Electrostatic and shape-based methods have become a standard part of the drug design workflow, used in hit-finding, hit to lead and lead optimization strategies. The methods build on earlier pharmacophore approaches to offer a detailed description of the way ligands behave in a biological context. The success of these methods is evident in their widespread adoption, but they are not magic bullets and failures do occur. Understanding their limitations is vital if we are to improve them in the future. The initial application of electrostatic and shape methods was in virtual screening. Methods such as ROCS, Blaze and Phase Shape all have numerous publications detailing their successful application. We present one such success using Blaze, but also some of the cases where it has been unsuccessful. Analysis of these provides some important lessons for virtual screening best practice, as well as some insights into when it is unlikely to work. As the methods matured, their utility widened to include hit-to-lead and lead optimization. One of the major applications is scaffold-hopping. A known active can be transformed into new chemical matter by finding chemical transformations which preserve its essential shape and electrostatic character. The outcome of these computational processes are all new molecules to be made. As such the synthetic accessibility, the chemical stability, the relative conformation energies, and the protonation state of the suggestions all need to be considered. We examine all of these in the context of successful and unsuccessful scaffold hops that we have tried. One of the first applications of electrostatics and shape was CoMFA, originally published in 1988. The field of 3D-QSAR has gone in and out of fashion since then: it has great intuitive appeal, but the effort required to get a consistent alignment of the training set made it hard to use. Modern methods have largely alleviated this difficulty, but 3D-QSAR still divides opinion as to its utility. We share some of our experiences in developing and supporting a 3D-QSAR method, with insights gained from our large user base. All of these approaches look at the drug discovery problem with a focus purely on the ligand. With the increasing availability of X-ray and cryo-EM structures, do they have a future? The electrostatics of the ligand is only half the story, so what can we gain by considering the electrostatic characteristic of the protein active site as well as its shape?

Finding new chemical matter using electrostatic and shape-based approaches: Successes and failures

2019 Spring National Meeting

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