The structure of the cationic 8,9-dihydro-8-(N7-guanyl)-9-hydroxy-aflatoxin B(1) adduct embedded in a 5'-CpG-3' sequence context and paired with deoxycytosine in the oligodeoxynucleotide d(ACATC(AFB)GATCT) x d(AGATCGATGT) was refined using molecular dynamics calculations restrained by NOE data and dihedral angle restraints obtained from NMR data. The aflatoxin moiety intercalated above the 5' face of the modified guanine. It stacked between C(5) x G(16) and (AFB)G(6) x C(15). The AFB(1) H5, OCH(3), and methylene protons faced into the minor groove, with the methylene protons oriented between the C(15) and G(16) nucleobases. The aflatoxin B(1) H6a, H8, H9, and H9a protons faced the major groove, with H6a and H9a pointing toward the 5' direction from the lesion site. The refined structure was compared to the structure of the aflatoxin B(1) adduct embedded in a 5'-ATGCAT-3' sequence in the oligodeoxynucleotide d(TAT(AFB)GCATA)(2) [Jones, W. R., Johnston, D. S., and Stone, M. P. (1998) Chem. Res. Toxicol.11, 873-881]. The structure of the intercalated aflatoxin B(1) lesion in the ATC(AFB)GAT sequence is similar to its structure in the d(AT(AFB)GCAT) sequence. This is consistent with a mechanism in which the precovalent intercalation of aflatoxin-8,9-exo-epoxide on the 5' face of guanine places the epoxide in close proximity and in the proper orientation to the N7 position of guanine, thus facilitating an S(N)2 reaction. The data provides additional insight into the nature of the disruption of the B-DNA duplex induced by aflatoxin B(1) intercalation. Overall, the results suggest that sequence contributes a minor role in modulating the structure of the cationic guanine N7 AFB(1) lesion in duplex DNA. On the other hand, structural differences are observed when the correctly paired structure is compared to the structure of the cationic AFB(1) adduct mispaired with dA [Giri, I., Johnston, D. S., and Stone, M. P. (2002) Biochemistry 41, 5462-5472].

The trans-8,9-dihydro-8-(N7-guanyl)-9-hydroxyaflatoxin B(1) cationic guanine N7 adduct of aflatoxin B(1) thermally stabilizes the DNA duplex, as reflected in increased T(m) values upon adduction. The magnitude of the increased T(m) value is characteristically 2-3 degrees C. The major rotamer of the neutral guanine N7 adduct trans-8,9-dihydro-8-(2,6-diamino-4-oxo-3,4-dihydropyrimid-5-yl-formamido)-9-hydroxy aflatoxin B(1) (the FAPY major adduct) exhibits a 15 degrees C increase in T(m) in 5'-d(CTAT(FAPY)GATTCA)-3'-5'-d(TGAATCATAG)-3'. Site-specific mutagenesis experiments reveal the FAPY major adduct induces G-->T mutations in Escherichia coli at a frequency six times higher than that of the cationic adduct (Smela, M. E.; Hamm, M. L.; Henderson, P. T.; Harris, C. M.; Harris, T. M.; Essigmann, J. M. Proc Natl Acad Sci USA, 99, 6655-6660). Thus, the FAPY major lesion may account substantially for the genotoxicity of AFB(1). Structural studies for cationic and FAPY adducts of aflatoxin B(1) suggest both adducts intercalate above the 5'-face of the modified deoxyguanosine and that in each instance the aflatoxin moiety spans the DNA helix. Intercalation of the aflatoxin moiety, accompanied by favorable stacking with the neighboring base pairs, is thought to account for the increased thermal stability of the aflatoxin cationic guanine N7 and the FAPY major adducts. However, the structural basis for the large increase in thermal stability of the FAPY major adduct in comparison to the cationic guanine N7 adduct of aflatoxin B(1) is not well understood. In light of the site-specific mutagenesis studies, it is of considerable interest. For both adducts, the intercalation structures are similar, although improved stacking with neighboring base pairs is observed for the FAPY major adduct. In addition, the presence of the formamido group in the aflatoxin B(1) FAPY major adduct may enhance duplex stability, perhaps via intrastrand sequence-specific hydrogen bonding interactions within the duplex.

The structure of 5'-d(ACATC(AFB)GATCT)-3'.5'-d(AGATCAATGT)-3', containing the C(5).A(16) mismatch at the base pair 5' to the modified (AFB)G(6), was determined by NMR. The characteristic 5'-intercalation of the AFB(1) moiety was maintained. The mismatched C(5).A(16) pair existed in the wobble conformation, with the C(5) imino nitrogen hydrogen bonded to the A(16) exocyclic amino group. The wobble pair existed as a mixture of protonated and nonprotonated species. The pK(a) for protonation at the A(16) imino nitrogen was similar to that of the C(5).A(16) wobble pair in the corresponding duplex not adducted with AFB(1). Overall, the presence of AFB(1) did not interfere with wobble pair formation at the mismatched site. Molecular dynamics calculations restrained by distances derived from NOE data and torsion angles derived from (1)H (3)J couplings were carried out for both the protonated and nonprotonated wobble pairs at C(5).A(16). Both sets of calculations predicted the A(16) amino group was within 3 A of the C(5) imino nitrogen. The calculations suggested that protonation of the C(5).A(16) wobble pair should shift C(5) toward the major groove and shift A(16) toward the minor groove. The NMR data showed evidence for the presence of a minor conformation characterized by unusual NOEs between T(4) and (AFB)G(6). T(4) is two nucleotides in the 5'-direction from the modified base. These NOEs suggested that in the minor conformation nucleotide T(4) was in closer proximity to (AFB)G(6) than would be expected for duplex DNA. Modeling studies examined the possibility that T(4) transiently paired with the mismatched A(16), allowing it to come within NOE distance of (AFB)G(6). This model structure was consistent with the unusual NOEs associated with the minor conformation. The structural studies are discussed in relationship to nontargeted C --> T transitions observed 5' to the modified (AFB)G in site-specific mutagenesis experiments [Bailey, E. A., Iyer, R. S., Stone, M. P., Harris, T. M., and Essigmann, J. M. (1996) Proc. Natl. Acad. Sci. U.S.A. 93, 1535-1539].

The G --> T transversion is the dominant mutation induced by the cationic trans-8,9-dihydro-8-(N7-guanyl)-9-hydroxy-aflatoxin B(1) adduct. The structure of d(ACATC(AFB)GATCT).d(AGATAGATGT), in which the cationic adduct was mismatched with deoxyadenosine, was refined using molecular dynamics calculations restrained by NOE data and dihedral restraints obtained from NMR spectroscopy. Restrained molecular dynamics calculations refined structures with pairwise rmsd<1 A and a sixth root R1x factor between the refined structure and NOE data of 10.5 x 10-2. The mismatched duplex existed in a single conformation at neutral pH. The aflatoxin moiety intercalated above the 5' face of the modified (AFB)G. The mismatched dA was in the anti conformation about the glycosyl bond. It extruded toward the major groove and did not participate in hydrogen bonding with (AFB)G. The structure was compared with that of d(ACATCGATCT).d(AGATAGATGT) containing the corresponding unmodified G.A mismatch and with d(ACATC(AFB)GATCT).d(AGATCGATGT) containing the aflatoxin lesion in the correctly paired (AFB)G.C context. The correctly paired oligodeoxynucleotide exhibited Watson-Crick-type geometry at the (AFB)G.C pair. It melted at higher temperature than the mismatched (AFB)G.A duplex. The unmodified mismatched G.A duplex exhibited spectral line broadening at neutral pH, suggesting a mixture of conformations. It exhibited a lower melting temperature than did the mismatched (AFB)G.A duplex. These differences correlated with replication bypass experiments performed in vitro utilizing DNA polymerase I exo- [Johnston, D. S., and Stone, M. P. (2000) Chem. Res. Toxicol. 13, 1158-1164]. Those experiments showed that correct insertion of dC opposite (AFB)G blocked replication by the enzyme, whereas incorrect insertion of dA opposite (AFB)G allowed full-length replication of the adducted template strand.

This study probes the potential of variable-temperature high performance liquid chromatography (VT-HPLC) as a tool for dissecting and modulating nucleic acid structural transitions, using as a model the duplex-hairpin-coil transitions of d(CGCGAATTCGCG). It is demonstrated that VT-HPLC, combined with diode-array detection of the uv signal, enables, for the first time, a physical separation of spectroscopically distinct species that can be assigned to the duplex, hairpin, and coil forms of d(CGCGAATTCGCG). Although the species are spectroscopically distinguishable, they are not readily isolated. Hence, if fractions from the peaks for hairpin or duplex forms are collected and subsequently reinjected onto the cartridge, reequilibration occurs, and both hairpin and duplex peaks are observed. Area integration of the peaks corresponding to duplex and hairpin species provides a means to monitor the duplex to hairpin transition at effective concentrations in the nanomolar range, well below that accessible by conventional spectrophotometers. Concentration-dependent equilibrium constants, melting temperatures, and standard state enthalpies extracted from our measurements compare very well with previous literature results, and with our own results that take into account the effect of our solvent conditions [100 mM TEAA (triethylammonium acetate) and variable acetonitrile] on the melting behavior. By combining precise temperature control with separation based on size, physical behavior, and interaction free energies, VT-HPLC provides a powerful tool for both the modulation and the separation of nucleic acid conformations.