Module 2: Structure Based Ph4 Design MOE provides several applications to analyze protein information in absence of ligands: • The Site Finder (using the binding site of a receptor to generate the query) • Contact preferences In part b) of the SBDD course: • The MCSS algorithm for de-novo design • Docking Alpha Site Finder Objective The Site Finder screens the surface of a protein for potential binding sites. In addition to locating cavities it also indicates preferred locations for hydrophilic or non-hydrophilic interaction points. Methodology Site Finder considers the relative positions and accessibility of the receptor atoms along with a rough classification of chemical type. The method applies alpha spheres. This is a special case of a contact sphere that circumscribes 4 atoms on its boundary and contain no internal atoms. Centers of alpha spheres are clustered into hydrophobic and hydrophilic areas. Binding Site Identification Geometric analysis of a molecule • Alpha spheres identify cavities • Sphere size is related to degree of exposure • Small spheres indicate "tight" cavities • The sphere size is not reflected in the graphical representation of centroids • Centroids are clustered for display • Hydrophilic contact points are marked by red centroids • Hydrophobic (defined as non-hydrophilic) are colored white. P33 Protein Kinase Exercise: Query generation and Search from Receptor Site - Binding Site Identification I Use alpha spheres to predict ligand positions 1. (MOE | File | Open) for the file 1ke6.moe 2. Open (MOE | Compute | Site Finder) 3. Click “Apply” on the Site Finder panel 17 sites are found … Site Finder Panel Atoms included in the calculation Display mode of alpha spheres Display mode of site(s); Residue selection option Creates dummy atoms Minimum sphere radii to detect (non-) LP-active atoms Settings for alpha sphere clustering Site list: Site: Site number Size: Number of contributing spheres Hyd: Number of hydrophobic atoms contacted in receptor Side: Number of sidechain contact atoms Residues: Residues at local surface Distance filter before clustering Exercise: Query generation and Search from Receptor Site - Binding Site Identification II 4. Examine the different sites. The 2nd site has more receptor contacts but fewer hydrophobic contacts. 5. Select the first site. To restrict the view to the immediate environment by selecting Isolate: “Atoms” and enable SE Residues. Ensure that (MOE | Selection | Synchronize) is ON. Invert the selection (MOE | Selection | Invert) and delete all hidden residues. 6. Keep the positions of the Alpha Centers by pressing “Dummies” and close the panel. 7. To increase the size of the dummy atoms <Ctrl>-click on one of the dummy atoms to select all of them. Select (Render | Space Filling). Red: Potential hydrophilic contact areas Exercise: Query generation and Search from Receptor Site - Binding Site Identification III 8. Create a surface for the pocket using (MOE | Compute | Surfaces and Maps). 9. In the panel, keep default settings but select Near: Dummy Atoms Click: Apply The colours of the surface displays those regions of the receptor surface suitable for a hydrogen bond or metal-lone-pair interactions. Try also different surface color schemes to compare the results or modify the transparency (TF, TB) settings. 10. Save the pocket as 1ke6_pocket.moe Receptor Contact Preferences Objective This approach complements the site finder information by identifying preferred areas for hydrophobic or hydrophilic interactions based on statistical preferences derived from non-bonding contacts in high resolution protein structures. Methodology Non-bonded protein-ligand interactions are analyzed with respect to distance, angle and out-of-plane preferences. The receptor and ligand atoms are classified into atom types and the experimental histograms of the contacts are fitted by analytical functions. Contour maps display likelihood ratios for hydrophobic over hydrophilic preferences (green) or vice versa (red). Receptor Atom Classification • For each receptor atom, A, define a coordinate system – Define vectors u and v derived from hybridization and heavy neighbors – Some atom types do not have a u or a v (taken to be zero) u A v v u A u A v taken from pi system • Define polar coordinate system from u and v – Distance from atom A r “distance” – Angle with u vectora “lone pair angle” – Angle with u in u-v plane p “out-of-plane angle” v u p a • A contact atom, B, is mapped to (r,a,p) local coordinates A r B Receptor Atom Classification - Atom Typing 0,16 T_nQ2: HYD (50%) LPA (50%) r 0,12 0,08 0,04 0 1,5 1,74 1,98 2,22 2,46 2,7 2,94 3,18 3,42 3,66 3,9 4,14 4,38 0,32 a 0,24 0,16 0,08 r a p HYD lognorm gamma gamma LPA 12-6 cauchy cauchy 0 2,5 22,5 42,5 62,5 82,5 102,5 122,5 142,5 162,5 N 0,32 NH p 0,24 0,16 O HN N 0,08 0 2,5 17,5 32,5 47,5 62,5 77,5 NH Exercise: Query generation and Search from Receptor Site - Contact Preferences Contact preferences can be used to generate or refine a Ph4 query. 1. Hide the molecular surface from the previous exercise in (MOE | Window | Graphic Objects) to concentrate on contact potentials. 2. Select Surface: “Contact Preference” in the Surface and Maps panel and press Apply. 3. Play with different contour levels and display styles. Exercise: Query generation and Search from Receptor Site - Pharmacophore Query Editor I The suggestions for preferred interactions in the binding site may be used to derive a Ph4 query in absence of any known ligands. 1. 2. 3. 4. Open (MOE | File | New | Pharmacophore Query), select an Ph4 scheme, e.g. PPCH_all. Features are created in clusters of hydrophilic or hydrophobic groups. Select individual dummy atoms to place the new feature. In the Query Editor, press Feature. A generic ‘Any’ feature will be created. Adjust the feature positions by using <Shift><Alt><middle mouse button>. Furthermore modify and reassign the feature types adjusting the radii, expressions, etc. according to chemical intuition. Exercise: Query generation and Search from Receptor Site - Pharmacophore Query Editor II 5. Once finished creating features, an excluded volume may be added.*) Use a surface representation to guide the adjustment of tolerances for the excluded volume. Exercise: Query generation and Search from Receptor Site - Pharmacophore Search Save and use this query in the same way as the ones generated in the previous exercises. 6. Save the query as PPCH_ALL_sitefindV.ph4 7. Press Search and define the databases to be searched, consecutively double clicking on them, or select and then press Add.*) 8. Once the list is complete, press OK 9. The Search panel will reflect multiple databases. Continue with search process as previous exercises. 10. Note that the databases should be pre-annotated with the ph4 scheme Case Studies in Pharmacophore Search Module 3: Structure-based design with known actives and structural binding site details Case studies Small molecules Yes no No Protein structure Module 1 Yes Module 3 Module 2 Module 3: Structure Based Ph4 Design If structural information about both proteins and their ligands is available, the essential (conserved) protein ligand interactions (e.g. H-bonds) can be identified and used to focus on the key Ph4 features. Since those interactions include “projected” protein interaction sites, the results should be more meaningful than a small molecule alignment in itself.*) MOE provides several applications to analyze protein-ligand information: • Alignment and superposition of proteins • Ph4 consensus analysis • Surface properties and contact preferences In the Protein course: • Homology modeling (if only the amino acid sequence of a protein is available) Workflow of Structure Based Ph4 Design 1. Align a set of proteins with their co-crystallized ligands. • Ligands should already be docked in the binding site 2. Generate the relevant pharmacophore features based on SAR • Identify conserved Ph4 features with a Ph4 consensus analysis • Define the features based on biological/chemical knowledge of interactions • Identify key features based on interactions using Contact Statistics, Molecular Surfaces or Hydrogen Bonding interactions 3. Save pharmacophore model • Refine using Excluded volumes, or Interior Volumes for SMILES strings 4. Search conformational database(s) for ligands that contain the relevant pharmacophores 5. Re-design the pharmacophore model, if necessary and search again