In silico screening for natural ligands to non-structural nsp7 conformers of SARS coronaviruses

The non-structural protein 7 (nsp7) of Severe Acute Respiratory Syndrome (SARS) coronaviruses was selected as a new target to potentially interfere with viral replication. The nsp7s are uniquely conserved small coronavirus proteins having a critical, yet intriguing participation on the replication of the long viral RNA genome after complexing with nsp8 and nsp12. Drugs with potential to interfere with nsp7s have not been described yet. Despite the difficulties of having no previously defined binding pocket, high-throughput blind screening of more than one hundred thousand natural compounds < 400 Dalton of molecular weight docked against the nsp7.1ysy conformer identified hundreds of leads displaying predicted high binding-affinities by AutoDockVina. The leads were then docked to 14 nsp7 available conformers by two different binding scoring algorithms ( AutoDockVina-PyRx and HYDE-seeSAR), to identify consensus top-leads. Further predictive analysis of their physiological/toxicity ADMET criteria (chemical properties, adsorption, metabolism, toxicity) narrowed top-leads to a few drug-like ligands, most of them showing steroid-like structures closely related to some of those being actually used in clinical work. A final optimization by search for structural similarities to the drug-like top-lead, yielded a collection of novel steroid-like ligands with ~100-fold higher-affinity whose antiviral activity may be experimentally validated since they are available. Additionally, these nsp7-interacting ligands and/or their further optimized derivatives, may offer new tools to investigate the intriguing role of nsp7 on replication of coronaviruses