Method for binding site identification by molecular dynamics simulation (silcs: site identification by ligand competitive saturation)

Abstract

The invention describes an explicit solvent all-atom molecular dynamics methodology (SILCS: Site Identification by Ligand Competitive Saturation) that uses small aliphatic and aromatic molecules plus water molecules to map the affinity pattern of a large molecule for hydrophobic groups, aromatic groups, hydrogen bond donors, and hydrogen bond acceptors. By simultaneously incorporating ligands representative of all these functionalities, the method is an in silico free energy-based competition assay that generates three-dimensional probability maps of fragment binding (FragMaps) indicating favorable fragment:large molecule interactions. The FragMaps may be used to qualitatively inform the design of small-molecule ligands or as scoring grids for high-throughput in silico docking that incorporates both an atomic-level description of solvation and the large molecule's flexibility.

Description

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH[0001]This invention was made with government support under Grant Numbers GM 51501, CA107331, CA120215 (ADM), and F32CA1197712 (OG) awarded by the National Institutes of Health. The government has certain rights in the invention.BACKGROUND OF THE INVENTION[0002]1. Field of the Invention[0003]The present invention is directed to a method of computational chemistry for identifying binding sites by molecular dynamics simulations using ligand competitive saturation. In particular, the method overcomes the problem of small nonpolar molecule aggregation to allow competitive saturation in an aqueous solution at physiological conditions. More particularly, the method, when used in a two-tier approach, may determine which one of several multiple fragment molecules has a highest probability of improving binding to a surface region of a large molecule that is proximate to a bound ligand, in order to produce an optimized lead compound for drug dis...

AI Technical Summary

Benefits of technology

[0013]Toward overcoming present costs and limitations of fragment-based computational drug design, the inventors describe a method that combines ideas from experimental fragment-based drug discovery, using NMR and crystallographic methods, with all-atom explicit-solvent MD simulations, using small aliphatic and aromatic molecules plus water molecules to map the affinity pattern for hydrophobic groups, aromatic groups, hydrogen bond donors, and hydrogen bond acceptors to a large molecule. The method, SILCS: Site Identification by Ligand Competitive Saturation, involves computationally “immersing” a large molecule, e.g., a protein, a glycoprotein, DNA, RNA, a carbohydrate, or a glycolipid, in an aqueous solution containing different types of small molecules having concentrations sufficient for competitive saturation, such that convergence of the simulation is anticipated. The large molecule+small molecules+water system is then subjected to multiple MD simulations allowing for competitive binding of the small molecules to the large molecule. Snapshots from the multiple MD simulation trajectories are combined to generate 3-dimensional (3D) probability maps, i.e., FragMaps, which reveal what molecular functionalities, e.g., aliphatic molecules, aromatic molecules, hydrogen bond donor molecules, and hydrogen bond acceptor molecules, bind most strongly to different parts of the large molecule's surface. Because the SILCS FragMaps are generated from MD simulations, they incorporate both the large molecule's flexibility with a Boltzmann distribution of conformations, as well as atomic-level solvation effects, resulting in 3D SILCS FragMaps that represent rigorous free energy distributions. Notably, the invention of the present method requires minimal time, labor, and materials compared to the conventional experimental approaches.

[0017]Additionally, an important part of the SILCS method of the invention is its computational feasibility. In an exemplary embodiment of the invention, each 5-ns SILCS simulation of an exemplary BCL-6 oncoprotein took less than three days on a single 2×4-core node of a commodity-computing cluster, and because each of the ten simulations was independent, they were all run simultaneously to yield converged FragMaps in under three days. The ability to achieve converged FragMaps probability maps in such a short time is a very important result, since MD simulations are often limited by the computational cost for simulations beyond the nanosecond regime, which in turn limits their utility in computer-aided drug discovery.

[0028]Yet another embodiment of the method of the invention further includes chemically linking one of the selected multiple fragment molecules, having a highest probability of improving binding to the surface region of the protein that is proximate to the bound ligand, to the bound ligand to design an optimized drug.

Problems solved by technology

This limit means that even the best fragments, having an LE value of 0.4-0.5 kcal*mol−1 per heavy atom, still have weak affinities for their target regions, making their screening by traditional assays difficult.

However, despite the utility of NMR spectroscopy and x-ray crystallography to detect fragment binding, there are significant time, labor, and materials costs associated with these two biophysical fragment-based drug discovery approaches.

As a result, these computational approaches are limited in their ability to accurately account for the exemplary protein molecule's conformational heterogeneity and solvation effects, contributions that are essential to compute free energies of binding of small molecules to the target region of the exemplary protein molecule (Guvench O, et al.

Nonetheless, approximations used in computational approaches to date still limit the accuracy of fragment placement and fragment scoring, which relates to affinity of the fragment for the binding site, and, ultimately, the determination of the most suitable fragment for a selected binding-site region of the protein molecule.

However, such MD simulated free-energy calculations are computationally expensive, limiting MD simulations from being used directly for high-throughput in silico screening.

Images