Nature of the SOS mutator activity: genetic characterization of untargeted mutagenesis in Escherichia coli

Summary In Escherichia coli, induction of the SOS functions by UV irradiation or by mutation in the recA gene promotes an SOS mutator activity which generates mutations in undamaged DNA. Activation of RecA protein by the recA730 mutation increases the level of spontaneous mutation in the bacterial DNA. The number of recA730-induced mutations is greatly increased in mismatch repair deficient strains in which replication errors are not corrected. This suggests that the majority of recA730-induced mutations (90%) arise through correctable, i.e. non-targeted, replication errors. This recA730 mutator effect is suppressed by a mutation in the umuC gene. We also found that dam recA730 double mutants are unstable, segregating clones that have lost the dam or the recA mutations or that have acquired a new mutation, probably in one of the genes involved in mismatch repair. We suggest that the genetic instability of the dam recA730 mutants is provoked by the high level of replication errors induced by the recA730 mutation, generating killing by coincident mismatch repair on the two unmethylated DNA strands. The recA730 mutation increases spontaneous mutagenesis of phage poorly. UV irradiation of recA730 host bacteria increases phage untargeted mutagenesis to the level observed in UV-irradiated recA ⁺ strains. This UV-induced mutator effect in recA730 mutants is not suppressed by a umuC mutation. Therefore UV and the recA730 mutation seem to induce different SOS mutator activities, both generating untargeted mutations.     

Genetic characterization of the inducible SOS-like system of Bacillus subtilis.

The SOS-like system of Bacillus subtilis consists of several coordinately induced phenomena (e.g., cellular filamentation, prophage induction, and Weigle reactivation of UV-damaged bacteriophage) which are expressed after cellular insult such as DNA damage or inhibition of DNA replication. Mutagenesis of the bacterial chromosome and the development or maintenance of competence also appear to be involved in the SOS-like response in this bacterium. The genetic characterization of the SOS-like system has involved an analysis of (i) the effects of various DNA repair mutations on the expression of inducible phenomena and (ii) the tsi-23 mutation, which renders host strains thermally inducible for each of the SOS-like functions. Bacterial filamentation was unaffected by any of the DNA repair mutations studied. In contrast, the induction of prophage after thermal or UV pretreatment was abolished in strains carrying the recE4, recA1, recB2, or recG13 mutation. The Weigle reactivation of UV-damaged bacteriophage was also inhibited by the recE4, recA1, recB2, or recG13 mutation, whereas levels of Weigle reactivation were lower in strains which carried the uvrA42, polA5, or rec-961 mutation than in the DNA repair-proficient strain. Strains which carried the recE4 mutation were incapable of chromosomal DNA-mediated transformation, and the frequency of this event was decreased in strains carrying the recA1, recB2, or tsi-23 mutation. Plasmid DNA transformation efficiency was decreased only in strains carrying the tsi-23 mutation in addition to the recE4, recA1, or recB2 mutation. The results indicate that the SOS-like system of B. subtilis is regulated at different levels by two or more gene products. In this report, the current data regarding the genetic regulation of inducible phenomena are summarized, and a model is proposed to explain the mechanism of SOS-like induction in B. subtilis.     

Cloning and DNA sequence of a mycoplasmal recA gene.

Mycoplasmas are wall-less prokaryotes phylogenetically related to gram-positive bacteria. In order to investigate DNA recombination in these organisms, we have cloned the recA gene from the mycoplasma Acholeplasma laidlawii. DNA sequence data indicate extensive homology between the A. laidlawii recA gene and recA genes from other bacteria, particularly Bacillus subtilis. The recA sequences from three A. laidlawii strains (strains JA1, K2, and 8195) were compared, and surprisingly, the gene from A. laidlawii 8195 was found to contain a nonsense mutation that results in truncation of 36 amino acids from the carboxyl terminus of the RecA protein. By using sensitivity to UV irradiation as a measure of DNA repair, strain 8195 had an apparent RecA- phenotype. When carried on a multicopy plasmid, the wild-type A. laidlawii recA gene was detrimental to growth of Escherichia coli, perhaps because of improper regulation of the RecA protein.     

Effects of biological DNA precursor pool asymmetry upon accuracy of DNA replication in vitro

Deoxyguanosine triphosphate is underrepresented among the four common deoxyribonucleoside triphosphates (dNTPs), typically accounting for just 5-10% of the total dNTP pool. We have asked whether this pool asymmetry affects the fidelity of DNA replication, by use of an in vitro assay in which an M13 phagemid containing the Escherichia coli lacZalpha gene and an SV40 replication origin is replicated by extracts of human cells. By monitoring reversion of either a TGA or TAA codon within the lacZalpha gene, we found that replication in "biologically biased" dNTPs, representing our estimate of the concentrations in HeLa cell nuclei, is not significantly more accurate than when measured in reaction mixtures containing the four dNTPs at equimolar concentrations. However, sequence analysis of revertants revealed significantly different patterns of mispairing events leading to mutation. During replication at biased dNTP levels, mutations at the site 5' to C in the template strand for the TGA triplet were less frequent than seen in equimolar reaction mixtures, suggesting that extension from mismatches at this site is relatively slow, and proofreading efficiency high, when dGTP is the next nucleotide to be incorporated. Mismatches opposite template C, which might have been favored by the low physiological concentrations of dGTP, were not favored in our in vitro system, although one particular substitution at this site, TGA-->TTA, was strongly favored at low [dGTP]. An excess of one dNTP was found in our system to be more mutagenic than a corresponding deficiency. We also estimated dNTP concentrations in non-transformed human fibroblasts and found that in vitro replication at these levels caused significantly fewer mutations than we observed under equimolar conditions (100 microM each dNTP). This increased replication fidelity may result from increased proofreading efficiency at the lower dNTP levels; however, replication rates were decreased only slightly at these non-transformed fibroblast concentrations.     

Replication mutations differentially enhance RecA-dependent and RecA-independent recombination between tandem repeats in Bacillus subtilis

We have studied DNA recombination between 513 bp tandem direct repeats present in a kanamycin resistance gene inserted in the Bacillus subtilis chromosome. Tandem repeat deletion was not significantly affected by a recA mutation. However, recombination was stimulated by mutations in genes encoding replication proteins, including the primosomal proteins DnaB, DnaD and the DnaG primase, the putative DNA polymerase III subunits PolC, DnaN and DnaX, as well as the DNA polymerase DnaE. Hyper-recombination was found to be dependent on RecA in the dnaE, dnaN and dnaX mutants, whereas the dnaG and dnaD mutants stimulated recombination independently of RecA. Altogether, these data show that both RecA-dependent and RecA-independent mechanisms contribute to recombination between tandem repeats in B. subtilis and that both types of recombination are stimulated by replication mutations.     

Growth-dependent DNA breakage and cell death in a gyrase mutant of Salmonella.

A class of gyrase mutants of Salmonella enterica mimics the properties of bacteria exposed to quinolones. These mutants suffer spontaneous DNA breakage during normal growth and depend on recombinational repair for viability. Unlike quinolone-treated bacteria, however, they do not show accumulation of cleavable gyrase-DNA complexes. In recA or recB mutant backgrounds, the temperature-sensitive (ts) allele gyrA208 causes rapid cell death at 43 degrees. Here, we isolated "suppressor-of-death" mutations, that is, secondary changes that allow a gyrA208 recB double mutant to survive a prolonged exposure to 43 degrees and subsequently to form colonies at 28 degrees. In most isolates, the secondary change was itself a ts mutation. Three ts alleles were mapped in genes coding for amino acyl tRNA synthetases (alaS, glnS, and lysS). Allele alaS216 completely abolished DNA breakage in a gyrA208 recA double mutant. Likewise, treating this mutant with chloramphenicol prevented death and DNA damage at 43 degrees. Additional suppressors of gyrA208 lethality include rpoB mutations and, surprisingly, icd mutations inactivating isocitrate dehydrogenase. We postulate that the primary effect of the gyrase alteration is to hamper replication fork movement. Inhibiting DNA replication under conditions of continuing macromolecular synthesis ("unbalanced growth") activates a mechanism that causes DNA breakage and cell death, reminiscent of "thymineless" lethality.     

Roles of RecA protein in spontaneous mutagenesis in Escherichia coli

To verify the extent of contribution of spontaneous DNA lesions to spontaneous mutagenesis, we have developed a new genetic system to examine simultaneously both forward mutations and recombination events occurring within about 600 base pairs of a transgenic rpsL target sequence located on Escherichia coli chromosome. In a wild-type strain, the recombination events were occurring at a frequency comparable to that of point mutations within the rpsL sequence. When the cells were UV-irradiated, the recombination events were induced much more sharply than point mutations. In a recA null mutant, no recombination event was observed. These data suggest that the blockage of DNA replication, probably caused by spontaneous DNA lesions, occurs often in normally growing E. coli cells and is mainly processed by cellular functions requiring the RecA protein. However, the recA mutant strain showed elevated frequencies of single-base frameshifts and large deletions, implying a novel mutator action of this strain. A similar mutator action of the recA mutant was also observed with a plasmid-based rpsL mutation assay. Therefore, if the recombinogenic problems in DNA replication are not properly processed by the RecA function, these would be a potential source for mutagenesis leading to single-base frameshift and large deletion in E. coli. Furthermore, the single-base frameshifts induced in the recA-deficient cells appeared to be efficiently suppressed by the mutS-dependent mismatch repair system. Thus, it seems likely that the single-base frameshifts are derived from slippage errors that are not directly caused by DNA lesions but made indirectly during some kind of error-prone DNA synthesis in the recA mutant cells.     

RNase H and replication of ColE1 DNA in Escherichia coli

Amber mutations within the rnh (RNase H) gene of Escherichia coli K-12 were isolated by selecting for bacteria capable of replicating in a sup+ background replication-defective cer-6 mutant of the ColE1 replicon. The cer-6 mutation is an alteration of one base pair located 160 nucleotides upstream of the unique replication origin of this plasmid. Subsequently, we determined the DNA alterations present within these mutants. ColE1 DNA replicated in rnh(Am) recA cells, indicating that (i) RNase H, which has been shown to be absolutely required for in vitro initiation of ColE1 DNA replication, is dispensable in vivo, and (ii) ColE1 replication in the absence of RNase H is not dependent on "stable DNA replication," which has been reported to be an alternative mode of chromosomal DNA replication. Another class of bacterial mutations was also isolated. These mutations, named herB, suppressed cer-6 replication in rnh+ bacteria. herB mutations mapped close to the polA gene on the E. coli chromosome and increased the activity of DNA polymerase I. These findings suggest that when the DNA polymerase I has an opportunity to initiate DNA synthesis before RNase H acts, the replication-defective cer-6 mutant or the wild-type ColE1 replicates in E. coli.     

Therefore,what are recombination proteins there for?

The order of discovery can have a profound effect upon the way in which we think about the function of a gene. In E. coli, recA is nearly essential for cell survival in the presence of DNA damage. However, recA was originally identified, as a gene required to obtain recombinant DNA molecules in conjugating bacteria. As a result, it has been frequently assumed that recA promotes the survival of bacteria containing DNA damage by recombination in which DNA strand exchanges occur. We now know that several of the processes that interact with or are controlled by recA, such as excision repair and translesion synthesis, operate to ensure that DNA replication occurs processively without strand exchanges. Yet the view persists in the literature that recA functions primarily to promote recombination during DNA repair. With the benefit of hindsight and more than three decades of additional research, we reexamine some of the classical experiments that established the concept of DNA repair by recombination, and we consider the possibilities that recombination is not an efficient mechanism for rescuing damaged cells, and that recA may be important for maintaining processive replication in a manner that does not generally promote recombination.     

Mutation and killing of Escherichia coli expressing a cloned Bacillus subtilis gene whose product alters DNA conformation.

Expression of the Bacillus subtilis gene coding for SspC, a small, acid-soluble protein, caused both killing and mutation in a number of Escherichia coli B and K-12 strains. SspC was previously shown to bind E. coli DNA in vivo, and in vitro this protein binds DNA and converts it into an A-like conformation. Analysis of revertants of nonsense mutations showed that SspC caused single-base changes, and a greater proportion of these were at A-T base pairs. Mutation in the recA gene abolished the induction of mutations upon synthesis of SspC, but the killing was only slightly greater than in RecA+ cells. Mutations in the umuC and umuD genes eliminated most of the mutagenic effect of SspC but not the killing, while the lexA mutation increased mutagenesis but did not appreciably affect the killing. Since there was neither killing nor mutation of E. coli after synthesis of a mutant SspC which does not bind DNA, it appears likely that the binding of wild-type SspC to DNA, with the attendant conformational change, was responsible for the killing and mutation. A strain containing the B. subtilis gene that is constitutive for the RecA protein at 42 degrees C showed a lower frequency of mutation when that temperature was used to induce the RecA protein than when the temperature was 30 degrees C, where the RecA level is low, suggesting that at the elevated temperature the high RecA level could be inhibiting binding of the B. subtilis protein to DNA.     

recA-dependent and recA-independent N-ethyl-N-nitrosourea mutagenesis at a plasmid-encoded herpes simplex virus thymidine kinase gene in Escherichia coli

We have compared isogenic recA13/recA+ Escherichia coli K-12 strains for the induction by N-ethyl-N-nitrosourea (ENU) of forward mutations at a plasmid-encoded herpes simplex virus type 1 thymidine kinase (HSV-tk) gene. Treatment of plasmid-bearing bacteria with ENU resulted in a dose-dependent increase in the mutant frequencies of the chromosomal udk locus and of the plasmid HSV-tk locus in both recA13 and recA+ strains. Although the recA13 strain was considerably more sensitive to the cytotoxic effects of ENU treatment than was the recA+ strain, the ENU-induced mutation frequency at both loci was greater for the recA+ strain than for the recA13 strain. When plasmid DNA modified by in vitro reaction with ENU was used to transform recA13, recA+, and UV pre-irradiated recA+ strains, an increase in the HSV-tk mutant frequency was observed in all 3 cases. The induction of mutations in recA13 and recA+ strains followed a similar dose-response, while the ENU-induced HSV-tk mutant frequency was significantly greater for UV pre-irradiated recA+ bacteria. These results indicate that fixation of ENU-induced premutagenic lesions can occur by both recA-dependent and recA-independent pathways.     

Genetic recombination destabilizes (CTG)n.(CAG)n repeats in E. coli

The expansion of trinucleotide repeats has been implicated in 17 neurological diseases to date. Factors leading to the instability of trinucleotide repeat sequences have thus been an area of intense interest. Certain genes involved in mismatch repair, recombination, nucleotide excision repair, and replication influence the instability of trinucleotide repeats in both Escherichia coli and yeast. Using a genetic assay for repeat deletion in E. coli, the effect of mutations in the recA, recB, and lexA genes on the rate of deletion of (CTG)n.(CAG)n repeats of varying lengths were examined. The results indicate that mutations in recA and recB, which decrease the rate of recombination, had a stabilizing effect on (CAG)n.(CTG)n repeats decreasing the high rates of deletion seen in recombination proficient cells. Thus, recombination proficiency correlates with high rates of genetic instability in triplet repeats. Induction of the SOS system, however, did not appear to play a significant role in repeat instability, nor did the presence of triplet repeats in cells turn on the SOS response. A model is suggested where deletion during exponential growth may result from attempts to restart replication when paused at triplet repeats.     

The Bacillus subtilis SftA (YtpS) and SpoIIIE DNA translocases play distinct roles in growing cells to ensure faithful chromosome partitioning

In several bacterial species, the faithful completion of chromosome partitioning is known to be promoted by a conserved family of DNA translocases that includes Escherichia coli FtsK and Bacillus subtilis SpoIIIE. FtsK localizes at nascent division sites during every cell cycle and stimulates chromosome decatenation and the resolution of chromosome dimers formed by recA -dependent homologous recombination. In contrast, SpoIIIE localizes at sites where cells have divided and trapped chromosomal DNA in the membrane, which happens during spore development and under some conditions when DNA replication is perturbed. SpoIIIE completes chromosome segregation post-septationally by translocating trapped DNA across the membrane. Unlike E. coli , B. subtilis contains a second uncharacterized FtsK/SpoIIIE-like protein, SftA (formerly YtpS). We report that SftA plays a role similar to FtsK during each cell cycle but cannot substitute for SpoIIIE in rescuing trapped chromosomes. SftA colocalizes with FtsZ at nascent division sites but not with SpoIIIE at sites of chromosome trapping. SftA mutants divide over unsegregated chromosomes more frequently than wild-type unless recA is inactivated, suggesting that SftA, like FtsK, stimulates chromosome dimer resolution. Having two FtsK/SpoIIIE paralogues is not conserved among endospore-forming bacteria, but is highly conserved within several groups of soil- and plant-associated bacteria.     

Characterization of the promoter of the Rhizobium etli recA gene.

The promoter of the Rhizobium etli recA gene has been identified by primer extension and by making deletions affecting several regions located upstream of its coding region. A gel mobility shift assay carried out with crude extracts of cells of R. etli has been used to show that a DNA-protein complex is formed in the R. etli recA promoter region in vitro. Analysis of the minimal region of the recA promoter giving rise to this DNA-protein complex revealed the presence of an imperfect palindrome corresponding to the sequence TTGN11CAA. Site-directed mutation of both halves of this palindrome indicated that both motifs, TTG and CAA, are necessary for both normal DNA-protein complex formation in vitro and full DNA damage-mediated inducibility of the recA gene in vivo. However, the TTG motif seems to be more dispensable than the CAA one. The presence of this same palindrome upstream of the recA genes of Rhizobium meliloti and Agrobacterium tumefaciens, whose expression is also regulated in R. etli cells, suggests that this TTGN11CAA sequence may be the SOS box of at least these three members of the Rhizobiaceae.     

Tus-mediated arrest of DNA replication in Escherichia coli is modulated by DNA supercoiling

In the absence of RecA, expression of the Tus protein of Escherichia coli is lethal when ectopic Ter sites are inserted into the chromosome in an orientation that blocks completion of chromosome replication. Using this observation as a basis for genetic selection, an extragenic suppressor of Tus-mediated arrest of DNA replication was isolated with diminished ability of Tus to halt DNA replication. Resistance to tus expression mapped to a mutation in the stop codon of the topA gene (topA869), generating an elongated topoisomerase I protein with a marked reduction in activity. Other alleles of topA with mutations in the carboxyl-terminal domain of topoisomerase I, topA10 and topA66, also rendered recA strains with blocking Ter sites insensitive to tus expression. Thus, increased negative supercoiling in the DNA of these mutants reduced the ability of Tus-Ter complexes to arrest DNA replication. The increase in superhelical density did not diminish replication arrest by disrupting Tus-Ter interactions, as Tus binding to Ter sites was essentially unaffected by the topA mutations. The topA869 mutation also relieved the requirement for recombination functions other than recA to restart replication, such as recC, ruvA and ruvC, indicating that the primary effect of the increased negative supercoiling was to interfere with Tus blockage of DNA replication. Introduction of gyrB mutations in combination with the topA869 mutation restored supercoiling density to normal values and also restored replication arrest at Ter sites, suggesting that supercoiling alone modulated Tus activity. We propose that increased negative supercoiling enhances DnaB unwinding activity, thereby reducing the duration of the Tus-DnaB interaction and leading to decreased Tus activity.     

SOS-like induction in Bacillus subtilis: induction of the RecA protein analog and a damage-inducible operon by DNA damage in Rec+ and DNA repair-deficient strains.

We quantitated the induction of the Bacillus subtilis Rec protein (the analog of Escherichia coli RecA protein) and the B. subtilis din-22 operon (representative of a set of DNA damage-inducible operons in B. subtilis) following DNA damage in Rec+ and DNA repair-deficient strains. After exposure to mitomycin C or UV irradiation, each of four distinct rec (recA1, recB2, recE4, and recM13) mutations reduced to the same extent the rates of both Rec protein induction (determined by densitometric scanning of immunoblot transfers) and din-22 operon induction (determined by assaying beta-galactosidase activity in din-22::Tn917-lacZ fusion strains). The induction deficiencies in recA1 and recE4 strains were partially complemented by the E. coli RecA protein, which was expressed on a plasmid in B. subtilis; the E. coli RecA protein had no effect on either induction event in Rec+, recB2, or recM13 strains. These results suggest that (i) the expression of both the B. subtilis Rec protein and the din-22 operon share a common regulatory component, (ii) the recA1 and recE4 mutations affect the regulation and/or activity of the B. subtilis Rec protein, and (iii) an SOS regulatory system like the E. coli system is highly conserved in B. subtilis. We also showed that the basal level of B. subtilis Rec protein is about 4,500 molecules per cell and that maximum induction by DNA damage causes an approximately fivefold increase in the rate of Rec protein accumulation.     

Molecular characterization of regulatory elements controlling expression of the Bacillus subtilis recA+ gene.

Expression of the Bacillus subtilis recA gene is induced following DNA damage as well as during the development of the competent state. DNA damage-induction of the recA gene occurs by a RecA-dependent mechanism, whereas competence-induction occurs by a RecA-independent mechanism. To examine the molecular mechanisms that control the expression of the recA gene, a deletion analysis of the recA promoter region was performed. A regulatory region that is required for repression of recA expression was identified upstream of the recA promoter. Deletion of this regulatory region derepressed expression and abolished damage-induction of the recA promoter. Within this region are sequences similar to the consensus sequence that has been identified within DNA damage-inducible promoter regions of other B subtilis genes. Another regulatory region was identified that is required for the RecA-independent, competence-specific induction of the recA gene. Deletion of these sequences significantly reduced competence-induction of the recA promoter.     

Concatemer formation of ColE1-type plasmids in mutants of Escherichia coli lacking RNase H activity

rnh mutants harboring pBR322 were found to contain several slowly migrating DNA species when examined by agarose gel electrophoresis. The plasmid DNA from rnh mutants included large molecules, i.e. plasmids two, three or four times the size of a single plasmid unit. That this DNA contained concatemeric plasmid joined in a head-to-tail fashion was determined by digestion with restriction endonucleases that cleaved the monomeric plasmid DNA at a unique site. This treatment resulted in migration of the plasmid DNA at a mobility identical to that of linearized monomeric plasmid by agarose gel electrophoresis. This was confirmed by electron microscopy. Plasmid concatemer formation was detected with several high-copy-number (relaxed type) plasmids but not with low-copy-number (stringent) plasmids. Concatemer formation was dependent on RecA+ and RecF+ functions since several recA and recF mutations abolished concatemer formation. ColE1-type plasmids were previously shown to replicate in rnh mutants in the absence of DNA polymerase I (PolI) activity. This DNA PolI-independent plasmid replication was also examined for its dependence on the recF and recA gene products. rnh- polA(Ts) recF- strains were efficiently transformed with these plasmids at 30 degrees C and 42 degrees C, indicating the presence of DNA PolI-independent replication under recF- conditions. The presence or absence of plasmid replication in rnh- polA- recA(Ts) strains was also examined by measuring the increase in total amounts of plasmid. The results indicated that DNA PolI-independent replication occurred in these triple mutants at 42 degrees C as well as at 30 degrees C. It was concluded that the recombination event giving rise to concatemer formation was not essential for DNA PolI-independent replication in rnh mutants.     

Purification of an SOS repressor from Bacillus subtilis.

We have identified in Bacillus subtilis a DNA-binding protein that is functionally analogous to the Escherichia coli LexA protein. We show that the 23-kDa B. subtilis protein binds specifically to the consensus sequence 5'-GAACN4GTTC-3' located within the putative promoter regions of four distinct B. subtilis DNA damage-inducible genes: dinA, dinB, dinC, and recA. In RecA+ strains, the protein's specific DNA binding activity was abolished following treatment with mitomycin C; the decrease in DNA binding activity after DNA damage had a half-life of about 5 min and was followed by an increase in SOS gene expression. There was no detectable decrease in DNA binding activity in B. subtilis strains deficient in RecA (recA1, recA4) or otherwise deficient in SOS induction (recM13) following mitomycin C treatment. The addition of purified B. subtilis RecA protein, activated by single-stranded DNA and dATP, abolished the specific DNA binding activity in crude extracts of RecA+ strains and strains deficient in SOS induction. We purified the B. subtilis DNA-binding protein more than 4,000-fold, using an affinity resin in which a 199-bp DNA fragment containing the dinC promoter region was coupled to cellulose. We show that B. subtilis RecA inactivates the DNA binding activity of the purified B. subtilis protein in a reaction that requires single-stranded DNA and nucleoside triphosphate. By analogy with E. coli, our results indicate that the DNA-binding protein is the repressor of the B. subtilis SOS DNA repair system.