Synthesis of Oligonucleotides Containing the N²-Deoxyguanosine Adduct of the Dietary Carcinogen 2-Amino-3-methylimidazo[4,5-f]quinoline

Abstract

2-Amino-3-methylimidazo[4,5-f]quinoline (IQ) is a highly mutagenic heterocyclic amine formed in all cooked meats. IQ has been found to be a potent inducer of frameshift mutations in bacteria and carcinogenic in laboratory animals. Upon metabolic activation, IQ forms covalent adducts at the C8- and N²-positions of deoxyguanosine with a relative ratio of up to ~4:1. We have previously incorporated the major dGuo-C8-IQ adduct into oligonucleotides through the corresponding phosphoramidite reagent. We report here the sequence-specific synthesis of oligonucleotides containing the minor dGuo-N²-IQ adduct. Thermal melting analysis revealed that the dGuo-N²-IQ adduct significantly destabilizes duplex DNA.

Introduction

2-Amino-3-methylimidazo[4,5-f]quinoline (IQ, 1)¹ is one of the most mutagenic members of a family of heterocyclic amines (HCAs) that are formed in parts-per-billion quantities during the cooking of all meats, fish, and poultry (Figure 1) (1–5). Others members of this family include 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), 2-amino-3,4-dimethylimidazo[4,5-f]quinoline (MeIQ), and 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx), among others. Although human intake of HCAs is relatively modest (6, 7), the exposure usually occurs over a lifetime. Some of the HCAs are potent inducers of frameshift mutations in the Ames assays, and many have been shown to be carcinogenic in laboratory animals (3, 5, 8). Epidemiological studies have linked red meat consumption to a higher risk of colorectal and other cancers in humans, and it is likely that the HCAs contribute to these findings (4, 9–12).

Figure 1.

IQ and its C8- and N²-dGuo adducts.

Aromatics amines require a two-stage metabolic activation to covalently modify DNA (2, 13). The amino group undergoes initial N-oxidation by a cytochrome P450 to the corresponding hydroxylamine. The hydroxylamine is esterified by N-acetyl transferase or sulfotransferase to give the O-acetyl or O-sulfonyl derivative, respectively. The ultimate carcinogenic species is a highly reactive aryl nitrenium ion, which is generated through the rapid solvolysis of the hydroxylamine esters. Aryl nitrenium ions are ambident electrophiles and can react with nucleophilic species at the exocyclic nitrogen or a ring carbon. The nitrenium ion of IQ, as well as other heterocyclic and simple aromatic amines, reacts predominantly at the C8-positions of dGuo (dGuo-C8-IQ, 2) (Figure 1). A minor N²-adduct of IQ (dGuo-N²-IQ, 3) (14, 15), MeIQx (15), and some simple arylamines have also been characterized (16–23).

There is evidence that the N²-arylamine adducts are more persistent than the C8 adducts, making this minor regioisomer of particular interest. Feeding studies with rats and monkeys found that the major DNA lesion was the C8 adduct 24 h after a single dose of IQ (24, 25). The relative ratio of the C8- to N²-dGuo adducts was in the range of 4:1–3:2, which roughly parallels the natural reactivity of the IQ nitrenium ion toward DNA. However, in chronically treated monkeys undergoing long-term carcinogen bioassays, a sharp increase in the proportion of N² adduct was observed after 3.6 years. This increase was found in slowly dividing tissue such as the liver, pancreas, and kidney. There was no increased persistence of the N² adduct in colon tissue, which divides at a high rate. These observations suggest that the C8 adduct is repaired more efficiently in slow-growing tissue. Similar results have been reported for 2-aminofluorene (AF) and N-acetyl-2-aminofluorene (AAF) (26, 27). This has led to speculation that the minor N²-dGuo adducts may play a more significant role in the tumorigenic properties of the HCAs and other aromatic amines.

We report here the synthesis of a phosphoramidite reagent of the minor dGuo-N²-IQ adduct and its site-specific incorporation into oligonucleotides. The key synthetic step involved a Buchwald–Hartwig palladium-catalyzed cross-coupling reaction. The oligonucleotides were characterized by matrix-assisted laser desorption mass spectrometry (MALDI-MS) and thermal melting analysis (T_m). The presence of the dGuo-N²-IQ significantly destabilized the duplex when compared to the corresponding dGuo-C8-IQ adducts.

Experimental Procedures

Anhydrous pyridine was purchased from Aldrich in sure-seal bottles and used as received. Methylene chloride was freshly distilled from CaH₂. IQ was purchased from Toronto Research Chemicals and used as received. Melting points are uncorrected. Proton and carbon-13 NMR data were recorded at 300 or 400 MHz and 75 or 100 MHz in the solvent noted. High-resolution fast atom bombardment (FAB) mass spectra were obtained from the University of Notre Dame Mass Spectrometry Center using nitrobenzyl alcohol (NBA) as the matrix. MALDI time-of flight (TOF) mass spectrometry of oligonucleotides was performed at the Vanderbilt University Mass Spectrometry Resource Center using a 3-hydroxypicolinic acid (HPA) matrix containing ammonium hydrogen citrate (7 mg/mL) to suppress sodium and potassium adducts.

N²-{2-[2-(4-tert-Butylphenoxy)acetylamino]-3-methyl-3H-imidazo[4,5-f]quinolin-5-yl}-5′-O-(4,4′-dimethoxytrityl)-2′-deoxyguanosine (10)

To a stirred solution of 9 (28) (50 mg, 0.076 mmol) in 1.0 mL of anhydrous pyridine was added 4,4′-dimethoxytrityl chloride (39 mg, 0.011 mmol) as a solid, and the reaction was allowed to stir for 3 h at room temperature. The reaction was quenched by the addition of methanol (1.0 mL), and the mixture was stirred for 10 min at room temperature and then concentrated under reduced pressure. The residue was purified by flash chromatography on silica, eluting with 1.0% methanol in methylene chloride containing 0.5% triethylamine, and provided 10 (48 mg, 78% yield) as a yellow solid; mp 222–224 °C; [α]_D²³ 18.5° (c 0.5, 1:1 MeOH/CHCl ). ¹H NMR (DMSO-d₆): δ 10.2 (bs, 1H), 8.94 (s, 1H), 8.81 (bs, 1H), 7.99 (s, 1H), 7.81 (dd, J = 7.20, 3.51, 1H), 7.32 (m, 2H), 7.22 (m, 2H), 7.15 (m, 7H), 6.95 (m, 2H), 6.71 (m, 4H), 6.45 (t, J = 10.72, 4.85, 1H), 5.36 (bs, 1H), 4.89 (bs, 2H), 4.41 (m, 1H), 3.99 (m, 1H), 3.75 (s, 3H), 3.65 (m, 6H), 3.12 (m, 3H), 2.85 (m, 3H), 1.23 (m, 10H). ¹³C NMR (DMSO-d₆): δ 158.85, 157.81, 156.29, 148.48, 145.01, 140.24, 136.32, 130.32, 130.22, 128.90, 128.41, 127.80, 127.39, 123.97, 120.82, 117.14, 112.32, 87.31, 85.74, 74.54, 65.31, 55.37, 42.21, 37.18, 28.56. HRMS (FAB, NBA) m/z calcd for C₅₄H₅₄N₉O₈ (M + H), 956.4095; found, 956.4418.

N²-{2-[2-(4-tert-Butylphenoxy)acetylamino]-3-methyl-3H-imidazo[4,5-f]quinolin-5-yl}-5′-O-(4,4′-dimethoxytrityl)-3′-O[N,N-diisopropylamino(2-cyanoethoxy)phosphinyl]-2′-deoxyguanosine (11)

To a stirred suspension of 10 (50 mg, 0.052 mmol) and 1H-tetrazole (5 mg, 0.07 mmol) in anhydrous methylene chloride (3 mL) was added 2-cyanoethyl N,N,N′,N′-tetraisopropylphosphorodiamidite (0.6 mg, 0.030 mL, 0.085 mmol). The reaction was stirred for 1.5 h at room temperature. The solvent was then removed under reduced pressure, and the residue was purified by flash chromatography on silica, eluting with 2% methanol in methylene chloride containing 0.5% pyridine, and provided 11 (50 mg, 83% yield) as a light yellow solid; mp 188–190 °C; [α]_D²³ +79° (c 0.5, CH₂Cl₂). ¹H NMR (CD₂Cl₂): δ 10.1 (bs, 1H), 8.74 (s, 1H), 8.65 (bs, 1H), 8.52 (bs, 1H), 7.78 (s, 1H), 7.40–7.15 (m, 10H), 6.95 (m, 2H), 6.79 (m, 4H), 6.41 (t, J = 10.60, 4.55, 1H), 4.72 (bs, 2H), 4.25 (m, 1H), 3.71 (m, 6H), 3.65 (m, 2H), 3.25 (m, 2H), 3.21 (m, 14H), 2.75 (m, 2H), 2.55 (m, 2H), 1.33 (m, 14H). ¹³C NMR (CD₂Cl₂): δ 158.85, 157.81, 156.29, 152.25, 149.20, 149.00, 148.48, 145.01, 140.24, 136.67, 136.32, 130.83, 130.61, 130.32, 130.22, 128.90, 128.41, 128.36, 127.80, 127.39, 123.97, 123.90, 120.82, 117.14, 112.32, 87.31, 86.78, 85.74, 74.92, 74.54, 65.31, 59.15, 58.76, 55.37, 44.21, 42.21, 38.15, 37.18, 28.56, 25.21, 23.45, 21.10. ³¹P NMR (CD₂Cl₂, 121 MHz): δ 150.30, 149.90. HRMS (FAB, NBA) m/z calcd for C₆₃H₇₁N₁₁O₉P (M + H), 1156.5174; found, 1156.2138.

General Procedure for the Synthesis of Oligonucleotides

The adducted oligonucleotides were synthesized on an Expedite 8909 DNA synthesizer (PerSeptive Biosystems) on a 1 μmol scale using the UltraMild line of phosphoramidites [phenoxyacetyl-protected dAdo, (4-isopropylphenoxy)acetyl-protected dGuo, acetyl-protected dCyd, and dThd phosphoramidites] and solid supports (Glen Research). The manufacturer’s standard synthesis protocol was followed except at the incorporation of the modified phosphoramidites, which was accomplished manually, off-line. At this point, the column was removed from the instrument and sealed with two syringes: One contained 250–300 μL of the manufacturer’s 1H-tetrazole activator solution (1.9–4.0% in acetonitrile), and the other contained 250 μL of the phosphoramidite solution (15 mg, 0.098 M in anhydrous methylene chloride). The 1H-tetrazole and the phosphoramidite solutions were sequentially drawn through the column (1H-tetrazole first), and this procedure was repeated periodically over 30 min. After this time, the column was washed with manufacturer’s grade anhydrous acetonitrile and returned to the instrument for the capping, oxidation, and detritylation steps. The remainder of the synthesis was carried out as normal.

Oligonucleotide Purification

The analysis and purification of modified oligonucleotides were performed on a Beckman HPLC system with a UV diode array detector (model 166) monitoring at 254 nm utilizing 32 Karat software (Version 3.1). The HPLC solvents were pH 7.0, 20 mM sodium phosphate, and methanol. The solvent gradient was as follows: initially 1% methanol; then a 27.5 min linear gradient to 65% methanol; 2.5 min linear gradient to 50% methanol; followed by 5 min isocratic at 50% methanol; then a 5 min linear gradient to initial conditions.

Capillary Gel Electrophoresis (CGE)

Electrophoretic analyses were carried out using a Beckman P/ACE Instrument System 5500 Series monitored at 260 nm. The P/ACE MDQ instrument used a 31.2 cm × 100 μm eCAP capillary with samples applied at 10 kV and run at 9 kV. The column was packed with the manufacturer’s 100-R gel (for ss DNA) using a Tris-borate buffer system containing 7.0 M urea.

5′-CTC GGC GCC ATC-3′ (12)

Oligonucleotide 12 was purified by the HPLC method above to give 3.8 A₂₆₀ units. MALDI-TOF MS (HPA) m/z calcd for (M – H), 3776.58; found, 3777.49.

5′-d(CTC GGC GCC ATC)-3′ (13)

Oligonucleotide 13 was purified by the HPLC method above to give 6.2 A₂₆₀ units. MALDI-TOF MS (HPA) m/z calcd for (M – H), 3775.58; found, 3776.54.

5′-CTC GGC GCC ATC-3′ (14)

Oligonucleotide 14 was purified by the HPLC method above to give 6.5 A₂₆₀ units. MALDI-TOF MS (HPA) m/z calcd for (M – H), 3776.58; found, 3776.26.

5′-G GCA GGT GTG G-3′ (15)

Oligonucleotide 15 was purified by the HPLC method above to give 8.2 A₂₆₀ units. MALDI-TOF MS (HPA) m/z calcd for (M – H), 3647.76; found, 3647.89.

Thermal Melting (T_m) Studies

Equal amounts of the adducted and the complementary strands (0.3 A₂₆₀ units each) were dissolved in 0.5 mL of buffer (100 mM NaCl, 0.05 mM EDTA, and 10 mM phosphate buffer, pH 7.1). The UV absorption at 260 nm was monitored as a function of temperature using a Cary 300 spectrometer. The temperature was increased at a rate of 1 °C/min from 5 to 90 °C. The melting temperatures of the native and modified oligonucleotides were calculated by determining the first derivatives of the melting curve. The cell path length was 1.0 cm.

Circular Dichroism (CD) Measurements

CD measurements were recorded at 23 °C using the same solutions and concentrations as the T_m studies. Samples were scanned from 410 to 220 nm and averaged over 1 s in a 300 μL strain-free quartz cuvette. The corresponding raw CD spectra were subtracted from a blank, converted to molar ellipticity, and processed using the manufacturer’s software package.

Results

Phosphoramidite and Oligonucleotide Synthesis

We previously reported the synthesis of a phosphoramidite reagent of the dGuo-C8-IQ adduct and the dGuo-N²-IQ nucleoside (28–30). The synthesis of the dGuo-N²-IQ phosphoramidite is shown in Scheme 1. The key step is a Buchwald–Hartwig reaction between the N²-amino group of protected dGuo derivative 4 and 5-bromo-3-methyl-2-nitroimidazo[4,5-f′]quinoline (5). Each of these starting substrates was prepared in two steps from dGuo and IQ, respectively.²⁹ Palladium-catalyzed cross-coupling of 4 and 5 was achieved in 72% yield under typical Buchwald–Hartwig conditions. Hydrogenation of the coupling product 6 removed the O⁶-benzyl group and simultaneously reduced the nitro group to an amine, affording 7 in 80% yield. Intermediate 7 could be deprotected with tetrabutylammonium fluoride (TBAF) to give the N²-IQ-modified nucleoside (3) in 83% yield.³⁰ Alternatively, the IQ-amino group of 7 was protected with p-t-butylphenoxyacetyl (t-BuPAC) chloride, and the ribose hydroxyl groups were then deprotected with fluoride ion. Protection of the 5′-hydroxyl group as a dimethoxytrityl (DMTr) ether and phosphitylation of the 3′-hydroxyl group provided the desired phosphoramidite reagent 11.

Scheme 1.

The N²-IQ adduct was incorporated into the four oligonucleotide sequences shown in Table 1. A manual coupling protocol was utilized for incorporation of the modified base as described previously (30). Oligonucleotides 12–14 contain the recognition sequence for the NarI restriction enzyme (5′-G₁G₂CG₃CC-3′), while oligonucleotide 15 contains codon-12 of the Ras protooncogene (5′-G₁G₂T-3′). The oligonucleotides were purified by HPLC and judged to be >92% pure by CGE and were characterized by MALDI-TOF mass spectrometry (Table 1).

Table 1.

Mass Spectral Characterization of the N²-IQ-Modified Oligonucleotides Used in This Study (G = dGuo-N²-IQ)^a

	Oligonucleotide	m/z (Da)
12	5′-CTC-G1GCGCC-ATC-3′	3111A9
13	5′-CTC-GG2CGCC-ATC-3′	3776.54
14	5′-CTC-GGCG3CC-ATC-3′	3776.26
15	5′-G-GCA-G1GT-GGT-G-3′	3647.89

^aThe calculated masses for oligonucleotides 12–14 and 15 (M – H) are 3776.58 and 3647.76 Da, respectively.

Thermal Melting Studies

Oligonucleotides containing the dGuo-N²-IQ adduct were hybridized to their complement, and their stability was assessed by UV thermal melting analysis (T_m). These data are summarized in Table 2 along with the T_m values for the corresponding dGuo-C8-IQ adduct in the same sequence context for comparison (30, 31). In every example, the dGuo-N²-IQ adducts destabilized the DNA duplex more than the corresponding C8 adduct. The T_m values were lowered by 12 °C for the Ras-12 duplex (15) and between 21 and 27 °C for the NarI duplexes when compared to the corresponding unmodified duplexes. Arylamine modification of the G₃-position of the NarI recognition sequence (14) is known to give rise to two-base frameshift deletions in prokaryotic systems (32). To mimic the deletion product, 12-mer oligonucleotide 14 was hybridized to a 10-mer complement that was missing the CpG dinucleotide unit opposite the modified G₃-position. This duplex was 16 °C more stable than the unmodified 12•10 duplex (Figure 2). In fact, the thermal stability of oligonucleotide 14 opposite a two-base deletion was 10° greater than when opposite dCyd of a full-length complement (entry d vs c). The melting of the modified oligonucleotide 14 opposite a two-base deletion occurs over a broader range than when paired with a full-length complement, which may be indicative of a less cooperative denaturation (33).

Table 2.

Thermal Melting Temperatures (T_m) of Oligonucleotide Duplexes Containing the N²-IQ Adduct

Oligonucleotides a			Tm (ΔTm)b,c
			dGuo-N2-IQd	dGuo-C8-IQe
a	5′-CTC GGC GCC ATC-3′	(12)	38° (−27°)	58° (−7°)
	3′-GAG CCG CGG TAG-5′
b	5′-CTC GGC GCC ATC-3′	(13)	40° (−25°)	60° (−5°)
	3′-GAG CCG CGG TAG-5′
c	5′-CTC GGC GCC ATC-3′	(14)	44° (−21°)	61° (−4°)
	3′-GAG CCG CGG TAG-5′
d	5′-CTC GGC GCC ATC-3′	(14)	54° (+16°)	48° (+10°)
	3′-GAG CCG --G TAG-5′
e	5′-G GCA GGT GGT G-3′	(15)	48° (−12°)	51° (−9°)
	3′-C CGT CCA CCA C-5′

^aG is the modified dGuo.

^bΔT_m = T_m (modified) − T_m (unmodified). The T_m for the unmodified NarI duplex is 65 °C. The T_m for the unmodified NarI oligonucleotide opposite a two-base deletion (entry d) is 38 °C.

^cConditions: 10 mM phosphate buffer (pH 7.0) containing 100 mM NaCl, 0.05 mM EDTA, and 0.5 A₂₆₀/mL of each oligonucleotide. The temperature was raised 1 °C min⁻¹.

^dThis work.

^eRefs 30 and 31.

Figure 2.

Normalized T_m curves for the dGuo-N ²-IQ-modified oligonucleotide 14 paired with a full-length complement (––,T_m = 44 °C) and a two-base deletion (– – –, T_m = 54 °C) and the unmodified oligonucleotide paired with a full-length complement (- - -, T_m = 65 °C) and a two-base deletion (– - –, T_m = 38 °C).

Absorption Spectroscopy

A representative UV absorption spectrum of a dGuo-N²-IQ containing oligonucleotide (14) is shown in Figure 3A. The intensity of the characteristic long wavelength absorbance of the IQ chromophore is very weak, and no discernable change was observed upon duplex formation. Similarly, we were unable to observe an induced CD absorbance for the IQ chromophore (Figure 3B) for duplexes 12–14. The presence of the dGuo-N²-IQ adduct resulted in modest shifts in the CD absorbances assigned to the DNA helix. In the case of duplexes 12 and 14, the intensities of these CD absorbances were virtually unchanged, while it is reduced by approximately 50% in duplex 13.

Figure 3.

(A) UV spectrum of oligonucleotide 14, single strand (– – –) and duplex (—). (B) CD spectrum of duplexes containing modified oligonucleotides 12 (- - -), 13 (– – –), and 14 (– - –) and the unmodified NarI sequence (—).

Discussion

Because of the wide exposure of HCAs through the diet, there is significant interest in the chemistry and biology of this class of DNA adducts. DNA adducts of the HCAs have been understudied as compared to simple carbocyclic arylamines such as AF and AAF (34–36). Site specifically modified oligonucleotides containing C8 adducts of simple arylamines such as AF have been prepared primarily by a biomimetic approach, in which the aryl nitrenium ion is reacted with an oligonucleotide and the products are separated (37–39). This process is laborious, and sufficient quantities of the minor N² adducts are particularly difficult to obtain from the biomimetic route.

The Buchwald–Hartwig palladium-catalyzed cross-coupling of primary and secondary amines with aryl halides and sulfonates has found wide applicability in organic synthesis (40, 41). This reaction has proven exceptionally useful for the preparation of modified nucleosides including carcinogen adducts at N²-dGuo, C8-dGuo, N⁶-dAdo, and C8-dAdo (42, 43). Our laboratory reported the synthesis of dGuo-C8-IQ adducts utilizing the Buchwald–Hartwig reactions and its incorporation into oligonucleotides (29, 30). Other laboratories subsequently prepared C8-dGuo adducts of PhIP, MeIQx, and simple arylamines via a similar approach (44–51).

Waterhouse and Rapoport showed that IQ undergoes electrophilic aromatic bromination at the C5-position, giving 5-bromo-IQ in good yield (52). The amino group of 5-bromo-IQ required protection for the cross-coupling to the N²-amino group of a suitably protected dGuo derivative such as 4. A number of amide, formamidine, and silyl protecting groups were tried, but none of these derivatives underwent cross-coupling with 4 in serviceable yield. Buchwald–Hartwig reactions involving the exocyclic amines of protected dAdo and dGuo derivatives have been noted to proceed in improved yields when electron-deficient aryl halides were used (42, 43, 53, 54). We oxidized the N²-amino group to a nitro (NaNO₂, H₂O, AcOH, 71%) to activate this aryl halide toward cross-coupling (55, 56); this transformation had the added benefit of acting as a protecting group for the IQ amino group. After cross-coupling, the nitro group was reduced to the amine under the same conditions used to remove the O⁶-benzyl ether. Intermediate 7 could then be deprotected (TBAF, THF, 84 %) to give the dGuo-N²-IQ nucleoside (3); the synthesis of 3 required five steps and proceeded in ~22% overall yield from IQ (28). Alternatively, 7 was converted to phosphoramidite reagent 9 under well-established protocols (57). The synthesis of phosphoramidite 9 required eight linear steps from either dGuo or IQ and proceeded in ~10% overall yield.

The synthesis of phosphoramidite 9 allowed for the site-specific incorporation of the dGuo-N²-IQ adduct into oligonucleotides. The adduct was incorporated at all three positions of the NarI recognition sequence (5′-G₁G₂CG₃CC-3′) (12–14). This sequence contains a frameshift-prone CG-dinucleotide repeat sequence and is a hotspot for arylamine modification (58). The dGuo-N²-IQ adduct was also incorporated into the G₁-position of codon-12 of the Ras proto-oncogene (5′-G₁G₂T-3′) (15); single base mutations in this sequence are associated with loss of Ras GTPase function and activation of the Ras signaling pathway (59, 60). Activated Ras oncogenes have been identified in a large number of human tumors and are particularly frequent in colorectal cancers (59). Modification of codon-12 has been observed in rats when their diet was supplemented with IQ and other HCAs (61–67). DNA adducts of polycyclic aromatic hydrocarbons (PAHs) have been well-studied in this sequence (60).

The thermal stability of the dGuo-N²-IQ-modified duplexes was examined by UV melting analysis and compared to the dGuo-C8-IQ adduct in the same sequence contexts (30, 31). In each sequence, the N² adduct was found to be more destabilizing than the dGuo-C8-IQ adduct. This observation is somewhat surprising in light of the data from animal feeding studies suggesting that the N² adduct is less efficiently repaired (24, 25). There is a significant disparity in the change in T_m for the Ras-12 and NarI duplexes. The T_m for the dGuo-N²-IQ-modified Ras-12 duplex (Table 2, entry e) is lowered by 12 °C as compared to the unmodified duplex, while the destabilization of the NarI duplexes (Table 2, entries a–c) was between 21 and 27 °C. These values can be compared to the dGuo-N²-AAF adduct, which was found to increase the T_m of the duplex in a 5′-CGTACGCATGC-3′ (G = dGuo-N²-AAF) sequence by 5 °C over unmodified (68). The structure of this duplex showed that the dGuo-N²-AAF adduct was positioned in the minor groove of a relatively unperturbed B-form helix with intact hydrogen bonding between the modified dGuo and its complementary dCyd. The less efficient repair of the dGuo-N²-AAF adduct can be rationalized by these structure and stability studies.

More data are available for the thermal stability of dGuo-N²-PAH adducts, and in some cases, conformational information accompanies the T_m studies (69). While there is a wide range of T_m values for the dGuo-N²-PAH adducts, they typically destabilize the duplex by 5–12 °C as compared to unmodified (69–73). The (−)- and (+)-dGuo-syn-BPDE adducts have been reported to lower the T_m of the duplex 5′-CCTATAGATATCC-3′ by 15.5 and 19 °C relative to unmodified (72). In general, the T_m values for a base-displaced intercalated conformation are more modestly depressed than for a minor groove-bound conformation. In the case of PAH adducts, the exocyclic nitrogen of the DNA base (N²-dGuo or N⁶-dAdo) is covalently linked to an sp³-hybridizied carbon of the adduct; this allows for greater conformational freedom for the PAH adducts than the N²-dGuo adducts of arylamine, which are linked to the nucleobase via an sp²-hybridized carbon. In some fjord-region PAH adducts such as benzo[c]phenanthrene, the added conformational freedom allows the adduct to intercalate into the base stack with minimal perturbation of the hydrogen bonding of the modified base pair (74, 75). This structural property has been attributed to the low rate of repair of fjord-region benzo[c]phenanthrene adducts and their increased carcinogenic potential (75–77). The destabilization of the dGuo-N²-IQ adducts in the NarI sequences is larger than typical for other bulky N² adducts. Using the dGuo-N²-PAH adducts as a guide, the destabilization of the dGuo-N²-IQ-containing duplexes when paired opposite dCyd are more typical of a minor groove-bound adduct; the larger than normal influence of the dGuo-N²-IQ adduct on the T_m may be indicative of a longer-range disruption of base pairing than is observed for the PAH adducts. It should be noted that the specific conformation of the dGuo-C8-IQ adduct in the NarI sequence was dependent on the local sequence context adopting a base-displaced intercalated conformation when situated at the iterated G₃-position and minor groove-bound conformations at the G₁- and G₂-positions (31, 78, 79).

NMR studies of the dGuo-C8-AF and dGuo-C8-AAF adducts positioned at the frameshift prone G₃-position of the NarI recognition sequence opposite a two-base deletion have been reported and revealed that the adduct was intercalated into the DNA stack (80, 81). The C8-modified dGuo and neighboring dCyd were displaced into the major groove as part of an extrahelical bulge. To our knowledge, there have been no reported structural studies of a dGuo-N²-arylamine adduct opposite a deletion site. The conformations of three stereoisomeric dGuo-N²-BPDE adducts (Figure 4) opposite a one-base deletion in a 5′-CCATCGCTACC-3′ sequence have been determined in which the modified dGuo was flanked by dCyds as is the modified dGuo of oligonucleotide 14 (82–84). In each case, the dGuo-N²-BPDE adduct adopted a base-displaced intercalated conformation; however, the relative orientation of the adduct and the displacement of the modified guanine differed depending on the adduct stereochemistry. The modified guanine was displaced into the major groove for the (+)-trans-anti isomer (82), while it was displaced into the minor groove for the (−)-trans-anti- and (−)-cis-anti isomers (83, 84). The modified duplexes opposite a one-base deletion were stabilized vs the unmodified according to T_m analysis. The (−)-trans-anti-and (−)-cis-anti isomers in which the modified guanines were displaced into the minor groove stabilized the duplex opposite a one-base deletion by +14 and +25 °C, respectively, while the (+)-trans-anti- isomer added +6 °C to the T_m vs unmodified (72). The (−)-cis-anti isomers were ~4 °C more stable than the adduct opposite dCyd in a full-length complement, while the others were 7–10 °C less stable under these conditions. The duplex containing the dGuo-N²-IQ opposite a two-base deletion was 16 °C more stable than the corresponding unmodified duplex and 10 °C more stable than the duplex paired with dCyd of a full-length complement. On the basis of the available literature, it is likely that the dGuo-N²-IQ adduct adopts a base-displaced intercalated conformation when opposite a deletion site and that the T_m data are more consistent with that of the dGuo-N²-BPDE adduct in which the modified dGuo is displaced into the minor groove.

Figure 4.

dGuo-N²-benzo[a]pyrene adducts.

We were able to make preliminary conformational assignments for the dGuo-C8-IQ adducts based on UV absorption and CD spectroscopy, and these predictions were subsequently confirmed by a full structural analysis by a combination of high-field NMR and restrained molecular dynamics (31, 78, 79). The conformation of dGuo-N²-PAH adducts was also correlated to their UV spectra (85). Unfortunately, the chemical linkage between C5 of the IQ ring system to the N²-position of dGuo significantly altered the IQ chromophore. As a result, conformational preferences of dGuo-N²-IQ adduct could not be discerned from the UV and CD spectra.

We have also reported the in vitro bypass of the N²- and C8-dGuo adducts of IQ when positioned at the noniterated G₁-and iterated G₃-positions of the NarI recognition sequence by prokaryotic and human DNA polymerases (32, 86). We found that replication of the dGuo-C8-IQ by the prokaryotic polymerases Escherichia coli DNA polymerase I Klenow fragment exo⁻ (Kf⁻), E. coli DNA polymerase II exo⁻ (pol II⁻), and Sulfolobus solfataricus P2 DNA polymerase IV (Dpo4) resulted in two-base deletions when positioned at the iterated G₃-positions, and error-free bypass and extension when positioned at the G₁-position. The dGuo-N²-IQ was a stronger block to replication, but error-free products were observed when bypassed. Many DNA polymerases make contact with the DNA duplex in the minor groove, and conformations of the dGuo-N²-IQ adduct that possess minor groove perturbations would represent significant hurdles to bypass. Of the human Y-family polymerases, only pol η was able to bypass the IQ adducts with any appreciable efficiency. Replication of the dGuo-C8-IQ resulted in error-free products independent of sequence. Interestingly, a −2 deletion product was observed from bypass of dGuo-N²-IQ at the iterated G₃-position by human pol η. While IQ is a potent inducer of frameshift mutations in bacterial systems, studies have shown that frameshift mutations become a minor mutational event when mammalian cells are treated with IQ and other arylamines (1–5, 87). Because the dGuo-N²-IQ adduct is the minor product accounting for 20–40% of the total IQ adduct level and gives rise to frameshifts only in a specific sequence context, it is not surprising that two-base deletions are observed at low frequency in mammalian systems.

Summary

We have described a reasonably efficient synthesis of a suitable phosphoramidite reagent for the solid-phase synthesis of oligonucleotides containing the minor N²-dGuo adduct of the dietary mutagen IQ. The dGuo-N²-IQ adduct was incorporated into all three dGuo positions of the NarI recognition sequence (12–14) and the G₁-position of codon-12 of the Ras gene (15). The modified duplexes were significantly destabilized vs the unmodified duplex as measured by T_m analysis; the destabilization was also larger than the corresponding dGuo-C8-IQ adduct in identical sequences. We were unable to assign the conformation of the dGuo-N²-IQ modified duplexes by UV and CD spectroscopy as previously described for the corresponding dGuo-C8-IQ adducts. A full structural analysis of the dGuo-N²-IQ modified duplexes would be of great interest.