Design of ligand binding pocket of RdRp:
The nsp12 domain is an RNA dependent RNA polymerase which plays pivotal role in catalyzing the RNA replication of coronavirus. The overall architecture of SARS-CoV-2 nsp12-nsp8-nsp7 complex is very similar to that of SARS-CoV with a Root Mean Square Deviation (RMSD) value of 0.82. Analysis of sequence conservation across the coronavirus family reveals that the template entry, template-primer exit, and NTP tunnels, as well as the polymerase-active site, are the most highly conserved surfaces on nsp121415. The superimposed images of RdRp structures 7BTF (SARS-CoV-2) and 6NRU (SARS-CoV) with conserved amino acid residues are shown in Figure 1. In the current study we targeted the NTP entry pocket and metal ion stabilizing triad and NTP binding and positioning residues as a possible site for ligand binding there by disrupting the RNA synthesis mechanism of RdRp. The grid box for docking was assigned to these conserved residues for both SARS-CoV and SARS-CoV-2.
Non-nucleotide analogs as RdRp inhibitors:
The classical drug candidates for RdRps are NAs and many such molecules are under various stages of development. One of the main hurdles in using NAs is the presence of proofreading and exonuclease activity for both SARS-CoV and SARS-CoV-2 in its nsp14 domain (ExoN) 16. There are also two reported mutations in sub-domain of the nsp12 RdRp (F480L and V557L SARS-CoV numbering). Neither of the resistance mutations directly impact the catalytic site nor substrate-binding pocket, but rather cause minor structural alterations which likely impact an NTP ‘checking step’ performed by the polymerase before catalysis 17. These factors increase the fidelity of RNA synthesis by SARS-CoV and thus pose a threat to the use of NAs as possible inhibitors for RdRps. In this study we explored the potential of phytochemicals from common medicinal herbs as non-nucleotide analog inhibitors for both SARS-CoV and SARS-CoV-2 RdRp. Sixty nine active constituents were selected from herbal extracts like Withania somnifera root extract, Hyssopus officinalis, Glycyrrhiza glabra root extract, Ocimum sanctum aerial part extract, Camellia sinensis leaf extract, Prunella vulgaris, Hedychium spicatum rhizome extract, Cyperus scariosus, Nigella sativa seed extract and Triphala fruit and fruit rind extract based on literature.
Molecular docking of phytochemical compounds:
AutoDock Vina results represent the docking scores as Gibbs free energy of binding (ΔG (kcal/mol)) which approximates the sum of all interactions ligand/receptor minus desolvation energies. The docking scores of phytochemicals against both SARS-CoV-2 and SARS-CoV are given in Table 1. The control drug molecule remdesivir was shown to have docking scores of -7.2 kcal/mol and -7.5 kcal/mol against RdRp of SARS-CoV-2 and SARS-CoV respectively. Our results showed that thirty seven phytochemicals had ΔG values less than the cut-off value -6 kcal/mol and hence they were considered to be potential inhibitors 18. The dissociation constant for protein ligand binding (Ki) were calculated for thirty-seven potent phytochemical compounds and remdesivir and the data is given in Table 2 in their order of decreasing potency. Many of the phytochemicals in our study showed lower Ki values than remdesivir, with the Ki values for later being 5.048 µM and 3.037 µM for SARS-CoV-2 and SARS-CoV respectively. As discussed earlier, lower the Ki value better the inhibitor. The predicted Ki values calculated here gives an approximate estimation of the potency of inhibitors and caution should be exercised in extrapolating this data because the experimental Ki values can be different from the predicted values depending on various biophysical parameters involved in enzyme inhibition 19.
In our study, lowest docking scores for both SARS-CoV-2 (-9 to -7.7 kcal/mol) and SARS-CoV (-9.3 to -7.8 kcal/mol) were obtained for the phytochemicals from Withania somnifera root extract such as withaferin A, withanolide A, B, D,E and F, withasomniferol A, B and C and withanone. These phytochemicals also exhibited the least of the Ki values among all the phytochemical compounds tested. Withania somnifera, commonly known as Ashwagandha, is well established traditionally as a promoter of longevity, well-being, and disease prevention. Some constituents of Ashwagandha have reported anti-influenza properties and are also being prospected in silico against COVID-19 20,21. Ashwagandha extracts have been shown to inhibit the viral RNA replication by inducing nitric oxide in vitro. It is reported that NO induction can inhibit RdRps of viruses and can be used as a potential therapy 2223. Our results suggest that in addition to its immunomodulatory activity, Ashwagandha can be positioned as anti-viral herbal supplement against COVID-19.
Another medicinal herb with most phytoconstituents displaying potential inhibition was Hyssopus officinalis (Hyssop) with ΔG values in the range of -8.4 to -7.7 kcal/mol for SARS-CoV-2 and -8.7 to -7.7 kcal/mol for SARS-CoV. This herb is well known for its anti-microbial, anti-fungal, antiviral and immunomodulatory activities and has been shown to possess anti-HIV and anti-HSV activities 24–26. The major constituent luteolin has been reported to inhibit RNA synthesis in Enterovirus 71 and Coxsackievirus A16 27. There are no reported studies on inhibition of RdRps by the phytochemical constituents of Hyssop. The flavonoid glycoside diosmin showed the maximum potency with docking scores of -8.4 kcal/mol and -8.7 kcal/mol.
The polyphenols catechins and its isomers from Camellia sinensis leaf extract showed good docking scores against both the CoVs in our study. The antiviral activity of catechins is evaluated against many viruses and the major mode of action is inhibition of RNA replication 28. In our study, epigallocatechin gallate (EGCG) gave the lowest binding energy (-8.3 kcal/mol for SARS-CoV-2 and -8.5 kcal/mol for SARS-CoV) amongst the catechins and this is in line with the literature report that EGCG potentially inhibits the viral replication in vitro when compared to other catechins 29.
The phytochemicals oleanolic acid and ursolic acid have been reported to exhibit anti-Hepatitis C virus (HCV) activity by acting as noncompetitive inhibitors of HCV NS5B RdRp 29–31. In our study also, these two molecules sourced from herbs Prunella vulgaris and Ocimum sanctum aerial part extract showed good affinity towards both SARS-CoV and SARS-CoV-2 RdRp. In addition, several other molecules from different herbs showed moderate docking scores and Ki values against RdRp.
Interaction profile of phytochemical compounds and RdRp – molecular visualization:
The 3D and 2D visualization of the interaction of nine phytochemicals showing highest binding affinities (low Ki values) and control drug remdesivir for SARS-CoV-2 is given in Figure 2. Similarly, the 2D and 3D visualization of interactions of phytochemicals and remdesivir with SARS-CoV are depicted in Figure 3. Most of the phytochemicals have at least one hydrogen bond forming interaction with the residues Lys545-Arg553-Arg555 of NTP entry site and Asp760-Asp761-Asp618 of the Mg2+ ion chelating site. The 2D illustration also depicts the non-bonded interactions like van der Waals forces, pi-cation, pi-alkyl etc. between phytochemicals and amino acid residues Asp623 and Asn691, Thr680, Ser682 and Thr687 (involved in NTP recognition and positioning). The control drug remdesivir is a nucleotide analog that binds to RdRp in similar way as a nucleotide and then leads to chain termination. The interaction of various amino acid residues with remdesivir is discussed in literature 32. The ligand binding profile of the highest ranked phytochemicals in our screening is in line with the study of Yin et al. for FDA approved antiviral agents like remdesivir, ribavirin and favipiravir 33.