Investigation of 2′-Deoxyadenosine-derived Adducts Specifically Formed in Rat Liver and Lung DNA by N′-Nitrosonornicotine (NNN) Metabolism

Abstract

The International Agency for Research on Cancer has classified the tobacco-specific nitrosamines N´-nitrosonornicotine (NNN) and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) as “carcinogenic to humans” (Group 1). To exert its carcinogenicity, NNN requires metabolic activation to form reactive intermediates which alkylate DNA. Previous studies have identified cytochrome P450s-catalyzed 2′-hydroxylation and 5′-hydroxylation of NNN as major metabolic pathways, with preferential activation through the 5′-hydroxylation pathway in some cultured human tissues and patas monkeys. So far, the only DNA adducts identified from NNN 5′-hydroxylation in rat tissues are 2-[2-(3-pyridyl)-N-pyrrolidinyl]-2′-deoxyinosine (Py-Py-dI), 6-[2-(3-pyridyl)-N-pyrrolidinyl]-2′-deoxynebularine (Py-Py-dN), and N⁶-[4-hydroxy-1-(pyridine-3-yl)butyl]-2′-deoxyadenosine (N⁶-HPB-dAdo) after reduction. To expand the DNA adduct panel formed by NNN 5′-hydroxylation and identify possible activation biomarkers of NNN metabolism, we investigated the formation of dAdo-derived adducts using a new highly sensitive and specific LC-NSI-HRMS/MS method. Two types of NNN-specific dAdo-derived adducts, N⁶-[5-(3-pyridyl)tetrahydrofuran-2-yl]-2′-deoxyadenosine (N⁶-Py-THF-dAdo) and 6-[2-(3-pyridyl)-N-pyrrolidinyl-5-hydroxy]-2′-deoxynebularine (Py-Py(OH)-dN), were observed for the first time in calf thymus DNA incubated with 5′-acetoxyNNN. More importantly, Py-Py(OH)-dN was also observed in relatively high abundance in the liver and lung DNA of rats treated with racemic NNN in the drinking water for 3 weeks. These new adducts were characterized using authentic synthesized standards. Both NMR and MS data agreed well with the proposed structures of N⁶-Py-THF-dAdo and Py-Py(OH)-dN. Reduction of Py-Py(OH)-dN by NaBH₃CN led to the formation of Py-Py-dN both in vitro and in vivo, which was confirmed by its isotopically labelled internal standard [pyridine-D₄]Py-Py-dN. The NNN-specific dAdo adducts Py-THF-dAdo and Py-Py(OH)-dN formed by NNN 5′-hydroxylation provide a more comprehensive understanding of the mechanism of DNA adduct formation by NNN.

Graphical Abstract

INTRODUCTION

Smokeless tobacco including chewing tobacco, snuff, snus, and related products comprise noncombustible tobacco products. While the tobacco industry promotes smokeless tobacco as less harmful than combustible tobacco, many of these products contain significant amounts of the tobacco-specific nitrosamine carcinogens N´-nitrosonornicotine (NNN) and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) (Scheme 1). Levels ranged from 2.2 to 42.6 μg/g for NNN and 0.38 to 9.9 μg/g for NNK in the 40 top-selling brands of moist snuff – the most popular smokeless tobacco product – sold in the United States in 2004.¹ An updated survey in 2015 suggested a general decrease but still relatively high NNN levels in 34 smokeless tobacco products, ranging from 0.64 to 12.0 μg/g dry weight.² The U.S. FDA has proposed regulating levels of NNN in finished smokeless tobacco products at a maximum of 1 μg/g dry weight.^3,4 Current use of smokeless tobacco is still alarming. According to the 2019 National Youth Tobacco Survey, 3.5% of high school and middle school students, or a total of 940,000 students in the United States reported current use of smokeless tobacco products.⁵

Scheme 1.

Mechanisms of dAdo-derived adduct formation from NNN and NNK metabolic activation catalyzed by cytochrome P450 enzymes.

Multiple epidemiological studies demonstrate that smokeless tobacco use increases the risk of precancerous lesions of the oral cavity^6,7 and leads to oral, pharyngeal and esophageal cancers.^6,8–10 NNN is considered the chemical carcinogen likely responsible for these types of cancer.^11,12 In a carcinogenicity study in rats, NNN caused esophageal cancer when administered at 5 ppm in the drinking water.¹³ Similarly, chronic administration of racemic NNN (28 ppm) in the drinking water for 17 months to rats produced a high incidence of oral and esophageal cancer.¹² Nasal tumors predominated in rats treated with NNN by subcutaneous injection.^14–16 Tracheal and nasal tumors were also mainly observed in Syrian golden hamsters regardless of the administration pathways.^17–20 In a prospective nested case-control study carried out in Shanghai, cigarette smokers’ urinary total NNN (free NNN plus NNN-N-glucuronide) strongly predicted the future incidence of esophageal cancer.²¹ Thus, NNN and the related carcinogen NNK, which always occur together in tobacco products, have been categorized as Group 1 carcinogens, “carcinogenic to humans” by the International Agency for Research on Cancer.²²

Multiple studies have investigated mechanisms of carcinogenesis by tobacco-specific nitrosamines including NNN and NNK.^23–25 Cytochrome P450s-catalyzed metabolism is required to exert their carcinogenicity (Scheme 1).^23,26 For NNN, hydroxylation at the 5′- or 2′-position forms hydroxyNNNs 4 and 5, respectively. They both spontaneously produce reactive intermediates (oxonium ion 8 and diazonium ion 9 from the 5′-hydroxylation pathway and diazonium ion 10 from the 2′-hydroxylation pathway), which are highly electrophilic and alkylate DNA to form different types of nucleobase adducts^27–33 and phosphate adducts.³⁴ Similarly, 𝛼-methyl and 𝛼-methylene hydroxylation metabolically activate NNK to form pyridyloxobutyl (POB) diazonium ion 10 and methyl diazonium ion 11, respectively. They are both strongly electrophilic and form the corresponding POB and methyl DNA adducts.^35–42 NNK is also enzymatically reduced to its major metabolite 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL), which also exerts strong carcinogenicity through a similar metabolic mechanism as described for NNK.²⁶

Previous studies on the NNN 5′-hydroxylation pathway identified a panel of DNA adducts (Figure 1), mainly from the reaction of 2′-deoxynucleosides or calf thymus DNA with 5′-acetoxy-N′-nitrosonornicotine (5′-acetoxyNNN, 3).^27,29 All the unreduced adducts 12, 22 and 23 were only observed in the reaction mixture of 2′-deoxyadenosine (dAdo) or 2′-deoxyguanosine (dGuo) with 3; the reduced adducts 19, 24, 25 and 21, 26 were observed in calf thymus DNA incubated with 3. However, 2-[2-(3-pyridyl)-N-pyrrolidinyl]-2′-deoxyinosine (Py-Py-dI, 26) was the only quantifiable adduct found in the tissues of rats treated with racemic NNN in the drinking water for 3 weeks. 6-[2-(3-Pyridyl)-N-pyrrolidinyl]-2′-deoxynebularine (Py-Py-dN, 21) was barely observable in the lung and nasal mucosa from the same rats.³³ In the accompanying manuscript, we identified a new type of dAdo adduct N⁶-[4-hydroxy-1-(pyridine-3-yl)butyl]-2´-deoxyadenosine (N⁶-HPB-dAdo, 20) in vivo after reduction, indicating the existence of its precursor adduct 13 before reduction.⁴³ This finding sparked our interest in performing a thorough investigation of all possible dAdo-derived adducts formed by NNN 5′-hydroxylation.

Figure 1.

Summary of DNA adducts formed by NNN 5′-hydroxylation in vitro and in vivo. dR = 2′-deoxyribose. Structures of oxonium ion 8 and diazonium ion 9 are shown in Scheme 1.

In the study presented here, we found that adducts N⁶-[5-(3-pyridyl)tetrahydrofuran-2-yl]-2′-deoxyadenosine (N⁶-Py-THF-dAdo, 12, Scheme 1) and 6-[2-(3-pyridyl)-N-pyrrolidinyl-5-hydroxy]-2′-deoxynebularine (Py-Py(OH)-dN, 14) were both formed in calf thymus DNA incubated with 3. These two new adducts were structurally confirmed and characterized using authentic chemical standards. Importantly, adduct 14 was also observed in the liver and lung DNA of rats treated with NNN in the drinking water for 3 weeks. The formation of its reduced form Py-Py-dN (21) was also confirmed upon NaBH₃CN reduction using the authentic chemical standard. The dAdo-derived adducts 12 and 14 specifically formed by NNN 5′-hydroxylation provide new insights pertinent to our understanding of mechanisms of carcinogenesis by NNN.

MATERIALS AND METHODS

Caution:

NNN and 5′-acetoxyNNN are strong carcinogens. They should be handled in a well-ventilated fume hood with extreme caution and with appropriate protective equipment.

Chemicals and supplies:

5′-AcetoxyNNN was obtained from our accompanying study.⁴³ Chemical standards of N⁶-Py-THF-dAdo and Py-Py-dN were synthesized using the method described before.²⁹ Their corresponding isotopically labeled internal standards [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo and [pyridine-D₄]Py-Py-dN were synthesized as described below, as was Py-Py(OH)-dN. [¹⁵N₅]Py-Py-dI was obtained from our previous study.³³ 6-Chloropurine-2′-deoxyriboside (CAS number 4594–45-0) was purchased from Alfa Aesar (Tewksbury, MA). (R,S)-Nornicotine and [pyridine-D₄](R,S)-nornicotine and [¹³C₁₀¹⁵N₅]2′-deoxyadenosine were procured from Toronto Research Chemicals Inc. (Toronto, Ontario, Canada). Purified DNA hydrolysis enzymes and porcine liver esterase were from our accompanying study.⁴³ All other chemicals and supplies were purchased from Sigma-Aldrich or Fisher Scientific. Milli-Q H₂O was routinely used unless otherwise mentioned.

N⁶-[5-(3-Pyridyl)tetrahydrofuran-2-yl]-2′-deoxyadenosine (12, N⁶-Py-THF-dAdo):

N⁶-Py-THF-dAdo was synthesized using the method described before.²⁹ The synthetic scheme is shown in supplementary Scheme S1. NMR data agreed with most of the reported values except for a few revisions including peak assignments of adenine-N⁶-NH, THF-H₅ and THF-H_4b (Figure S1). High-resolution MS (HRMS) also confirmed the structure. The newly synthesized N⁶-Py-THF-dAdo was essentially identical to our previous standard²⁹ but was quantified by quantitative ¹H NMR.^{44 1}H NMR (500 MHz, DMSO-d₆) δ 8.56 (d, J = 2.2 Hz, 1H, pyr-H₂), 8.48 (dd, J = 4.8, 1.7 Hz, 1H, pyr-H₆), 8.44 (d, J = 1.5 Hz, 1H, ade-H₈), 8.29 (s, 1H, ade-H₂), 7.77 (dt, J = 7.9, 2.1 Hz, 1H, pyr-H₄), 7.37 (dd, J = 7.9, 4.7 Hz, 1H, pyr-H₅), 6.48 (brs, 1H, -NH), 6.38 (t, J = 6.9 Hz, 1H, 1′-H), 5.32 (d, J = 4.0 Hz, 1H, 3′-OH), 5.16 (q, J = 5.6, 4.9 Hz, 1H, 5′-OH), 5.12 – 5.05 (m, 1H, THF-H₂), 5.06 – 4.92 (m, 1H, THF-H₅), 4.53 – 4.36 (m, 1H, 3′-H), 3.88 (t, J = 3.7 Hz, 1H, 4′-H), 3.63 (dd, J = 11.1, 5.2 Hz, 1H, 5′-H_a), 3.53 (q, J = 6.7, 5.8 Hz, 1H, 5′-H_b), 2.72 (ddd, J = 13.3, 7.8, 5.7 Hz, 1H, 2′-H_a), 2.54 (d, J = 1.5 Hz, 1H, THF-H_3a), 2.42 – 2.34 (m, 1H, THF-H_4a), 2.28 (ddd, J = 9.2, 6.5, 3.5 Hz, 1H, 2′-H_b), 2.19 – 1.94 (m, 1H, THF-H_4b), 1.78 (dq, J = 12.1, 8.6 Hz, 1H, THF-H_3b). HRMS (Orbitrap Fusion): [M+H]⁺ calc’d 399.1775; found 399.1775.

[¹³C₁₀¹⁵N₅]N⁶-[5-(3-Pyridyl)tetrahydrofuran-2-yl]-2′-deoxyadenosine ([¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo):

[¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo was synthesized analogously to N⁶-Py-THF-dAdo. Its structure was confirmed by NMR and HRMS. Reaction details and ¹H NMR spectrum can be found in Scheme S1 and Figure S2, respectively. The product was obtained as a colorless oil (0.29 mg, 0.9%). ¹H NMR (500 MHz, DMSO-d₆) δ 8.65 (d, J = 11.4 Hz, 1H, pyr-H₂), 8.56 (d, J = 2.2 Hz, 1H, pyr-H₆), 8.51 – 8.41 (m, 1H, ade-H₈), 8.30 – 7.98 (m, 1H, ade-H₂), 7.77 (dt, J = 8.1, 2.0 Hz, 1H, pyr-H₄), 7.37 (dd, J = 7.9, 4.8 Hz, 1H, pyr-H₅), 6.37 (d, 0.5 H, 0.5 H, 1′-H), 5.09 (dd, J = 8.5, 6.2 Hz, 1H, THF-H₂), 5.06 – 4.91 (m, 1H, THF-H₅), 4.41 (d, 0.5 H, 0.5 H, 1′-H), 4.05 – 3.69 (m, 1H, 4′-H), 3.69 (dd, J = 28.7, 9.6 Hz, 1H, 5′-H_a), 3.50 – 3.36 (m, 1H, 5′-H_b), 2.85 (s, 1H, 2′-H_a), 2.50 (overlapped, 1H, THF-H_3a)2.45 – 2.41 (m, 1H, THF-H_4a), 2.39 – 2.19 (m, 1H, 2′-H_b), 2.19 – 1.93 (m, 1H, THF-H_4b), 1.78 (dq, J = 12.3, 8.2, 7.7 Hz, 1H, THF-H_3b). HRMS (Orbitrap Fusion): [M+H]⁺ calc’d 414.1963; found 414.1962.

6-[2-(3-Pyridyl)-N-pyrrolidinyl-5-hydroxy]-2′-deoxynebularine (14, Py-Py(OH)-dN):

Py-Py(OH)-dN was synthesized as illustrated in Scheme 2, starting with the newly identified dAdo adduct N⁶-HPB-dAdo (20).⁴³ To a solution of 20 (0.03 mmol, 12 mg) in DMSO (0.5 mL) were added Dess-Martin periodinane (0.033 mmol, 14 mg) and Et₃N (0.15 mmol, 21 μL). The mixture was stirred at room temperature for 24 h. After reaction, the mixture was subjected directly to reverse phase HPLC purification as described in our accompanying study.⁴³ The desired product was collected at retention time 27.4 min. However, it was converted to 4 peaks after evaporation of the solvents (MeOH and H₂O) (Figure S3A). This mixture was separated into 3 fractions (7.1 min, 8.5 min, 14.9 min) under isocratic conditions (50% MeOH in H₂O, 15 min running time for each injection), the first two of which were collected and evaporated. Re-injections of the fractions collected for the peaks at 7.1 min and 8.5 min using the same LC method showed the separation of the same 4 peaks (Figure S3B). This strongly indicated that the synthesized compound Py-Py(OH)-dN contained multiple isomers likely including 4 diastereomers of Py-Py(OH)-dN (14) and its equilibrated ring open form N⁶-OPB-dAdo (13) which consisted of 2 diastereomers. The desired product Py-Py(OH)-dN was collected as a colorless oil (0.57 mg, 5%). NMR spectra of Py-Py(OH)-dN are shown in Figure S4. ¹H NMR (500 MHz, DMSO-d₆) δ 8.62 (1H, pyr-H₂), 8.43 – 8.34 (1H, pyr-H₆), 8.28 (1H, ade-H₈), 8.12 (1H, ade-H₂), 7.75 (1H, pyr-H₄), 7.29 (1H, pyr-H₅), 6.51 – 6.33 (m, 1H, 1′-H), 6.23 (1H, pyrrolidine-CH(OH)-), 5.51 (s, 1H, pyrrolidine-CH-pyr), 5.31 (s, 1H, 3′-OH), 5.09 (s, 1H, 5′-OH), 4.41 (s, 1H, 3′-H), 3.87 (s, 1H, 4′-H), 3.60 (s, 1H, 5′-H_a), 3.48 (s, 1H, 5′-H_b), 2.80 – 2.60 (m, 1H, 2′-H_a), 2.45 (s, 1H, pyrrolidine-CH_2aCH-pyr), 2.33 – 2.19 (m, 1H, 2′-H_b), 2.11 (s, 1H, pyrrolidine-CH_2aCH(OH)-), 2.02 – 1.89 (m, 2H, pyrrolidine-CH_2bCH-pyr and pyrrolidine-CH_2bCH(OH)-). HRMS (Orbitrap Fusion): [M+H]⁺ calc’d 399.1775; found 399.1778.

Scheme 2.

Synthetic route for Py-Py(OH)-dN (14).

6-[2-(3-Pyridyl)-N-pyrrolidinyl]-2′-deoxynebularine (21, Py-Py-dN).

Py-Py-dN was synthesized as described.²⁹ The reaction details and NMR data are presented in Scheme S2 and Figure S5. The desired compound was obtained as a colorless oil (11 mg, 73%) and quantified by quantitative ¹H NMR.^{44 1}H NMR (500 MHz, DMSO-d₆) δ 8.45 (s, 1H, pyr-H₂), 8.38 (s, 1H, pyr-H₆), 8.38 (s, 0.5H, ade-H₈), 8.29 (s, 0.5H, ade-H₂), 8.19 (s, 0.5H, ade-H_8′), 8.05 (s, 0.5H, ade-H_2′), 7.55 (d, J = 7.3 Hz, 1H, pyr-H₄), 7.27 (dd, J = 7.5, 5.0 Hz, 1H, pyr-H₅), 6.33 (s, 1H, 1′-H), 6.20 (s, 0.5H, pyrrolidine-CH_a-pyr), 5.52 (s, 0.5H, pyrrolidine-CH_b-pyr), 5.28 (s, 1H, 3′-OH), 5.13 (s, 1H, 5′-OH), 4.42 (s, 0.5H, pyrrolidine-CH_2a(N)-), 4.38 (s, 1H, 3′-H), 4.23 (s, 0.5H, pyrrolidine-CH_2a′(N)-), 4.11– 4.01 (m, 0.5H, pyrrolidine-CH_2b(N)-), 3.85 (s, 1H, 4′-H), 3.79 (s, 0.5H, pyrrolidine-CH_2b′(N)-), 3.58 (s, 1H, 5′-H_a), 3.50 (s, 1H, 5′-H_b), 2.67 (s, 1H, 2′-H_a), 2.42 (s, 1H, pyrrolidine-CH_2aCH-pyr), 2.24 (s, 1H, 2′-H_b), 2.11 – 1.82 (m, 3H, pyrrolidine-CH_2bCH-pyr and pyrrolidine-CH₂CH₂(N)-). ¹³C NMR (126 MHz, DMSO-d₆) δ 152.4 (ade-C₆), 152.0 (ade-C₂), 149.3 (ade-C₄), 147.6 (pyr-C₂), 147.5 (pyr-C₆), 139.1 (overlapped, ade-C₈ and pyr-C₃), 133.2 (pyr-C₄), 123.3 (pyr-C₅), 119.9 (ade-C₅), 87.9 (C_4′), 83.7 (C_1′), 70.9 (doublet peak, C_3′), 61.8 (C_5′), 61.8 (pyrrolidine-CH(N)-), 59.4 (pyrrolidine-CH-pyr), 39.8 (C_2′), 34.9 (pyrrolidine-CH₂CH-pyr), 23.7 (pyrrolidine-CH₂CH₂(N)-). HRMS (Orbitrap Fusion): [M+H]⁺ calc’d 383.1826; found 383.1828.

[Pyridine-D₄]6-[2-(3-pyridyl)-N-pyrrolidinyl]-2′-deoxynebularine (21, [pyridine-D₄]Py-Py-dN).

[Pyridine-D₄]Py-Py-dN was synthesized in the same way as Py-Py-dN (Scheme S2), and was obtained as a colorless oil (11 mg, 95%). NMR spectra are available in Figure S6. ¹H NMR (500 MHz, DMSO-d₆) δ 8.38 (s, 0.5H, ade-H₈), 8.29 (s, 0.5H, ade-H₂), 8.18 (s, 0.5H, ade-H_8′), 8.05 (s, 0.5H, ade-H_2′), 6.33 (s, 1H, 1′-H), 6.20 (s, 0.5H, pyrrolidine-CH_a-pyr), 5.52 (s, 0.5H, pyrrolidine-CH_b-pyr), 5.28 (s, 1H, 3′-OH), 5.13 (s, 1H, 5′-OH), 4.42 (s, 0.5H, pyrrolidine-CH_2a(N)-), 4.38 (s, 1H, 3′-H), 4.23 (s, 0.5H, pyrrolidine-CH_2a′(N)-), 4.04 (s, 0.5H, pyrrolidine-CH_2b(N)-), 3.85 (s, 1H, 4′-H), 3.77 (s, 0.5H, pyrrolidine-CH_2b′(N)-), 3.58 (s, 1H, 5′-H_a), 3.50 (s, 1H, 5′-H_b), 2.67 (s, 1H, 2′-H_a), 2.42 (s, 1H, pyrrolidine-CH_2aCH-pyr), 2.24 (s, 1H, 2′-H_b), 2.08 – 1.80 (m, 3H, pyrrolidine-CH_2bCH-pyr and pyrrolidine-CH₂CH₂(N)-). ¹³C NMR (126 MHz, DMSO-d₆) δ 152.4 (ade-C₆), 151.8 (ade-C₂), 149.8 (ade-C₄), 147.3 (overlapped, pyr-C₂ and pyr-C₆), 139.2 (overlapped, ade-C₈ and pyr-C₃), 132.8 (pyr-C₄), 123.3 (pyr-C₅), 119.6 (ade-C₅), 87.9 (C_4′), 83.7 (C_1′), 70.9 (doublet peak, C_3′), 61.8 (C_5′), 61.8 (pyrrolidine-CH(N)-), 59.2 (pyrrolidine-CH-pyr), 39.8 (C_2′), 33.6 (pyrrolidine-CH₂CH-pyr), 23.4 (pyrrolidine-CH₂CH₂(N)-). HRMS (Orbitrap Fusion): [M+H]⁺ calc’d 387.2077; found 387.2078.

DNA sample preparation without reduction:

Calf thymus DNA incubated with 5′-acetoxyNNN and liver and lung DNA of rats treated with racemic NNN in their drinking water for 3 weeks were obtained from our previous studies.^33,43 They were prepared similarly as described before except for not adding the reducing reagent NaBH₃CN or NaBH₄. The internal standard [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo (10 fmol) was added prior to enzyme hydrolysis. To prepare the spiked samples for the analysis of Py-Py(OH)-dN, calf thymus DNA or rat DNA was analyzed by MS prior to adding synthesized Py-Py(OH)-dN. The spiked samples were then re-analyzed under the same conditions.

DNA sample preparation with reduction:

DNA samples were prepared essentially as described in our accompanying study.⁴³ Briefly, when using NaBH₃CN as the reducing agent, to the isolated DNA (~50 μg) dissolved in 500 μL of 10 mM sodium succinate buffer containing 5 mM CaCl₂ was added 100 μL 20 mg/mL NaBH₃CN in the succinate buffer. The mixture was incubated at 37 °C for 1 h. Then the internal standards [pyridine-D₄]Py-Py-dN (10 fmol) and [¹⁵N₅]Py-Py-dI (10 fmol) [pyridine-D₄]Py-Py-dN were added, followed by 3 purified hydrolytic enzymes deoxyribonuclease I (0.4 units), phosphodiesterase I (0.1 units), and alkaline phosphatase (0.8 units). The mixture was allowed to incubate at 37 °C overnight for a complete DNA hydrolysis. The hydrolysate was filtered with 10K filters (Microcon-10 filter, Millipore) at 14,000 g for 1 h at 4 °C. When using NaBH₄ as the reducing agent, 100 μL 20 mg/mL NaBH₄ in the succinate buffer was added to the DNA hydrolysate and incubated for 1 h. The resulting mixture was adjusted to pH ~8.0 by adding ~50 μL 1 N HCl solution. The hydrolysate was similarly filtered as described above. A small fraction of the hydrolysate (10 μL) was taken for dGuo quantitation by HPLC.⁴³ The remaining hydrolysate was purified with 33 mg Strata-X cartridges and the analyte was finally dissolved in 10 μL H₂O (Fisher, Optima^®) prior to the MS analysis.

Mass spectrometric analysis of dAdo-derived adducts:

DNA hydrolysates were similarly analyzed by the liquid chromatography-nano-electrospray ionization-high-resolution tandem mass spectrometry (LC-NSI-HRMS/MS) method described in our accompanying study.⁴³ Table S1 describes the LC conditions that successfully resolved the newly identified dAdo-derived adducts, and in particular the 4 diastereomers of N⁶-Py-THF-dAdo. New LC conditions to form sharp peaks of Py-Py(OH)-dN and to allow simultaneous detection of Py-Py-dN and Py-Py-dI are described in Table S2. Parameters of the high-resolution mass spectrometer with Orbitrap detector (Thermo Scientific^™ Orbitrap Fusion^™ Tribrid^™) were basically the same except for using the masses of precursor ions of targeted analytes in this study. Precursor ions and MS² and MS³ product ions characteristic of each analyte are summarized in Table 1. Proposed fragmentation patterns corresponding to Table 1 data can be found in Figures S7 and S8.

Table 1.

Precursor ions and MS² and MS³ product ions of dAdo-derived adducts.

	Fragment ions	N6-Py-THF-dAdo	[13C1015N5]N6-Py-THF-dAdo	Py-Py(OH)-dN	Py-Py-dN	[pyridine-D4]Py-Py-dN
Precursor ion	[M+H]+	399.1775	414.1963	399.1775	383.1826	387.2077
MS2 product ions	[M+H-dR]+	283.1302	293.1321	283.1302	267.1353	271.1604
	[M+H-dR-H2O]+	ND	ND	265.1196	ND	ND
MS3 product ions	[M+H-dAdo]+	148.0757	148.0757	148.0757	132.0808	136.1059
	[M+H-dAdo-H2O]+	ND	ND	130.0651	ND	ND
	[Ade+H]+	136.0618	146.0637	136.0618	136.0618	136.0618

ND: not detected; dR: 2'-deoxyribose

RESULTS

Our accompanying study characterized a structurally unique dAdo adduct N⁶-HPB-dAdo (20, Scheme 1) in vitro and in vivo after reduction.⁴³ This adduct was hypothesized to result from the reduction of its precursor N⁶-[4-oxo-1-(pyridine-3-yl)butyl]-2′-deoxyadenosine (N⁶-OPB-dAdo, 13). In addition, a putative adduct N⁶-Py-THF-dAdo (12) was initially observed when investigating new DNA phosphate adducts formed by NNN 5′-hydroxylation (unpublished data). Collectively, those results encouraged us to perform a detailed study of dAdo-derived adducts formed by NNN 5′-hydroxylation.

Synthesis and characterization of chemical standards

N⁶-Py-THF-dAdo (12) and [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo were readily synthesized using the method reported before.²⁹ Both one- and two-dimensional NMR data confirmed their structures (Figures S1 and S2). HRMS analysis suggested that the most abundant MS² product ions were m/z 283.1302 and 293.1321 for N⁶-Py-THF-dAdo and [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo, respectively (Figure S7). They were both formed by the neutral loss of 2′-deoxyribose. Product ions from MS³ transitions were also similar for both compounds, with the most abundant ions of m/z 148.0757 [M+H-dAdo]⁺ and 136.0618 [Ade+H]⁺ for N⁶-Py-THF-dAdo, and 148.0757 [M+H-dAdo]⁺ and 146.0637 [[¹³C₁₀¹⁵N₅]Ade+H]⁺ for [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo (Table 1). Chromatographic baseline resolution of the 4 diastereomers of each compound was successfully achieved under the LC conditions described in Table S1 (as shown in Figure 3 and Figure 5). Absolute configuration of the stereochemistry of each peak had been assigned previously,²⁹ however, was not pursued in this study.

Figure 3.

Representative extracted product ion chromatograms of MS² transitions for the analysis of N⁶-Py-THF-dAdo and Py-Py(OH)-dN formation in calf thymus DNA incubated with 5′-acetoxyNNN. Four diastereomer peaks of N⁶-Py-THF-dAdo were observed from the MS² transition of 399.2 → 283.1302, co-eluting with the synthesized internal standard [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo. However, a broad peak nearly co-eluting with N⁶-Py-THF-dAdo was observed. It can form a unique MS² fragment ion 265.1195, which suggests its structure to be Py-Py(OH)-dN (see Figure 2 for the structure and fragmentation pattern of Py-Py(OH)-dN).

Figure 5.

Py-Py(OH)-dN (in red) but not N⁶-Py-THF-dAdo (in blue) was observed in the lung DNA of rats treated with 500 ppm racemic NNN for 3 weeks. LC conditions are critical for resolving N⁶-Py-THF-dAdo peaks from the massive peak of Py-Py(OH)-dN. (A) LC conditions depicted in Table S1 can separate the 4 diastereomers of N⁶-Py-THF-dAdo and partially resolve it from Py-Py(OH)-dN, however causing a significantly broad peak of Py-Py(OH)-dN. (B) LC conditions depicted in Table S2 afford a sharper peak of Py-Py(OH)-dN but fail to resolve N⁶-Py-THF-dAdo.

Py-Py(OH)-dN (14) is the cyclic equilibrated form of N⁶-OPB-dAdo (13) (Scheme 1). They were together hypothesized as products of NNN 5′-hydroxylation based on characterization of N⁶-HPB-dAdo (20) after reduction.⁴³ Py-Py(OH)-dN was synthesized as illustrated in Scheme 2. Oxidation of N⁶-HPB-dAdo (20) using Dess-Martin periodinane yielded N⁶-OPB-dAdo, which spontaneously cyclized to Py-Py(OH)-dN during solvent evaporation. As shown in Figure S3, HPLC purification afforded a mixture of at least 4 peaks, which likely contained both the 4 diastereomers of Py-Py(OH)-dN and its equilibrated ring open form N⁶-OPB-dAdo with 2 diastereomers. The ¹H NMR spectrum showed typical proton resonances consistent with the pyridine ring and nucleobase moiety of Py-Py(OH)-dN (Figure S4). However, absolute quantitation was not achieved due to the difficulty of accurately integrating each spectral peak for quantitative ¹H NMR analysis.

HRMS data agreed well with the proposed fragmentation pattern of Py-Py(OH)-dN (Figure 2). Interestingly, we noted that an ion of m/z 265.1197 [M+H-dR-H₂O]⁺ was the second most predominant ion in the MS² transitions of Py-Py(OH)-dN. This ion was not observed in the MS² spectra of N⁶-Py-THF-dAdo (Figure S7). Fragmentation of the most abundant ion m/z 283.1302 [M+H-dR]⁺ led to the formation of m/z 136.0618 [Ade+H]⁺ and m/z 148.0757 [M+H-dAdo]⁺, both of which were also observed in the MS³ transitions of N⁶-Py-THF-dAdo. However, fragmentation of the unique ion m/z 265.1196 formed a product ion m/z 130.0651 [M+H-dAdo-H₂O]⁺, which appeared to be characteristic of Py-Py(OH)-dN. Thus, transitions of m/z 399.2 → m/z 265.1196 and m/z 265.2 → m/z 130.0651 were chosen for the characterization of Py-Py(OH)-dN, distinguishing it from the nearly co-eluting peak of N⁶-Py-THF-dAdo in the DNA samples (Figure 3).

Figure 2.

Representative chromatograms and MS² and MS³ product ion spectra of Py-Py(OH)-dN. The fragmentation patterns of MS² and MS³ transitions of Py-Py(OH)-dN agreed well with proposed patterns.

Py-Py-dN (21), the reduced form of Py-Py(OH)-dN, and its isotopically labeled internal standard [pyridine-D₄]Py-Py-dN were synthesized as reported previously.²⁹ NMR data were consistent with reported values (Figures S5 and S6). Analysis by HRMS showed good agreement with the proposed fragmentation patterns (Figure S8). The predominant MS² transition of Py-Py-dN was m/z 383.2 [M+H]⁺ → 267.1353 [M+H-dR]⁺; the predominant MS² transition of [pyridine-D₄]Py-Py-dN was the corresponding m/z 387.2 to m/z 271.1604. It is noteworthy that MS³ fragmentation of each compound required higher collision energy than usual. With an HCD collision energy setting at 75%, there were still substantial amounts of precursor ions unfragmented with both compounds (data not shown).

Detection of N⁶-Py-THF-dAdo and Py-Py(OH)-dN in vitro without reduction

A highly sensitive and specific LC-NSI-HRMS/MS method was used for the analysis of dAdo-derived adducts without reduction. Analysis of hydrolysates of calf thymus DNA incubated with 5′-acetoxyNNN clearly showed the formation of N⁶-Py-THF-dAdo (Figure 3). Four diastereomers of this adduct were resolved nicely, each of which co-eluted with the corresponding isomer of the internal standard [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo. The representative MS² fragmentation pattern of N⁶-Py-THF-dAdo agreed well with expectations. Both MS² and MS³ fragmentation patterns of each of the 4 major peaks were essentially identical (Figure S9), further indicating the formation of N⁶-Py-THF-dAdo containing 4 diastereomers in calf thymus DNA incubated with 5′-acetoxyNNN.

However, as depicted in Figure 3, there was a broad peak nearly co-eluting with N⁶-Py-THF-dAdo. Mass spectrometric differences were clearly observed between this broad peak and N⁶-Py-THF-dAdo. The presence of the precursor ion m/z 399.1771 and the formation of the unique product ions m/z 265.1195 and m/z 130.0650 suggested that this broad peak was Py-Py(OH)-dN. The extracted ion chromatogram of MS² transition m/z 399.2 → 265.1196 distinguished this peak nicely from N⁶-Py-THF-dAdo (middle MS trace of Figure 3). An additional experiment with synthesized Py-Py(OH)-dN spiked into the calf thymus DNA showed the same but augmented peak area of Py-Py(OH)-dN (Figure 4). This strongly implied that Py-Py(OH)-dN was formed as the broad peak in vitro as shown in Figure 3. It is noteworthy that Py-Py(OH)-dN peak shown in Figure 4 is relatively sharper than in Figure 3 when new LC conditions were applied as described in Table S2. However, this new method could not resolve the peaks of N⁶-Py-THF-dAdo from the massive peak of Py-Py(OH)-dN (Figure 5B).

Figure 4.

Py-Py(OH)-dN was formed in calf thymus DNA incubated with 5′-acetoxyNNN and in the liver and lung DNA of rats treated with 500 ppm racemic NNN in the drinking water for 3 weeks. Synthesized Py-Py(OH)-dN spiked into the DNA samples co-eluted with the observed peak. No such Py-Py(OH)-dN peak was observed in an untreated control calf thymus DNA sample.

Detection of Py-Py(OH)-dN but not N⁶-Py-THF-dAdo in vivo without reduction

Similarly, the hydrolysates of liver and lung DNA of rats treated with 500 ppm racemic NNN in the drinking water for 3 weeks were analyzed by the LC-NSI-HRMS/MS method. As shown in Figure 4, clear peaks of Py-Py(OH)-dN, determined by its characteristic MS² transition of m/z 399.2 → 265.1196 and MS³ transition of m/z 265.1 → 130.0651, were observed in the rat liver and lung DNA. A similar spiking experiment as described above also confirmed the formation of Py-Py(OH)-dN in the rat tissues. No such Py-Py(OH)-dN peak was observed in the liver and lung DNA of control rats from the same study given tap water only.

However, it was interesting that no such N⁶-Py-THF-dAdo peaks were observed in any in vivo DNA samples used for this study. As shown in Figure 5A, [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo was at least partially resolved from the massive peak of Py-Py(OH)-dN. No peaks co-eluted with the isomers of [¹³C₁₀¹⁵N₅]N⁶-Py-THF-dAdo in the rat lung DNA, which suggested the absence of N⁶-Py-THF-dAdo. Considering that N⁶-Py-THF-dAdo was readily formed in vitro (Figure 3), this could indicate a high efficiency of repair towards this type of DNA damage in vivo, at least in the liver and lung tissues of rats examined in this study.

Detection of Py-Py-dN in vitro and in vivo upon reduction

Py-Py(OH)-dN has been detected in relatively high MS abundance both in vitro and in vivo, as shown in Figures 4 and 5. However, it was difficult to accurately quantify due to its broad peak shape (W_baseline = ~10 min) and the complexity of the synthesized standard. Based on our experience with Py-Py-dI quantitation,³³ it seemed reasonable to reduce Py-Py(OH)-dN to Py-Py-dN for quantitation. However, Py-Py(OH)-dN was not fully reduced using 2 mg NaBH₃CN as in our previous studies of Py-Py-dI.^29,33 Increased amounts of NaBH₃CN (up to 10 mg) improved the conversion rate but significantly affected the enzymatic hydrolysis of DNA. Instead, 2 mg NaBH₄ nearly completely converted Py-Py(OH)-dN to its reduced form Py-Py-dN when added into the hydrolysate after enzymatic hydrolysis (data not shown).

Surprisingly, Py-Py-dN, the reduced form of Py-Py(OH)-dN which has been readily observed both in vitro and in vivo, was barely detected in the same DNA samples reduced by NaBH₄. However, it was clearly observed in NaBH₃CN-reduced samples (Figure 6 and Figure S10). The formation of the two diastereomers of Py-Py-dI in the same DNA samples suggested that this unexpected result was not due to experimental errors with the reduction procedure. The equilibrium between Py-Py(OH)-dN and N⁶-OPB-dAdo (Scheme 1) is hypothesized to be responsible for the missing detection of Py-Py-dN in samples treated with NaBH₄. This will be discussed further in the Discussion section.

Figure 6.

Py-Py-dN was clearly observed in the lung DNA of rats reduced by NaBH₃CN but not NaBH₄. (A) Untreated calf thymus DNA reduced with 2 mg NaBH₃CN; (B) lung DNA of rats treated with 500 ppm NNN reduced with 2 mg NaBH₃CN; (C) lung DNA of rats treated with 500 ppm NNN reduced with 2 mg NaBH₄.

DISCUSSION

In this study, we characterized two types of NNN-specific DNA adducts - N⁶-Py-THF-dAdo (12, Scheme 1) and Py-Py(OH)-dN (14). Both are formed in calf thymus DNA treated with 5′-acetoxyNNN and, more importantly, Py-Py(OH)-dN was for the first time observed in the DNA of rats treated with racemic NNN. The structure and presence of both adducts were confirmed by the authentic chemical standards. Py-Py-dN, the reduced form of Py-Py(OH)-dN, was also detected in rat DNA samples reduced by NaBH₃CN. The two new adducts characterized here, together with N⁶-HPB-dAdo (20), characterized in the companion study⁴³, provide a more comprehensive understanding of the mechanism of DNA damage by NNN.

N⁶-Py-THF-dAdo was previously characterized in the reaction mixture of dAdo with 5′-acetoxyNNN;²⁹ however, it was not observed in treated calf thymus DNA until now. An important aspect of the current study is the use of HRMS, which provides significantly improved sensitivity and specificity to distinguish the peaks of targeted analytes from background noise. The observed peaks of N⁶-Py-THF-dAdo in our calf thymus DNA sample were unique, with 4 baseline-resolved peaks, each of which showed essentially identical MS² and MS³ product ion patterns (Figures S9). This strongly indicated that those peaks represented 4 isomers of one single compound. The MS² product ion m/z 283.1302 (formed by neutral loss of 2′-deoxyribose) and its corresponding MS³ product ion m/z 136.0617 [Ade+H]⁺ suggested that this compound contained 2′-dAdo. Thus, a possible dAdo-derived adduct N⁶-Py-THF-dAdo which contains 2 chiral centers with an exact mass of m/z 399.1775 was proposed to be the observed adduct in the treated calf thymus DNA sample. Using the authentic synthesized standards, those peaks were confirmed unambiguously as the 4 diastereomers of N⁶-Py-THF-dAdo. However, no such peaks were observed from any in vivo DNA samples. This may imply that N⁶-Py-THF-dAdo, if formed in rat DNA, is efficiently repaired by a mechanism yet to be studied.

The tetrahydrofuranyl dAdo adducts formed by NNN were undetectable in rat liver and lung DNA in this study. This was similar to the results we obtained in hepatic DNA of rats treated with N-nitrosopyrrolidine (NPYR). As a strong rat hepatocarcinogen and a model compound for the study of NNN, NPYR has been found to form tetrahydrofuranyl (THF) adducts with all 4 nucleobases in vitro and in vivo.^45–48 However, the overall levels of all those THF-type base adducts (quantified after reduction) were significantly lower than other adducts formed by NPYR.⁴⁹ The dAdo-derived adduct N⁶-(4-hydroxybut-1-yl)-2′-deoxyadenosine (N⁶-(4-HOB)dAdo) was in particularly low abundance compared to the other nucleobase-derived adducts, occurred at levels of 0.02 – 0.04 μmol/mol dGuo (or 20 – 40 fmol/μmol dGuo) in the liver DNA of rats treated with 200 or 600 ppm NPYR in the drinking water for various times.⁴⁸ Albeit the low levels of N⁶-(4-HOB)dAdo in rat liver DNA, it might be of greater biological importance than other analogous adducts to cause the predominant AT to GC transition in the mutants of NPYR-treated Sprague–Dawley gpt δ transgenic rats.⁵⁰

Initial characterization of Py-Py(OH)-dN was difficult due to its very similar chromatographic behavior to N⁶-Py-THF-dAdo. After extensive optimization of the LC conditions, a very shallow linear gradient as depicted in Table S1 successfully resolved the Py-Py(OH)-dN peak from the 4 diastereomers of N⁶-Py-THF-dAdo (Figure 3). Unique features of m/z 265.1196 in MS² transitions and m/z 130.0618 in MS³ transitions were found to be characteristic to Py-Py(OH)-dN. This was likely due to the hemiaminal moiety of Py-Py(OH)-dN that could lose a molecule of H₂O during fragmentation (Figure 2).

For the quantitation of Py-Py(OH)-dN in rat liver and lung DNA samples, the reducing strategy used for the Py-Py-dI study was adopted here. However, reported conditions using NaBH₃CN did not achieve a complete reduction of Py-Py(OH)-dN.^{28, 32} The optimized reducing condition of adding NaBH₄ into the DNA hydrolysate after enzymatic hydrolysis led to a nearly quantitative conversion. It is noteworthy that reducing agents can interfere with enzymatic hydrolysis under the conditions described before.^28,32 NaBH₄ strongly inhibited the hydrolysis process (with pH adjusted) at the amount of 2 mg if added prior to the enzymatic hydrolysis step. Extra washing steps by filtering undigested DNA with H₂O or succinate buffer did not solve the problem.⁵¹ On the contrary, NaBH₃CN was well tolerated by the hydrolytic enzymes at the amount of 2 mg but started showing inhibitory effect at higher amounts. The underlying mechanism is unclear but seems to be related with the boronate ions formed by the solvolysis of these two reducing agents.

The formation of Py-Py-dN (21, Scheme 1) in calf thymus DNA incubated with 5′-acetoxyNNN and rat liver and lung DNA was expected upon NaBH₃CN reduction, since its precursor Py-Py(OH)-dN was readily detected in those samples. However, it was surprisingly unexpected to barely detect this adduct in the same DNA samples upon NaBH₄ reduction as shown in Figure 6 and Figure S10. Considering that Py-Py(OH)-dN is in equilibrium with N⁶-OPB-dAdo (Scheme 1) in the DNA hydrolysates, it seems possible that a stronger reducing agent such as NaBH₄ favors the reduction of N⁶-OPB-dAdo versus Py-Py(OH)-dN, and thus leads to the predominant formation of N⁶-HPB-dAdo. In contrast, NaBH₃CN is likely to reduce both N⁶-OPB-dAdo and the Schiff base 18 (hypothetically formed by Py-Py(OH)-dN) to a similar extent to form both of their corresponding reduced products N⁶-HPB-dAdo (20) and Py-Py-dN (21). This is consistent with the results obtained in our accompanying study, in which levels of N⁶-HPB-dAdo were generally lower in DNA samples reduced by NaBH₃CN versus NaBH₄.⁴³ Considering that Py-Py-dI was formed in a similar mechanism,³³ its decreased levels in DNA samples reduced by NaBH₄ compared to those reduced by NaBH₃CN as observed in Figure 6 and Figure S10 might also be the same cause.

One of the biomarkers used for monitoring tobacco carcinogen uptake in smokeless tobacco users is NNAL, a reliable biomarker specifically for NNK exposure which can also reflect relevant uptake of other tobacco carcinogens.⁵² However, this biomarker does not directly reflect the level of NNN, which is frequently present in relatively high quantities in smokeless tobacco products.⁵³ NNN is the most abundant tobacco-specific carcinogen in unburned tobacco.^1,2 One direct biomarker for NNN exposure which was applied in the Shanghai Cohort Study is urinary total NNN.²¹ However, the level of urinary total NNN does not reflect the metabolic fate of NNN, and only accounted for less than 1% of dosed NNN in the patas monkey.⁵⁴ Thus, identification of new NNN-specific biomarkers such as NNN-specific DNA adducts that can reflect NNN uptake and metabolic activation is of high interest to us. As depicted in Scheme 1, NNN 2′-hydroxylation forms the same diazonium ion 10 as the 𝛼-methyl hydroxylation of NNK. The convergence of this alkylating intermediate makes the origin of resulting DNA adducts such as 7-(4-(3-pyridyl)-4-oxobut-1-yl)guanine (7-POB-Gua) or O²-(4-(3-pyridyl)-4-oxobut-1-yl)thymidine (O²-POB-Thd) indistinguishable from the two tobacco-specific nitrosamine carcinogens. However, NNN 5′-hydroxylation forms oxonium ion 8 and diazonium ion 9, both of which are structurally unique. DNA adducts formed by alkylation of those two reactive intermediates are structurally specific to NNN metabolism. Besides, the 5′-hydroxylation pathway of NNN is more prevalent than 2′-hydroxylation in some cultured human tissues^15,55–59 and in the patas monkey.⁵⁴ Thus, products formed by NNN 5′-hydroxylation such as N⁶-Py-THF-dAdo and Py-Py(OH)-dN characterized in this study have potential to be NNN-specific metabolic activation biomarkers in people who use tobacco products.

In summary, we have characterized the formation of two NNN-specific DNA adducts N⁶-Py-THF-dAdo and Py-Py(OH)-dN in calf thymus DNA incubated with 5′-acetoxyNNN using authentic synthesized chemical standards. More importantly, Py-Py(OH)-dN was for the first time detected in the liver and lung DNA of rats treated with racemic NNN. Its reduced form Py-Py-dN was also confirmed in the same DNA samples reduced by NaBH₃CN. The results of this study provide a better understanding of the mechanism of carcinogenesis caused by the tobacco-specific carcinogen NNN.

ABBREVIATIONS

NNN

N′-nitrosonornicotine

NNK

4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone

FDA

Food and Drug Administration

POB

pyridyloxobutryl

NNAL

4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol

5′-acetoxyNNN

5′-acetoxy-N′-nitrosonornicotine

dAdo

2′-deoxyadenosine

dGuo

2′-deoxyguanosine

Py-Py-dI

2-[2-(3-pyridyl)-N-pyrrolidinyl]-2′-deoxyinosine

Py-Py-dN

6-[2-(3-pyridyl)-N-pyrrolidinyl]-2′-deoxynebularine

N⁶-HPB-dAdo

N⁶-[4-hydroxy-1-(pyridine-3-yl)butyl]-2´-deoxyadenosine

N⁶-Py-THF-dAdo

N⁶-[5-(3-pyridyl)tetrahydrofuran-2-yl]-2′-deoxyadenosine

Py-Py(OH)-dN

6-[2-(3-pyridyl)-N-pyrrolidinyl-5-hydroxy]-2′-deoxynebularine

NaBH₃CN

sodium cyanoborohydride

NaBH₄

sodium borohydride

N⁶-PHB-dAdo

N⁶-[4-(3-pyridyl)-4-hydroxy-1-butyl]-2′-deoxyadenosine

MS

mass spectrometry

HRMS

high-resolution mass spectrometry

NaBH₄

sodium borohydride

LC-NSI-HRMS/MS

liquid chromatography-nano-electrospray ionization-high-resolution tandem mass spectrometry

SPE

solid phase extraction

HCD

higher-energy collisional dissociation

AGC

automatic gain control

LOD

limit of detection

LOQ

limit of quantitation

CV

coefficient of variation

NYPR

N-nitrosopyrrolidine

N⁶-(4-HOB)dAdo

N⁶-(4-hydroxybut-1-yl)-2′-deoxyadenosine