Structural studies of an oligodeoxynucleotide containing a trimethylene interstrand cross-link in a 5'-(CpG) motif: model of a malondialdehyde cross-link.

Malondialdehyde (MDA), a known mutagen and suspected carcinogen, is a product of lipid peroxidation and byproduct of eicosanoid biosynthesis. MDA can react with DNA to generate potentially mutagenic adducts on adenine, cytosine, and particularly guanine. In addition, repair-dependent frame shift mutations in a GCGCGC region of Salmonella typhimurium hisD3052 have been attributed to formation of interstrand cross-links (Mukai, F. H. and Goldstein, B. D. Science 1976, 191, 868--869). The cross-linked species is unstable and has never been characterized but has been postulated to be a bis-imino linkage between N(2) positions of guanines. An analogous linkage has now been investigated as a stable surrogate using the self-complementary oligodeoxynucleotide sequence 5'-d(AGGCG*CCT)(2,) in which G* represents guanines linked via a trimethylene chain between N(2) positions. The solution structure, obtained by NMR spectroscopy and molecular dynamics using a simulated annealing protocol, revealed the cross-link only minimally distorts duplex structure in the region of the cross-link. The tether is accommodated by partially unwinding the duplex at the lesion site to produce a bulge and tipping the guanine residues; the two guanines and the tether attain a nearly planar conformation. This distortion did not result in significant bending of the DNA, a result which was confirmed by gel electrophoresis studies of multimers of a 21-mer duplex containing the cross-link.

Influence of the R(61,2)- and S(61,2)-alpha-(N6-adenyl)styrene oxide adducts on the A.C mismatched base pair in an oligodeoxynucleotide containing the human N-ras codon 61.

Conformational studies of R- and S-alpha-(N6-adenyl)styrene oxide adducts mismatched with deoxycytosine at position X6 in d(CGGACXAGAAG).d(CTTCTCGTCCG), incorporating codons 60, 61 (underlined), and 62 of the human N-ras protooncogene, are described. These were the R- and S(61,2)C adducts. The S(61,2)C adduct afforded a stable solution structure, while the R(61,2)C adduct resulted in a disordered structure. Distance restraints for the S(61, 2)C adduct were calculated from NOE data using relaxation matrix analysis. These were incorporated as effective potentials into the total energy equation. The structures were refined using restrained molecular dynamics calculations which incorporated a simulated annealing protocol. The accuracy of the emergent structures was evaluated by complete relaxation matrix methods. The structures refined to an average rms difference of 1.07 A, determined by pairwise analysis. The experimentally determined structure was compared to NOE intensity data using complete relaxation matrix back-calculations, yielding an R1x value of 11.2 x 10(-)2. The phenyl ring of the styrene in the S(61,2)C adduct was in the major groove and remained oriented in the 3'-direction as observed for the corresponding S(61,2) adduct paired with thymine [Feng, B., Zhou, L., Pasarelli, M., Harris, C. M., Harris, T. M., and Stone, M. P. (1995) Biochemistry 34, 14021-14036]. A shift of the modified adenine toward the minor groove resulted in the styrenyl ring stacking with nucleotide C5 on the 5'-side of the lesion, which shifted toward the major groove. Unlike the unmodified A.C mismatch, neither the S(61,2)C nor the R(61,2)C adduct formed protonated wobble A.C hydrogen bonds. This suggests that protonated wobble A.C pairing need not be prerequisite to low levels of alpha-SO-induced A --> G mutations. The shift of the modified adenine toward the minor groove in the S(61,2)C structure may play a more important role in the genesis of A --> G mutations. The disordered structure of the R(61,2)C adduct provides a potential explanation as to why that adduct does not induce A --> G mutations.

Multiple conformations of an intercalated (-)-(7S,8R,9S, 10R)-N6-[10-(7,8,9,10-tetrahydrobenzo[a]pyrenyl)]-2'-deoxyadenosyl adduct in the N-ras codon 61 sequence.

The structure of the (-)-(7S,8R,9S,10R)-N6-[10-(7,8,9, 10-tetrahydrobenzo[a]pyrenyl)]-2'-deoxyadenosyl adduct at A7 of 5'-d(CGGACAAGAAG)-3'.5'-d(CTTCTTGTCCG)-3', derived from trans addition of the exocyclic N6-amino group of dA to (-)-(7S,8R,9R, 10S)-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene [(-)-DE2], was determined using molecular dynamics simulations restrained by 532 NOEs from 1H NMR. This was named the SRSR(61,3) adduct, derived from the N-rasprotooncogene at and adjacent to the nucleotides encoding amino acid 61 (underlined) of the p21 gene product. The solution structure of this adduct was best described as a mixture of two conformations in rapid equilibrium on the NMR time scale. The two populations differed in the pseudorotation angle of the sugar ring for the 5'-neighboring base A6, as determined from scalar coupling data. One population, estimated to be present at 53%, had the A6 deoxyribose in the C2'-endo conformation, while in the second conformation the A6 deoxyribose was in the C3'-endo conformation. NOEs between C5, A6, and SRSRA7 were either disrupted or weakened, as were those in the complementary strand between C15, T16, and T17. Major groove NOEs were observed between the benzo[a]pyrene aromatic protons, H1, H2, H3, H4, H5, and H6, and T16 CH3. Minor groove NOEs were observed between H1, H2, and H3 of benzo[a]pyrene and T16 H1' and H2' and T17 H1' and H2'. The benzo[a]pyrene protons H10, H11, and H12 showed NOEs to A6 H1', H2', and H2". The chemical shifts of the pyrenyl moiety were dispersed over a 1.9 ppm range. Upfield chemical shifts of 2.4 ppm for T16 N3H, 1.1 ppm for T17 N3H, 1.3 and 1.0 ppm for T16 H6 and CH3, 0.85 ppm for T16 H1', and 0.80 and 0.90 ppm for C15 H2' and H2" were observed. These observations were consistent with intercalation of the pyrenyl moiety toward the 5' direction of SRSRA7. The results were compared to the isomeric SRSR(61,2) adduct [I. S. Zegar, S. J. Kim, T. N. Johansen, P. J. Horton, C. M. Harris, T. M. Harris, and M. P. Stone (1996) Biochemistry 35, 6212-6224] and revealed the role of DNA sequence in modulating the conformation of this benzo[a]pyrene adduct.

Adduction of the human N-ras codon 61 sequence with (-)-(7S,8R,9R,10S)-7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a] pyrene: structural refinement of the intercalated SRSR(61,2) (-)-(7S,8R,9S,10R)-N6-[10-(7,8,9,10- tetrahydrobenzo[a]pyrenyl)]-2'-deoxyadenosyl adduct from 1H NMR.

The structure of the (-)-(7S,8R,9S,10R)-N6-[10-(7,8,910-tetrahydrobenzo [a]pyrenyl)]-2'-deoxyadenosyl adduct at X6 of 5'-d(CGGACXAGAAG)-3'-5'-d(CTTCTTGTCCG)-3', derived from trans addition of the exocyclic N6-amino group of dA to (-)-(7S,8R,9R,10S)-7, 8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene [(-)-DE2], was determined using molecular dynamics simulations restrained by 369 NOEs from 1H NMR. This was named the SRSR(61,2) adduct, derived from the N-ras protooncogene at and adjacent to the nucleotides encoding amino acid 61 (underlined) of the p21 gene product. NOEs between C5, S.R.S.R A6, and A7 were disrupted, as were those between T17 and G18. NOEs between benzo[a]pyrene and DNA protons were localized on the two faces of the pyrenyl ring. The benzo[a]pyrene H3-H6 protons showed NOEs to T17 CH3, while H1, H2, and H3 showed NOEs to T17 deoxyribose; the latter protons and H4 showed NOEs to T17 H2', H2" and to T17 H6. Noes were observed between H11 and H12 and C5 H]',H2', H2". G18 N1H showed NOEs to both faces of benzo[a]pyrene. Upfield shifts of 2.6 ppm for T17 N3H and 1.8 ppm for G18 N1H. 1 ppm for T17 H6 and CH3, and 0.75 ppm for C5 H5, with a smaller shift for C5 H6, and a 1.5 ppm dispersion of the pyrenyl protons suggested that benzo[a]pyrene intercalated above the 5'-face of S.R.S.R A6. The precision of the refined structures was monitored by pairwise root mean square deviations. which were < 1.5 A; accuracy was measured by complete relaxation matrix calculations, which yielded a sixth root R factor of 8.1 x 10(-2). Interstrand stacking between the pyrenyl ring and the T17 pyrimidine and G18 purine rings was enhanced by the bay ring. Changes of +30 degrees and -25 degrees in buckle for C5.G18 and S.R.S.R A6.T17, respectively, were calculated, as was a -40 degrees change in propeller twist for C5.G18. The rise between C5.G18 and S.R.S.R A6.T17 was calculated to be 7 A. The work extended the pattern for adenine N6 benzo[a]pyrene adducts, in which the R stereochemistry at C10 predicted 5'-intercalation of the pyrenyl moiety.

Styrene oxide adducts in an oligodeoxynucleotide containing the human N-ras codon 12 sequence: structural refinement of the minor groove R(12,2)- and S(12,2)-alpha-(N2-guanyl) stereoisomers from 1H NMR.

The structures of the (R)- and (S)-alpha-(N2-guanyl)styrene oxide adducts at X6 in d(GGCAGXTGGTG).d(CACCACCTGCC), encompassing codon 12 of the human n-ras protooncogene (underlined), were refined from 1H NMR data. These were the R(12,2) and S(12,2) adducts. For the R(12,2) adduct, upfield chemical shifts were observed for the T7 H6, H1', and N3H resonances. At 30 degrees C, R-SOG 6 N1H, T7 N3H, and T10 N3H disappeared due to exchange with solvent. For the S(12,2) adduct, S-SOG6 H1' shifted upfield 0.33 ppm, but all imino resonances were observed. The styrene methylene protons were nonequivalent for both adducts, suggesting hydrogen bonding between the hydroxyl and C18 O2 or O4' in the R(12,2) adduct and C17 O2 in the S(12,2) adduct. The styrene aromatic protons appeared as three signals in the R(12,2) adduct and as two signals in the S(12,2) adduct, suggesting rapid rotation of the styrene ring on the NMR time scale. NOE data revealed that the phenyl ring was oriented in the 3'-direction relative to R-SOG6 for the R(12,2) adduct and in the 5'-direction relative to S-SOG6 for the S(12,2) adduct. A total of 253 and 221 interproton distances were obtained from relaxation matrix analyses of the R(12,2) and S(12,2) adducts, respectively. NOE-restrained molecular dynamics calculations converged with root mean square deviations of 0.8-1.2 A for the R(12,2) adduct and 0.82-1.4 A for the S(12,2) adduct. Complete relaxation matrix analyses of the nine inner base pairs yielded sixth root residual indices between calculated and experimental NOE intensities of 8.8 x 10(-2) for the R(12,2) adduct and 7.9 x 10(-2) for the S(12,2) adduct. The refined structure for the R(12,2) adduct showed a 0.4 A increase in the stretch of R-SOG6.C17 and T7.A16, and a 1-2 A widening of the minor groove at and adjacent to the SO lesion, with the styrene ring oriented edgewise in the minor groove. Smaller minor groove disturbances were observed for the S(12,2) adduct, which had the styrene ring oriented flat in the minor groove. No DNA bending was predicted by the calculated structures.

Solution structure of an oligodeoxynucleotide containing the human N-ras codon 12 sequence refined from 1H NMR using molecular dynamics restrained by nuclear Overhauser effects.

The structure of d(GGCAGGTGGTG).d(CACCACCTGCC), consisting of codons 11, 12 (underlined), and 13 of the human n-ras protooncogene, was refined from 1H NMR data. Patterns of internucleotide NOEs consistent with a B-form helix were observed for each strand. NOE intensities between purine H8 and H1' protons were small compared to intensities between cytosine H5 and H6 protons, indicative of glycosyl torsion angles in the anti range. Cross-peaks were observed between purine H8 and pyrimidine H5 and CH3 protons on adjacent bases in the direction of purine (5'-->3')pyrimidine, but not in the direction pyrimidine(5'-->3')purine. Watson-Crick hydrogen bonding between bases was intact. A total of 232 experimental distance restraints were obtained. Of these, 143 were intra-residue restraints and 89 were inter-residue restraints. A restrained molecular dynamics/simulated annealing approach yielded 6 MD structures calculated from a B-form starting structure and 6 MD structures from an A-form starting structure. These refined to an average pairwise rms difference of 0.92 angstrom, with maximum pairwise rmsd of 1.35 angstroms. Accuracy of emergent structures was assessed by relaxation matrix back-calculation. The sixth-root residual index of 7.0 x 10(-2) measured between the refined structures and the NOE intensity data suggested that the former were in reasonable agreement with the NOE data. The refined solution structures were in the B-family. Similar to the human n-ras codon 61 sequence [Feng, B., & Stone, M.P. (1995) Chem. Res. Toxicol. 8, 821-832], the ras12 sequence contained local variations in B-like conformation which did not confer large structural alterations upon the duplex, but perhaps modulated the reactivity of the first as compared to the second guanine in codon 12.

Formation of hemoglobin-benzo[a]pyrene adducts in human erythrocytes incubated with benzo[a]pyrene and hamster embryo cells.

Evidence is accumulating that the levels of covalent carcinogen-macromolecule adducts, including adducts with hemoglobin, reflect biologically effective levels of carcinogen exposure. The purposes of the present study were (a) to establish a cellular system for obtaining adducts between intracellular human hemoglobin and metabolites of polycyclic aromatic hydrocarbons (PAH), and (b) to evaluate techniques for chromatographic characterization of the adducts. We showed that hemoglobin-benzo[a]pyrene adducts were formed when human erythrocytes were treated with [3H]benzo[a]pyrene (BP) in the presence of hamster embryo fibroblasts, which are known to be effective for BP metabolism. After lysis of the erythrocytes, noncovalently bound BP and its metabolites were effectively removed from hemoglobin under mild conditions by using hydrophobic interaction and size-exclusion liquid chromatography. Three to five distinct adducts were resolved by reversed-phase and ion-exchange liquid chromatography. As determined by a two-step, reversed-phase liquid chromatographic procedure, trypsin treatment of globin from the cellular system yielded at least three of the four 7,8,9,10-tetrahydro-7,8,9,10-tetrahydroxy BP tetrols known to arise from mammalian metabolism of BP. This observation is consistent with both (a) the recently described formation of labile carboxyl esters via reaction of BP-7,8-dihydrodiol-9,10-epoxide (BPDE) with hemoglobin and (b) the known formation of both anti- and syn-BPDE in hamster embryo fibroblasts. In addition, high-performance liquid chromatographic analysis demonstrated the presence of other products presumed to be BP-peptide adducts because of their susceptibility to thermolysin treatment.

A comparison of the intercalative binding of non-reactive benzo[a]pyrene metabolites and metabolite model compounds to DNA.

The reversible DNA physical binding of a series of non-reactive metabolites and metabolite model compounds derived from benzo[a]pyrene (BP) has been examined in UV absorption and in fluorescence emission and fluorescence lifetime studies. Members of this series have steric and pi electronic properties similar to the highly carcinogenic metabolite trans-7,8-dihydroxy-anti-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) and the less potent metabolite 4,5-epoxy-4,5-dihydrobenzo(a)pyrene (4,5-BPE). The molecules examined are trans-7,8-dihydroxy-7,8-dihydrobenzo[a]-pyrene (7,8-di(OH)H2BP), 7,8,9,10-tetrahydroxytetrahydrobenzo[a]pyrene (tetrol) 7,8,9,10-tetrahydrobenzo[a]pyrene (7,8,9,10-H4BP), pyrene, trans-4,5-dihydroxy-4,5-dihydrobenzo[a]pyrene (4,5-di(OH)H2BP) and 4,5-dihydrobenzo[a]pyrene (4,5-H2BP). In 15% methanol at 23 degrees C the intercalation binding constants of the molecules studied lie in the range 0.79-6.1 X 10(3) M-1. Of all the molecules examined the proximate carcinogen 7,8-di(OH)-H2BP is the best intercalating agent. The proximate carcinogen has a binding constant which in UV absorption studies is found to be 2.8-6.0 times greater than that of the other hydroxylated metabolites. Intercalation is the major mode of binding for 7,8-di(OH)H2BP and accounts for more than 95% of the total binding. Details concerning the specific role of physical bonding in BP carcinogenesis remain to be elucidated. However, the present studies demonstrate that the reversible binding constants for BP metabolites are of the same magnitude as reversible binding constants which arise from naturally occurring base-base hydrogen bonding and pi stacking interactions in DNA. Furthermore, previous autoradiographic studies indicate that in human skin fibroblasts incubated in BP, pooling of the unmetabolized hydrocarbons occurs at the nucleus. The high affinity of 7,8-di(OH)H2BP for DNA may play a role in similarly elevating in vivo nuclear concentrations of the non-reactive proximate carcinogen.

A comparison of the DNA intercalative binding of bay versus K region metabolites of benzo[a]pyrene.

The DNA intercalating properties of trans-7,8-dihydroxy-7,8-dihydrobenzo[a]pyrene (1) and of trans-4,5-dihydroxy-4,5-dihydrobenzo[a]pyrene (2) have been compared in UV absorption and in fluorescence emission and fluorescence lifetime studies. Molecules 1 and 2 represent steric models of the two epoxide containing metabolites of benzo[a]pyrene, trans-7,8-dihydroxy-anti-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE) and benzo[a]pyrene-4,5-oxide. The former of these metabolites is a highly carcinogenic bay region metabolite, the latter is a much less carcinogenic K region metabolite. The association constant for intercalation for model 1 is 5,226 M-1. This is more than 2.7 times greater than that for molecule 2. These results taken together with results form previous studies of bay and K region metabolite models of benz[a]anthracene suggest that intercalation is important to the overall carcinogenic activity of polycyclic aromatic hydrocarbons.

Intercalative DNA binding of model compounds derived from metabolites of 7,12-dimethylbenz[a]anthracene.

The DNA binding of nonreactive model compounds of metabolites of 7,12-dimethylbenz[a]-anthracene (DMBA)1 was studied in fluorescence quenching and fluorescence lifetime experiments. The model compounds examined were DMA and 8,9,10,11-tetrahydro-BA. DMA is a pi electron model of a highly carcinogenic bay region epoxide of DMBA, 8,9,10,11-tetrahydro-BA is a model compound of a less carcinogenic DMBA epoxide. The results indicate that the binding of DMA occurs primarily via intercalation. In 15% methanol the binding constant is 3.1 x 10(3) M-1. In 15% methanol and at DNA phosphate levels of 5.0 x 10(-4) M the intercalative binding of DMA is reduced by a factor of 6.2 when 5.0 x 10(-4) M Mg+2 is added. The DMA binding constant for intercalation is reduced by more than a factor of 4 when the methanol content of the solvent is increased from 0% to 20%. Finally DMA binding arising from pi interactions with the DNA bases is reduced more than 15 times when the DNA is denatured. For 8,9,10,11-tetrahydro-BA in 15% methanol the binding constant for intercalation is 6 times lower than that for DMA. These results along with previously reported binding data on other model compounds suggest that bay region metabolites of DMBA readily participate in physical pi stacking interactions with DNA.

Fluorescence and photoelectron studies of the intercalative binding of benz(a)anthracene metabolite models to DNA.

DNA binding of nonreactive metabolite models derived from benz(a)anthracene was studied. The molecules investigated include 1,2,3,4-tetrahydrobenz(a)anthracene (1), 5,6-dihydrobenz(a)anthracene (2), and 8,9,10,11-tetrahydrobenz(a)anthracene (3), as well as anthracene and phenanthrene. Measurements of the effects of DNA binding upon fluorescence intensities and fluorescence lifetimes indicate that molecules 1 and 3 (KA = 1.5 - 2.5 x 10(3) M-1) bind more strongly to native DNA than does molecule 2 (KA congruent to 0.5 x 10(3) M-1). Furthermore, molecules 1 and 3 bind to DNA much more effectively than do the two less sterically hindered pi electron metabolite models, anthracene and phenanthrene. Photoelectron data suggests that the enhanced binding of molecules 1 and 3 is due to increases in polarizability. Experiments carried out with denatured DNA indicate that the binding of molecule 1 entails the greatest intercalation.