Intuitive Density Functional Theory-Based Energy Decomposition Analysis for Protein-Ligand Interactions.

First-principles quantum mechanical calculations with methods such as density functional theory (DFT) allow the accurate calculation of interaction energies between molecules. These interaction energies can be dissected into chemically relevant components such as electrostatics, polarization, and charge transfer using energy decomposition analysis (EDA) approaches. Typically EDA has been used to study interactions between small molecules; however, it has great potential to be applied to large biomolecular assemblies such as protein-protein and protein-ligand interactions. We present an application of EDA calculations to the study of ligands that bind to the thrombin protein, using the ONETEP program for linear-scaling DFT calculations. Our approach goes beyond simply providing the components of the interaction energy; we are also able to provide visual representations of the changes in density that happen as a result of polarization and charge transfer, thus pinpointing the functional groups between the ligand and protein that participate in each kind of interaction. We also demonstrate with this approach that we can focus on studying parts (fragments) of ligands. The method is relatively insensitive to the protocol that is used to prepare the structures, and the results obtained are therefore robust. This is an application to a real protein drug target of a whole new capability where accurate DFT calculations can produce both energetic and visual descriptors of interactions. These descriptors can be used to provide insights for tailoring interactions, as needed for example in drug design.

Predicting catalysis: understanding ammonia synthesis from first-principles calculations.

Here, we give a full account of a large collaborative effort toward an atomic-scale understanding of modern industrial ammonia production over ruthenium catalysts. We show that overall rates of ammonia production can be determined by applying various levels of theory (including transition state theory with or without tunneling corrections, and quantum dynamics) to a range of relevant elementary reaction steps, such as N(2) dissociation, H(2) dissociation, and hydrogenation of the intermediate reactants. A complete kinetic model based on the most relevant elementary steps can be established for any given point along an industrial reactor, and the kinetic results can be integrated over the catalyst bed to determine the industrial reactor yield. We find that, given the present uncertainties, the rate of ammonia production is well-determined directly from our atomic-scale calculations. Furthermore, our studies provide new insight into several related fields, for instance, gas-phase and electrochemical ammonia synthesis. The success of predicting the outcome of a catalytic reaction from first-principles calculations supports our point of view that, in the future, theory will be a fully integrated tool in the search for the next generation of catalysts.

Energy Decomposition Analysis Based on Absolutely Localized Molecular Orbitals for Large-Scale Density Functional Theory Calculations in Drug Design.

We report the development and implementation of an energy decomposition analysis (EDA) scheme in the ONETEP linear-scaling electronic structure package. Our approach is hybrid as it combines the localized molecular orbital EDA (Su, P.; Li, H. J. Chem. Phys., 2009, 131, 014102) and the absolutely localized molecular orbital EDA (Khaliullin, R. Z.; et al. J. Phys. Chem. A, 2007, 111, 8753-8765) to partition the intermolecular interaction energy into chemically distinct components (electrostatic, exchange, correlation, Pauli repulsion, polarization, and charge transfer). Limitations shared in EDA approaches such as the issue of basis set dependence in polarization and charge transfer are discussed, and a remedy to this problem is proposed that exploits the strictly localized property of the ONETEP orbitals. Our method is validated on a range of complexes with interactions relevant to drug design. We demonstrate the capabilities for large-scale calculations with our approach on complexes of thrombin with an inhibitor comprised of up to 4975 atoms. Given the capability of ONETEP for large-scale calculations, such as on entire proteins, we expect that our EDA scheme can be applied in a large range of biomolecular problems, especially in the context of drug design.