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.

1 . A method of computational chemistry for identifying binding sites, by molecular dynamics (MD) simulations using ligand competitive saturation, said method comprising:
 inputting a 3-dimensional (3D) structure of a protein;   preparing multiple solutions for ligand competitive saturation, each of said multiple solutions corresponding to a volumetric grid in which either a benzene, a propane, or a water molecule is randomly placed on a grid point, such that, a spacing of said grid points and said random placing of said benzene, propane, or water molecules corresponds to a concentration sufficient for competitive saturation of said benzene and propane molecules;   overlaying said 3D structure of said protein with each of said multiple solutions and removing all benzene, propane and water molecules that overlay said protein in said volumetric grid;   introducing a repulsive interaction energy between said benzene and said propane molecules to prevent non-polar aggregation;   minimizing energies and equilibrating each of said multiple solutions with said protein by MD simulations to produce multiple equilibrated systems;   propagating MD simulations for each of said multiple equilibrated systems;   binning atoms from each of said propagated MD simulations into voxels,
 wherein benzene and propane carbon atoms are binned as aromatic and aliphatic carbons, respectively, and 
 wherein water oxygen and hydrogen atoms are binned as hydrogen bond acceptors and hydrogen bond donors, respectively; 
   using bin counts for each of said aromatic and aliphatic carbons, and said hydrogen bond acceptors and hydrogen bond donors, based on their free energies of binding, from said propagated MD simulations to identify surface regions of said protein having a high probability to bind one of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom; and   using said identified surface regions of said protein as docking grids for high-throughput in silico docking of drug-like compounds.   
     
     
         2 . The method of  claim 1 , further comprising, assigning an orientation of each of His, Gln, and Asn sidechains to said 3D structure of said protein. 
     
     
         3 . The method of  claim 1 , further comprising aligning said protein from a trajectory of said propagated MD simulations. 
     
     
         4 . The method of  claim 1 , further comprising:
 introducing harmonic positional restraints to all atoms of said protein before equilibrating said multiple systems; and   subsequently replacing said harmonic positional restraints with weak restraints on protein backbone carbon atoms to prevent denaturation before propagating said MD simulations.   
     
     
         5 . The method of  claim 1 , further comprising creating fragment binding maps (FragMaps) of said surface regions of said protein, each of said FragMaps corresponding to said high probability of binding one of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom, wherein said FragMap represents a single isocontour value, empirically chosen, to optimize visualization of each of said FragMaps. 
     
     
         6 . The method of  claim 1 , wherein said protein is a glycoprotein. 
     
     
         7 . The method of  claim 1 , wherein said concentration sufficient for competitive saturation of said benzene and propane molecules is 1 M benzene and 1 M propane. 
     
     
         8 . The method of  claim 1 , wherein said concentration sufficient for competitive saturation of said benzene and propane molecules is greater than 0.01 mM and less than 10 M benzene and greater than 0.01 mM and less than 10 M propane. 
     
     
         9 . A method of computational chemistry for identifying binding sites, by molecular dynamics (MD) simulations using ligand competitive saturation, said method comprising:
 inputting a 3-dimensional (3D) structure of a large molecule,
 wherein said large molecule comprises one of DNA, RNA, a carbohydrate and a glycolipid; 
   preparing multiple solutions for ligand competitive saturation, each of said multiple solutions corresponding to a volumetric grid in which either a small aromatic molecule, a small aliphatic molecule, or a water molecule is randomly placed on a grid point, such that, a spacing of said grid points and said random placing of small aromatic molecules, small aliphatic molecules, or water molecules corresponds to a concentration sufficient for competitive saturation of said small aromatic molecules and said small aliphatic molecules;   overlaying said 3D structure of said large molecule with each of said multiple solutions and removing all said small aromatic molecules, said small aliphatic molecules, and said water molecules that overlay said large molecule in said volumetric grid;   introducing a repulsive interaction energy between said small aromatic molecules and said small aliphatic molecules to prevent non-polar aggregation;   minimizing energies and equilibrating each of said multiple solutions with said large molecule by MD simulations to produce multiple equilibrated systems;   propagating MD simulations for each of said multiple equilibrated systems;   binning atoms from each of said propagated MD simulations into voxels,
 wherein carbon atoms of said small aromatic molecules and said small aliphatic molecules are binned as aromatic and aliphatic carbons, respectively, and 
 wherein water oxygen and hydrogen atoms are binned as hydrogen bond acceptors and hydrogen bond donors, respectively; 
   using bin counts for each of said aromatic and aliphatic carbons, and said hydrogen bond acceptors and hydrogen bond donors and their free energies of binding from said propagated MD simulations to identify surface regions of said large molecule having a high probability to bind one of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom; and   using said identified surface regions of said large molecule as docking grids for high-throughput in silico docking of drug-like compounds.   
     
     
         10 . A method of computational chemistry for identifying binding sites, by molecular dynamics (MD) simulations using ligand competitive saturation, for drug design optimization, said method comprising:
 inputting a 3-dimensional (3D) structure of a complex of a protein and a bound ligand;   preparing multiple solutions for ligand competitive saturation, each of said multiple solutions corresponding to a volumetric grid in which either a benzene, a propane, or a water molecule is randomly placed on a grid point, such that, a spacing of said grid points and said random placing of said benzene, propane, or water molecules corresponds to a concentration sufficient for competitive saturation of said benzene and propane molecules;   overlaying said 3D structure of said protein and said bound ligand with each of said multiple solutions and removing all benzene, propane and water molecules that overlay said protein and said bound ligand in said volumetric grid;   introducing a repulsive interaction energy between said benzene and said propane molecules to prevent non-polar aggregation;   minimizing energies and equilibrating each of said multiple solutions with said protein and said bound ligand by MD simulations to produce multiple equilibrated systems;   propagating a first tier of MD simulations for each of said multiple equilibrated systems;   binning atoms from each of said propagated MD simulations into voxels,
 wherein benzene and propane carbon atoms are binned as aromatic and aliphatic carbons, respectively, and 
 wherein water oxygen and hydrogen atoms are binned as hydrogen bond acceptors and hydrogen bond donors, respectively; 
   using bin counts for each of said aromatic and aliphatic carbons, and said hydrogen bond acceptors and hydrogen bond donors, based on their free energies of binding, from said first tier of propagated MD simulations to identify surface regions of said protein having a high probability to bind one of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom;   identifying a surface region of said protein that is proximate to said bound ligand and has a high probability of binding to at least two of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom;   selecting multiple fragment molecules, containing said at least two of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom, for a second tier of MD simulations;   preparing multiple solutions for ligand competitive saturation of said selected multiple fragment molecules;   introducing an intermolecular repulsive interaction energy between nonpolar regions of said selected multiple fragment molecules to prevent non-polar aggregation;   minimizing energies and equilibrating each of said multiple solutions with said protein and said bound ligand by MD simulations to produce multiple equilibrated systems;   propagating said second tier of MD simulations for each of said multiple equilibrated systems; and   analyzing said second tier of MD simulations to determine which one of said selected multiple fragment molecules has a highest probability of improving binding to said surface region of said protein that is proximate to said bound ligand.   
     
     
         11 . The method of  claim 10 , further comprising, chemically linking said one of said selected multiple fragment molecules, having a highest probability of improving binding to said surface region of said protein that is proximate to said bound ligand, to said bound ligand to design an optimized drug. 
     
     
         12 . The method of  claim 10 , further comprising, assigning an orientation of each of His, Gln, and Asn sidechains to said 3D structure of said protein. 
     
     
         13 . The method of  claim 10 , further comprising, aligning said protein from a trajectory of said first and second tier propagated MD simulations. 
     
     
         14 . The method of  claim 10 , further comprising:
 introducing harmonic positional restraints to all atoms of said protein before equilibrating said multiple systems in said first and second tier MD simulations; and   subsequently replacing said harmonic positional restraints with weak restraints on protein backbone carbon atoms to prevent denaturation before propagating said first and second tier MD simulations.   
     
     
         15 . The method of  claim 10 , wherein said multiple fragment molecules comprise any of methanol, phenol, imidazole, aniline, pyridine, pyrrole, and acetone. 
     
     
         16 . The method of  claim 10 , wherein said protein is a glycoprotein. 
     
     
         17 . The method of  claim 10 , wherein said concentration sufficient for competitive saturation of said multiple fragment molecules is 1 M. 
     
     
         18 . The method of  claim 11 , wherein said concentration sufficient for competitive saturation of said fragment molecules is greater than 0.01 mM and less than 10 M. 
     
     
         19 . The method of  claim 10 , wherein analysis of said second tier of MD simulations is based on binning atoms of said selected multiple fragment molecules from each of said propagated MD simulations into voxels for each of said multiple equilibrated systems. 
     
     
         20 . A method of computational chemistry for identifying binding sites, by molecular dynamics (MD) simulations using ligand competitive saturation, for drug design optimization, said method comprising:
 inputting a 3-dimensional (3D) structure of a complex of a large molecule and a bound ligand,
 wherein said large molecule comprises one of DNA, RNA, a carbohydrate and a glycolipid; 
   preparing multiple solutions for ligand competitive saturation, each of said multiple solutions corresponding to a volumetric grid in which either a benzene, a propane, or a water molecule is randomly placed on a grid point, such that, a spacing of said grid points and said random placing of said benzene, propane, or water molecules corresponds to a concentration sufficient for competitive saturation of said benzene and propane molecules;   overlaying said 3D structure of said large molecule and said bound ligand with each of said multiple solutions and removing all benzene, propane and water molecules that overlay said large molecule and said bound ligand in said volumetric grid;   introducing a repulsive interaction energy between said benzene and said propane molecules to prevent non-polar aggregation;   minimizing energies and equilibrating each of said multiple solutions with said large molecule and said bound ligand by MD simulations to produce multiple equilibrated systems;   propagating a first tier of MD simulations for each of said multiple equilibrated systems;   binning atoms from each of said propagated MD simulations into voxels,
 wherein benzene and propane carbon atoms are binned as aromatic and aliphatic carbons, respectively, and 
 wherein water oxygen and hydrogen atoms are binned as hydrogen bond acceptors and hydrogen bond donors, respectively; 
   using bin counts for each of said aromatic and aliphatic carbons, and said hydrogen bond acceptors and hydrogen bond donors and their free energies of binding from said first tier of propagated MD simulations to identify surface regions of said large molecule having a high probability to bind one of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom;   identifying a surface region of said large molecule that is proximate to said bound ligand and has a high probability of binding to at least two of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom;   selecting multiple fragment molecules, containing said at least two of an aromatic carbon atom, an aliphatic carbon atom, a hydrogen bond acceptor oxygen atom, and a hydrogen bond donor hydrogen atom, for a second tier of MD simulations;   preparing multiple solutions for ligand competitive saturation of said selected multiple fragment molecules;   introducing an intermolecular repulsive interaction energy between nonpolar regions of said selected multiple fragment molecules to prevent non-polar aggregation;   minimizing energies and equilibrating each of said multiple solutions with said large molecule and said bound ligand by MD simulations to produce multiple equilibrated systems;   propagating said second tier of MD simulations for each of said multiple equilibrated systems; and   analyzing said second tier of MD simulations to determine which one of said selected multiple fragment molecules has a highest probability of improving binding to said surface region of said large molecule that is proximate to said bound ligand.