Evaluating Potential Broad-Spectrum Antiviral Activity using ColabFold
and Docking
Abstract
Broad-spectrum antivirals that work against many viruses provide an
immediate treatment for diseases caused by novel pathogenic viruses.
Notably, there is no universal drug against all four genera of the
coronaviridae family, in particular d-coronaviruses, which have recently
spilled over from pigs to humans. Here, we present and illustrate an
in-silico strategy to evaluate potential broad-spectrum activity of an
EUA-approved drug; viz., nirmatrelvir, for the porcine d-coronavirus
(PDCoV) that has infected humans. First, we show that the sequence-based
protein structure prediction method, ColabFold, can provide structures
for the M pro dimer of a-, b-, and g-coronaviruses
that are highly similar to the respective X-ray structures. Next, we
validated the performance of the docking software, AutoDock Vina 1.2.3
on ColabFold-predicted SARS-CoV-2 and MERS-CoV M pro
structures by showing that AutoDock Vina 1.2.3 can yield poses of
nirmatrelvir that are near the catalytic Cys, as seen in the respective
nirmatrelvir-bound X-ray structures. By using AutoDock Vina 1.2.3 to
dock nirmatrelvir to the ColabFold-predicted M pro
structure of PDCoV, we provide evidence that nirmatrelvir may inhibit
PDCoV M pro. These results show the feasibility of
using state-of-the-art sequence-based protein structure prediction and
docking methods to assess broad-spectrum antivirals for known viruses
against novel viruses lacking solved structures but sharing highly
similar conserved viral domains.