Computational Design of Small Molecules with Druglike Properties

Abstract

Many cancers overexpress antiapoptotic proteins, which can lead to poor chemotherapy response. This work reports on computational design of druglike small molecules that could potentially facilitate apoptosis by forming complexes with antiapoptotic proteins. Drugs based on small molecules that target a single protein can lead to drug resistance. On the other hand, a drug that operates too broadly may harm healthy tissue. An alternative is a drug that can bind two or more of the proteins. A single drug would be more economical and lead to fewer side effects than a combination of drugs with each one targeting a single protein. In this work, structures of experimentally known small molecules were used as templates to design new molecules. Common structural features of the experimental molecules were identified. The Osiris Property Explorer program was employed to study how the features influenced molecular druglike properties. Atomic substitutions and structural modifications were done to design new small molecules. Drug-related properties and potential toxicities of the molecules were determined and compared to those of commercial drugs. Molecules with no indicated toxic risks and optimal values of drug-related properties were used for docking studies in the ArgusLab program. Binding energies of stable configurations of the designed molecules and antiapoptotic proteins were calculated. Designed molecules that made the most stable complexes with two or more antiapoptotic proteins were identified. The potential to use the designed molecules in anticancer drug design is discussed.