Enhanced mutagenic potential of 8-oxo-7,8-dihydroguanine when present
http://www.100md.com
《核酸研究医学期刊》
Medical Research Council, Radiation and Genome Stability Unit, Harwell, Oxon OX11 0RD, UK and 1 Department of Ion Beam Applied Biology, Japan Atomic Energy Research Institute, Watanuki-machi 1233, Takasaki, Gunma 370-1292, Japan
*To whom correspondence should be addressed. Tel: +44 1235 841017; Fax: +44 1235 841200; Email: p.oneill@har.mrc.ac.uk
ABSTRACT
The formation of clustered DNA damage sites is a unique feature of ionizing radiation. Recent studies have shown that the repair of lesions within clusters may be compromised, but little is understood about the mutagenic consequences of such damage sites. Using a plasmid-based method, damaged DNA containing uracil positioned at 1–5 bp separations from 8-oxo-7,8-dihydroguanine on the complementary strand was transfected into wild-type Escherichia coli or into strains lacking the DNA glycosylases Fpg and MutY. Mutation frequencies were found to be significantly higher for clustered damage sites than for single lesions. The loss of MutY gave a large relative increase in mutation frequency and a strain lacking both Fpg and MutY showed even higher mutation frequencies, up to nearly 40% of rescued plasmid. In these strains, the mutation frequency decreases with increasing spacing of the uracil from the 8-oxo-7,8-dihydroguanine site. Sequencing of plasmid DNA carrying clustered damage, following rescue from bacteria, showed that almost all of the mutations are GCTA transversions. The data suggest that at clustered damage sites, depending on lesion spacing, the action of Fpg is compromised and post-replication processing of lesions by MutY is the most important mechanism for protection against mutagenesis.
INTRODUCTION
Ionizing radiation leads to the formation of various types of DNA damage in cells. Many of the DNA lesions caused are chemically indistinguishable from those produced endogenously by reactive oxygen species (ROS), including superoxide, hydrogen peroxide and hydroxyl radicals, which are produced during normal, aerobic metabolism and known to cause continual, low levels of damage to DNA (1). Estimates put the steady-state level of endogenous damage at about 2000 lesions/cell (2). Environmental levels of radiation produce only a few lesions against this plethora of endogenous damage, suggesting a minimal degree of impact upon the stability of the genome. However, the true damaging potential of environmental radiation can be revealed through a more qualitative view. Relative to endogenous processes, ionising radiation appears to be uniquely responsible for the production of clustered DNA damage sites, defined as two or more elemental lesions induced within 10–20 bp, or 1–2 turns of the DNA helix, by a single radiation track. Track structure simulations have led to the prediction that energy deposition by a single track in or near to cellular DNA produces both isolated DNA lesions and clustered DNA damage (3–5). It has been hypothesised from biological modelling that radiation-induced clustered DNA damage sites are less repairable (4) and therefore are particularly detrimental to the maintenance of genome stability, thereby distinguishing them from readily repairable endogenous damage. For instance, 40% of single-strand breaks (SSB) induced by sparsely ionizing radiation (-radiation) have been shown experimentally to have a base lesion/abasic site within 10 bp of the SSB (6), and this value is estimated to increase to 80% for densely ionizing -radiation (5). While DNA double-strand breaks (DSB) are commonly considered to be one of the most harmful effects of irradiation, they only constitute a minor proportion compared with the yield of clustered DNA damage sites. Therefore, environmental levels of radiation exposure are more likely to induce a clustered damage site rather than a DSB in irradiated cells.
Recent studies have verified that clustered DNA damage sites are induced in mammalian cells (7–10) and Escherichia coli (11) by ionizing radiation. Evidence is accumulating, using model approaches, in support of the hypothesis that the processing of lesions within the clustered damage site by base excision repair (BER) is indeed compromised relative to repair of individual lesions (12–20). However, it seems that there is a significant degree of variation in the way particular clustered damage sites are processed. For instance, two closely spaced, bi-stranded abasic (AP) sites have been shown to yield DSB or inhibit one anothers incision by HAP1 and exonuclease III depending upon their orientation to each other (13,19,20). The indications to date are that a bi-stranded cluster of damage containing a SSB/AP site within a few base pairs opposite to a 8-oxo-7,8-dihydroguanine (8-oxoG) is one of the main types of clusters that is difficult to repair (21,22), whereas a base lesion only inhibits excision of 8-oxoG when one base pair away on the opposite strand (12,14,15). As a consequence, there is an increased probability of this type of clustered damage containing 8-oxoG persisting until the DNA is replicated (M.Lomax and P.O’Neill, unpublished work). The excision of an 8-oxoG lesion by formamidopyrimidine glycosylase (Fpg, also known as MutM) is strongly inhibited by the presence of a SSB on the opposite strand with a 1 bp separation. However, increasing the inter-lesion gap to 3–5 nt allows Fpg catalysis to occur, albeit at a reduced efficiency relative to that of a single 8-oxoG (15), and may lead to the formation of DSB (14,17). The AP lyase activity of Fpg is independent of the presence of an 8-oxoG on the complementary strand (15). The effect of 8-oxoG on the efficiency of incision of an AP site within a cluster by endonuclease IV or exonuclease III has not been studied in detail.
BER is a key repair pathway involved in countering the deleterious effects of ionizing radiation. Important enzymes in this pathway that give protection against the mutagenic effect of one of the most common lesions, 8-oxoG, are the aforementioned Fpg and MutY glycosylases, in addition to MutT, which hydrolyses 8-oxoGTP in the nucleotide pool (1). Fpg recognises 8-oxoG in DNA and promotes its excision through cleavage of the N-glycosylic bond linking the base to the sugar–phosphate backbone. Should 8-oxoG persist through to replication, it has been found that DNA polymerases associated with DNA replication predominantly catalyse the incorporation of adenine opposite 8-oxoG; polymerases involved primarily in DNA repair tend to insert cytosine opposite 8-oxoG (23). The incorporation of adenine opposite 8-oxoG following replication is countered by the action of MutY, a mismatch glycosylase, which excises the adenine base allowing the correct base, cytosine, to replace it. An 8-oxoG lesion is weakly mutagenic in E.coli (24), but a higher frequency of G:CT:A transversions is commonly associated with E.coli strains deficient in Fpg and MutY (23). An fpg mutY double mutation acts synergistically with respect to G:CT:A mutations (25), demonstrating that Fpg and MutY combine to protect E.coli from the mutagenic effects of 8-oxoG (23).
Since clustered damage is a unique attribute of ionizing radiation, its impact on genetic instability and its relevance to radiation carcinogenesis necessitates significant attention. In this study we have addressed the issue of whether clustered damage has biological consequences and, specifically, whether the mutability of clustered damage is greater than that of individual lesions and whether it depends on their spatial separation. We also addressed the importance of defence against 8-oxoG by Fpg relative to post-replication repair by MutY for the mutability of clustered damage. We report a novel assay to test the mutability of specific bi-stranded clustered damage sites. The clustered damage sites were created within oligonucleotides and consist of 8-oxoG with AP lesions, formed by glycosylase action on uracil, on the opposite strand at 1–5 bp separation, with the 8-oxoG lesion positioned within a restriction site. Each oligonucleotide was ligated into a plasmid for transfer into wild-type or repair-deficient E.coli and following damage processing, mutations were determined by inability to cut the restriction site, as described in Materials and Methods. The types of mutations present were also determined by sequencing the recovered plasmid DNA. This study provides novel insights by showing that clustered DNA damage sites are highly mutagenic in E.coli, relative to isolated lesions, and that post-replicative repair involving MutY and not initial removal of 8-oxoG by Fpg is the major influence on mutagenesis.
MATERIALS AND METHODS
Escherichia coli strains
Isogenic strains CC104 (wild-type) (26), BH540 (fpg::KanR), BH980 (mutY::KanR) and BH990 (fpg::KanR mutY::KanR) double mutant were a kind gift from Dr S.Boiteux (CEA, France). The genotypes of these strains was verified by testing for spontaneous reversion frequency to lactose utilization (25).
Preparation of oligonucleotides
Oligonucleotides (40mer) carrying either a uracil lesion at different positions or an 8-oxoG lesion at a fixed position (Fig. 1) were synthesized by MWG Biotech and purified by HPLC. Aliquots of 20 pmol of each of the complementary oligonucleotides were annealed in Tris–EDTA buffer at pH 8 through heating at 80°C followed by cooling to room temperature over 2–3 h. The annealed oligonucleotides were phosphorylated at the 5' termini using 10 U T4 polynucleotide kinase in 25 mM ATP with 6x Forward exchange buffer (Invitrogen) at 37°C for 30 min. The oligonucleotides were retrieved by passage through a Qiagen nucleotide extraction column as per the manufacturer’s instructions.
Figure 1. Oligonucleotides (40mer) carrying either a uracil lesion at different positions or an 8-oxoG lesion at a fixed position. X, 8-oxoG; Y, Uracil. Position denotes the base pair separation of the uracil from the 8-oxoG on the complementary strand where 5' is – and 3' is +. The BsmAI restriction site is underlined.
Plasmid preparation and ligation
pUC18 plasmid DNA (10 μg) in Tris–EDTA was treated with 100 U SmaI (NEB) at room temperature for 3 h. Following electrophoresis, the DNA was retrieved then purified using a Qiagen gel purification system. The linearized plasmid in Tris–EDTA was incubated with 20 U calf intestine phosphatase (NEB) for 15 min at 37°C, followed by a second incubation at 55°C for 45 min. The plasmid DNA was again gel purified and the concentration determined spectrophotometrically. Aliquots of 200 fmol linearized pUC18 DNA were ligated to 5 pmol annealed oligonucleotide for 10 min at room temperature using a Quick Ligase and Quick ligation reaction buffer (NEB), followed by dialysis using 0.025 μm Millipore nitrocellulose filters. Under these conditions, 20–50% of plasmid molecules were rejoined (open circular form).
Transformation of electrocompetent E.coli
An aliquot of 50 ng of the ligation product was added to 60 μl of electrocompetent bacteria (10% v/v glycerol in distilled H2O) in an electroporation cuvette and electroporated using a Bio-Rad E.coli pulser, set at 1.8 mV with time constants of 5.0 ms. Immediately after electroporation, 500 μl of SOB was added to the transformed bacteria, which were incubated for 1 h at 37°C. Transformants were selected in 5 ml of LB broth containing ampicillin (100 μg/ml) at 37°C for 16 h. Transformation frequencies were very similar with all plasmid constructs (undamaged or damaged) and bacterial strains, showing that there was little or no loss of plasmid due to the damage present.
Quantification of mutation of the DNA damage site
Plasmid DNA was retrieved in 50 μl of Tris–EDTA from 1.5 ml of the overnight miniculture using a Qiakit-spin mini-prep kit. A 15 μl aliquot of the plasmid DNA eluate was added to 5 U of BsmAI in a total volume of 20 μl. Initial restriction took place over 3 h at 55°C, followed by incubation with a further 5 U of BsmAI at 55°C for 16 h. The samples were electrophoresed for 6 h at room temperature on a 1% agarose gel (1 μg/ml ethidium bromide) at 4 V/cm. Following electrophoresis, the gel image was captured under UV light using a CCD camera. Images were analysed using Bio-Rad Quantity One? Quantitation Software. A quantitative value for the mutation frequency of the clustered damage site (represented by the 1755 bp band, due to non-cutting by BsmAI) was obtained using the following equation:
mutation frequency (%) = x 1001
where I is total pixel intensity minus local background.
The significance of the data was determined by analysis of variance. Comparisons between individual experiments were made, with an allowance for multiple comparisons.
Verification of lesion conversion and restriction site cutting
Prior to bacterial transformation the oligonucleotides carry a uracil base opposed to 8-oxoG and the method relies on endogenous bacterial glycosylase activity to remove a uracil base rapidly, creating an abasic site. To check that the uracil lesion in a clustered damage site is removed successfully by the bacteria, the 8-oxoG/uracil oligonucleotide (with uracil at either position –1, –3 or –5) was treated with uracil DNA glycosylase (15) or untreated and ligated to the plasmid. Following bacterial transformation, the mutation frequency of the UDG-treated and untreated damage sites was assessed and found to be similar (Supplementary Material, Fig. S1).
The mutation assay is based on the inability to cut a BsmAI restriction site within the oligonucleotide (Fig. 1) following bacterial transformation and plasmid retrieval. To ensure the validity of the assay, we confirmed that the presence of 8-oxoG within the BsmAI restriction site does inhibit incision. 32P-5'-labelled oligonucleotides containing either no damage or a single 8-oxoG lesion (X in Fig. 1) in the BsmAI restriction site were treated with excess BsmAI. This treatment did not change the size of the band in the case of the 8-oxoG-bearing oligonucleotide, but the undamaged oligonucleotide showed almost complete cutting to yield a band of the appropriate size (Supplementary Material, Fig. S2).
Sequence analysis of plasmid DNA
After transformation (as above), the bacteria were incubated for 1 h at 37°C in 500 μl of SOB. A 5 μl aliquot was removed and plated onto LB agar containing ampicillin (100 μg/ml). The bacteria were incubated at 37°C overnight and 10 colonies were picked at random and used to inoculate 10 separate 5 ml aliquots of LB broth containing ampicillin (1 μg/ml) and grown at 37°C overnight. The plasmid DNA was retrieved from the bacteria as previously described and prepared for sequencing using an Applied Biosystems Big Dye sequencing kit using the following forward and reverse primers, 5'-CTTCGCTATTACGCCAGCTG-3' and 5'-GGCACGACA GGTTTCCCGACTGGA-3', respectively. These primers amplify sequence across the site of damage in the cloned oligonucleotide. The sequencing data were analysed using the online multiple sequence alignment tool CLUSTALW (http://www.ebi.ac.uk/clustalw/index.html).
RESULTS
The mutation assay described above is based on the inability to cut a BsmAI restriction site within the cloned oligonucleotide (Fig. 2), following bacterial transformation and plasmid retrieval. As noted in the Introduction, the position of a second lesion within a clustered damage site can influence the repair of that lesion. Therefore, we investigated whether bi-stranded damage constructs with uracil positioned at –5, –3 or –1 or at +1, +3 or +5 relative to 8-oxoG (Fig. 1) would change the mutation frequency relative to that of 8-oxoG alone. We have also used wild-type bacteria and strains mutated in specific base excision repair proteins to understand the role of pre- and post-replicative processes in mutagenesis by clustered damage.
Figure 2. Plasmid construct with BsmAI restriction sites and fragment lengths highlighted. (A) pUC18 and (B) pUC18 with cloned oligonucleotide containing BsmAI restriction site. Misrepair of the 8-oxoG lesion within the cloned BsmAI restriction site results in the ‘mutant’ fragment –1755 bp, whereas successful repair yields fragments of 1352 and 403 bp.
The mutagenic potential of clustered DNA damage in wild-type bacteria
Following transfection of plasmid constructs into wild-type E.coli, the band indicating mutation (1755 bp) was weak or absent for oligonucleotides containing no damage or a single 8-oxoG (Fig. 3A, lanes 4 and 5). The absence of this band indicates that repair had been successful since the BsmAI restriction site had been cleaved. However, bands are visible at the 1755 bp mark when the clustered damage constructs were used (Fig. 3A, lanes 1–3). From a comparison of the mutation frequencies determined using equation 1 and shown in Figure 4A, all the clustered damage constructs give an increase in the mutation frequency compared with that of 8-oxoG alone. There is little indication of a difference in mutation frequency as the interlesion distance increases, although the difference between the +1 position and the +5 position is significant at the 1% level.
Figure 3. Gel images of BsmAI restricted plasmid DNA containing the oligonucleotides in Figure 1, following processing by (A) wild-type and (B) double mutant (fpg mutY) strains. The specific damage within the plasmid constructs is indicated above each lane. The intensity of the band at the 1755 bp position indicates the extent of mutation.
Figure 4. Histograms showing the mutation frequency of the clustered damage site in plasmid DNA recovered from the (A) wild-type, (B) fpg, (C) mutY and (D) double mutant fpg mutY strains. The types of damage (see Fig. 1 for description) are shown along the x-axis. Error bars display the standard errors from three experiments.
The importance of mutY in determining mutation frequency
Using the plasmid system we separately tested the effect of the absence of the major DNA glycosylases which process 8-oxoG or adenine opposite 8-oxoG, Fpg or MutY, respectively, on the mutation frequency of clustered damage. In the fpg strain (Fig. 4B) the average mutation frequency does not rise above 7% for even the most compact clustered damage (8-oxoG/+1U) and overall the frequencies are similar to those in the wild-type strain (it should be noted that the fpg strain retained its spontaneous mutator phenotype; see Materials and Methods). However, there is clear evidence for a significant increase in the mutation frequency of clustered damage in the mutY strain relative to that in the wild-type strain. For instance, the mutation frequency for clustered damage with 1 bp separation is 10-fold greater than that for the single 8-oxoG lesion (Fig. 4C), emphasizing the importance of the MutY protein in combating the effects of the 8-oxoG lesion when in a clustered damage site. In Figure 4C it can be further seen that the large increase found for the mutY strain is present whichever orientation the uracil is in relative to that of the 8-oxoG lesion, but that the mutation frequency declines as the inter-lesion distance increases.
Deficiency in both Fpg and MutY glycosylases leads to high levels of mutation
Although Fpg seems to have a relatively small influence on mutation frequency, we wished to see if a strain lacking both Fpg and MutY glycosylases would have a simple additive or a synergistic effect on mutation frequency. As seen in Figure 3B, even for the 8-oxoG lesion alone (lane 4) a relatively intense band representing mutation is formed, although these bands are still more intense for the clustered damage (Fig. 3B, lanes 1–3). Again, there is a significant decrease in the relative mutation frequency with increasing separation of the uracil from the 8-oxoG lesion, although this effect is of lower magnitude for the – orientation than for the + orientation (Fig. 4D). The decrease in mutation frequency between the +1 and the +5 positions is significant at the 0.1% level, whereas the decrease in mutation frequency between the –1 and the –5 positions is only significant at the 5% level (Fig. 4D). The difference in mutation frequency between that for the single 8-oxoG lesion and the +5 position of clustered damage is not significant at 5%. However, the mutation frequencies for clustered damage seen with the double mutant strain are in all cases greater than those with the single mutant strains.
GCTA is the most common mutation arising from the inefficient repair of 8-oxoG
The mutation frequencies described above were derived from bulk plasmid isolates. To verify the mutagenic nature of the clustered damage, the plasmid construct containing the 8-oxoG/+1U clustered damage was used to transform the fpg mutY strain and individual colonies were picked and grown for DNA isolation and sequencing. In general, each colony will contain the progeny of a single plasmid molecule (27,28). Out of 98 plasmids sequenced from individual colonies, 60% were correctly repaired, 37% showed a GCTA transversion at the original 8-oxoG position within the BsmAI restriction site and 3% had single base deletions (two plasmids with loss of the guanine base at the site of the 8-oxoG lesion and one plasmid with loss of the thymine base at the site of the uracil lesion). It should be noted that the 40% mutation frequency found by sequencing 8-oxoG/+1-derived plasmids is in very good agreement with the estimates of mutation frequency found by gel band analysis.
DISCUSSION
The plasmid-based method we have devised for measurement of the mutagenic potential of clustered damage in DNA has yielded several important findings. Mutation frequencies are consistently higher for bi-stranded clustered damage sites than for single lesions with all E.coli strains studied. A very recent report (29) has also shown that clustered damage containing two opposed 8-oxoG sites (at +1 spacing only) is more effective in mutation induction than a single damage site. In that report (29), oligonucleotides carrying opposed lesions also acted as the substrate for mutation, but the method used to identify mutants was very inefficient: in wild-type bacteria, at best only 20% of selected ‘mutants’ were considered to be caused by the opposed lesions (sequencing had to be used to identify true mutants, which were deemed to be GT transversions, while all other mutations were considered oligonucleotide synthesis errors). While we always detected a >2-fold increase in mutation frequency for clustered damage at the +1 position, relative to a single 8-oxoG lesion, they failed to find an increase for one clustered damage site in wild-type cells (29). Similar to our study, they showed that clustered damage gave large relative increases in mutation frequency in MutY-deficient bacteria, although their relative increases (6–8-fold) were about half those that we found at the +1 position (Fig. 4). However, our data also go much further in defining the roles of both Fpg and MutY in processing clustered damage and in assessing the effects of interlesion distance on mutation frequency. As seen most clearly in the mutY and double mutant strains, the effect of spacing of the two lesions within the cluster on mutation is important (Fig. 4). There is commonly a decline in mutation frequency with an increase in spatial distance between lesions, and sometimes at a spacing of 5 bp no difference in the mutation frequency is seen between the clustered damage and the single 8-oxoG lesion. Interestingly, there is some evidence that the – orientation of clustered damage is less prone to reduced mutation frequency than the + orientation as the lesion spacing increases. Lastly, it is clear that absence of Fpg has less effect on mutation frequencies than the absence of MutY for clustered damage sites, although in the double mutant strain the absence of Fpg appears to have some impact even for a single 8-oxoG lesion.
Clustered DNA damage is more mutagenic than the single lesions
As noted in the Introduction from in vitro studies, a lesion(s) within clustered DNA damage is more difficult to excise than isolated lesions, due to the inhibition of BER enzymes (12,13,15–17,19,20,30,31). However, there is considerable variability in the levels of inhibition of the BER glycosylases/AP endonucleases depending on the types of lesion as well as the interlesion spacing in the bi-stranded cluster. This effect of interlesion distance will be influential only before plasmid replication, since the lesions will separate onto different DNA molecules following replication. The difficulty in excising lesions from within a cluster containing an AP site and a base lesion may arise in particular where the AP site is rapidly converted (Fig. 5, step 3) into a SSB, i.e. before excision of the base lesion (19,22) by AP endonucleases or through the AP lyase activity of Fpg. The latter activity is not affected by the presence of 8-oxoG (15). The specific activity for excision of 8-oxoG by Fpg (Fig. 5, steps 1 or 4) decreases as the distance to a closely spaced AP site/SSB decreases (17). However, if a DSB is formed in step 1, this would lead to loss of the plasmid and therefore not contribute to the mutations seen.
Figure 5. Diagram of mutagenic consequences of clustered DNA damage in different BER-deficient strains. The damaged site is shown at the top, followed by its progress through two replication cycles. At replication 1, it is proposed that the SSB-carrying strand is lost (see text), while the other, if unrepaired, will have mostly an adenine inserted opposite 8-oxoG. At the second replication, two outcomes of damage processing are shown, either with or without the action of MutY. The bold arrows indicate the main pathway of mutagenesis in an fpg mutY strain.
Additionally, repair of the SSB (Fig. 5, step 4) is retarded in the presence of 8-oxoG (14,22; M.Lomax and P.O’Neill, unpublished data). These reciprocal effects, depending on interlesion distance, may lead to persistence of the clustered damage site, allowing 8-oxoG to endure through to replication where its propensity to mispair with adenine will lead to an increased level of mutagenesis. Further, where the inhibitory effects of a SSB on 8-oxoG excision by Fpg have been assessed in vitro, the pattern of inhibition with distance of the SSB from the 8-oxoG lesion (17) is similar to that seen in the present mutation study. Firstly, this pattern shows a reduction with distance over 1–5 bp between the two lesions and, secondly, the reduction is most dramatic in the + orientation. These distinctions are also apparent in the results of studies on the effect of a SSB on the excision of 8-oxoG with eukaryotic BER enzymes and cell extracts (16,17).
MutY plays a more significant role than Fpg in countering the mutagenic effects of clustered DNA damage containing 8-oxoG
The results shown in Figure 4B and C indicate that the role of Fpg is significantly less important than that of MutY in processing clustered damage in E.coli. It is only with the clustered damaged constructs that this is notable, as neither single mutant strain appears to produce significant levels of mutation when dealing with either 8-oxoG or an AP site. This shows that the presence of either Fpg or MutY is sufficient to depress the potential mutability of the single 8-oxoG lesion. With the fpg strain, the levels of mutation arising from clustered damage sites are as low as those for the wild-type, whereas the corresponding data set for the mutY strain indicates a significant increase in levels of mutation. Figure 5 illustrates the potential mutagenic consequences of clustered DNA damage in different BER-deficient strains. Step 1 (Fig. 5) cannot occur in an fpg strain, so that steps 2 and 3 are the main pathways for clustered damage processing prior to replication. Following replication of the plasmid, when the 8-oxoG lesion is no longer associated with a cluster, the incorporation of adenine opposite 8-oxoG becomes potentially mutagenic and can only be countered by MutY. The efficiency of step 3 (Fig. 5) may be reduced in a fpg mutant strain relative to that in the wild-type since the AP site cannot be incised by Fpg, but the effect of 8-oxoG on the incision of an AP site by AP endonucleases is as yet not known. If a SSB is formed, its repair is retarded in wild-type cells, as discussed above, although retardation does not occur in a BER reconstructed system in the absence of Fpg protein (14). Therefore, the lower mutation frequency for clustered damage seen in the fpg mutant strain relative to the wild-type may reflect the relative importance of steps 2 and 3, in part through absence of the AP lyase activity of Fpg, but also through increased efficiency of repair of SSB in step 4 (Fig. 5). However, it is also possible that there is back-up capacity for oxidized guanine processing in E.coli; for example endonuclease VIII has weak 8-oxoG glycosylase activity and its absence in an fpg mutY strain increases the mutation frequency by a factor of 2–3 (32). It is interesting that absence of the mammalian analogue of Fpg, OGG1, in mice does not increase the incidence of tumours (33) and, unlike the ablation of subsequent steps in BER, is not embryonically lethal (34).
High mutation frequencies in the fpg mutY double mutant strain
It is evident that, following the processing of clustered damage sites, the mutation frequencies in the fpg mutY double mutant strain are higher than those in the mutY strain alone. Strikingly, only in the double mutant does the single 8-oxoG lesion give rise to a significant level of mutation (16%). Therefore, it appears that the presence of Fpg is important when processing a single 8-oxoG lesion, but only in the absence of MutY. While our data suggest that Fpg has less impact on the processing of clustered damage sites prior to replication of the plasmid DNA, it will be able to lower the probability of mutation by acting on 8-oxoG after replication when it pairs correctly with cytosine . If most of the 8-oxoG lesions are paired with adenine following replication, then MutY will be most important in protecting against mutation whether Fpg is present or not. In Figure 5 we show the anticipated fate of clustered damage, with emphasis on the formation of mutations in the fpg mutY strain (thick arrows). Thus, before the first plasmid DNA replication there will be some opportunity for repair of the damage through back-up repair, yielding a fraction of non-mutant plasmids. Once replication occurs, however, there will be little further chance of non-mutant plasmid forming in the fpg mutY strain since the replicative polymerase will mostly place an adenine opposite 8-oxoG (23), leading to mutation at the next or a subsequent replication.
GCTA is the major mutation arising from the presence of 8-oxoG
The mutation system we have devised shows remarkably good agreement of mutation frequency estimates from gel image analysis with those derived from plasmid DNA sequencing . The almost exclusive occurrence of transversion mutations in the sequence data reinforces the above suggestion that a clustered damage site is not processed at random, but sequentially in order to minimize the formation of an adjacent SSB by excision of 8-oxoG leading to more severe damage. If this processing led to a DSB, for example, then it would be anticipated that deletions would occur rather than base pair substitutions at the site of damage (36), as shown recently for closely opposed uracil lesions (37). However, we should note that to date we have sequenced mutations derived from clustered damage at the +1 position in the fpg mutY strain and it is possible that the mutation spectrum will differ at other positions or in different mutant backgrounds. Interestingly, all of the mutations detected with the fpg mutY strain are ‘complete’, i.e. from the sequence data there is no evidence that both an unmutated and a mutated DNA sequence coexist in a single bacterial cell. It might be expected that a mixed sequence would occur since the 8-oxoG lesion exists on one strand only and the other strand will have a cytosine residue which will pair normally at replication. The phenomenon of complete mutations induced by radiation and other agents has been observed in earlier mutagenesis studies with several different microorganisms, leading to a variety of interpretations (38–42). In the present study, a possible explanation is that when plasmids carrying a SSB adjacent to 8-oxoG are replicated, the SSB-carrying strand becomes unstable and is lost (Fig. 5). In plasmid DNA replication, one strand is displaced into a so-called R-loop (43), which may promote abortive replication when the strand contains a SSB.
In summary, clustered DNA damage sites containing only two lesions have greater mutagenic potential than those of the component lesions. Even these simple types of clustered DNA damage, which represent 60% of the overall number of clustered sites produced by -radiation (44) are highly mutagenic. Clustered DNA damage sites are produced by radiation (7–10) in a higher yield than that of DSB, so that at environmentally low doses of radiation, the probability of forming a DSB is less than that of a clustered DNA damage site in a given cell. This study therefore provides evidence that clustered DNA damage sites have a greater potential to contribute to the deleterious health effects of radiation than single lesions, particularly at low doses.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
We are very grateful to David Papworth for statistical advice, the Medical Research Council for a studentship to Colin Pearson, the Japan Atomic Energy Research Institute for a study grant to Naoya Shikazono and the European Commission (contract FIGH-CT2002-00207) for partial financial support.
REFERENCES
Friedberg,C.E., Walker,G.C. and Siede,W. (1995) DNA Repair and Mutagenesis. ASM Press, Washington, DC.
Pouget,J.P., Douki,T., Richard,M.J. and Cadet,J. (2000) DNA damage induced in cells by gamma and UVA radiation as measured by HPLC/GC-MS and HPLC-EC and Comet assay. Chem. Res. Toxicol., 13, 541–549.
Ward,J.F. (1994) The complexity of DNA damage: relevance to biological consequences. Int. J. Radiat. Biol., 66, 427–432.
Goodhead,D.T. (1994) Initial events in the cellular effects of ionizing radiations: clustered damage in DNA. Int. J. Radiat. Biol., 65, 7–17.
Nikjoo,H., O’Neill,P., Terrissol,M. and Goodhead,D.T. (1999) Quantitative modelling of DNA damage using Monte Carlo track structure method. Radiat. Environ. Biophys., 38, 31–38.
Cunniffe,S. and O’Neill,P. (1999) The complexity of radiation-induced DNA damage as revealed by exposure to cell extracts. Radiat. Res., 152, 421–427.
Sutherland,B.M., Bennett,P.V., Sidorkina,O. and Laval,J. (2000) Clustered DNA damages induced in isolated DNA and in human cells by low doses of ionizing radiation. Proc. Natl Acad. Sci. USA, 97, 103–108.
Jenner,T.J., Fulford,J. and O’Neill,P. (2001) Contribution of base lesions to radiation-induced clustered DNA damage: implication for models of radiation response. Radiat. Res., 156, 590–593.
Gulston,M., Fulford,J., Jenner,T., de Lara,C. and O’Neill,P. (2002) Clustered DNA damage induced by gamma radiation in human fibroblasts (HF19), hamster (V79-4) cells and plasmid DNA is revealed as Fpg and Nth sensitive sites. Nucleic Acids Res., 30, 3464–3472.
Sutherland,B.M., Bennett,P.V., Sutherland,J.C. and Laval,J. (2002) Clustered DNA damages induced by x rays in human cells. Radiat. Res., 157, 611–616.
Blaisdell,J.O. and Wallace,S.S. (2001) Abortive base-excision repair of radiation-induced clustered DNA lesions in Escherichia coli. Proc. Natl Acad. Sci. USA, 98, 7426–7430.
Chaudhry,M.A. and Weinfeld,M. (1995) The action of Escherichia coli endonuclease III on multiply damaged sites in DNA. J. Mol. Biol., 249, 914–922.
Chaudhry,M.A. and Weinfeld,M. (1997) Reactivity of human apurinic/apyrimidinic endonuclease and Escherichia coli exonuclease III with bistranded abasic sites in DNA. J. Biol. Chem., 272, 15650–15655.
Harrison,L., Hatahet,Z. and Wallace,S.S. (1999) In vitro repair of synthetic ionizing radiation-induced multiply damaged DNA sites. J. Mol. Biol., 290, 667–684.
David-Cordonnier,M.H., Laval,J. and O’Neill,P. (2000) Clustered DNA damage, influence on damage excision by XRS5 nuclear extracts and Escherichia coli Nth and Fpg proteins. J. Biol. Chem., 275, 11865–11873.
David-Cordonnier,M.H., Boiteux,S. and O’Neill,P. (2001) Efficiency of excision of 8-oxo-guanine within DNA clustered damage by XRS5 nuclear extracts and purified human OGG1 protein. Biochemistry, 40, 11811–11818.
David-Cordonnier,M.H., Laval,J. and O’Neill,P. (2001) Recognition and kinetics for excision of a base lesion within clustered DNA damage by the Escherichia coli proteins Fpg and Nth. Biochemistry, 40, 5738–5746.
Budworth,H., Dianova,I.I., Podust,V.N. and Dianov,G.L. (2002) Repair of clustered DNA lesions. Sequence-specific inhibition of long-patch base excision repair by 8-oxoguanine. J. Biol. Chem., 277, 21300–21305.
David-Cordonnier,M.H., Cunniffe,S.M., Hickson,I.D. and O’Neill,P. (2002) Efficiency of incision of an AP site within clustered DNA damage by the major human AP endonuclease. Biochemistry, 41, 634–642.
Tian,K., McTigue,M. and de los Santos,C. (2002) Sorting the consequences of ionizing radiation: processing of 8-oxoguanine/abasic site lesions. DNA Repair (Amst.), 1, 1039–1049.
Weinfeld,M., Rasouli-Nia,A., Chaudhry,M.A. and Britten,R.A. (2001) Response of base excision repair enzymes to complex DNA lesions. Radiat. Res., 156, 584–589.
Lomax,M.E., Gulston,M.K. and O’Neill,P. (2002) Chemical aspects of clustered DNA damage induction by ionising radiation. Radiat. Prot. Dosimetry, 99, 63–68.
Grollman,A.P. and Moriya,M. (1993) Mutagenesis by 8-oxoguanine: an enemy within. Trends Genet., 9, 246–249.
Moriya,M., Ou,C., Bodepudi,V., Johnson,F., Takeshita,M. and Grollman,A.P. (1991) Site-specific mutagenesis using a gapped duplex vector: a study of translesion synthesis past 8-oxodeoxyguanosine in E. coli. Mutat. Res., 254, 281–288.
Michaels,M.L., Cruz,C., Grollman,A.P. and Miller,J.H. (1992) Evidence that MutY and MutM combine to prevent mutations by an oxidatively damaged form of guanine in DNA. Proc. Natl Acad. Sci. USA, 89, 7022–7025.
Cupples,C.G. and Miller,J.H. (1989) A set of lacZ mutations in Escherichia coli that allow rapid detection of each of the six base substitutions. Proc. Natl Acad. Sci. USA, 86, 5345–5349.
Weston,A., Brown,M.G., Perkins,H.R., Saunders,J.R. and Humphreys,G.O. (1981) Transformation of Escherichia coli with plasmid deoxyribonucleic acid: calcium-induced binding of deoxyribonucleic acid to whole cells and to isolated membrane fractions. J. Bacteriol., 145, 780–787.
Hanahan,D. (1983) Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol., 166, 557–580.
Malyarchuk,S., Youngblood,R., Landry,A.M., Quillin,E. and Harrison,L. (2003) The mutation frequency of 8-oxo-7,8-dihydroguanine (8-oxodG) situated in a multiply damaged site: comparison of a single and two closely opposed 8-oxodG in Escherichia coli. DNA Repair (Amst.), 2, 695–705.
Harrison,L., Hatahet,Z., Purmal,A.A. and Wallace,S.S. (1998) Multiply damaged sites in DNA: interactions with Escherichia coli endonucleases III and VIII. Nucleic Acids Res., 26, 932–941.
David-Cordonnier,M.H., Boiteux,S. and O’Neill,P. (2001) Excision of 8-oxoguanine within clustered damage by the yeast OGG1 protein. Nucleic Acids Res., 29, 1107–1113.
Blaisdell,J.O., Hatahet,Z. and Wallace,S.S. (1999) A novel role for Escherichia coli endonuclease VIII in prevention of spontaneous GT transversions. J. Bacteriol., 181, 6396–6402.
Klungland,A., Rosewell,I., Hollenbach,S., Larsen,E., Daly,G., Epe,B., Seeberg,E., Lindahl,T. and Barnes,D.E. (1999) Accumulation of premutagenic DNA lesions in mice defective in removal of oxidative base damage. Proc. Natl Acad. Sci. USA, 96, 13300–13305.
Wilson,S.H. and Kunkel,T.A. (2000) Passing the baton in base excision repair. Nature Struct. Biol., 7, 176–178.
Tchou,J., Bodepudi,V., Shibutani,S., Antoshechkin,I., Miller,J., Grollman,A.P. and Johnson,F. (1994) Substrate specificity of Fpg protein. Recognition and cleavage of oxidatively damaged DNA. J. Biol. Chem., 269, 15318–15324.
Dianov,G.L., O’Neill,P. and Goodhead,D.T. (2001) Securing genome stability by orchestrating DNA repair: removal of radiation-induced clustered lesions in DNA. Bioessays, 23, 745–749.
D’Souza,D,I. and Harrison,L. (2003) Repair of clustered uracil DNA damages in Escherichia coli. Nucleic Acids Res., 31, 4573–4581.
Holliday,R. (1962) Mutation and replication in Ustilago maydis. Genet. Res., 3, 472–486.
Kubitschek,H.E. (1964) Mutation without segregation. Proc. Natl Acad. Sci. USA, 52, 1374–1381.
Munson,R.J. and Bridges,B.A. (1964) Segregation of radiation-induced mutations in Escherichia coli. Nature, 203, 270–273.
Witkin,E.M. and Sicurella,C. (1964) Pure clones of lactose-negative mutants obtained in Escherichia coli after treatment with 5-bromouracil. J. Mol. Biol., 8, 610–613.
Nasim,A. and Auerbach,C. (1967) The origin of complete and mosaic mutants from mutagenic treatment of single cells. Mutat. Res., 4, 1–14.
Kuzminov,A. (1999) Recombination repair of DNA damage in Escherichia coli and bacteriophage . Microbiol. Mol. Biol. Rev., 63, 751–813.
Nikjoo,H., O’Neill,P., Wilson,W.E. and Goodhead,D.T. (2001) Computational approach for determining the spectrum of DNA damage induced by ionizing radiation. Radiat. Res., 156, 577–583.(Colin G. Pearson, Naoya Shikazono1, John)
*To whom correspondence should be addressed. Tel: +44 1235 841017; Fax: +44 1235 841200; Email: p.oneill@har.mrc.ac.uk
ABSTRACT
The formation of clustered DNA damage sites is a unique feature of ionizing radiation. Recent studies have shown that the repair of lesions within clusters may be compromised, but little is understood about the mutagenic consequences of such damage sites. Using a plasmid-based method, damaged DNA containing uracil positioned at 1–5 bp separations from 8-oxo-7,8-dihydroguanine on the complementary strand was transfected into wild-type Escherichia coli or into strains lacking the DNA glycosylases Fpg and MutY. Mutation frequencies were found to be significantly higher for clustered damage sites than for single lesions. The loss of MutY gave a large relative increase in mutation frequency and a strain lacking both Fpg and MutY showed even higher mutation frequencies, up to nearly 40% of rescued plasmid. In these strains, the mutation frequency decreases with increasing spacing of the uracil from the 8-oxo-7,8-dihydroguanine site. Sequencing of plasmid DNA carrying clustered damage, following rescue from bacteria, showed that almost all of the mutations are GCTA transversions. The data suggest that at clustered damage sites, depending on lesion spacing, the action of Fpg is compromised and post-replication processing of lesions by MutY is the most important mechanism for protection against mutagenesis.
INTRODUCTION
Ionizing radiation leads to the formation of various types of DNA damage in cells. Many of the DNA lesions caused are chemically indistinguishable from those produced endogenously by reactive oxygen species (ROS), including superoxide, hydrogen peroxide and hydroxyl radicals, which are produced during normal, aerobic metabolism and known to cause continual, low levels of damage to DNA (1). Estimates put the steady-state level of endogenous damage at about 2000 lesions/cell (2). Environmental levels of radiation produce only a few lesions against this plethora of endogenous damage, suggesting a minimal degree of impact upon the stability of the genome. However, the true damaging potential of environmental radiation can be revealed through a more qualitative view. Relative to endogenous processes, ionising radiation appears to be uniquely responsible for the production of clustered DNA damage sites, defined as two or more elemental lesions induced within 10–20 bp, or 1–2 turns of the DNA helix, by a single radiation track. Track structure simulations have led to the prediction that energy deposition by a single track in or near to cellular DNA produces both isolated DNA lesions and clustered DNA damage (3–5). It has been hypothesised from biological modelling that radiation-induced clustered DNA damage sites are less repairable (4) and therefore are particularly detrimental to the maintenance of genome stability, thereby distinguishing them from readily repairable endogenous damage. For instance, 40% of single-strand breaks (SSB) induced by sparsely ionizing radiation (-radiation) have been shown experimentally to have a base lesion/abasic site within 10 bp of the SSB (6), and this value is estimated to increase to 80% for densely ionizing -radiation (5). While DNA double-strand breaks (DSB) are commonly considered to be one of the most harmful effects of irradiation, they only constitute a minor proportion compared with the yield of clustered DNA damage sites. Therefore, environmental levels of radiation exposure are more likely to induce a clustered damage site rather than a DSB in irradiated cells.
Recent studies have verified that clustered DNA damage sites are induced in mammalian cells (7–10) and Escherichia coli (11) by ionizing radiation. Evidence is accumulating, using model approaches, in support of the hypothesis that the processing of lesions within the clustered damage site by base excision repair (BER) is indeed compromised relative to repair of individual lesions (12–20). However, it seems that there is a significant degree of variation in the way particular clustered damage sites are processed. For instance, two closely spaced, bi-stranded abasic (AP) sites have been shown to yield DSB or inhibit one anothers incision by HAP1 and exonuclease III depending upon their orientation to each other (13,19,20). The indications to date are that a bi-stranded cluster of damage containing a SSB/AP site within a few base pairs opposite to a 8-oxo-7,8-dihydroguanine (8-oxoG) is one of the main types of clusters that is difficult to repair (21,22), whereas a base lesion only inhibits excision of 8-oxoG when one base pair away on the opposite strand (12,14,15). As a consequence, there is an increased probability of this type of clustered damage containing 8-oxoG persisting until the DNA is replicated (M.Lomax and P.O’Neill, unpublished work). The excision of an 8-oxoG lesion by formamidopyrimidine glycosylase (Fpg, also known as MutM) is strongly inhibited by the presence of a SSB on the opposite strand with a 1 bp separation. However, increasing the inter-lesion gap to 3–5 nt allows Fpg catalysis to occur, albeit at a reduced efficiency relative to that of a single 8-oxoG (15), and may lead to the formation of DSB (14,17). The AP lyase activity of Fpg is independent of the presence of an 8-oxoG on the complementary strand (15). The effect of 8-oxoG on the efficiency of incision of an AP site within a cluster by endonuclease IV or exonuclease III has not been studied in detail.
BER is a key repair pathway involved in countering the deleterious effects of ionizing radiation. Important enzymes in this pathway that give protection against the mutagenic effect of one of the most common lesions, 8-oxoG, are the aforementioned Fpg and MutY glycosylases, in addition to MutT, which hydrolyses 8-oxoGTP in the nucleotide pool (1). Fpg recognises 8-oxoG in DNA and promotes its excision through cleavage of the N-glycosylic bond linking the base to the sugar–phosphate backbone. Should 8-oxoG persist through to replication, it has been found that DNA polymerases associated with DNA replication predominantly catalyse the incorporation of adenine opposite 8-oxoG; polymerases involved primarily in DNA repair tend to insert cytosine opposite 8-oxoG (23). The incorporation of adenine opposite 8-oxoG following replication is countered by the action of MutY, a mismatch glycosylase, which excises the adenine base allowing the correct base, cytosine, to replace it. An 8-oxoG lesion is weakly mutagenic in E.coli (24), but a higher frequency of G:CT:A transversions is commonly associated with E.coli strains deficient in Fpg and MutY (23). An fpg mutY double mutation acts synergistically with respect to G:CT:A mutations (25), demonstrating that Fpg and MutY combine to protect E.coli from the mutagenic effects of 8-oxoG (23).
Since clustered damage is a unique attribute of ionizing radiation, its impact on genetic instability and its relevance to radiation carcinogenesis necessitates significant attention. In this study we have addressed the issue of whether clustered damage has biological consequences and, specifically, whether the mutability of clustered damage is greater than that of individual lesions and whether it depends on their spatial separation. We also addressed the importance of defence against 8-oxoG by Fpg relative to post-replication repair by MutY for the mutability of clustered damage. We report a novel assay to test the mutability of specific bi-stranded clustered damage sites. The clustered damage sites were created within oligonucleotides and consist of 8-oxoG with AP lesions, formed by glycosylase action on uracil, on the opposite strand at 1–5 bp separation, with the 8-oxoG lesion positioned within a restriction site. Each oligonucleotide was ligated into a plasmid for transfer into wild-type or repair-deficient E.coli and following damage processing, mutations were determined by inability to cut the restriction site, as described in Materials and Methods. The types of mutations present were also determined by sequencing the recovered plasmid DNA. This study provides novel insights by showing that clustered DNA damage sites are highly mutagenic in E.coli, relative to isolated lesions, and that post-replicative repair involving MutY and not initial removal of 8-oxoG by Fpg is the major influence on mutagenesis.
MATERIALS AND METHODS
Escherichia coli strains
Isogenic strains CC104 (wild-type) (26), BH540 (fpg::KanR), BH980 (mutY::KanR) and BH990 (fpg::KanR mutY::KanR) double mutant were a kind gift from Dr S.Boiteux (CEA, France). The genotypes of these strains was verified by testing for spontaneous reversion frequency to lactose utilization (25).
Preparation of oligonucleotides
Oligonucleotides (40mer) carrying either a uracil lesion at different positions or an 8-oxoG lesion at a fixed position (Fig. 1) were synthesized by MWG Biotech and purified by HPLC. Aliquots of 20 pmol of each of the complementary oligonucleotides were annealed in Tris–EDTA buffer at pH 8 through heating at 80°C followed by cooling to room temperature over 2–3 h. The annealed oligonucleotides were phosphorylated at the 5' termini using 10 U T4 polynucleotide kinase in 25 mM ATP with 6x Forward exchange buffer (Invitrogen) at 37°C for 30 min. The oligonucleotides were retrieved by passage through a Qiagen nucleotide extraction column as per the manufacturer’s instructions.
Figure 1. Oligonucleotides (40mer) carrying either a uracil lesion at different positions or an 8-oxoG lesion at a fixed position. X, 8-oxoG; Y, Uracil. Position denotes the base pair separation of the uracil from the 8-oxoG on the complementary strand where 5' is – and 3' is +. The BsmAI restriction site is underlined.
Plasmid preparation and ligation
pUC18 plasmid DNA (10 μg) in Tris–EDTA was treated with 100 U SmaI (NEB) at room temperature for 3 h. Following electrophoresis, the DNA was retrieved then purified using a Qiagen gel purification system. The linearized plasmid in Tris–EDTA was incubated with 20 U calf intestine phosphatase (NEB) for 15 min at 37°C, followed by a second incubation at 55°C for 45 min. The plasmid DNA was again gel purified and the concentration determined spectrophotometrically. Aliquots of 200 fmol linearized pUC18 DNA were ligated to 5 pmol annealed oligonucleotide for 10 min at room temperature using a Quick Ligase and Quick ligation reaction buffer (NEB), followed by dialysis using 0.025 μm Millipore nitrocellulose filters. Under these conditions, 20–50% of plasmid molecules were rejoined (open circular form).
Transformation of electrocompetent E.coli
An aliquot of 50 ng of the ligation product was added to 60 μl of electrocompetent bacteria (10% v/v glycerol in distilled H2O) in an electroporation cuvette and electroporated using a Bio-Rad E.coli pulser, set at 1.8 mV with time constants of 5.0 ms. Immediately after electroporation, 500 μl of SOB was added to the transformed bacteria, which were incubated for 1 h at 37°C. Transformants were selected in 5 ml of LB broth containing ampicillin (100 μg/ml) at 37°C for 16 h. Transformation frequencies were very similar with all plasmid constructs (undamaged or damaged) and bacterial strains, showing that there was little or no loss of plasmid due to the damage present.
Quantification of mutation of the DNA damage site
Plasmid DNA was retrieved in 50 μl of Tris–EDTA from 1.5 ml of the overnight miniculture using a Qiakit-spin mini-prep kit. A 15 μl aliquot of the plasmid DNA eluate was added to 5 U of BsmAI in a total volume of 20 μl. Initial restriction took place over 3 h at 55°C, followed by incubation with a further 5 U of BsmAI at 55°C for 16 h. The samples were electrophoresed for 6 h at room temperature on a 1% agarose gel (1 μg/ml ethidium bromide) at 4 V/cm. Following electrophoresis, the gel image was captured under UV light using a CCD camera. Images were analysed using Bio-Rad Quantity One? Quantitation Software. A quantitative value for the mutation frequency of the clustered damage site (represented by the 1755 bp band, due to non-cutting by BsmAI) was obtained using the following equation:
mutation frequency (%) = x 1001
where I is total pixel intensity minus local background.
The significance of the data was determined by analysis of variance. Comparisons between individual experiments were made, with an allowance for multiple comparisons.
Verification of lesion conversion and restriction site cutting
Prior to bacterial transformation the oligonucleotides carry a uracil base opposed to 8-oxoG and the method relies on endogenous bacterial glycosylase activity to remove a uracil base rapidly, creating an abasic site. To check that the uracil lesion in a clustered damage site is removed successfully by the bacteria, the 8-oxoG/uracil oligonucleotide (with uracil at either position –1, –3 or –5) was treated with uracil DNA glycosylase (15) or untreated and ligated to the plasmid. Following bacterial transformation, the mutation frequency of the UDG-treated and untreated damage sites was assessed and found to be similar (Supplementary Material, Fig. S1).
The mutation assay is based on the inability to cut a BsmAI restriction site within the oligonucleotide (Fig. 1) following bacterial transformation and plasmid retrieval. To ensure the validity of the assay, we confirmed that the presence of 8-oxoG within the BsmAI restriction site does inhibit incision. 32P-5'-labelled oligonucleotides containing either no damage or a single 8-oxoG lesion (X in Fig. 1) in the BsmAI restriction site were treated with excess BsmAI. This treatment did not change the size of the band in the case of the 8-oxoG-bearing oligonucleotide, but the undamaged oligonucleotide showed almost complete cutting to yield a band of the appropriate size (Supplementary Material, Fig. S2).
Sequence analysis of plasmid DNA
After transformation (as above), the bacteria were incubated for 1 h at 37°C in 500 μl of SOB. A 5 μl aliquot was removed and plated onto LB agar containing ampicillin (100 μg/ml). The bacteria were incubated at 37°C overnight and 10 colonies were picked at random and used to inoculate 10 separate 5 ml aliquots of LB broth containing ampicillin (1 μg/ml) and grown at 37°C overnight. The plasmid DNA was retrieved from the bacteria as previously described and prepared for sequencing using an Applied Biosystems Big Dye sequencing kit using the following forward and reverse primers, 5'-CTTCGCTATTACGCCAGCTG-3' and 5'-GGCACGACA GGTTTCCCGACTGGA-3', respectively. These primers amplify sequence across the site of damage in the cloned oligonucleotide. The sequencing data were analysed using the online multiple sequence alignment tool CLUSTALW (http://www.ebi.ac.uk/clustalw/index.html).
RESULTS
The mutation assay described above is based on the inability to cut a BsmAI restriction site within the cloned oligonucleotide (Fig. 2), following bacterial transformation and plasmid retrieval. As noted in the Introduction, the position of a second lesion within a clustered damage site can influence the repair of that lesion. Therefore, we investigated whether bi-stranded damage constructs with uracil positioned at –5, –3 or –1 or at +1, +3 or +5 relative to 8-oxoG (Fig. 1) would change the mutation frequency relative to that of 8-oxoG alone. We have also used wild-type bacteria and strains mutated in specific base excision repair proteins to understand the role of pre- and post-replicative processes in mutagenesis by clustered damage.
Figure 2. Plasmid construct with BsmAI restriction sites and fragment lengths highlighted. (A) pUC18 and (B) pUC18 with cloned oligonucleotide containing BsmAI restriction site. Misrepair of the 8-oxoG lesion within the cloned BsmAI restriction site results in the ‘mutant’ fragment –1755 bp, whereas successful repair yields fragments of 1352 and 403 bp.
The mutagenic potential of clustered DNA damage in wild-type bacteria
Following transfection of plasmid constructs into wild-type E.coli, the band indicating mutation (1755 bp) was weak or absent for oligonucleotides containing no damage or a single 8-oxoG (Fig. 3A, lanes 4 and 5). The absence of this band indicates that repair had been successful since the BsmAI restriction site had been cleaved. However, bands are visible at the 1755 bp mark when the clustered damage constructs were used (Fig. 3A, lanes 1–3). From a comparison of the mutation frequencies determined using equation 1 and shown in Figure 4A, all the clustered damage constructs give an increase in the mutation frequency compared with that of 8-oxoG alone. There is little indication of a difference in mutation frequency as the interlesion distance increases, although the difference between the +1 position and the +5 position is significant at the 1% level.
Figure 3. Gel images of BsmAI restricted plasmid DNA containing the oligonucleotides in Figure 1, following processing by (A) wild-type and (B) double mutant (fpg mutY) strains. The specific damage within the plasmid constructs is indicated above each lane. The intensity of the band at the 1755 bp position indicates the extent of mutation.
Figure 4. Histograms showing the mutation frequency of the clustered damage site in plasmid DNA recovered from the (A) wild-type, (B) fpg, (C) mutY and (D) double mutant fpg mutY strains. The types of damage (see Fig. 1 for description) are shown along the x-axis. Error bars display the standard errors from three experiments.
The importance of mutY in determining mutation frequency
Using the plasmid system we separately tested the effect of the absence of the major DNA glycosylases which process 8-oxoG or adenine opposite 8-oxoG, Fpg or MutY, respectively, on the mutation frequency of clustered damage. In the fpg strain (Fig. 4B) the average mutation frequency does not rise above 7% for even the most compact clustered damage (8-oxoG/+1U) and overall the frequencies are similar to those in the wild-type strain (it should be noted that the fpg strain retained its spontaneous mutator phenotype; see Materials and Methods). However, there is clear evidence for a significant increase in the mutation frequency of clustered damage in the mutY strain relative to that in the wild-type strain. For instance, the mutation frequency for clustered damage with 1 bp separation is 10-fold greater than that for the single 8-oxoG lesion (Fig. 4C), emphasizing the importance of the MutY protein in combating the effects of the 8-oxoG lesion when in a clustered damage site. In Figure 4C it can be further seen that the large increase found for the mutY strain is present whichever orientation the uracil is in relative to that of the 8-oxoG lesion, but that the mutation frequency declines as the inter-lesion distance increases.
Deficiency in both Fpg and MutY glycosylases leads to high levels of mutation
Although Fpg seems to have a relatively small influence on mutation frequency, we wished to see if a strain lacking both Fpg and MutY glycosylases would have a simple additive or a synergistic effect on mutation frequency. As seen in Figure 3B, even for the 8-oxoG lesion alone (lane 4) a relatively intense band representing mutation is formed, although these bands are still more intense for the clustered damage (Fig. 3B, lanes 1–3). Again, there is a significant decrease in the relative mutation frequency with increasing separation of the uracil from the 8-oxoG lesion, although this effect is of lower magnitude for the – orientation than for the + orientation (Fig. 4D). The decrease in mutation frequency between the +1 and the +5 positions is significant at the 0.1% level, whereas the decrease in mutation frequency between the –1 and the –5 positions is only significant at the 5% level (Fig. 4D). The difference in mutation frequency between that for the single 8-oxoG lesion and the +5 position of clustered damage is not significant at 5%. However, the mutation frequencies for clustered damage seen with the double mutant strain are in all cases greater than those with the single mutant strains.
GCTA is the most common mutation arising from the inefficient repair of 8-oxoG
The mutation frequencies described above were derived from bulk plasmid isolates. To verify the mutagenic nature of the clustered damage, the plasmid construct containing the 8-oxoG/+1U clustered damage was used to transform the fpg mutY strain and individual colonies were picked and grown for DNA isolation and sequencing. In general, each colony will contain the progeny of a single plasmid molecule (27,28). Out of 98 plasmids sequenced from individual colonies, 60% were correctly repaired, 37% showed a GCTA transversion at the original 8-oxoG position within the BsmAI restriction site and 3% had single base deletions (two plasmids with loss of the guanine base at the site of the 8-oxoG lesion and one plasmid with loss of the thymine base at the site of the uracil lesion). It should be noted that the 40% mutation frequency found by sequencing 8-oxoG/+1-derived plasmids is in very good agreement with the estimates of mutation frequency found by gel band analysis.
DISCUSSION
The plasmid-based method we have devised for measurement of the mutagenic potential of clustered damage in DNA has yielded several important findings. Mutation frequencies are consistently higher for bi-stranded clustered damage sites than for single lesions with all E.coli strains studied. A very recent report (29) has also shown that clustered damage containing two opposed 8-oxoG sites (at +1 spacing only) is more effective in mutation induction than a single damage site. In that report (29), oligonucleotides carrying opposed lesions also acted as the substrate for mutation, but the method used to identify mutants was very inefficient: in wild-type bacteria, at best only 20% of selected ‘mutants’ were considered to be caused by the opposed lesions (sequencing had to be used to identify true mutants, which were deemed to be GT transversions, while all other mutations were considered oligonucleotide synthesis errors). While we always detected a >2-fold increase in mutation frequency for clustered damage at the +1 position, relative to a single 8-oxoG lesion, they failed to find an increase for one clustered damage site in wild-type cells (29). Similar to our study, they showed that clustered damage gave large relative increases in mutation frequency in MutY-deficient bacteria, although their relative increases (6–8-fold) were about half those that we found at the +1 position (Fig. 4). However, our data also go much further in defining the roles of both Fpg and MutY in processing clustered damage and in assessing the effects of interlesion distance on mutation frequency. As seen most clearly in the mutY and double mutant strains, the effect of spacing of the two lesions within the cluster on mutation is important (Fig. 4). There is commonly a decline in mutation frequency with an increase in spatial distance between lesions, and sometimes at a spacing of 5 bp no difference in the mutation frequency is seen between the clustered damage and the single 8-oxoG lesion. Interestingly, there is some evidence that the – orientation of clustered damage is less prone to reduced mutation frequency than the + orientation as the lesion spacing increases. Lastly, it is clear that absence of Fpg has less effect on mutation frequencies than the absence of MutY for clustered damage sites, although in the double mutant strain the absence of Fpg appears to have some impact even for a single 8-oxoG lesion.
Clustered DNA damage is more mutagenic than the single lesions
As noted in the Introduction from in vitro studies, a lesion(s) within clustered DNA damage is more difficult to excise than isolated lesions, due to the inhibition of BER enzymes (12,13,15–17,19,20,30,31). However, there is considerable variability in the levels of inhibition of the BER glycosylases/AP endonucleases depending on the types of lesion as well as the interlesion spacing in the bi-stranded cluster. This effect of interlesion distance will be influential only before plasmid replication, since the lesions will separate onto different DNA molecules following replication. The difficulty in excising lesions from within a cluster containing an AP site and a base lesion may arise in particular where the AP site is rapidly converted (Fig. 5, step 3) into a SSB, i.e. before excision of the base lesion (19,22) by AP endonucleases or through the AP lyase activity of Fpg. The latter activity is not affected by the presence of 8-oxoG (15). The specific activity for excision of 8-oxoG by Fpg (Fig. 5, steps 1 or 4) decreases as the distance to a closely spaced AP site/SSB decreases (17). However, if a DSB is formed in step 1, this would lead to loss of the plasmid and therefore not contribute to the mutations seen.
Figure 5. Diagram of mutagenic consequences of clustered DNA damage in different BER-deficient strains. The damaged site is shown at the top, followed by its progress through two replication cycles. At replication 1, it is proposed that the SSB-carrying strand is lost (see text), while the other, if unrepaired, will have mostly an adenine inserted opposite 8-oxoG. At the second replication, two outcomes of damage processing are shown, either with or without the action of MutY. The bold arrows indicate the main pathway of mutagenesis in an fpg mutY strain.
Additionally, repair of the SSB (Fig. 5, step 4) is retarded in the presence of 8-oxoG (14,22; M.Lomax and P.O’Neill, unpublished data). These reciprocal effects, depending on interlesion distance, may lead to persistence of the clustered damage site, allowing 8-oxoG to endure through to replication where its propensity to mispair with adenine will lead to an increased level of mutagenesis. Further, where the inhibitory effects of a SSB on 8-oxoG excision by Fpg have been assessed in vitro, the pattern of inhibition with distance of the SSB from the 8-oxoG lesion (17) is similar to that seen in the present mutation study. Firstly, this pattern shows a reduction with distance over 1–5 bp between the two lesions and, secondly, the reduction is most dramatic in the + orientation. These distinctions are also apparent in the results of studies on the effect of a SSB on the excision of 8-oxoG with eukaryotic BER enzymes and cell extracts (16,17).
MutY plays a more significant role than Fpg in countering the mutagenic effects of clustered DNA damage containing 8-oxoG
The results shown in Figure 4B and C indicate that the role of Fpg is significantly less important than that of MutY in processing clustered damage in E.coli. It is only with the clustered damaged constructs that this is notable, as neither single mutant strain appears to produce significant levels of mutation when dealing with either 8-oxoG or an AP site. This shows that the presence of either Fpg or MutY is sufficient to depress the potential mutability of the single 8-oxoG lesion. With the fpg strain, the levels of mutation arising from clustered damage sites are as low as those for the wild-type, whereas the corresponding data set for the mutY strain indicates a significant increase in levels of mutation. Figure 5 illustrates the potential mutagenic consequences of clustered DNA damage in different BER-deficient strains. Step 1 (Fig. 5) cannot occur in an fpg strain, so that steps 2 and 3 are the main pathways for clustered damage processing prior to replication. Following replication of the plasmid, when the 8-oxoG lesion is no longer associated with a cluster, the incorporation of adenine opposite 8-oxoG becomes potentially mutagenic and can only be countered by MutY. The efficiency of step 3 (Fig. 5) may be reduced in a fpg mutant strain relative to that in the wild-type since the AP site cannot be incised by Fpg, but the effect of 8-oxoG on the incision of an AP site by AP endonucleases is as yet not known. If a SSB is formed, its repair is retarded in wild-type cells, as discussed above, although retardation does not occur in a BER reconstructed system in the absence of Fpg protein (14). Therefore, the lower mutation frequency for clustered damage seen in the fpg mutant strain relative to the wild-type may reflect the relative importance of steps 2 and 3, in part through absence of the AP lyase activity of Fpg, but also through increased efficiency of repair of SSB in step 4 (Fig. 5). However, it is also possible that there is back-up capacity for oxidized guanine processing in E.coli; for example endonuclease VIII has weak 8-oxoG glycosylase activity and its absence in an fpg mutY strain increases the mutation frequency by a factor of 2–3 (32). It is interesting that absence of the mammalian analogue of Fpg, OGG1, in mice does not increase the incidence of tumours (33) and, unlike the ablation of subsequent steps in BER, is not embryonically lethal (34).
High mutation frequencies in the fpg mutY double mutant strain
It is evident that, following the processing of clustered damage sites, the mutation frequencies in the fpg mutY double mutant strain are higher than those in the mutY strain alone. Strikingly, only in the double mutant does the single 8-oxoG lesion give rise to a significant level of mutation (16%). Therefore, it appears that the presence of Fpg is important when processing a single 8-oxoG lesion, but only in the absence of MutY. While our data suggest that Fpg has less impact on the processing of clustered damage sites prior to replication of the plasmid DNA, it will be able to lower the probability of mutation by acting on 8-oxoG after replication when it pairs correctly with cytosine . If most of the 8-oxoG lesions are paired with adenine following replication, then MutY will be most important in protecting against mutation whether Fpg is present or not. In Figure 5 we show the anticipated fate of clustered damage, with emphasis on the formation of mutations in the fpg mutY strain (thick arrows). Thus, before the first plasmid DNA replication there will be some opportunity for repair of the damage through back-up repair, yielding a fraction of non-mutant plasmids. Once replication occurs, however, there will be little further chance of non-mutant plasmid forming in the fpg mutY strain since the replicative polymerase will mostly place an adenine opposite 8-oxoG (23), leading to mutation at the next or a subsequent replication.
GCTA is the major mutation arising from the presence of 8-oxoG
The mutation system we have devised shows remarkably good agreement of mutation frequency estimates from gel image analysis with those derived from plasmid DNA sequencing . The almost exclusive occurrence of transversion mutations in the sequence data reinforces the above suggestion that a clustered damage site is not processed at random, but sequentially in order to minimize the formation of an adjacent SSB by excision of 8-oxoG leading to more severe damage. If this processing led to a DSB, for example, then it would be anticipated that deletions would occur rather than base pair substitutions at the site of damage (36), as shown recently for closely opposed uracil lesions (37). However, we should note that to date we have sequenced mutations derived from clustered damage at the +1 position in the fpg mutY strain and it is possible that the mutation spectrum will differ at other positions or in different mutant backgrounds. Interestingly, all of the mutations detected with the fpg mutY strain are ‘complete’, i.e. from the sequence data there is no evidence that both an unmutated and a mutated DNA sequence coexist in a single bacterial cell. It might be expected that a mixed sequence would occur since the 8-oxoG lesion exists on one strand only and the other strand will have a cytosine residue which will pair normally at replication. The phenomenon of complete mutations induced by radiation and other agents has been observed in earlier mutagenesis studies with several different microorganisms, leading to a variety of interpretations (38–42). In the present study, a possible explanation is that when plasmids carrying a SSB adjacent to 8-oxoG are replicated, the SSB-carrying strand becomes unstable and is lost (Fig. 5). In plasmid DNA replication, one strand is displaced into a so-called R-loop (43), which may promote abortive replication when the strand contains a SSB.
In summary, clustered DNA damage sites containing only two lesions have greater mutagenic potential than those of the component lesions. Even these simple types of clustered DNA damage, which represent 60% of the overall number of clustered sites produced by -radiation (44) are highly mutagenic. Clustered DNA damage sites are produced by radiation (7–10) in a higher yield than that of DSB, so that at environmentally low doses of radiation, the probability of forming a DSB is less than that of a clustered DNA damage site in a given cell. This study therefore provides evidence that clustered DNA damage sites have a greater potential to contribute to the deleterious health effects of radiation than single lesions, particularly at low doses.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
We are very grateful to David Papworth for statistical advice, the Medical Research Council for a studentship to Colin Pearson, the Japan Atomic Energy Research Institute for a study grant to Naoya Shikazono and the European Commission (contract FIGH-CT2002-00207) for partial financial support.
REFERENCES
Friedberg,C.E., Walker,G.C. and Siede,W. (1995) DNA Repair and Mutagenesis. ASM Press, Washington, DC.
Pouget,J.P., Douki,T., Richard,M.J. and Cadet,J. (2000) DNA damage induced in cells by gamma and UVA radiation as measured by HPLC/GC-MS and HPLC-EC and Comet assay. Chem. Res. Toxicol., 13, 541–549.
Ward,J.F. (1994) The complexity of DNA damage: relevance to biological consequences. Int. J. Radiat. Biol., 66, 427–432.
Goodhead,D.T. (1994) Initial events in the cellular effects of ionizing radiations: clustered damage in DNA. Int. J. Radiat. Biol., 65, 7–17.
Nikjoo,H., O’Neill,P., Terrissol,M. and Goodhead,D.T. (1999) Quantitative modelling of DNA damage using Monte Carlo track structure method. Radiat. Environ. Biophys., 38, 31–38.
Cunniffe,S. and O’Neill,P. (1999) The complexity of radiation-induced DNA damage as revealed by exposure to cell extracts. Radiat. Res., 152, 421–427.
Sutherland,B.M., Bennett,P.V., Sidorkina,O. and Laval,J. (2000) Clustered DNA damages induced in isolated DNA and in human cells by low doses of ionizing radiation. Proc. Natl Acad. Sci. USA, 97, 103–108.
Jenner,T.J., Fulford,J. and O’Neill,P. (2001) Contribution of base lesions to radiation-induced clustered DNA damage: implication for models of radiation response. Radiat. Res., 156, 590–593.
Gulston,M., Fulford,J., Jenner,T., de Lara,C. and O’Neill,P. (2002) Clustered DNA damage induced by gamma radiation in human fibroblasts (HF19), hamster (V79-4) cells and plasmid DNA is revealed as Fpg and Nth sensitive sites. Nucleic Acids Res., 30, 3464–3472.
Sutherland,B.M., Bennett,P.V., Sutherland,J.C. and Laval,J. (2002) Clustered DNA damages induced by x rays in human cells. Radiat. Res., 157, 611–616.
Blaisdell,J.O. and Wallace,S.S. (2001) Abortive base-excision repair of radiation-induced clustered DNA lesions in Escherichia coli. Proc. Natl Acad. Sci. USA, 98, 7426–7430.
Chaudhry,M.A. and Weinfeld,M. (1995) The action of Escherichia coli endonuclease III on multiply damaged sites in DNA. J. Mol. Biol., 249, 914–922.
Chaudhry,M.A. and Weinfeld,M. (1997) Reactivity of human apurinic/apyrimidinic endonuclease and Escherichia coli exonuclease III with bistranded abasic sites in DNA. J. Biol. Chem., 272, 15650–15655.
Harrison,L., Hatahet,Z. and Wallace,S.S. (1999) In vitro repair of synthetic ionizing radiation-induced multiply damaged DNA sites. J. Mol. Biol., 290, 667–684.
David-Cordonnier,M.H., Laval,J. and O’Neill,P. (2000) Clustered DNA damage, influence on damage excision by XRS5 nuclear extracts and Escherichia coli Nth and Fpg proteins. J. Biol. Chem., 275, 11865–11873.
David-Cordonnier,M.H., Boiteux,S. and O’Neill,P. (2001) Efficiency of excision of 8-oxo-guanine within DNA clustered damage by XRS5 nuclear extracts and purified human OGG1 protein. Biochemistry, 40, 11811–11818.
David-Cordonnier,M.H., Laval,J. and O’Neill,P. (2001) Recognition and kinetics for excision of a base lesion within clustered DNA damage by the Escherichia coli proteins Fpg and Nth. Biochemistry, 40, 5738–5746.
Budworth,H., Dianova,I.I., Podust,V.N. and Dianov,G.L. (2002) Repair of clustered DNA lesions. Sequence-specific inhibition of long-patch base excision repair by 8-oxoguanine. J. Biol. Chem., 277, 21300–21305.
David-Cordonnier,M.H., Cunniffe,S.M., Hickson,I.D. and O’Neill,P. (2002) Efficiency of incision of an AP site within clustered DNA damage by the major human AP endonuclease. Biochemistry, 41, 634–642.
Tian,K., McTigue,M. and de los Santos,C. (2002) Sorting the consequences of ionizing radiation: processing of 8-oxoguanine/abasic site lesions. DNA Repair (Amst.), 1, 1039–1049.
Weinfeld,M., Rasouli-Nia,A., Chaudhry,M.A. and Britten,R.A. (2001) Response of base excision repair enzymes to complex DNA lesions. Radiat. Res., 156, 584–589.
Lomax,M.E., Gulston,M.K. and O’Neill,P. (2002) Chemical aspects of clustered DNA damage induction by ionising radiation. Radiat. Prot. Dosimetry, 99, 63–68.
Grollman,A.P. and Moriya,M. (1993) Mutagenesis by 8-oxoguanine: an enemy within. Trends Genet., 9, 246–249.
Moriya,M., Ou,C., Bodepudi,V., Johnson,F., Takeshita,M. and Grollman,A.P. (1991) Site-specific mutagenesis using a gapped duplex vector: a study of translesion synthesis past 8-oxodeoxyguanosine in E. coli. Mutat. Res., 254, 281–288.
Michaels,M.L., Cruz,C., Grollman,A.P. and Miller,J.H. (1992) Evidence that MutY and MutM combine to prevent mutations by an oxidatively damaged form of guanine in DNA. Proc. Natl Acad. Sci. USA, 89, 7022–7025.
Cupples,C.G. and Miller,J.H. (1989) A set of lacZ mutations in Escherichia coli that allow rapid detection of each of the six base substitutions. Proc. Natl Acad. Sci. USA, 86, 5345–5349.
Weston,A., Brown,M.G., Perkins,H.R., Saunders,J.R. and Humphreys,G.O. (1981) Transformation of Escherichia coli with plasmid deoxyribonucleic acid: calcium-induced binding of deoxyribonucleic acid to whole cells and to isolated membrane fractions. J. Bacteriol., 145, 780–787.
Hanahan,D. (1983) Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol., 166, 557–580.
Malyarchuk,S., Youngblood,R., Landry,A.M., Quillin,E. and Harrison,L. (2003) The mutation frequency of 8-oxo-7,8-dihydroguanine (8-oxodG) situated in a multiply damaged site: comparison of a single and two closely opposed 8-oxodG in Escherichia coli. DNA Repair (Amst.), 2, 695–705.
Harrison,L., Hatahet,Z., Purmal,A.A. and Wallace,S.S. (1998) Multiply damaged sites in DNA: interactions with Escherichia coli endonucleases III and VIII. Nucleic Acids Res., 26, 932–941.
David-Cordonnier,M.H., Boiteux,S. and O’Neill,P. (2001) Excision of 8-oxoguanine within clustered damage by the yeast OGG1 protein. Nucleic Acids Res., 29, 1107–1113.
Blaisdell,J.O., Hatahet,Z. and Wallace,S.S. (1999) A novel role for Escherichia coli endonuclease VIII in prevention of spontaneous GT transversions. J. Bacteriol., 181, 6396–6402.
Klungland,A., Rosewell,I., Hollenbach,S., Larsen,E., Daly,G., Epe,B., Seeberg,E., Lindahl,T. and Barnes,D.E. (1999) Accumulation of premutagenic DNA lesions in mice defective in removal of oxidative base damage. Proc. Natl Acad. Sci. USA, 96, 13300–13305.
Wilson,S.H. and Kunkel,T.A. (2000) Passing the baton in base excision repair. Nature Struct. Biol., 7, 176–178.
Tchou,J., Bodepudi,V., Shibutani,S., Antoshechkin,I., Miller,J., Grollman,A.P. and Johnson,F. (1994) Substrate specificity of Fpg protein. Recognition and cleavage of oxidatively damaged DNA. J. Biol. Chem., 269, 15318–15324.
Dianov,G.L., O’Neill,P. and Goodhead,D.T. (2001) Securing genome stability by orchestrating DNA repair: removal of radiation-induced clustered lesions in DNA. Bioessays, 23, 745–749.
D’Souza,D,I. and Harrison,L. (2003) Repair of clustered uracil DNA damages in Escherichia coli. Nucleic Acids Res., 31, 4573–4581.
Holliday,R. (1962) Mutation and replication in Ustilago maydis. Genet. Res., 3, 472–486.
Kubitschek,H.E. (1964) Mutation without segregation. Proc. Natl Acad. Sci. USA, 52, 1374–1381.
Munson,R.J. and Bridges,B.A. (1964) Segregation of radiation-induced mutations in Escherichia coli. Nature, 203, 270–273.
Witkin,E.M. and Sicurella,C. (1964) Pure clones of lactose-negative mutants obtained in Escherichia coli after treatment with 5-bromouracil. J. Mol. Biol., 8, 610–613.
Nasim,A. and Auerbach,C. (1967) The origin of complete and mosaic mutants from mutagenic treatment of single cells. Mutat. Res., 4, 1–14.
Kuzminov,A. (1999) Recombination repair of DNA damage in Escherichia coli and bacteriophage . Microbiol. Mol. Biol. Rev., 63, 751–813.
Nikjoo,H., O’Neill,P., Wilson,W.E. and Goodhead,D.T. (2001) Computational approach for determining the spectrum of DNA damage induced by ionizing radiation. Radiat. Res., 156, 577–583.(Colin G. Pearson, Naoya Shikazono1, John)