A Structure-Based 3D-QSAR(CoMSIA) Study on a Series of Aryl Diketoacids (ADK) as Inhibitors of HCV RNA-dependent RNA Polymerase

Abstract

The diketoacid moiety of aryl α,γ-diketoacids (ADKs, Fig. 1) proved essential for antiviral activity against hepatitis C virus (HCV) RNA-dependent RNA polymerase (RdRp), and structure-activity optimization studies have led to the identification of an ADK inhibitor as one of the most potent HCV RdRp inhibitors reported. However, in spite of the extensive structure-activity relationship study, it has not been clear what controls the binding affinity of ADK analogues to the target enzyme. 3D-QSAR techniques, such as comparative molecular force field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA), are based on the experimental structure-activity relationship on specific biomolecule-ligand pair. This method is based only on the ligand structure and thus the spatial arrangement (or alignment) is crucial in determining the accuracy of these approaches. In case of ADK analogues, it is obvious that the common structural unit, diketoacid moiety, could be used as a template for atom based alignment. However, the atombased fit (using atoms of diketoacid) produced poor CoMFA and CoMSIA results (data not shown) presumably due to the highly flexible alkyl linker which connects two aromatic rings of ADK analogues (Fig. 1). This result implies that ADK analogues bind to the target enzyme in a characteristic active conformation, which cannot be identified by investigation of the ligand structures alone. Thus, ADK analogues should be docked into the binding site of the target enzyme to provide the active conformations, which can be used for the structure-based alignment. However, the lack of information about the binding site of ADK analogues at the target enzyme limits this approach. In this study, the binding site of ADK analogues in the HCV RdRp was proposed by using an unusual crystal structure of rUTP-HCV RdRp complex (PDB ID 1GX6) and structural similarity between rUTP and ADK. The ADK analogues were aligned by docking into the binding site, and a structure-based 3D-QSAR study was performed to correlate the biological activities of ADKs with their three-dimensional structures. The diketoacid moiety is famous for its metal-binding ability, and ADK analogues are known to bind the divalent metal ions at the active site of HCV RdRp. However, it remains unsolved how ADK analogues bind to the active site of a polymerase enzyme without formation of the complementary base pairing with the RNA template chain. Additionally, it has been speculated that ADK analogues might have different binding site around the active site of HCV RdRp. Recently, Bressanelli et al. reported the unusual crystal structure of rUTP bound to HCV RdRp and it shows that rUTP binding site is quite different from the active site with the base (uracil) hydrogen atoms bonded to the polypeptide main chain (PDB ID 1GX6). This alternative binding mode of rUTP is artificial in the sense that rUTP cannot bind in such a way in the presence of template, but it is conceivable that molecules with higher binding affinity to the rUTP-binding site in this mode can inhibit the catalytic activity of the enzyme. Thus, we set out to investigate the characteristic binding mode of rUTP, which resulted in construction of a pharmacophore model composed of three key interactions between rUTP and HCV RdRp (Fig. 2): (a) electrostatic interaction with two divalent metal ions (Mn), (b) H-bonding of triphosphate moiety to nearby amino acid residues (Phe224, Asp225), (c) H-bonding between uracil