Spontaneous Mutation at a 5-Methylcytosine Hotspot Is Prevented by Very Short Patch (Vsp) Mismatch Repair

In many strains of Escherichia coli, the product of gene dcm methylates the internal cytosines in the sequence 5'CC(A or T)GG. Spontaneous deamination of 5-methylcytosine produces thymine which, if not corrected, can result in a transition mutation. 5-Methylcytosines in the lacI gene are hotspots for spontaneous C to T mutations. dcm is linked to vsr, a gene required for very short patch (VSP) repair. VSP repair corrects T.G mispairs in the following contexts:CTAAGGGGTCC, CTTGGGGACC, TAGGGTCC and CTAGGGTC. I have investigated the relationships between cytosine methylation, mutation, and VSP repair. Spontaneous mutations in the repressor (cI) gene of lambda prophage were isolated in wild-type and mutant lysogens. A hotspot for spontaneous mutation that corresponds with a 5-methylcytosine was observed in wild-type lysogens but was not present in bacteria lacking both methylase and VSP repair activity. Introduction of a plasmid containing dcm+ and vsr+ restored the mutation hotspot. If the added plasmid carried only dcm+, the frequency of spontaneous mutations at the 5-methylcytosine was over 10-fold higher than in Dcm+Vsr+ lysogens. The addition of vsr on a plasmid to a wild-type lysogen resulted in a 4-fold reduction in mutation at the hotspot. These findings support the previously untested hypothesis that VSP repair prevents mutations resulting from deamination of 5-methylcytosine.     

Very short patch repair: reducing the cost of cytosine methylation

In Escherichia coli and related bacteria, the product of gene dcm methylates the second cytosine of 5'-CCWGG sequences (where W is A or T). Deamination of 5-methylcytosine (5meC) results in C to T mutations. The mutagenic potential of 5meC is reduced by a system called very short patch (VSP) repair, which replaces T with C. T:G and U:G mispairs in the methylatable sequence and in related sequences are recognized by the product of vsr , a gene adjacent to dcm . Vsr creates a nick just 5' of the mispaired pyrimidine to initiate the repair. Additional products known to be required for VSP repair are DNA polymerase I and DNA ligase. MutS and MutL have a stimulatory role but are not required. The ability of Vsr to recognize T:G mispairs in sequences related to CCWGG is probably responsible for over- and under-representation of certain tetranucleotides in the E. coli genome. Although VSP repair reduces spontaneous mutations at 5meCs in replicating bacteria, mutation hot-spots persist at these sites. Under conditions that more accurately mimic the natural environment of E. coli , VSP repair appears to be effective in preventing mutation at 5meC.     

Bacterial genes mutL, mutS, and dcm participate in repair of mismatches at 5-methylcytosine sites.

Certain amber mutations in the cI gene of bacteriophage lambda appear to recombine very frequently with nearby mutations. The aberrant mutations included C-to-T transitions at the second cytosine in 5'CC(A/T)GG sequences (which are subject to methylation by bacterial cytosine methylase) and in 5'CCAG and 5'CAGG sequences. Excess cI+ recombinants arising in crosses that utilize these mutations are attributable to the correction of mismatches by a bacterial very-short-patch (VSP) mismatch repair system. In the present study I found that two genes required for methyladenine-directed (long-patch) mismatch repair, mutL and mutS, also functioned in VSP mismatch repair; mutH and mutU (uvrD) were dispensable. VSP mismatch repair was greatly reduced in a dcm Escherichia coli mutant, in which 5-methylcytosine was not methylated. However, mismatches in heteroduplexes prepared from lambda DNA lacking 5-methylcytosine were repaired in dcm+ bacteria. These results indicate that the product of gene dcm has a repair function in addition to its methylase activity.     

DNA mismatch correction by Very Short Patch repair may have altered the abundance of oligonucleotides in the E. coli genome.

A base mismatch correction process in E. coli K-12 called Very Short Patch (VSP) repair corrects T:G mismatches to C:G when found in certain sequence contexts. Two of the substrate mismatches (5'-CTWGG/3'-GGW'CC; W = A or T) occur in the context of cytosine methylation in DNA and reduce the mutagenic effects of 5-methylcytosine deamination to thymine. However, VSP repair is also known to repair T:G mismatches that are not expected to arise from 5-methylcytosine deamination (example--CTAG/GGT-C). In these cases, if the original base pair were a T:A, VSP repair would cause a T to C transition. We have carried out Markov chain analysis of an E. coli sequence database to determine if repair at the latter class of sites has altered the abundance of the relevant tetranucleotides. The results are consistent with the prediction that VSP repair would tend to deplete the genome of the 'T' containing sequences (example--CTAG), while enriching it for the corresponding 'C' containing sequences (CCAG). Further, they provide an explanation for the known scarcity of CTAG containing restriction enzyme sites among the genomes of enteric bacteria and identify VSP repair as a force in shaping the sequence composition of bacterial genomes.     

A DNA repair process in Escherichia coli corrects U:G and T:G mismatches to C:G at sites of cytosine methylation

Escherichia coli contains a base mismatch correction system called VSP repair that is known to correct T:G mismatches to C:G when they occur in certain sequence contexts. The preferred sequence context for this process is the site for methylation by the E. coli DNA cytosine methylase (Dcm). For this reason, VSP repair is thought to counteract potential mutagenic effects of deamination of 5-methylcytosine to thymine. We have developed a genetic reversion assay that quantitates the frequency of C to T mutations at Dcm sites and the removal of such mutations by DNA repair processes. Using this assay, we have studied the repair of U: G mismatches in DNA to C: G and have found that VSP repair is capable of correcting these mismatches. Although VSP repair substantially affects the reversion frequency, it may not be as efficient at correcting U: G mismatches as the uracil DNA glycosylase-mediated repair process.     

A DNA repair process in Escherichia coli corrects U:G and T:G mismatches to C:G at sites of cytosine methylation

Escherichia coli contains a base mismatch correction system called VSP repair that is known to correct T:G mismatches to C:G when they occur in certain sequence contexts. The preferred sequence context for this process is the site for methylation by the E. coli DNA cytosine methylase (Dcm). For this reason, VSP repair is thought to counteract potential mutagenic effects of deamination of 5-methylcytosine to thymine. We have developed a genetic reversion assay that quantitates the frequency of C to T mutations at Dcm sites and the removal of such mutations by DNA repair processes. Using this assay, we have studied the repair of U: G mismatches in DNA to C: G and have found that VSP repair is capable of correcting these mismatches. Although VSP repair substantially affects the reversion frequency, it may not be as efficient at correcting U: G mismatches as the uracil DNA glycosylase-mediated repair process.     

Very Short Patch Mismatch Repair in Phage Lambda: Repair Sites and Length of Repair Tracts

Five amber mutations in the repressor (cI) gene of bacteriophage lambda recombine anomalously with nearby cI mutations. When any of these markers is used in four-factor crosses, cI+ recombinants that are expected to require three cross-overs occur at high frequencies. These recombinants are attributable to very-short-patch (VSP) repair of specific mismatches in DNA heteroduplexes formed during recombination between the markers flanking cI. The sites of the repair-prone mutations and the lengths of repair tracts have now been determined. Amber mutations subject to VSP repair are C to T transitions in 5'CCATGG, the sequence methylated by the product of gene dcm, and also in the related 5'CAGG or 5'CCAG sequences. Ambers arising in CAG sequences found in other contexts, or in codons other than CAG, were not subject to VSP repair. Repair tracts rarely, if ever, exceed ten nucleotides in length, and can be as short as two nucleotides. A repair-prone mutation does not stimulate recombination between flanking cI markers.     

DNA binding and cleavage selectivity of the Escherichia coli DNA G:T-mismatch endonuclease (vsr protein).

The Escherichia coli vsr endonuclease recognises T:G base-pair mismatches in double-stranded DNA and initiates a repair pathway by hydrolysing the phosphate group 5' to the incorrectly paired T. The gene encoding the vsr endonuclease is next to the gene specifying the E. coli dcm DNA-methyltransferase; an enzyme that adds CH3 groups to the first dC within its target sequence CC[A/T]GG, giving C5MeC[A/T]GG. Deamination of the d5MeC results in CT[A/T]GG in which the first T is mis-paired with dG and it is believed that the endonuclease preferentially recognises T:G mismatches within the dcm recognition site. Here, the preference of the vsr endonuclease for bases surrounding the T:G mismatch has been evaluated. Determination of specificity constant (kst/KD; kst = rate constant for single turnover, KD = equilibrium dissociation constant) confirms vsr's preference for a T:G mismatch within a dcm sequence i.e. CT[A/T]GG (the underlined T being mis-paired with dG) is the best substrate. However, the enzyme is capable of binding and hydrolysing sequences that differ from the dcm target site by a single base-pair (dcm star sites). Individual alteration of any of the four bases surrounding the mismatched T gives a substrate, albeit with reduced binding affinity and slowed turnover rates. The vsr endonuclease has a much lower selectivity for the dcm sequence than type II restriction endonucleases have for their target sites. The results are discussed in the light of the known crystal structure of the vsr protein and its possible physiological role.     

Statistical evaluation and biological interpretation of non-random abundance in the E. coli K-12 genome of tetra- and pentanucleotide sequences related to VSP DNA mismatch repair.

The abundance of all tetra- and pentanucleotide sequences is calculated for a set of DNA sequence data comprising 767,393 nucleotides of the E. coli K-12 genome. Observed frequencies are compared to those expected from a Markov chain prediction algorithm. Systematic and extreme non-random representations are found for special sets of sequences. These are interpreted as arising from incorporation of a 2'-deoxyguanosine residue opposite thymidine during replication which, in special sequence contexts, leads to a T/G mismatch that is simultaneously substrate for two competing DNA mismatch repair systems: the mutHLS and the VSP pathway. Processing by the former leads to error correction, by the latter to mutation fixation. The significance of the latter process, as demonstrated here, makes it unlikely that VSP repair has evolved mainly as a mutation avoidance mechanism. It is proposed that in E. coli K-12, VSP repair, together with DNA cytosine methylation, constitutes a mutagenesis/recombination system capable of promoting gene-conversion-like unidirectional transfer of short stretches of DNA sequence.     

A gene required for very short patch repair in Escherichia coli is adjacent to the DNA cytosine methylase gene.

Deamination of 5-methylcytosine in DNA results in T/G mismatches. If unrepaired, these mismatches can lead to C-to-T transition mutations. The very short patch (VSP) repair process in Escherichia coli counteracts the mutagenic process by repairing the mismatches in favor of the G-containing strand. Previously we have shown that a plasmid containing an 11-kilobase fragment from the E. coli chromosome can complement a chromosomal mutation defective in both cytosine methylation and VSP repair. We have now mapped the regions essential for the two phenotypes. In the process, we have constructed plasmids that complement the chromosomal mutation for methylation, but not for repair, and vice versa. The genes responsible for these phenotypes have been identified by DNA sequence analysis. The gene essential for cytosine methylation, dcm, is predicted to code for a 473-amino-acid protein and is not required for VSP repair. It is similar to other DNA cytosine methylases and shares extensive sequence similarity with its isoschizomer, EcoRII methylase. The segment of DNA essential for VSP repair contains a gene that should code for a 156-amino-acid protein. This gene, named vsr, is not essential for DNA methylation. Remarkably, the 5' end of this gene appears to overlap the 3' end of dcm. The two genes appear to be transcribed from a common promoter but are in different translational registers. This gene arrangement may assure that Vsr is produced along with Dcm and may minimize the mutagenic effects of cytosine methylation.     

Processing of mispaired and unpaired bases in heteroduplex DNA in E. coli

Bacteriophage lambda and phi X 174 DNAs, carrying sequenced mutations, have been used to construct in vitro defined species of heteroduplex DNA. Such heteroduplex DNAs were introduced by transfection, as single copies, into E. coli host cells. The progeny of individual heteroduplex molecules from each infective center was analyzed. The effect of the presence of GATC sequences (phi X 174 system) and of their methylation (lambda system) was tested. The following conclusions can be drawn: some mismatched base pairs trigger the process of mismatch repair, causing a localized strand-to-strand information transfer in heteroduplex DNA: transition mismatches G:T and A:C are efficiently repaired, whereas the six transversion mismatches are not always readily recognized and/or repaired. The recognition of transversion mismatches appears to depend on the neighbouring nucleotide sequence; single unpaired bases (frameshift mutation "mismatches") are recognized and repaired, some equally efficiently on both strands (longer and shorter), some more efficiently on the shorter (-1) strand; large non-homologies (about 800 bases) are not repaired by the Mut H, L, S, U system, but some other process repairs the non-homology with a relatively low efficiency; full methylation of GATC sequences inhibits mismatch repair on the methylated strand: this is the chemical basis of strand discrimination (old/new) in mismatch correction; unmethylated GATC sequences appear to improve mismatch repair of a G:T mismatch in phi X 174 DNA, but there may be some residual mismatch repair in GATC-free phi X 174, at least for some mismatches.(ABSTRACT TRUNCATED AT 250 WORDS)     

Repair of a Mismatch Is Influenced by the Base Composition of the Surrounding Nucleotide Sequence

Heteroduplexes with single base pair mismatches of known sequence were prepared by annealing separated strands of bacteriophage lambda DNA and used to transfect Escherichia coli. A series of transition (G:T and A:C) and transversion (G:A and C:T) mismatches located throughout most of the bacteriophage lambda cI gene has been examined. The results suggest that the transition mismatches are generally better repaired than the transversion mismatches and that, at least for the transversion mismatches studied, repair efficiency increases with increasing G:C content in the neighboring nucleotide sequence. This specificity of the E. coli mismatch repair system can account, in part, for the similar frequencies of base substitution mutations throughout the E. coli genome.     

Spontaneous Mutations Occur near Dam Recognition Sites in a dam-Escherichia coli Host

The mismatch repair system of Escherichia coli K12 removes mispaired bases from DNA. Mismatch repair can occur on either strand of DNA if it lacks N6-methyladenines within 5'-GATC-3' sequences. In hemimethylated heteroduplexes, repair occurs preferentially on the unmethylated strand. If both strands are fully methylated, repair is inhibited. Mutant (dam-) strains of E. coli defective in the adenine methylase that recognizes 5'-GATC-3' sequences (Dam), and therefore defective in mismatch repair, show increased spontaneous mutation rates compared to otherwise isogenic dam+ hosts. We have isolated and characterized 91 independent mutations that arise as a consequence of the Dam- defect in a plasmid-borne phage P22 repressor gene, mnt. The majority of these mutations are A:T----G:C transitions that occur within six base pairs of the two 5'-GATC-3' sequences in the mnt gene. In contrast, the spectrum of mnt- mutations in a dam+ host is comprised of a majority of insertions of IS elements and deletions that do not cluster near Dam recognition sites. These results show that Dam-directed post-replicative mismatch repair plays a significant role in the rectification of potential transition mutations in vivo, and suggest that sequences associated with Dam recognition sites are particularly prone to replication or repair errors.     

Escherichia Coli Mutator Mutd5 Is Defective in the Muthls Pathway of DNA Mismatch Repair

We have previously reported that the Escherichia coli mutator strain mutD5 was defective in the correction of bacteriophage M13mp2 heteroduplex DNA containing a T.G mismatch. Here, this defect was further investigated with regard to its interaction with the mutHLS pathway of mismatch repair. A set of 15 different M13mp2 heteroduplexes was used to measure the mismatch-repair capability of wild-type, mutL and mutD5 cells. Throughout the series, the mutD5 strain proved as deficient in mismatch repair as the mutL strain, indicating that the repair defect is similar in the two strains in both extent and specificity. [One exception was noted in the case a T.G mispair that was subject to VSP (Very Short Patch) repair. VSP repair was abolished by mutL but not by mutD.] Variation in the dam-methylation state of the heteroduplex molecules clearly affected repair in the wild-type strain but had no effect on either the mutD or mutL strain. Finally, mutDmutL or mutDmutS double-mutator strains were no more deficient in mismatch repair as were the single mutator strains. The combined results strongly argue that the mismatch-repair deficiency of mutD5 cells resides in the mutH,L,S-dependent pathway of mismatch repair and that the high mutation rate of mutD strains derives in part from this defect.     

DNA-substrate sequence specificity of human G:T mismatch repair activity.

G:T mispairs in DNA originate spontaneously via deamination of 5-methylcytosine. Such mispairs are restored to normal G:C pairs by both E. coli K strains and human cells. In this study we have analyzed the repair by human cell extracts of G:T mismatches in various DNA contexts. We performed two sets of experiments. In the first, repair was sequence specific in that G:T mispairs at CpG sites at four different CpG sites were repaired, but a G:T mismatch at a GpG site was not. Cytosine hemimethylation did not block repair of a substrate containing a CpG/GpT mismatch. In the second set of experiments, substrates with a G:T mismatch at a fixed position were constructed with an A, T, G, or C 5' to the mismatched G, and alterations in the complementary strand to allow otherwise perfect Watson-Crick pairing. All were incised just 5' to the mismatched T and competed for repair incision with a G:T substrate in which a C was 5' to the mismatched G. Thus human G:T mismatch activity shows sequence specificity, incising G:T mismatched pairs at some DNA sites, but not at others. At an incisable site, however, incision is little influenced by the base 5' to the mismatched G.     

Chromosomal model for analysis of a long CTG/CAG tract stability in wild-type Escherichia coli and its nucleotide excision repair mutants

Many human hereditary neurological diseases, including fragile X syndrome, myotonic dystrophy, and Friedreich's ataxia, are associated with expansions of the triplet repeat sequences (TRS) (CGG/CCG, CTG/CAG, and GAA/TTC) within or near specific genes. Mechanisms that mediate mutations of TRS include DNA replication, repair, and gene conversion and (or) recombination. The involvement of the repair systems in TRS instability was investigated in Escherichia coli on plasmid models, and the results showed that the deficiency of some nucleotide excision repair (NER) functions dramatically affects the stability of long CTG inserts. In such models in which there are tens or hundreds of plasmid molecules in each bacterial cell, repetitive sequences may interact between themselves and according to a recombination hypothesis, which may lead to expansions and deletions within such repeated tracts. Since one cannot control interaction between plasmids, it is also sometimes difficult to give precise interpretation of the results. Therefore, using modified lambda phage (lambdaInCh), we have constructed a chromosomal model to study the instability of trinucleotide repeat sequences in E. coli. We have shown that the stability of (CTG/CAG)68 tracts in the bacterial chromosome is influenced by mutations in NER genes in E. coli. The absence of the uvrC or uvrD gene products greatly enhances the instability of the TRS in the chromosome, whereas the lack of the functional UvrA or UvrB proteins causes substantial stabilization of (CTG/CAG) tracts.     

Isolation and characterization of an Escherichia coli strain with a high frequency of C-to-T mutations at 5-methylcytosines.

We used a genetic selection system to isolate a strain of Escherichia coli with a high frequency of C-to-T transition mutations at the second C of the sequence CCAGG. Cytosines in other sequences do not mutate to thymine at a high frequency in this strain, and the frequencies of other base substitution mutations are not increased to the same extent. The gene responsible for the mutator phenotype has been mapped to 43 min on the E. coli chromosome. Several lines of evidence indicate that this gene is distinct from the very short patch repair gene vsr.     

The role of the Escherichia coli mug protein in the removal of uracil and 3,N(4)-ethenocytosine from DNA.

The human thymine-DNA glycosylase has a sequence homolog in Escherichia coli that is described to excise uracils from U.G mismatches (Gallinari, P., and Jiricny, J. (1996) Nature 383, 735-738) and is named mismatched uracil glycosylase (Mug). It has also been described to remove 3,N(4)-ethenocytosine (epsilonC) from epsilonC.G mismatches (Saparbaev, M., and Laval, J. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 8508-8513). We used a mug mutant to clarify the role of this protein in DNA repair and mutation avoidance. We find that inactivation of mug has no effect on C to T or 5-methylcytosine to T mutations in E. coli and that this contrasts with the effect of ung defect on C to T mutations and of vsr defect on 5-methylcytosine to T mutations. Even under conditions where it is overproduced in cells, Mug has little effect on the frequency of C to T mutations. Because uracil-DNA glycosylase (Ung) and Vsr are known to repair U.G and T.G mismatches, respectively, we conclude that Mug does not repair U.G or T.G mismatches in vivo. A defect in mug also has little effect on forward mutations, suggesting that Mug does not play a role in avoiding mutations due to endogenous damage to DNA in growing E. coli. Cell-free extracts from mug(+) ung cells show very little ability to remove uracil from DNA, but can excise epsilonC. The latter activity is missing in extracts from mug cells, suggesting that Mug may be the only enzyme in E. coli that can remove this mutagenic adduct. Thus, the principal role of Mug in E. coli may be to help repair damage to DNA caused by exogenous chemical agents such as chloroacetaldehyde.     

Structures of mismatched base pairs in DNA and their recognition by the Escherichia coli mismatch repair system.

The Escherichia coli mismatch repair system does not recognize and/or repair all mismatched base pairs with equal efficiency: whereas transition mismatches (G X T and A X C) are well repaired, the repair of some transversion mismatches (e.g. A X G or C X T) appears to depend on their position in heteroduplex DNA of phage lambda. Undecamers were synthesized and annealed to form heteroduplexes with a single base-pair mismatch in the centre and with the five base pairs flanking each side corresponding to either repaired or unrepaired heteroduplexes of lambda DNA. Nuclear magnetic resonance (n.m.r.) studies show that a G X A mismatch gives rise to an equilibrium between fully helical and a looped-out structure. In the unrepaired G X A mismatch duplex the latter predominates, while the helical structure is predominant in the case of repaired G X A and G X T mismatches. It appears that the E. coli mismatch repair enzymes recognize and repair intrahelical mismatched bases, but not the extrahelical bases in the looped-out structures.     

SOS mutator effect in E. coli mutants deficient in mismatch correction.

We have used bacteriophage lambda to characterize the mutator effect of the SOS response induced by u.v. irradiation of Escherichia coli. Mutagenesis of unirradiated phages grown in irradiated or unirradiated bacteria was detected by measuring forward mutagenesis in the immunity genes or reversion mutagenesis of an amber codon in the R gene. Relative to the wild-type, the SOS mutator effect was higher in E. coli mismatch correction-deficient mutants (mutH, mutL and mutS) and lower in an adenine methylation-deficient mutant ( dam3 ). We conclude that a large proportion of SOS-induced 'untargeted' mutations are removed by the methyl-directed mismatch correction system, which acts on newly synthesized DNA strands. The lower SOS mutator effect observed in E. coli dam mutants may be due to a selective killing of mismatch-bearing chromosomes resulting from undirected mismatch repair. The SOS mutator effect on undamaged lambda DNA, induced by u.v. irradiation of the host, appears to result from decreased fidelity of DNA synthesis.