Sequence-dependent binding of flavonoids to duplex DNA.

Abstract

The whole family of structurally distinct flavonoids has been recognized as a valuable source of prospective anticancer agents. There is experimental evidence demonstrating that some flavonoids, like flavopiridol (FLP) and quercetin (QUE), bind to DNA influencing their key physiological function. FLP is involved in the combined mode of interaction (intercalation and minor groove binding), while QUE is viewed as a minor groove binder. From a physical standpoint, experimental and theoretical studies have not so far provided a sufficiently consistent picture of the nature of interaction with DNA. Herein the sequence-dependent binding of FLP and of QUE (two representative examples of the structurally different flavonoids) with duplex DNA, containing a variety of the sequences of eight nucleotides (I: GGGGCCCC, II: GGCCGGCC, III: AAAATTTT, IV: AAGCGCTT, V: GCGCGCGC) in the 5'-strand, is investigated using a sophisticated molecular dynamics (MD) approach. For various parts (helix, backbone, bases) of the DNA structure, the change of asymptotic (in terms of an infinite length of MD simulation) configurational entropy, being the thermodynamic consequence of DNA flexibility change due to ligand binding, is explored. As far as the sequence-dependent extent of DNA flexibility change upon QUE (or FLP) binding is concerned, for the entire double helix, increased flexibility is observed for I (or I ≈ II), while increased rigidity is found to be in the order of V > III > II > IV (or III > V > IV) for the rest of sequences. For the backbone, increased rigidity in the order of V > III > II > IV > I (or III > V > IV > I > II) is generally observed. For the nucleobases, increased flexibility is determined for I and II (I > II for both ligands), while increased rigidity in the order of V ≈ III > IV (or III > V > IV) is reported for the other sequences. Of the overall increased rigidity of the DNA structure upon ligand binding that is observed for the sequences III, IV, and V, about 50-70% comes from the sugar-phosphate backbone. Noteworthy is that the increased flexibility of the entire double helix and of the complete system of nucleobases upon ligand binding is only established for sequence I. The insights are further subtly substantiated by considering the configurational entropy contributions at the level of individual nucleobase pairs and of individual nucleo-base pair steps and by analyzing the sequence dependent estimates of intra-base pair entropy and inter-base pair entropy. The GGC triplet, which is part of the central tetramer (GGCC) of I, is concluded to be critical for binding of flavonoids, while the effect of the presence of ligand to the flexibility of nucleobases is localized through the intra-base pair motion of the intercalation site and its immediate vicinity. G-rich DNA sequences with consecutive Gs going before and/or after the critical GGC code (such as I: GGGGCCCC) are proposed to be uniquely specific for flavonoids. The configurational entropy contribution, as an upper bound of the true entropy contribution to the free energy in noncovalent binding, is demonstrated to influence the fundamental discrimination (intercalation vs groove binding) of DNA-flavonoid recognition modes. Some interesting implications for the structure-based design of optimal DNA binders are discussed.