Transcriptional inhibition and mutagenesis induced by N-nitroso compound-derived carboxymethylated thymidine adducts in DNA.

N-nitroso compounds represent a common type of environmental and endogenous DNA-damaging agents. After metabolic activation, many N-nitroso compounds are converted into a diazoacetate intermediate that can react with nucleobases to give carboxymethylated DNA adducts such as N3-carboxymethylthymidine (N3-CMdT) and O(4)-carboxymethylthymidine (O(4)-CMdT). In this study, we constructed non-replicative plasmids carrying a single N3-CMdT or O(4)-CMdT, site-specifically positioned in the transcribed strand, to investigate how these lesions compromise the flow of genetic information during transcription. Our results revealed that both N3-CMdT and O(4)-CMdT substantially inhibited DNA transcription mediated by T7 RNA polymerase or human RNA polymerase II in vitro and in human cells. In addition, we found that N3-CMdT and O(4)-CMdT were miscoding lesions and predominantly directed the misinsertion of uridine and guanosine, respectively. Our results also suggested that these carboxymethylated thymidine lesions may constitute efficient substrates for transcription-coupled nucleotide excision repair in human cells. These findings provided important new insights into the biological consequences of the carboxymethylated DNA lesions in living cells.

INTRODUCTION

The human genome is constantly threatened by DNA-damaging agents including N-nitroso compounds (NOCs), which can form from red meat consumption and be present in tobacco smoke as well as other environmental and endogenous sources (1–5). Multiple lines of evidence support a strong link between the exposure to NOCs and an elevated risk of gastrointestinal cancer and other human diseases (1,3–7). Metabolic activation of many NOCs results in the generation of a common reactive intermediate, diazoacetate, which can induce carboxymethylation of nucleobases in DNA (8–12). It has been suggested that diazoacetate-induced DNA carboxymethylation, but not methylation, may lead to the signature mutations in p53 gene found in human gastrointestinal cancer (13,14). In this respect, a yeast-based functional assay revealed that diazoacetate-induced mutation spectrum in human p53 gene at non-CpG sites is strikingly similar to those observed in mutated p53 in human stomach and colorectal tumors (13). Recent studies showed that N3-carboxymethylthymidine (N3-CMdT) and O4-carboxymethylthymidine (O4-CMdT) are the major carboxymethylated adducts of thymidine formed in isolated DNA upon exposure to diazoacetate (Figure 1), and these lesions may contribute to p53 mutations at non-CpG sites induced by diazoacetate (8,13,15,16).

Figure 1.

Chemical structures of dT, N3-CMdT and O4-CMdT.

Understanding the biological consequences of DNA lesions in living cells necessitates the examination about how these lesions compromise the transmission of genetic information during DNA replication and transcription (17–20). In the respect, it has been shown that N3-CMdT and O4-CMdT are both blocking and miscoding during DNA replication in vitro and in Escherichia coli cells (15,16). In addition, E. coli DNA polymerase V or its yeast ortholog (i.e. polymerase η) is involved in translesion synthesis of N3-CMdT and O4-CMdT (15,16). To date, the effects of carboxymethylated thymidine adducts on DNA transcription have not yet been studied. Herein, we placed a single N3-CMdT or O4-CMdT at a defined position in the transcribed DNA strand of non-replicative double-stranded plasmids and assessed how these carboxymethylated DNA lesions perturb DNA transcription in vitro and in human cells.

MATERIALS AND METHODS

Materials and cell culture conditions

All enzymes, chemicals, [γ-32P]ATP and unmodified oligodeoxyribonucleotides (ODNs), unless otherwise specified, were purchased from New England BioLabs, Sigma-Aldrich, Perkin-Elmer and Integrated DNA Technologies, respectively. All siRNAs used in this study were obtained from Thermo Scientific Dharmacon: CSB ON-TARGETplus SMARTpool (L-004888), XPC ON-TARGETplus SMARTpool (L-016040) and non-targeting control siRNA (D-001210). The 293T human embryonic kidney epithelial cells (ATCC) and human fibroblast cells (XP12RO and GM15876A) (21) were grown at 37°C with 5% CO2 in Dulbecco's Modified Eagle's medium supplemented with 10% fetal bovine serum (Invitrogen) and 100 U/ml penicillin, and 100 μg/ml streptomycin (ATCC).

Transcription template preparation

The undamaged control and competitor plasmid substrates, which were derived from a non-replicating pTGFP-Hha10 plasmid (22,23), were constructed in a recent study (24). The N3- and O4-CMdT-bearing plasmids were prepared following a well-established method as described elsewhere (Figure 2a) (22,23,25). Briefly, two nicks were produced in the transcribed strand of undamaged control plasmid by Nt.BstNBI, and the resulting 25-mer single-stranded ODN was removed from the nicked plasmid by annealing with an ‘excess amount’ of complementary ODN. The gapped vector was purified and ligated with a 13-mer ODN (5′-AATTGAGTCGATG-3′) and a 12-mer lesion-containing ODN (5′-ATGGCGXGCTAT-3′, X = N3-CMdT or O4-CMdT) (8). The ligated products were incubated with ethidium bromide and the supercoiled lesion-bearing constructs were gel purified as described previously (25). The lesion-bearing or undamaged control plasmids were premixed with the competitor vector at a molar ratio of 3:1 (lesion or control/competitor) and used as DNA templates for all transcription assays in this study.

Figure 2.

Experimental outline. (a) A schematic diagram illustrating the procedures for the construction of the plasmids harboring a site-specifically incorporated N3-CMdT or O4-CMdT. (b) Strategy for assessing the impact of the CMdT lesions on DNA transcription. The non-lesion competitor vector has the same sequence as the lesion-bearing vector except for three more bases near the lesion region. ‘X’ indicates dT, N3-CMdT or O4-CMdT, and the arrowheads indicate the +1 transcription start sites of the cytomegalovirus (CMV) and T7 promoters.

Transcription assay

In vitro transcription reactions mediated by T7 RNA polymerase (T7 RNAP) or human RNA polymerase II (hRNAPII) as well as transcription assays in human cells were conducted as described elsewhere (24). The human cells used in this study included XPA-deficient (XP12RO) and XPA-complemented (GM15876A) cell lines, and 293T cells that were treated with non-targeting control or siRNAs targeting CSB or XPC.

RNA extraction and reverse transcription-polymerase chain reaction

The RNA products arising from in vitro and in vivo transcription were extracted with Total RNA Kit I (Omega) and were further processed using a DNA-free kit (Ambion) following the manufacturer's instructions. The transcripts of interest were then reverse transcribed and polymerase chain reaction (PCR) amplified as described previously (23). The efficiency of siRNA knockdown was evaluated by real-time reverse transcription-PCR (RT-PCR) as described previously (23).

Polyacrylamide gel electrophoresis analysis

We employed NcoI/SfaNI-mediated restriction digestion/postlabeling assay, as described recently (24), to resolve the 13-mer non-mutagenic product d(CATGGCGAGCTAT) from the DNA fragments d(CATGGCGCGCTAT) and d(CATGGCGTGCTAT), which correspond to the transcription products carrying A→C and A→U mutations opposite the lesion, respectively (Figure 3a and b). Briefly, the RT-PCR products were treated with NcoI and shrimp alkaline phosphatase. After heat inactivation, the dephosphorylated restriction fragments were radiolabeled with the use of [γ-32P]ATP and T4 polynucleotide kinase (T4 PNK) and were further digested with SfaNI. The products were separated by 30% native acrylamide:bisacrylamide (19:1) gel and analyzed by a phosphorimager (23,25). Similarly, we employed MluCI/Cac8I-mediated restriction digestion/postlabeling assay to separate the 10-mer fragment d(AATTATAGCG) (i.e. with A→G mutation) from the corresponding non-mutagenic product or the products with an A→C or A→U mutation (Figure 2c and d). The relative bypass efficiency (RBE) and the frequency of base misincorporation were determined as described previously (23,26).

Figure 3.

The effects of N3-CMdT and O4-CMdT on DNA transcription mediated by T7 RNAP and HeLa nuclear extract (hRNAPII) in vitro. (a) Sample processing for NcoI/SfaNI-mediated restriction digestion/postlabeling assay (P* indicates a 32P-labeled phosphate group). (b) Representative gel images showing the NcoI/SfaNI-produced restriction fragments of interest. The restriction fragment arising from the competitor vector, i.e. d(CATGGCGATATGCTAT), is designated as ‘16mer-Comp’, ‘13mer-C’, ‘13mer-A’, ‘13mer-G’ and ‘13mer-T’ represent the standard synthetic ODNs d(CATGGCGNGCTAT), where ‘N’ is C, A, G and T, respectively. (c) Sample processing for MluCI/Cac8I-mediated restriction digestion/postlabeling assay. (d) Representative gel images showing the MluCI/Cac8I-generated restriction fragments of interest. ‘10mer-C’, ‘10mer-A’, ‘10mer-G’ and ‘10mer-T’ represent the standard synthetic ODNs d(AATTATAGCM), where ‘M’ is C, A, G and T, respectively. (e) The RBE values of N3-CMdT and O4-CMdT in in vitro transcription systems. (f and g) Mutagenic properties of N3-CMdT (f) and O4-CMdT (g) in in vitro transcription systems. The data represent the mean and standard error of results from three independent experiments.

Liquid chromatography-tandem mass spectrometry analysis

We employed liquid chromatography-tandem mass spectrometry (LC-MS/MS) for the identification of transcription products, as described elsewhere (24). Briefly, the RT-PCR products were treated with SfaNI and shrimp alkaline phosphatase. After heat inactivation, the dephosphorylated restriction fragments were further digested by NcoI. The proteins in the restriction mixture were removed by extraction with phenol/chloroform/isoamyl alcohol (25:24:1, v/v), and the aqueous portion was desalted with HPLC, and subjected to LC-MS/MS analysis (23,25).

RESULTS

We employed a well-established competitive transcription and adduct bypass assay (23) to systematically examine how N3-CMdT and O4-CMdT compromise the efficiency and fidelity of DNA transcription in vitro and in human cells. For this purpose, we incorporated site-specifically a single N3-CMdT or O4-CMdT into a non-replicative double-stranded plasmid using a previously described method (Figure 2a) (22,23,25). The lesions were placed on the template strand 39 and 58 nucleotides downstream from the transcriptional start sites of the T7 promoter and cytomegalovirus promoter, respectively (Figure 2b). We premixed the lesion-carrying or undamaged control plasmid with a lesion-free competitor construct and used them as DNA templates for in vitro or in vivo transcription. The transcripts of interest were amplified using RT-PCR, and the resulting RT-PCR products were digested with suitable restriction enzymes. The DNA fragments released from the restriction digestion reactions were subjected to polyacrylamide gel electrophoresis (PAGE) and LC-MS/MS analyses (Figure 2b).

We performed the in vitro transcription assays using single-subunit T7 RNA polymerase (T7 RNAP) or multisubunit human RNA polymerase II (hRNAPII) in HeLa cell nuclear extract. In this regard, because of the structural homology between T7 RNAP and eukaryotic mitochondrial RNAPs (27), we chose T7 RNAP as a model to assess the potential effects of N3-CMdT and O4-CMdT on mitochondrial transcription in eukaryotic cells. After in vitro transcription, the RNA products were purified, reverse-transcribed and amplified with PCR.

By digesting the RT-PCR products with two sets of restriction enzymes, i.e. NcoI/SfaNI and MlucI/Cac8I, we were able to unequivocally distinguish the restriction fragments with a single nucleotide difference at the lesion site by PAGE analysis (Figure 3a–d).

Our results showed that both N3-CMdT and O4-CMdT markedly inhibited DNA transcription catalyzed by T7 RNAP in vitro, with the RBE values being ∼46 and 29%, respectively (Figure 3e). Similarly, N3-CMdT and O4-CMdT exhibited strong inhibitory effects on the hRNAPII-mediated transcription in vitro, with the RBE values being ∼26 and 19%, respectively (Figure 3e). Moreover, we found that N3-CMdT was highly mutagenic during in vitro transcription mediated by T7 RNAP and hRNAPII, with uridine misincorporation opposite the lesion (A→U) occurring at frequencies of ∼83 and 70%, respectively (Figure 3f). Transcriptional bypass of O4-CMdT by T7 RNAP and hRNAPII was also mutagenic, with guanosine misincorporation opposite the lesion (A→G) occurring at frequencies of ∼33 and 17%, respectively (Figure 3g). The identities of the above mutant products were also confirmed by LC-MS/MS analysis (Supplementary Figures S1 and S2).

We next asked how N3-CMdT and O4-CMdT perturb the efficiency and fidelity of DNA transcription in human cells. To this end, we co-transfected the lesion-containing or lesion-free control plasmids together with a competitor vector into XPA-deficient XP12RO and XPA-complemented GM15876A cells. After a 24-h incubation, we extracted the RNA from the cells and analyzed the transcription products as described above (Figures 2b and 3a and c).

Our results showed that N3-CMdT and O4-CMdT substantially impeded DNA transcription in GM15876A cells, with the RBE values being ∼33 and 37%, respectively (Figure 4a and Supplementary Figure S4a). Moreover, we observed significantly lower RBE values for both N3-CMdT and O4-CMdT in XPA-deficient XP12RO cells (Figure 4a and Supplementary Figure S4a). Transcriptional bypass of N3-CMdT was observed to induce A→U mutation opposite the lesion site at a frequency of ∼40% in GM15876A cells, whereas such mutation occurred at a significantly higher frequency (∼65%) in XP12RO cells (Figure 4b and Supplementary Figure S4a and b). We also found that O4-CMdT induced a markedly higher degree of A→G mutation in XP12RO cells than in GM15876A cells (Figure 4c and Supplementary Figure S4a and b). These results indicated that XPA, a core component of the nucleotide excision repair (NER) pathway (28,29), is involved in the removal of N3-CMdT and O4-CMdT in human cells.

Figure 4.

The effects of N3-CMdT and O4-CMdT on DNA transcription in human cells. (a) The RBE values of N3-CMdT and O4-EtdT in XPA-deficient XP12RO and XPA-complemented GM15876A cells. (b and c) Mutagenic properties of N3-CMdT (b) and O4-CMdT (c) in XP12RO and GM15876A cells. (d) The RBE values of N3-CMdT and O4-CMdT in HEK293T cells treated with control non-targeting, CSB or XPC siRNAs. (e and f) Mutagenic properties of N3-CMdT (e) and O4-CMdT (f) in HEK293T cells treated with CSB or XPC siRNAs. The data represent the mean and standard error of results from three independent experiments. ‘*’, P < 0.05; ‘**’, P < 0.01; ‘***’, P < 0.001. The P values were calculated by using unpaired two-tailed Student's t-test.

Because CSB and XPC are specifically required for transcription-coupled NER (TC-NER) and global-genome NER (GG-NER), respectively (28,29), we investigated the effects of their depletion on transcriptional alternations induced by N3-CMdT and O4-CMdT in human cells. Our results showed that the siRNA-mediated downregulation of CSB resulted in a significant decrease in transcriptional bypass efficiency of N3-CMdT in human 293T cells (Figure 4d and Supplementary Figures S4 and S5a). In addition, depletion of CSB caused a significant elevation in transcriptional mutagenesis induced by N3-CMdT, with the frequency of A→U mutation being increased from ∼23% in control siRNA-treated cells to ∼39% in CSB-knockdown cells (Figure 4e and Supplementary Figure S5a and b). Similarly, we found that siRNA knockdown of CSB elicited a significant change in the effects of O4-CMdT on the efficiency and fidelity of transcription in human 293T cells (Figure 4d and f and Supplementary Figure S5a and b). On the other hand, siRNA-mediated downregulation of XPC did not significantly alter the transcriptional bypass efficiency or mutation frequency of N3-CMdT and O4-CMdT in 293T cells (Figure 4e and f and Supplementary Figures S4 and S5a and b). Taken together, these results indicated that TC-NER may play a major role in the removal of N3-CMdT and O4-CMdT from the transcribed strand of active genes in human cells.

DISCUSSION

Accurate and efficient transmission of genetic information is essential for maintaining normal cellular functions (17–20,30). Unrepaired DNA damage, apart from perturbing DNA replication, may also elicit cytotoxic effects by blocking transcription and/or confer genome instability by inducing aberrant transcripts during DNA transcription (17,18,30,31). Notably, emerging evidence suggests that transcriptional mutagenesis may cause stable phenotypic changes in cells and may ultimately contribute to carcinogenesis or other adverse health effects (17,32–35). Due to the abundant presence of carboxymethylating agents in the environment and within cells, carboxymethylation of nucleobases in DNA is generally unavoidable (2–4,8–12). In this vein, O6-carboxymethyl-2′-deoxyguanosine (O6-CMdG) was detected, by an immunoslot blot assay, at a level of three to eight lesions per 107 nucleosides in the blood DNA of healthy human subjects (12), and humans consuming red meat have elevated levels of O6-CMdG in colonic exfoliated cells than control group consuming vegetarian diet (36). Hence, it is important to examine how carboxymethylated DNA lesions compromise the efficiency and fidelity of DNA transcription.

It has been previously shown that two carboxymethylated thymidine lesions, i.e. N3-CMdT and O4-CMdT, are strong blockages to DNA replication machinery in E. coli cells and primarily induced T→A and T→C mutations at the lesion sites, respectively (16). In keeping with these findings, we showed here that both N3-CMdT and O4-CMdT markedly inhibited DNA transcription mediated by T7 RNA polymerase or human RNA polymerase II in vitro. In addition, we found that N3-CMdT was a strong miscoding lesion and predominantly induced mutant transcripts containing a uridine opposite the lesion, whereas O4-CMdT was a moderately miscoding lesion that exclusively induced misinsertion of guanosine opposite the lesion during transcription in vitro.

We further demonstrated that both N3-CMdT and O4-CMdT substantially impeded DNA transcription in human cells. In addition, we found that N3-CMdT and, to a much lesser degree, O4-CMdT were able to induce transcriptional mutagenesis in human cells. Moreover, our results revealed that depletion of TC-NER components could further exacerbate the deleterious effects of these carboxymethylated DNA lesions on transcriptional efficiency and fidelity in cells. In this respect, it was observed that NER-deficient cells are significantly more sensitive to carboxymethylating agents than repair-proficient cells, suggesting that one or more unidentified carboxymethylated DNA adduct(s) may serve as potential substrate(s) for NER, thereby conferring cytotoxic effects that are selective toward NER-deficient cells (37). Our results indicate that N3-CMdT and O4-CMdT may constitute such carboxymethylated DNA lesions that elicit elevated sensitivities of NER-deficient cells toward carboxymethylating agents. In addition, previous studies showed that TC-NER plays an important role in the repair of several non-bulky lesions, including abasic sites, in cells (23,24,38,39). Our present work added N3-CMdT and O4-CMdT to the pool of non-bulky DNA lesions that partially inhibit transcription and may trigger TC-NER, which is a subpathway of NER that efficiently removes transcription-blocking lesions in cells (24,28,29).

The N3 and O4 positions of thymine are among the major sites for DNA modification induced by alkylating agents in vitro (40,41); it is therefore of interest to compare the similar and distinct effects of N3-CMdT, O4-CMdT and their structurally related alkylated lesions on the efficiency and fidelity of DNA transcription (24). Similar as N3-CMdT, it was recently reported that N3-ethylthymidine (N3-EtdT) is both blocking and miscoding during transcription and is subjected to TC-NER in human cells (24). While N3-CMdT directs specifically the misinsertion of uridine, N3-EtdT was found to display promiscuous miscoding properties during transcription (24). The observed difference may be attributed to the unique chemical properties of these two lesions. In this context, the deprotonated carboxyl group in N3-CMdT may foster hydrogen-bonding interaction with the hydrogen atom on N3 of uridine, thereby promoting the misincorporation of uridine opposite the lesion, whereas the presence of an alkyl functionality on the N3 position of thymine may abolish the Watson–Crick hydrogen-bonding property, and thus the base-pairing capabilities of the nucleobase (24,42,43). On the other hand, both O4-CMdT and O4-EtdT lesions were observed to direct exclusively the misincorporation of guanosine opposite the lesion sites during DNA transcription; however, unlike O4-CMdT, O4-EtdT is not a strong blockage to DNA transcription machinery and cannot be efficiently repaired by TC-NER in human cells (24).

In summary, we investigated, for the first time, how N3-CMdT and O4-CMdT perturb the efficiency and fidelity of DNA transcription mediated by T7 RNA polymerase in vitro or human RNA polymerase II in vitro and in human cells. Our results also suggested that these carboxymethylated DNA lesions may constitute efficient substrates for TC-NER in cells. These findings provide important new insights into the biological consequences of DNA damage induced by carboxymethylating agents in living cells. In this vein, it is worth noting that a more complete understanding about the biological relevance of N3-CMdT, O4-CMdT and other carboxymethylated DNA lesions also requires a rigorous assessment about their formation in human cells and tissues. Our laboratory is in the process of developing LC-MS/MS-based methods for examining the occurrence of these lesions in cultured mammalian cells.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.