Structures of exocyclic R,R- and S,S-N(6),N(6)-(2,3-dihydroxybutan-1,4-diyl)-2'-deoxyadenosine adducts induced by 1,2,3,4-diepoxybutane.

1,3-Butadiene (BD) is an industrial and environmental chemical present in urban air and cigarette smoke, and is classified as a human carcinogen. It is oxidized by cytochrome P450 to form 1,2,3,4-diepoxybutane (DEB); DEB bis-alkylates the N(6) position of adenine in DNA. Two enantiomers of bis-N(6)-dA adducts of DEB have been identified: R,R-N(6),N(6)-(2,3-dihydroxybutan-1,4-diyl)-2'-deoxyadenosine (R,R-DHB-dA), and S,S-N(6),N(6)-(2,3-dihydroxybutan-1,4-diyl)-2'-deoxyadenosine (S,S-DHB-dA) [ Seneviratne , U. , Antsypovich , S. , Dorr , D. Q. , Dissanayake , T. , Kotapati , S. , and Tretyakova , N. ( 2010 ) Chem. Res. Toxicol. 23 , 1556 -1567 ]. Herein, the R,R-DHB-dA and S,S-DHB-dA adducts have been incorporated into the 5'-d(C(1)G(2)G(3)A(4)C(5)X(6)A(7)G(8)A(9)A(10)G(11))-3':5'-d(C(12)T(13)T(14)C(15)T(16)T(17)G(18)T(19)C(20)C(21)G(22))-3' duplex [X(6) = R,R-DHB-dA (R(6)) or S,S-DHB-dA (S(6))]. The structures of the duplexes were determined by molecular dynamics calculations, which were restrained by experimental distances obtained from NMR data. Both the R,R- and S,S-DHB-dA adducts are positioned in the major groove of DNA. In both instances, the bulky 3,4-dihydroxypyrrolidine rings are accommodated by an out-of-plane rotation about the C6-N(6) bond of the bis-alkylated adenine. In both instances, the directionality of the dihydroxypyrrolidine ring is evidenced by the pattern of NOEs between the 3,4-dihydroxypyrrolidine protons and DNA. Also in both instances, the anti conformation of the glycosyl bond is maintained, which combined with the out-of-plane rotation about the C6-N(6) bond, allows the complementary thymine, T(17), to remain stacked within the duplex, and form one hydrogen bond with the modified base, between the imine nitrogen of the modified base and the T(17) N3H imino proton. The loss of the second Watson-Crick hydrogen bonding interaction at the lesion sites correlates with the lower thermal stabilities of the R,R- and S,S-DHB-dA duplexes, as compared to the corresponding unmodified duplex. The reduced base stacking at the adduct sites may also contribute to the thermal instability.

Introduction

1,3-Butadiene (BD)[1] is used to manufacture styrene-butadiene rubber (SBR).[1,2] Several billion pounds of SBR is produced annually in the United States. It is also a combustion product from automobile emissions[3] and cigarette smoke.[4] BD is genotoxic and is carcinogenic in rodents,[5,6] particularly in mice,[7−9] and less potent in rats.[10−13] These species-specific differences in genotoxicity and carcinogenicity have been attributed to differences in various pathways of BD metabolism.[14] In the SBR industry, occupational exposures to BD are associated with increased risk for hematopoietic cancers.[15−30] Consequently, BD has been classified by the United States Environmental Protection Agency as “carcinogenic to humans by inhalation”,[31] and it has been also characterized as a known human carcinogen by the National Toxicology Program.[32] The International Agency for Cancer Research (IARC) lists BD as “carcinogenic to humans (Group 1).”[33,34] Accordingly, there has been interest in identifying biomarkers of exposures to BD.[35,36]

Delineating the molecular basis for BD-induced genotoxicity is complicated by the formation of multiple electrophilic species derived from BD metabolism, their abilities to alkylate multiple sites in DNA, and the stereochemistry of the resulting adducts.[30,37−39] BD is oxidized by cytochrome P450s to form 1,2-epoxy-3-butenes (EB).[40] As shown in Scheme 1, EB can either undergo hydrolysis to form 1,2-dihydroxy-3-butene or further oxidation to 1,2,3,4-diepoxybutane (DEB). Hydrolysis of DEB can form 1,2-dihydroxy-3,4-epoxybutane (EBD). Albertini, Kirman, and co-workers have reviewed BD metabolism and genotoxicity.[30,41,42] Although DEB is a minor metabolite of BD, it is 50- to 100-fold more genotoxic and mutagenic in human cells than the monoepoxide metabolites, EB and EBD.[43,44] Three stereoisomers of DEB, S,S, R,R, and meso, are generated metabolically.[45] Of these, the S,S isomer is the most cytotoxic[46,47] and mutagenic.[48] In mice, levels of DEB have been measured at ∼250 pmol/g in blood.[49] While the genotoxicity of DEB has been attributed to its ability to cross-link DNA via the N7 position of guanine,[50,51] it induces a significant number of A to T transversions,[44,52,53] suggesting the formation of adenine adducts. A number of these have been identified, including the stereospecific R,R- and S,S-N6,N6-(2,3-dihydroxybutan-1,4-diyl)-2′-deoxyadenosine adducts (N6,N6-DHB-dA),[54] which arise from bis-alkylation of the N6 exocyclic amino groups of adenines in DNA by R,R- and S,S -DEB (Chart 1).

Scheme 1

Cytochrome P450-Mediated Oxidation of BD to Stereoisomeric EB, DEB, and EBD Electrophiles

Chart 1

(A) Structures of the R,R-DHB-dA and S,S-DHB-dA Adducts, and (B) Sequences and Numbering of the Unmodified and Modified Duplexesa

In A, Hα and Hα′ are pro-R protons; and Hβ, Hβ′ are pro-S protons. In B, at the 6th position adenine was replaced with adduct R6 (R,R-DHB-dA) to form the R,R-duplex or adduct S6 (S,S-DHB-dA) to form the S,S-duplex.

Presently, we have used NMR spectroscopy to determine the structures of the R,R- and S,S-N6,N6-DHB-dA adducts[54] incorporated into the oligodeoxynucleotide duplexes 5′-d(C1G2G3A4CRAG8A9A10G11)-3′:5′-d(C12T13T14C15T16T17G18T19C20C21G22)-3′ (R = R,R-DHB-dA) and 5′-d(C1G2G3A4CSAG8A9A10G11)-3′:5′-d(C12T13T14C15T16T17G18T19C20C21G22)-3′ (S = S,S-DHB-dA) (Chart 1). The sequence of these duplexes arises from codon 61 and the surrounding sequence of the ras proto-oncogene, and the unmodified duplex has been spectroscopically characterized and examined as to its structure in solution.[55] As compared to the unmodified duplex, both the R,R-DHB-dA and S,S-DHB-dA adducts thermally destabilize the DNA. The structures of the duplexes containing either the R6 or S6 adducts shows that, in both instances, the DHB moiety rotates around the C6-N6 bond and is located in the major groove of DNA. DHB-dA adducts can only form one Watson–Crick hydrogen bond, between N1 of R6 or S6 and the T17 N3H imino proton in the complementary strand. These results reveal that the R6 and S6 DHB-dA adducts destabilize DNA by disrupting Watson–Crick hydrogen bonding and affecting stacking interactions with neighboring base pairs.

Materials and Methods

Synthesis of Modified Oligodeoxynucleotides

The 2′-deoxyribonucleoside-3′-phosphoramidites iPr-PAC-dG-CE, PAC-dA-CE, and Ac-dC-CE, dT-CE, the iPr-PAC-dG-CPG ABI columns, and other reagents necessary for automated DNA synthesis were purchased from Glen Research (Sterling, VA). The 5′-O-(4,4′-dimethoxytrityl)-3′-O-(2-cyanoethyl)-N,N-diisopropyl phosphoramidite of the 6-chloropurine-2′-deoxyriboside was purchased from ChemGenes Co. (Wilmington, MA). Oligodeoxynucleotides were synthesized by solid phase methods using an ABI 394 DNA synthesizer (Life Technologies, Carlsbad, CA). All solvents and chemical reagents were obtained from commercial sources and used without further purification. The modified oligodeoxynucleotides 5′-d(CGGACRAGAAG)-3′ and 5′-d(CGGACSAGAAG)-3′ containing site- and stereospecific N6,N6-DHB-dA lesions were synthesized by a postoligomerization approach,[56] which was originally developed by Harris, and co-workers.[57−59] Briefly, 11-mer oligodeoxynucleotides containing a site-specific 6-chloropurine at position R6 or S6 (100 nmol) were coupled with (R,R)- or (S,S)-pyrrolidine-3,4-diol (20 mg), respectively, in the presence of DIPEA (200 μL) in DMSO (300 μL) for 72 h at 37 °C. The resulting structurally modified oligodeoxynucleotides were purified and desalted by reverse phase HPLC, characterized by capillary HPLC-ESI-MS, and quantified by UV spectrophotometry. Sequence and site-specificity were confirmed by MALDI-MS of partial exonuclease digests.[56]

The unmodified oligodeoxynucleotides 5′-d(CGGACAAGAAG)-3′ and 5′-d(CTTCTTGTCCG)-3′ were synthesized by the Midland Reagent Company (Midland, TX) and purified by anion-exchange HPLC. The oligodeoxynucleotides were purified using a semipreparative scale reverse-phase HPLC column (YMC, Kyoto, Japan, Phenyl-Hexyl, 5 μm, 250 mm ×10.0 mm) equilibrated with 0.1 M ammonium formate (pH 7.0) using an acetonitrile gradient. The oligodeoxynucleotides were desalted by elution from Sephadex G-25, lyophilized, and characterized by MALDI-TOF-MS.

The concentrations of single-stranded oligodeoxynucleotides were determined by UV absorbance at 260 nm using extinction coefficients of 118,300 L M–1cm–1 for strands 5′-d(CGGACRAGAAG)-3′, 5′-d(CGGACSAGAAG)-3′, and 5′-d(CGGACAAGAAG)-3′ and assuming similar extinction coefficients for N6,N6-DHB-dA and dA, and 90,800 L M–1cm–1 for the complementary strand 5′-d(CTTCTTGTCCG)-3′.[60] To assemble the respective duplexes, the complementary strand 5′-d(CTTCTTGTCCG)-3′ was combined with equal molar quantities of either 5′-(CGGACAAGAAG)-3′, 5′-d(CGGACRAGAAG)-3′, or 5′-d(CGGACSAGAAG)-3′, heated to 75 °C for 15 min, and cooled to room temperature to form the unmodified, and R,R- and S,S-duplexes, respectively.

UV Melting Studies

Absorption vs temperature profiles (UV melts) for each duplex were measured using a Varian Cary 100 Bio spectrophotometer (Varian Associates, Palo Alto, CA). The concentrations of the duplexes were 2.1 μM. Samples were prepared in a buffered solution of 10 mM NaH2PO4 (pH 7.0) and 50 μM Na2EDTA containing either 0.1 M NaCl or 1 M NaCl. The temperature was increased from 5 to 85 °C for each duplex at a rate of 0.50 °C/min. The UV absorbance was monitored at 260 nm.

NMR Spectroscopy

The modified double-stranded duplexes containing the R6 or S6 adducts were prepared in 10 mM NaH2PO4 (pH 7.0) containing 0.1 M NaCl and 50 μM Na2EDTA at 0.7 mM and 0.4 mM concentrations, respectively. To observe the nonexchangeable protons, the duplexes were exchanged with D2O and dissolved in 99.96% D2O. For the observation of exchangeable protons, the samples were dissolved in 9:1 H2O:D2O. 1H NMR spectra were recorded using 800 MHz, 600 MHz, and 500 MHz spectrometers equipped with cryogenic probes (Bruker Biospin Inc., Billerica, MA). Chemical shifts were referenced to the chemical shift of water resonance at the corresponding temperature, with respect to 4,4-dimethyl-4-silapentane-1-sulfonic acid (DSS). Data were processed using the program TOPSPIN (Bruker Biospin Inc., Billerica, MA). The NOESY[61,62] and DQF-COSY[63] spectra in D2O were collected at 15 °C at 800 MHz and for the unmodified duplex only at 600 MHz; NOESY experiments were conducted at a mixing time of 60, 150, 200, and 250 ms. These experiments were performed with a relaxation delay of 2.0 s. The NOESY spectra in 9:1 H2O:D2O were collected at 5 °C at 500 MHz for modified duplexes and 600 MHz for the unmodified duplex, with a 250 ms mixing time. NMR experiments in 9:1 H2O:D2O were collected at 5, 10, 15, 20, 25, and 30 °C at 800 MHz for the S,S-duplex, and 600 MHz for the R,R and unmodified duplexes. These experiments were performed with a relaxation delay of 1.5 s. Water suppression was performed using the WATERGATE pulse sequence.[64]

NMR Experimental Restraints

The NOESY spectra of the duplexes measured in D2O were processed using the TOPSPIN software and were evaluated using the program SPARKY[65] to obtain the cross-peak assignments. The cross-peak intensities were measured by volume integrations. The peak volumes were divided into classes based on the confidence in integrations of the cross-peaks. For well-resolved strong nonoverlapping cross-peaks, a volume integration error of 10% was assigned. For weaker or more overlapping cross-peaks, for which the confidence in volume integration was lower, errors of 20, 30, 40, or 50% were applied. Experimental intensities were combined with intensities calculated from complete relaxation matrix analysis of a starting model to generate a hybrid intensity matrix.[66,67] Conversion from peak volumes into distances, including upper and lower bounds, was completed using the program MARDIGRAS,[68] which refined the hybrid intensity matrix.[68] For methyl protons, the JUMP 3[69] model was employed. Calculations were performed using 150, 200, and 250 ms mixing time data and 2, 3, and 4 ns isotropic correlation times. Evaluations of these results, e.g., for spin diffusion effects, provided distance restraints used for restrained molecular dynamics (rMD) calculations. Additional empirical restraints were obtained from canonical values derived from B-DNA.[70] Empirical Watson–Crick base pair restraints were employed with the exception of the modified base pairs. Pseudorotation restraints were also used, but with the exceptions of the terminal bases C1, G11, C12, G22, and the modified nucleotides R6 or S6. For unmodified nucleotides, empirical phosphodiester backbone angles were restrained by square well potentials with widths of ±60°. For the modified nucleotides R6 or S6, the widths of the square well potentials were increased to ±120°.

Restrained Molecular Dynamics Calculations

A B-type DNA duplex was constructed. The adenine at position A6 was replaced by either the R6 or the S6 adduct using the program INSIGHT II (Accelrys Inc., San Diego, CA). The program AMBER,[71] including the parm99 force field,[72] was used. Partial charges for the R,R- and S,S-DHB-dA adducts were calculated with the B3LYP/6-31G* basis set in GAUSSIAN,[73] and they were employed in the parameter files prepared in the program XLEAP.[74] Tables S1 and S2 in the Supporting Information provide the parametrization utilized for the R,R- and S,S-DHB-dA nucleotides. The modified duplexes were each subjected to 1000 cycles of potential energy minimization. The simulated annealing protocol[75] used for the rMD calculations utilized force constants of 32 kcal mol–1 Å–2 for distance restraints. The generalized Born model[76] was used for solvation. The salt concentration was 0.1 M. The molecule was coupled to the bath temperature to establish the temperature during simulated annealing calculations.[77] Initially, calculations were performed for 20 ps (20,000 steps). For the first 1,000 steps, the system was heated from 0 to 600 K with a coupling of 0.5 ps, followed by 1,000 steps at 600 K, followed by 16,000 steps in which the system was cooled to 100 K with a coupling of 4 ps. In the last 2,000 steps, additional cooling was applied from 100 to 0 K with a coupling of 1 ps. Subsequently, a 100,000 step calculation was performed over 100 ps. For the first 5,000 steps, the system was heated from 0 to 600 K with a coupling of 0.5 ps, followed by 5,000 steps at 600 K, followed by 80,000 during which the system was cooled to 100 K with a coupling of 4 ps, followed by additional cooling for the last 10,000 steps with a coupling of 1 ps. Structure coordinates were saved after each cycle. Complete relaxation matrix analysis (CORMA)[66,67] was used to compare NOE intensities calculated from these emergent structures with the experimentally measured distances. For each modified duplex, eight structures were chosen, based on the lowest deviations from the experimental distance and dihedral restraints. These were subjected to potential energy minimization, and used to obtain average refined structures. Helicoidal analyses were performed using the CURVES+ web server.[78]

Data Deposition

The structure factors and coordinates were deposited in the Protein Data Bank (www.rcsb.org). The PDB ID code for the duplex containing the R,R-DHB-dA adduct is 2MHX, and for the duplex containing the S,S-DHB-dA adduct the PDB ID code is 2MHZ.

Results

The unfolding of the duplexes was studied by temperature-dependent UV spectroscopy, monitored at 260 nm. The Tm values were determined by taking the first derivatives of the melting curves and shape analyses. For the unmodified duplex, at a concentration of 2.1 μM in 0.1 M NaCl at pH 7, the Tm was 44 °C. At the same concentration and under the same conditions, the Tm for the R,R-duplex was 28 °C and for the S,S-duplex was 27 °C. Thus, both the R,R- and S,S-duplexes resulted in similar thermal destabilization of the DNA, irrespective of adduct stereochemistry. The Tm experiments were repeated in 1 M NaCl at pH 7. Under these conditions, the Tm of the unmodified duplex was 53 °C, while the melting temperatures of the R,R- and S,S-duplexes were 36 and 35 °C, respectively. Again, both the R,R- and S,S-duplexes thermally destabilized the DNA, to approximately the same extent for both-duplexes, irrespective of adduct stereochemistry.

NMR Spectroscopy of the R,R-DHB-dA Duplex

Base Proton Assignments

Figure 1 (panels A and B) shows the region of the NOESY spectrum including the base aromatic proton resonances and deoxyribose H1′ proton resonances[79,80] for the modified strand C1 → G11 and the complementary strand C12 → G22. For the modified strand, the NOE cross-peaks adjacent to the R,R-DHB-dA adduct, C5 H6 → C5 H1′, C5 H1′ → R6 H8, R6 H8 → R6 H1′, and R6 H1′ → A7 H8 were of similar intensities as compared to the cross-peaks arising from distal nucleotides. For the complementary strand, among the cross-peaks of bases proximal to the DHB-dA adduct, those arising from T16, T17, and G18 were of similar intensities compared to the remainder of the cross-peaks from nucleotides distal to the adduct. The greatest chemical shift perturbations as compared to the unmodified-duplex (Figure S1 in the Supporting Information) were observed in the modified strand, for C5 H5, which shifted 0.1 ppm upfield, C5 H1′, which shifted 0.3 ppm downfield, R6 H8, which shifted 0.1 ppm upfield, and R6 H2, which shifted 0.1 ppm downfield. Minimal chemical shift perturbations were observed for the complementary strand.

Figure 1

NOESY spectra for the R,R- and S,S-DHB-dA modified-duplexes, showing sequential NOEs between the base aromatic and anomeric protons. (A) The modified strand for the R,R-duplex, showing nucleotides C1 to G11. (B) The complementary strand for the R,R-duplex, showing nucleotides C12 to G22. (C) The modified strand for the S,S-duplex, showing nucleotides C1 to G11. (D) The complementary strand for the S,S-duplex, showing nucleotides C12 to G22. Spectra were obtained at 800 MHz with a mixing time of 250 ms at 15 °C.

Imino Proton Assignments

The sequential pattern of cross-peaks between imino protons[81] was observed for base pairs G2:C21 → G3:C20 → A4:T19 → C5:G18. The T16 N3H → T17 N3H and T17 N3H → G18 N1H sequential cross-peaks were not observed, indicative of structural changes in the vicinity of the adduct. The sequential walk resumed from G8:C15 → A9:T14 → A10:T13 (Figure 2, panels A–C). The NOEs between the base imino and amino protons and adenine H2 protons showed cross-peaks for all base pairs, with an exception of the terminal base G22 (Figure 2B). The A7 H2 → T16 N3H cross-peak was of lower intensity compared to the other cross-peaks and overlapped with the G11 N1H → C12NH1 cross-peak. The R6 H2 → T17 N3H cross-peak (k, Figure 2, panel B) was visible when the contour levels of the spectrum were lowered; its intensity was significantly lower than the remainder of the A:T cross-peaks. The greatest chemical shift perturbation was observed for the T17 N3H proton, which shifted 0.35 ppm upfield as compared to that of the unmodified-duplex (Figure S2 in the Supporting Information). The T16 and T17 imino protons resonances (Figure S3 in the Supporting Information) were broader and of lower intensity as compared to that of the unmodified-duplex (Figure S4 in the Supporting Information) at 5 °C, and when the temperature was increased to 20 °C, both resonances disappeared (Figure S3 in the Supporting Information).

Figure 2

NOESY spectra for the R,R- and S,S-duplexes, showing NOEs between the base imino protons and the amino protons, and DHB protons. (A) Interstrand NOEs between the R6 adduct and the G18 base. The cross-peaks are assigned as a, R6 Hβ′/α′ → G18 N1H; b, R6 Hβ′/α′ → G18 N1H; c, R6 Hβ/α → G18 N1H; d, R6 Hβ/α → G18 N1H; e, R6 Hγ′ → G18 N1H; and f, R6 Hγ → G18 N1H. (B) Interstrand NOEs between complementary bases. The cross-peaks are assigned as g, A10 H2 → T13 N3H; h, A7 H2 → T16 N3H; i, A9 H2 → T14 N3H; j, A4 H2 → T19 N3H; k, R6 H2 → T17 N3H and C12N H1 → G11 N1H; l, C21NH1 → G2 N1H; m, C21NH2 → G2 N1H; n, C20NH1 → G3 N1H; o, C20NH2 → G3 N1H; p, C15NH1 → G8 N1H; q, C15NH2 → G8 N1H; r, C5NH1 → G18 N1H; and s, C5NH2 → G18 N1H. The intensity of the cross-peak k in the black box is lower than the rest of the cross-peaks, and the contour of this section was adjusted to be visible in the plot. (C) NOE connectivity for the imino protons for the base pairs G2:C21, G3:C20, A4:T19, C5:G18, A7:T16, G8:C15, A9:T14, and A10:T13. The cross-peaks are T13 N3H → T14 N3H, T14 N3H → G8 N1H, G8 N1H → T16 N3H, G18 N1H → T19 N3H, T19 N3H → G3 N1H, and G3 N1H → G2 N1H. For clarity, in panel A the contours are represented at 2× the intensity of those in panel B, in panel B the contour of cross-peak k (in black box) is presented at 4× the intensity of the remainder of the cross-peaks, and in panel C the contours are represented at 4× the intensity of those in panel B. (D) Interstrand NOEs between the S6 nucleotide and G18. The cross-peaks are assigned as a, S6 Hα′,β′, Hα,β → G18 N1H; and b, S6 Hα,β, Hγ′ → G18 N1H. (E) Interstrand NOEs between complementary bases. The cross-peaks are assigned as c, A10 H2 → T13 N3H; d, A7 H2 → T16 N3H; e, A9 H2 → T14 N3H; f, A4 H2 → T19 N3H; g, S6 H2 → T17 N3H and C12N H1 → G11 N1H; h, C21NH1 → G2 N1H; i, C21NH2 → G2 N1H; j, C20NH1 → G3 N1H; k, C20NH2 → G3 N1H; l, C15NH1 → G8 N1H; m, C15NH2 → G8 N1H; n, C5NH1 → G18 N1H; and o, C5NH2 → G18 N1H. (F) NOE connectivity for the imino protons for the base pairs G2:C21, G3:C20, A4:T19, C5:G18, A7:T16, G8:C15, A9:T14, and A10:T13. The cross-peaks are T13 N3H → T14 N3H, T14 N3H → G8 N1H, G8 N1H → T16 N3H, G18 N1H → T19 N3H, T19 N3H → G3 N1H, and G3 N1H → G2 N1H. For clarity, in panel D the contours are represented at 4× the intensity of those in panel E, in panel E the contour of cross-peak g (black box) is presented at 4× the intensity of the remainder of the cross-peaks, and in panel F the contours are represented at 4× the intensity of those in panel E. The spectra were obtained at 500 MHz with a mixing time of 250 ms at 5 °C.

R,R-DHB-dA Proton Assignments

The R,R-DHB-dA adduct is characterized by a 2-fold rotation axis about the C6-N6 bond. The pro-R and pro-S diastereotopic hydrogens on the two methylene groups are symmetry related with respect to rotation about this bond, as are the corresponding hydroxyl and hydrogen substituents on the stereoisomeric carbons of the DHB moiety. Accordingly, the two pro-R methylene hydrogens are designated Hα and Hα′, the two methylene pro-S hydrogens are designated Hβ and H β′, and the stereoisomeric hydrogens on the hydroxylated carbons are designated Hγ and Hγ′ (Chart 1). Six resonances were observed between 3.6 and 4.4 ppm (Figure 3A). This indicated that on the NMR time scale, the DHB moiety did not rapidly rotate about the C6-N6 bond and that the nonsymmetric DNA environment split the three sets of symmetry-related hydrogens into six resonances. This observation was consistent with previous calculations that indicated partial double bond character for the C6-N6 bond.[56] These resonances were observed in the same region of the spectrum as the resonances for the 1,4-bis(2′-deoxyadenosin-N6-yl)-2S,3S-butanediol intrastrand DEB-derived cross-link in the same sequence.[82] For the two sets of methylene protons, i.e., Hα,β and Hα′,β′ (Chart 1), it was not possible to establish unequivocally which was the resonance arising from the pro-R hydrogen and which was the resonance arising from the pro-S hydrogen. However, it was possible to assign the two sets of methylene resonances, based upon NOEs to DNA protons. One set of methylene resonances, assigned as the Hα,β hydrogens, was located at 3.88 and 3.95 ppm, and the other set, assigned as the Hα′,β′ hydrogens, was located at 3.64 and 3.73 ppm. The Hα′,β′ set of resonances exhibited a strong NOE to C5 H5 and weak NOEs to C5 H6 [not observed at the contour level plotted in Figure 3]. This set of methylene protons exhibited weak NOEs to A7 H2 in the 3′-direction. These methylene protons exhibited NOEs to the R6 H2 resonance at 7.27 ppm. They also exhibited strong NOEs to the major groove T16 CH3 protons, located in the 3′-neighbor A7:T16 base pair. The Hα,β methylene protons exhibited strong NOEs to C5 H5 and weaker NOEs to C5 H6. This set of methylene protons did not exhibit NOEs to R6 H2. The diastereomeric Hγ and Hγ′ protons were identified at 4.31 and 4.29 ppm. It was possible to make the assignments of the Hγ and Hγ′ based on the NOE intensities to the methylene protons. Hγ had stronger cross-peaks to the Hα,β and weaker cross-peaks to Hα′,β′. The opposite results were observed for Hγ′, which exhibited stronger cross-peaks to Hα′,β′ and weaker to Hα,β. NOEs were observed from both to the neighbor base A7 H2, C5 H6, C5 H5, and T16 CH3 protons. Weak interstrand NOEs were observed between Hα,α′, Hβ,β′ and Hγ,γ′ of R6 and the imino proton of G18 (Figure 2, panel A).

Figure 3

Expanded plot of the NOESY spectrum of the R,R- and S,S-DHB-dA modified-duplexes, showing assignments of the adduct protons and cross-peaks from the adduct protons to neighbor base protons. (A) The R,R-DHB-dA-duplex. The chemical shifts for each proton are A7 H2, 7.53 ppm; R6 H2, 7.27 ppm; C5 H6, 7.14 ppm; C5 H5, 5.10 ppm; R6 Hγ, 4.31 ppm; and R6 Hγ′, 4.29 ppm; R6 Hα and R6 Hβ could not be unequivocally assigned. They are observed at 3.95 and 3.88 ppm. R6 Hα′ and R6 Hβ′ could not be unequivocally assigned. They are observed at 3.73 and 3.64 ppm. T16 CH3, 1.68 ppm. The dashed lines show the NOE connectivity for each proton. For clarity, in the regions with black borders the contours are represented at 8× and 1× the intensity, which is indicated in the corner of each region. (B) The S,S-DHB-dA-duplex. The chemical shifts for each proton are A7 H2, 7.53 ppm; S6 H2, 7.30 ppm; C5 H6, 7.14 ppm; C5 H5, 5.14 ppm; S6 Hγ, 4.36 ppm; and S6 Hγ′, 4.25 ppm; S6 Hα and S6 Hβ could not be unequivocally assigned and were observed at 3.66 and 4.14 ppm; S6 Hα′ and S6 Hβ′ could not be unequivocally assigned and were observed at 3.62 and 3.70 ppm; T17 CH3, 1.82 ppm; and T16 CH3, 1.69 ppm. The dashed lines show connectivity for each proton. For picture clarity, in the regions the contours are represented at 32× and 4×, and 1× the intensity, which is indicated in the corner of each region. For both samples spectra were obtained at 800 MHz, with a mixing time of 250 ms. The temperature was 15 °C.

Structural Refinement of the R,R-DHB-dA Duplex

A total of 204 distance restraints obtained from the analyses of the NOESY spectra of nonexchangeable protons were used for restrained molecular dynamics (rMD) calculations, where 106 were internucleotide restraints and 98 were intranucleotide restraints. The average well width for the distance of the upper and lower bounds was 1.5 Å. A simulated annealing protocol was used. A total of 90 empirical backbone torsion angles, 42 empirical hydrogen bonding, and 17 empirical deoxyribose pseudorotations were included as experimental restraints in the calculations (Table 1). These were included based upon the observation that with the exception of the modified base pair R6:T17, and the neighboring base pairs C5:G18 and A7:T16, the remainder of the NMR data were consistent with a right-handed helical DNA-duplex similar to the canonical B-form DNA.[70] The distereotopic Hα,β and H α′,β′ protons of the DHB group could not be assigned unequivocally. In several instances, however, both exhibited similar NOE cross-peak intensities to specific DNA protons. Such distances were included in the rMD calculations. Additionally, several NOEs involving the exchangeable protons overlapped, precluding accurate distance calculations, so distances involving these exchangeable protons were not included in the rMD calculations. Table 1 summarizes the restraints that were used and the refinement statistics.

Table 1

NMR Restraints Used for the rMD Structural Refinement of the R,R- and S,S-Duplexes, and the Refinement Statistics

NMR restraints	R,R-duplex	S,S-duplex
NOE restraints
internucleotide	106	103
intranucleotide	98	96
total	204	199
backbone torsion angle restraints	90	90
hydrogen bonding restraints	42	42
deoxyribose pseudorotation restraints	17	17
total number of restraints	353	348
refinement statistics
number of distance restraint violations >0.025 Å	8	10
number of torsion restraint violations	8	7
total distance penalty/maximum penalty [kcal mol–1]	0.48/0.11	0.82/0.31
total torsion penalty/maximum penalty [kcal mol–1]	0.59/0.13	3.15/1.61
r.m.s. distances (Å)	0.012	0.012
r.m.s. angles (deg)	2.32	2.30
distance restraint force constant [kcal mol–1 Å–2]	32	32
torsion restraint force constant [kcal mol–1 deg–2]	32	32
average well width for distance restraints [Å]	1.5	1.5

Eight structures emergent from the rMD calculations were subjected to potential energy minimization. Figure S5, panel A of the Supporting Information shows these eight superimposed structures. Good convergence was observed, with a maximum pairwise rmsd between the eight structures of 0.67 Å (Table 2). These eight structures were averaged, and the resulting average structure was subjected to complete relaxation matrix analysis.[67] The results are shown in Figure 4A. In general, the sixth root residuals (R1x values) were consistently below 10%, for both intranucleotide NOEs and internucleotide NOEs. This indicated that the average of the refined structures was in reasonable agreement with the experimental NOE intensities. Base pairs A10:T13 and G11:C12, located at the 3′-end of the duplex with respect to R6, showed somewhat poorer agreement with the experimental NOE intensities. This was attributed to a combination of spectral overlap and increased fraying of the duplex at this end. Table 2 summarizes the structural statistics.

Table 2

Structural Statistics for the R,R- and S,S-Duplexes

average structure (obtained from 8 structures)	R,R-duplex	S,S-duplex
RMS pairwise difference between structures [Å]	0.67	0.66
RMS difference from average structure [Å]	0.45	0.44
CORMA analysis for average structurea
R1xb
intranucleotide	0.068	0.071
internucleotide	0.080	0.086
total	0.074	0.078
average errorc	0.016	0.016

The mixing time was 250 ms.

R1x is the sixth root R factor: Σ[((Io)i1/6) – ((Ic)i1/6)/Σ((Io)i1/6)].

Average error: Σ(Ic – Io)/n, where Ic values are NOE intensities calculated from refined structure, and Io values are experimental NOE intensities.

Figure 4

Complete relaxation matrix analysis results for internucleotide and intranucleotide NOEs for the R,R- and S,S-DHB-dA duplexes. (A) The R,R-DHB-dA duplex. (B) The S,S-DHB-dA duplex. R1x is the sixth root R factor: Σ[((Io)i1/6) – ((Ic)i1/6)/Σ((Io)i1/6)], where Ic values are NOE intensities calculated from the refined structure, and Io values are experimental NOE intensities.

Figure 5A shows the average structure of the R,R-duplex in the region of the C5:G18, R6:T17 and A7:T16 base pairs. The view is from the major groove. The R,R-DHB-dA nucleoside maintained the anti conformation about the glycosyl bond, and the adduct was located in the major groove. The DHB moiety rotated around the C6-N6 bond such that the diastereotopic R6 Hα,β hydrogens were oriented toward the 5′ direction and were in proximity to the major groove edge of the C5 base. This placed the diastereotopic R6 Hα′,β′ hydrogens in the 3′ direction, facing toward base pair A7:T16. The stereoisomeric hydroxyls of the R,R-DHB moiety both faced into the major groove. They did not participate in hydrogen bonding interactions with the DNA. With respect to base pairing, the R6 base was positioned such that it could maintain only one Watson–Crick hydrogen bond with the complementary thymine T17, between the N1 imine nitrogen and the T17 N3H imino proton (Figure S6 in the Supporting Information). The distance between R6 N1 and T17 N3H was 2.1 Å, and the distance between A7 N1 and T16 N3H was 2.0 Å, which was slightly longer than that for an unmodified A:T base pair, which is typically 1.9 Å. This was confirmed by helicoidal analysis,[78] where the opening parameter for the R6:T17 base pair was 19° (Figure S7 in the Supporting Information) but was 6° for the unmodified duplex.[55] The modified R6 base was tilted out of plane, as was also discerned by helicoidal analysis (Figure S8 in the Supporting Information). The exocyclic DHB ring of the R6 base was rotated out of plane and the N1-C6-N6-Cδ dihedral angle was −118°. Watson–Crick base pairing geometry at the neighboring C5:G18 and A7:T16 base pairs was maintained. However, the A7:T16 base pair exhibited lower stability as could be observed from NMR data collected as a function of temperature (Figure S3 of the Supporting Information). The R,R-DHB-dA adduct perturbed stacking interactions at base pairs C5:G18, R6:T17, and A7:T16 (Figure 6A,B) Thus, R6 stacked with its 5′ neighbor C5 but not with its 3′ neighbor A7. The complementary thymine T17 stacked well with its 5′ neighbor T16 but not with 3′ neighbor G18.

Figure 5

Average structures of the R,R- and S,S-DHB-dA modified duplexes. (A) Stereo drawing for the R,R-DHB-dA duplex in the region of the C5:G18, R6:T17, and A7:T16 base pairs. (B) Stereo drawing for the S,S-DHB-dA duplex in the region of the C5:G18, S6:T17, and A7:T16 base pairs. The modified nucleotides R6 or S6 are shown in green and red, respectively.

Figure 6

Stacking interactions for the R,R- and S,S-DHB-dA duplexes. (A) The R,R-DHB-dA duplex. Stacking of the C5:G18 base pair (black) above R6 (green) and T17 (black). (B) The S,S-DHB-dA duplex. Stacking of the R6:T17 pair (in green and black, respectively) above the base pair A7:T16 (black). (C) Stacking of the C5:G18 base pair (black) above S6 (red) and T17 (black). (D) Stacking of the S6:T17 pair (in red and black, respectively) above the base pair A7:T16 (black).

NMR Spectroscopy of the S,S-DHB-dA Duplex

The region of the NOESY spectrum showing the resonances between aromatic protons of the bases to the deoxyribose H1′ protons[79,80] for the modified strand C1 → G11 and for the complementary strand C12 → G22 is shown in Figure 1C,D. For the modified strand, the sequential NOE cross-peaks C5 H6 → C5 H1′, C5 H1′ → S6 H8, S6 H8 → S6 H1′, and S6 H1′ → A7 H8 were of similar intensities compared to those of the other cross-peaks arising from nucleotides distal to the adduct. Likewise, for the complementary strand, the sequential NOE cross-peaks for the bases proximal to the adduct T16 H6 → T16 H1′, T16 H1′ → T17 H6, T17 H6 → T17 H1′, and T17 H1′ → G18 H8 were of similar intensities as compared to the remainder of the cross-peaks arising from bases distal to the adduct. The greatest chemical shift perturbations compared to the unmodified duplex (Figure S1 of the Supporting Information) were observed for C5 H5, which shifted upfield 0.1 ppm, C5 H1′, which shifted downfield 0.3 ppm, S6 H8, which shifted upfield 0.1 ppm, and S6 H2, which shifted downfield 0.1 ppm. Minimal chemical shift perturbations were observed for the complementary strand.

Imino Proton Assignments

Figure 2 D–F shows an expansion of the NOESY spectrum, in the far downfield region in which the Watson–Crick base paired imino proton resonances are visible. The sequential NOE cross-peaks between imino protons[81] were observed for base pairs G2:C21 → G3:C20 → A4:T19 → C5:G18. The sequential NOE cross-peaks between imino protons for base pairs C5:G18 → S6:T17 and S6:T17 → A7:T16 were not observed. The sequential pattern of NOE cross-peaks could be continued from G8:C15 → A9:T14 → A10:T13 (Figure 2, panel F). The region of the spectrum showing NOEs between the base imino and the amino and adenine H2 protons exhibited the anticipated NOE cross-peaks for all base pairs, except for the terminal base G22 (Figure 2, panel E). The S6 H2 → T17 N3H cross-peak (g, Figure 2, panel E) was visible when the contour level of the spectrum was lowered and overlapped with the G11 N1H → C12N H1 cross-peak; its intensity was significantly lower than the remainder of the A:T cross-peaks. The A7 H2 → T16 N3H cross-peak was of lower intensity as compared to the other cross-peaks. The greatest chemical shift perturbation was observed for the T17 N3H resonance, which moved 0.35 ppm upfield as compared to the unmodified duplex (Figure S2 of the Supporting Information). The intensities of the T16 N3H and T17 N3H resonances were significantly lower (Figure S9 in the Supporting Information). These resonances were broader compared to their counterparts in the unmodified duplex (Figure S4 of the Supporting Information). Both resonances disappeared from the spectrum when the experiment was conducted at 20 °C.

S,S-DHB-dA Proton Assignments

Six S,S-DHB-dA adduct resonances were observed between 3.62 and 4.36 ppm (Figure 3B). This indicated that on the NMR time scale the DHB moiety did not rapidly rotate about the C6-N6 bond. Two sets of diastereotopic methylene proton resonances, i.e., Hα,β and Hα′,β′ (Chart 1), were identified, one located at 3.62 and 3.70 ppm and the other at 3.66 and 4.14 ppm. It was not possible to establish unequivocally which resonances were arising from the pro-R and from the pro-S hydrogen. It was possible to assign the two sets of methylene resonances, based upon NOEs to DNA protons. The set of methylene resonances at 3.66 and 4.14 ppm, assigned as the Hα,β hydrogens, exhibited intense NOEs to C5 H5 of the 5′-neighbor C5:G18 base pair at 5.14 ppm. They exhibited weaker NOEs to C5 H6. The set of methylene resonances at 3.62 and 3.70 ppm, assigned as the Hα′,β′ protons, exhibited NOEs to S6 H2 and a weak NOE to A7 H2. They also exhibited NOEs to the T16 CH3 protons at 1.69 ppm. The diastereomeric protons Hγ and Hγ′ were identified at 4.36 ppm and 4.25 ppm. The resonance at 4.36 ppm showed stronger NOEs to the Hα,β protons, whereas the resonance at 4.25 ppm exhibited NOEs to the Hα′,β′ protons. The Hα′,β′ protons gave more intense cross-peaks to the S6 H2 compared to that of other adduct protons. Additional weak interstrand cross-peaks were observed between the adduct protons S6 Hα′,β′ and Hα,β and the G18 N1H imino proton (cross-peaks a and b, Figure 2, panel D).

Structural Refinement of the S,S-DHB-dA Duplex

The structure of the S,S-duplex was determined using a simulated annealing rMD protocol, using distance restraints determined from the NOE data. A total of 199 NOE-based distance restraints were employed, which included 103 internucleotide restraints and 96 internucleotide restraints (Table 1). The diastereotopic DHB Hα,β and Hα′,β′ proton resonances could not be unequivocally assigned. As for the R,R-DHB-dA adduct, in some instances these protons exhibited similar NOE intensities to specific DNA protons, and in such cases, it was possible to include these distances in the rMD calculations. Also, the weak interstrand cross-peaks between DHB Hα,β and Hα′,β′ protons and the G18 N1H imino proton were not included in the structure calculations. The experimentally determined restraints were supplemented with 90 empirical phosphodiester backbone restraints, 42 Watson–Crick hydrogen bonding restraints, and 17 deoxyribose pseudorotation restraints. The inclusion of these empirical restraints was predicated upon the observation that the NMR data suggested that apart from the modified base pair S6:T17 and its 5′- and 3′-neighboring base pairs C5:G18 and A7:T16, the duplex maintained a right handed B-DNA-like conformation in solution. Table 1 shows the restraints used for rMD calculations. Eight structures that emerged from the rMD calculations were subjected to potential energy minimization. These are shown superimposed in Figure S5, panel B of the Supporting Information. These structures converged to a maximum pairwise rmsd value of 0.66 Å and were used to calculate an average structure, which was subjected to complete relaxation matrix analysis.[67] The overall sixth root residual (R1x) was 7.8% (Table 2). An evaluation of the R1x residuals for individual nucleotides (Figure 4B) indicated that these were consistently within the 10% range. Thus, the average structure was in good agreement with the observed NOE intensities. Table 1 shows the structural statistics.

Figure 5B shows the S,S-duplex in the region of the C5:G18, S6:T17, and A7:T16 base pairs, viewed from the major groove. The S,S-DHB-dA modified base adopted the anti conformation about the glycosyl bond. The DHB moiety was located in the major groove. The adduct rotated around the C6-N6 bond and was positioned such that the S6 Hα,β hydrogens oriented toward the 5′ direction and were proximate to the C5 base. This placed the S6 Hα′,β′ hydrogens toward the 3′-neighbor A7:T16 base pair. The stereoisomeric hydroxyl groups of the DHB moiety both faced into the major groove, and they did not form hydrogen bonds with the DNA. The Hγ proton faced toward the modified strand, whereas the Hγ′ proton faced toward the complementary strand. The S6 base formed one Watson–Crick hydrogen bond between the S6 N1 imine nitrogen and the T17 imino proton (Figure S10 in the Supporting Information). The distance between S6 N1 and N3H was 2.0 Å, which is slightly longer than that for an unmodified A:T base pair, which is typically 1.9 Å. It is also confirmed by helicoidal analysis, where the opening for the S6:T17 base pair was 12° higher than that for other base pairs (Figure S7 of the Supporting Information). The exocyclic DHB ring of the S6 base was rotated out of plane, and the N1-C6-N6-Cδ dihedral angle was −123°. Watson–Crick hydrogen bonding at the neighbor base pairs was maintained. However, the A7:T16 base pair exhibited lower stability as observed in the NMR spectra obtained as a function of temperature (Figure S9 of the Supporting Information). The S,S-DHB-dA base interfered with base stacking interactions. It did not stack well with its 5′ neighbor C5 but stacked well with its 3′ neighbor A7 (Figure 6C,D. The complementary thymine T17 did stack well with its 5′ neighbor T16 but not with 3′ neighbor G18.

Discussion

The potential for bis-alkylation of DNA appears to be critical for the mutagenicity and cytotoxicity of DEB,[83] as butadiene mono-oxidation products, e.g., EB and EBD, are much less genotoxic.[44] Adenine adducts are likely to play a major role in the genotoxicity of DEB since the number of A:T base pair substitutions equals or exceeds the number of mutations at the G:C base pairs,[84] despite the fact that DEB preferentially alkylates guanines in DNA.[83] Thus, one or more DEB-specific dA lesions must contribute to DEB-induced genotoxicity.[53,84,85] However, the specific DEB-dA adducts capable of inducing these A to T transversions remain obscure.[53,86,87] The double alkylation of the R,R- and S,S-DHB-dA adducts removes the N6-dA amino groups as hydrogen bond donors in base pairing with dT and raises the question as to how the bulky 3,4-dihydroxypyrrolidine rings are accommodated in the DNA duplex. UV melting studies have revealed that the R,R- and S,S-DHB-dA adducts significantly destabilize duplex DNA, evidenced by a 16–17 °C decrease in Tm as compared to that of unmodified DNA.[56] However, CD studies have suggested that R,R- and S,S-DHB-dA-modfied DNA duplexes maintain a B-type DNA conformation.[56] Consequently, it was of interest to determine the solution structures of the R,R- and S,S-DHB-dA adducts.

Structures of the R,R- and S,S-DHB-dA Adducts

The structures of the R,R- and S,S-DHB-dA adducts in DNA are similar, and both adducts are positioned in the major groove of DNA. In both instances, the bulky 3,4-dihydroxypyrrolidine rings are accommodated by an out-of-plane rotation about the C6-N6 bond of the bis-alkylated adenine such that the Hα,β protons orient in the 5′-direction, whereas the Hα′,β′ protons orient in the 3′-direction. In both instances, this directionality of the 3,4-dihydroxypyrrolidine ring is evidenced by the patterns of NOEs between 3,4-dihydroxypyrrolidine protons and DNA. For the R,R-DHB-dA adduct, NOEs between the Hα,β protons and C5 H5 (Figure 3A) confirm that the Hα,β protons orient in the 5′ direction and are proximate to the C5 base. The largest downfield chemical shift of 0.3 ppm compared to that of the unmodified duplex is observed for C5 H1′, which is attributed to a less planar stacking of the C5 and Y6 bases. Likewise, for the S,S-DHB-dA adduct, similar NOEs are observed between Hα,β protons and C5 H5 (Figure 3B). The conclusion that both the R,R- and S,S-DHB-dA modified bases maintain the anti conformation about the glycosyl bond is supported by the intensities of the R6 H8 → R6 H1′ and S6 H8 → S6 H1′ NOEs, which are similar to those of the other nucleobase aromatic proton → deoxyribose H1′ NOEs (Figure 1). In both instances, maintaining the anti conformation of the glycosyl bond, combined with the out-of-plane rotation about the C6-N6 bond, allows the complementary thymine, T17, to remain stacked within the duplex and form one hydrogen bond with the modified base pair, between the imine nitrogen of the modified base and the T17 N3H imino proton. For the R,R-DHB-dA adduct, the presence of this hydrogen bond is revealed by the weak NOE between R6 H2 and T17 N3H (cross-peak k, Figure 2, panel B). Likewise, the presence of the corresponding hydrogen bond at the modified S6:T17 base pair is indicated by the weak NOE between S6 H2 and T17 N3H (cross-peak g, Figure 2, panel E). The loss of the second Watson–Crick hydrogen bonding interaction at the lesion sites correlates with the significantly lower thermal stabilities of the R,R- and S,S-DHB-dA duplexes, as compared to that of the unmodified duplex. The reduced base stacking at the adduct sites may also contribute to the thermal destabilization. In particular, both the R6 and S6 bases stack with the 5′ neighbor C5 but not with the 3′ neighbor A7 (Figure 6). Likewise, in both instances, the complementary thymine T17 stacks well with the 5′ neighbor T16 but not with the 3′ neighbor G18.

In summary, these data suggest that out-of-plane rotations of the 3,4-dihydroxypyrrolidine rings allow the complementary thymine to be accommodated in the DNA duplex. Density functional calculations at the nucleoside level have revealed that the pyrrolidine nitrogen lone electron pair is shared with the π-electron system of the purine ring, resulting in partial double bond character and a barrier to rotation about the C6-N6 bond.[56] At the nucleoside level, the two sets of methylene protons for the pyrrolidine ring experience different chemical shift environments. The same is observed in duplex DNA (vide infra). Hence, the rMD calculations employed herein utilized an sp2 hybridized pyrrolidine nitrogen. These calculations suggest that the adenine C6-N6 bond possesses sufficient single bond character to facilitate its out-of-plane orientation with respect to the adenine nucleobase. Accordingly, a second set of rMD calculations was performed, which utilized an sp3 hybridized pyrrolidine nitrogen. The resulting structures were similar, with the pyrrolidine rings of both DHB-dA stereoisomers similarly orienting in the major groove. Complete relaxation matrix calculations performed on the refined structures emergent from the calculations involving the sp3 hybridized pyrrolidine nitrogen showed comparable agreement with the NOESY data at the modified base pairs, as was obtained from the refined structures involving the sp2 hybridized pyrrolidine nitrogen. Overall, this is consistent with the conclusion that the adenine C6-N6 bond possesses some degree of partial double bond character but retains sufficient single bond character to allow the 3,4-dihydroxypyrrolidine ring to rotate out of the plane of the adenine nucleobase.

While both the R,R- and S,S-DHB-dA adducts are oriented in the major groove, they differ as to the configurations of the hydroxyl groups at the stereoisomeric carbon atoms of the 3,4-dihydroxypyrrolidine ring. It was thus of interest to examine the possibility of differential hydrogen bonding patterns between the stereoisomeric hydroxyl groups and the DNA. However, the structural data does not support the notion that the R,R- and S,S-DHB-dA adducts exhibit differential hydrogen bonding patterns between the stereoisomeric hydroxyls and the DNA. Also, the thermodynamic data do not suggest the formation of interstrand hydrogen bonding interactions by these stereoisomeric lesions. In both instances, the DNA is destabilized, and for R,R-duplex, Tm is 28°, while for the S,S-duplex, Tm is 27°. This suggests that neither modified duplex is stabilized by hydrogen bonding interactions involving the DHB hydroxyl groups. However, differential orientations of the DHB hydroxyl groups might provide a mechanism to differentially interact with DNA polymerases during translesion bypass, thus modulating mutagenic outcomes. Consequently, it will be of interest to examine the structures of the R,R- and S,S-DHB-dA adducts in the context of translesion replication complexes.

Structure–Activity Relationships

The orientations of the R,R- and S,S-DHB-dA adducts in the major groove of DNA with respect to the bis-alkylated adenine nucleobase allows for thymine in the complementary strand to be accommodated in the DNA duplex with relatively little distortion. However, because the pyrrolidine nitrogen cannot participate in hydrogen bonding with thymine in the complementary strand, at the DNA damage site, the damaged DNA base pair is stabilized by a single hydrogen bond. Thus, the loss of the T O4 → A N6H Watson–Crick hydrogen bond, coupled with reduced base stacking interactions, probably accounts for the significant decrease in Tm observed for both the R,R- and S,S-DHB-dA adducts. Structural studies of a series of other monodentate N6-dA adducts arising from styrene oxide and butadiene epoxides suggest that they also are contained within the major groove but induce less structural perturbation of Watson−Crick base pairing in DNA.[88−93] Repair mechanisms responsible for the removal of the R,R- and S,S-DHB-dA lesions in cells are currently under investigation. Our initial results using extracts from human fibrosarcoma (HT1080) cells suggest that DHB-dA is rapidly repaired.[94] Site-specific mutagenesis studies conducted in E. coli have revealed that the R,R- and S,S-DHB-dA adducts are weakly mutagenic,[95] which could potentially be due to the efficient repair of these adducts in vivo. However, our replication studies in vitro have revealed that DHB-dA is bypassed in an error-prone manner by human translesion synthesis (TLS) polymerases κ and η. These studies suggest that DHB-dA adducts can induce both base substitution and deletion mutations (Kotapati and Tretyakova, unpublished results).[96]

Summary

Both R,R- and S,S- stereoisomers of DEB-induced N6,N6-DHB-dA adducts are oriented in the major groove of DNA. For both adducts, this allows thymine in the complementary strand of the DNA to be inserted into the duplex. However, the complementary thymine forms only a single hydrogen bond with the damaged adenine due to the fact that the N6-protons of the modified bases are replaced by the pyrrolidine ring, which cannot participate in hydrogen bonding with the complementary thymine. Moreover, the complementary thymine is displaced toward the DNA minor groove, weakening its stacking interactions with the neighboring base pairs. This may explain the thermal destabilization induced into DNA by the R,R- and S,S-DHB-dA adducts and predicts that they might be good candidates for base excision repair. The structural data do not support the notion that the R,R- and S,S-DHB-dA adducts show differential hydrogen bonding patterns with the complementary strand involving the stereoisomeric pyrrolidine ring hydroxyl groups.