Subtle structural differences of nucleotide analogs may impact SARS-CoV-2 RNA-dependent RNA polymerase and exoribonuclease activity

Fig. 1. Relevant positions for binding in the RdRp – exonuclease system. The catalytic and i + 1 positions are shown in panel A a 90 rotation in the z-axis is shown in panel B. Images underneath show each of the binding poses obtained for the reference nucleotides at the three relevant positions of the protein complex.

Fig. 2. Substituent analysis. The chemical structure of ATP is shown above, highlighting in red the common scaffold for the molecules in the database and the substitution sites, and with the nitrogenous base shown in blue (note that adenine is a purine, i.e., it has two connected rings). The box plots depict the distribution in the different consensus scores. Consensus scores include the three docked sites (RdRp-cat, RdRp-p4 and ExoN-cat), and the score when using a certain complimentary base (uracil, cytosine, guanine, or adenine). Different subgroups are compared in the boxplots: compounds with one vs more rings in the nitrogenous base (i.e., similar to pyrimidine or purine, respectively), and substitutions in the ribose carbons other than the canonical (being the canonical form the one shown for ATP above). To ensure a better graphical representation, all variables were standardized by subtracting the mean and dividing by the standard deviation. For all variables, a lower value implies a better fit. In white, molecules with the canonical substitution; in red, molecules with other substitutions; in cyan, molecules with a cyano group in C1′ or a fluorine in C2′. The scores of reference…

Fig. 3. Pairwise plots of the consensus scores at the three docking sites. RDV-MP is marked in red, and the other references in orange. Note that RDV-MP has excellent performance in all docking sites. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4. Binding modes of selected nucleotide analogs. Upper panel: the top seven selected molecules, identified by its PubChem CID. Lower panel: details of binding mode for (A) analogue 68,658,112 (cyan sticks) and (B) analogue 24,763,257 (olive sticks) into RdRp-Cat system (grey cartoon and white sticks); (C) analogue 70,587,122 (green sticks) into RdRp-p4 system (grey cartoon and grey sticks); and (D) analogue 70,907,629 (orange sticks) into the ExoN-Cat system (grey cartoon and grey sticks). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5. Factorial design for the structural equations model used for estimating site consensus variables. This is a compact depiction presenting only 11 variables for one nitrogenous base system, but each of the four nitrogenous bases in the template RNA would have an identical structure (hence the dashed arrows). The rectangles represent observed variables, while circles are latent variables. Double-headed arrows represent admitted correlations; single-headed arrows represent factorial loadings. Cat: RdRp-cat; i + 4: RdRp-p4; ExoN: ExoN-cat; NB: nitrogenous base; ura: uracil; cyt: cytosine; gua: guanine; ade: adenine. D. score: docking score, APF: atomic property field, HB: hydrogen bonds score.