Mutational spectra analysis reveals significance of RdRp and 3CLpro for drug targeting:
All 75 sequences retrieved from December 2019 to 14th May 2020 were analysed and mutations at nsp 3, nsp 5 and nsp 12 were identified. Among those three locations at orf1ab, nsp 3 or PLpro is found to be most vulnerable harboring a total of 14 mutations. Out of these 14 mutations, some with very low frequency (< 5%) and thus, we considered only those mutations, represented at least 4 out of those 75 studied sequences (i.e> 5%).Four mutations at nsp 3 (i.e. 1159 I>M, 1534 S>I, 2016 T>K and 2376 P>L) were found to be prominent for Indian strain(s) representing a distribution of 5.3 to 13.3%. At nsp 12 or RdRp, a total of 11 mutations are observed till now, but except 4489 A>V (18.6%) and 4715 P>L (74.6%), rest are present in a very low abundance (<5%) (Fig. 1).Nsp 5 or the main protease 3CLpro is found to be least vulnerable to mutation in Indian strain(s) of SARS-CoV-2 with none observed at>5% abundance. Thus the wild type 3CLpro is predominant in the Indian population. The mutation with only >5% abundance in the population was considered for latter analysis and the detailed list of mutations at nsp 3, nsp 5 and nsp 12 are shown in Table Ӏ.
Homology modeling predicts mutation induced altered protein structure:
The findings suggest, 3CLpro or main protease is the most suitable target for drugs due to highly conserved sequence and its corresponding structure in Indian population whereas PLpro is not suitable for drug targeting. Therefore nsp 5 (3CLpro) and nsp 12 (RdRp) were further analysed in this study and modeled structures were generated by Swiss Model. GMQE and QMEAN score was found to be 0.99 (out of 1) and 0.45 (around zero indicate the best quality and ≤ -4.0 indicate low quality) for the wild type structure of nsp 5 (or 3CLpro) whereas it was 0.96 and -1.49 respectively for wild type nsp 12 (or RdRp). High level quality estimation depicted the reliability of all the modeled structures. Superimposing all the mutant structures over the wild type in case of RdRp illustrated (Fig. 2) the structural variation and established the necessity of considering all the mutant structures of the population for drug targeting.
Identification of suitable drug molecules targeting RdRp using molecular docking:
Among 13 repurposed drugs tested against RdRp following our methodology, Elbasvir has come up as the best suitable drug to inhibit the activity of RdRp in Indian population. FDA approved Elbasvir is used as an antiviral against hepatitis C virus (HCV) to treat hepatitis, an infectious, chronic liver disease. The estimated binding free energy (ΔG) of Elbasvir against the RdRp of the wild type structure was observed to be -8.84 kcal/mol, and the estimated Inhibition Constant (Ki) was 363 (nM). The ΔG and Ki values for all the mutant structures were even better in case of Elbasvir indicating slightly greater binding affinity than that of wild type and requirement of smaller dose to inhibit the enzyme.
Remdesivir, the most popular drug of choice at recent times found to be the 2nd best option to inhibit RdRp in India. The ΔG and Ki values of Remdesivir against wild type 3CLpro was -8.52 kcal/mol and 620 (nM). The ΔG values for the mutant structures harboring 4489 A>V and 4715 P>L mutations alone or in combination were -8.02, -7.78 and -7.66 kcal/mol respectively. Beside ΔG and Ki values, the number of polar bonds formed between the interacting residues may be an important feature in receptor-drug association study. In case of wild type and two mutant types harboring either of 4489 A>V or 4715 P>L mutation, 6 polar bonds were formed but when both these mutations occur simultaneously, 5 polar bonds were formed; indicating no such alteration in interaction. The bond lengths ranged from approximately 1.9 Åto 5.5 Å and the detailed interactions are shown in Table ӀӀ. The results revealed that binding affinity of Remdesivir although decreased, alteration was not significantly different for mutant structures, indicating the suitability of this drug in Indian population even under mutational background.
The 3rd suitable option to inhibit RdRp is Methylprednisolone, a corticosteroid medicine that is used to prevent inflammation.The ΔG value for wild type structure was -8.18 kcal/mol and for the mutant structures, it ranged from -7.84 to -7.96 Kcal/mol. The Ki values ranged between 1000 and 2000 (nM).
Cinnamaldehyde(ΔG = -4.9 kcal/mol) and Thymoquinone(ΔG = -4.94 kcal/mol) were used as negative control for all the drugs tested here and statistical test further confirms significantly higher (p<0.0001) binding affinity for all the chosen drugs towards RdRp compared to negative control and their interaction with RdRp is shown in Fig. 3.
Molecular docking identifies promising repurposed drugs targeting 3CLpro
3CLpro or nsp 5 is the most appropriate drug target area for the viral strain(s) of India as it is least vulnerable to mutation and out of 15 repurposed drug molecules tested against it, Elbasvir, TMC 310911 and Lopinavir are found to be the best candidate to inhibit the 3CLpro and it is shown in Fig. 4. According to the observed ΔG and Ki value, Elbasvir(ΔG = -10.44 kcal/mol and Ki = 25nM) is the best one but the number of polar bonds formed during the interaction with 3CLpro is only three, among which one has bond distance ≥ 4 Å. The ΔG and Ki value of TMC 310911 was observed to be -9.98 kcal/mol and 54 (nM) and the number of polar bonds formed was 5, among which all had bond length ≤ 3 Å. In case of well-known protease blocker Lopinavir, the ΔG and Ki value was noted to be -9.26 kcal/mol and 179 (nM) and 4 polar bonds were observed between the drug and the enzyme, in which bond distance ranged between 2.1 Å to 2.8 Å. Here also, Cinnamaldehyde and Thymoquinone were used as negative control for 3CLpro showed ΔG value of -4.96 and -5.32 kcal/mol respectively and statistical test confirms significantly higher (p<0.0001) binding affinity of above mentioned three drugs towards 3CLpro compared to negative control. The binding affinity and interaction of all the selected drugs against 3CLpro are shown in Table ӀӀӀ.