Computational Drug Design
- 1. Computational Drug DesignNoel M. O’BoyleApr 2010Postgrad course on Comp Chem
- 2. ReferencesComputational Drug Design, David C. YoungMy EBI lecture on protein-ligand docking, http://slideshare.net/baoilleachAn Introduction to Cheminformatics, AR Leach, VJ Gillet
- 3. Rational Drug DesignUse knowledge of protein or ligand structuresDoes not rely on trial-and-error or screeningComputer-aided drug design (CADD) now plays an important role in rational designStructure-based drug designUses protein structure directlyCADD: Protein-ligand dockingLigand-based drug designDerive information from ligand structuresProtein structure not always available40% of all prescription pharmaceuticals target GPCRsProtein structure has large degree of flexibilityStructure deforms to accommodate ligands or gross movements occur on bindingCADD: Pharmacophore approach, Quantitative structure-activity relationship (QSAR)
- 4. Computer-aided drug design (CADD)Known ligand(s)No known ligandStructure-based drug design (SBDD)Protein-ligand dockingDe novo designKnown protein structureLigand-based drug design (LBDD)1 or more ligands Similarity searchingSeveral ligands Pharmacophore searchingMany ligands (20+) Quantitative Structure-Activity Relationships (QSAR)CADD of no useNeed experimental data of some sortCan apply ADMET filtersUnknown protein structure
- 5. Virtual screeningVirtual screening is the computational or in silico analogue of biological screeningThe aim is to score, rank or filter a set of structures using one or more computational proceduresIt can be usedto help decide which compounds to screen (experimentally)which libraries to synthesisewhich compounds to purchase from an external companyto analyse the results of an experiment, such as a HTS run
- 6. Virtual screeningAR Leach, VJ Gillet, An Introduction to Cheminformatics
- 7. What is a Pharmacophore?Two somewhat distinct usages:
- 8. That substructure of a molecule that is responsible for its pharmacological activity (c.f. chromophore)
- 9. A set of geometrical constraints between specific functional groups that enable the molecule to have biological activityBojarski, Curr. Top. Med. Chem. 2006, 6, 2005.
- 10. Overview of Pharmacophore-based Drug DesignGenerate pharmacophoreActivity dataTest activitySearch compound library for activesBuy or synthesise ‘hits’See also John Van Drie’s http://pharmacophore.org
- 11. Pharmacophore generation and searchingProtein structure not requiredThere are also approaches that create pharmacophores from the active siteAssumes that all (or the majority) of the known actives bind to the same locationPharmacophore generationIdentify pharmacophoric features (hydrogen bond donors and acceptors, lipophilic groups, charges)Find a geometrical arrangement of pharmacophoric features that all actives that match with a low-energy conformationPharmacophore searchingGiven a pharmacophore, find all molecules in a database that can match it in a low-energy conformationSome pharmacophore software gives an estimate of activity, but most just give true or false for a matchScaffold-hopping possibleDoesn’t require structural similarityJust needs to match the pharmacophore
- 12. Protein-ligand dockingA Structure-Based Drug Design (SBDD) method“structure” means “using protein structure”Computational method that mimics the binding of a ligand to a proteinGiven...Predicts...
- 13. The pose of the molecule in the binding site
- 14. The binding affinity or a score representing the strength of bindingImage credit: CharakaGoonatilake, Glen Group, University of Cambridge. http://www-ucc.ch.cam.ac.uk/research/cg369-research.html
- 15. Protein-ligand docking IITypically, protein-ligand docking software consist of two main components which work together:1. Search algorithmGenerates a large number of poses of a molecule in the binding site2. Scoring functionCalculates a score or binding affinity for a particular poseThe difficulty with protein–ligand docking is in part due to the fact that it involves many degrees of freedomThe translation and rotation of one molecule relative to another involves six degrees of freedomThere are in addition the conformational degrees of freedom of both the ligand and the proteinThe solvent may also play a significant role in determining the protein–ligand geometryThe search algorithm generates poses, orientations of particular conformations of the molecule in the binding siteTries to cover the search space, if not exhaustively, then as extensively as possibleThere is a tradeoff between time and search space coverage
- 16. Ligand conformationsConformations are different three-dimensional structures of molecules that result from rotation about single bondsThat is, they have the same bond lengths and angles but different torsion anglesFor a molecule with N rotatable bonds, if each torsion angle is rotated in increments of θ degrees, number of conformations is (360º/ θ)NQuestionIf the torsion angles are incremented in steps of 30º, how many conformations does a molecule with 5 rotatable bonds have, compared to one with 4 rotatable bonds?Having too many rotatable bonds results in “combinatorial explosion”Also ring conformationsTaxolImage: IUPAC Gold BookImage: http://www.orgachemistryhelp.com/articles/organic-chemistry-synthesis.htmlLakdawalaet al.BMC Chemical Biology 2001 1:2
- 17. Types of search algorithmsClassified based on the degrees of freedom that they considerRigid dockingThe ligand is treated as a rigid structure during the dockingOnly the translational and rotational degrees of freedom are consideredTo deal with the problem of ligand conformations, a large number of conformations of each ligand are generated in advance and each is docked separatelyFlexible docking is more common todayConformations of each molecule are generated on-the-fly by the search algorithm during the docking processAvoids considering conformations that do not fitExhaustive (systematic) searching computationally too expensive as the search space is very largeOne common approach is to use stochastic search methodsThese don’t guarantee optimum solution, but good solution within reasonable length of timeStochastic means that they incorporate a degree of randomnessIncludes genetic algorithms (GOLD), simulated annealing (AutoDock)
- 18. Handling protein conformationsMost docking software treats the protein as rigidRigid Receptor ApproximationThis approximation may be invalid for a particular protein-ligand complex as...the protein may deform slightly to accommodate different ligands (ligand-induced fit)protein side chains in the active site may adopt different conformationsSome docking programs allow protein side-chain flexibility
- 19. For example, selected side chains are allowed to undergo torsional rotation around acyclic bonds
- 20. Increases the search space
- 21. Larger protein movements can only be handled by separate dockings to different protein conformationsImage: Cláudio M. Soares, Protein Modelling Laboratory, http://www.itqb.unl.pt/labs/protein-modelling/activities/psccip-pf
- 22. The perfect scoring function will…Accurately calculate the binding affinityWill allow actives to be identified in a virtual screenBe able to rank actives in terms of affinityScore the poses of an active higher than poses of an inactiveWill rank actives higher than inactives in a virtual screenScore the correct pose of the active higher than an incorrect pose of the activeWill allow the correct pose of the active to be identifiedBroadly speaking, scoring functions can be divided into the following classes:Forcefield-basedBased on terms from molecular mechanics forcefieldsGoldScore, DOCK, AutoDockEmpiricalParameterised against experimental binding affinitiesChemScore, PLP, Glide SP/XPKnowledge-based potentialsBased on statistical analysis of observed pairwise distributionsPMF, DrugScore, ASP
- 23. Böhm’s empirical scoring functionIn general, scoring functions assume that the free energy of binding can be written as a linear sum of terms to reflect the various contributions to binding
- 24. Bohm, J. Comput.-Aided Mol. Des., 1994, 8, 243
- 25. Bohm’sscoring function included contributions from hydrogen bonding, ionic interactions, lipophilic interactions and the loss of internal conformational freedom of the ligand.The hydrogen bonding and ionic terms are both dependent on the geometry of the interaction, with large deviations from ideal geometries (ideal distance R, ideal angle α) being penalised.The lipophilic term is proportional to the contact surface area (Alipo) between protein and ligand involving non-polar atoms.The conformational entropy term is the penalty associated with freezing internal rotations of the ligand. It is largely entropic in nature. Here the value is directly proportional to the number of rotatable bonds in the ligand (NROT).The ∆G values on the right of the equation are all constants determined using multiple linear regression on experimental binding data for 45 protein–ligand complexesHence “empirical”
- 26. Pose prediction accuracyGiven a set of actives with known crystal poses, can they be docked accurately?Accuracy measured by RMSD (root mean squared deviation) compared to known crystal structuresRMSD = square root of the average of (the difference between a particular coordinate in the crystal and that coordinate in the pose)2Within 2.0Å RMSD considered cut-off for accuracyMore sophisticated measures have been proposed, but are not widely adoptedIn general, the best docking software predicts the correct pose about 70% of the timeNote: it’s always easier to find the correct pose when docking back into the active’s own crystal structureMore difficult to cross-dock
- 27. Assess performance of a virtual screenNeed a dataset of Nact known actives, and inactivesDock all molecules, and rank each by scoreIdeally, all actives would be at the top of the listIn practice, we are interested in any improvement over what is expected by chanceDefine enrichment, E, as the number of actives found (Nfound) in the top X% of scores (typically 1% or 5%), compared to how many expected by chanceE = Nfound / (Nact * X/100)E > 1 implies “positive enrichment”, better than randomE < 1 implies “negative enrichment”, worse than randomWhy use a cut-off instead of looking at the mean rank of the actives?Typically, the researchers might test only have the resources to experimentally test the top 1% or 5% of compoundsMore sophisticated approaches have been developed (e.g. BEDROC) but enrichment is still widely used
- 28. Preparing the protein structureThe Protein Data Bank (PDB) is a repository of protein crystal structures, often in complexes with inhibitorsPDB structures often contain water moleculesIn general, all water molecules are removed except where it is known that they play an important role in coordinating to the ligandPDB structures are missing all hydrogen atomsMany docking programs require the protein to have explicit hydrogens. In general these can be added unambiguously, except in the case of acidic/basic side chainsAn incorrect assignment of protonation states in the active site will give poor results
- 29. Glutamate, Aspartate have COO- or COOH
- 30. OH is hydrogen bond donor, O- is not
- 31. Histidine is a base and its neutral form has two tautomersPreparing the protein structureFor particular protein side chains, the PDB structure can be incorrectCrystallography gives electron density, not molecular structureIn poorly resolved crystal structures of proteins, isoelectronic groups can give make it difficult to deduce the correct structureAffects asparagine, glutamine, histidineImportant? Affects hydrogen bonding patternMay need to flip amide or imidazoleHow to decide? Look at hydrogen bonding pattern in crystal structures containing ligands
Editor's Notes
- #13: 12^5/12^4 = 12