JUNE 2004 VOLUME 17, NUMBER 6 © Copyright 2004 by the American Chemical Society
Articles Structure of a Site Specific Major Groove (2S,3S)-N6-(2,3,4-Trihydroxybutyl)-2′-deoxyadenosyl DNA Adduct of Butadiene Diol Epoxide Tandace A. Scholdberg, Lubomir V. Nechev,† W. Keither Merritt, Thomas M. Harris, Constance M. Harris, R. Stephen Lloyd,‡ and Michael P. Stone* Department of Chemistry, Center in Molecular Toxicology, Vanderbilt-Ingram Cancer Center, Vanderbilt University, Nashville, Tennessee 37235 Received December 30, 2003
The solution structure of the (2S,3S)-N6-(2,3,4-trihydroxybutyl)-2′-deoxyadenosyl adduct arising from the alkylation of adenine N6 at position X6 in d(CGGACXAGAAG)‚d(CTTCTTGTCCG), by butadiene diol epoxide, was determined. This oligodeoxynucleotide contains codon 61 (underlined) of the human N-ras protooncogene. This oligodeoxynucleotide, containing the adenine N6 adduct butadiene triol (BDT) adduct at the second position of codon 61, was named the ras61 S,S-BDT(61,2) adduct. NMR spectroscopy revealed modest structural perturbations localized to the site of adduction at X6‚T17, and its nearest-neighbor base pairs C5‚G18 and A7‚T16. All sequential NOE connectivities arising from DNA protons were observed. Torsion angle analysis from COSY data suggested that the deoxyribose sugar at X6 remained in the C2′-endo conformation. Molecular dynamics calculations using a simulated annealing protocol restrained by a total of 442 NOEderived distances and J coupling-derived torsion angles refined structures in which the BDT moiety oriented in the major groove. Relaxation matrix analysis suggested hydrogen bonding between the hydroxyl group located at the β-carbon of the BDT moiety and the T17 O4 of the modified base pair X6‚T17. The minimal perturbation of DNA induced by this major groove adduct correlated with its facile bypass by three Escherichia coli DNA polymerases in vitro and its weak mutagenicity [Carmical, J. R., Nechev, L. V., Harris, C. M., Harris, T. M., and Lloyd, R. S. (2000) Environ. Mol. Mutagen. 35, 48-56]. Overall, the structure of this adduct is consistent with an emerging pattern in which major groove adenine N6 alkylation products of styrene and butadiene oxides that do not strongly perturb DNA structure are not strongly mutagenic.
Introduction 1,3-BD1
(CAS RN 106-99-0) is used in the manufacture of SBR (1, 2); several billion lbs/year are produced in the * To whom correspondence should be addressed. Tel: 615-322-2589. Fax: 615-322-7591. E-mail:
[email protected]. † Current address: Alnylam Pharmaceuticals, Inc., 790 Memorial Drive Suite 202, Cambridge, MA 02139. ‡ Center for Research on Occupational and Environmental Toxicology, Oregon Health and Science University, Portland, OR 97239.
United States. It is a combustion product from automobile emissions (3) and cigarette smoke (4). BD is genotoxic and is a carcinogen in rodents, particularly in mice (57), and also in rats (8). Recently, BD was classified by the United States Environmental Protection Agency as “carcinogenic to humans by inhalation” (9). The International Agency for Cancer Research (IARC) lists BD as a “probable human carcinogen” (group 2A) (10). Chronic human exposure in the SBR industry may induce geno-
10.1021/tx034271+ CCC: $27.50 © 2004 American Chemical Society Published on Web 05/08/2004
718
Chem. Res. Toxicol., Vol. 17, No. 6, 2004 Scheme 1. Reactive Metabolites of BD
toxic effects (11-13) and is correlated with an increased risk for leukemia (1, 14-22). BD is epoxidized primarily by cytochrome P450 2E1, but also by cytochrome P450 2A6, to form BDOs (Scheme 1) (23, 24). These may be further oxidized by cytochrome P450 2E1 or 3A4 to form BDO2s (23, 25-29). Hydrolysis of BDO mediated by epoxide hydrolase forms 1,2-dihydroxy-3-butenes (27, 30, 31), which are metabolized by cytochrome P450 to hydroxymethylvinyl ketone (HMVK) (32). Either BDO2 or the 1,2-dihydroxy-3-butenes undergo cytochrome P450-mediated oxidation to form 1,2-dihydroxy-3,4-epoxybutanes (BDE) (27, 30, 33). Thus, proximate electrophiles arising from BD metabolism include BDO, BDO2, BDE, and potentially, HMVK (34). Of these species, BDO2 is highly genotoxic (2, 10, 35), probably due to its potential to form DNA-DNA and DNA-protein cross-links, which were observed in mice (36, 37). The greater sensitivity of mice than rats upon exposure to BD is attributed to their efficient oxidation of BD to BDO2 (38, 39). BDO and BDE are less genotoxic (35, 40, 41); however, BDE is of particular concern because it may represent the most abundant metabolite of BD produced in humans (42). The reaction of DNA with BDE (42, 43) leads to the formation of trihydroxy (BDT) adducts. These represent the most prevalent adducts isolated from humans (44, 45) and also from rodents exposed to BD (42, 46, 47). While BDT adducts at N7-dG predominate (42, 46), there is considerable interest in BDT adducts at other sites in DNA, particularly dA (48-50). The BDT dA-N1 adduct was identified in humans exposed occupationally to BD (51). The dA-N6 BDT adduct was identified in Chinese hamster ovary cells (52). The latter likely results from Dimroth rearrangement of the corresponding dA-N1 adduct (53, 54), as opposed to direct alkylation at dAN6. Analogous chemistry was described for the reactions of styrene oxide with adenine N1 (55, 56). In the present work, the (2S,3S)-N6-(2,3,4-trihydroxybutyl)-2′-deoxyadenosyl adenine N6 BDT adduct was site specifically incorporated into the ras61 oligodeoxynucleotide (57). The resulting modified oligodeoxynucleotide, 5′-d(CGGACXAGAAG)-3′‚5′-d(CTTCTTGTCCG)-3′, contained the N6-dA BDT adduct at the second position of codon 61 (Scheme 2). This was named the dA N6 S,SBDT-(61,2) lesion. Structural refinement from NMR data demonstrates that the S,S-BDT-(61,2) moiety is located 1 Abbreviations: BD, butadiene; BDE, 3,4-epoxy-1,2-butanediol; BDO, butadiene monoepoxide (1,2-epoxy-3-butene); BDO2, butadiene diepoxide (1,2:3,4-diepoxybutane); CPK, Corey-Pauling-Koltun spacefilling models; DQF-COSY, double quantum filtered correlation spectroscopy; HMBC, heteronuclear multiple bond correlation spectroscopy; R1x, sixth root residual; rMD, restrained molecular dynamics; rmsd, root mean square deviation; SBR, styrene-butadiene rubber; TOCSY, total correlation spectroscopy; TPPI, time proportional phase increment.
Scholdberg et al. Scheme 2. ras61 Oligodeoxynucleotide, the Chemical Structure of the (2S,3S)-N6-(2,3,4Trihydroxybutyl)-2′-deoxyadenosyl Adduct, and Nomenclature
in the major groove and causes minimal localized structural perturbation. A series of rMD calculations suggest that the S,S-BDT-(61,2) adduct is stabilized by the formation of a hydrogen bond between the β-hydoxyl group of the BDT moiety and the T17 O4. The results of bacterial site specific mutagenesis experiments using this S,S-BDT-(61,2) adduct revealed low levels of A f C transversions (58). Earlier studies showed that adenine N6 adducts of R- (59-61) and β-styrene oxide (62) in the ras61 oligodeoxynucleotide (63) also oriented in the major groove of the DNA with minor structural perturbation and were weakly mutagenic. The combined results suggest that the major groove monodentate adenine N6 adducts of styrene and BD are weakly mutagenic.
Materials and Methods Oligodeoxynucleotide Synthesis. The unmodified oligodeoxynucleotides were synthesized by the Midland Certified Reagent Co. (Midland, TX) and purified by anion exchange chromatography. The concentration of the single-stranded unmodified oligonucleotide (57) was determined from the extinction coefficient of 9.08 × 104 cm-1 at 260 nm (64). An excess of unmodified strand was combined with the S,S-BDT-(61,2) modified strand, in a buffer consisting of 10 mM NaH2PO4, 0.1 M NaCl, and 50 µM Na2EDTA at pH 7.0. The solution was heated at 95 °C for 5 min and then allowed to cool slowly to room temperature. The resulting mixture of single-stranded and duplex DNA was equilibrated in 10 mM NaH2PO4, 0.1 M NaCl, and 50 µM Na2EDTA at pH 7.0 on a column containing DNA Grade Biogel hydroxylapatite (Bio-Rad Laboratories, Richmond, CA). The DNA was eluted off of the hydroxylapatite with a gradient from 10 to 200 mM NaH2PO4, pH 7.0, to separate excess single-stranded from duplex DNA. The duplex was lyophilized, resuspended in 1 mL of H2O, and desalted on a Sephadex G-25 column. The sample was lyophilized. NMR Spectroscopy. The S,S-BDT-(61,2) modified duplex was prepared at a concentration of 2 mM. For observation of nonexchangeable protons, the sample was dissolved in 0.5 mL of 10 mM NaH2PO4, 0.1 M NaCl, and 50 µM Na2EDTA at pH 7.0. The sample was exchanged three times with 99.9% D2O and dissolved in 99.96% D2O. For observation of exchangeable protons, the sample was dissolved in 0.5 mL of 10 mM NaH2PO4, 0.1 M NaCl, and 50 µM Na2EDTA at pH 7.0. The sample
Adducts of Butadiene Diol Epoxide
Chem. Res. Toxicol., Vol. 17, No. 6, 2004 719
Figure 1. Expanded plot showing sequential 1H NOE connectivities for the ras61 S,S-BDT-(61,2) oligodeoxynucleotide. (A) The modified strand. (B) The complementary strand. was lyophilized and suspended in 0.5 mL of 9:1 H2O/D2O. 1H NMR spectra were recorded at 800.23 MHz. The temperature was controlled at 20 ( 0.5 °C for the observation of exchangeable and nonexchangeable protons. Chemical shifts were referenced to the water resonance at 4.71 ppm at 20 °C. NOESY and DQF-COSY experiments were performed at a frequency of 800.23 MHz. Phase sensitive NOESY spectra used for resonance assignment were recorded using TPPI phase cycling with mixing times of 150, 200, and 250 ms. For examination of exchangeable protons, phase sensitive NOESY were carried out using a field gradient Watergate pulse sequence for water suppression (65). The spectra were recorded at 20 °C and 150 ms. These experiments were recorded with 1024 real data points in the d1 dimension and 2048 real data points in the d2 dimension. A relaxation delay of 2.0 s was used. To derive distance restraints, NOESY spectra were recorded at mixing times of 150, 200, and 250 ms. A TOCSY experiment was performed at a frequency of 800.23 MHz with mixing times of 90 and 150 ms, utilizing the homonuclear Hartman-Hahn transfer with MLEV17 sequence for mixing. Data from the TOCSY spectrum allowed for the assignment of the BD protons via bond magnetization transfer and allowed for the observance of all BD protons in the aliphatic chain. The NMR data were processed using FELIX (Accelrys, Inc., San Diego, CA) on an Octane workstation (Silicon Graphics, Inc., Mountain View, CA). The data in the d1 dimension were zero-filled to give a matrix of 1024 × 2048 real points. A skewed sine-bell square apodization function with a 90° phase shift and a skew factor of 1.0 was used in both dimensions. rMD Calculations. Calculations were performed using XPLOR (66). The force field was derived from CHARMM (67) and adapted for rMD calculations for nucleic acids. Classical A-DNA and B-DNA were used as the reference structures to create starting structures for the refinement (68). The BD adduct was constructed at the sixth position using the BUILDER module of INSIGHT II (Accelrys, Inc.). The A and B structures were energy-minimized by the conjugate gradients method for 200 iterations without experimental restraints to give starting IniA and IniB used for the subsequent relaxation matrix analysis and molecular dynamics calculations. Footprints were drawn around the NOE cross-peak for the NOESY spectrum measured with a mixing time of 250 ms to define the size and the shape of the individual cross-peak using FELIX. The same set of footprints was applied to spectra measured with the two other mixing times. Cross-peak intensities were determined by integrating the volume of the areas under the footprints. The intensities were combined as necessary with intensities generated from complete relaxation matrix analysis of a starting DNA structure to generate two hybrid intensity matrixes using
Figure 2. Expanded plot showing the sequential 1H NOE connectivities for the Watson-Crick base-paired imino protons of the ras61 S,S-BDT-(61,2) oligodeoxynucleotide. Table 1. rmsd, Excluding the End Base Pairs, between Various Initial Structures and Intermediate Structures of the S,S-BDT-(61,2) Oligodeoxynucleotide atomic rmsd (Å) NMR restraints total no. of distance restraints interresidue distance restraints intraresidue distance restraints DNA-BDT distance restraints deoxyribose pseudorotation restraints backbone torsion angle restraints hydrogen-bonding restraints initial structures IniA vs IniB rms shifts IniA vs 〈rMDA〉a IniB vs 〈rMDB〉b 〈rMDB〉 vs 〈rMDB〉 〈rMDA〉 vs 〈rMDA〉 rms pairwise difference (rMDB) standard deviation (rMDB) rms difference from mean structure (rMDB) standard deviation (rMDB)
442 91 351 18 90 62
5.92 5.70 ( 0.15 1.25 ( 0.10 0.74 ( 0.12 0.87 ( 0.10 1.03 0.18 0.66 0.14
a 〈rMDA〉 represents the set of six structures that emerged from rMD calculations starting with IniA. b 〈rMDB〉 represents the set of six structures that emerged from rMD calculations starting from IniB.
MARDIGRAS (69, 70) for each of the three mixing times. Calculations using DNA starting models generated by INSIGHT II, NOE experiments with three mixing times, and isotropic
720
Chem. Res. Toxicol., Vol. 17, No. 6, 2004
Scholdberg et al.
Figure 3. Expanded plots showing the assignments of the BD resonances of the ras61 S,S-BDT-(61,2) oligodeoxynucleotide. (A) A NOESY spectrum. (B) A DQF-COSY spectrum. correlation times of 2, 3, and 4 ns yielded 18 sets of distances. These data were pooled; the average values of all minimum and maximum distances were used in setting error bounds to give the experimental NOE restraints used in subsequent molecular dynamics calculations (71). These distance restraints were divided into classes on the basis of the confidence factor obtained from them. The program X-PLOR was used for energy minimization and rMD calculations (66). The calculations were based on an energy function approach where the total energy was the sum of the empirical energy of the molecule and the effective energy, comprised of the restraint energy terms. The CHARMM force field was utilized, which provided the terms for the bonds, bond angles, torsion angles, tetrahedral and planar geometry, hydrogen bonding, and nonbonding interactions, including van der Waals and electrostatic forces. The electrostatic term used the Coulombic function based on a reduced charge set of partial charges and a distance-dependent dielectric constant of 4.0, to mimic solvent screening of charge. The van der Waals term was approximated using the Lennard-Jones potential energy function. The effective energy function was comprised of two terms describing distances and dihedral restraints, both of which were in the form of a standard square-well potential (72). Bond lengths involving hydrogens were fixed with the SHAKE algorithm (73) during molecular dynamics calculations. These calculations were all performed in vacuo. The simulated annealing procedure consisted of a total of 50 ps of rMD. The protocol consisted of 30 ps of high-temperature rMD, 5 ps of cooling, and 15 ps of low-temperature rMD. The high temperature was 2500 K, and the low temperature was 300 K. The temperature was controlled by coupling the molecules to a temperature bath with a coupling constant of 0.05 ps (74). Ten structures were calculated from each starting structure, each being assigned a different random seed for the generation of the initial atomic velocity vectors. The rMD calculations were initialized by assigning a random set of velocities to all of the atoms that fit a Maxwell-Boltzmann distribution at 2500 K. The rMD calculations were carried out for 3000 steps at 2500 K, then cooled to 300 K over 5000 steps, and continued at 300 K for an additional 15 000 steps. An initial force constant of 50.0 kcal mol-1 was used for class 1 distance restraints. Throughout the rMD calculation, the force constants for classes 2, 3, 4, and 5 were set to 90, 80, 70, and 60%, respectively, of the value for class 1. The initial value of the force constant for the base-pairing restraints was set at 50.0 kcal mol-1. The force constants were maintained at the initial value for the first 10 000 steps of the rMD calculations; class 1 force constants were increased to 200 kcal mol-1, and basepairing force constants were increased to 150 kcal mol-1 over the next 10 000 steps. The force constants were maintained for 17 000 steps, scaled down to 70 and 50 kcal mol-1 for class 1 and base-pairing restraints, respectively, over 3000 steps, and
Figure 4. Expanded tile plots showing 1H NOEs between the DNA and the BD protons. (A) NOEs between the T16 and the T17 CH3 protons and HR,R′ and Hδ,δ′ of the BD moiety. (B) NOEs between C5 H5 and BD. (C) NOEs between C5 H6 and BD. (D) NOEs between T16 and T17 N3H and BD. remained at these values for the final 10 000 steps of the rMD calculations. Structure coordinates were archived every 0.1 ps over the final 10 ps of the rMD simulation. Structure coordinates extracted from the final 4 ps of each rMD calculation were averaged and energy-minimized for 200 iterations using a conjugate gradients algorithm. The final refined structure was obtained by minimizing the averaged structure generated from 10 structures from two rMD calculations, five from an A form starting structure, iniA, and five from a B form starting structure, iniB. Back calculation of NMR data was performed using CORMA (75) on the final refined structure resulting in an R1x of 0.12.
Results Assignments of Nonexchangeable DNA Protons. The sequential assignment for the S,S-BDT-(61,2) duplex was accomplished using standard protocols (76, 77). The sequential NOEs were uninterrupted in both the modi-
Adducts of Butadiene Diol Epoxide
Chem. Res. Toxicol., Vol. 17, No. 6, 2004 721
Figure 5. Chemical shift differences of aromatic and anomeric protons of the S,S-BDT-(61,2) adduct relative to the unmodified ras61 oligodeoxynucleotide. (A) The modified strand of the S,S-BDT-(61,2) adduct. (B) The complementary strand of the S,S-BDT(61,2) adduct.
fied and the unmodified strands of the duplex. A contour plot of the NOESY spectrum collected at a mixing time of 250 ms is shown in Figure 1. The complete assignment of the deoxyribose H2′, H2′′, H3′, and H4′ protons was achieved. Because of overlap, partial assignments were made for the deoxyribose H5′ and H5′′ protons. The spectrum was similar to that of the unmodified ras61 oligodeoxynucleotide (63), suggesting that the presence of the S,S-BDT-(61,2) adduct had minimal effect on the ras61 oligodeoxynucleotide. Assignments of Exchangeable DNA Protons. The assignments for the base imino protons were made at 5 °C (78) (Figure 2). The thymine imino protons, T13 N3H
and T14 N3H, resonated at 14.3 and 13.78 ppm. T16 N3H and T17 N3H cross-peaks were observable at 13.53 and 13.54 ppm. T19 N3H resonated at 13.71 ppm. The imino resonances from G2, G3, G8, G11, and G18 were individually resolved. The nonhydrogen-bonded amino resonance of C5, the base 5′ to the adducted nucleotide, resonated at 6.73 ppm, while the hydrogen-bonded resonance was located at 8.21 ppm. G18 showed an NOE to T17 N3H, indicating that the modified base pair X6‚T17 was intact. The spectrum was also similar to that of the unmodified ras61 oligodeoxynucleotide (63), suggesting that the presence of the S,S-BDT-(61,2) adduct had minimal effect on the ras61 oligodeoxynucleotide.
722
Chem. Res. Toxicol., Vol. 17, No. 6, 2004
Scholdberg et al.
Figure 6. Stereoview of six superimposed structures emergent from the simulated annealing rMD protocol; the structures resulted from randomly seeded calculations.
BDT Protons. The BDT protons were observed between 3.4 and 4.0 ppm. The HR and HR′ protons were observed at δ 3.10 and δ 3.74 ppm, respectively. The difference in chemical shift between the two protons reflects the difference in chemical shift environments. The Hβ proton was observed at δ 3.95 ppm. The Hγ proton was observed at δ 3.93 ppm. The Hδ and Hδ′ protons were observed at δ 3.64 and δ 3.62 ppm. The stereotopic assignment2 of the HR,R′ and Hδ,δ′ proton resonances were elucidated by analysis of 3J coupling constants and NOE intensities, the latter in comparison with predicted intensities generated by complete relaxation matrix analysis (75). NOE Connectivities between BDT and DNA. Table 1 tabulates cross-peaks observed between BDT protons and DNA. A total of 18 NOEs were observed between BDT and X6, its adducted base, as well as its base-paired T17. NOEs between the T17 N3H and the BDT protons were observed in the 1H NOESY spectrum. The BDT protons HR and Hγ displayed a NOE to H6 of the adjacent 5′ neighboring cytosine. No NOE was observed between BDT HR′ and either C5 H5 or H6 of the 5′ neighboring cytosine, thereby distinguishing between HR and HR′. Similarly, a different set of NOEs was observed for Hδ and Hδ′, with the Hδ showing a strong cross-peak to T16 N3H and a weaker cross-peak to T17 N3H, whereas the 2 The definitions of the prochiral protons at C and C of the BDT R δ moiety are based upon the Cahn, Ingold, and Prelog nomenclature. The proton HR is defined as the pro-R proton at CR; HR′ is defined as the pro-S proton at CR. Likewise, Hδ is the pro-R proton at Cδ; Hδ′ is the pro-S proton at Cδ.
Hδ′ did not exhibit these NOEs. The NOEs between BDT and DNA protons established the orientation of the BDT moiety. Torsion Angle Restraints. A series of 1H DQF-COSY and E-COSY experiments showed that all of the deoxyribose sugars in the S,S-BDT-(61,2) oligodeoxynucleotide existed predominantly in the C2′-endo conformation, characteristic of a B type DNA helix (79). The 31P spectrum showed no unusual chemical shifts, suggesting that the phosphodiester backbone was not significantly perturbed by the presence of the S,S-BDT moiety. This observation was confirmed by a 1H-31P HMBC experiment using an IBURP-shaped pulse (80) in the 1H dimension. The 3J couplings between H3′ and 31P were consistent with backbone torsion angles in the normal range for B type duplexes (81). The glycosyl torsion angles of all nucleotides were examined using 1H NOESY experiments. These revealed weak NOEs between the purine H8 or pyrimidine H6 protons and the anomeric H1′ protons of the attached deoxyribose sugars, consistent with glycosyl torsion angles in the anti conformational range, also consistent with a B form helix (82). Chemical Shift Effects. Small chemical shift perturbations were observed when the ras61 S,S-BDT-(61,2) oligodeoxynucleotide was compared to the unmodified ras61 oligodeoxynucleotide (63). Figure 5 shows the chemical shift perturbations for the nucleobase aromatic protons. The observed chemical shifts were clustered at the modified base pair X6‚T17 and its nearest neighbor base pairs in either direction, C5‚G18 and A7‚T16. The largest of these was on the order of 0.1 ppm, suggesting
Adducts of Butadiene Diol Epoxide
Chem. Res. Toxicol., Vol. 17, No. 6, 2004 723
Figure 7. Complete relaxation matrix calculations (75) on the average structure emergent from the simulated annealing rMD protocol showing R1x for each nucleotide. (A) The adducted strand. (B) The complementary strand.
that the S,S-BDT adduct introduced only minimal perturbation to the magnetic field environment at and adjacent to the lesion site. The remainder of the base pairs in the S,S-BDT-(61,2) oligodeoxynucleotide showed essentially no chemical shift perturbations as compared to the unmodified ras61 oligodeoxynucleotide. rMD Calculations. The NOE and torsion angle data were used to restrain a series of molecular dynamics calculations. These rMD calculations employed a simulated annealing protocol. They incorporated 442 NOEbased distance restraints, including 18 between BDT protons and DNA protons. An additional 90 pseudorotation and 62 phosphodiester backbone angle measurements were also included. The orientations of the hydroxyl groups at the β- and γ-carbons were of particular interest. Initial rMD calculations predicted that the β-hydroxyl would be approximately 2 Å from T17 O4, suggesting the potential for formation of a hydrogen
bond. Consequently, it was decided to initiate a second series of rMD calculations in which an additional hydrogen-bonding restraint was incorporated between the β-hydroxyl proton and the T17 O4. Figure 6 shows a stereoview of an ensemble of six structures that emerged from six randomly seeded rMD calculations including this hydrogen-bonding restraint. Table 1 details the structural refinement parameters. These calculations were initiated using both B form DNA and A form DNA starting structures. The structures that emerged from the simulated annealing protocol converged with a rmsd of
