Epitope‐based peptide vaccine design and target site depiction against Middle East Respiratory Syndrome Coronavirus: an immune-informatics study

Primary sequence of Saudi Arabia isolate MERS-COVS protein was retrieved from NCBI database using accession number ALW82742.1 [18]. Experimentally known 3D structure of MERS-COV S protein was retrieved by using PDB ID: 5X59 from Protein-Data-Bank [19]. Protein sequence was analysed for its chemicals and physical properties including GRAVY (Grand average of hydropathicity), half-life, molecular weight, stability index and amino acid atomic composition via an online tool Protparam [20]. Secondary structure of MERS-COV S protein was analysed through PSIPRED [21]. TMHMM an online tool (http://​www.​cbs.​dtu.​dk/​services/​TMHMM/​), used to examine the transmembrane topology of S protein. Existence of disulphide-bonds were examined through an online tool DIANNA v1.1. It makes prediction based on trained neural system [22]. Antigenicity testing carried out through vaxijen v2.0 [23]. Allergenicity of query sequence was checked through AllerTOP v2.0 [24].

Freely online accessible servers IEDB (Immune-Epitope-Database And Analysis-Resource) [25] and BCPRED [26] were used to for B-cell epitopes forecast. Criteria was set to have 75% specificity and 14 residue lengthy epitopes were viewed as adequate to persuade defensive immune reaction. Only those epitopes were chosen that were visible on outer surface and other intracellular epitopes were eliminated. Vaxijen 2.0 server was utilized for antigenicity study of chosen epitopes [23]. Recognition of B-cell epitopes was depended on; antigenicity, accessibility of surface, flexibility, hydrophilicity and predictions of linear epitope [27]. Hydrophilicity, isolation of linear epitope, accessibility of surface and Flexibility analysis were performed through Bepipred linear epitope prediction and Parker hydrophilicity prediction algorithms, Kolaskar and Tongaonkar antigenicity scale, Emini surface accessibility prediction tool and Karplus and Schulz flexibility prediction tool [28]. Forecast of beta turns in polyprotein was done by utilizing Chou and Fasman beta-turn prediction algorithm [29]. As the discontinuous epitopes are increasingly explicit and have higher dominant attributes over linear epitopes [30, 31], so, the forecast of discontinuous epitopes have additionally been carried out via DiscoTope server [32]. Parameter was set at ≥ 0.5 which indicated 90% specificity and 23% sensitivity. This method relies on surface accessibility and amino acid statistics in a collected form dataset of discontinuous epitopes found out by X-ray crystallography of antigen/antibody protein buildings. At last, position of predicted epitopes clusters (positional affirmation) on 3D structure of S protein was observed via PepSurf [32]. Pymol was utilized to examine the positions of forecast epitopes on the 3D structure of MERS-COV S protein [33].

Cytotoxic T-lymphocyte (CTL) epitopes play a crucial role in vaccine designation. Most significant, it decreases the cost and time as compared with wet lab experiments [34]. By utilizing two distinctive online accessible tools Propred-1 [35] and Propred tool [36], CTL epitopes of target protein of MHC class-I and MHC class-II were predicted respectively. The outcomes of these tools are quite substantial because they utilize vast number of alleles of HLAs (human-leukocyte-antigens) during computation. The sequence was given in plain format and all alleles were chosen for prediction. For propred-1 proteasome and Immuno-proteasome filters with a threshold value of 5% were kept on.

After finalizing the epitopes of both MHC class-1 and MHC class-II alleles, their important features including digestion, mutation, toxicity, allergenicity, hydro and physiochemical were checked via vaxijen 2.0 [23], protein digest server (http://​db.​Systemsbiology.​net:​8080/​proteomicsToolki​t/​proteinDigest.​html), AllergenFP 1.0 [37] server, Aller Hunter server (https://​omictools.​com/​allerhunter-tool) and ToxinPred server (http://​crdd.​osdd.​net/​raghava/​toxinpred/​). AllergenFP 1.0 is generally utilized for the prediction of allergenicity of epitopes for vaccine development [37]. Aller Hunter server compares peptide’s query sequences opposed to the database of previously reported allergens to give significant outcomes. An in silico method, ToxinPred is used to predict Non-Toxic/Toxic peptides. For further analysis, only NonToxic epitopes were chosen.

S protein sequences of 8 distinctive countries were taken from an open access Genbank database [38]. By utilizing CLC work bench, the multiple-sequence-alignment (MSA) was carried out to perceive the conservation of chosen epitopes [39]. The aligned files (.aln) were additionally utilized to make phylogenetic tree via MEGA7 software [40]. By analysing the multiple-sequence-alignment results and with IEDB conservation-analysis-tool, all the chosen epitopes were checked for their variability and conservation.

All the predicted peptides 3D structures were modelled via PEPFOLD server at RPBS MOBYL portal [41], from Protein databank (PDB ID: 3VCL) at a resolution of 1.7 Å, the 3D structure of human HLA-B7 allele crystallized was taken [42] and utilized for further molecular docking purpose. Through Molecular Operating Environment (MOE) tool, the peptide models (antigenic determinants) were docked against their respective HLA-B7 allele to analyse their inhibitory potential. Procedure for molecular docking using MOE has already been described in various studies [13, 43, 44]. Docking procedure utilized in those studies include protonation, expulsion of already bound peptide and energy reduction followed by expulsion of water particles. Triangular matcher algorithm was applied as default peptide placement methods dependent on the receptor shape which without energy optimization rapidly produces 1000 best poses of docked peptide [13]. By applying London-dG scoring function, the energy approximation of the imitated poses was rescored. For every peptide, top ten positioned poses of London-dG were additionally reduced by Force field refinement algorithm. Protein peptide connection were than examined via LigX tool of MOE. UCSF Chimera and Pymol tools were utilized to produce figures of docked complexes [33, 45].

The physiochemical properties of MERS-COV S protein computed via protparam demonstrates that it contained 1353 amino acids (aa) with molecular weight of 149,479.23 kDa, which reflects good antigenic nature. Theoretical isoelectric point (PI) of subject protein was 5.80 which indicate its negative in nature. An isoelectric point under 7 shows negatively charged protein. Briefly, out of 1353 residue, 112 aa were found as negatively charged whereas others found as positively charged. Protparam computed instability-index (II) 36.81, this categories protein as stable. Aliphatic-index 82.79, which devotes a thought of proportional volume hold by aliphatic side chain and GRAVY value for protein sequence is 0.078. Half-life of protein depicted as the total time taken for its vanishing after it has been synthesized in cell, which was computed as 30 h for mammalian-reticulocytes, > 20 h for yeast, > 10 h for Escherichia coli. Total number of Carbon (C), Oxygen (O), Nitrogen (N), Hydrogen (H) and Sulfur (S) were entitled by formulaC₆₆₈₇H₁₀₂₅₈N₁₇₄₀O₂₀₂₇S₆₃. Protparam computed details of physiochemical properties enlisted in Additional file 2: Table S1.

Secondary and 3-D structure examination of S protein via PSIPRED [21], UCSF Chimera [45] and Pymol [33] respectively showed that (50%) Beta sheets, (10%) Helixes and (40%) Loops are present in structure as shown in Additional file 1: Figure S1. Two different conformations of structure of MERS-COVS protein shown in Additional file 1: Figure S2.

Furthermore in target protein, DiANNA1.1 tool [22] calculated 21 disulphides bond (S–S) positions and assign them a score given in Additional file 2: Table S2. Antigenicity of protein was evaluated via Vaxijen 2.0 [23] by setting the threshold at ≥ 0.5, for higher specificity. Antigenicity analysis of full-length protein showed antigenicity 0.4808 for S protein showing it as an expected antigen. An online tool TMHMM used to checked the transmembrane protein topology (http://​www.​cbs.​dtu.​dk/​services/​TMHMM/​) and it was found that residue from 1 to 1295 were exposed on the surface, while residue from 1296 to 1318 were inside transmembrane-region and residues from 1319 to 1353 were buried within the core-region of the S protein.

B-cell epitopes are significant for defence against viral disease. Potential B-cell epitopes have different features that direct B-cell to recognize and activate the rich defence responses against distinct viral infection. Primary sequence of S protein was scanned via IEDB server [25] and BCPRED [26] to predict B-cell epitopes. Total 59 B-cell epitopes were predicted. From all predicted epitopes, just6 epitopes (Table 1) were selected which were exposed on the surface of S protein and have high antigenicity score. Vaxijen 2.0 was used to compute antigenicity score and TMHMM server was utilized to check the surface availability. Among these selected epitopes, ‘QLQMGFGITVQYGT’ predicted at position 566 showed highest antigenicity and predicted scores.

Moreover, it is essential to check out the surface availability of possible B-cell epitopes. Kolaskar and Tongaonkar antigenicity measurement tools analysed the S protein for prediction of B-cell epitopes by assessing the physiochemical properties of the amino acid and their abundance in already known B-cell epitopes. Higher antigenicity score has proposed that it can play a vital role in starting of immune response. The threshold value of tool was adjusted at 1.045 and window size was kept 7. It estimated the antigenic tendency value of protein 1.045 (average), 0.872 (minimum) and 1.258 (maximum). Result of kolaskar and Tongaonkar analysis are shown in Fig. 2a. Hydrophilic region of protein is generally uncovered on the surface and play a significant part in eliciting the immune response. BCPRED-score and calculated antigenicity outcomes of vaxijen surely manifest that all predicted peptides are part of extracellular area of transmembrane-protein and capable to maximize a defence response inside the host during MERS-COV infection. Therefore, to find the surface availability of possible B-cell epitopes and hydrophilicity, parker-hydrophilicity with threshold value 1.279 and Emini surface accessibility prediction tools with threshold value 1.000 were utilized. The visual representation of outcome of both tools is shown in Fig. 2b, c respectively. Values calculated by both these tools were 1.279 (average), − 8.486 (minimum), 6.543 (maximum); and 1.000 (average), 0.033 (minimum), 7.392 (maximum), respectively. Emini surface accessibility analysing tool’s outcomes are given in additional file 2: table S3. Chou and Fasman beta turn analysing algorithm was utilized to predict beta-turn in S protein because beta turn is exposed on the surface and hydrophilic in nature and play a vital role in starting the defence response. Tool’s threshold was adjusted at 1.009, it computed the values which are 1.009 (average), 0.581 (minimum), and 1.414 (maximum). Chou and fasman’s result’s graphical representation is shown in Fig. 2d. The result indicates that region from 213 to 220 amino acid and from 641 to 650 are more disposed to persuade Bturns in peptide structure. It is described by an experimental information that the parts of epitope which connect with antibodies or alleles are mainly elastic in nature. Karplus and schulz flexibility analysing tool represented that the area from amino acid from 854 to 860 sequence positions are highly versatile as shown in Fig. 2e. Position of every predicted epitope on surface of 3-D structure of S protein was confirmed by Pepsurf [32] and shown in Fig. 3 using Pymol [33].

To further increase the specificity and range of B-cell epitopes, Discotope 2.0 server was used which calculate surface availability in term of residue contact number and novel tendency amino acid score was utilized to predict the discontinuous epitopes. 3D structure of S protein (PDB ID: 5X59) [19] was used for discontinuous epitopes prediction, 90% specificity, − 3.700 threshold and 22.000 Angstroms propensity score radius. Total 22 discontinuous epitopes were calculated at different exposed surface areas (Table 2). Position of each predicted epitope on surface of 3D structure of S protein shown in Fig. 4 using Pymol [33].

Propred-I (47 MHC class-I alleles) [35] and Propred (51 MHC class-II alleles) [36] were utilized for prediction of T-cell epitopes for the S protein. Propred-I utilizes a matrix base approach to scan and predict the peptides against library of 47 MHC class-1 alleles. The S protein sequence in FASTA format was transferred to the propred-I server, whereas choosing all the alleles with higher scoring peptide with 4% threshold and keeping the proteasome filter and immune proteasome filter at on mode. Additionally, antigenicity testing and screening of peptides were finished with assistance of vaxijen 2.0 [23]. Just 6 potential peptides were chosen for next processing on the basis of their antigenicity-score (Table 3). A peptide which has capacity to attach with larger number of alleles is observed as most important peptide due to its potential to bring a powerful defense response. Between MHC class-I predicted epitopes, the peptide ‘YKLQPLTFL’ indicated higher antigenicity score 1.5335 attaching with number of alleles including MHC-Db, HLA-Cw*0301, HLA-B*51, HLA-B*5401, HLA-B*5301, HLA-B*3902, HLA-B*3901, HLA-B*3701, HLA-B7, HLA-B14, HLA-A2.1, HLA-A20 Cattle, HLA-A2 and HLA-A*0201.

Propred, a quantitative matrix base method was used for prediction of peptides, which can interact with MHC class-II alleles. Sequence was given in FASTA format to Propred. Screening was done with the help of vaxijen 2.0 and just 6 high scoring epitopes were chosen (Table 4). The peptide ‘YCILEPRSG’ was considered more antigenic for its higher antigenicity score 1.7889 and it demonstrated virtual attachment with larger number of alleles (almost 15) including, DRB5_0105, DRB5_0101, DRB1_1328, DRB1_1327, DRB1_1323, DRB1_1307, DRB1_1305, DRB1_1302, DRB1_1301, DRB1_1128, DRB1_1120, DRB1_1114, DRB1_1101, DRB1_0802 and DRB1_0101.

Some important features of selected epitopes were analysed to support our findings. The peptides that can be digested by several enzymes are usually non-stable. On the other hand, peptides digested by fewer enzymes are highly stable and more favourable vaccine candidates. Peptides digesting enzymes were predicted through Protein digest server. Allergen FP 1.0 was used for allergenicity prediction of epitopes. ToxinPred was utilized for toxicity prediction of chosen epitopes. Toxinpred is based on support vector machine (SVM) used to predict toxicity along with mutations, hydropathicity, hydrophilicity, hydrophobicity, and charge. All T-cell epitopes along with their digestion, mutation, toxicity, allergenicity, hydro and physiochemical results are given in Table 5.

Sequence of MERS-COV S protein from 8 different countries isolates including Saudi Arabia (ALW82742.1), Abu Dhabi (ASU90340.1), Jordan (ASY99842.1), Qatar (AHX71946.1), South Korea (AKL59401.1), Thailand (ALD51904.1), United Kingdom (AJD81440.1) and United State (AHZ58501.1) were subjected to multiple-sequence-alignment through CLC workbench to analyse the conservation of chosen epitopes. It was noticed that all the chosen epitopes are mostly conserved in all sequences utilized for analysis as shown in Additional file 1: Figure S3. A phylogenetic tree was created to indicate the evolutionary relationship of MERS-COV of 8 distinct countries as shown in Fig. 5.

The epitope-conservancy study through IEDB epitope conservancy analysis tool shows that all of selected B-cell and T-cell (MHC class-I and II) epitopes have 100% identity and conserved in all isolates of distinct countries (Additional file 2: Table S4).

3D structures of all 6 MHC class-I attaching peptides were predicted via PEPFOLD [41]. It created 5 models of every peptide; one best model was chosen for every peptide (Additional file 1: Figure S4). At first models were refined via energy minimization in MOE and peptide library involved of 6 peptides was made to dock with explained structure of HLA-B7 allele.

Crystal structure of human HLA-B7 (PDB ID: 3VCL) protein was previously accessible with co-crystallized peptide in PDB [42]. So, rigid/focused docking was performed by utilizing same active pocket to dock our peptide library. 10 confirmations for every epitope were produced and top positioned conformations dependent on their dock scores and interactions with HLA-B7 residues were chosen (Table 6). Afterward, interaction examination by ligX tool of MOE was done (additional file 1: figure S5) which displayed that the peptide ‘AGYKVLPPL’ with highest dock score (-20.9793 kcal/mol) is connecting with key catalytic residues. Human HLA-B7 is a hetero-dimer structure, from the interaction analysis it was showed that Asp-114, Gln-115, Lys-146, Glu-152 and Arg-156 from A chains were making stable hydrogen bonds with the previously mentioned peptide (Fig. 6a). Peptide ‘WPRPIDVSK’ was docked (dock score -20.4007 kcal/mol) inside the catalytic pocket of receptor protein through 4 hydrogen bonds with Arg-62, Glu-152, Glu-163 and Trp-167 (Fig. 6b). Peptide ‘ESAALSAQL’ has -19.9914 kcal/mol of dock score with 5 stable hydrogen bonds between peptide and Arg-62, Asn-63, Gln-70, Glu-152 and Gln-155 (Fig. 6c). Similarly, other peptides also show strong and stable bonding with human HLA-B7 residues and shown in Table 6 and Fig. 6d–f.