E-Novo: An Automated Workflow for Efficient Structure-Based Lead Optimization

Citations Over TimeTop 10% of 2009 papers

Abstract

An automated E-Novo protocol designed as a structure-based lead optimization tool was prepared through Pipeline Pilot with existing CHARMm components in Discovery Studio. A scaffold core having 3D binding coordinates of interest is generated from a ligand-bound protein structural model. Ligands of interest are generated from the scaffold using an R-group fragmentation/enumeration tool within E-Novo, with their cores aligned. The ligand side chains are conformationally sampled and are subjected to core-constrained protein docking, using a modified CHARMm-based CDOCKER method to generate top poses along with CDOCKER energies. In the final stage of E-Novo, a physics-based binding energy scoring function ranks the top ligand CDOCKER poses using a more accurate Molecular Mechanics-Generalized Born with Surface Area method. Correlation of the calculated ligand binding energies with experimental binding affinities were used to validate protocol performance. Inhibitors of Src tyrosine kinase, CDK2 kinase, beta-secretase, factor Xa, HIV protease, and thrombin were used to test the protocol using published ligand crystal structure data within reasonably defined binding sites. In-house Respiratory Syncytial Virus inhibitor data were used as a more challenging test set using a hand-built binding model. Least squares fits for all data sets suggested reasonable validation of the protocol within the context of observed ligand binding poses. The E-Novo protocol provides a convenient all-in-one structure-based design process for rapid assessment and scoring of lead optimization libraries.

Related Papers

• → Advances and Challenges in Protein-Ligand Docking(2010)604 cited
• → Protein-ligand docking(2011)22 cited
• → CSAlign and CSAlign-Dock: Structure alignment of ligands considering full flexibility and application to protein–ligand docking(2022)14 cited
• → Role of Docking in Anticancer Drug Discovery(2022)8 cited
• → Protein–Ligand Docking in Drug Design(2001)8 cited