In Silico Prediction of Epitopes in Virulence Proteins of Mycobacterium ulcerans for Vaccine Designing.

Background: Mycobacterium ulcerans is the fundamental agent of the third most common Mycobacterial disease known as Buruli Ulcer (BU). It is an infection of the skin and soft tissue affecting the human population worldwide. Presently, the vaccine is not available against BU. Objective: This study aimed to investigate the vaccine potential of virulence proteins of M. ulcerans computationally.
Methods: Chromosome encoded virulence proteins of Mycobacterium ulcerans strain Agy99 were selected, which were available at the VFDB database. These proteins were analyzed for their subcellular localization, antigenicity, and human non-homology analysis. Ten virulence factors were finally chosen and analyzed for further study. Three-dimensional structures for selected proteins were predicted using Phyre2. B cell and T cell epitope analysis was done using methods available at Immune Epitope Database and Analysis Resource. Antigenicity, allergenicity, and toxicity analysis were also done to predict epitopes. Molecular docking analysis was done for T cell epitopes, those showing overlap with B cell epitopes.
Results: Selected virulence proteins were predicted with B cell and T cell epitopes. Some of the selected proteins were found to be already reported as antigenic in other mycobacteria. Some of the predicted epitopes also had similarities with experimentally identified epitopes of M. ulcerans and M. tuberculosis which further supported our predictions.
Conclusion: In-silico approach used for the vaccine candidate identification predicted some virulence proteins that could be proved important in future vaccination strategies against this chronic disease. Predicted epitopes require further experimental validation for their potential use as peptide vaccines.

INTRODUCTION

Buruli ulcer (BU), caused by Mycobacterium ulcerans is the disease of subcutaneous tissue [1]. World Health Organization (WHO) has classified it as one of the skin-related neglected tropical diseases [2].

Cases of BU have been reported in more than 33 countries worldwide [3]. It has become the most common mycobacterial disease after tuberculosis and leprosy. BU causes deep skin lesions and is more prevalent in children less than 15 years of age than adults. Because of the late diagnosis, excision of the lesion combined with skin grafting is used for the treatment of the disease [4].

Vaccination is an important strategy to induce an immune response against the pathogen by specifically inducing the adaptive immune system. Limited attempts have been made to identify vaccine candidates against BU [5]. The growth rate of this bacilli, is slow (doubling time >48 h) and this makes experimental identification of vaccine candidates difficult. Currently, an effective vaccine against this disease is not available. BCG vaccine is used against this disease, but the role of Mycobacterium BCG vaccination against BU is controversial [6]. Effective BU vaccine would be helpful in protecting the populations in highly endemic areas. Identification of microbial proteins involved in immune response may prove an important step in the development of ‘effective vaccine’. The use of the proteins having immunogenic potential may be better because it can eliminate the side effects of whole-cell vaccines.

Prediction of immunogenic peptide epitopes by computational analysis is effective in vaccine designing [7] and can reduce efforts and costs associated with the experimental trials. An attempt has been made for computational identification of vaccine candidates in proteome of M. ulcerans [8]. Virulence factors (VFs) have been explored to design vaccines against several pathogens [9].

In this study, we explored virulence factors of M. ulcerans to identify the potential vaccine candidates through a reverse vaccinology approach. Several in silico tools were used for the identification of B- and T-cell epitopes from the selected proteins, which also included their allergenic, antigenic and antitoxic predictions. Antigenic, non-allergenic and non-toxic epitopes were selected. Docking simulation was also performed for MHC Class I and MHC Class II epitopes with their MHC alleles. Selected peptides would be helpful in the future vaccine designing against Buruli ulcers.

MATERIALS AND METHODS

Retrieval of Mycobacterium ulcerans Virulence Factors

Virulence factor database VFDB (http://www.mgc.ac.cn/ VFs/) [10] provides data of ‘bacterial virulence factors’ of various pathogenic microbes. Chromosome encoded virulence factors of Mycobacterium ulcerans strain Agy99 were already available in VFDB database. Proteins sequences of these virulence factors were retrieved from VFDB in FASTA format.

Prediction of Sub-cellular Localization

Surface exposed proteins are an important target of the immune system as cells of the immune system recognize these proteins of the pathogens very efficiently. Surface proteins that are also virulence factors may be important in therapeutics design. For vaccine designing, the identification of subcellular localization of proteins plays a very important role. All the virulence factor proteins retrieved were analyzed by CELLO v.2.5 (http://cello.life.nctu.edu.tw/) [11] for their sub-cellular localization. The prediction was based on protein sequences only. CELLO is based on two-level SVM classifiers. This program identifies the proteins as extracellular, outer membrane, cytoplasmic, periplasmic, and inner membrane.

Antigenicity Evaluation

Surface located protein sequences predicted as a result of CELLO analysis were analyzed for their antigenicity. Vaxijen (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/ Vaxi-Jen.html) [12] is the server for the prediction of protective antigens based on alignment-free approach. It classifies antigens based on the physicochemical properties of proteins. The threshold value of Vaxijen was set to 0.5 Proteins with a Vaxijen score equal to or above 0.5 were considered to be antigenic. Predicted antigenic proteins were chosen for further study.

Human Non-Homology Analysis

Human non-homologous proteins are regarded as suitable vaccine candidates. The selection of such proteins as vaccine targets reduces the chances of cross-reactivity. BLASTp was used for the analysis of such proteins against host Homo sapiens (taxid: 9606). E-value (expectation value) was set to 0.005. Homologous proteins were excluded out for further analysis. Non-homologous top ten antigenic virulent proteins were selected for the reverse vaccinology approach.

Structural Modeling

To predict structural B cell epitopes, three-dimensional structure of selected virulent proteins was required. Phyre2 (http://www.sbg.bio.ic.ac.uk/phyre2) [13] was used for structural modelling. phyre2 uses a homology approach to build structure and it uses ab initio folding simulation to predict structure in the region of non-homology. Predicted structures were refined by the GalaxyRefine tool at GalaxyWEB [14] and were checked by PROCHECK analysis generated by PDBsum Generate at PDBsum server (http://www. ebi.ac.uk/thornton-srv/databases/pdbsum/Generate.html).

B cell Epitope Prediction

B cell lymphocytes interact with B cell epitopes and provide long-term immunity to the host. Web-based tool BepiPred-2.0 [15] (http://tools.iedb.org/bcell/) hosted at IEDB [16] was used to predict sequence-based B-cell epitope prediction at threshold value 0.5 The BepiPred-2.0 is based on a random forest algorithm which is trained on epitope data derived from antibody-antigen protein structures. ElliPro, a web-based tool at IEDB [17] (http://tools.iedb.org/ellipro/) was used to predict linear structural B cell epitopes on default parameters (minimum score value 0.5 and maximum distance of 6Å). Ellipro is based on an algorithm like MODELLER, Jmol viewer etc. to predict and visualize the epitopes in the given protein structure. It calculates residue protrusion index (PI) and clusters neighboring residues according to their PI values. The antigenicity of the predicted epitopes, was calculated using Vaxijen at a threshold value 0.5.

Prediction of MHC Class I Epitopes

MHC Class I epitope prediction was made using server ANN 4.0 [18] at IEDB (http://tools.iedb.org/main/tcell/) with epitope length 9 mer. FASTA formatted protein sequences were used as input for the prediction tool. Most frequently occurring MHC alleles were selected for analysis. Based on IC50 values, T cell epitopes are classified as high-affinity (IC50s < 50 nM), intermediate-affinity (IC50s < 500 nM) and low-affinity (IC50s < 5,000 nM). Predicted high affinity epitopes i.e., epitopes with IC50s of < 50 were selected for further analysis. Selected epitopes were analyzed for class I immunogenicity using the IEDB Class I Immunogenicity tool (http://tools.iedb.org/immunogenicity/) [19] and non-immunogenic epitopes (epitopes with negative score) were eliminated. Recognition of MHC complex bound peptides by T cells is predicted by this tool. A high immunogenicity score is associated with the induction of cellular immunity cells with high potency. This tool has only been validated for 9-mers.

Prediction of MHC Class II Epitopes

MHC class II binding predictions for selected antigenic proteins was made using the NetMHCIIpan4.0 EL prediction approach [20] at IEDB-AR with epitope length 15 mer. Alleles of full HLA reference set were used. NetMHCIIpan 4.0 EL prediction approach covers methods that are trained with data retrieved from mass spectrometry (MS) experiments, known as eluted ligands (EL). According to the IEDB recommendations, on the basis of a percentile rank of ≤ 10%, epitopes were selected for analysis. MHC class II epitopes that bound with at least 50% alleles (13 alleles) were selected for further analysis.

Epitope Shortlisting by Antigenicity, Allergenicity and Toxicity Analysis

Predicted linear sequence-based B-cell epitopes, MHC class I and MHC class II T cell epitopes were subjected to a series of analyses. These predicted epitopes were analyzed to determine their antigenicity, allergenicity and toxicity properties through VaxiJen v2.0, AllergenFP v1.0 (https://ddg-pharmfac.net/AllergenFP/) [21] and ToxinPred (http:// crdd.osdd.net/raghava/toxinpred/) [22] respectively. AllergenFP predicts the allergens and non-allergens in the given peptide sequence by using an alignment-free, descriptor-based fingerprint approach. ToxinPred is based on the Support Vector Machine (SVM) method to predict the toxicity of peptide sequences based on their physio-chemical properties.

Structural Modeling of Epitopes

MHC Class I and II epitopes which were also part of B cell epitopes or had overlapping with B cell epitopes, were considered as best epitopes and 3D structures of these epitopes were predicted using online tool PEP-FOLD 3.5 (https://bioserv.rpbs.univ-paris-diderot.fr/services/PEP-FOLD3/). PEP-FOLD 3.5 is based on de novo approach, which aim at predicting peptide structures from amino acid sequences [23].

Molecular Docking of Epitopes with MHC Class I and Class II Alleles

Molecular docking is an in-silico approach that involves the interaction between the ligand and receptor molecule. Modeled epitope ligands were docked to their specific MHC allele receptors by ClusPro Server (]. The docking steps in ClusPro are based on rigid body docking, clustering of lowest energy structure and structural refinement. Best docked complexes were selected on the basis of their lowest energy scores.

Binding Affinity Analysis

Binding affinities of best-docked complex of epitope and receptor molecule resulting from ClusPro analysis were predicted using PRODIGY web server [25]. It predicts the binding affinity in biological complexes in terms of Gibbs free energy (∆G). Negative values of free energy indicate the greater likelihood of complex formation in specific conditions. In this study, MHC alleles and epitope molecules were defined as chains A and B respectively and the temperature was set to 25°C.

RESULTS

Retrieval of Virulence Factors

VFDB consisted of total of 173 proteins as chromosome-based virulence factors in Mycobacterium ulcerans strain Agy99. These protein sequences were retrieved & downloaded in FASTA format.

Sub-cellular Localization Analysis

All the retrieved 173 virulence factors were analyzed by CELLO programs for identification of their putative sub-cellular location. 101 proteins were predicted as either extracellular or membrane-bound. The protein sequences predicted at extracellular or membrane-bound were taken for antigenicity analysis. Those proteins that were predicted exclusively as cytoplasmic were discarded.

Antigenicity Analysis

101 virulence factors that were predicted as either extracellular or membrane-bound proteins were analyzed for their antigenicity evaluation. Based on the VaxiJen 2.0 antigenicity score of ≥ 0.5, 38 proteins were predicted as antigenic. These proteins were analyzed further.

Elimination of Homologous Human Proteins

A subset of 38 antigenic proteins was checked to determine homology with humans using the Basic Local Alignment Search Tool (BLAST) to exclude cross-reactivity of the vaccine. Proteins that were showing homology with human proteins were eliminated, leaving 27 candidate proteins. Of these 27 proteins, top 10 antigenic proteins based on the Vaxijen score were finally selected for B cell and T cell epitope analysis. These proteins, with their Vaxijen score and subcellular localization, are given in Table 1.

3D Structure Prediction

Phyre2 predicted structures of the selected protein sequences, and the model that was based on the highest identity and best alignment was selected for each protein. Model refinement and Ramachandran plot analysis were done for the predicted structures. Refined predicted protein structures are given in Fig. ( and Ramachandran plot analysis of each protein is given in Table , and Ramachandran plot of predicted protein structures are given in Fig. (.

Fig. (1)

Predicted protein structures. (A higher resolution / colour version of this figure is available in the electronic copy of the article).

Fig. (2)

Ramachandran plot of predicted protein structures. (A higher resolution / colour version of this figure is available in the electronic copy of the article).

Selected B-cell Epitopes

A total of 73 sequence-based B-cell epitopes were predicted from selected proteins by the BepiPred 2, of which 52 epitopes were with the length of more than five (>5) residues. On the basis of antigenicity, allergenicity and toxicity analysis of these epitopes, 28 epitopes were selected and are given in Supplementary Table . Graphical representation of predicted sequence-based B cell epitopes is given in

Fig. (. A total of 74 linear structural B-cell epitopes were predicted by Ellipro, of which 69 epitopes were above the length of 5 residues. The 47 structural B cell epitopes were shortlisted (Supplementary Table ) on the basis of antigenicity analysis of these epitopes by Vaxijen 2.0. Significant overlapping was observed between linear structural epitopes predicted by Ellipro with epitopes predicted by the BepiPred 2.0 method. These overlapping epitopes were finally selected as B cell epitopes and are given in Table .

Fig. (3)

Graphical representation of B cell epitopes of antigenic proteins using Bepipred 2.0 linear epitope server. (A higher resolution / colour version of this figure is available in the electronic copy of the article).

MHC Class I Epitopes

After analysis of selected protein sequences with ANN 4.0 program for MHC Class I binding prediction, epitopes in all three categories high-affinity (IC50s < 50 nM), intermediate-affinity (IC50s < 500 nM) low-affinity (IC50s < 5,000 nM) were predicted in all sequences. Epitopes of high affinity category were selected and characterized for immunogenicity using class I MHC immunogenicity predictor and epitopes with positive scores were filtered for further analysis. Total 356 epitopes had a positive immunogenicity score. Filtering of these epitopes by antigenicity, allergenicity and toxicity analysis, 54 epitopes were finally selected for MHC class I and are given in Table 4.

MHC Class II Epitopes

Selected proteins were subjected to analysis for the HLA Class II specific T- cell epitopes binders. Total 27 human MHC class II alleles were used for analysis. NetMHCIIpan-4.0 EL approach predicted MHC class II epitopes in each protein analyzed. Epitopes showing affinity with at least 50% of selected HLA Class II alleles were selected at a threshold value of 10% percentile. 7 proteins of 10 were found to have the MHC Class II binding sites at selected criteria. Three proteins (MUL_2383, MUL_4591, MUL_1207) did not show MHC II class epitopes at defined thresholds. MUL_1208 showed the highest number and MUL_4389 showed the minimum number of selected Class II epitopes. Total 82 epitopes were predicted at the above threshold. Filtering of these epitopes by antigenicity, allergenicity and toxicity analysis, 44 epitopes were finally selected for MHC class II and are given in Table .

Overlapping MHC Class I and MHC Class II Epitopes

Five proteins (MUL_4389, MUL_3082, MUL_1208, MUL_4591, MUL_4987) had overlapping epitopes having binding affinity for both MHC Class I and MHC Class II epitopes. These overlapping epitopes are given in Table . These identified overlapping epitopes may play an important role in eliciting both CD8(+) and CD4(+) T cell immune response.

Overlapping B-cell and T-cell Epitopes

T cell epitopes that were part of B cell epitopes or had overlapping regions between B cell and T cell epitopes were also identified for more effective vaccine designing. Five proteins (MUL_4389, MUL_3082, MUL_1208, MUL_0252, MUL_2013) showed overlapped epitopes for B cell and T cell. These overlapped B cells and T cell epitopes are given in Table 7.

Molecular Docking Analysis

An epitope that can elicit both T cells and B cells could prove very important in peptide vaccine designing. Thus, T cell epitopes having a common region with B cell epitopes were analyzed further by docking analysis with their MHC alleles to identify their suitability as vaccine candidates. The three-dimensional model of epitopes was modeled by PEP- FOLD 3.5 For docking analysis of MHC Class I epitopes, three-dimensional structures of receptor alleles that were specific to each epitope were retrieved from RCSB PDB (https://www.rcsb.org/)[26]. Three MHC Class I epitopes could not be docked with their MHC-I allele because the structures of their respective alleles were not available in PDB. Thus, docking analysis was done for total 10 MHC Class I epitopes. For MHC Class II docking analysis, HLA-DRB1*04:01 was used as a receptor as this allele had an affinity with all 17 predicted MHC Class II epitopes and the three-dimensional structure of this allele was also available at PDB. Docking analysis was done by ClusPro 2.0 and the generated best-docked complexes were selected on the basis of the lowest binding energy. Docking results showed the lowest binding energy in the range between -504.7 kJ/mol to -850.2 kJ/mol for MHC Class I analysis and -592.8 kJ/mol to -815.2kJ/mol for MHC Class II analysis. Selected docked complexes were then submitted to the PRODIGY server. The Gibbs free energies (∆G) given by the PRODIGY server ranged between -11.7 kcal/mol to -14.2 kcal/mol for MHC Class I analysis and -19.4 kcal/mol to -19.6 kcal/mol for MHC Class II analysis Lowest energy score of ClusPro 2.0 docked complexes and their free-binding energies given by PRODIGY server are given in Table . All peptides had low free-binding energy in docking with MHC I and II alleles which suggests that these epitopes could prove good vaccine candidates.

DISCUSSION

Buruli ulcers, if untreated, can cause permanent disability. Specific therapeutics or vaccination is not available against this disease and current disease control measures mainly depend on early case diagnosis and antibiotic treatment. Vaccine against this disease would be certainly beneficial to the worldwide community and especially children that have the highest percentage of the affected population. Therefore, we aimed to identification of proteins that can be potentially used for peptide vaccine designing.

The reverse vaccinology approach was used to screen virulence proteins to filter out antigenic proteins. Vaccine potential analysis of plasmid-encoded virulence factors has been done extensively in M. ulcerans [27]. Thus, in this study, we analyzed the antigenic potential of chromosome encoded virulence proteins of M. ulcerans which were available at the VFDB database. Virulence factors play a very important role in the establishment of infection in the host. Proteins localized on the cell membrane or extracellularly could be good antigens as they are readily exposed to cells of the host immune system. The first objective was to identify cell surface or extracellular virulence factors. Antigenicity and subcellular localization of these virulence proteins was determined by Vaxijen and CELLO program respectively. A study of the similarity of protein sequences between the host and the pathogen is necessary because proteins with high similarity could induce autoimmune responses. Antigenic proteins which were found to be surface localized were further analyzed for homology analysis with the host.

B and T cell epitope analysis was done for selected virulence proteins to identify immunogenic regions in these proteins. Structures of these proteins were not available in any structural database, so the protein structure of these proteins was predicted by Phyre2. Both sequence-based and structural B cell epitope analysis was done and the common B cell epitopes predicted by both the methods were finally selected. All selected proteins were predicted to have B cell epitopes. Linear B cell epitope YQFGDVDA5-12 of MUL_3082 (EsxN) had similarities with the published epitope of EsxN protein of M. tuberculosis [28].

T cell epitope analysis was done using an artificial neural network for MHC Class I and NetMHCIIpan-4.0 EL method for MHC Class II at IEDB. MHC Class I epitopes LTDANPPEV91-99, AVAGAAILV8-16 of MUL_4389 (LpqH) was shown similarity with the previously identified epitopes of protein LpqH of M. tuberculosis [29, 30]. Peptide ANPPEV94-99 was also predicted with B cell epitope, which indicated its induction of both cellular and humoral immunity. LpqH has been shown to have vaccine potential in tuberculosis [31]. In a study, dendritic cells with phagocytose apoptotic macrophages containing mycobacterial protein LpqH has been shown to activate CD8 T cells [32].

MHC Class II epitopes TSNMKFQAAYNAAGG278-292, NMKFQAAYNAAGGHN280-294, MKFQAAYNAAGGH- NA281-295, THSWEYWGAQLNAMR304-318 of MUL_4987 (secreted antigen 85-A FbpA) have shown similarity with experimentally identified epitopes of FbpA proteins of M. ulcerans [33]. Peptide NAAGGHNAV288-296 was also predicted in MHC Class I analysis. Studies have shown that recombinant BCG vaccine expressing MU-Ag85A (BCG MU-Ag85A) exhibited the highest protection in MU challenged mice in comparison to BCG vaccination alone [34]. DNA vaccine with antigen 85A of Mycobacterial species, including M. ulcerans [33, 35] has shown improved protection against Mycobacterial diseases. MHC Class II epitope ATNFFGLNAIPIALN128-142 of MUL_1208 (PPE family protein) has shown similarity/overlapping with the experimentally identified epitope of protein of M. tuberculosis [36]. This peptide was also found overlapping between MHC Class I and Class II epitope. MUL_1208 was predicted with maximum numbers of T cell epitopes. Proteins of PE/PPE family in Mycobacteria have been explored as potential vaccine candidates [37]. Previous studies have shown that many PE/PPE proteins can induce a strong T cell response [38]. T cell response against the M. ulcerans infection in humans have been identified in which untreated lesions displayed higher expression of TNF-α, IFN-g etc. [39].

Analysis also identified some virulence factors with overlapped MHC Class I and MHC Class II epitopes which could induce both cytotoxic T cells and helper T cell immune response, which is required to eliminate the intracellular pathogen. Our analysis also revealed overlapping B cell and T cell epitopes in some selected proteins, which can provide both cellular and humoral immune responses against infection and could prove to be strongly immunogenic. Docking analysis of these T cell epitopes confirmed the interaction among the predicted T cell epitopes and MHC-I/II alleles.

CONCLUSION

Buruli ulcers could be prevented with an efficient vaccine which can be the lasting solution to this chronic disease. Peptide-based vaccine candidates identified using the in-silico approach could be helpful in this direction and could reduce the number of in vitro and in vivo analyses in future vaccine identification strategies.

Table 1

Selected proteins with their predicted Vaxijen score and sub-cellular localization.

S. No.	Protein Locus Tag	Protein Name	Vaxijen Score	Sub-cellular Localization
1	MUL_4389	LpqH	1.257	Extracellular
2	MUL_3082	EsxN	0.7033	Extracellular
3	MUL_1208	PPE family protein	0.6964	Extracellular
4	MUL_0252	OmpA	0.6815	Extracellular/membrane
5	MUL_2383	MmpS4	0.6761	Extracellular
6	MUL_4591		0.6735	Extracellular
7	MUL_1207	PE family protein	0.665	Extracellular
8	MUL_1214	Conserved transmembrane protein	0.6587	Membrane
9	MUL_4987	Secreted antigen 85-A FbpA	0.6493	Extracellular
10	MUL_2013	DrrB	0.6468	Membrane

Table 2

Ramachandran plot analysis summary.

Protein	No. of Amino Acids	Most Favored Regions [A, B, L]	Additional Allowed Regions [a, b, l, p]	Generously Allowed Regions [~a, ~b, ~l, ~p]	Disallowed Regions [XX]
MUL_4389	161	81.0	14.3	1.6	3.2
MUL_3082	94	96.3	3.7	0.0	0.0
MUL_1208	517	78.6	14.9	2.7	3.9
MUL_0252	332	88.4	10.2	0.4	1.1
MUL_2383	156	84.4	13.3	0.0	2.3
MUL_4591	104	92.4	4.3	0.0	3.3
MUL_1207	102	95.2	4.8	0.0	0.0
MUL_1214	316	84.7	12.7	0.7	1.8
MUL_4987	337	91.1	7.7	0.0	1.1
MUL_2013	289	90.7	8.4	0.8	0.0

Table 3

Selected B cell epitopes.

Protein	Sequence-Based Linear B Cell Epitope			Structure-Based Linear B Cell Epitope
	S. No	Position	Sequence	S. No	Position	Sequence
MUL_4389	1	23-51	SSNKSTTSGGESSSTGSTSASTGGGQASG	1	19-47	LSGCSSNKSTTSGGESSSTGSTSASTGGG
	2	94-99	ANPPEV	2	92-99	TDANPPEV
	3	114-128	YTSGTGQGNASATKD	3	116-122	SGTGQGN
				4	127-133	KDGSSYK
MUL_3082	1	5-12	YQFGDVDA	1	1-12	MTINYQFGDVDA
	2	65-90	YEQANSHGQKVQAAGNNMAQTDSAVG	2	80-94	NNMAQTDSAVGSSWA
MUL_1208	1	170-254	SAPRTGPPPKVVNPGGGQVAHLAADAARAQSGAAAAGGSDPWQLLWQLLQYLWNAYTGFNTWMFDLIWEFLQDPIGNTIKIIIAF	1	170-212	SAPRTGPPPKVVNPGGGQVAHLAADAARAQSGAAAAGGSDPWQ
	2	278-495	FNLVGWPTWALILSSPFLLPAALGLALSAIAFTPIQIAAEVPAAAAPLAAAAVTAKSVFPAASLASPGTGSAGAPASGAGAGAPGTAPPGPGAPAPASFLYAVGGSGDWGPSLGPTVGGRSGAKAPAVTIPAAGAAAASRASARSRRRRRTEMRDYGDEFLDMDANGVTGPAATDDAARASERGAGQFGFAGTVSTEAVLQAAGLTELAVDEFGNGPR	2	352-374	PASGAGAGAPGTAPPGPGAPAPA
				3	380-432	VGGSGDWGPSLGPTVGGRSGAKAPAVTIPAAGAAAASRASARSRRRRRTEMRD
				4	458-465	SERGAGQF
MUL_0252	1	48-91	GAYQRPTSVNGPTGELPTVTASSAKPDRPELSLSLLSVSRSGNT	1	46-82	GYGAYQRPTSVNGPTGELPTVTASSAKPDRPELSLSL
	2	128-151	DPKVRALDFSNAEPVFSASGPIPD	2	115-139	LSQNVTVVDQIQIDPKVRALDFSNA
	3	306-323	LGSANPIASNDTAEGRIK	3	303-325	AKGLGSANPIASNDTAEGRIKNR
MUL_2383	1	5-11	PRGGGCT	1	1-22	MPAGPRGGGCTEIVTVLKRAWL
	2	43-67	MFGSQQLPTYADSTPAGSVSSSGPK	2	53-68	ADSTPAGSVSSSGPKR
	3	105-111	EVETDAP	3	105-113	EVETDAPSM
	4	135-146	EVKDERTSSAVS	4	132-138	SGGEVKD
MUL_4591	1	5-24	LTLQPDVISRLSQGHDATVT	1	1-11	MNEILTLQPDV
MUL_1207	1	42-48	AVDPVSL	2	38-53	VVPAAVDPVSLQTAAG
MUL_1214	1	296-313	VWFTPGAGLHRAAPTETA	1	292-316	LMRLVWFTPGAGLHRAAPTETASRT
MUL_4987	1	120-139	SFYSDWYNPACGKAGCTTYK	1	125-138	WYNPACGKAGCTTY
	2	213-233	DAGGYKASDMWGPKDDPAWAR	2	212-221	GDAGGYKASD
				3	223-232	WGPKDDPAWA
	3	257-270	GKPSDLGGDNLPAK	4	255-271	GNGKPSDLGGDNLPAKF
	4	326-335	GATPNTGDTQ	5	325-337	LGATPNTGDTQGA
MUL_2013	1	5-30	AVDPTPVPTFKGAGPSAAKPRLSTLQ	1	1-34	MSISAVDPTPVPTFKGAGPSAAKPRLSTLQQWWV
	2	70-87	PWNHYVGGGASGVPSSLG	2	61-81	VGFYIPFAIPWNHYVGGGASG

Table 4

Selected MHC class I epitopes.

Protein	Position	Peptide	Vaxijen Score
MUL_4389	8-16	AVAGAAILV	0.8755
	83-91	AATGIAAVL	0.5918
	91-99	LTDANPPEV	0.8028
	133-141	KISGTATGV	1.4876
MUL_3082	55-63	TQLGRNFQV	0.7566
MUL_1208	43-51	EYALTAAEL	0.7317
	114-122	TELGANHAL	0.7149
	126-134	LVATNFFGL	0.8695
	160-168	YQATAGAAL	1.0632
	207-215	GSDPWQLLW	1.4657
	262-270	IITYGPLLF	0.7831
	263-271	ITYGPLLFA	0.8768
	264-272	TYGPLLFAL	1.0191
	292-300	SPFLLPAAL	0.8046
	294-302	FLLPAALGL	0.7752
	296-304	LPAALGLAL	0.6705
	317-325	EVPAAAAPL	0.5713
	407-415	IPAAGAAAA	0.7259
	418-426	ASARSRRRR	1.8636
	419-427	SARSRRRRR	2.117
MUL_0252	36-44	VVIPLLLAA	1.4773
	70-78	SAKPDRPEL	0.7163
	72-80	KPDRPELSL	0.7873
	252-260	KLRACPDAK	1.2443
	321-329	RIKNRRVEI	1.5031
MUL_2383	18-26	KRAWLPLVL	0.5502
	22-30	LPLVLAVVF	1.2528
	30-38	FAVGGFAVS	0.534
MUL_4591	62-70	RARAGQALQ	1.2543
MUL_1207	5-13	VVPEGLAAA	0.1486
	12-20	AASAAVEAL	0.7269
	35-43	ISAVVPAAV	0.5463
	61-69	HAAVAAEGV	0.8884
MUL_1214	5-13	VMAYPWHSR	1.0159
	13-21	RRDYWLIGI	0.7775
	17-25	WLIGIAAFV	0.9953
	26-34	VIVLFAWWR	1.495
	52-60	RPAEPGAEA	0.7328
	88-96	YLNCYGIRV	1.6856
	120-128	SAATNLAAL	0.5837
	129-137	QARAARIPL	1.0631
	139-147	ETARVTARR	1.0283
	151-159	HLREIGWEV	0.691
MUL_4987	10-18	AATGTSRRL	1.7137
	97-105	NTPAFEWYY	1.9458
	203-211	GPSLIGLAM	0.6079
	288-296	NAAGGHNAV	1.9957
MUL_2013	59-67	FTVGFYIPF	2.2144
	63-71	FYIPFAIPW	3.0233
	67-75	FAIPWNHYV	1.0055
	156-164	YAIGFHFNR	1.3973
	262-270	VMTPTLAWL	0.9321
	267-275	LAWLAGFAL	0.7423
	269-277	WLAGFALFL	1.1617

Table 5

Selected MHC Class II epitopes.

Protein	Position	Peptide	No. of MHC Class II Alleles	Vaxijen Score
MUL_4389	128-142	DGSSYKISGTATGVD	13	1.8416
MUL_3082	14-28	GALIRAQAASLEAEH	15	0.6047
	58-72	GRNFQVIYEQANSHG	16	0.9052
	59-73	RNFQVIYEQANSHGQ	13	0.7598
	61-75	FQVIYEQANSHGQKV	14	1.1473
	62-76	QVIYEQANSHGQKVQ	16	1.1696
	71-85	HGQKVQAAGNNMAQT	13	1.1013
	72-86	GQKVQAAGNNMAQTD	13	1.1272
MUL_1208	5-19	AGPIWIASPPEVHSA	17	0.7617
	128-142	ATNFFGLNAIPIALN	17	0.5962
	184-198	GGGQVAHLAADAARA	14	0.9866
	185-199	GGQVAHLAADAARAQ	16	0.8100
	186-200	GQVAHLAADAARAQS	15	0.6923
	187-201	QVAHLAADAARAQSG	15	0.8660
	188-202	VAHLAADAARAQSGA	13	0.8587
	304-318	LSAIAFTPIQIAAEV	13	0.8326
	308-322	AFTPIQIAAEVPAAA	15	0.6518
	309-323	FTPIQIAAEVPAAAA	20	0.7089
	310-324	TPIQIAAEVPAAAAP	19	0.6859
	311-325	PIQIAAEVPAAAAPL	13	0.6569
	434-448	GDEFLDMDANGVTGP	21	0.7678
MUL_0252	121-135	VVDQIQIDPKVRALD	14	1.0375
	122-136	VDQIQIDPKVRALDF	14	1.3367
	131-145	VRALDFSNAEPVFSA	13	0.5181
	309-323	ANPIASNDTAEGRIK	14	1.2768
MUL_4591	21-25	ATVTGLQAATAAPAG	14	0.5859
	22-36	TVTGLQAATAAPAGI	15	0.5771
	23-37	VTGLQAATAAPAGIS	16	0.6895
	53-67	NALSEFEAVRARAGQ	15	0.7164
	54-68	ALSEFEAVRARAGQA	14	0.9178
	55-69	LSEFEAVRARAGQAL	14	0.7494
MUL_4987	2-16	KLVDRFRGAATGTSR	14	0.5541
	3-17	LVDRFRGAATGTSRR	15	0.6718
	4-18	VDRFRGAATGTSRRL	15	1.3548
	49-63	LPVEYLQVPSVAMGR	14	0.5470
	278-292	TSNMKFQAAYNAAGG	19	1.3548
	280-294	NMKFQAAYNAAGGHN	18	1.6168
	281-295	MKFQAAYNAAGGHNA	14	1.4559
	304-318	THSWEYWGAQLNAMR	15	0.5408
	305-319	HSWEYWGAQLNAMRP	13	0.5040
MUL_2013	10-24	PVPTFKGAGPSAAKP	15	0.8990
	11-25	VPTFKGAGPSAAKPR	15	0.9586
	105-119	SSAFRAATDSLQGIN	17	0.5609
	119-133	NRRFRYMPIAPLTPV	22	0.8870

Table 6

Overlapped MHC class I and MHC class II epitopes.

Protein	MHC Class I Epitope			MHC Class II Epitope
	S. No.	Position	Sequence	S. No.	Position	Sequence
MUL_4389	1	133-141	KISGTATGV	1	128-142	DGSSYKISGTATGVD
MUL_3082	1	55-63	TQLGRNFQV	1	58-72	GRNFQVIYEQANSHG
				2	59-73	RNFQVIYEQANSHGQ
MUL_1208	1	126-134	LVATNFFGL	1	128-142	ATNFFGLNAIPIALN
	2	317-325	EVPAAAAPL	2	311-325	PIQIAAEVPAAAAPL
MUL_4591	1	62-70	RARAGQALQ	1	53-67	NALSEFEAVRARAGQ
				2	54-68	ALSEFEAVRARAGQA
				3	55-69	LSEFEAVRARAGQAL
MUL_4987	1	10-18	AATGTSRRL	1	2-16	KLVDRFRGAATGTSR
				2	3-17	LVDRFRGAATGTSRR
				3	4-18	VDRFRGAATGTSRRL
	2	288-296	NAAGGHNAV	4	278-292	TSNMKFQAAYNAAGG
				5	280-294	NMKFQAAYNAAGGHN
				6	281-295	MKFQAAYNAAGGHNA

Table 7

Overlapped B cell and T cell epitopes.

Protein	Sequential Linear B Cell Epitope			MHC Class I Epitope		MHC Class II Epitope
	S.No.	Position	Sequence	Position	Sequence	Position	Sequence
MUL_4389	1	94-99	ANPPEV	91-99	LTDANPPEV
MUL_3082	1	65-90	YEQANSHGQKVQAAGNNMAQTDSAVG			71-85	HGQKVQAAGNNMAQT
						72-86	GQKVQAAGNNMAQTD
MUL_1208	1	170-254	SAPRTGPPPKVVNPGGGQVAHLAADAARAQSGAAAAGGSDPWQLLWQLLQYLWNAYTGFNTWMFDLIWEFLQDPIGNTIKIIIAF	207-215	GSDPWQLLW	184-198	GGGQVAHLAADAARA
						185-199	GGQVAHLAADAARAQ
						186-200	GQVAHLAADAARAQS
						187-201	QVAHLAADAARAQSG
						188-202	VAHLAADAARAQSGA
	2	278-495	FNLVGWPTWALILSSPFLLPAALGLALSAIAFTPIQIAAEVPAAAAPLAAAAVTAKSVFPAASLASPGTGSAGAPASGAGAGAPGTAPPGPGAPAPASFLYAVGGSGDWGPSLGPTVGGRSGAKAPAVTIPAAGAAAASRASARSRRRRRTEMRDYGDEFLDMDANGVTGPAATDDAARASERGAGQFGFAGTVSTEAVLQAAGLTELAVDEFGNGPR	292-300	SPFLLPAAL	304-318	LSAIAFTPIQIAAEV
				294-302	FLLPAALGL	308-322	AFTPIQIAAEVPAAA
				296-304	LPAALGLAL	309-323	FTPIQIAAEVPAAAA
				317-325	EVPAAAAPL	310-324	TPIQIAAEVPAAAAP
				407-415	IPAAGAAAA	311-325	PIQIAAEVPAAAAPL
				418-426	ASARSRRRR	434-448	GDEFLDMDANGVTGP
				419-427	SARSRRRRR
MUL_0252	1	48-91	GAYQRPTSVNGPTGELPTVTASSAKPDRPELSLSLLSVSRSGNT	70-78	SAKPDRPEL
				72-80	KPDRPELSL
	2	128-151	DPKVRALDFSNAEPVFSASGPIPD			131-145	VRALDFSNAEPVFSA
	3	306-323	LGSANPIASNDTAEGRIK	321-329	RIKNRRVEI	309-323	ANPIASNDTAEGRIK
MUL_2013	1	5-30	AVDPTPVPTFKGAGPSAAKPRLSTLQ			10-24	PVPTFKGAGPSAAKP
						11-25	VPTFKGAGPSAAKPR
	2	70-87	PWNHYVGGGASGVPSSLG	67-75	FAIPWNHYV

Table 8

Molecular docking analysis with MHC Class I and MHC Class II alleles.

Protein	Epitope Position	Epitope	MHC Class I Alleles	PDB ID (MHC Allele)	Lowest Energy Score (kJ/mol) (Cluspro Server)	Binding Affinity ΔG (kcal/ mol) (PRODIGY Server)
MUL_4389	91-99	LTDANPPEV	HLA-C*05:01	5VGD	-517.4	-13.1
MUL_1208	207-215	GSDPWQLLW	HLA-B*58:01	5VWH	-713.3	-13.5
	292-300	SPFLLPAAL	HLA-B*39:01	4O2E	-734.0	-12.2
			HLA-B*07:02	7LG0	-641.5	-13.4
			HLA-B*35:01	3OXR	-640.0	-13.3
	294-302	FLLPAALGL	HLA-A*02:06	3OXR	-709.2	-13.0
			HLA-A*02:01	4U6Y	-632.8	-11.8
	296-304	LPAALGLAL	HLA-B*07:02	7LG0	-640.2	-13.9
			HLA-B*35:01	4LNR	-608.8	-12.8
			HLA-B*39:01	4O2E	-646.7	-12.5
	317-325	EVPAAAAPL	HLA-A*68:02	4I48	-507.4	-13.1
	407-415	IPAAGAAAA	HLA-B*07:02	7LG0	-551.0	-13.8
			HLA-B*35:01	4LNR	-559.6	-13.5
MUL_0252	72-80	KPDRPELSL	HLA-B*07:02	7LG0	-504.7	-13.9
	321-329	RIKNRRVEI	HLA-B*08:01	3X13	-639.7	-14.2
MUL_2013	67-75	FAIPWNHYV	HLA-A*02:06	3OXR	-850.2	-13.4
			HLA-A*02:01	4U6Y	-820.5	-11.7
MUL_3082	71-85	HGQKVQAAGNNMAQT	HLA-DRB1*04:01	5LAX	-618.3	-19.5
	72-86	GQKVQAAGNNMAQTD			-597.1	-19.6
MUL_1208	184-198	GGGQVAHLAADAARA			-592.8	-19.4
	185-199	GGQVAHLAADAARAQ			-668.3	-19.5
	186-200	GQVAHLAADAARAQS			-649.5	-19.5
	187-201	QVAHLAADAARAQSG			-640.8	-19.5
	188-202	VAHLAADAARAQSGA			-683.1	-19.4
	304-318	LSAIAFTPIQIAAEV			-815.2	-19.6
	308-322	AFTPIQIAAEVPAAA			-708.2	-19.5
	309-323	FTPIQIAAEVPAAAA			-734.6	-19.6
	310-324	TPIQIAAEVPAAAAP			-652.0	-19.5
	311-325	PIQIAAEVPAAAAPL			-712.7	-19.4
	434-448	GDEFLDMDANGVTGP			-719.0	-19.4
MUL_0252	131-145	VRALDFSNAEPVFSA			-799.3	-19.5
	309-323	ANPIASNDTAEGRIK			-697.6	-19.5
MUL_2013	10-24	PVPTFKGAGPSAAKP			-746.4	-19.6
	11-25	VPTFKGAGPSAAKPR			-757.3	-19.5