Skip to main content
Science Areas
Functional and Systems Biology
Computing, Analytics, and Modeling

Screening Existing Compound Libraries for Potential Coronavirus Therapeutics Identifies Several Compounds with Antiviral Activity

High-throughput biochemical assays targeting a vital viral protein identified one molecule out of more than 13,000 with promising antiviral activity against SARS-CoV-2. 

multi-colored illustration of coronavirus
High-throughput screening methodologies to search molecular libraries for antivirals active against SARS-CoV-2 may also be useful to develop potent antivirals in future pandemics. (Illustration by Timothy Holland | Pacific Northwest National Laboratory)

The Science  

Vaccines for the novel coronavirus are becoming available around the world, but it will take time for a vaccination strategy to slow the spread of the virus. Therefore, medicines to treat COVID-19 are still needed. To speed the development of new antivirals, a multi-institutional team of scientists screened more than 13,000 compounds from existing drug libraries for the ability to inhibit a nonstructural protein of SARS-CoV-2 called nsp15. Three hits were confirmed as potent nsp15 inhibitors in vitro. Native mass spectrometry and molecular docking simulations confirmed that one of those hits bound to nsp15. Cell-based assays confirmed the one hit possessed modest antiviral activity against SARS-CoV-2.  

The Impact  

The coronavirus nonstructural protein nsp15 is highly conserved among coronaviruses. It is also a key component for viral replication with no corresponding counterpart in host cells. These factors make it an intriguing candidate for drug development using two common approaches: high-throughput screening and structure-guided drug discovery. The screening methodologies described in this work to confirm, characterize, and validate molecules that inhibit a SARS-CoV-2 protein could also be useful to rapidly develop potent antivirals in future pandemics. 


A multi-institutional team of scientists, including researchers from the University of Washington School of Medicine and the U.S. Department of Energy’s (DOE) Pacific Northwest National Laboratory (PNNL), screened more than 13,000 compounds in existing libraries of drug and lead repurposing compound libraries for activity against nsp15 from SARS-CoV-2, the novel coronavirus that causes COVID-19. Their assays identified three hits as inhibiting nsp15 activity in vitro.  

Using native mass spectrometry capabilities at EMSL, the Environmental Molecular Sciences Laboratory, the team confirmed that one of those hits bound to nsp15. This candidate is a molecule called Exebryl-1, a ß-amyloid anti-aggregation molecule, and was designed for Alzheimer’s disease therapy. Exebryl-1 did not have sufficient antiviral activity in cell-based assays for immediate drug repurposing efforts. However, artificial intelligence-based lead optimization using the Exebryl-1 scaffold together with in silico molecular docking calculations onto the crystal structure of nsp15 are being used to optimize the medicinal chemistry of this compound to improve its antiviral properties. 


Mowei Zhou, EMSL, 

Wesley Van Voorhis, University of Washington School of Medicine, 


This project has been funded in whole or in part from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services, and by the DOE Office of Science through the National Virtual Biotechnology Laboratory, a consortium of DOE national laboratories focused on response to COVID-19, with the later funding provided by the Coronavirus CARES Act. Part of the research was conducted at the EMSL, a national scientific user facility sponsored by Office of Science, Biological and Environmental Research program located at PNNL.  


R. Choi, et al., “High-throughput screening of the ReFRAME, Pandemic Box, and COVID Box drug repurposing libraries against SARS-CoV2 nsp15 endoribonuclease to identify small-molecule inhibitors of viral activity,” PLoS One 16, e0250019 (2021). [DOI: 10.1371/journal.pone.0250019] 

Related Link

From Molecule to Medicine via Machine Learning