Drug-like molecules are often weak acid or bases with multiple accessible protonations and tautomeric states that can contribute significantly to the binding process. Therefore an accurate description of protonation states and tautomeric states are very important in binding affinity predictions, for example, free energy perturbation (FEP). When selecting the most relevant tautomers is preceded by scoring, a large set of reasonable tautomers needs to be generated first. The quantity of possible tautomers increases binomially with the number of protonation sites in the molecule and often includes non-trivial structures, and therefore algorithms for generating possible tautomers have been proposed in the literature. This work describes <b>a novel algorithm for tautomer generation</b>, which involves a combined use of molecular mechanics, semi-empirical quantum chemistry, and density functional theory, and which foregoes the use of hard-coded rules of proton distribution. We utilize the notion of the complete active space (CAS) of tautomers, which contains all possible tautomeric species for the identified protonation sites and the mobile protons in the molecule. Working with CASs allows us to generate tautomers in ever-growing systematic sets. Our construction of CASsis based on the predicted micro-pKa’s that are computed for every single atom in the molecule and its nearest protonated and deprotonated forms. Thus, our approach to tautomer generation is grounded in computational prediction of physical properties in contrast to other approaches that rely on user intuition and the previously observed tautomerization patterns from the known chemical space. This approach has been applied to FXa system in combination with FFP method to predict the binding affinity of 10 ligands with multiple protonation and tautomeric states, as shown in Figure 1.

Generation of complete active spaces of tautomers for drug discovery

2019 Spring National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.