ABSTRACT
The Hepatitis B virus (HBV) core protein is an attractive target for preventing capsid assembly and viral replication. Drug repurposing strategies have introduced several drugs targeting HBV core protein. This study used a fragment-based drug discovery (FBDD) approach to reconstruct a repurposed core protein inhibitor to some novel antiviral derivatives. Auto Core Fragment in silico Screening (ACFIS) server was used for deconstruction-reconstruction of Ciclopirox in complex with HBV core protein. The Ciclopirox derivatives were ranked based on their free energy of binding (ΔGB). A quantitative structure affinity relationship (QSAR) was established on the Ciclopirox derivatives. The model was validated by a Ciclopirox-property-matched decoy set. A principal component analysis (PCA) was also assessed to define the relationship of the predictive variable of the QSAR model. 24-derivatives with a ΔGB (-16.56±1.46 Kcal.mol-1) more than Ciclopirox was highlighted. A QSAR model with a predictive power of 88.99% (F-statistics = 9025.78, corrected df(25), Pr > F = 0.0001) was developed by four predictive descriptors (ATS1p, nCs, Hy, F08[C-C]). The model validation showed no predictive power for the decoy set (Q2 = 0). No significant correlation was observed between predictors. By directly attaching to the core protein carboxyl-terminal domain, Ciclopirox derivatives may be able to suppress HBV virus assembly and subsequent viral replication inhibition. Hydrophobic residue Phe23 is a critical amino acid in the ligand binding domain. These ligands share the same physicochemical properties that lead to the development of a robust QSAR mode. The same strategy may also be used for future drug discovery of viral inhibitors.
