As of 24^th May, 2021, there were 825 ORF1ab protein sequences of Indian SARS-CoV-2 deposited in the NCBI Virus database. All these ORF1ab protein sequences (7096 residues) comprising nonstructural proteins (Nsp1-16) were retrieved from the NCBI-Virus database. The protein accession numbers of ORF1ab sequences used in the study are mentioned in the Table S1. The first reported sequence from China was used as a reference or wild type ORF1ab sequence (accession number: YP_009724389) for mutational analysis in the study. The polypeptide sequences of ORF1ab were exported in the FASTA format, in which the Nsp13 of 601 residues starting from 5325 to 5925 were extracted from the ORF1ab sequence. Nsp13 is proteolytically cleaved by the protease encoded by the SARS-CoV-2 genome. The Clustal Omega online web server tool was used for multiple sequence alignment (MSA) [22, 23] to identify mutations in the amino acid sequences of Nsp13 among the Indian isolates. Clustal Omega is a widely used online program for efficiently generating alignments using a fast and reliable algorithm. For this analysis, the Indian sequences were compared with the reference sequence of Nsp13 (accession number: YP_009724389), and then the identified mutant residues were marked carefully and noted for further analysis.

To understand the impact of mutational changes on the protein stability and dynamicity of Nsp13, protein modelling studies were performed by using DynaMut, a web server tool [24]. The 3D structure of the target protein Nsp13 with PDB ID-6ZSL was used for protein modelling studies. DynaMut program is widely used to analyse and visualise the changes in protein dynamics and stability caused by mutations in terms of the difference in vibrational entropy (∆∆S) and free energy (∆∆G) between the wild type and mutant proteins. The effect of mutations on changes in protein stability was estimated by the predicted ∆∆G(Kcal/mol) between wild-type and mutant proteins. The positive values (more than zero) indicated the stabilization of protein structure and negative values (below zero) showed the destabilization of the protein structure. The protein flexibility and rigidity were calculated by the difference in vibrational entropy (∆∆S) between wild-type and mutant proteins of Nsp13. The positive ∆∆S values corresponded to an increase in the flexibility of the Nsp13 protein, whereas the negative ∆∆S values represented a decrease in the flexibility of the protein.

To understand the effect of amino acid variations on the function of Nsp13, PROVEAN (Protein Variation Effect Analyzer) tool was used [25]. This prediction tool was used to generate a PROVEAN score for each variant. PROVEAN provides an approach to predicting the functionally important sequence variants. Based on the predicted PROVEAN score, mutants are classified as deleterious or neutral. The protein variant is predicted to have a deleterious effect if the PROVEAN score is equal to or less than the default threshold score (-2.5) and if the PROVEAN score is more than the threshold score, the variant is predicted to have a neutral effect. PROVEAN tool is useful in measuring the functional effect of protein sequence variations, including amino acid substitution, insertion, and deletions.

B-cell epitope predictions were performed using an online web server IEDB (Immune Epitope Database) prediction tool [26], as described by Jesperson et al. [27]. This prediction tool predicts the linear continuous B-cell epitopes of Nsp13 based on the BepiPred linear epitope prediction method by using all the standard parameters, such as flexibility, hydrophilicity, accessibility, and beta turns at a threshold value of 0.48. The identified epitopes of Nsp13 were further assessed for immunological parameters, such as antigenicity, allergenicity, and toxicity (by Toxin Pred). Antigenicity and allergenicity of peptides were predicted by freely available web servers, VaxiJen (http://www.jenner.ac.uk/VaxiJen) [28] and Allergen FP v.1.0 (http://ddg-pharmfac.net/AllergenFP) [29]. Moreover, the discontinuous B-cell epitopes were also predicted using Discotope 2.0 web server tool [30] at a threshold value of -6.6 followed by representing the identified epitopes in the three-dimensional structure of Nsp13 protein. The Discotope score of residues above the threshold value indicates positive prediction and those below the threshold value indicate negative prediction, respectively.

The Nsp13 SARS-CoV-2 sequences reported from India till 24^th May, 2021, were aligned with the reference sequence reported from China (accession number: YP_009724389) using Clustal omega-mediated multiple sequence alignment for mutational analysis. The amino acid variations in the Nsp13 protein sequences among Indian isolates were identified carefully and noted for analysis. The analysis revealed the presence of 38 mutations distributed all over the Nsp13 protein. The details of each mutation are listed in Table 1 and their location observed are represented in the three-dimensional structure of Nsp13 (PDB ID-6ZSL) using the Pymol visualisation tool (Fig. 1). Our data also demonstrated a change in polarity and charge upon amino acid mutations in Nsp13. It was found that most of the mutations did not contribute to any changes (Neutral to Neutral). Although some of these mutations led to an alteration in charge from basic to neutral (H164Y, H290Y, R392C, and R442Q), only one mutation at position 142 (H142V) altered the charge from acidic to neutral, as depicted in Table 1.

The mutational analysis for evaluating the impact of mutations on the stability and flexibility of Nsp13 protein was conducted by the DynaMut web server. This web server calculated the difference in free energy(∆∆G) between wild-type and mutants. The positive ΔΔG corresponded to an increase in stability, while the negative ΔΔG corresponded to a decrease in stability. Our data demonstrated that a total of 38 mutations caused an increase or decrease in the stability of Nsp13 protein. The positive ∆∆G values in the range of 0.013 to 0.975 kcal/mol showed a stabilising effect (Table 1). We observed the maximum positive ΔΔG for H164Y (0.975), A237T (0.924), T214I (0.881), C309Y (0.844), S236I (0.750), and P419S (0.693), indicating stabilising mutations (Fig. 2). While, V305 (-2.53), G54S (-1.69), and A308Y (-1.306) showed a high negative ∆∆G value, indicating the destabilization effect on the Nsp13 protein (Fig. 2). Similarly, the flexibility was also calculated by the difference in vibrational entropy (∆∆S) between the wild-type and mutant protein of Nsp13. The maximum positive ∆∆S (0.896) for H290Y caused an increase in the flexibility of Nsp13 protein whereas, the maximum negative ∆∆S (-0.725) for A308Y represented a decrease in flexibility of the protein (Table 2). Collectively, our data suggested that mutations identified in the Nsp13 protein could alter its stability and dynamicity.

The PROVEAN score predicts the functional impact of mutations in Nsp13. PROVEAN scores of mutants equal to or less than the default threshold score (-2.5) were predicted to have a deleterious effect on protein whereas PROVEAN scores more than the threshold score were predicted to have a neutral effect on the respective protein. A total of 38 mutations were reported in Nsp13 among Indian SARS-CoV-2 sequences, including 15 mutants as neutral and 23 mutants as deleterious based on their predicted PROVEAN scores, as mentioned in Table 3.

Linear continuous B-cell epitopes of SARS-CoV-2 Nsp13 were predicted by the IEDB web server tool, as shown in Fig. (3A). The BepiPred score of Nsp13 residues depicted a minimum score of -0.161 and a maximum score of -0.651 at a threshold value of 0.482. The data set identified the top seven linear epitopes with at least eight amino acid residues, of which both their sequences and locations are mentioned in Fig. (3B). These epitopes were further analyzed for antigenicity, allergenicity, and toxicity by in-silico based approaches. The predicted data showed that all of the seven peptides are non-toxic based on the Toxin Pred prediction tool. Out of the seven peptides, three peptides were identified as allergens and four as non-allergens based on allergenicity prediction (Fig. 3B). Similarly, two of these predicted linear epitopes exhibited antigenicity and the remaining five peptides were non-antigenic as analyzed by the VaxiJen prediction tool (Fig. 3B). In addition, discontinuous B-cell epitopes of Nsp13 were predicted using Discotope 2.0 server tool. The available 3D structure of Nsp13 in PDB format (PDB ID: 6ZSL) with a chain A of protein was used as input in the Discotope 2.0 server tool (Fig. 3C). The residue with Discotope score equal to or greater than the threshold value (-6.6) was identified as discontinuous epitope residues of Nsp13. The data sets identified 18 B cell epitope residues out of the total residues of Nsp13 protein at a threshold value of -6.6 (Fig. 3D). Histidine and valine at positions 245 and 247, out of 18 predicted residues were significantly positively predicted with high DiscoTope score values, as depicted in Fig. (2D). Overall, this data revealed B-cell epitopes contributed by Nsp13.

Fig. (1). A The diagrammatic representation of SARS-CoV-2 Nsp13 (601 amino residues) indicating the mutations identified in the study. B Three-dimensional structure of SARS-CoV-2 Nsp13 (PDB ID: 6ZSL) depicting the location of identified mutations in this study.