Identification of three major DNA adducts formed by the carcinogenic air pollutant 3-nitrobenzanthrone in rat lung at the C8 and N2 position of guanine and at the N6 position of adenine
Article first published online: 5 DEC 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 118, Issue 9, pages 2139–2146, 1 May 2006
How to Cite
Arlt, V. M., Schmeiser, H. H., Osborne, M. R., Kawanishi, M., Kanno, T., Yagi, T., Phillips, D. H. and Takamura-Enya, T. (2006), Identification of three major DNA adducts formed by the carcinogenic air pollutant 3-nitrobenzanthrone in rat lung at the C8 and N2 position of guanine and at the N6 position of adenine. Int. J. Cancer, 118: 2139–2146. doi: 10.1002/ijc.21622
- Issue published online: 21 FEB 2006
- Article first published online: 5 DEC 2005
- Manuscript Accepted: 15 SEP 2005
- Manuscript Received: 30 MAY 2005
- Cancer Research UK
- DNA adducts;
- air pollution;
- diesel exhaust
3-Nitrobenzanthrone (3-NBA) is a potent mutagen and potential human carcinogen identified in diesel exhaust and ambient air particulate matter. Previously, we detected the formation of 3-NBA-derived DNA adducts in rodent tissues by 32P-postlabeling, all of which are derived from reductive metabolites of 3-NBA bound to purine bases, but structural identification of these adducts has not yet been reported. We have now prepared 3-NBA-derived DNA adduct standards for 32P-postlabeling by reacting N-acetoxy-3-aminobenzanthrone (N-Aco-ABA) with purine nucleotides. Three deoxyguanosine (dG) adducts have been characterised as N-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-C8-N-ABA), 2-(2′-deoxyguanosin-N2-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-N2-ABA) and 2-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-C8-C2-ABA), and a deoxyadenosine (dA) adduct was characterised as 2-(2′-deoxyadenosin-N6-yl)-3-aminobenzanthrone-3′-phosphate (dA3′p-N6-ABA). 3-NBA-derived DNA adducts formed experimentally in vivo and in vitro were compared with the chemically synthesised adducts. The major 3-NBA-derived DNA adduct formed in rat lung cochromatographed with dG3′p-N2-ABA in two independent systems (thin layer and high-performance liquid chromatography). This is also the major adduct formed in tissue of rats or mice treated with 3-aminobenzanthrone (3-ABA), the major human metabolite of 3-NBA. Similarly, dG3′p-C8-N-ABA and dA3′p-N6-ABA cochromatographed with two other adducts formed in various organs of rats or mice treated either with 3-NBA or 3-ABA, whereas dG3′p-C8-C2-ABA did not cochromatograph with any of the adducts found in vivo. Utilizing different enzymatic systems in vitro, including human hepatic microsomes and cytosols, and purified and recombinant enzymes, we found that a variety of enzymes [NAD(P)H:quinone oxidoreductase, xanthine oxidase, NADPH:cytochrome P450 oxidoreductase, cytochrome P450s 1A1 and 1A2, N,O-acetyltransferases 1 and 2, sulfotransferases 1A1 and 1A2, and myeloperoxidase] are able to catalyse the formation of 2-(2′-deoxyguanosin-N2-yl)-3-aminobenzanthrone, N-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone and 2-(2′-deoxyadenosin-N6-yl)-3-aminobenzanthrone in DNA, after incubation with 3-NBA and/or 3-ABA. © 2005 Wiley-Liss, Inc.
Environmental factors and individual susceptibility play an important role in many human cancers.1 Epidemiological studies have shown that exposure to diesel exhaust and ambient air particulate matter is associated with an increased risk of lung cancer.2, 3, 4, 5 Although traditional industrial emission levels are tending to decrease in Western countries, vehicular exhaust is becoming an increasing problem.6 The particulate phase of diesel exhaust is the predominant source of human exposure to nitrated polycyclic aromatic hydrocarbons (nitro-PAHs),7 many of which are mutagens and rodent carcinogens.2, 8 3-Nitrobenzanthrone (3-NBA, 3-nitro-7H-benz[de]anthracen-7-one; Fig. 1) was discovered in diesel exhaust and in airborne particulate matter.9, 10, 11, 12, 13 It has also been detected in surface soil and rainwater,14, 15, 16 probably as a consequence of atmospheric washout. More recently, 3-NBA has been detected in domestic coal-burning-derived particulates,17 suggesting that it is a ubiquitous environmental contaminant. 3-NBA is genotoxic in various short-term bioassays,9, 18, 19, 20, 21 being one of the most potent direct-acting mutagens ever recorded in the Salmonella typhimurium Ames assay.9 Its genotoxicity has been further documented by the detection of specific DNA adducts formed in various chemical and enzymatic in-vitro assays, in cells and in vivo in rodents treated with 3-NBA.22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35 The main metabolite of 3-NBA, 3-aminobenzanthrone (3-ABA, Fig. 1), has been found in urine samples of underground salt mine workers occupationally exposed to diesel emissions,10 and this metabolite forms the same DNA adducts as those formed by its nitroaromatic counterpart 3-NBA.28, 29, 36, 37 Human exposure to 3-NBA is thought to occur primarily via the respiratory tract. It has been shown recently that the uptake of 3-NBA by the lungs of rats treated with 3-NBA by intratracheal instillation results in the formation of high levels of specific DNA adducts in several organs and also in blood.33 Furthermore, 3-NBA is carcinogenic in rats after intratracheal administration,38 suggesting that 3-NBA is a potential human carcinogen.
3-NBA requires metabolism to reactive electrophilic species to exert its genotoxic activity. The activation of 3-NBA to N-hydroxy-3-aminobenzanthrone (N-OH-ABA, Fig. 1) is catalysed primarily by cytosolic reductases, in particular NAD(P)H:quinone oxidoreductase (NQO1) and xanthine oxidase (XO), although microsomal NADPH:cytochrome P450 oxidoreductase (POR) can also be involved in the reduction of 3-NBA.25, 26, 30, 31, 32N-OH-ABA formed by nitroreduction can be further activated by phase II enzymes, such as N,O-acetyltransferases (NATs) and sulfotransferases (SULTs), leading to the formation of reactive N-acetoxy- or sulfooxy esters capable of reacting with DNA to form adducts.27, 29, 32 It has been established that the multiple DNA adducts formed in vivo are derived from purine bases reacted with reductive metabolites of 3-NBA, which do not posses an N-acetyl group.21, 26, 28, 32, 33 Although analyses by 32P-postlabeling of in-vitro and in-vivo 3-NBA-derived DNA adducts have shown that they contain deoxyadenosine (dA) and deoxyguanosine (dG),26, 30 the definitive structural characterisation of these adducts has not been reported.35
We have now prepared 3-NBA-derived DNA adducts by reacting the putative ultimate reactive metabolite, N-acetoxy-3-aminobenzanthrone (N-Aco-ABA, Fig. 1) with purine nucleotides, and used them as authentic standards for the 32P-postlabeling assay to determine the identity of adducts formed in vivo in rodents. Three dG adducts and one dA adduct have been characterised. We have compared their chromatographic properties with those of the adducts formed by 3-NBA in vivo and in vitro, and have thereby been able to identify the structure of three of the latter. We also used specific DNA binding detected by 32P-postlabeling as an end point for studying enzymes involved in the formation of these adducts by 3-NBA and its human metabolite 3-ABA.
Material and methods
3-NBA was synthesised as reported previously.9, 22, 27 3-ABA was synthesised as described.29N-OH-ABA was prepared as reported.22, 32 The authenticity of 3-NBA and 3-ABA was confirmed by UV, liquid chromatography–mass spectrometry (LC-MS), and high field proton nuclear magnetic resonance (NMR) spectroscopy. The authenticity of N-OH-ABA and N-Aco-ABA was confirmed by UV and LC-MS analyses.
LC-MS measurements were made on a Waters ZQ 2000 Mass Spectrometer, equipped with an electrospray ion source (ESI) and Agilent 1100 high performance liquid chromatography (HPLC) system. Electrospray ionisation was set at 3.0 kV capillary-voltage with nitrogen used as the sheath gas and set at 350 l/hr for desolvation gas flow and 350°C for desolvation gas temperature. Data acquisition was performed in sequential positive/negative ion modes with a cone voltage of 30 V and cone gas flow of 50 l/hr. Ion source temperature was set at 350°C. The electron multiplier was operated at 650 V for all analyses, and HPLC–UV chromatograms were recorded at 260 nm. 32P-Postlabeling HPLC analysis was carried out with a Waters 2690 HPLC system.
Preparation and characterisation of 2-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-C8-C2-ABA),N-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-C8-N-ABA) and 2-(2′-deoxyguanosin-N2-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-N2-ABA)
Standard samples of 2′-deoxyguanosine-3′-phosphate (dG3′p) adducts of 3-NBA were prepared by reacting N-Aco-ABA with dG3′p. N-Aco-ABA was prepared from N-OH-ABA by methods reported previously with slight modifications.39, 40, 41 Briefly, to a solution of N-OH-ABA (0.01 mmol) in tetrahydrofuran (400 μl) was added 10 mg of polymer bound 1,5,7-triazobicycle[4.4.0]dec-5-ene (TBD; TBD-methyl polystyrene 2% divinylbenzene; Novabiochem, Germany), followed by 5 μl of pyruvonitrile (Aldrich Chemical Co., WI). After the reaction was stirred at room temperature for 10 min, 10 μl of methanol was added and the solution was filtered. The filtrate was then injected directly into a prepared solution (200 μl) of dG3′p (0.02 mmol) in 100 mM sodium phosphate buffer, pH 7.4. If precipitation occurred, dimethylformamide (200 μl) was further added. The reaction mixture was then heated at 37°C for 24 hr and extracted twice with chloroform. The aqueous layer was then subjected to HPLC equipped with a photodiode array detector where a column of Cosmosil-AR II (4.6 × 250 mm, Nacalai, Kyoto, Japan) was used with the elution system of 18% methanol in 0.25% of triethylamine-acetate, pH 7.0, at 40°C column temperature. The desired compounds were eluted at retention times of 12 min for 2-(2′-deoxyguanosin-N2-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-N2-ABA), 16 min for 2-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-C8-C2-ABA) and 21 min for N-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone-3′-phosphate (dG3′p-C8-N-ABA) (Fig. 2a). Identification of these specimens was done with direct comparison of UV spectra of the authentic samples without a phosphate moiety synthesized independently (compare Fig. 4). The synthesis of dG-C8-C2-ABA and dG-C8-N-ABA was reported previously.22, 34 Details concerning the structural identification of dG-N2-ABA will be published elsewhere (Takamura-Enya et al., manuscript in preparation). Yields of dG3′p-N2-ABA, dG3′p-C8-C2-ABA and dG3′p-C8-N-ABA based on the initial amount of N-OH-ABA were about 0.6%, 1.1% and 5.5%, respectively. The yields were estimated from the absorbance coefficient from 3-ABA at 500 nm. MS analysis of the eluate showed that dG3′p-N2-ABA, dG3′p-C8-C2-ABA and dG3′p-C8-N-ABA had the protonated molecular ion of m/z 591 with a daughter ion of m/z 395 arising from neutral loss of the deoxyribose-3′-phosphate moiety (Fig. 2b). Samples of each nucleotide was enzymatically digested by bacterial alkaline phosphatase (Sigma) in 5 mM Tris-HCl buffer (pH 7.9) containing 10 mM sodium chloride, 2 mM magnesium chloride and 0.2 mM dithiothreitol at 37°C for 4 hr and then subjected to HPLC (Fig. 5).
Preparation and characterisation of 2-(2′-deoxyadenosin-N6-yl)-3-aminobenzanthrone-3-phosphate (dA3′p-N6-ABA)
Synthetic methods were the same as those of dG3′p-ABA adducts described earlier except that 2′-deoxyadenosine-3′-phosphate (dA3′p; 0.02 mmol) was used instead of dG3′p. HPLC conditions were a Cosmosil-C18-AR II column (4.6 × 250 mm, Nacalai, Kyoto, Japan) with a linear gradient of 15–30% acetonitrile in 0.25% of triethylamine-acetate pH 7.0 over the course of 30 min. The desired 2-(2′-deoxyadenosine-N6-yl)-3-aminobenzanthrone-3′-phosphate (dA3′p-N6-ABA) was eluted at a retention time of 19 min (Fig. 3a), and showed the same UV spectrum as the independently synthesised dA-N6-ABA (compare Fig. 4). Details concerning the structural identification of dA-N6-ABA will be published elsewhere (Takamura-Enya et al., manuscript in preparation). Yields of dA3′p-N6-ABA were 0.5% calculated from initial amount of N-OH-ABA. The yields were estimated as described earlier. LC-MS of this compound showed the predicted m/z 575 [M + 1]+ involved with a daughter ion of m/z 379 (Fig. 3b). The nucleotide was enzymatically digested by bacterial alkaline phosphatase as described earlier and analysed by HPLC (Fig. 5).
Animal experiments with 3-NBA and its metabolites
Rats were treated as follows: (i) Female Sprague-Dawley rats (n = 3) were treated with a single dose of 2 mg/kg b.w. 3-NBA by intratracheal instillation and sacrificed 48 hr after administration,33 (ii) female Sprague-Dawley rats (n = 3) were treated with a single dose of 2 mg/kg b.w. 3-NBA by gavage and sacrificed 4 hr after treatment and26 (iii) female Wistar rats (n = 3) were treated with a single i.p. dose of 2 mg/kg b.w. 3-NBA, 3-ABA, 3-Ac-ABA or N-Ac-N-OH-ABA and sacrificed 24 hr after treatment.28 Mice were treated as follows: (i) Male C57BL/6 mice (n = 3) were treated with a single i.p. dose of 2 mg/kg b.w. 3-NBA or 3-ABA and sacrificed 24 hr after administration and32, 37 (ii) male Muta™ Mouse animals (n = 5) were treated i.p. with 25 mg/kg b.w. 3-NBA once a week for 4 weeks and sacrificed 3 days after the last treatment.21 In all cases, control rats and mice were treated with vehicle, tricaprylin, only. Lung and liver tissues were collected and stored at −80°C until DNA isolation.
Enzyme incubations with 3-NBA and its metabolites
To resolve which enzymes are able to catalyse the formation of 2-(2′-deoxyadenosin-N6-yl)-3-aminobenzanthrone (dA-N6-ABA), 2-(2′-deoxyguanosin-N2-yl)-3-aminobenzanthrone (dG-N2-ABA), N-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone (dG-C8-N-ABA) and/or 2-(2′-deoxyguanosin-8-yl)-3-aminobenzanthrone (dG-C8-C2-ABA) induced by 3-NBA and its metabolites (3-ABA, 3-Ac-ABA and/or N-Ac-N-OH-ABA) different experimental approaches were used according to previously described procedures: (i) human B-lymphoblastoid MCL-5 cells expressing native and human recombinant cytochrome P450s (CYPs);20, 30, 36 (ii) rat and human hepatic microsomes;30, 36 (iii) rat and human hepatic cytosols;32 (iv) purified and recombinant enzymes (human NQO1, human myeloperoxidase (MPO), rat CYP1A1 and rat NQO1);32, 36, 37 (v) human recombinant NQO1 in combination with human recombinant NATs or SULTs expressed in a heterologous baculovirus expression systems (Supersomes);32 (vi) genetically engineered Chinese hamster lung V79 cells expressing human recombinant CYPs, POR, NATs and/or SULTs27, 29, 30, 31, 36 and (vii) human recombinant CYPs and/or POR in Supersomes.30, 36
Preparation of reference compounds for 32P-postlabeling
As reference compounds, dA3′p and dG3′p (4 μmol/ml) were incubated with 3-NBA (300 μM) activated by buttermilk xanthine oxidase (XO, Sigma, UK; 1 U/ml) in the presence of hypoxanthine (Sigma, UK).26
32P-Postlabeling thin layer chromatography (TLC) analysis
DNA adduct analysis using the butanol extraction enrichment version of the 32P-postlabeling assay and TLC were performed as described recently.27 Enrichment by butanol extraction has been shown to yield more adduct spots and a better recovery of 3-NBA-derived DNA adducts than using enrichment by nuclease P1 digestion. For resolution of DNA adduct spots, the following chromatographic conditions were used: D1, 1.0 M sodium phosphate, pH 6.0; D3, 4 M lithium formate, 7 M urea, pH 3.5 and D4, 0.8 lithium chloride, 0.5 M Tris, 8.5 M urea, pH 8.0. Detection and quantitation of DNA adducts were performed using a Packard Instant Imager (Dowers Grove, IL). DNA adduct spots were numbered as reported.26, 27, 32, 36 DNA adduct levels (RAL, relative adduct labeling) were calculated from the adduct cpm, the specific activity of [γ-32P]ATP and the amount of DNA (pmol of DNA-P) used.
HPLC analysis of 32P-labeled 3′,5′-deoxyribonucleoside bisphosphate adducts
Individual adduct spots detected by the 32P-postlabeling TLC assay were excised from the TLC plates, extracted with pyridinium formate buffer, and co-chromatographed with standard bisphosphate adduct.28 For resolution of DNA adducts, the following chromatographic conditions were used: a phenyl-modified reversed-phase column (Luna 5μ phenyl-hexyl, 4.6 × 150 mm, Phenomenex, UK) with a linear gradient of methanol (from 30 to 55% in 45 min) in aqueous 0.5 M sodium phosphate (pH 3.5) at a flow rate of 1 ml/min. Radioactivity eluting from this column was measured by monitoring Cerenkov radiation with a Flow Scintillation Analyzer (Packard, Dowers Grove, IL).
Preparation and characterisation of dG3′p-C8-C2-ABA,dG3′p-C8-N-ABA, dG3′p-N2-ABA and dA3′p-N6-ABA
Authentic adduct standards for 32P-postlabeling were prepared by reaction of N-Aco-ABA with nucleotides, with subsequent isolation by HPLC. Compounds were identified by LC-MS techniques and the direct comparison of their UV spectra with those of corresponding nucleosides, which were prepared independently. Because of the low stability of N-Aco-ABA, it was used immediately after preparation without further purification. For that purpose, a polymer-bound base was used to be removed by simple filtration from the reaction mixture. Under the conditions of TBD-methyl-polystyrene with pyruvonitrile, N-OH-ABA was smoothly converted to N-Aco-ABA within 10 min at room temperature, which was confirmed by a change of mobility in TLC analysis, using silica gel developed with chloroform/methanol (10:1). N-Aco-ABA, thus obtained, was allowed to react with nucleotides under the conditions of pH 7.4 buffered solution at 37°C. The reaction was monitored intermittently by HPLC and was found to be completed within 24 hr.
LC-MS analysis of the reaction mixture of dG3′p with N-Aco-ABA showed the presence of 3 prominent peaks (assigned p1–p3) at retention times of 12 (p1), 16 (p2) and 21 min (p3), which were not present in the control reaction separately prepared without dG3′p (Fig. 2a). The material in all these 3 peaks had a molecular ion of m/z 591, the presumed mass of a dG3′p adduct with 3-ABA (Fig. 2b). The UV spectra of these peaks (Fig. 4) were identical with those of independently synthesised dG nucleoside adducts of dG-N2-ABA, dG-C8-C2-ABA and dG-C8-N-ABA, respectively. The identity of dG3′p-N2-ABA, dG3′p-C8-C2-ABA and dG3′p-C8-N-ABA as the 3′-phosphate derivatives of dG-N2-ABA, dG-C8-C2-ABA and dG-C8-N-ABA, respectively, was further confirmed by HPLC analysis after converting them to the latter by dephosphorylation with alkaline phosphatase (Fig. 5). In the same reaction mixture, some minor peaks with a molecular ion of m/z 591 were also identified by LC-MS, indicating the presence of other structurally unidentified dG adducts.
In the reaction mixture of dA3′p with N-Aco-ABA, dA3′p-N6-ABA (assigned p4, retention time 19 min) with a protonated molecular ion of m/z 575 was clearly identified by LC-MS analysis (Fig. 3). The UV spectrum of dA3′p-N6-ABA (Fig. 4) was the same as that of the authentic specimen of dA-N6-ABA. The identity of dA3′p-N6-ABA as the 3′-phosphate derivative of dA-N6-ABA was further confirmed by HPLC analysis after converting it to the latter with alkaline phosphatase (Fig. 5). One other peak (16.8 min retention time) with protonated molecular ion m/z 575 accompanied by m/z 379 was also found, but its molecular structure is still unknown because of the isolation of insufficient material from the HPLC peak.
Identification of DNA adducts formed by 3-NBA and its metabolites in vivo
DNA adduct formation in lung and liver tissue of rats treated with a single dose of 3-NBA by intratracheal instillation,33 by gavage26 or by intraperitoneal injection28 was analysed employing the butanol enrichment version of the 32P-postlabeling method. As illustrated in Figure 6e, after intratracheal instillation, a cluster of 5 adducts (spots 1–5) was induced by 3-NBA.33 As reported before,26, 28 the same DNA adduct pattern was observed in tissue of rats treated either orally or intraperitoneally with 3-NBA. Similarly, these adducts were observed in mice treated with 3-NBA by intraperitoneal injection, either after single or multiple dosing.21, 32 Using the same approach as reported before,20, 26, 28, 33 cochromatographic analysis by HPLC confirmed that all major 3-NBA-derived DNA adducts found in vivo were derived from reductive metabolites of 3-NBA bound to either dA (spot 1 and 2) or dG (spots 3–5) and did not have an N-acetyl group (Figs. 6f and 6g). Essentially, the same DNA adduct pattern was induced by 3-ABA in rats and mice28, 37 as well as by 3-Ac-ABA or N-Ac-N-OH-ABA in rats.28
Aliquots of all standards were analysed directly by 32P-postlabeling. As shown in Figure 6, dG3′p-C8-C2-ABA (spot SG1; Fig. 6a), dG3′p-N2-ABA (spot SG2; Fig. 6b) and dA3′p-N6-ABA (spot SA1; Fig. 6d) each resulted in 1 major spot on TLC. Labeling of adduct standard dG3′p-C8-N-ABA resulted in 1 major (spot SG3) and 1 minor spot (spot SG4, Fig. 6c). To investigate whether these authentic adducts are formed in vivo in rodents treated with 3-NBA or its metabolites, comparative chromatographic analyses were performed on TLC. Of note, a recent study conducted in our laboratory showed that dG-C8-C2-ABA is not formed in vivo.35 Digests of DNA isolated from rodent tissue were spiked with aliquots of the authentic adduct standards and analysed by 32P-postlabeling. On TLC, spot SG2 was indistinguishable from adduct spot 3 found in vivo (data not shown). To confirm this observation with a second, independent chromatographic procedure, the individual spots were also subjected to HPLC analysis. Equal amounts of radioactivity from spot 3 and spot SG2 were mixed and analysed by HPLC (Fig. 7a). They eluted with the same retention time (25.5 min; Fig. 7a) thereby demonstrating that adduct 3 is dG-N2-ABA. Similarly, comparative analysis on TLC and HPLC showed that adduct spot SG2 is chromatographically indistinguishable from the dG-derived adduct 3 prepared in vitro with XO (Fig. 7a; compare Fig. 6g), which has been used as reference compound to detect 3-NBA-derived DNA adduct in previous studies by 32P-postlabeling.21, 26, 28, 33, 37
Spot SG3 was indistinguishable from adduct spot 4 found in vivo using both TLC and HPLC analysis (Fig. 7b; compare Fig. 6e). It is also clear from Figure 7b that adduct spot SG3 is chromatographically indistinguishable from the dG-derived adduct 4 prepared in vitro with XO (compare Fig. 6g). Similarly, we found that both on TLC and HPLC, spot SG4 is indistinguishable from adduct spot 5 found in vivo (Fig. 7c; compare Fig. 6e) and the dG-derived adduct 5 prepared with XO (Fig. 7c; compare Fig. 6g). Since C8-dG arylamine-DNA adducts can be instable under 32P-postlabeling conditions,42 it is not known which spot, adduct spot SG3 or SG4, represents the dG-C8-N-ABA adduct or is a degradation product of it. However, collectively these findings clearly demonstrate that dG-C8-N-ABA is formed in vivo.
Spot SA1 was chromatographically indistinguishable from adduct spot 1 found in vivo, both on TLC and HPLC (Fig. 7d; compare Fig. 6e), demonstrating that adduct 1 is identical to dA-N6-ABA. Similarly, 1 of the reference compounds prepared with XO, dA-derived adduct 1, was indistinguishable from spot SA1 (Fig. 7d; compare Fig. 6f).
Formation of dG-C8-N-ABA, dG-N2-ABA and dA-N6-ABA by human enzymes
Enzymes and enzyme systems capable of catalysing the formation of dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA are summarized in Figure 1.
The results of our study demonstrate for the first time that dA-N6-ABA (adduct spot 1), dG-N2-ABA (adduct spot 3) and dG-C8-N-ABA (adduct spots 4 and 5) are formed in DNA of rats and mice treated with 3-NBA and its metabolites (3-ABA, 3-Ac-ABA and N-Ac-N-OH-ABA) (Fig. 1). Authentic standards of these adducts for 32P-postlabeling were obtained by reacting the ultimate metabolite N-Aco-ABA with dA3′p or dG3′p and subsequent HPLC purification. Comparative analyses by 32P-postlabeling were performed on TLC and HPLC, two independent chromatographic systems. Although a variety of unusual dA and dG adducts were characterised previously from reaction with N-acetoxy-N-acetyl-3-aminobenzanthrone (N-Aco-N-Ac-ABA) as the reactive metabolite, these adducts were different to those found in vivo,22, 26, 35 indicating that DNA adducts derived from N-Aco-N-Ac-ABA and activation pathways reported earlier based onin-vitro synthesis might not entirely represent the situation in vivo.22, 35 Two of these adducts have been characterised as N-acetyl-3-amino-2-(2′-deoxyguanosin-8-yl)benzanthrone (dG-C8-C2-N-Ac-ABA) and an unusual 3-Ac-ABA adduct of dA, which involves a double linkage between adenine and benzanthrone (N1–C1, N6–C11b), creating a five-membered imidazo-type ring system.35 Although treatment of 3′-phosphates of both these novel adducts with alkali generated two authentic nonacetylated standards for 32P-postlabeling, these were also shown to differ from those formed in vivo.35 The latter dG3′p-C8-C2-ABA adduct was also obtained after reaction of N-Aco-ABA with dG3′p (present study).
Most of the adducts derived from nitroaromatics and aromatic amines that have been reported to date are dG adducts with the C8 position of dG linked to the N atom of the amine or, to a lesser extent, with the amino group of dG linked to an aromatic C atom.43, 44 The most abundant adduct formed in vivo both after treatment with 3-NBA or 3-ABA is dG-N2-ABA. Adduct dG-C8-N-ABA was also clearly identified in vivo. It has been shown in previous studies23, 26 that dG3′p-C8-N-ABA is sensitive to digestion with nuclease P1 (data not shown), which is also indicative of C8-dG arylamine-DNA adducts in the 32P-postlabeling assay;45 this property has been used as a presumptive test for the presence of nitro-PAHs in complex mixtures.46 The formation of C8- and N6-substituted dA derivatives has also been reported for other nitroaromatics and aromatic amines.43, 44 Although dA-N6-ABA was clearly detectable in vivo, the formation of N-(2′-deoxyadenosin-8-yl)-3-aminobenzanthrone (dA-C8-N-ABA) still needs to be explored. Unfortunately, the 3′-phosphate of dA-C8-N-ABA, as an authentic standard for 32P-postlabeling, could not be prepared by reaction of N-Aco-ABA with dA3′p. However, it is tempting to speculate that the as-yet-unidentified adduct spot 2 formed in vivo may be dA-C8-N-ABA.
The principal pathway of metabolic activation of 3-NBA in rodents appears to be reduction and esterification to give reactive acetate and sulfate esters of N-OH-ABA, which form covalent adducts with DNA.26, 28, 29, 32 Comparative analyses showed that dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA are also present in DNA modified in vitro with N-OH-ABA,30, 35 confirming that N-OH-ABA is a critical intermediate in the formation of these DNA adducts in vivo. In addition, preliminary data indicate that these adducts are also formed in DNA modified in vitro with N-Aco-ABA (Takamura-Enya et al., unpublished results). Determining the capability of humans to metabolise 3-NBA and understanding which human enzymes are involved in its activation are important in the assessment of individual susceptibility.30, 32 To resolve which human enzyme systems are able to catalyse the formation of dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA from 3-NBA and its main human metabolite 3-ABA, we used different experimental approaches to study the enzymology as described previously.20, 27, 29, 30, 31, 32, 36, 37 As outlined in Figure 1, POR and NQO1 are mainly responsible for nitroreduction of 3-NBA to form N-OH-ABA.29, 30, 32 Recent data indicate that in vivo 3-NBA is predominantly activated by cytosolic nitroreductases rather than microsomal POR.32 On the other hand, CYP1A1 and CYP1A2 are primarily responsible for N-hydroxylation of 3-ABA.36N-OH-ABA can be further activated by NAT1 and NAT2 as well as SULT1A1 and SULT1A2 to reactive acetate and sulfate esters capable of reacting with DNA to form covalent DNA adducts (Fig. 1).27, 29, 32 Moreover, recent data indicate that 3-ABA is also activated by different model peroxidases in vitro, e.g. human MPO.37 In the present study, we demonstrate that all these enzymes participate in the metabolic activation of 3-NBA and 3-ABA, and thus are capable of catalysing the formation of dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA. As discussed previously,27, 29, 30, 32, 36 genetic polymorphisms, levels of expression and activities of these enzymes in humans may contribute to an individual's susceptibility to 3-NBA and could be important determinants of a possible cancer risk of 3-NBA in humans.
DNA adduct formation is a critical event in mutagenesis and in the initiation of carcinogenesis.47 Besides its extreme mutagenic activity in the Salmonella typhimurium Ames assay,9 3-NBA is also an effective mutagen in human MCL-5 cells18 in which it forms dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA, as shown in the present study.20, 30 Moreover, recent data on the mutagenic specificity of 3-NBA indicate that it induces mainly G:C to T:A transversion mutations in the liver cII gene of transgenic mice (Muta™Mouse) after i.p. treatment.21 The increase in mutation frequency in the liver cII gene was associated with strong DNA binding by 3-NBA in liver DNA (70–80% of total DNA binding was at dG), whereas in other tissues, in which there was no increase in mutation frequency, a low level of DNA binding (20–30-fold lower) was observed.21 DNA adduct formation occurred predominantly at guanine residues in liver DNA (adducts 3, 4 and 5), and it was proposed that the observed G:C to T:A transversion mutations are caused by misreplication of adducted guanine bases through incorporation of adenine opposite the adduct (the “A”-rule).21 The present study shows that dG-N2-ABA and dG-C8-N-ABA are probably the guanine adducts responsible for the induction of G:C to T:A transversion mutations observed in Muta™Mouse. Site-specific mutagenesis studies in the future may help to investigate how these individual DNA adducts are converted into mutations.
3-NBA is carcinogenic in F344 rats after intratracheal administration.38 Squamous cell carcinoma was found in the lungs after 7–9 months in the high-dose group (total dose 2.5 mg 3-NBA) and after 10–12 months in the low-dose group (total dose 1.5 mg 3-NBA). We suggest that some or all of the major DNA adducts formed in vivo (adducts 1–5) not only represent premutagenic lesions involved in the mutagenic process but are also critical to the mechanism of 3-NBA carcinogenicity. Indeed, high levels of DNA adducts are found in the lungs of Sprague-Dawley rats treated with a single intratracheal instillation of 3-NBA.33 In the present study, we show that dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA are formed in rat lung DNA after intratracheal instillation with 3-NBA, suggesting that some or all of these lesions are promutagenic and probably initiate tumour development.
In summary, this is the first report on the preparation and characterisation of the 3′-phosphates of dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA to use as authentic standards for the 32P-postlabeling assay. We demonstrate for the first time that these adducts are formed in rodents treated with 3-NBA and its human metabolite 3-ABA. Our results suggest that some of these adducts not only induce specific G:C to T:A transversion mutations in vivo, but may also be critical to the mechanism of 3-NBA carcinogenicity in rodents. A variety of different human phase I and II enzymes are able to catalyse the formation of dA-N6-ABA, dG-N2-ABA and dG-C8-N-ABA in incubations with 3-NBA and/or 3-ABA, and this is important for the assessment of the potential for 3-NBA genotoxicity (carcinogenicity) in humans.
We are most grateful to Dr. T. Suzuki (National Institute of Health Sciences, Tokyo, Japan) for the treatment of Muta™Mouse with 3-NBA and Dr. C.J. Henderson (Biomedical Research Center, Dundee, UK) for the treatment of C57BL/6 mice with 3-NBA and 3-ABA.
- 2IARC. Monographs on the evaluation of carcinogenic risks to humans. Diesel and gasoline engine exhausts and some nitroarenes, vol. 46. Lyon: IARC, 1989.
- 3U.S. Environmental Protection Agency. Health assessment document for diesel engine exhaust. EPA/600/8-90/057F, 2002.
- 6World Health Organisation. World cancer report. Lyon: IARC Press, 2003.
- 363-Aminobenzanthrone, a human metabolite of the environmental pollutant 3-nitrobenzanthrone, forms DNA adducts after metabolic activation by human and rat liver microsomes: evidence for activation by cytochrome P450 1A1 and P450 1A2. Chem Res Toxicol 2004; 17: 1092–1101., , , , , .
- 37Bioactivation of 3-aminobenzanthrone, a human metabolite of the environmental pollutant 3-nitrobenzanthrone: evidence for DNA adduct formation mediated by cytochrome P450 enzymes and peroxidases. Cancer Lett 2005, Epub ahead of print 7 May 2005., , , , , .
- 43DNA adducts of nitropolycyclic aromatic hydrocarbons. In: HemminkiK, DippleA, ShukerDEG, KadlubarFF, SegerbäckD, BartschH, eds. DNA adducts, identification and biological significance. IARC Sci Publ No. 125. Lyon: IARC Press, 1994. 229–44., .
- 44DNA adducts of carcinogenic aromatic amines. In: HemminkiK, DippleA, ShukerDEG, KadlubarFF, SegerbäckD, BartschH, eds. DNA adducts, identification and biological significance. IARC Sci Publ No. 125. Lyon: IARC Press, 1994. 199–216..