Molecular Dynamics Simulations Indicate the SARS-CoV-2 Mpro Is Not a Viable Target for Small-Molecule Inhibitors Design

Maria Bzówka; Karolina Mitusinska; Agata Raczynska; Aleksandra Samol; Jack Tuszynski; Artur Góra

The comparison of MD simulations of both main proteases initiated from different starting conformations (with and without N3 inhibitor) suggests that beside plasticity of the whole protein, there can be large differences between the accessibility to the binding cavity and/or the accommodation of the shape of the cavity in response to the inhibitor that can be bound. There are also differences in the outer pockets' maximal accessible volumes between the two structures of SARS-CoV main proteases; the apo SARS-CoV Mpro structure used as a starting point of MD simulations has shown the largest MAV of all the analysed systems. These results suggest that the SARS-CoV main proteases' binding cavity is highly flexible and changes both in volume and shape significantly alter the ligand binding. This finding indicates a serious obstacle for a classical virtual screening approach and drug design in general. Numerous novel compounds that are considered as potential inhibitors of SARS-CoV, have not reached the stage of clinical trials. The lack of success might be related to the above-mentioned plasticity of the binding cavity. Some of these compounds have been used for docking and virtual screening research aimed not only at SARS-CoV [42, 43] but also at the novel SARS-CoV-2 [15] [16] [17] [18] [19] [20] [21] . Such an approach focuses mostly on the structural similarity between the binding pockets, but ignores the fact that the actual available binding space differs significantly. In general, a rational drug design can be a very successful tool in the identification of possible inhibitors in cases where the atomic resolution structure of the target protein is known. For a new target, when a highly homologous structure is available, a very logical strategy would be seeking chemically similar compounds or creating derivatives of this inhibitor, and finding those compounds that are predicted to have a higher affinity for the new target structure than the original one. This would be expected to work for SARS-CoV-2 proteins (such as Mpro) using SARS-CoV proteins as templates. However, our in-depth analysis indicates a very different situation taking place, with major shape and size differences emerging due to the binding site flexibility. The continuous effort of Diamond Light Source group [44] performing massive XChem crystallographic fragment screen against Mpro supports our finding on large binding site flexibility (Supplementary Figure S4) . Therefore, repurposing SARS drugs against COVID-19 may not be successful due to major shape and size differences, and despite docking methods, the enhanced sampling should be considered.