Structural and computational biology of COVID-19: Targeting SARS-CoV-2 proteins using virtual screening and fragment-based drug discovery.
The Blundell Group are using multifaceted approaches to identify potential drug targets and inhibitors against SARS-CoV-2. Drug candidates against various targets are necessary to combat the emergence of resistance.
Ali Alsulami has developed an extensively annotated SARS-CoV2 3D proteome database which assembles experimental structures of gene products and has models of the remainder including higher-order assemblies. The user-friendly web interface allows users to navigate, inspect and download proteome data. Hotspot mapping, druggability and binding-pocket prediction software allow identification of potential targets for structure-based drug discovery, and to understand impacts of mutations. Sherine Thomas has initiated experimental work on the viral endoribonuclease (Nsp15) involved in evading host viral RNA sensors, and on Nsp10 co-factor, a molecular switch in activation of multiple viral replicative enzymes. Sherine is testing potential drug candidates for fragment-based drug discovery workflow, developed earlier by Blundell and colleagues in Astex for cancer research and in our Department for mycobacterial infections.
Pedro Torres, now in Brazil, runs virtual screening on crystallographic structures of SARS-CoV-2 main protease using the Enamine REAL database comprising 1.2 billion compounds, selecting 104 compounds for purchase and in vitro testing. Sundeep Chaitanya exploits virtual screening on RNA-dependent RNA polymerase from the REAL database, selecting five top hits. Liviu Copoiu uses in-house protocols for filtering and selecting ligands based on substructure match and virtual docking for NSP14 and NSP12. Arian Jamasb has built complexes for known viral-human protein-protein interactions to identify druggable sites. Chris Beaudoin is using computational bioinformatics, such as protein-protein docking and homology modelling, to investigate profiles of Spike protein interactions with human cell surface proteins other than ACE2. Bridget Bannerman, with Jorge Julvez, Spain, carries out structural and dynamical analyses of integrated human/virus metabolic models in order to predict new treatment regimens against COVID-19. A new tool, findCPcli, predicts new drug targets selected from 15 essential reactions.
Publications
Alsulami A. et al. (2020). An extensively annotated SARS-CoV2-3D proteome database, that assembles experimental structures of gene products with models of the remainder including higher-order assemblies. Database and manuscript in preparation.