Computational method for predicting passive permeability of naturally-sourced drugs ABSTRACT The major obstacle to optimization of bioavailability of drug leads is the lack of a conceptual framework that adequately describes different aspects of drug-membrane interactions and predicts their ability to cross the hydrophobic barrier of the lipid bilayer. There is a critical need for development of a conceptual model and an efficient computational tool for quantification of passive permeability across membranes of diverse molecules, including naturally-sourced compounds with properties that lie outside physicochemical parameters of drug-like molecules. The proposed research is in line with our long-term goal: to develop computational methods and resources for studying solute-solvent interactions and spatial distributions of molecules in biological membranes, which can assist in design of new therapeutics to treat human diseases. The objective in this proposal is to develop and validate a novel physics-based approach for predicting Permeability of Molecules through Membranes (PerMM) and organelle targeting. The method will perform all-atom modeling of binding and passive diffusion of biomolecules in natural membranes with highly anisotropic and asymmetric properties. The first specific aim is to create the PerMM method and implement it in an open-access web server that will calculate the following parameters for a specified compound and a selected membrane type: (1) the membrane binding energy and penetration depth; (2) the lowest energy pathway across the membrane; (3) the permeability coefficient; and (4) the molecular flux across the membrane in the presence of surface and transmembrane electric potentials and pH gradient. The results obtained for various biological membranes will allow prediction of compound accumulation in a specific cell compartment.
The second aim i s to apply the newly designed server for the systematic evaluation of permeability through different biomembranes, including the blood-brain barrier, of naturally-sourced drugs with focus on peptide-derived compounds. The server will also assist in optimization of bioavailability of opioid peptides and peptidomimetics with analgesic activity, which are currently under development in the laboratory of Dr. Mosberg (UM). The proposed research is expected to have a high impact on the field of drug design, optimization, and targeted delivery using lipid- based systems.
The proposed project is relevant to public health because the development of a new physics-based computational method and a web server will assist investigators in design and optimization of permeability of new therapeutics based on natural products. The planned computational resources are relevant to the NIH's mission because they will help to ensure the Nation's capability to discover new drugs to cure human diseases.