Towards more effective acetylcholinesterase inhibitors: a comprehensive modelling study based on human acetylcholinesterase protein–drug complex

Epik is a computer program for predicting pK(a) values for drug-like molecules. Epik can use this capability in combination with technology for tautomerization to adjust the protonation state of small drug-like molecules to automatically generate one or more of the most probable forms for use in further molecular modeling studies. Many medicinal chemicals can exchange protons with their environment, resulting in various ionization and tautomeric states, collectively known as protonation states. The protonation state of a drug can affect its solubility and membrane permeability. In modeling, the protonation state of a ligand will also affect which conformations are predicted for the molecule, as well as predictions for binding modes and ligand affinities based upon protein-ligand interactions. Despite the importance of the protonation state, many databases of candidate molecules used in drug development do not store reliable information on the most probable protonation states. Epik is sufficiently rapid and accurate to process large databases of drug-like molecules to provide this information. Several new technologies are employed. Extensions to the well-established Hammett and Taft approaches are used for pK(a) prediction, namely, mesomer standardization, charge cancellation, and charge spreading to make the predicted results reflect the nature of the molecule itself rather just for the particular Lewis structure used on input. In addition, a new iterative technology for generating, ranking and culling the generated protonation states is employed.

Ligand enrichment among top-ranking hits is a key metric of molecular docking. To avoid bias, decoys should resemble ligands physically, so that enrichment is not simply a separation of gross features, yet be chemically distinct from them, so that they are unlikely to be binders. We have assembled a directory of useful decoys (DUD), with 2950 ligands for 40 different targets. Every ligand has 36 decoy molecules that are physically similar but topologically distinct, leading to a database of 98,266 compounds. For most targets, enrichment was at least half a log better with uncorrected databases such as the MDDR than with DUD, evidence of bias in the former. These calculations also allowed 40x40 cross-docking, where the enrichments of each ligand set could be compared for all 40 targets, enabling a specificity metric for the docking screens. DUD is freely available online as a benchmarking set for docking at http://blaster.docking.org/dud/.