Exploration of SARS-CoV‑2 Mpro Noncovalent Natural Inhibitors Using Structure-Based Approaches

With the emergence of antibody-evasive omicron subvariants
(BA.2.12.1,
BA.4, and BA.5), which can compromise the efficacy of vaccination,
it is of utmost importance to widen the finite therapeutic options
for COVID-19. Although more than 600 co-crystal complexes of Mpro
with inhibitors have been revealed, utilizing them to search for novel
Mpro inhibitors remains limited. Although there were two major groups
of Mpro inhibitors, covalent and noncovalent inhibitors, noncovalent
inhibitors were our main focus due to the safety concerns with their
covalent counterparts. Hence, this study aimed to explore Mpro noncovalent
inhibition ability of phytochemicals extracted from Vietnamese herbals
by combining multiple structure-based approaches. By closely inspecting
223 complexes of Mpro with noncovalent inhibitors, a 3D-pharmacophore
model representing typical chemical features of Mpro noncovalent inhibitors
was generated with good validation scores (sensitivity = 92.11%, specificity
= 90.42%, accuracy = 90.65%, and goodness-of-hit score = 0.61). Afterward,
the pharmacophore model was applied to explore the potential Mpro
inhibitors from our in-house Vietnamese phytochemical database, revealing
18 substances, 5 of which were in vitro assayed. The remaining 13 substances were then examined by induced-fit
molecular docking, revealing 12 suitable compounds. A machine-learning
activity prediction model was developed to rank the hit, suggesting
nigracin and calycosin-7-O-β-glucopyranoside
as promising Mpro natural noncovalent inhibitors.