PSCDD · Protein-Specific Charges for Drug Discovery

Protein-Specific Charges for Drug Discovery

Optimisation of the inhibitory potency of a drug candidate requires an accurate description of the interactions between the compound and its target protein receptor. State-of-the-art computational simulations of drug potency rely on classical molecular mechanics force fields with model parameters that have been collected over many decades from experiments or computational analysis of small molecules. Yet recent advances in linear-scaling density functional theory software allow accurate quantum mechanical simulations of biomolecules comprising many thousands of atoms. The goal of the Marie Curie IOF project, “Protein-Specific Charges for Drug Discovery (PSCDD)” is to enhance the accuracy of computational predictions of drug binding by deriving parameters of the target receptor directly from a single quantum mechanical simulation of the entire protein, thus incorporating the polarisation of the protein's native state into the force field derivation procedure. In this way, we hope to improve the efficiency and accessibility of the drug discovery process by allowing medical researchers to focus their design efforts on synthesising only the most promising compounds. To this end, the density derived electrostatic and chemical (DDEC) charge derivation method has been implemented in the linear-scaling density functional theory (DFT) code ONETEP as a post-processing tool. The work allows the straightforward derivation of atomic charges for systems comprising thousands of atoms, including entire proteins. The derived charges have been extensively validated for use in flexible force fields: i) the ONETEP charges agree with benchmark calculations using quantum chemistry software; ii) the charges are computed with linear-scaling computational cost; iii) they reproduce the quantum mechanical electrostatic properties of small and large molecules, including proteins and nanorods; iv) dynamical simulations of proteins agree with experimental NMR data. An additional unforeseen benefit of the atoms-in-molecule method is that it may also be used to derive the Lennard-Jones parameters of the force field, which are used to describe the attractive dispersive interactions and repulsive interactions between overlapping electron clouds. Thus, remarkably, all non-bonded parameters of molecular mechanics force fields for biological molecules may now be derived directly from quantum mechanical simulations, rather than fit to experimental data. The DDEC analysis scheme is available to users of ONETEP and is documented on the website (www.onetep.org/Main/Documentation). Early training and drug design efforts were focused on inhibition of the HIV-1 reverse transcriptase protein, with factors affecting protein-ligand binding being investigated using two complementary software packages, MCPRO and Desmond. An enhanced sampling scheme has been implemented in MCPRO to improve the reliability and accuracy of computed free energies, and the scheme is included in the latest software release. In the final year, efforts focused on developing a library of bonded parameters that are consistent with the derived non-bonded force field parameters. These new parameters are currently being validated against experimentally measured benchmark potencies of fifteen p38 MAP kinase inhibitors and will be published this year. During the project, eleven papers have been published in peer-reviewed journals, including three in the Journal of Chemical Theory and Computation on the development of the protein-specific force field and two in the field of drug discovery. The results have also been presented at fourteen seminars, pharmaceutical company meetings, and international conferences, including the American Chemical Society March meeting and the Psi-K electronic structure conference. In addition, the researcher has organised the first Cavendish Laboratory pharmaceutical industry engagement symposium for 20 industrial and academic participants with the aim of showcasing medical research at the University of Cambridge physics department. Ongoing collaborations with the Cambridge MRC Cancer Unit and Yale University will allow us to continue to extensively validate the protein-specific force fields against a range of benchmark systems and to search for novel pharmaceutical compounds. The researcher has recently been appointed to a lectureship in computational medicinal chemistry at Newcastle University, which was made possible by the training that he received during his Marie Curie fellowship. With the rising cost of developing pharmaceuticals (around $1.8 billion for every new molecule that reaches the consumer), computational pre-screening methods that guide traditional medicinal Chemistry optimisation processes will prove to be extremely valuable. The software and protocols developed in this project improve the accuracy and reliability of computational lead optimisation, whilst remaining user-friendly and relatively inexpensive to run. A topical review concerning the use of large-scale density functional theory calculations in biology, including applications in medicinal chemistry has been prepared by the researcher, and more details are available at: http://www.tcm.phy.cam.ac.uk/~djc56/marie_curie.html

Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз