Chem. Res. Toxicol. 2005, 18, 1271-1278
1271
Mutagenesis Induced by Oxidized DNA Precursors: Roles of Y Family DNA Polymerases in Escherichia coli Kazuya Satou,† Masami Yamada,‡ Takehiko Nohmi,‡ Hideyoshi Harashima,† and Hiroyuki Kamiya*,† Graduate School of Pharmaceutical Sciences, Hokkaido University, Kita-12, Nishi-6, Kita-ku, Sapporo 060-0812, Japan, and Division of Genetics and Mutagenesis, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku, Tokyo 158-8501, Japan Received February 21, 2005
To reveal the roles of Y family DNA polymerases in the mutagenesis induced by oxidatively damaged DNA precursors, 2-hydroxy-dATP (2-OH-dATP) and 8-hydroxy-dGTP (8-OH-dGTP) were introduced into Escherichia coli strains deficient in the Y family polymerases, DNA polymerase IV (pol IV, encoded by the dinB gene) and DNA polymerase V (pol V, encoded by the umuDC locus). The mutation induced by 2-OH-dATP, but not that induced by 8-OH-dGTP, occurred less frequently in the dinB- strain than in the wild-type (wt) strain, suggesting the involvement of pol IV in the mutagenesis by 2-OH-dATP. Expression of pol IV from plasmid enhanced the mutagenesis by 2-OH-dATP in the dinB- strain. This enhancement depends on the polymerase activity since the expression of a mutant pol IV lacking the polymerase activity did not increase the mutations induced by 2-OH-dATP. In contrast, both 2-OH-dATP and 8-OHdGTP caused mutations more efficiently in the umuDC- strain than in the wt strain, suggesting that the umuDC gene products suppressed the mutagenesis by these oxidized DNA precursors. The DNA polymerase activity was not required for the suppressive effects because expression of the umuDC gene products lacking the polymerase activity also suppressed the mutagenesis. These results suggest that the E. coli pol IV was involved in mutagenesis by 2-OH-dATP and that the umuDC gene products play suppressive role(s) in the mutagenesis by damaged nucleotides.
Introduction (ROS)1
Reactive oxygen species are generated endogenously and exogenously and are believed to be an important source of the mutations that cause various diseases, aging, and neurodegeneration (1). Many of the lesions formed on DNA by ROS result in the alteration of genetic information (2). 2-Hydroxyadenine (2-OH-Ade) and 8-hydroxyguanine (8-OH-Gua), which are mutagenic oxidative DNA lesions (3), are likely to be produced in the DNA through the following two pathways. One is the direct modification of DNA residues, and the other is the incorporation by DNA polymerase(s) (pol) of oxidatively damaged DNA precursors, 2-hydroxy-2′-deoxyadenosine 5′-triphosphate (2-OH-dATP) and 8-hydroxy-2′-deoxyguanosine 5′-triphosphate (8-OH-dGTP), which are produced in the nucleotide pool. Both of the pathways contribute similarly to the formation of 8-OH-Gua (4). In addition, the involvement of 8-OH-dGTP as well as 8-OH-Gua in DNA in mutagenesis was indicated by analysis of bacterial strains lacking MutT, MutM, and/or MutY (5). The incorporation of oxidatively damaged precursor might be more important in the case of 2-OH-Ade, because its * To whom correspondence should be addressed. Tel: +81-11-7063733. Fax: +81-11-706-4879. E-mail:
[email protected]. † Hokkaido University. ‡ National Institute of Health Sciences. 1 Abbreviations: HE, holoenzyme; 2-OH-Ade, 2-hydroxyadenine; 2-OH-dATP, 2-hydroxy-2′-deoxyadenosine 5′-triphosphate; 8-OH-Gua, 8-hydroxyguanine; 8-OH-dGTP, 8-hydroxy-2′-deoxyguanosine 5′-triphosphate; pol, polymerase; ROS, reactive oxygen species; TLS, translesion synthesis; wt, wild-type.
formation by the Fenton type reaction is much more efficient in the monomeric form than in DNA in vitro (6). The significance of these damaged nucleotides is supported by the presence of their specific hydrolyzing enzymes (7-9). The MutT protein removes 8-OH-dGTP from the nucleotide pool by hydrolysis to the corresponding monophosphate derivative (7). In fact, a mutT deficient Escherichia coli strain shows a 1000-fold increase in the mutation rate as compared to the wild-type (wt) strain (4). In addition, the human MutT homologue, hMTH1, hydrolyzes 2-OH-dATP as well as 8-OH-dGTP (8, 10), and the knock-out mice of this gene exhibit increased occurrences of carcinogenesis in the lung, liver, and stomach (11). Moreover, both the 2-OH-dATPase and the 8-OH-dGTPase activities of hMTH1 suppressed H2O2induced cell death (12). Thus, the incorporation of 2-OHdATP and 8-OH-dGTP seems to strongly affect cellular functions. These oxidized DNA precursors are misincorporated by DNA pol(s) opposite incorrect bases and form mispairs. 8-OH-dGTP is incorporated opposite A, thus inducing A:TfC:G transversions in E. coli (7). In contrast, the misincorporation mode of 2-OH-dATP is highly DNA pol specific. E. coli DNA pols misincorporate 2-OH-dATP opposite G, whereas mammalian DNA pol R incorporates 2-OH-dATP opposite C on template DNA in vitro (6, 13). These variations generate the different mutation spectra of 2-OH-dATP, G:CfT:A transversions, and G:CfA:T transitions in E. coli and in mammalian nuclear extracts, respectively (14, 15), although the actual mutation
10.1021/tx050046b CCC: $30.25 © 2005 American Chemical Society Published on Web 07/12/2005
1272
Chem. Res. Toxicol., Vol. 18, No. 8, 2005
spectrum of 2-OH-dATP in living mammalian cells is unknown. The Y family of DNA pols, a new class that lacks intrinsic exonuclease activity, has been found in various species (16). This family of DNA pols seems to be involved in error-free or error-prone translesion synthesis (TLS) (17). Recently, the involvement of Y family DNA pols in the incorporation of damaged DNA precursors was also suggested by in vitro experiments using purified DNA pols. The archeal Y family DNA pols P1 and P2 and the human pol η exclusively incorporate 8-OH-dGTP opposite A in the template DNA and incorporate 2-OH-dATP opposite G more efficiently than opposite T (18). Thus, it would be quite interesting to examine the in vivo roles of Y family DNA pols in the incorporation of oxidatively damaged deoxyribonucleotides. Here, we examined the effects of E. coli Y family DNA pols, DNA pol IV (pol IV, encoded by the dinB gene) (19), and DNA pol V (pol V, encoded by the umuDC locus) (20, 21) on the mutagenesis induced by 2-OH-dATP and 8-OH-dGTP. 2-OH-dATP induced mutations less efficiently in the dinB- strain than in the wt strain, and both 2-OH-dATP and 8-OH-dGTP induced mutations more efficiently in the umuDC- strain than in the wt strain. Interestingly, these effects of pol IV and the umuDC gene products seemed to be DNA pol activitydependent and -independent, respectively. These results suggest that the E. coli Y family DNA pols, pol IV, and the umuDC gene products, enhanced and suppressed, respectively, the mutagenesis induced by oxidatively damaged DNA precursors.
Materials and Methods Materials. The FPLC grade nucleoside triphosphates were from Amersham Bioscience (Piscataway, NJ). 2-OH-dATP and 8-OH-dGTP were prepared from dATP and dGTP, respectively, and were purified by HPLC as described (6, 22). The purified nucleotides showed the same ultraviolet spectra as those of the corresponding deoxyribonucleosides in the literature (23, 24) and were eluted as a single peak in both reverse phase and anion exchange HPLC (data not shown). Their purities were estimated to be more than 99%. Purified oligonucleotides were from Hokkaido System Science (Sapporo, Japan) and Sigma Genosys Japan (Ishikari, Japan). The E. coli strain YG7207 (∆dinB::kan) is a derivative of the AB1157 strain (25). The E. coli strains YG7209 and YG7210 are derivatives of AB1157 and YG7207, respectively, in which ∆umuDC::Cat (26) was transferred by P1 transduction. Plasmids pYG768 and pYG779, respectively, carry dinB and dinB003, of which the latter encodes a mutant (D103N) pol IV lacking the pol activity (26, 27). The amount of dinB mRNA in AB1157 containing pYG768 was more than 10-fold as abundant as that in AB1157 without the plasmid (data not shown). The pSE117 plasmid contains the entire umuDC operon (28), and the amount of umuC mRNA in AB1157 containing pSE117 was more than 10-fold as abundant as that in AB1157 without the plasmid (data not shown). The conversion of the 101st codon from GAT (Asp) to AAT (Asn) of the umuC gene in pSE117 was performed according to the protocol for the Altered Sites II in vitro Mutagenesis System (Promega, Madison, WI). The mutagenic oligonucleotide used was 5′-dAGATTTACAGTATTAATGAGGCATTCTGC-3′, where the mismatched base is underlined. The mutation was confirmed by DNA sequencing. This plasmid (pUE101) expresses a mutant UmuD′2C protein (pol V), which cannot perform TLS (20). Plasmid pUE102 was constructed by deleting the Hind III fragment (about 3 kb) including 2/3 of the umuC gene (amino acid residues 149-422) from pSE117. Plasmids pKMTa1 and pKMTb1, derivatives of pYG768, were
Satou et al. Table 1. Plasmids Used in This Study plasmids
relevant genotype or characteristics
source
pYG768 pYG779 pKMTa1 pKMTb1 pSE117 pUE101 pUE102 pGW2101 pGW2133
pWSK29 derivative, DinB pWSK29 derivative, DinB (D103N) pYG768 derivative, lexA-, dinBpYG768 derivative, dinBpBR322 derivative, UmuDC pSE117 derivative, UmuDC (D101N) pSE117 derivative, UmuD, UmuC (1-148) pBR322 derivative, UmuDC, roppGW2101 derivative, UmuD′C
27 25 this study this study 28 this study this study 29 29
constructed by Sac I-EcoR I (including the lexA binding region and the dinB gene) and SnaB I-EcoR I (including the dinB gene) digestions, respectively, followed by end filling of the 3′ overhangs using a Blunting High kit (Toyobo, Osaka, Japan) prior to religation of the blunt-ended DNA. Plasmids pGW2101 and pGW2133 are rop- plasmids encoding UmuDC and UmuD′C, respectively (29). The plasmid DNAs used in this study are shown in Table 1. Mutation Assays. The introduction of 2-OH-dATP and 8-OH-dGTP was performed essentially as described (14). Single colonies of each strain on optimal selection plates were suspended in 7.5 mL of prewarmed LB medium and incubated at 37°C until the cultures were growing exponentially (ca. 0.6 OD610), and competent cells were prepared by a treatment with 0.1 M calcium chloride. The nucleotide solution (1.25 µL) was added to a final concentration of 125 or 250 µM to 50 µL of the E. coli suspension, which was placed on ice for 20-25 min. After heat shock treatment (42 °C for 45 s and then 0 °C for 2 min), SOC medium (450 µL) was added, and the cells were incubated at 37 °C for 2 h with shaking. LB and SOC media containing 150 µg/mL ampicillin were used for E. coli strains harboring the vector plasmids in this paper. A portion of the culture was diluted with ice-cold LB, plated on an LB agar plate, and incubated at 37 °C overnight (the titer plate). Another portion of the culture was plated on rifampicin (100 µg/mL) plates, which were incubated at 37 °C for 24 h to select rpoB mutants. The mutant frequency was calculated according to the numbers of colonies on the titer and selection plates. The statistical significance of the results was examined by the Student’s t-test. Sequence Analysis of the rpoB Gene. Portions of the rpoB gene were amplified by PCR, using Taq DNA pol (Toyobo) and sets of primers as follows: 5′-dACAGGATATGATCAACGCCAA3′ and 5′-dCGATACGGAGTCTCAAGGAA-3′ for positions 14671775 and 5′-dTGGTGATCTATGAGCGCGAA-3′ and 5′-dACCAGTTCCATCTGCAGCTT-3′ for positions 305-725, as described (30, 31). The closest 2-5 colonies to the center of a rifampicin plate were selected for sequencing analysis, to avoid any picking bias based on colony size. The amplified PCR products were sequenced using the former primers, an ABI PRISM Big Dye Terminator Cycle Sequencing Kit, and an ABI model 377 DNA sequencer (Applera, Norwalk, CT).
Results Effects of DNA Pols IV and V on Mutations by Damaged DNA Precursors. 2-OH-dATP and 8-OHdGTP, the oxidized forms of dATP and dGTP, respectively, show high mutagenicity in E. coli (14). To investigate the effects of Y family DNA pols on mutagenesis by these oxidatively damaged DNA precursors, we introduced 2-OH-dATP and 8-OH-dGTP by the CaCl2 method into E. coli strains deficient in DNA pol IV (pol IV, encoded by the dinB gene) and/or DNA pol V (pol V, encoded by the umuDC locus). Moreover, E. coli strains containing the episomal dinB gene and the umuDC locus were used. After 2-OH-dATP and 8-OH-dGTP were added to a CaCl2-treated E. coli suspension, the cells were cultured at 37 °C for 2 h after heat shock. The chromosomal rpoB gene was used as the mutagenesis target,
Oxidized DNA Precursors and Y Family DNA Polymerases
Chem. Res. Toxicol., Vol. 18, No. 8, 2005 1273
Table 2. Effects of Pol IV and Pol V Expression on Mutagenesis by Oxidized Nucleotides mutant frequency (×10-8)a 2-OH-dATP AB1157 (wt) YG7207 (dinB-) YG7209 (umuDC-) YG7210 (dinB-, umuDC-) AB1157/pYG768 (DinB) AB11157/pSE117 (UmuDC)
addedb
8-OH-dGTP addedb
0 µM
125 µM
250 µM
0 µM
125 µM
250 µM
5.7 (2.9) 2.8 (1.6) 3.8 (1.6) 3.2 (2.1) 5.1 (1.1) 11.1 (2.7)
9.0 (4.4) 5.1 (2.8) 12.7 (6.7) 11.4 (6.5) 10.9 (3.9) 13.1 (3.6)
21.5 (8.1) 11.4 (6.4) 32.7 (10.4) 31.6 (13.6) 23.6 (7.6) 19.4 (6.6)
6.0 (2.2) 3.0 (1.3) 3.9 (1.3) 5.1 (2.9) 7.0 (2.6) 14.1 (5.2)
13.0 (5.8) 9.5 (3.6) 19.1 (4.5) 18.2 (7.6) 17.1 (7.9) 19.3 (4.9)
18.1 (7.6) 15.8 (7.2) 31.8 (8.3) 25.6 (8.9) 23.4 (9.4) 21.8 (6.1)
a All data are represented as means (SD) of 11-26 independent experiments. b Amount of 1.25 µL, to a final concentration of 125 or 250 µM of each oxidized nucleoside triphosphate, added to 50 µL of a 0.1 M calcium chloride-treated E. coli suspension. The control experiment (0 µM) was acomplished with the same volume of water.
and single base substitutions in various sequence contexts could be monitored by the selection with rifampicin (30). In the control experiment, in which an equal volume of water instead of the damaged nucleotide solution was added to the bacteria, the observed mutant frequency was 2-fold lower in the pol IV deficient strain, YG7207, as compared to that in AB1157, the wt strain (Table 2). The pol V deficient strain YG7209 also showed a slightly decreased mutant frequency. Conversely, the additional expression of pol V by the pSE117 plasmid containing the umuDC genes in the wt strain exhibited more than a 2-fold increase in the mutant frequency in the control experiments. Although the overexpression of pol IV reportedly causes large elevations of the mutant frequency (25, 32), only a slight increase was observed in this study, possibly because of the presence of the lexA operating region on the pYG768 plasmid. LexA keeps the expression of dinB low by transcriptional regulation, even on an episomal plasmid (27). The mutagenic effect of pol V expressed from pSE117 might be limited similarly. We used these lexA binding site-containing plasmids to avoid too much of an increase in the background mutant frequencies by the error-prone DNA pols, for the evaluation of the mutagenesis by damaged DNA precursors. Next, we compared the frequencies of the mutations induced by these damaged nucleotides in various strains. The mutant frequencies were increased by the addition of 2-OH-dATP and 8-OH-dGTP in all of the strains used, in a dose-dependent manner (Table 2). In contrast, the treatment with unmodified dATP and dGTP did not increase mutant frequencies in all of the strains tested (data not shown), as described previously (14). Thus, mutations by the addition of 2-OH-dATP and 8-OH-dGTP were not due to mere nucleotide imbalance. To evaluate the actual frequency of mutations induced by the exogenous nucleotides, we subtracted the mutant frequency of the control experiments from the value obtained in the presence of 250 µM of an oxidized nucleotide. This subtracted mutant frequency was effective when strains with different background mutant frequencies were compared (33). Each subtracted frequency was calculated from the 250 µM nucleotide and control experiments using the same original single colonies. The subtracted mutant frequencies in the wt strain were 15.8 × 10-8 and 12.1 × 10-8 with the addition of 2-OH-dATP and 8-OH-dGTP, respectively (Figure 1). 2-OH-dATP induced mutations less efficiently (8.6 × 10-8, P < 0.001, examined by the Student’s t-test) in the pol IV deficient strain than in the wt strain, suggesting that pol IV enhanced mutagenesis by 2-OH-dATP. The statistical significance (P < 0.01) was also obtained by
Figure 1. Mutant frequencies induced by (A) 2-OH-dATP and (B) 8-OH-dGTP in various E. coli strains. Values were calculated by subtracting the mutant frequencies of control experiments from the values obtained in the presence of 250 µM 2-OH-dATP or 8-OH-dGTP. This subtraction was done for the 250 µM nucleotide and control experiments, using the same original single colonies, and each subtracted frequency was considered as a single datum. The presented data are the means of 11-26 independent experiments. Error bars represent SD. *P < 0.05; **P < 0.01; ***P < 0.001 (significant difference vs the wt strain).
the Mann-Whitney test. However, the additional expression of pol IV from plasmid in the wt strain did not
1274
Chem. Res. Toxicol., Vol. 18, No. 8, 2005
Satou et al.
Table 3. Spectra of Mutations Induced by Damaged Nucleotides in E. colia YG7207 (dinB-)
AB1157 (wt) control single base substitution transition A:TfG:C 7 (5) G:CfA:T 98 (64) transversion A:TfT:A 15 (10) A:TfC:G 9 (6) G:CfT:A 20 (13) G:CfC:G 3 (2) others 0 (0) total 152 (100)
2-OH-dATP
8-OH-dGTP
control
2-OH-dATP
8-OH-dGTP
control
2-OH-dATP
8-OH-dGTP
2 (2) 26 (31)
0 (0) 22 (28)
18 (13) 57 (38)
2 (3) 15 (19)
1 (1) 7 (8)
25 (20) 34 (28)
2 (3) 5 (6)
0 (0) 4 (5)
0 (0) 1 (1) 54 (64) 1 (1) 0 (0) 84 (100)
1 (1) 55 (71) 0 (0) 0 (0) 0 (0) 78 (100)
15 (11) 31 (19) 26 (17) 4 (2) 0 (0) 151 (100)
3 (4) 6 (8) 53 (66) 1 (1) 0 (0) 80 (100)
1 (1) 71 (83) 2 (2) 3 (4) 0 (0) 85 (100)
20 (16) 20 (16) 17 (14) 3 (3) 3b (3) 122 (100)
1 (1) 3 (4) 65 (83) 2 (3) 0 (0) 78 (100)
2 (3) 66 (89) 2 (3) 0 (0) 0 (0) 74 (100)
YG7210 (dinB-,umuDC-) control single base substitution transition A:TfG:C 10 (11) G:CfA:T 23 (24) transversion A:TfT:A 16 (17) A:TfC:G 30 (32) G:CfT:A 10 (11) G:CfC:G 4 (4) others 0 (0) total 93 (100) a
YG7209 (umuDC-)
AB1157 /pYG768 (DinB)
AB1157 /pSE117 (UmuDC)
2-OH-dATP
8-OH-dGTP
control
2-OH-dATP
8-OH-dGTP
control
2-OH-dATP
8-OH-dGTP
0 (0) 1 (2)
2 (3) 4 (6)
9 (10) 44 (51)
0 (0) 10 (17)
1 (2) 11 (17)
9 (7) 68 (52)
1 (1) 29 (40)
3 (6) 25 (48)
1 (2) 7 (12) 50 (83) 1 (2) 0 (0) 60 (100)
5 (7) 55 (81) 2 (3) 0 (0) 0 (0) 68 (100)
11 (13) 16 (18) 5 (6) 1 (1) 1b(1) 87 (100)
1 (2) 1 (2) 48 (80) 0 (0) 0 (0) 60 (100)
2 (3) 47 (73) 1 (2) 2 (3) 0 (0) 64 (100)
8 (6) 0 (0) 33 (25) 14 (11) 0 (0) 132 (100)
2 (3) 0 (0) 39 (54) 1 (1) 0 (0) 72 (100)
1 (2) 16 (31) 6 (12) 1 (2) 0 (0) 52 (100)
All data are represented as cases found (%). b See the Supporting Information Table 1 legend.
significantly increase the mutagenesis by 2-OH-dATP (Figure 1A). In contrast, neither the deficiency nor the additional expression of pol IV affected the mutation frequency of 8-OH-dGTP (Figure 1B). The deficiency in pol V significantly facilitated the mutagenicity of 2-OH-dATP (P < 0.001) and 8-OH-dGTP (P < 0.001) (Figure 1). A similar enhancement was obtained in the dinB umuDC mutant strain YG7210. This result suggests that pol V suppressed the mutagenesis induced by these oxidized nucleotides. When pol V was additionally expressed from plasmid in the wt strain, the subtracted mutant frequencies were decreased to 8.3 × 10-8 and 7.7 × 10-8 (P < 0.001 and P < 0.05), in the cases of 2-OH-dATP and 8-OH-dGTP, respectively. To rule out the possibility that the permeation efficiency of the damaged nucleotides into cells causes the difference in the mutant frequencies, we introduced 33Plabeled dATP into the wt, pol IV and pol V deficient strains to check the membrane permeability. As expected, there were no differences in the incorporation amount among the strains tested (data not shown). In addition, cell division occurred about twice in all of the strains, from the time of the introduction of damaged nucleotides to the plating, indicating that the growth rate did not vary considerably between the wt and its derivative strains. Thus, the difference in the induced mutant frequency was not due to the frequency of DNA replication. These results suggest that the mutagenicity of 2-OHdATP was enhanced and suppressed in the presence of pol IV and pol V, respectively. The mutagenesis by 8-OHdGTP was also suppressed by pol V, although a deficiency in pol IV did not affect it in this case. Mutation Spectra of 2-OH-dATP and 8-OH-dGTP. We then analyzed the sequences of the rpoB genes in a total of 1628 colonies, obtained with the control, 2-OHdATP, and 8-OH-dGTP experiments in various strains. Between two and five colonies per plate were picked and sequenced, based on the number of mutants on the
rifampicin plates. When two colonies were picked from one plate, each colony had different mutations. Five colonies were picked from one plate when more than 50 mutant colonies appeared on the plate. Because the E. coli suspension seems to have been plated before the third cell division as described above and because the mutations are fixed during the second replication, most of the nucleotide-induced mutants would be derived from different E. coli cells, which incorporated the nucleotide. Supporting Information Tables 1-3 and Table 3 show the results of the sequencing analysis of the rpoB mutants. In the control experiment, 64% of the mutants contained G:CfA:T transitions in the case of the wt strain (Supporting Information Table 1 and Table 3). The ratio of these mutations was greatly decreased in the pol IV and pol V deficient strains. Only the additional expression of pol V enhanced this type of mutation, when we consider the increase in the total mutant frequency. The absence of an obvious effect of pol IV expression from plasmid in the wt strain may be due to the existence of a relatively large number of pol IV molecules in cells (250 molecules/cell), even without induction by the SOS response (27). The number of A:TfC:G and A:TfG:C mutations, which were shown to be induced on the cII gene by the overexpression of pol IV (32), seemed to increase in the AB1157/pYG768 combination. The frequencies of G:CfC:G and G:CfT:A transversions were increased in the pol V-overexpressing strain, as reported previously (34). A deficiency in pol V increased the occurrence of A:TfC:G transversions, while this type of mutation was not observed when pol V was additionally expressed in the wt strain. Unexpectedly, one and three examples of 3 or 9 bp deletions and insertions were detected in the pol IV-overexpressing strain (TAT insertion at positions between 1541 and 1542) and the pol V deficient strain (TTT insertion at positions between 1540 and 1543, 9 bp deletion from 1546 to 1554, and GCA deletion from 1594 to 1596), respectively.
Oxidized DNA Precursors and Y Family DNA Polymerases
Chem. Res. Toxicol., Vol. 18, No. 8, 2005 1275
Table 4. Effects of the dinB and umuDC Gene Products mutant frequency (×10-8)a 2-OH-dATP added plasmid
0 µM
250 µM
0 µM
250 µM
YG7207
none pYG768 (DinB) pYG779 (DinB D103N) none pSE117 (UmuDC) pUE101 (UmuDC D101N) pUE102 (UmuD)b pGW2101 (UmuDC) pGW2133 (UmuD′C)
2.8 (1.6) 3.1 (1.6) 4.1 (2.1) 3.8 (1.6) 6.3 (3.8) 2.8 (1.6) 2.4 (0.9) 2.4 (0.8) 14.0 (6.9)
11.4 (6.4) 20.2 (9.6) 5.7 (2.1) 32.7 (10.4) 18.9 (7.8) 12.4 (6.9) 34.2 (5.6) 12.0 (1.4) 32.8 (5.9)
3.0 (1.3) 4.4 (2.1) 4.4 (2.2) 3.9 (1.3) 5.8 (4.3) 3.0 (1.6) 2.4 (1.1) 5.1 (4.5) 10.5 (4.6)
15.8 (7.2) 23.3 (4.0) 8.0 (3.6) 31.8 (8.3) 24.3 (5.2) 17.8 (9.3) 28.3 (7.7) 16.7 (5.4) 24.6 (3.6)
YG7209
a
8-OH-dGTP added
strain
All data are represented as means (SD) of 3-26 independent experiments. b UmuD, UmuC (1-148).
The addition of 250 µM 2-OH-dATP specifically induced G:CfT:A transversions in all strains examined (Supporting Information Table 2 and Table 3). Similarly, 8-OH-dGTP induced A:TfC:G transversions (Supporting Information Table 3 and Table 3). These results suggest that 2-OH-dATP and 8-OH-dGTP were misincorporated opposite G and A, respectively, independent of the expression of pol IV or pol V. The G:CfT:A transversions induced by 2-OH-dATP formed some minor hotspots (Supporting InformationTable 2). It has been reported that 2-OH-dATP and 8-OH-dGTP preferentially induce transversions in 5′-GG-3′ and 5′- TAA-3′ sequences, respectively (14). The sites of the G:CfT:A mutations induced by 2-OH-dATP were mostly within 5′-GG-3′ sequences (25/54 in the wt strain), whereas 8-OH-dGTP induced A:TfC:G transversions mostly in 5′-C/GA-3′ sequences. The reason for this contradictory result with 8-OH-dGTP might be due to the paucity of detectable A sites with T in the 5′-flanking sequence in the rpoB gene. No remarkable difference in the incidence of G:CfT:A or A:TfC:G mutations in these hotspots was observed with the expression of pol IV and pol V. Pol Activity-Dependent and -Independent Effects on Mutagenesis by Damaged DNA Precursors. The plasmids pYG779 (19) and pUE101, respectively, encode mutant pol IV and pol V, with an amino acid substitution in the highly conserved region of the UmuC analogues (Asp-103fAsn and Asp-101fAsn, respectively). To examine the mechanism by which pol IV and pol V affect the mutagenesis by damaged DNA precursors, we introduced these plasmids into the pol IV and polV deficient strains YG7207 and YG7209, respectively, and the mutations induced by 2-OH-dATP and 8-OH-dGTP were measured (Table 4). Because these mutant proteins lack their pol activities, the expression of their derivatives cannot induce SOS mutagenesis (19-21, 35). The mutant UmuC (D101N) product forms the UmuD2C (D101N) complex with two molecules of UmuD and confers the cold sensitivity in a lexA deficient strain (35), which means that the folding of the mutant protein is probably normal. We measured the mutant frequency in strains harboring various plasmids, including pYG779 and pUE101, with and without the treatment with damaged nucleotides (Figure 2). The expression of pol IV in YG7207 restored the mutant frequency to a level comparable to that in the wt strain, when 2-OH-dATP was added (8.6-17.1 × 10-8) (Figure 2A). Surprisingly, the mutant pol IV (DinB D103N) reduced the mutant frequency, as compared to the pol IV deficient strain without the plasmid. This reduction also occurred in the case of 8-OH-dGTP.
Figure 2. Effect of the pol IV and pol V pol activities. Subtracted mutant frequencies, calculated as in the Figure 1 legend, were examined (A) in a pol IV deficient strain (YG7207) expressing DinB (pol IV) and DinB D103N (mutant DinB lacking the pol activity) and (B) in a pol V deficient strain (YG7209) expressing UmuDC (pol V), UmuDC D101N (mutant UmuC with UmuD lacking the pol activity), and UmuD with UmuC (1-148). Data are the means of at least five independent experiments. Error bars represent SD. *P < 0.05; **P < 0.01; ***P < 0.001.
Control mock vectors, constructed by the deletion of only the dinB coding region, and dinB and lexA from pYG768 (pKMTb1 and pKMTa1, respectively), did not influence
1276
Chem. Res. Toxicol., Vol. 18, No. 8, 2005
Satou et al.
decreased the mutant frequency, in good agreement with the expression from another vector, pSE117 (Figure 2). Intriguingly, the expression of UmuD′C, the activated form of UmuDC that actually carries out TLS, also suppressed the mutagenesis by 2-OH-dATP and 8-OHdGTP, albeit less effectively than that of UmuDC. This unexpected outcome suggests the possibility that both the UmuD2C and the UmuD′2C complexes had the suppressing activity or that UmuC suppressed the mutations.
Discussion
Figure 3. Subtracted mutant frequencies, calculated as in the Figure 1 legend, induced by (A) 2-OH-dATP and (B) 8-OH-dGTP in a pol V deficient strain (YG7209) harboring plasmids encoding UmuDC or UmuD′C. Data are the means of at least three independent experiments. Error bars represent SD. *P < 0.05; **P < 0.01; ***P < 0.001 (significant difference vs YG7209).
the mutagenicity of the two damaged nucleotides (data not shown). Thus, pol IV exerted the specific facilitation of the mutagenesis by 2-OH-dATP, probably via its pol activity. The reasons why the mutant pol IV (DinB D103N) reduced the mutant frequency will be discussed later. In contrast, pol V suppressed the mutations induced by 2-OH-dATP and 8-OH-dGTP (Figure 2B). Because the mutant pol V (UmuD2C D101N) also decreased the subtracted mutant frequency to a similar extent, pol V suppressed the induction of mutations by damaged DNA precursors in a pol activity-independent manner. The expression of UmuD did not affect the mutant frequencies induced by 2-OH-dATP and 8-OH-dGTP. Thus, most of the suppressive effect of pol V was not due to the action(s) of UmuD without UmuC. Moreover, other plasmids encoding UmuDC (pGW2101) and UmuD′C (pGW2133) were introduced into the pol V deficient strain, and the mutant frequencies were measured (Table 4 and Figure 3). The UmuDC expression from pGW2101 considerably
The incorporation of oxidatively damaged DNA precursors, 2-OH-dATP and 8-OH-dGTP, by DNA pols appears to be one of the major pathways of 2-OH-Ade and 8-OHGua formation in DNA, respectively, which would cause mutations and cancers. These damaged nucleotides are believed to be incorporated mainly by the replicative DNA pol, the DNA pol III holoenzyme (pol III HE) in E. coli, since pol III HE and the pol III catalytic subunit incorporate both damaged nucleotides with high error frequencies in vitro (7, 13, 36). Recently, the highly erroneous incorporation of 2-OH-dATP and 8-OH-dGTP in vitro by Y family DNA pols, such as human DNA pol η or archeal DNA pols P1 and P2, was reported (18). Thus, the E. coli Y family pols may also incorporate these damaged DNA precursors in vivo. In this experiment, we investigated the mutations induced by 2-OH-dATP and 8-OH-dGTP in E. coli strains deficient in the Y family pols, DNA pol IV and DNA pol V. In this study, 2-OH-dATP, but not 8-OH-dGTP, induced mutations less efficiently in the pol IV deficient strain than in the wt strain, suggesting that pol IV may contribute to the mutagenesis by 2-OH-dATP. Because the expression of the pol IV D101N mutant, which lacks the pol activity, in the pol IV deficient strain did not restore the lower mutagenesis level with 2-OH-dATP, the enhancement by pol IV is dependent on its pol activity. On the basis of these results, pol IV seems to facilitate the mutagenicity of 2-OH-dATP through (i) the misincorporation of 2-OH-dATP and/or (ii) the extension from the 3′-terminal 2-OH-dAMP residue. DNA pol IV incorporates 2-OH-dATP opposite G and T in the template at an almost 1:1 ratio in vitro (Shimizu, M., Gruz, P, and Nohmi, T. Unpublished results). Alternatively, pol IV may be involved in the fixation of mutations by (iii) the incorporation of dTTP opposite 2-OH-Ade during the second round of replication (TLS) and/or (iv) extension from the 3′-terminal dTMP residue incorporated opposite 2-OH-Ade. The transcriptional regulation of dinB by lexA is not very strict, so the cellular level of pol IV without SOS induction (250 molecules/cell) is somewhat higher than that of pol III (
