Aristolochic acids (AAI and AAII), produced by the Aristolochiaceae family of plants, are classified as group I (human) carcinogens by the International Agency for Research on Cancer. These acids are metabolized in cells to yield aristolactams (ALI and ALII, respectively), which further form bulky adducts with the purine nucleobases. Specifically, the adenine lesions are more persistent in cells and have been associated with chronic renal diseases and related carcinogenesis. To understand the structural basis of the nephrotoxicity induced by AAs, the ALI-N(6)-dA and ALII-N(6)-dA lesions are systematically studied using computational methods. Density functional theory calculations indicate that the aristolactam moiety intrinsically prefers a planar conformation with respect to adenine. Nucleoside and nucleotide models suggest that the anti and syn orientations about the glycosidic bond are isoenergetic for both adducts. Molecular dynamics simulations and free energy calculations reveal that the anti base-displaced intercalated conformation is the most stable conformer for both types of AL-N(6)-dA adducted DNA, which agrees with previous experimental work on the ALII-N(6)-dA adduct and thereby validates our approach. Interestingly, this conformer differs from the dominant conformations adopted by other N6-linked adenine lesions, including those derived from polycyclic aromatic hydrocarbons. Furthermore, the second most stable syn base-displaced intercalated conformation lies closer in energy to the anti base-displaced intercalated conformation for ALI-N(6)-dA compared to ALII-N(6)-dA. This indicates that a mixture of conformations may be detectable for ALI-N(6)-dA in DNA. If this enhanced conformational flexibility of double-stranded DNA persists when bound to a lesion-bypass polymerase, this provides a possible structural explanation for the previously observed greater nephrotoxic potential for the ALI versus ALII-N(6)-dA adduct. In addition, the structural characteristics of the preferred conformations of adducted DNA explain the resistance of these adducts to repair and thereby add to our current understanding of the toxicity of AAs within living cells.

There were 1765 contacts identified between DNA nucleobases or deoxyribose and cyclic (W, H, F, Y) or acyclic (R, E, D) amino acids in 672 X-ray structures of DNA-protein complexes. In this first study to compare π-interactions between the cyclic and acyclic amino acids, visual inspection was used to categorize amino acid interactions as nucleobase π-π (according to biological edge) or deoxyribose sugar-π (according to sugar edge). Overall, 54% of contacts are nucleobase π-π interactions, which involve all amino acids, but are more common for Y, F, and R, and involve all DNA nucleobases with similar frequencies. Among binding arrangements, cyclic amino acids prefer more planar (stacked) π-systems than the acyclic counterparts. Although sugar-π interactions were only previously identified with the cyclic amino acids and were found to be less common (38%) than nucleobase-cyclic amino acid contacts, sugar-π interactions are more common than nucleobase π-π contacts for the acyclic series (61% of contacts). Similar to DNA-protein π-π interactions, sugar-π contacts most frequently involve Y and R, although all amino acids adopt many binding orientations relative to deoxyribose. These DNA-protein π-interactions stabilize biological systems, by up to approximately -40 kJ mol(-1) for neutral nucleobase or sugar-amino acid interactions, but up to approximately -95 kJ mol(-1) for positively or negatively charged contacts. The high frequency and strength, despite variation in structure and composition, of these π-interactions point to an important function in biological systems.