- Email: [email protected]
Involvement of DNA polymerase β in DNA replication and mutagenic consequences1
doi:10.1006/jmbi.2001.5307 available online at http://www.idealibrary.com on
J. Mol. Biol. (2002) 315, 1039±1047
Involvement of DNA Polymerase b in DNA Replication and Mutagenic Consequences Laurence Servant1, Anne Bieth1, Hiroshi Hayakawa2, Christophe Cazaux1* and Jean-SeÂbastien Hoffmann1* 1
The Institut de Pharmacologie et Biologie Structurale, UMR CNRS 5089, 31077 Toulouse ceÂdex 4, France 2
Department of Medical Biochemistry, Graduate School of Medical Sciences, Kyushu University, Fukuoka 812-8582, Japan
Overexpression in mammalian cells of the error-prone DNA polymerase b (Pol b) has been found to increase the spontaneous mutagenesis. Here, we investigated a possible mechanism used by Pol b to be a genetic instability enhancer: its interference in replicative DNA synthesis, which is normally catalysed by the DNA polymerases a, d and e. By taking advantage of the ability to incorporate ef®ciently into DNA the chain terminator ddCTP as well as the oxidised nucleotide 8-oxo-dGTP, we show here that puri®ed Pol b can compete with the replicative DNA polymerases during replication in vitro of duplex DNA when added to human cell extracts. We found that involvement of Pol b lowers replication ®delity and results in a modi®ed error-speci®city. Furthermore, we demonstrated that involvement of Pol b occurred during synthesis of the lagging strand. These in vitro data provide one possible explanation of how overexpression of the enzyme could perturb the genetic instability in mammalian cells. We discuss these ®ndings within the scope of the up-regulation of Pol b in many cancer cells. # 2002 Elsevier Science Ltd.
*Corresponding authors
Keywords: DNA replication; mutagenesis; DNA polymerases; cancer
Introduction Mammalian cells contain many DNA polymerases involved in different DNA transactions. While DNA polymerases (Pol) a, d, and e are devoted to the genomic DNA replication process, other DNA polymerases are con®ned to additional speci®c roles. For example, in mammalian somatic cells, Pol b, a monomeric 39 kDa enzyme, is involved in the the single-nucleotide base excision repair (BER) pathway,1 which processes small DNA lesions such as oxidized or alkylated bases. This DNA polymerase can be distinguished from the DNA polymerases involved in the elongation step of the genome replication (Pol d and Pol e) by its high level of in®delity in replicating DNA in vitro, due to the lack of associated proofreading activity and a poor ability to discriminate nucleotides. Probably because of these error-prone features, its expression is tightly regulated in vivo, being constant and low throughout the cell-cycle.2 Limitation of the inaccuracy of Pol b in the BER Abbreviations used: Pol, DNA polymerase a; BER, base excision repair. E-mail addresses of the corresponding authors: [email protected]; [email protected] 0022-2836/02/051039±9 $35.00/0
pathway has been suggested as a mechanism to avoid the high mutation rate subsequent to its repair action: DNA ligase III, which catalyses the joining step, can discriminate strongly between a correctly paired versus a mispaired residue at the 30 position of a nick in DNA, and thus delay the joining reaction, allowing the removal of the mismatched terminal residue by a distinct 30 exonuclease.3 By using conventional methodologies testing the appearance of mutational events at the Na-KATPase locus as well as at the locus encoding the purine salvage enzyme hypoxanthine guanine phosphoribosyl transferase (HPRT), leading to ouabain-resistance and 6-thioguanine-resistance phenotypes respectively, we demonstrated previously that excess Pol b in mammalian cells resulted in a fourfold to 12-fold increased rate of spontaneous mutagenesis.4 Moreover, we demonstrated recently that ectopic expression of Pol b in mammalian cells induces chromosomal instability and potentiates tumorigenesis in nude immunode®cient mice (unpublished results). Thus, it appears that alteration of Pol b expression may induce major genetic changes that contribute to the malignant phenotype. # 2002 Elsevier Science Ltd.
1040 Based on these data, we hypothesized that an excess of the error-prone Pol b can perturb the well-de®ned speci®c functions of error-free DNA polymerases during DNA transactions such as repair, replication or recombination pathways.5 Here, we have investigated whether such a mutagenic interference could be observed during in vitro DNA replication of duplex DNA, which in a normal somatic cell, requires, the coordinated activity of the replicative error-free Poly a, d and e, but not b.6 To conduct the experiments, we took advantage of the fact that dideoxycytidine (ddC), which terminates chain elongation when incorporated into DNA,7 as well as 8-oxo-dGTP, an oxidized form of dGTP, are substrates of Pol b but are incorporated poorly by the replicative polymerases.8 By using the archetypal simian virus 40 (SV40) origin-containing plasmid replication assay, which reconstitutes the mammalian DNA replication fork,9 we present evidence that addition of Pol b to cell extracts renders the process hypersensitive to ddCTP and enhances misincorporation of the 8oxo-dGTP analogue, demonstrating that the enzyme can be associated with the replicative event in the context of its overexpression. Furthermore, we showed that this involvement occurred during the synthesis of the lagging strand and resulted in an increased frequency of mutation. The modi®ed error-speci®city spectrum induced by Pol b that we observed is in accordance with the mutagenic features of the enzyme. The consequences of the interference in DNA replication by error-prone DNA polymerases such as Pol b is discussed in relation to the design of the putative role in cell transformation of the misregulation of newly discovered error-prone DNA polymerases.
Results Incorporation of ddCTP into DNA by Pol b during SV40 DNA replication T-antigen-dependent replication of a DNA plasmid using cell-free extracts from human Hela cells was performed at 37 C with increasing amount of puri®ed Pol b in the absence or presence of the nucleotide analogue ddCTP, at a mass ratio of ddCTP to dCTP equal to 0.5. DNA was then puri®ed, linearized by BamHI, and subjected to electrophoresis in a 1 % (w/v) agarose gel, followed by ethidium bromide staining and autoradiography (Figure 1). We veri®ed that semi-conservative replication was occurring, since the products were resistant to DpnI, which cuts only fully methylated 50 -GATC sequences. We veri®ed that without T-antigen, no incorporation was detectable in the presence of various amounts of Pol b during the reaction, meaning that no background repair synthesis occurred under our conditions. We found that, without ddCTP, the presence of excess Pol b itself did not affect the ef®ciency of the replication reaction quantitatively (Figure 1(a)). When we added ddCTP, known to inhibit strongly the
Role of Excess Pol in DNA Replication
polymerisation of DNA in vitro catalysed by Pol b but not by the replicases Pol d and Pol e, the presence of Pol b induced a hypersensitivity of DNA synthesis to this nucleotide analogue (Figure 1(a), upper part). We measured a 66 % inhibition with 0.012 unit of Pol b, while we were not able to detect a radioactive signal in the presence of 0.024 unit, suggesting that, for this concentration, Pol b interfered with the replication forks of most plasmids that undergo DNA replication (see quantitative data in Figure 1(b)). An internal control DNA incorporated into the reaction mixture after the six hour replication incubation was used to verify the ef®ciency and homogeneity of DNA puri®cation after the reaction (see Figure 1(a), lower part). We did not observe any effect on SV40 replication by the same concentrations of ddCTP when heatdenatured Pol b was added, demonstrating that the polymerase activity is necessary for the inhibition (not shown). It is important to note that in all these experiments the enzyme concentrations had a physiological relevance and did not represent a great excess of enzyme that could introduce artefacts. Firstly, by using western blotting experiments (not shown), we observed that the protein amounts were less than physiological ranges measured in several human tumoral cells.10 Secondly, we conducted calibration assays in order to determine the speci®c activity of Pol b relative to replicative cell extracts, by measuring nucleotide incorporation on activated calf thymus DNA preincubated with DNase I (Figure 2). These assays showed that concentration ranges of puri®ed Pol b added to the cell extracts represented 3 to 20 % compared to the global activity of the cell extracts. In the SV40 replication assay, where the enzyme was mixed with the extracts, this relative range of activity should be lowered considerably, since Pol b most probably did not interfere at the replicative forks with an ef®ciency comparable to that of the other replicative DNA polymerases. This was con®rmed by using primer extension reactions on oligonucleotide in the presence of the chainterminator ddCTP (not shown). Incorporation of 8-oxo-dGTP into DNA by Pol b during SV40 DNA replication The mutagenic analogue 8-oxo-dGTP, resulting from the conversion of dGTP by reactive oxygen species, has been shown to be incorporated ef®ciently into DNA in vitro by Pol b. We took advantage of this feature to further demonstrate the speci®c involvement of Pol b during the elongation of the replicative forks using the SV40 replication assay. We conducted DNA replication reactions comparable with the ddCTP experiments described above, but in the presence of 32P-labelled 8-oxodGTP that we prepared (see Materials and Methods). Figure 3 shows that in the absence of cold dGTP, incorporation of the labelled oxidized nucleotide is enhanced by threefold when 0.012 unit of Pol b is added to the replicative extracts.
Role of Excess Pol in DNA Replication
1041
Figure 1. (a) T-antigen-dependent replication of 100 ng of pCineo using 400 mg of cell-free extracts from human Hela cells. DNA replication was performed using the conventional SV40 replication assay at 37 C for six hours with increasing amounts of puri®ed Pol b (0.0024u, 0.0048u, 0.012 and 0.024 unit), in the presence or absence of the nucleotide analogue ddCTP at a ratio of dCTP to ddCTP equal to 0.5. DNA was puri®ed in the presence of an internal pCDNA2 control DNA, linearized by BamHI, and subjected to electrophoresis in a 1 % agarose gel, followed by ethidium bromide staining (lower part) and autoradiography (upper part). The arrow indicates the position of the linearized product. (b) Quanti®cation analysis of the resolved radioactive bands on the gel seen in (a) was achieved by the PhosphoImager Storm-system analysis using Imagequant software.
Such an increase in labelled 8-oxo-dGTP incorporation is strengthened in the presence of 5 mM cold dGTP, con®rming the poor discrimination of Pol b between the correct dGTP and its 8-oxo-dGTP competitor antagonist, in contrast to the replicative polymerases. We cannot rule out the possibility that hydolysis of the incorporated 8-oxo-dGTP by signi®cant levels of the 8-oxoG DNA glycosylases (hOGG1, hOGG2, and hNTH1) in the cell extracts occurred during the six hour incubation, but this activity would have the same ef®ciency for both conditions. Together with the ddCTP incorporation data, this demonstrates that physiological level of Pol b found in several human cancer cells can interfere with SV40 replication. Fidelity of SV40 DNA replication in the presence of Pol b In order to appreciate the mutagenic impact of the Pol b involvement in DNA replication, the ®delity of SV40 DNA replication was investigated by using the pBK-CMV and M13mp2SV ori Left and ori Right vectors containing the target reporter lacZa sequence as a mutagenic target. The replication-induced mutations occurring in the target gene, which resulted in the expression of a non-
functional-b-gal enzyme, were analysed (see Tables 1-3). We determined the mutant frequencies for the replicated DNA samples after their electroporation into a mutS Escherichia coli strain to ®x the errors, and their subsequent insertion into an indicator E. coli strain to score colourless or light blue mutants among total clones after plating the cells in the presence of 5-bromo-4-chloro-3-indolyl-b,Dgalactopyranoside (Xgal) and IPTG. The mutant frequencies found after SV40 DNA replication of the pBK-CMV and M13mp2SV vectors were of the order of magnitude of those measured previously.11,12 In the presence of 0.012 unit of Pol b, the number of mutants was increased signi®cantly and reproductively by twofold to threefold (Table 1) in three independent reactions, depending on which substrate is considered. This result is an additional argument demonstrating that Pol b interferes with DNA replication, since it affects the accuracy of this process. Error speci®city was analysed by sequencing 50 replicated pBK-CMV mutants from the control replication and the replication in the presence of Pol b (Figure 4). In contrast to the spectrum observed for the control conditions, where some mutational hot-spots were detected, notably at positions 1159 and 1173-1182 (a G C-rich sequence propitious for slippage
1042
Role of Excess Pol in DNA Replication
described previously for gapped DNA by using the lacZ gene target.13 Strand-specificity of Pol b - induced mutations during SV40 replication
Figure 2. Calibration of DNA synthesis activity. Relative activities of 0.010 unit of added Pol b (open circles) compared to 45 mg of cell extracts (®lled squares) quanti®ed on activated calf thymus DNA by measuring the incorporation of [a-32-P]dCTP into acid-insoluble materials at 37 C for the indicated incubation times. One unit is the amount of Pol b required to catalyse the incorporation of 1 pmol of [a-32P]dGTP in one hour at 37 C.
during DNA polymerisation), the errors occurring when Pol b was present in the cell extract were distributed essentially all along the sequence of the a-peptide, while one hot-spot remained targeted. This suggests some Pol b intervention during the elongation step of the replication reaction. Base substitutions were detected at a frequency of 61.5 % and 45.5 % for control and excess Pol b reactions, respectively, while frameshifts represented the majority of the changes for the excess Pol b reaction (54.5 %) against 38.5 % for the control replication. This change in error speci®city favouring the frameshift mutations is in accordance with the mutational speci®city of the enzyme in vitro
In order to de®ne more precisely on which strand excess Pol b could interfere during SV40 replication, we used the M13mp2 ori Left and M13mp2 ori Right vectors that allow strandspeci®c errors analysis, since these templates contain the SV40 origin located on opposite sides relative to the LacZa mutational target and close to it. The 8-oxo-dGTP nucleotide analogue is known to be misincorporated during SV40 replication opposite adenine,14 leading to A T ! C G transversions, produced rarely in the presence of normal dNTPs.15 Therefore, when 8-oxo-dGTP is added to SV40 reactions, its miscoding de®nes the strand into which the 8-oxo-dGTP is incorporated. By taking advantage of the fact that Pol b is capable of ef®cient incorporation of 8-oxo-dGTP with a preferential mispairing with a template dA compared to a template dC (24/1),16 we got information about strand speci®city of Pol b-dependent synthesis during M13mp2-ori-Left/Right replication by analysing the sequence of mutant phages. According to previous experiments,11 we measured a two- to threefold increase in mutation frequency by the sole addition of 8-oxo-dGTP during the replication of both vectors (Table 2). Addition of puri®ed Pol b to the reactions resulted in an additional 2.4-fold enhancement of the mutant frequency, con®rming the capacity of the enzyme for ef®cient 8-oxo-dGTP incorporation. For determination of strand speci®city, we performed DNA sequence spectra of collections of independent ori Left mutants after replication in the presence of 8-oxo-dGTP, analysing the LacZacomplementation target sequence from positions ÿ45 to 197. In all, 23 mutants were sequenced from control reactions and 33 mutants from reaction in the presence of excess Pol b. A T ! C G transversions were the primary mutations recovered, consistent with the expected misincorpora-
Table 1. Fidelity of SV40 DNA replication in the absence or in the presence of excess Pol b Number of plaques scored DNA vector pBK-CMV M13mp2SV ori Left M13mp2SV ori Right
Conditions Unreplicated Replicated without Pol b Replicated with Pol b Replicated without Pol b Replicated with Pol b Replicated without Pol b Replicated with Pol b
Total
Mutants
Mutant frequencies (10ÿ4)
28,640 141,390 209,762 43,040 24,018 115,611 60,956
184 918 2953 18 31 63 95
64 65 140 4.2 12.9 5.4 15.6
Error ratea (10ÿ5) <0.05 4.0 <0.03 0.39 <0.08 0.5
SV40 reactions were performed as described in Materials and Methods from three independent reactions. Error rate was calculated by subtracting the background mutant frequency from the average mutant frequency. This value was divided by 0.5 and then by the number of detectable mispairs sites for each vector: 482 for M13mp2SV9 and 362 for pBB-CMV.10 In the case of M13 vectors, the background mutant frequency found was 3.5 10ÿ4. a
1043
Role of Excess Pol in DNA Replication
Figure 3. Incorporation of 8-oxodGTP into DNA by excess Pol b during in vitro SV40 replication reaction. The replication conditions are identical with those used for the ddCTP experiment in Figure 2 except that 40 mM [a-32P]8-oxodGTP were added in the presence (right part) or the absence (left part) of 5 mM cold dGTP.
tion of 8-oxo-dGTP opposite the template adenine residue. We also detected in three mutants (one from the control reaction and two from the reaction Pol b) C G ! A T transversions, which could represent misincorporation of dAMP opposite the template 8-oxo-dGMP in a second round of DNA replication. We found that, in the presence of Pol b, most of the mutations arose from 8-oxo-dGTP misincorporation opposite adenine residues on the lagging strand (Table 3), whereas these modi®cations were distributed quite equivalently on both strands for the control reaction. We calculated a fourfold increase in mispair rate on the lagging strand and we observed no variation on the leading strand. Taken together, these ®ndings suggest strongly that the lagging strand is the preferential target of Pol b when interfering with the DNA replication fork elongation.
Discussion We show here that Pol b, an enzyme required in somatic cells in the BER pathway,1 can interfere with duplex DNA replication in vitro when up-represented and renders the process inaccurate. Such a result was obtained by using the well-calibrated SV40 replication assay, which reconstitutes the mammalian DNA replication fork by providing all the factors required for bidirectional replication of double-stranded DNA in a human cell extract except for the SV40 T-antigen. Cell extracts include DNA polymerase a and its associated primase for the initiating step and for starting the synthesis of Okazaki fragments, DNA polymerase d assisted by replication factor C (RF-C) and proliferating cell nuclear antigen (PCNA), and eventually DNA
polymerase e, for the synthesis of the leading and lagging strands. By adding to the reaction nucleotide analogues such as ddC and 8-oxo-dG, incorporated into DNA preferentially by Pol b, we present evidence here for some intervention of Pol b during the elongation process. We further employed a strategy based on the replication of vector templates bearing both orientations of the SV40 origin in order to examine whether Pol b competes for primer termini on the leading strand, the lagging strand, or both. We were able to demonstrate that the lagging strand was the preferred target of Pol b during the fork elongation process. This is in accordance with the properties of the enzyme, whose preferential substrate has been demonstrated to be gapped DNA,17 since the synthesis of the lagging strand, which requires the formation and the joining of Okazaki fragments, produces gapped DNA that needs to be ®lled. This result is supported by a previous observation showing that ectopic expression of rat Pol b in E. coli can restore growth in a DNA polymerase Idefective bacterial mutant by increasing the rate of joining of Okazaki fragments.18 Moreover, we showed recently that gap-®lling during the nucleotide excision repair (NER), a pathway processed normally by DNA polymerases d and e, can be achieved by Pol b when over-regulated in cells.19 Although the in vitro SV40 DNA replication assay did not re¯ect completely the in vivo situation, the error-prone involvement of excess Pol b in DNA replication has biological importance. First, it may explain why an increase in spontaneous mutation frequency occurred in mammalian cells overexpressing the enzyme.4 It may represent a mechanistic basis for our recent data
Table 2. Effect of excess Pol b on 8-oxo-dGTP-induced mutation frequency during SV40 replication Conditions Replicated without Pol b Mispairs on leading strand Mispairs on lagging strand Replicated with Pol b Mispairs on leading strand Mispairs on lagging strand
Transversion number
Mutant ratio (%)
Error ratea (10ÿ6)
13 10
56 43
41.5 35.8
10 23
30 70
58.1 149.6
a Error rate was calculated as described in the legend to Table 1, except that the number of detectable AT ! CG transversion sites for each M13 vector was 53 for ori Left and 59 for ori Right.11
1044
Role of Excess Pol in DNA Replication
Figure 4. Spectrum of mutations on the target reporter lacZa sequence during SV40 replication in the absence (above) or in the presence (below) of 0.012 unit of puri®ed Pol b: 50 colourless mutant colonies from each replication reaction were sequenced. The ÿ1 one base deletions are shown as open triangles, insertions are indicated as ®lled triangles together with the inserted base as a bold character, and single-base substitutions are indicated by the nucleotide concerned.
showing that excess Pol b provided these cells with a new phenotype that creates competent tumour cells when injected into immunode®cient mice ( . Bergoglio et al., unpublished results). High levels of Pol b have been found in various tumours such as glioma and lymphoma, breast, colon and prostate,10 as well as in ovarian,19 leukemia and lymphoma tumour cell lines (unpublished results). It is possible that, like the de®ciency in an error-correcting mechanism, error-prone DNA synthesis during replication may contribute to the global genetic instabilities and the great variability displayed by these cancer cells. Such a ``mis-invol-
vement'' of Polb could take place through the context of its overexpression in other error-free DNA transactions pathways involving a long patch DNA synthesis such as NER,19 and DNA homologous recombination or mismatch repair. New cellular error-prone DNA polymerases have been described recently.20,21 These enzymes, which belong to the UmuC/DinB/Rev1/Rad30 superfamily of DNA polymerases that are able to bypass lesions in DNA,22 are part of stressinducible processes that allow them to function only when high mutation rates are advantageous. Among these DNA mutases23 that have been
Table 3. Effect of excess Pol b on the speci®city of leading or lagging strand errors for dA 8-oxo-dGMP mispairs after M13mp2SV-ori-Left replication Number of plaques scored
DNA vector M13mp2SV ori Left 8-oxo-dGTP M13mp2SV ori Right 8-oxo-dGTP
Conditions Replicated Replicated Replicated Replicated
without Pol b with Pol b without Pol b with Pol b
Total
Mutants
Mutant frequencies (10ÿ4)
30,967 49,151 9598 4476
32 120 11 12
10.3 24.4 11.4 26.8
Error ratea (10ÿ5) 2.6 7.9 2.7 7.9
a Error rate was calculated as described in the legend to Table 1, except that the number of detectable mutation sites for each strand was 28 for the leading strand and 25 for the lagging strand.11 The error rate obtained was then multiplyed by the corresponding mutant ratio.
1045
Role of Excess Pol in DNA Replication
isolated in human cells, Pol k,24 Pol Z21 and Pol i25 produce errors at even higher rates than Pol b in copying undamaged DNA in vitro. A recent report demonstrated that Pol Z can also compete with the replicative DNA polymerases during replication of duplex DNA, and lowers replication ®delity by a factor of 5 to 8.26 Interestingly, those authors showed that the proofreading exonuclease activities reduce the mutagenic potential of Pol Z during the replicative process. It is possible that such an editing process of the Pol b-induced errors described in our work may be performed by an exonuclease activity associated with the replication machinery. The low processivity of Pol b, in addition to its ability to extend mispairs less ef®ciently than matched termini, giving more chance for extrinsic proofreading, may explain why excess Pol b lowered replication ®delity by only a factor of 2 to 3. In addition, misincorporation of 8-oxodGTP mediated by excess Pol b during SV40 replication may be underestimated, since dA 8-oxodGMP mispairs have been demonstrated to be proofread exonucleolytically by cell extracts during synthesis of both strands. Nevertherless, such mutagenicity rates in vivo have been shown to have dramatic effects on tumour development.27 To our knowledge, no direct evidence has been described for a possible relationship between upregulation of error-prone DNA polymerases and the genetic variability of cancer cells, although the impact of such enzymes could be of great importance within the scope of tumour development. Our data presented here with Pol b, which need to be enlarged to other members of this enzymatic family, open the possibility to explore the putative role of mutases in carcinogenesis by interfering with the DNA replication process.
In vitro SV40 DNA replication reactions
Materials and Methods
32 P-radiolabelled 8-oxodGTP was prepared as described29 with some modi®cations. Brie¯y, 6 mM dGTP was incubated at 37 C for two hours in the dark in a reaction mixture containing 100 mM sodium phosphate (pH 6.8), 30 mM ascorbic acid, 100 mM H2O2. A portion (50 ml) of the reaction mixture was loaded onto a MonoQ column and eluted with a linear gradient (0.01 M-1 M) of triethylammonium hydrogen carbonate (pH 7.0) at a ¯ow-rate of 1 ml/minute. 8-Oxo-dGTP was monitored with a UV detector (254 nm) and eluted with a retention time about one minute later than that for dGTP. Fractions containing 8-oxo-dGTP were combined, lyophilized, dissolved in 12.5 mM citric acid, 25 mM sodium acetate, 30 mM NaOH, 10 mM acetic acid (pH 5.4) and applied to a C18 reverse phase column TOSOH TSK-GEL ODS-80T equibrated with the same buffer at a ¯ow-rate of 1 ml/minute. 8-Oxo-dGTP was eluted with a retention about ®ve minutes later than that for dGTP. The sample was further puri®ed with the MonoQ column as described above. The 8-oxo-dGTP was lyophilized, dissolved in 20 mM sodium phosphate (pH 6.8) and stored at ÿ20 C. Purity and quantity of the preparation were examined with HPLC and UV spectrum analyses. No dGTP was detectable in the ®nal preparation of 8-oxo-dGTP. SV40 replication conditions were the same as described above, except that 44 mM 32 P-radiolabelled 8-oxodGTP was added to the reaction
Calibration of DNA polymerization Calf thymus DNA pre-incubated with DNase I (0.4 mg) was used as template in assay reactions (20 ml) at 37 C for various incubation times with different concentrations of either puri®ed Pol b or replicative Hela cell extracts in 50 mM Tris-HCl (pH 7.5), 6 mM MgCl2, 0.5 mM dithiothreitol, 300 mM each dATP, dGTP, dTTP, and 0.2 mCi of [32P]dCTP. At the end of the reaction, 10 mM EDTA was added and aliquots were spotted on Whatman glass ®lters (GF/C). Filters were washed with a 5 % (w/v) trichloroacetic acid (TCA), 1 % (w/v) sodium pyrophosphate solution and radioactivity was determined. Preparation of cell-free extracts for SV40 replication HeLa cells were grown as spinner cultures in RPMI 1640 (Biomedia) with 10 % (v/v) fetal calf serum, 4 mM glutamine, and antibiotics (50 units/ml of penicillin and 50 mg/ml of streptomycin) at 37 C in a humidi®ed 5 % (v/v) CO2 atmosphere. The cells were harvested by centrifugation in the upper part of their log phase growth and extracts were prepared as described.9
Replication reaction mixtures (25 ml) contained 30 mM Hepes (pH 7.8), 7 mM MgCl2, 200 mM each CTP, GTP, UTP, 4 mM ATP, 100 mM each dATP, dCTP, dTTP, 10 mM dGTP, and [a-32P]dGTP (4000 cpm/pmol; Amersham), 40 mM creatine phosphate (Sigma), 100 mg/ml of creatine phosphokinase (Sigma), 50-100 ng of DNA substrate, 0.5 mg of SV40 large T-antigen (Molecular Biology Resources), 400 mg of Hela cell extract, and various concentrations of rat DNA Pol b, puri®ed as described.28 One unit of Pol b corresponds to 1 nmol of dNTP incorporated into acid-insoluble materials at 37 C in 60 minutes, using activated calf thymus DNA pre-incubated with DNase I as substrate. Reactions without T-antigen were used as negative controls. SV40 replication reactions in presence of ddCTP After incubation at 37 C for the periods indicated in Figure 2 with or without 50 mM ddCTP (ddCTP to dCTP ratio of 0.5), reactions were quenched by adding an equal volume of stop solution (2 % (w/v) SDS, 2 mg/ml of proteinase K, 50 mM EDTA) and further incubation for one hour at 55 C. DNA (0.5 mg of pc-DNA II, In vitroGen) was added to each sample as internal puri®cation controls. Reaction products were puri®ed by extraction with phenol/chloroform/(25:24:1 by vol.) isoamyl alcohol followed by precipitation in ethanol. The DNA was resuspended in distilled water. The samples were then treated with BamHI and DpnI (New England Biolabs), and the restriction digests were separated on a 1 % (w/v) agarose gel. After staining the gel with ethidium bromide, internal control DNAs were quanti®ed. The gel was then dried, and autoradiography performed. Quanti®cation of the resolved radioactive bands on the gel was achieved by PhosphoImager Storm-system analysis using Imagequant software. SV40 replication reactions in presence of radiolabelled 8-Oxo-dGTP
1046 mixture in the presence or the absence of cold dGTP as noted. Mutagenesis assays As SV40 replication templates, we used the pBK-CMV vector (Stratagene), which contains the full SV40 origin of replication and the lacZa gene (at positions 810-1183), or the M13mp2SVori Right/Left vectors kindly provided by Dr Thomas Kunkel (NIEHS, NC), which contain the SV40 origin of replication located either side of the wildtype LacZa sequence.30 DNA substrate (40 to 50 ng) was replicated during six hours as described above, except for the 8-oxo-dGTP mutation-induced replication assay where 100 mM 8-oxo-dGTP and 10 mM dGTP were added to the reactions. After replication, the products were precipitated and subjected to DpnI digestion to eliminate unreplicated DNA templates. Products of a single round of replication carrying mutations that exist as heteroduplexes, pBKCMV replication products were electroporated into a MC1061mutS E. coli host (kindly provided by Professor Masson, Toulouse) and M13mp2SV ori Left/Right products were electroporated into NR9162 E. coli host (kindly provided by T. Kunkel, NIEHS, NC). Both bacterial strains being de®cient for mismatch repair, mutations generated during SV40 replication were ®xed. In the case of the pBK-CMV vector, DNA was extracted, electroporated in JM109 E. coli bacteria and then plated onto indicator plates containing Xgal and IPTG. In the case of M13 DNA, the NR9162-infected cells were plated immediately after electroporation in soft agar containing XGal, IPTG and a log-phase culture of CSH50. The inactivation of the a-complementation gene due to a mutation in the pBK-CMV/M13mp2 LacZa gene resulted in a colorless or light-blue phenotype of the colonies or of the plaques. To determine the speci®city of nucleotide errors, automated sequencing of the mutant colonies has been performed (MilleGen Co., Toulouse).
Acknowledgments This work was supported by La ``Ligue contre le cancer'' (Equipe LabelliseÂe).
References 1. Sobol, R., Horton, J., KuÈhn, R., Gu, H., Singhal, R., Prasad, R. et al. (1996). Requirement of mammalian DNA polymerase-beta in base-excision repair. Nature, 379, 183-186. 2. Zmudzka, B., Fornace, A., Collins, J. & Wilson, S. (1988). Characterization of DNA polymerase beta mRNA: cell-cycle and growth response in cultured human cells. Nucl. Acids Res. 16, 9589-9596. 3. Bhagwat, A., Sanderson, R. & Lindahl, T. (1999). Delayed DNA joining at 30 mismatches by human DNA ligases. Nucl. Acids Res. 27, 4028-4033. 4. Canitrot, Y., Cazaux, C., Frechet, M., Bouayadi, K., Lesca, C., Salles, B. & Hoffmann, J. S. (1998). Overexpression of DNA polymerase beta in cell results in a mutator phenotype and a decreased sensitivity to anticancer drugs. Proc. Natl Acad. Sci. USA, 95, 12586-12590.
Role of Excess Pol in DNA Replication 5. Canitrot, Y., FreÂchet, M., Servant, L., Cazaux, C. & Hoffmann, J. S. (1999). Overexpression of DNA polymerase beta: a genomic instability enhancer process. FASEB J. 13, 1107-1111. 6. Kornberg, A. & Baker, T. A. (1992). DNA Replication, W. H. Freeman and Co., New York. 7. Bouayadi, K., Hoffmann, J., Fons, P., Tiraby, M., Reynes, J. & Cazaux, C. (1997). Overexpression of DNA polymerase b sensitizes mammalian cells to 20 ,30 -dideoxyxytidine and 30 -azido-30 -deoxythymidine. Cancer Res. 57, 110-116. 8. Kamath-Loeb, A., Hizi, A., Kasai, H. & Loeb, L. (1997). Incorporation of the guanosine triphosphate analogs 8-oxo-dGTP and 8-NH2-dGTP by reverse transcriptases and mammalian DNA polymerases. J. Biol. Chem. 272, 5892-5898. 9. Roberts, J., Nguyen, D. & Kunkel, T. (1993). Frameshift ®delity during replication of double-stranded DNA in HeLa cell extracts. Biochemistry, 32, 40834089. 10. Srivastava, D., Husain, I., Arteaga, C. & Wilson, S. (1999). DNA polymerase beta expression differences in selected human tumors and cell lines. Carcinogenesis, 20, 1049-1054. 11. Minnick, D. T., Pavlov, Y. I. & Kunkel, T. A. (1994). The ®delity of the human leading and lagging strand DNA replication apparatus with 8-oxodeoxyguanosine triphosphate. Nucl. Acids Res. 22, 56585664. 12. Sekowski, J., Malkas, L., Schnaper, L., Bechtel, P., Long, B. & Hickey, R. (1998). Human breast cancer cells contain an error-prone DNA replication apparatus. Cancer Res. 58, 3259-3263. 13. Kunkel, T. (1985). The mutational speci®cty of DNA polymerase-beta during in vitro DNA synthesis. Production of frameshift, base substitution, and deletion mutations. J. Biol. Chem. 260, 5787-5796. 14. Pavlov, Y. I., Minnick, D. T., Izuta, S. & Kunkel, T. A. (1994). DNA replication ®delity with 8-oxodeoxyguanosine triphosphate. Biochemistry, 33, 46954701. 15. Thomas, D., Roberts, J. & Kunkel, T. (1991). Heteroduplex repair in extracts of human Hela cells. J. Biol. Chem. 266, 3744-3751. 16. Miller, H., Prasad, R., Wilson, S. H., Johnson, F. & Grollman, A. P. (2000). 8-OxodGTP incorporation by DNA polymerase beta is modi®ed by active-site residue Asn279. Biochemistry, 39, 1029-1033. 17. Wilson, S. (1998). Mammalian base excision repair and DNA polymerase beta. Mutat. Res. 407, 203-215. 18. Sweasy, J. B. & Loeb, L. A. (1992). Mammalian DNA polymerase beta can substitute for DNA polymerase I during DNA replication in Escherichia coli. J. Biol. Chem. 267, 1407-1410. 19. Canitrot, Y., Hoffmann, J. S., Calsou, P., Hayakawa, H., Salles, B. & Cazaux, C. (2000). Nucleotide excision repair DNA synthesis by excess DNA polymerase beta: a potential source of genetic instability in cancer cells. FASEB J. 14, 1765-1774. 20. Washington, M., Johnson, R., Prakash, S. & Prakash, L. (1999). Fidelity and processivity of Saccharomyces cerevisiae of DNA polymerase eta. J. Biol. Chem. 274, 36835-36838. 21. Matsuda, T., Bebeneck, K., Masutani, C., Hanaoka, F. & Kunkel, T. (2000). Low ®delity DNA synthesis by human DNA polymerase-eta. Nature, 404, 10111013. 22. Woodgate, R. (1999). A plethora of lesion-replicating DNA polymerases. Genes Dev. 13, 2191-2195.
Role of Excess Pol in DNA Replication 23. Radman, M. (1999). Enzymes of evolutionary change. Nature, 401, 866-868. 24. Ohashi, E., Ogi, T., Kusumoto, R., Iwai, S., Masutani, C., Hanaoka, F. & Ohmori, H. (2000). Error-prone bypass of certain DNA lesions by the human DNA polymerase kappa. Genes Dev. 14, 1589-1594. 25. Tissier, A., McDonald, J., Franck, E. & Woodgate, R. (2000). Poliota, a remarkably error-prone human DNA polymerase. Genes Dev. 14, 1642-1650. 26. Bebenek, K., Matsuda, T., Masutani, C., Hanaoka, F. & Kunkel, T. A. (2001). Proofreading of DNA polymerase eta-dependent replication errors. J. Biol. Chem. 276, 2317-2320. 27. Yao, X., Buernmeyer, A., Narayanan, L., Tran, D., Baker, S., Prolla, T. et al. (1999). Different mutator
1047 phenotypes in Mlh1- versus Pms2- de®cient mice. Proc. Natl Acad. Sci. USA, 96, 6850-6855. 28. Kumar, A., Widen, S., Williams, K., Kedar, P., Karpel, R. & Wilson, S. (1990). Studies of the domain structure of mammalian DNA polymerase beta. J. Biol. Chem. 265, 2124-2131. 29. Mo, J. Y., Maki, H. & Sekiguchi, M. (1992). Hydrolytic elimination of a mutagenic nucleotide, 8-oxodGTP, by human 18-kilodalton protein: sanitization of nucleotide pool. Proc. Natl Acad. Sci. USA, 89, 11021-11025. 30. Roberts, J. & Kunkel, T. (1996). Fidelity of DNA replication. In DNA Replication in Eukaryotic Cells Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Edited by M. Yaniv (Received 14 September 2001; received in revised form 26 November 2001; accepted 26 November 2001)
J. Mol. Biol. (2002) 315, 1039±1047
Involvement of DNA Polymerase b in DNA Replication and Mutagenic Consequences Laurence Servant1, Anne Bieth1, Hiroshi Hayakawa2, Christophe Cazaux1* and Jean-SeÂbastien Hoffmann1* 1
The Institut de Pharmacologie et Biologie Structurale, UMR CNRS 5089, 31077 Toulouse ceÂdex 4, France 2
Department of Medical Biochemistry, Graduate School of Medical Sciences, Kyushu University, Fukuoka 812-8582, Japan
Overexpression in mammalian cells of the error-prone DNA polymerase b (Pol b) has been found to increase the spontaneous mutagenesis. Here, we investigated a possible mechanism used by Pol b to be a genetic instability enhancer: its interference in replicative DNA synthesis, which is normally catalysed by the DNA polymerases a, d and e. By taking advantage of the ability to incorporate ef®ciently into DNA the chain terminator ddCTP as well as the oxidised nucleotide 8-oxo-dGTP, we show here that puri®ed Pol b can compete with the replicative DNA polymerases during replication in vitro of duplex DNA when added to human cell extracts. We found that involvement of Pol b lowers replication ®delity and results in a modi®ed error-speci®city. Furthermore, we demonstrated that involvement of Pol b occurred during synthesis of the lagging strand. These in vitro data provide one possible explanation of how overexpression of the enzyme could perturb the genetic instability in mammalian cells. We discuss these ®ndings within the scope of the up-regulation of Pol b in many cancer cells. # 2002 Elsevier Science Ltd.
*Corresponding authors
Keywords: DNA replication; mutagenesis; DNA polymerases; cancer
Introduction Mammalian cells contain many DNA polymerases involved in different DNA transactions. While DNA polymerases (Pol) a, d, and e are devoted to the genomic DNA replication process, other DNA polymerases are con®ned to additional speci®c roles. For example, in mammalian somatic cells, Pol b, a monomeric 39 kDa enzyme, is involved in the the single-nucleotide base excision repair (BER) pathway,1 which processes small DNA lesions such as oxidized or alkylated bases. This DNA polymerase can be distinguished from the DNA polymerases involved in the elongation step of the genome replication (Pol d and Pol e) by its high level of in®delity in replicating DNA in vitro, due to the lack of associated proofreading activity and a poor ability to discriminate nucleotides. Probably because of these error-prone features, its expression is tightly regulated in vivo, being constant and low throughout the cell-cycle.2 Limitation of the inaccuracy of Pol b in the BER Abbreviations used: Pol, DNA polymerase a; BER, base excision repair. E-mail addresses of the corresponding authors: [email protected]; [email protected] 0022-2836/02/051039±9 $35.00/0
pathway has been suggested as a mechanism to avoid the high mutation rate subsequent to its repair action: DNA ligase III, which catalyses the joining step, can discriminate strongly between a correctly paired versus a mispaired residue at the 30 position of a nick in DNA, and thus delay the joining reaction, allowing the removal of the mismatched terminal residue by a distinct 30 exonuclease.3 By using conventional methodologies testing the appearance of mutational events at the Na-KATPase locus as well as at the locus encoding the purine salvage enzyme hypoxanthine guanine phosphoribosyl transferase (HPRT), leading to ouabain-resistance and 6-thioguanine-resistance phenotypes respectively, we demonstrated previously that excess Pol b in mammalian cells resulted in a fourfold to 12-fold increased rate of spontaneous mutagenesis.4 Moreover, we demonstrated recently that ectopic expression of Pol b in mammalian cells induces chromosomal instability and potentiates tumorigenesis in nude immunode®cient mice (unpublished results). Thus, it appears that alteration of Pol b expression may induce major genetic changes that contribute to the malignant phenotype. # 2002 Elsevier Science Ltd.
1040 Based on these data, we hypothesized that an excess of the error-prone Pol b can perturb the well-de®ned speci®c functions of error-free DNA polymerases during DNA transactions such as repair, replication or recombination pathways.5 Here, we have investigated whether such a mutagenic interference could be observed during in vitro DNA replication of duplex DNA, which in a normal somatic cell, requires, the coordinated activity of the replicative error-free Poly a, d and e, but not b.6 To conduct the experiments, we took advantage of the fact that dideoxycytidine (ddC), which terminates chain elongation when incorporated into DNA,7 as well as 8-oxo-dGTP, an oxidized form of dGTP, are substrates of Pol b but are incorporated poorly by the replicative polymerases.8 By using the archetypal simian virus 40 (SV40) origin-containing plasmid replication assay, which reconstitutes the mammalian DNA replication fork,9 we present evidence that addition of Pol b to cell extracts renders the process hypersensitive to ddCTP and enhances misincorporation of the 8oxo-dGTP analogue, demonstrating that the enzyme can be associated with the replicative event in the context of its overexpression. Furthermore, we showed that this involvement occurred during the synthesis of the lagging strand and resulted in an increased frequency of mutation. The modi®ed error-speci®city spectrum induced by Pol b that we observed is in accordance with the mutagenic features of the enzyme. The consequences of the interference in DNA replication by error-prone DNA polymerases such as Pol b is discussed in relation to the design of the putative role in cell transformation of the misregulation of newly discovered error-prone DNA polymerases.
Results Incorporation of ddCTP into DNA by Pol b during SV40 DNA replication T-antigen-dependent replication of a DNA plasmid using cell-free extracts from human Hela cells was performed at 37 C with increasing amount of puri®ed Pol b in the absence or presence of the nucleotide analogue ddCTP, at a mass ratio of ddCTP to dCTP equal to 0.5. DNA was then puri®ed, linearized by BamHI, and subjected to electrophoresis in a 1 % (w/v) agarose gel, followed by ethidium bromide staining and autoradiography (Figure 1). We veri®ed that semi-conservative replication was occurring, since the products were resistant to DpnI, which cuts only fully methylated 50 -GATC sequences. We veri®ed that without T-antigen, no incorporation was detectable in the presence of various amounts of Pol b during the reaction, meaning that no background repair synthesis occurred under our conditions. We found that, without ddCTP, the presence of excess Pol b itself did not affect the ef®ciency of the replication reaction quantitatively (Figure 1(a)). When we added ddCTP, known to inhibit strongly the
Role of Excess Pol in DNA Replication
polymerisation of DNA in vitro catalysed by Pol b but not by the replicases Pol d and Pol e, the presence of Pol b induced a hypersensitivity of DNA synthesis to this nucleotide analogue (Figure 1(a), upper part). We measured a 66 % inhibition with 0.012 unit of Pol b, while we were not able to detect a radioactive signal in the presence of 0.024 unit, suggesting that, for this concentration, Pol b interfered with the replication forks of most plasmids that undergo DNA replication (see quantitative data in Figure 1(b)). An internal control DNA incorporated into the reaction mixture after the six hour replication incubation was used to verify the ef®ciency and homogeneity of DNA puri®cation after the reaction (see Figure 1(a), lower part). We did not observe any effect on SV40 replication by the same concentrations of ddCTP when heatdenatured Pol b was added, demonstrating that the polymerase activity is necessary for the inhibition (not shown). It is important to note that in all these experiments the enzyme concentrations had a physiological relevance and did not represent a great excess of enzyme that could introduce artefacts. Firstly, by using western blotting experiments (not shown), we observed that the protein amounts were less than physiological ranges measured in several human tumoral cells.10 Secondly, we conducted calibration assays in order to determine the speci®c activity of Pol b relative to replicative cell extracts, by measuring nucleotide incorporation on activated calf thymus DNA preincubated with DNase I (Figure 2). These assays showed that concentration ranges of puri®ed Pol b added to the cell extracts represented 3 to 20 % compared to the global activity of the cell extracts. In the SV40 replication assay, where the enzyme was mixed with the extracts, this relative range of activity should be lowered considerably, since Pol b most probably did not interfere at the replicative forks with an ef®ciency comparable to that of the other replicative DNA polymerases. This was con®rmed by using primer extension reactions on oligonucleotide in the presence of the chainterminator ddCTP (not shown). Incorporation of 8-oxo-dGTP into DNA by Pol b during SV40 DNA replication The mutagenic analogue 8-oxo-dGTP, resulting from the conversion of dGTP by reactive oxygen species, has been shown to be incorporated ef®ciently into DNA in vitro by Pol b. We took advantage of this feature to further demonstrate the speci®c involvement of Pol b during the elongation of the replicative forks using the SV40 replication assay. We conducted DNA replication reactions comparable with the ddCTP experiments described above, but in the presence of 32P-labelled 8-oxodGTP that we prepared (see Materials and Methods). Figure 3 shows that in the absence of cold dGTP, incorporation of the labelled oxidized nucleotide is enhanced by threefold when 0.012 unit of Pol b is added to the replicative extracts.
Role of Excess Pol in DNA Replication
1041
Figure 1. (a) T-antigen-dependent replication of 100 ng of pCineo using 400 mg of cell-free extracts from human Hela cells. DNA replication was performed using the conventional SV40 replication assay at 37 C for six hours with increasing amounts of puri®ed Pol b (0.0024u, 0.0048u, 0.012 and 0.024 unit), in the presence or absence of the nucleotide analogue ddCTP at a ratio of dCTP to ddCTP equal to 0.5. DNA was puri®ed in the presence of an internal pCDNA2 control DNA, linearized by BamHI, and subjected to electrophoresis in a 1 % agarose gel, followed by ethidium bromide staining (lower part) and autoradiography (upper part). The arrow indicates the position of the linearized product. (b) Quanti®cation analysis of the resolved radioactive bands on the gel seen in (a) was achieved by the PhosphoImager Storm-system analysis using Imagequant software.
Such an increase in labelled 8-oxo-dGTP incorporation is strengthened in the presence of 5 mM cold dGTP, con®rming the poor discrimination of Pol b between the correct dGTP and its 8-oxo-dGTP competitor antagonist, in contrast to the replicative polymerases. We cannot rule out the possibility that hydolysis of the incorporated 8-oxo-dGTP by signi®cant levels of the 8-oxoG DNA glycosylases (hOGG1, hOGG2, and hNTH1) in the cell extracts occurred during the six hour incubation, but this activity would have the same ef®ciency for both conditions. Together with the ddCTP incorporation data, this demonstrates that physiological level of Pol b found in several human cancer cells can interfere with SV40 replication. Fidelity of SV40 DNA replication in the presence of Pol b In order to appreciate the mutagenic impact of the Pol b involvement in DNA replication, the ®delity of SV40 DNA replication was investigated by using the pBK-CMV and M13mp2SV ori Left and ori Right vectors containing the target reporter lacZa sequence as a mutagenic target. The replication-induced mutations occurring in the target gene, which resulted in the expression of a non-
functional-b-gal enzyme, were analysed (see Tables 1-3). We determined the mutant frequencies for the replicated DNA samples after their electroporation into a mutS Escherichia coli strain to ®x the errors, and their subsequent insertion into an indicator E. coli strain to score colourless or light blue mutants among total clones after plating the cells in the presence of 5-bromo-4-chloro-3-indolyl-b,Dgalactopyranoside (Xgal) and IPTG. The mutant frequencies found after SV40 DNA replication of the pBK-CMV and M13mp2SV vectors were of the order of magnitude of those measured previously.11,12 In the presence of 0.012 unit of Pol b, the number of mutants was increased signi®cantly and reproductively by twofold to threefold (Table 1) in three independent reactions, depending on which substrate is considered. This result is an additional argument demonstrating that Pol b interferes with DNA replication, since it affects the accuracy of this process. Error speci®city was analysed by sequencing 50 replicated pBK-CMV mutants from the control replication and the replication in the presence of Pol b (Figure 4). In contrast to the spectrum observed for the control conditions, where some mutational hot-spots were detected, notably at positions 1159 and 1173-1182 (a G C-rich sequence propitious for slippage
1042
Role of Excess Pol in DNA Replication
described previously for gapped DNA by using the lacZ gene target.13 Strand-specificity of Pol b - induced mutations during SV40 replication
Figure 2. Calibration of DNA synthesis activity. Relative activities of 0.010 unit of added Pol b (open circles) compared to 45 mg of cell extracts (®lled squares) quanti®ed on activated calf thymus DNA by measuring the incorporation of [a-32-P]dCTP into acid-insoluble materials at 37 C for the indicated incubation times. One unit is the amount of Pol b required to catalyse the incorporation of 1 pmol of [a-32P]dGTP in one hour at 37 C.
during DNA polymerisation), the errors occurring when Pol b was present in the cell extract were distributed essentially all along the sequence of the a-peptide, while one hot-spot remained targeted. This suggests some Pol b intervention during the elongation step of the replication reaction. Base substitutions were detected at a frequency of 61.5 % and 45.5 % for control and excess Pol b reactions, respectively, while frameshifts represented the majority of the changes for the excess Pol b reaction (54.5 %) against 38.5 % for the control replication. This change in error speci®city favouring the frameshift mutations is in accordance with the mutational speci®city of the enzyme in vitro
In order to de®ne more precisely on which strand excess Pol b could interfere during SV40 replication, we used the M13mp2 ori Left and M13mp2 ori Right vectors that allow strandspeci®c errors analysis, since these templates contain the SV40 origin located on opposite sides relative to the LacZa mutational target and close to it. The 8-oxo-dGTP nucleotide analogue is known to be misincorporated during SV40 replication opposite adenine,14 leading to A T ! C G transversions, produced rarely in the presence of normal dNTPs.15 Therefore, when 8-oxo-dGTP is added to SV40 reactions, its miscoding de®nes the strand into which the 8-oxo-dGTP is incorporated. By taking advantage of the fact that Pol b is capable of ef®cient incorporation of 8-oxo-dGTP with a preferential mispairing with a template dA compared to a template dC (24/1),16 we got information about strand speci®city of Pol b-dependent synthesis during M13mp2-ori-Left/Right replication by analysing the sequence of mutant phages. According to previous experiments,11 we measured a two- to threefold increase in mutation frequency by the sole addition of 8-oxo-dGTP during the replication of both vectors (Table 2). Addition of puri®ed Pol b to the reactions resulted in an additional 2.4-fold enhancement of the mutant frequency, con®rming the capacity of the enzyme for ef®cient 8-oxo-dGTP incorporation. For determination of strand speci®city, we performed DNA sequence spectra of collections of independent ori Left mutants after replication in the presence of 8-oxo-dGTP, analysing the LacZacomplementation target sequence from positions ÿ45 to 197. In all, 23 mutants were sequenced from control reactions and 33 mutants from reaction in the presence of excess Pol b. A T ! C G transversions were the primary mutations recovered, consistent with the expected misincorpora-
Table 1. Fidelity of SV40 DNA replication in the absence or in the presence of excess Pol b Number of plaques scored DNA vector pBK-CMV M13mp2SV ori Left M13mp2SV ori Right
Conditions Unreplicated Replicated without Pol b Replicated with Pol b Replicated without Pol b Replicated with Pol b Replicated without Pol b Replicated with Pol b
Total
Mutants
Mutant frequencies (10ÿ4)
28,640 141,390 209,762 43,040 24,018 115,611 60,956
184 918 2953 18 31 63 95
64 65 140 4.2 12.9 5.4 15.6
Error ratea (10ÿ5) <0.05 4.0 <0.03 0.39 <0.08 0.5
SV40 reactions were performed as described in Materials and Methods from three independent reactions. Error rate was calculated by subtracting the background mutant frequency from the average mutant frequency. This value was divided by 0.5 and then by the number of detectable mispairs sites for each vector: 482 for M13mp2SV9 and 362 for pBB-CMV.10 In the case of M13 vectors, the background mutant frequency found was 3.5 10ÿ4. a
1043
Role of Excess Pol in DNA Replication
Figure 3. Incorporation of 8-oxodGTP into DNA by excess Pol b during in vitro SV40 replication reaction. The replication conditions are identical with those used for the ddCTP experiment in Figure 2 except that 40 mM [a-32P]8-oxodGTP were added in the presence (right part) or the absence (left part) of 5 mM cold dGTP.
tion of 8-oxo-dGTP opposite the template adenine residue. We also detected in three mutants (one from the control reaction and two from the reaction Pol b) C G ! A T transversions, which could represent misincorporation of dAMP opposite the template 8-oxo-dGMP in a second round of DNA replication. We found that, in the presence of Pol b, most of the mutations arose from 8-oxo-dGTP misincorporation opposite adenine residues on the lagging strand (Table 3), whereas these modi®cations were distributed quite equivalently on both strands for the control reaction. We calculated a fourfold increase in mispair rate on the lagging strand and we observed no variation on the leading strand. Taken together, these ®ndings suggest strongly that the lagging strand is the preferential target of Pol b when interfering with the DNA replication fork elongation.
Discussion We show here that Pol b, an enzyme required in somatic cells in the BER pathway,1 can interfere with duplex DNA replication in vitro when up-represented and renders the process inaccurate. Such a result was obtained by using the well-calibrated SV40 replication assay, which reconstitutes the mammalian DNA replication fork by providing all the factors required for bidirectional replication of double-stranded DNA in a human cell extract except for the SV40 T-antigen. Cell extracts include DNA polymerase a and its associated primase for the initiating step and for starting the synthesis of Okazaki fragments, DNA polymerase d assisted by replication factor C (RF-C) and proliferating cell nuclear antigen (PCNA), and eventually DNA
polymerase e, for the synthesis of the leading and lagging strands. By adding to the reaction nucleotide analogues such as ddC and 8-oxo-dG, incorporated into DNA preferentially by Pol b, we present evidence here for some intervention of Pol b during the elongation process. We further employed a strategy based on the replication of vector templates bearing both orientations of the SV40 origin in order to examine whether Pol b competes for primer termini on the leading strand, the lagging strand, or both. We were able to demonstrate that the lagging strand was the preferred target of Pol b during the fork elongation process. This is in accordance with the properties of the enzyme, whose preferential substrate has been demonstrated to be gapped DNA,17 since the synthesis of the lagging strand, which requires the formation and the joining of Okazaki fragments, produces gapped DNA that needs to be ®lled. This result is supported by a previous observation showing that ectopic expression of rat Pol b in E. coli can restore growth in a DNA polymerase Idefective bacterial mutant by increasing the rate of joining of Okazaki fragments.18 Moreover, we showed recently that gap-®lling during the nucleotide excision repair (NER), a pathway processed normally by DNA polymerases d and e, can be achieved by Pol b when over-regulated in cells.19 Although the in vitro SV40 DNA replication assay did not re¯ect completely the in vivo situation, the error-prone involvement of excess Pol b in DNA replication has biological importance. First, it may explain why an increase in spontaneous mutation frequency occurred in mammalian cells overexpressing the enzyme.4 It may represent a mechanistic basis for our recent data
Table 2. Effect of excess Pol b on 8-oxo-dGTP-induced mutation frequency during SV40 replication Conditions Replicated without Pol b Mispairs on leading strand Mispairs on lagging strand Replicated with Pol b Mispairs on leading strand Mispairs on lagging strand
Transversion number
Mutant ratio (%)
Error ratea (10ÿ6)
13 10
56 43
41.5 35.8
10 23
30 70
58.1 149.6
a Error rate was calculated as described in the legend to Table 1, except that the number of detectable AT ! CG transversion sites for each M13 vector was 53 for ori Left and 59 for ori Right.11
1044
Role of Excess Pol in DNA Replication
Figure 4. Spectrum of mutations on the target reporter lacZa sequence during SV40 replication in the absence (above) or in the presence (below) of 0.012 unit of puri®ed Pol b: 50 colourless mutant colonies from each replication reaction were sequenced. The ÿ1 one base deletions are shown as open triangles, insertions are indicated as ®lled triangles together with the inserted base as a bold character, and single-base substitutions are indicated by the nucleotide concerned.
showing that excess Pol b provided these cells with a new phenotype that creates competent tumour cells when injected into immunode®cient mice ( . Bergoglio et al., unpublished results). High levels of Pol b have been found in various tumours such as glioma and lymphoma, breast, colon and prostate,10 as well as in ovarian,19 leukemia and lymphoma tumour cell lines (unpublished results). It is possible that, like the de®ciency in an error-correcting mechanism, error-prone DNA synthesis during replication may contribute to the global genetic instabilities and the great variability displayed by these cancer cells. Such a ``mis-invol-
vement'' of Polb could take place through the context of its overexpression in other error-free DNA transactions pathways involving a long patch DNA synthesis such as NER,19 and DNA homologous recombination or mismatch repair. New cellular error-prone DNA polymerases have been described recently.20,21 These enzymes, which belong to the UmuC/DinB/Rev1/Rad30 superfamily of DNA polymerases that are able to bypass lesions in DNA,22 are part of stressinducible processes that allow them to function only when high mutation rates are advantageous. Among these DNA mutases23 that have been
Table 3. Effect of excess Pol b on the speci®city of leading or lagging strand errors for dA 8-oxo-dGMP mispairs after M13mp2SV-ori-Left replication Number of plaques scored
DNA vector M13mp2SV ori Left 8-oxo-dGTP M13mp2SV ori Right 8-oxo-dGTP
Conditions Replicated Replicated Replicated Replicated
without Pol b with Pol b without Pol b with Pol b
Total
Mutants
Mutant frequencies (10ÿ4)
30,967 49,151 9598 4476
32 120 11 12
10.3 24.4 11.4 26.8
Error ratea (10ÿ5) 2.6 7.9 2.7 7.9
a Error rate was calculated as described in the legend to Table 1, except that the number of detectable mutation sites for each strand was 28 for the leading strand and 25 for the lagging strand.11 The error rate obtained was then multiplyed by the corresponding mutant ratio.
1045
Role of Excess Pol in DNA Replication
isolated in human cells, Pol k,24 Pol Z21 and Pol i25 produce errors at even higher rates than Pol b in copying undamaged DNA in vitro. A recent report demonstrated that Pol Z can also compete with the replicative DNA polymerases during replication of duplex DNA, and lowers replication ®delity by a factor of 5 to 8.26 Interestingly, those authors showed that the proofreading exonuclease activities reduce the mutagenic potential of Pol Z during the replicative process. It is possible that such an editing process of the Pol b-induced errors described in our work may be performed by an exonuclease activity associated with the replication machinery. The low processivity of Pol b, in addition to its ability to extend mispairs less ef®ciently than matched termini, giving more chance for extrinsic proofreading, may explain why excess Pol b lowered replication ®delity by only a factor of 2 to 3. In addition, misincorporation of 8-oxodGTP mediated by excess Pol b during SV40 replication may be underestimated, since dA 8-oxodGMP mispairs have been demonstrated to be proofread exonucleolytically by cell extracts during synthesis of both strands. Nevertherless, such mutagenicity rates in vivo have been shown to have dramatic effects on tumour development.27 To our knowledge, no direct evidence has been described for a possible relationship between upregulation of error-prone DNA polymerases and the genetic variability of cancer cells, although the impact of such enzymes could be of great importance within the scope of tumour development. Our data presented here with Pol b, which need to be enlarged to other members of this enzymatic family, open the possibility to explore the putative role of mutases in carcinogenesis by interfering with the DNA replication process.
In vitro SV40 DNA replication reactions
Materials and Methods
32 P-radiolabelled 8-oxodGTP was prepared as described29 with some modi®cations. Brie¯y, 6 mM dGTP was incubated at 37 C for two hours in the dark in a reaction mixture containing 100 mM sodium phosphate (pH 6.8), 30 mM ascorbic acid, 100 mM H2O2. A portion (50 ml) of the reaction mixture was loaded onto a MonoQ column and eluted with a linear gradient (0.01 M-1 M) of triethylammonium hydrogen carbonate (pH 7.0) at a ¯ow-rate of 1 ml/minute. 8-Oxo-dGTP was monitored with a UV detector (254 nm) and eluted with a retention time about one minute later than that for dGTP. Fractions containing 8-oxo-dGTP were combined, lyophilized, dissolved in 12.5 mM citric acid, 25 mM sodium acetate, 30 mM NaOH, 10 mM acetic acid (pH 5.4) and applied to a C18 reverse phase column TOSOH TSK-GEL ODS-80T equibrated with the same buffer at a ¯ow-rate of 1 ml/minute. 8-Oxo-dGTP was eluted with a retention about ®ve minutes later than that for dGTP. The sample was further puri®ed with the MonoQ column as described above. The 8-oxo-dGTP was lyophilized, dissolved in 20 mM sodium phosphate (pH 6.8) and stored at ÿ20 C. Purity and quantity of the preparation were examined with HPLC and UV spectrum analyses. No dGTP was detectable in the ®nal preparation of 8-oxo-dGTP. SV40 replication conditions were the same as described above, except that 44 mM 32 P-radiolabelled 8-oxodGTP was added to the reaction
Calibration of DNA polymerization Calf thymus DNA pre-incubated with DNase I (0.4 mg) was used as template in assay reactions (20 ml) at 37 C for various incubation times with different concentrations of either puri®ed Pol b or replicative Hela cell extracts in 50 mM Tris-HCl (pH 7.5), 6 mM MgCl2, 0.5 mM dithiothreitol, 300 mM each dATP, dGTP, dTTP, and 0.2 mCi of [32P]dCTP. At the end of the reaction, 10 mM EDTA was added and aliquots were spotted on Whatman glass ®lters (GF/C). Filters were washed with a 5 % (w/v) trichloroacetic acid (TCA), 1 % (w/v) sodium pyrophosphate solution and radioactivity was determined. Preparation of cell-free extracts for SV40 replication HeLa cells were grown as spinner cultures in RPMI 1640 (Biomedia) with 10 % (v/v) fetal calf serum, 4 mM glutamine, and antibiotics (50 units/ml of penicillin and 50 mg/ml of streptomycin) at 37 C in a humidi®ed 5 % (v/v) CO2 atmosphere. The cells were harvested by centrifugation in the upper part of their log phase growth and extracts were prepared as described.9
Replication reaction mixtures (25 ml) contained 30 mM Hepes (pH 7.8), 7 mM MgCl2, 200 mM each CTP, GTP, UTP, 4 mM ATP, 100 mM each dATP, dCTP, dTTP, 10 mM dGTP, and [a-32P]dGTP (4000 cpm/pmol; Amersham), 40 mM creatine phosphate (Sigma), 100 mg/ml of creatine phosphokinase (Sigma), 50-100 ng of DNA substrate, 0.5 mg of SV40 large T-antigen (Molecular Biology Resources), 400 mg of Hela cell extract, and various concentrations of rat DNA Pol b, puri®ed as described.28 One unit of Pol b corresponds to 1 nmol of dNTP incorporated into acid-insoluble materials at 37 C in 60 minutes, using activated calf thymus DNA pre-incubated with DNase I as substrate. Reactions without T-antigen were used as negative controls. SV40 replication reactions in presence of ddCTP After incubation at 37 C for the periods indicated in Figure 2 with or without 50 mM ddCTP (ddCTP to dCTP ratio of 0.5), reactions were quenched by adding an equal volume of stop solution (2 % (w/v) SDS, 2 mg/ml of proteinase K, 50 mM EDTA) and further incubation for one hour at 55 C. DNA (0.5 mg of pc-DNA II, In vitroGen) was added to each sample as internal puri®cation controls. Reaction products were puri®ed by extraction with phenol/chloroform/(25:24:1 by vol.) isoamyl alcohol followed by precipitation in ethanol. The DNA was resuspended in distilled water. The samples were then treated with BamHI and DpnI (New England Biolabs), and the restriction digests were separated on a 1 % (w/v) agarose gel. After staining the gel with ethidium bromide, internal control DNAs were quanti®ed. The gel was then dried, and autoradiography performed. Quanti®cation of the resolved radioactive bands on the gel was achieved by PhosphoImager Storm-system analysis using Imagequant software. SV40 replication reactions in presence of radiolabelled 8-Oxo-dGTP
1046 mixture in the presence or the absence of cold dGTP as noted. Mutagenesis assays As SV40 replication templates, we used the pBK-CMV vector (Stratagene), which contains the full SV40 origin of replication and the lacZa gene (at positions 810-1183), or the M13mp2SVori Right/Left vectors kindly provided by Dr Thomas Kunkel (NIEHS, NC), which contain the SV40 origin of replication located either side of the wildtype LacZa sequence.30 DNA substrate (40 to 50 ng) was replicated during six hours as described above, except for the 8-oxo-dGTP mutation-induced replication assay where 100 mM 8-oxo-dGTP and 10 mM dGTP were added to the reactions. After replication, the products were precipitated and subjected to DpnI digestion to eliminate unreplicated DNA templates. Products of a single round of replication carrying mutations that exist as heteroduplexes, pBKCMV replication products were electroporated into a MC1061mutS E. coli host (kindly provided by Professor Masson, Toulouse) and M13mp2SV ori Left/Right products were electroporated into NR9162 E. coli host (kindly provided by T. Kunkel, NIEHS, NC). Both bacterial strains being de®cient for mismatch repair, mutations generated during SV40 replication were ®xed. In the case of the pBK-CMV vector, DNA was extracted, electroporated in JM109 E. coli bacteria and then plated onto indicator plates containing Xgal and IPTG. In the case of M13 DNA, the NR9162-infected cells were plated immediately after electroporation in soft agar containing XGal, IPTG and a log-phase culture of CSH50. The inactivation of the a-complementation gene due to a mutation in the pBK-CMV/M13mp2 LacZa gene resulted in a colorless or light-blue phenotype of the colonies or of the plaques. To determine the speci®city of nucleotide errors, automated sequencing of the mutant colonies has been performed (MilleGen Co., Toulouse).
Acknowledgments This work was supported by La ``Ligue contre le cancer'' (Equipe LabelliseÂe).
References 1. Sobol, R., Horton, J., KuÈhn, R., Gu, H., Singhal, R., Prasad, R. et al. (1996). Requirement of mammalian DNA polymerase-beta in base-excision repair. Nature, 379, 183-186. 2. Zmudzka, B., Fornace, A., Collins, J. & Wilson, S. (1988). Characterization of DNA polymerase beta mRNA: cell-cycle and growth response in cultured human cells. Nucl. Acids Res. 16, 9589-9596. 3. Bhagwat, A., Sanderson, R. & Lindahl, T. (1999). Delayed DNA joining at 30 mismatches by human DNA ligases. Nucl. Acids Res. 27, 4028-4033. 4. Canitrot, Y., Cazaux, C., Frechet, M., Bouayadi, K., Lesca, C., Salles, B. & Hoffmann, J. S. (1998). Overexpression of DNA polymerase beta in cell results in a mutator phenotype and a decreased sensitivity to anticancer drugs. Proc. Natl Acad. Sci. USA, 95, 12586-12590.
Role of Excess Pol in DNA Replication 5. Canitrot, Y., FreÂchet, M., Servant, L., Cazaux, C. & Hoffmann, J. S. (1999). Overexpression of DNA polymerase beta: a genomic instability enhancer process. FASEB J. 13, 1107-1111. 6. Kornberg, A. & Baker, T. A. (1992). DNA Replication, W. H. Freeman and Co., New York. 7. Bouayadi, K., Hoffmann, J., Fons, P., Tiraby, M., Reynes, J. & Cazaux, C. (1997). Overexpression of DNA polymerase b sensitizes mammalian cells to 20 ,30 -dideoxyxytidine and 30 -azido-30 -deoxythymidine. Cancer Res. 57, 110-116. 8. Kamath-Loeb, A., Hizi, A., Kasai, H. & Loeb, L. (1997). Incorporation of the guanosine triphosphate analogs 8-oxo-dGTP and 8-NH2-dGTP by reverse transcriptases and mammalian DNA polymerases. J. Biol. Chem. 272, 5892-5898. 9. Roberts, J., Nguyen, D. & Kunkel, T. (1993). Frameshift ®delity during replication of double-stranded DNA in HeLa cell extracts. Biochemistry, 32, 40834089. 10. Srivastava, D., Husain, I., Arteaga, C. & Wilson, S. (1999). DNA polymerase beta expression differences in selected human tumors and cell lines. Carcinogenesis, 20, 1049-1054. 11. Minnick, D. T., Pavlov, Y. I. & Kunkel, T. A. (1994). The ®delity of the human leading and lagging strand DNA replication apparatus with 8-oxodeoxyguanosine triphosphate. Nucl. Acids Res. 22, 56585664. 12. Sekowski, J., Malkas, L., Schnaper, L., Bechtel, P., Long, B. & Hickey, R. (1998). Human breast cancer cells contain an error-prone DNA replication apparatus. Cancer Res. 58, 3259-3263. 13. Kunkel, T. (1985). The mutational speci®cty of DNA polymerase-beta during in vitro DNA synthesis. Production of frameshift, base substitution, and deletion mutations. J. Biol. Chem. 260, 5787-5796. 14. Pavlov, Y. I., Minnick, D. T., Izuta, S. & Kunkel, T. A. (1994). DNA replication ®delity with 8-oxodeoxyguanosine triphosphate. Biochemistry, 33, 46954701. 15. Thomas, D., Roberts, J. & Kunkel, T. (1991). Heteroduplex repair in extracts of human Hela cells. J. Biol. Chem. 266, 3744-3751. 16. Miller, H., Prasad, R., Wilson, S. H., Johnson, F. & Grollman, A. P. (2000). 8-OxodGTP incorporation by DNA polymerase beta is modi®ed by active-site residue Asn279. Biochemistry, 39, 1029-1033. 17. Wilson, S. (1998). Mammalian base excision repair and DNA polymerase beta. Mutat. Res. 407, 203-215. 18. Sweasy, J. B. & Loeb, L. A. (1992). Mammalian DNA polymerase beta can substitute for DNA polymerase I during DNA replication in Escherichia coli. J. Biol. Chem. 267, 1407-1410. 19. Canitrot, Y., Hoffmann, J. S., Calsou, P., Hayakawa, H., Salles, B. & Cazaux, C. (2000). Nucleotide excision repair DNA synthesis by excess DNA polymerase beta: a potential source of genetic instability in cancer cells. FASEB J. 14, 1765-1774. 20. Washington, M., Johnson, R., Prakash, S. & Prakash, L. (1999). Fidelity and processivity of Saccharomyces cerevisiae of DNA polymerase eta. J. Biol. Chem. 274, 36835-36838. 21. Matsuda, T., Bebeneck, K., Masutani, C., Hanaoka, F. & Kunkel, T. (2000). Low ®delity DNA synthesis by human DNA polymerase-eta. Nature, 404, 10111013. 22. Woodgate, R. (1999). A plethora of lesion-replicating DNA polymerases. Genes Dev. 13, 2191-2195.
Role of Excess Pol in DNA Replication 23. Radman, M. (1999). Enzymes of evolutionary change. Nature, 401, 866-868. 24. Ohashi, E., Ogi, T., Kusumoto, R., Iwai, S., Masutani, C., Hanaoka, F. & Ohmori, H. (2000). Error-prone bypass of certain DNA lesions by the human DNA polymerase kappa. Genes Dev. 14, 1589-1594. 25. Tissier, A., McDonald, J., Franck, E. & Woodgate, R. (2000). Poliota, a remarkably error-prone human DNA polymerase. Genes Dev. 14, 1642-1650. 26. Bebenek, K., Matsuda, T., Masutani, C., Hanaoka, F. & Kunkel, T. A. (2001). Proofreading of DNA polymerase eta-dependent replication errors. J. Biol. Chem. 276, 2317-2320. 27. Yao, X., Buernmeyer, A., Narayanan, L., Tran, D., Baker, S., Prolla, T. et al. (1999). Different mutator
1047 phenotypes in Mlh1- versus Pms2- de®cient mice. Proc. Natl Acad. Sci. USA, 96, 6850-6855. 28. Kumar, A., Widen, S., Williams, K., Kedar, P., Karpel, R. & Wilson, S. (1990). Studies of the domain structure of mammalian DNA polymerase beta. J. Biol. Chem. 265, 2124-2131. 29. Mo, J. Y., Maki, H. & Sekiguchi, M. (1992). Hydrolytic elimination of a mutagenic nucleotide, 8-oxodGTP, by human 18-kilodalton protein: sanitization of nucleotide pool. Proc. Natl Acad. Sci. USA, 89, 11021-11025. 30. Roberts, J. & Kunkel, T. (1996). Fidelity of DNA replication. In DNA Replication in Eukaryotic Cells Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Edited by M. Yaniv (Received 14 September 2001; received in revised form 26 November 2001; accepted 26 November 2001)