Tobacco-specific nitrosamine-derived O2-alkylthymidines are potent mutagenic lesions in SOS-induced Escherichia coli.

To investigate the biological effects of the O(2)-alkylthymidines induced by the tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), we have replicated a plasmid containing O(2)-methylthymidine (O(2)-Me-dT) or O(2)-[4-(3-pyridyl-4-oxobut-1-yl]thymidine (O(2)-POB-dT) in Escherichia coli with specific DNA polymerase knockouts. High genotoxicity of the adducts was manifested in the low yield of transformants from the constructs, which was 2-5% in most strains but increased 2-4-fold with SOS. In the SOS-induced wild type E. coli, O(2)-Me-dT and O(2)-POB-dT induced 21% and 56% mutations, respectively. For O(2)-POB-dT, the major type of mutation was T → G followed by T → A, whereas for O(2)-Me-dT, T → G and T → A occurred in equal frequency. For both lesions, T → C also was detected in low frequency. The T → G mutation was reduced in strains with deficiency in any of the three SOS polymerases. By contrast, T → A was abolished in the pol V(-) strain, while its frequency in other strains remained unaltered. This suggests that pol V was responsible for the T → A mutations. The potent mutagenicity of these lesions may be related to NNK mutagenesis and carcinogenesis.

The tobacco-specific nitrosamines 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and N′-nitrosonornicotine (NNN) are potent carcinogens in laboratory animals, inducing tumors at sites comparable to those found in smokers.[1,2] NNK is a potent lung carcinogen, but it also induces tumors in the liver, nasal cavity, and pancreas.[3,4] NNN induces tumors in the esophagus, nasal cavity, and respiratory tract.[1,5] Metabolic activation of both NNK and NNN by cytochrome P450 is required for their DNA binding, mutagenicity, and carcinogenicity.(1) NNK is metabolized to generate either a methylating agent or a pyridyloxobutylating agent, whereas NNN is metabolized only to the latter. The methylation pathway gives rise to multiple methyl (Me) adducts. 7-Me-dG and O6-Me-dG, have been identified in NNK-treated rodents,[2,6,7] but other methylation products,(8) including O2-Me-dC and O2-Me-dT, are also formed. O2-Me-pyrimidines are repaired in vitro by E. coli AlkA,[9,10] but otherwise, their biological properties are largely unknown. The pyridyloxobutylation pathway leads to four 4-(3-pyridyl)-4-oxobutyl (POB) adducts in vivo: O6-POB-dG, 7-POB-dG, O2-POB-dC, and O2-POB-dT.[2,11−13]

It is noteworthy that O6-POB-dG has been shown to be mutagenic in E. coli and mammalian cells,(14) but it is present in very low levels in NNK-treated A/J mice and rats,[15,16] In contrast, O2-POB-dT is the most persistent POB adduct in the lung and liver of male F344 rats.(16) When the mutagenicity of a model pyridyloxobutylating agent, 4-(acetoxymethylnitrosamino)-1-(3-pyridyl)-1-butanone (NNKOAc), was investigated in CHO cells, it induced point mutations primarily at AT base pairs, suggesting that O2-POB-dT might be mutagenic.(17)

In order to determine the replication properties of the two O2-alkyl-dT adducts formed by the methylation and pyridyloxobutylation pathway, we have constructed single-stranded pMS2 plasmids containing a single O2-Me-dT or O2-POB-dT (Chart 1 shows the structures), which were replicated in several isogenic strains of E. coli with specific DNA polymerase knockouts. The lesion repair capability of the strains remains unaltered, but single-stranded plasmids are inefficient substrates for DNA repair prior to the first round of replication. Viability was determined by a comparison of the colony-forming units obtained per microgram of the adducted construct relative to the control, which also reflected the lesion bypass efficiency. As shown in Figure 1, the yield of transformants from each adducted construct dropped significantly, with the bulkier O2-POB-dT being the more toxic. Upon induction of SOS, the yield of transformants increased about 2–4-fold in most E. coli strains. For example, in the wild type strain, transformants from the O2-Me-dT construct were 4.5 ± 0.7 and 15.1 ± 2.6% relative to the control, without and with SOS, respectively, whereas the O2-POB-dT construct generated 2.7 ± 0.2 and 6.7 ± 1.7% progeny, respectively, for the same.

Chart 1

Structures of O2-Methylthymidine and O2-Pyridyloxobutylthymidine

Figure 1

Viability of O2-Me-dT and O2-POB-dT without (open bars) and with (closed bars) SOS in different E. coli strains. The data represent the means and standard deviations of at least three independent experiments.

That the SOS polymerases are responsible for survival was confirmed in the strain that lacks pol II, pol IV, and pol V. The yield of transformants from both lesion-containing constructs was approximately 1% in this strain, either with or without prior UV irradiation of the host (Figure 1). We conclude that the O2-alkyl-dT adducts are replication blocking lesions, but increased TLS occurs with the SOS DNA polymerases. DNA alkylation products, including the extensively studied O6-alkyl-dG adducts, have been reported to partially block DNA synthesis.(18) Several other alkylated nucleosides, including 1-Me-dA, 3-Me-dC, 3-ethyl-dC, 1-Me-dG, and 3-Me-dT, are also blocks of DNA replication.(19) However, the blockages of the first three are completely removed in strains expressing AlkB, whereas the last two exhibited the strongest blocks.(19) Although these studies used different methods of analysis, the data in the current work taken together with the earlier studies imply that the O2-Me-dT and O2-POB-dT adducts are two of the strongest replication blocking DNA alkylation products.

To determine the frequency of miscoding, we analyzed the progeny plasmid by oligonucleotide hybridization followed by DNA sequencing. In the wild type strain, without SOS, 96–99% progeny contained a T at the O2-alkyl-dT site, indicating predominantly correct read-through by a DNA polymerase, most likely pol III (Supporting Information, Table S1 and S2). With SOS, mutation frequency (MF) increased to 21% and 56% for O2-Me-dT and O2-POB-dT, respectively, which indicates a high frequency of errors in TLS by the SOS DNA polymerases (Supporting Information, Table S1 and S2). Most mutations were targeted base substitutions, though a low frequency of targeted T deletions and semitargeted mutations also occurred (Supporting Information, Table S1 and S2). Figure 2 shows the relative population of each type of base substitution mutants relative to unaltered progeny in various SOS-induced strains, and it is apparent from this figure how each SOS DNA polymerase influences the mutational outcome. In contrast to high level of mutagenesis in the SOS-induced wild type strain, no mutants were isolated from the strain that lacks pol II, pol IV, and pol V. In the wild type strain, for O2-Me-dT, both T → G and T → A occurred at approximately 9% frequency (Figure 2 and Supporting Information, Table S1), but T → G was the dominant mutation at 37% compared to 12% T → A induced by O2-POB-dT (Figure 2 and Supporting Information, Table S2). For both lesions, T → A mutations were completely eliminated in the pol V-deficient strain, even though it remained approximately the same in pol II- and pol IV-deficient strains relative to the wild type (Figure 2 and Supporting Information, Table S1 and S2). The frequency of T → G, however, was reduced in each strain with a deficiency in any of the SOS polymerases, but the reduction was more pronounced in pol IV- and pol V-deficient strains. T → C mutations occurred only in the 3–5% and 5–7% frequency for O2-Me-dT and O2-POB-dT, respectively, but they dropped significantly in pol IV- and pol V-deficient strains. We conclude that for both lesions, T → A is induced by pol V, whereas all three SOS DNA polymerases contribute to T → G mutations. T → C mutations were likely induced by both pol IV and pol V.

Figure 2

Progeny analysis of the replication of O2-Me-dT and O2-POB-dT constructs in different E. coli strains with SOS. The bases A (black), G (blue), C (green), and T (red) at the lesion site show the percentage of each base substitution mutant and correct base, T.

To our knowledge, this is the first investigation of the replication of site-specific O2-alkylthymidines in a cell. However, in vitro replication studies of O2-ethyl-dT have been reported,[20,21] which showed that it blocks replication by T7 DNA polymerase and the Klenow fragment of the E. coli DNA polymerase I. These investigations also showed that incorporation of dA opposite O2-ethyl-dT inhibits DNA synthesis, whereas the DNA chain is more efficiently extended when dT is incorporated opposite the lesion. The current work in E. coli demonstrates that in the absence of TLS polymerases, bypass of O2-alkyl-dT is inefficient but accurate, whereas increased bypass by the TLS polymerases accompanies error-prone replication. Our study also suggests that pol V is the most error-prone of the three SOS polymerases and that only pol V incorporates dT opposite O2-alkyl-dT. Regardless of the types of mutations, our observation that O2-POB-dT is strongly mutagenic in E. coli underscores the risk posed by NNK and NNN since it is the most persistent adduct in experimental animals. Future studies of the repair kinetics of the O2-alkyl-dT in comparison to O6-alkyl-dG adducts in various organs may provide deeper insight into the roles of these mutagenic adducts in the carcinogenicity of NNK and NNN.