Mutagenesis during in vitro DNA synthesis.

The error frequency of in vitro DNA synthesis using a natural DNA template has been measured with a biological assay for nucleotide substitutions. phiX174 DNA containing an amber mutation was copied in vitro by Escherichia coli DNA polymerase I, and the reversion frequency of the progeny DNA was determined by transfection of E. coli spheroplasts. E. coli polymerase I makes less than 1 mistake at the am3 locus for every 7700 nucleotides incorporated under standard reaction conditions. Substitution of Mn2+ for Mg2+ and unequal concentrations of deoxynucleoside triphosphate substrates raises this mutation frequency to greater than 1 in 1000. Thus, E. coli DNA polymerase I can copy natural DNA templates with high fidelity and its accuracy can be affected by alterations in reaction conditions.     

Depurination causes mutations in SOS-induced cells.

Introduction of apurinic sites into phi X174 am3 DNA leads to loss of biological activity when measured in a transfection assay. For single-stranded DNA, approximately one apurinic site constitutes a lethal hit; for double-stranded (RFI) DNA, approximately 3.5 hits per strand are lethal. When the reversion frequency of am3 DNA is measured, no increase due to depurination is observed above the background level. However, a large increase in reversion frequency is observed when the same DNA is assayed by using spheroplasts derived from bacteria previously exposed to UV light. The results suggest that apurinic sites are impediments to a replicating DNA polymerase; however, nucleotides can be incorporated opposite these sites under SOS-induced conditions. We estimate the frequency of mutagenesis per apurinic site to be less than 1 in 1400 in normal spheroplasts and 1 in 100 in SOS-induced spheroplasts.     

Bypass and termination at apurinic sites during replication of single-stranded DNA in vitro: a model for apurinic site mutagenesis.

Mutations produced in Escherichia coli by apurinic sites are believed to arise via SOS-assisted translesion replication. Analysis of replication products synthesized on depurinated single-stranded DNA by DNA polymerase III holoenzyme revealed that apurinic sites frequently blocked in vitro replication. Bypass frequency of an apurinic site was estimated to be 10-15%. Direct evidence for replicative bypass was obtained in a complete single-stranded----replicative form replication system containing DNA polymerase III holoenzyme, single-stranded DNA binding protein, DNA polymerase I, and DNa ligase, by demonstrating the sensitivity of fully replicated products to the apurinic endonuclease activity of E. coli exonuclease III. Termination at apurinic sites, like termination at pyrimidine photodimers, involved dissociation of the polymerase from the blocked termini, followed by initiations at available primer templates. When no regular primer templates were available, the polymerase underwent repeated cycles of dissociation and rebinding at the blocked termini and, while bound, carried out multiple polymerization-excision reactions opposite the apurinic sites, leading to turnover of dNTPs into dNMPs. From the in vitro turnover rates, we could predict with striking accuracy the specificity of apurinic site mutagenesis, as determined in vivo in depurinated single-stranded DNA from an M13-lac hybrid phage. This finding is consistent with the view that DNA polymerase III holoenzyme carries out the mutagenic "misinsertion" step during apurinic site mutagenesis in vivo and that the specificity of the process is determined primarily by the polymerase. SOS-induced proteins such as UmuD/C might act as processivity-like factors to stabilize the polymerase-DNA complex, thus increasing the efficiency of the next stage of past-lesion polymerization required to complete the bypass reaction.     

Mutagenic potential of O4-methylthymine in vivo determined by an enzymatic approach to site-specific mutagenesis.

O4-Alkylthymine-DNA adducts have been implicated as causative lesions in chemical mutagenesis and carcinogenesis. To directly assess the mutagenic potential of these adducts in vivo, we have designed an enzymatic technique for introducing nucleotide analogues at predetermined sites of biologically active DNA. Escherichia coli DNA polymerase I was used in vitro to incorporate a single O4-methylthymine residue at the 3' terminus of an oligonucleotide primer opposite the adenine residue of the amber codon in bacteriophage phi X174 am3 DNA. After further extension of the primer with unmodified nucleotides, the partial-duplex product was transfected into E. coli spheroplasts. Replication of the site-specifically methylated DNA in E. coli deficient in O4-methylthymine-DNA methyltransferase (ada-) yielded 10-fold more mutant progeny phage than replication of nonmethylated DNA; no increase in mutation frequency was observed after replication in repair-proficient (ada+) E. coli. The DNA from 20 independently isolated mutant plaques all contained A.T----G.C transitions at the original site of O4-methylthymine incorporation. These data demonstrate that O4-methylthymine induces base-substitution mutations in E. coli and suggest that this adduct may be involved in mutagenesis by N-nitroso methylating agents. This enzymatic technique for site-specific mutagenesis provides an alternative to the chemical synthesis of oligonucleotides containing altered bases.     

Deoxynucleoside [1-thio]triphosphates prevent proofreading during in vitro DNA synthesis.

The contribution of proofreading to the fidelity of catalysis by DNA polymerases has been determined with deoxyribonucleoside [1-thio]triphosphate substrates. These analogues, which contain a sulfur in place of an oxygen on the alpha phosphorus, are incorporated into DNA by DNA polymerases at rates similar to those of the corresponding unmodified deoxynucleoside triphosphates. The fidelity of DNA synthesis was measured with phi X174 am3 DNA; reversion to wild type occurs most frequently by a single base substitution, a C for a T at position 587. By using avian myeloblastosis virus DNA polymerase and DNA polymerase beta (enzymes without a proofreading 3' leads to 5' exonucleolytic activity), substitution of deoxycytidine thiotriphosphate in the reaction mixture did not alter fidelity. In contrast, with DNA polymerases from E. coli (DNA polymerase I) and bacteriophage T4 (enzymes containing a proofreading activity), fidelity was markedly reduced with deoxycytidine [1-thio]triphosphate. DNA containing phosphorothioate nucleotides is insensitive to hydrolysis by the exonuclease associated with these prokaryotic DNA polymerases. These combined results indicate that the deoxynucleoside [1-thio]triphosphates have normal base-pairing properties; however, once misinserted by a polymerase, they are not excised by proofreading. Proofreading of a C:A mismatch at position 587 is thereby found to contribute 20-fold to the fidelity of E. coli DNA polymerase I and a greater amount to the fidelity of bacteriophage T4 DNA polymerase.     

Apurinic/apyrimidinic endonucleases in repair of pyrimidine dimers and other lesions in DNA.

The characteristics of the nicks (single-strand breaks) introduced into damaged DNA by Escherichia coli endonucleases III, IV, and VI and by phage T4 UV endonuclease have been investigated with E. coli DNA polymerase I (DNA nucleotidyltransferase). Nicks introduced into depurinated DNA by endonuclease IV or VI provide good primer termini for the polymerase, whereas nicks introduced into depurinated DNA by endonuclease III or into irradiated DNA by T4 UV endonuclease do not. This result suggests that endonuclease IV nicks depurinated DNA on the 5' side of the apurinic site, as does endonuclease VI, whereas endonuclease III has a different incision mechanism. T4 UV endonuclease also possesses apurinic endonuclease activity that generates nicks in depurinated DNA with low priming activity for the polymerase. The priming activity of DNA nicked with endonuclease III or T4 UV endonuclease can be enhanced by an additional incubation with endonuclease VI and, to a lesser extent, by incubation with endonuclease IV. These results indicate that endonuclease III and T4 UV endonuclease (acting upon depurinated and irradiated DNA, respectively) generate nicks containing apurinic/apyrimidinic sites at their 3' termini and that such sites are not rapidly excised by the 3' leads to 5' activity of DNA polymerase I. However, endonuclease IV or VI apparently can remove such terminal apurinic/apyrimidinic sites as well as cleave on the 5' side of the unnicked sites. These results suggest roles for endonucleases III, IV, and VI in the repair of apurinic/apyrimidinic sites as well as pyrimidine dimer sites in DNA. Our results with T4 UV endonuclease suggest that the incision of irradiated DNA by T4 UV endonuclease involves both cleavage of the glycosylic bond at the 5' half of the pyrimidine dimer and cleavage of the phosphodiester bond originally linking the two nucleotides of the dimer. They also imply that the glycosylic bond is cleaved before the phosphodiester bond.     

Escherichia coli host factor required specifically for the phi X174 stage III reaction: in vitro identification and partial purification.

A cell-free extract prepared from phi X174-infected Escherichia coli cells sustained in vitro synthesis of viral DNA (stage III reaction) when supplemented with fraction II from uninfected cells. The reaction was dependent upon deoxyribonucleoside triphosphate, ATP, added phi X174 replicative form I DNA template, and the fraction II from uninfected cells. This reaction differed from the stage II reaction (semiconservative replication of duplex replicative form DNA) by the production of stable viral protein-DNA complexes sensitive to anti-phi X174 antiserum. Three types of protein-DNA complexes were identified, 50S, 92S, and a 114S complex that cobanded in CsCl and cosedimented in neutral sucrose gradients with a phi X174 phage marker. The sensitivity of these complexes to anti-phi X174 antiserum and Staphylococcus aureus provided a relatively rapid biochemical assay for direct measurement of the amount of DNA synthesized by the stage III reaction. With this assay, an E. coli factor (SIII) required specifically for the synthesis of viral protein-DNA complexes was identified and purified 200-fold from uninfected E. coli cells. The partially purified SIII factor was required for the synthesis of DNA and viral protein-DNA complexes in the phi X174-infected cell extracts and could not be replaced by rep protein, single-strand binding protein, or DNA polymerase III holoenzyme.     

In vitro DNA replication of recombinant plasmid DNAs containing the origin of progeny replicative form DNA synthesis of phage phi X174.

The origin of phage phi X174 progeny replicative form (RF) DNA synthesis has been inserted into the plasmid vector pBR322 and cloned. In direct contrast to pBR322, the recombinant superhelical plasmids can substitute for phi X174 RFI DNA as template in phi X174-specific reactions in vitro. We have shown that the recombinant plasmids: (i) are cleaved by the phi X174 A protein; (ii) support net synthesis of unit-length single-stranded circular DNA in the presence of the phi X174 A protein and Escherichia coli rep protein, DNA-binding protein, and DNA polymerase III elongation system; (iii) support replication of duplexes catalyzed by the phi X174 A protein and extracts of E. coli.     

Stepwise biosynthesis in vitro of globin genes from globin mRNA by DNA polymerase of avian myeloblastosis virus.

Two approaches have been explored for the synthesis of double-stranded DNA from single-stranded DNA template complementary to rabbit 9S globin mRNA (cDNA). (i) cDNA was elongated with dCMP or dTMP homopolymeric tracts using terminal deoxynucleotidyltransferase (EC 2.7.7.31; nucleosidetriphosphate:DNA deoxynucleotidylexotransferase). cDNA-dC, in the presence of an oligo(dG)10 primer, was an efficient template with either DNA polymerase of Escherichia coli (EC 2.7.7.7; deoxynucleosidetriphosphate:DNA deoxynucleotidyltransferase) or RNA-directed DNA polymerase of avian myeloblastosis virus. cDNA-dT [ with an oligo(dA)10 primer] functioned as template only with E. coli polymerase. (ii) cDNA, without homopolymeric tails, was also efficiently copied in the absence of oligonucleotide primer, by DNA polymerase of avian myeloblastosis virus or of E. coli. The product of the reaction consisted of long hairpin molecules which could be converted into DNA duplex (melting temperature, 93 degrees) by digestion with single-strand nuclease S1. The data indicate that a loop structure on the 3' end of cDNA allowed DNA synthesis to take place by a "self-priming" mechanism. Some of the double-stranded DNA synthesized corresponded to the entire sequence of the 9S mRNA template. The synthesis of full-length double-stranded DNA from mouse globin mRNA and immunoglobulin light chain mRNA is also discussed.     

Mechanism of ultraviolet-induced mutagenesis: Extent and fidelity of in vitro DNA synthesis on irradiated templates

The effect of UV irradiation on the extent and fidelity of DNA synthesis in vitro was studied by using homopolymers and primed single-stranded varphiX174 phage DNA as substrates. Unfractionated and fractionated cell-free extracts from Escherichia coli pol(+) and polA1 mutants as well as purified DNA polymerase I were used as sources of enzymatic activity. (DNA polymerases, as used here, refer to deoxynucleosidetriphosphate:DNA deoxynucleotidyltransferase, EC 2.7.7.7.) The extent of inhibition of DNA synthesis on UV-irradiated varphiX174 DNA suggested that pyrimidine dimers act as an absolute block for chain elongation by DNA polymerases I and III. Experiments with an irradiated poly(dC) template failed to detect incorporation of noncomplementary bases due to pyrimidine dimers. A large increase in the turnover of nucleoside triphosphates to free monophosphates during synthesis by DNA polymerase I on irradiated varphiX174 DNA has been observed. We propose that this nucleotide turnover is due to idling by DNA polymerase (i.e., incorporation and subsequent excision of nucleotides opposite UV photolesions, by the 3'-->5' "proofreading" exonuclease) thus preventing replication past pyrimidine dimers and the potentially mutagenic event that should result. In support of this hypothesis, DNA synthesis by DNA polymerase from avian myeloblastosis virus and by mammalian DNA polymerase alpha, both of which are devoid of any exonuclease activity, was found to be only partially inhibited, but not blocked, by UV irradiation of the template and accompanied by an increased incorporation of noncomplementary nucleotides. It is suggested that UV mutagenesis in bacteria requires an induced modification of the cellular DNA replication machinery, possibly an inhibition of the 3'-->5' exonuclease activity associated with DNA polymerases.     

Endonuclease-resistant apyrimidinic sites formed by neocarzinostatin at cytosine residues in DNA: evidence for a possible role in mutagenesis.

When defined-sequence DNA from the lacl region of plasmid pMC1 was treated with the nonprotein chromophore of neocarzinostatin in the presence of various thiols, the predominant lesions were direct strand breaks, occurring primarily at thymine and adenine residues. In the presence of glutathione, however, alkali-dependent strand breaks, occurring at certain cytosine residues, were also detected but were virtually absent when other thiols were used. Chromophore-induced release of free cytosine base from [3H]cytosine-labeled DNA was 2- to 3-fold greater with glutathione than with the other thiols. These results suggest that the alkali-dependent strand break is some form of apyrimidinic site. These sites were substrates for endonuclease IV of Escherichia coli, although a 5-fold greater concentration of enzyme was required for their cleavage than was required for cleavage of apurinic sites in depurinated DNA. These sites were also less sensitive to E. coli endonuclease VI (exonuclease III) by a factor of at least 5 and less sensitive to E. coli endonuclease III by a factor of at least 10. These and other results suggest that these sites are chemically different from normal apurinic/apyrimidine sites. When chromophore-induced apyrimidinic sites were quantitated as alkali-dependent breaks at 11 specific sites in the lacl gene, a correlation was found between occurrences of these lesions and the reported frequencies of G-C to A X T transitions at the same sites. All occurrences of the trinucleotide sequence A-G-C, including the ochre 21 mutational hot spot, were particularly prominent sites. The selective formation of endonuclease-resistant apyrimidinic sites at specific cytosine residues may explain the high frequency of G X C to A X T transitions in the mutational spectrum of neocarzinostatin.     

Mutagenesis by the autoxidation of iron with isolated DNA.

Oxygen free radicals are highly reactive species generated by many cellular oxidation-reduction processes. These radicals damage cellular constituents and have been causally implicated in the pathogenesis of many human diseases. We report here that oxygen free radicals generated by Fe2+ in aqueous solution are mutagenic. Aerobic incubation of luminal diameter X174 am3 (amber 3 mutation) DNA with Fe2+ results in decreased phage survival when the treated DNA is transfected into Escherichia coli spheroplasts. Transfection of the treated DNA into SOS-induced spheroplasts results in an increase in mutagenesis as great as 50-fold. Both killing and mutagenesis can be prevented by binding of Fe2+ with deferoxamine or by the addition of catalase or mannitol. These results suggest that DNA damage and mutagenesis brought about by Fe2+ are likely to occur by a Fenton-type mechanism that involves the generation of (i) hydrogen peroxide by the autoxidation of iron and (ii) hydroxyl radicals by the interaction of the hydrogen peroxide with Fe2+. DNA sequence analysis of the Fe2+-induced mutants indicates that reversion of the phage phenotype to wild type occurs largely by a transversion type of mutation involving substitution of deoxyadenosine for thymidine opposite a template deoxyadenosine. Mutagenesis is not abolished by incubation of Fe2+-treated luminal diameter X174 am3 DNA with an apurinic endonuclease and only partially abolished by incubation with alkali, suggesting that a large fraction of the mutagenesis by oxygen free radicals is not caused by formation of apurinic sites but instead involves an as-yet-to-be-defined alteration in deoxyadenosine. These findings raise the possibility that free iron localized in cellular DNA may cause mutations by the generation of oxygen free radicals.     

Nucleoside Diphosphokinase Activity Associated with DNA Polymerases

Nucleoside diphosphokinase activity is present in highly purified preparations of DNA polymerase from Micrococcus luteus and Escherichia coli, and in a partially purified DNA polymerase from avian myeloblastosis virus. The activity is also observed in the protein fragment of molecular weight 76,000 that is produced by subtilisin cleavage of DNA polymerase I from E. coli. The NDP kinase activity in DNA polymerase preparations from M. luteus uses various ribo- and deoxyribonucleoside di- and triphosphates as substrates. The presence of this activity in preparations of DNA polymerase results in the apparent use of deoxyribonucleoside diphosphates as substrates for DNA synthesis, provided that some triphosphate is present to serve as a phosphate donor.     

The in vivo mutagenic frequency and specificity of O6-methylguanine in phi X174 replicative form DNA.

A bacteriophage phi X174-based site-specific mutagenesis system for the study of the in vivo mutagenic frequency and specificity of carcinogen-induced modification in DNA is presented. A (-)-strand primer containing O6-methylguanine in a specific site was hybridized to a single-stranded region in gene G of phi X gapped duplex DNA. The hybrid was enzymatically converted to replicative form DNA and was used to transform Escherichia coli cells. All gene G mutants generated by the modification were rescued by genetic complementation. An amber mutation in lysis gene E of the (+) strand of the replicative form DNA prevented lytic growth of wild-type phage derived from this strand. In each mutant-containing infective center produced from the transformed cells, gene G mutant phage were present in a 3:1 ratio compared to wild type. Thus, in vivo, O6-methylguanine in replicating phi X DNA has a mutagenic frequency of 75%. When repair of O6 methylguanine occurred, it was prereplicative. The mutations were due exclusively to the misincorporation of thymine.     

In vitro synthesis of bacteriophage phi X174 by purified components.

An in vitro system capable of synthesizing infectious phi X174 phage particles was reconstituted from purified components. The synthesis required phi X174 supercoiled replicative form DNA, phi X174-encoded proteins A, C, J, and prohead, Escherichia coli DNA polymerase III holoenzyme, rep protein, and deoxyuridinetriphosphatase (dUTPase, dUTP nucleotidohydrolase, EC 3.6.1.23) as well as MgCl2, four deoxyribonucleoside triphosphates, and ATP. Phage production was coupled to the synthesis of viral single-stranded DNA. More than 70% of the synthesized particles sedimented at the position of mature phage in a sucrose gradient and associated with the infectivity. The simple requirement of the host proteins suggests that the mechanism of viral strand synthesis in the phage-synthesizing reaction resembles that of viral strand synthesis during the replication of replicative form DNA.     

Rearrangements of DNA mediated by terminal transferase.

To assess the involvement of terminal transferase in generating immunoglobulin diversity, the mutagenic potential of this enzyme has been measured. The frequency of single base substitutions during copying of phi X174 DNA by DNA polymerase beta is increased by, at most, 3-fold upon the addition of terminal transferase. However, terminal transferase is highly mutagenic, either alone or with DNA polymerase beta, in a forward mutation assay using M13mp2 DNA. The frequency of complex mutants, as determined by DNA sequence, is increased by greater than 100-fold. These mutants involve the deletion of a variable number of bases initially present in the template sequence and the addition of a sequence of nucleotides rich in guanine residues. Analysis of these mutants suggests an antibody diversity model implicating terminal transferase in the imprecise linkage of variable, joining, and diversity segments during the formation of functional immunoglobulin genes.     

Conversion of ϕX174 Viral DNA to Double-Stranded Form by Purified Escherichia coli Proteins

The E. coli proteins that catalyze the conversion of varphiX174 single-stranded DNA to duplex DNA have now been purified extensively. The reaction depends on dnaB, dnaC(D), dnaE, and dnaG gene products, DNA elongation factors I and II, E. coli DNA binding protein, and two additional E. coli proteins, replication factors X and Y. DNA synthesis by these proteins requires varphiX174 viral DNA, dNTPs, Mg(+2), and ATP. The product synthesized is full-length linear varphiX174 DNA. The reaction has been resolved into two steps. The first step involves the interaction of ATP and varphiX174 DNA with dnaB and dnaC(D) gene products, E. coli DNA binding protein, and replication factors X and Y in the absence of dNTPs. Subsequent dNMP incorporation requires the addition of DNA polymerase III, DNA elongation factors I and II, dnaG gene product, and dNTPs.     

The dnaN gene codes for the beta subunit of DNA polymerase III holoenzyme of escherichia coli.

An Escherichia coli mutant, dnaN59, stops DNA synthesis promptly upon a shift to a high temperature; the wild-type dnaN gene carried in a transducing phage encodes a polypeptide of about 41,000 daltons [Sakakibara, Y. & Mizukami, T. (1980) Mol. Gen. Genet. 178, 541-553; Yuasa, S. & Sakakibara, Y. (1980) Mol. Gen. Genet. 180, 267-273]. We now find that the product of dnaN gene is the beta subunit of DNA polymerase III holoenzyme, the principal DNA synthetic multipolypeptide complex in E. coli. The conclusion is based on the following observations: (i) Extracts from dnaN59 cells were defective in phage phi X174 and G4 DNA synthesis after the mutant cells had been exposed to the increased temperature. (ii) The enzymatic defect was overcome by addition of purified beta subunit but not by other subunits of DNA polymerase III holoenzyme or by other replication proteins required for phi X174 DNA synthesis. (iii) Partially purified beta subunit from the dnaN mutant, unlike that from the wild type, was inactive in reconstituting the holoenzyme when mixed with the other purified subunits. (iv) Increased dosage of the dnaN gene provided by a plasmid carrying the gene raised cellular levels of the beta subunit 5- to 6-fold.