borne on a transfected single-stranded M13 DNA vector. Subsequent studies demonstrated that UVM is also induced by alkylating agents, and is distinct from both the SOS response and the adaptive response to alkylation damage. Because of the increasing significance being attributed to oxidative DNA damage, it is interesting to ask whether this class of DNA damage can also induce UVM. By transfecting M13 vector DNA bearing a site-specific epsilon C lesion into cells pretreated with inducing agents, we show here that the oxidative agent H2O2 is a potent inducer of UVM, and that the induction of UVM by H2O2 does not require oxyR-regulated gene expression. UVM induction by H2O2 appears to be mediated by DNA damage, as indicated by the observation of a concomitant reduction in cellular toxicity and UVM response in OxyRc cells. Available evidence suggests that UVM represents a generalized cellular response to a broad range of chemical and physical genotoxicants, and that DNA damage constitutes the most likely signal for its induction.