Accurate, Efficient, and Insightful Quantum Chemistry Calculations of Non-Covalent Interactions: A Potential Method for Drug Design
Loading...
Files
Date
2016-02
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
One of the most promising ways to find drug candidates in drug design is to study the interactions between target biomolecules and drug candidates accurately. Although experimental screening methodologies keep improving, it is still costly and time-consuming to experimentally screen large numbers of potential compounds suitable to a target protein. On the other hand, computational screening technologies are an alternative method that can alleviate these challenges. However, these fast computational methods have lower accuracy than experimental approaches due to the approximations needed to make them computationally feasible. Hence, an efficient and accurate computational screening method for large numbers of potential compounds is in urgent need. For this reason, our group has recently developed a fragment-based quantum chemistry method called XSAPT (extended symmetry-adapted perturbation theory) to decompose the binding region (supersystem) between biomolecule and drug into biomolecular subsystems (fragments) in order to greatly reduce the computational cost without sacrificing accuracy. The main attractive feature of the XSAPT method is its ability to capture many-body polarization effects (important for systems with many fragments) which are omitted in traditional computational screening functions in drug design. Furthermore, the XSAPT interaction energy can be decomposed into physically meaningful energy components, and we can explore how chemical modification of a potential drug molecule may change its binding affinity by studying the interplay of various energy components. From a computational point of view, XSAPT is "embarrassingly parallelizable", consisting of independent tasks that can be distributed across processors to reduce the scaling to only linear with respect to the number of fragments as opposed to the typical quantum mechanical methods which are at least cubic scaling with respect to the total system size. This fast theoretical method gives accurate binding energies for a variety of challenging non-covalent complexes, and these impressive results indicate that XSAPT is suitable for different binding environments. For example, XSAPT predicts a qualitatively-correct binding trend for a series of ionic-organic complexes as compared to experiment. Thus, XSAPT provides a route to understanding and controlling the ion-macromolecular binding property by modifying the structure of the macromolecule (a protein for example). The target of our XSAPT method is to predict accurate interactions between biomolecules and drug candidates. XSAPT has been employed in studying the interactions between an anti-cancer drug and DNA, and this binding complex consists of 157 atoms. A benchmark binding energy of −33.6 ± 0.9 kcal/mol is available from quantum Monte Carlo (QMC) calculations. Our XSAPT method yields a binding energy of −33.4 kcal/mol, within the statistical error bars of the QMC benchmark. Hence, accurate binding energies between DNA and drug molecules can be achieved by our XSAPT method. In summary, we demonstrated that XSAPT not only reduces the computational cost but also affords chemically-accurate interaction energies between molecules. These characteristics make XSAPT a promising method for use in fragment-based drug design to pre-screen large numbers of potential drug molecules.
Description
Mathematical and Physical Sciences: 1st Place (The Ohio State University Edward F. Hayes Graduate Research Forum)
Keywords
drug design, fragmentation, electronic structure theory, perturbation theory, many-body polarization, linear scaling