Multi-epitope vaccine construct against bovine tuberculosis: insights from immunoinformatics and molecular dynamics simulations

Present Address:

Truc Ly Nguyen 1,2✉ Emailnttly@uhsvnu.edu.vn

1 University of Health Sciences, Vietnam National University Ho Chi Minh City Ho Chi Minh City Vietnam

2 Vietnam National University Ho Chi Minh City Ho Chi Minh City Vietnam

Truc Ly Nguyen^1,2*

¹ University of Health Sciences, Vietnam National University Ho Chi Minh City, Ho Chi Minh City, Vietnam

² Vietnam National University Ho Chi Minh City, Ho Chi Minh City, Vietnam

*Corresponding author, E-mail: nttly@uhsvnu.edu.vn, ORCID: 0000-0001-8843-1072

Multi-epitope vaccine construct against bovine tuberculosis: insights from immunoinformatics and molecular dynamics simulations

Abstract

Bovine tuberculosis is a chronic zoonotic disease that continues to threaten the livestock industry worldwide, particularly in developing countries. Despite extensive control efforts, there is still no licensed vaccine available for effective prevention of bTB. This study aimed to design a multi-epitope vaccine candidate against Mycobacterium bovis using comprehensive immunoinformatics and molecular dynamics approaches. Eighteen high-confidence epitopes were predicted from antigenic bovine tuberculosis proteins through machine learning–based algorithms and subsequently assembled into a single vaccine construct using GPGPG linkers and an HSP70 adjuvant to enhance immunogenicity. The designed multi-epitope vaccine was predicted to be non-toxic and non-allergenic, as validated using CSM-Toxin and AllerTOP v2.0 servers. The tertiary structure of the vaccine was modeled and docked with Toll-like receptor 4 to assess molecular interactions and stability. Molecular dynamics simulations for 100 ns at 300 K revealed a stable vaccine–receptor complex with consistent hydrogen bonding and low RMSD fluctuations in a solvated environment. Overall, this study provides theoretical evidence supporting the rational design of a safe, stable, and potentially immunogenic multi-epitope vaccine candidate for bovine tuberculosis, warranting further in vitro and in vivo validation.

Keywords:

Bovine tuberculosis

Multi-epitope vaccine

Molecular dynamics simulation

Immunoinformatics

Computational vaccinology

1. Introduction

Bovine tuberculosis (bTB) is one of the deadliest infectious diseases globally, causing an estimated 1.2 million deaths (range: 1.1–1.3 million) in 2019 alone. Of these deaths, approximately 11,400 (range: 4,470–21,600) were attributed to Mycobacterium bovis [1], a zoonotic pathogen that infects a variety of wild animals, including badgers, deer, hares, swine, and nonhuman primates[2, 3]. While the Bacille Calmette-Guérin (BCG) vaccine has been the only licensed vaccine for tuberculosis (TB) in humans since 1921[4], there is still no vaccine available for bTB. Although the BCG vaccine was reported to provide protection against bTB in cattle a decade before human trials[5], subsequent field studies showed variable efficacy, ranging from 0% to 80%[6]. Furthermore, the field use of BCG complicates infection control programs, as vaccinated animals may test positive in purified protein derivative (PPD) assays, which are used to differentiate infected from vaccinated animals (DIVA)[7]. Attempts have been made to use major antigenic targets of Mycobacterium bovis that are deleted in BCG as new DIVA candidates[8, 9]; however, the absence of these antigenic targets reduces the overall efficacy of TB vaccines[10, 11].

With advancements in computational resources, multi-epitope vaccines (MEV) present a promising solution to these challenges. MEV is an in silico process that uses the genetic information of a pathogen for vaccine discovery [12]. Compared to traditional live attenuated vaccines, subunit vaccines developed via MEV are faster to produce, more stable, and associated with fewer side effects[13]. This approach has been used to target numerous pathogens[14–17] including Mycobacterium Ulcerans[18], Mycobacterium Avium[19], and Mycobacterium tuberculosis[20], although further validation is still needed. The first successful application of MEV was a vaccine against invasive meningococcal disease [21], and recent MEV-based mRNA vaccines against COVID-19 have saved millions of lives[22].

As bTB continues to pose a serious threat to both public health and the livestock industry, resulting in considerable economic losses worldwide, the development of an effective vaccine remains a critical priority. Despite the global implementation of control measures and the availability of the Bacillus Calmette–Guérin (BCG) vaccine, its efficacy in cattle remains inconsistent, and it does not fully prevent infection or transmission. Therefore, novel vaccination strategies are urgently needed to complement existing control programs. In this study, we aimed to design a MEV candidate for bTB using in silico immunoinformatics approaches. The Mycobacterium bovis strain AF2122/97, which is frequently used as a reference genome in tuberculosis and bovine tuberculosis vaccine design studies [23–25], was employed as the source for epitope prediction and vaccine construction.

2. Materials and Methods

2.1. Protein sequence retrieval

Epitope prediction and the analyses throughout the study were conducted on Outer membrane protein A (OmpA), 6 kDa early secretory antigenic target (Esat-6), and Esat-6 like protein EsxB (Cfp-10). Outer membrane protein A (OmpA) is a highly conserved and abundant virulence factor in Gram-negative bacteria[26]. Esat-6 and Cfp-10 are two of the major antigenic targets of the RD1 region, which is absent in BCG strains. RD1 region is responsible for the virulence in Mycobacterium bovis[27] and showed promising immunological responses in animal and clinical trials[28–32]. The protein sequences of bTB were obtained from Uniprot with the following entry identifiers: A0A679LDF2 (OmpA), P0A565 (Esat-6), P0A567 (Cfp-10), P0A5C0 (Hsp70).

2.2. Epitopes prediction

The protein sequences were submitted to NetMHCpan v.4.1 (https://services.healthtech.dtu.dk/services/NetMHCpan-4.1/) and NetMHCIIpan v.4.3 (https://services.healthtech.dtu.dk/services/NetMHCIIpan-4.3/) to predict MHC-Ⅰ and MHC-Ⅱ epitopes respectively. All the bovine MHC alleles in each version were used, which is 105 alleles for BoLA in NetMHCpan v.4.1 and 365 alleles for BoLa-DRB3 in NetMHCIIpan v.4.3. Epitopes with peptide length of 9 were chosen for MHC-Ⅰ and 15mer was used for MHC-Ⅱ. Only the peptides with BindLevel of Weak binding (WB) and Strong binding (SB) were sorted by appearance counts in each MHC type (Supplementary Table S1). B cell epitopes were predicted by ABCpred (https://webs.iiitd.edu.in/raghava/abcpred/) [33] with a default threshold of 0.51 and peptide length of 16. Non-overlapping peptides were kept for further analysis.

2.3. Construction of vaccine candidate sequence

The top two non-overlapping epitope sequences were chosen by the appearance count for T cell epitopes and probability score for B cell epitopes. Epitopes were joined with the adjuvant protein HSP70 through the GPGPG linker. HSP70 is a highly conserved TLR4 agonist[34] that activates the innate immune system, especially the T helper cell response in TB and bTB [35–37].

2.4. Allergenicity, antigenicity, toxicity, and physicochemical properties prediction of vaccine candidate

As high allergenicity and toxicity can cause adverse health issues, the vaccine candidate sequence was tested with AllerTop 2.0 server (https://www.ddg-pharmfac.net/AllerTOP/index.html) and CSM-Toxin (https://biosig.lab.uq.edu.au/csm_toxin/) respectively. To predict the antigenicity, VaxiJen v.2.0 (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html) was used with a threshold of 0.4.

To improve the efficacy of the vaccine candidate, solubility was evaluated with SOLpro (https://scratch.proteomics.ics.uci.edu/) [38] and the physiochemical properties were estimated by the ProtParam server (https://web.expasy.org/protparam/) [39].

2.6. Vaccine’s three-dimensional structure prediction, refinement, and validation

For the 3D structure prediction, AlphaFold2 [40] was used on the Neurosnap server (https://neurosnap.ai/service/AlphaFold2?gclid=CjwKCAiA5L2tBhBTEiwAdSxJX1F8veoqgaIyza6QseaxhBJuD8pxeh3Xtdkkle9TjIi_-mOaZ9TkKhoCYN4QAvD_BwE) with the default settings. AlphaFold2 showed remarkable results in the 14th critical Assessment of protein Structure prediction (CASP). The rank1 model was then refined with GalaxyRefine (http://galaxy.seoklab.org/refine) to increase the accuracy of the ab initio prediction method[41]. The validity of the final model was examined by ProSA-Web (https://prosa.services.came.sbg.ac.at/prosa.php) and Molprobity v4.4 (https://swissmodel.expasy.org/assess).

2.7. Molecular docking studies between the vaccine and immune receptor

Toll-like receptor 4 (TLR4) 3D structure of Bos taurus from UniProt (Entry identifier: Q9GL65) was used as a docking target for the vaccine candidate. Toll-like receptor 4 is known to activate the host immune system by the recognition of our adjuvant HSP70 and other antigenic proteins of TB and bTB [36, 42–45]. Docking prediction was conducted with Cluspro (https://cluspro.bu.edu/login.php), which presents the energy-minimizing complex model by rotating the ligand against the receptor. The model with the highest number of cluster members for each option was considered the best-fitting model [46]. The interactions between the molecules after docking were shown with PDBsum (https://www.ebi.ac.uk/thornton-srv/databases/pdbsum/) [47].

2.8. Molecular dynamics simulations of the vaccine-receptor complex

Molecular dynamics (MD) simulation was conducted to investigate the stability of the vaccine-receptor complex using GROMACS 2023 software on a Linux operating system, which provides real-life environmental conditions for various biological models[48]. The OPLSAA force field and the SPC/E water model were employed to create of topology file. The MEV-TLR4 complex was then enclosed in a cubic box to remain intact, surrounded by 599,883 solvent molecules. Subsequently, to neutralize the charge, 43 Na⁺ ions were added. The energy minimization was performed using the steepest descent algorithm with 50,000 steps, and the minimization process ceased when the maximum force reached < 1000.0 kJ/mol/nm. Following that, position restraints were applied during the equilibration process. NVT equilibration was executed at 300 K with 50,000 steps (100 ps), followed by NPT equilibration at 1 bar reference pressure with an additional 50,000 steps (100 ps). Afterward, production simulations for all-atom (1,826,988 atoms) were achieved using the NPT ensemble for 50,000,000 steps (100 ns). After the efficient completion of a 100 ns MD simulation, calculations were performed for the root mean square deviation (RMSD) of backbone residues, root mean square fluctuation (RMSF) of C-alpha, radius of gyration_total (Rg), and solvent accessible surface area (SASA).

Additionally, to evaluate the stability of the complex as a whole during the simulation, we calculated the RMSD of the MEV-TLR4 complex at various time steps. By comparing the stability of the MEV-TLR4 complex, we were able to determine their respective stability. Afterward, the COCOMAPS tool[49] was used to conduct a thorough study and visualization of the contacts at the MEV-TLR4 complex interface.

3. Results

3.1. Epitope screening and vaccine candidate construction

Total of 2,456 (OmpA: 1,462, Esat-6: 440, Cfp-10: 554) and 11,289 (OmpA: 8,357, Esat-6: 1,584, Cfp-10: 1,348) potential epitopes were predicted by NetMHCpan v.4.1 and NetMHCIIpan v.4.3 respectively. After sorting the epitopes by the appearance count, the top two non-overlapping epitopes were selected from each protein (Supplementary Data Sheets) for vaccine construction. From the total of 39 (OmpA: 24, Esat-6: 7, Cfp-10: 8) B cell epitopes sorted by the probability score of ABCpred, the top two non-overlapping epitopes were merged with the selected T cell epitopes, HSP70 adjuvant and GPGPG linkers (Fig. 1).

Fig. 1

The sequential overview of the vaccine construct.

Figure 1 The sequential overview of the vaccine construct.

3.2. Allergenicity, antigenicity, and toxicity prediction

The merged sequence was tested for allergenicity with the k nearest neighbours (kNN) based classifier AllerTop 2.0 server[50]. It was predicted to be a non-allergen with ARID1B (UniprotKB accession number: Q8NFD5) being the nearest protein. There was no sign of toxicity according to the Bidirectional Encoder Representations from Transformers (BERT) based CSM-Toxin[51]. Protective antigenicity was expected with a prediction value of 0.7838 using VaxiJen v2.0. Vaxijen classifies probable antigens by discriminant analysis using the partial least squares technique (DA-PLS)[52]. Considering these aspects as a safe antigen, our sequence complex was kept for further analysis.

3.3. Physiochemical properties and solubility prediction

Physiochemical properties estimated by ProtParam gave insights into the vaccine candidate’s composition, stability, and hydropathicity (Table 1). The approximately 98kDA complex was estimated to have an instability index of 28.21, and the grand average of hydropathicity (GRAVY) index of -0.396. The probability of predicted solubility upon overexpression was 0.985522 according to SOLpro (http://scratch.proteomics.ics.uci.edu/).

Table 1

Physiochemical properties and overall traits of the vaccine construct.
Properties	Values
Number of amino acids	955
Molecular weight	98573.67 Da
Theoretical pI	4.91
Total number of negatively charged residues (Asp + Glu)	122
Total number of positively charged residues (Arg + Lys)	87
Formula	C₄₂₆₆H₆₈₇₂N₁₂₄₈O₁₄₁₁S₁₁
Total number of atoms	13,808
Estimated half-life (mammalian reticulocytes, in vitro)	30 hours
Estimated half-life (yeast, in vivo)	> 20 hours
Estimated half-life (Escherichia coli, in vivo)	> 10 hours
Instability index	28.21 (stable)
Aliphatic index	79.30
Grand average of hydropathicity (GRAVY)	-0.396 (hydrophilic)
Solubility	0.985522 (probability)
Allergenicity	Non-allergen
Antigenicity	Probable antigen
Toxicity	Non-toxic

3.4. Validation of vaccine structure prediction

The resulting models from AlphaFold2 prediction showed a mean predicted Local Distance Difference Test (pLDDT) score of 64.73 to 69.57 (Table S1). The Rank1 model with the highest pLDDT score was then refined with GalaxyRefine. Model 1 was the most improved model after refinement as stated by the lowest GDT-HA, RMSD, and less problematic values (Table S2). The validity of model 1 was inspected with ProSA-Web and Molprobity. Z-score value of the model was − 13.35, which is normal compared to other X-ray and NMR results (Fig. 2a). The residues within the Ramachandran favoured region were expected to be 97.59% (Fig. 2b).

Fig. 2

Validation of vaccine structure. (a) ProSA-Web validation; (b) Molprobity Ramachandran plot.

Figure 2 Validation of vaccine structure. (a) ProSA-Web validation; (b) Molprobity Ramachandran plot.

3.5. Molecular docking of the vaccine and TLR4 receptor

From the molecular docking results of Cluspro (Fig. 3a), the model with the highest number of cluster members for the Van der Waals + electrostatics option was chosen regarding our input structures. Our selected model had 33 members in the cluster with a -268.4 weighted energy score calculated by the PIPER algorithm (Table S3). There were 13 hydrogen bonds and 10 salt bridges after docking (Fig. 3b). The resulting model was used as an input for molecular dynamics simulation.

Fig. 3

Molecular docking of the vaccine and TLR4 receptor. (a) Docking complex; (b) PDBsum summary of the residue interactions between the vaccine and TLR4 receptor.

Figure 3 Molecular docking of the vaccine and TLR4 receptor. (a) Docking complex; (b) PDBsum summary of the residue interactions between the vaccine and TLR4 receptor.

3.6. Molecular dynamics simulation studies

The RMSD of backbone residues over the course of a 100 ns trajectory was used to assess the lifetime of protein-protein complexes resulting in an average RMSD of 2.56 ± 0.62 nm (Fig. 4a). Based on the RMSF analysis, residues of MEV showed considerable flexibility, with fluctuations of about 1.08 ± 0.44 nm, as shown in (Fig. 4b). With an average of 6.90 ± 0.31 nm, the Rg analysis results, as shown in (Fig. 4c), indicate that the MEV-TLR4 complex maintains a slight stability. Furthermore, (Fig. 4d) shows the average SASA value of MEV-TLR4 complex was 949.37 ± 52.20 nm².

Fig. 4

Molecular dynamics simulation studies of docking complex. (a) Root mean square deviation (RMSD); (b) Root mean square fluctuation (RMSF); (c) Radius of gyration (Rg); (d) Solvent-accessible surface area (SASA).

Figure 4 Molecular dynamics simulation studies of docking complex. (a) Root mean square deviation (RMSD); (b) Root mean square fluctuation (RMSF); (c) Radius of gyration (Rg); (d) Solvent-accessible surface area (SASA).

Furthermore, by using intermolecular contact maps to find hot spot residues, COCOMAPS makes it possible to analyze and visualize the interface of interaction in protein complexes. While the results of RMSD and interaction map at various time steps show an interaction pattern with the particular contacts maintained (Figs. 5a,b), it is clear that overall contacts remained steady throughout the simulation time. All the above analyses indicate the stability of interface interactions between the vaccine construct and TLR4.

Fig. 5

(a) Superimposition of the vaccine-TLR4 complex and it’s respective RMSD of different timelines; (b) COCOMAPS contact map shows residue distances for each timeline. The dots in red, yellow, green and blue represent the residue distances below 7, 10, 13 and 16 angstrom.

Figure 5 (a) Superimposition of the vaccine-TLR4 complex and it’s respective RMSD of different timelines; (b) COCOMAPS contact map shows residue distances for each timeline. The dots in red, yellow, green and blue represent the residue distances below 7, 10, 13 and 16 angstrom.

4. Discussion

Currently, there are no available vaccines for bTB due to challenges such as differentiating infected from vaccinated animals, as well as the low protective efficacy resulting from the lack of virulence regions in the tested strains. As such, designing an in silico-based MEV could be a cost-efficient approach, either as a substitute vaccine or to complement existing vaccines.

In this study, we predicted the epitope regions of the major antigens ESAT-6, CFP-10, and OmpA from the Mycobacterium bovis strain AF2122/97 and combined them to create a stable vaccine candidate. Non-overlapping sequences predicted across multiple MHC alleles were used as epitopes to enhance vaccine coverage. Only epitopes with peptide lengths of 9 and 15, which are known to be optimal for MHCⅠ and MHCⅡ binding, respectively, were selected. Given the critical importance of vaccine safety, the vaccine candidate sequence was tested for toxicity and allergenicity, and it passed both assessments.

The pLDDT score of the best model from AlphaFold2's 3D structure prediction of the vaccine was 69.57, which is slightly below the confidence threshold of 70. Since side-chain structures are challenging to predict using ab initio methods, side-chain refinement was performed to enhance the accuracy of the vaccine's 3D structure. The refined structure exhibited improved side-chain folding and fell well within normal parameters when evaluated using ProSA-Web and Molprobity.

To assess the dynamic stability of the protein-protein binding complex, a 100 ns molecular dynamics (MD) simulation was conducted. The main goal was to measure the difference between the protein’s backbone conformation at the start and end of the simulation. The dynamical stability of a biological molecule can be inferred from the observed variations over the simulation period. A stable complex is typically indicated by minimal divergence between the initial and final states. Additionally, RMSFs of C-alpha were calculated to provide insights into the dynamics-function relationship, derived from the divergence of protein motions during evolution. Residual flexibility or rigidity is crucial for biological processes such as molecular recognition, binding/unbinding, and catalysis. A region with a high RMSF score is considered more flexible, whereas regions with lower RMSF scores show minimal movement during the experiment.

According to the RMSD and RMSF results, residues in the TLR4 portion exhibited minor fluctuations during the simulation, but they quickly stabilized. Following this, the radius of gyration (Rg) was calculated as a simulation-based metric for structural compactness. Rg, which represents the root mean square distance between the center of mass of each atom cluster, was plotted against time to show the degree of protein compactness during the MD simulation. Variable Rg values indicate protein unfolding or folding, while constant values suggest no significant change in protein folding. Solvent-accessible surface area (SASA) values gradually decreased throughout the simulation, indicating that the complexes became more compact or less solvent-exposed. Significant structural rearrangement likely occurred within the first 40 ns of the simulation, as evidenced by the rapid decrease in SASA during this period. After this, the structure remained stable for the remainder of the simulation.

5. Conclusion

This study employed an immunoinformatics-driven pipeline to design a safe, stable, and potentially immunogenic MEV candidate against bTB. Eighteen highly antigenic epitopes were selected from Mycobacterium bovis proteins and assembled into a single construct incorporating a bTB-specific HSP70 adjuvant and GPGPG linkers to enhance both T-cell and B-cell responses. The designed MEV was predicted to be non-toxic, non-allergenic, and physiochemically stable based on multiple bioinformatics evaluations. Molecular docking and molecular dynamics simulations further confirmed a strong and stable interaction between the vaccine construct and the TLR4, indicating its potential to elicit robust immune activation. Overall, the findings provide theoretical evidence supporting the rational design of an effective MEV candidate against bTB. However, comprehensive in vitro and in vivo validations remain essential to confirm its immunogenicity and protective efficacy before clinical or veterinary application.

Acknowledgement

The authors thank the laboratory of Bioinformatics and Population genetics, Seoul National University, for providing the computing infrastructure to implement and execute the Molecular dynamics simulations.

Funding

No funding was received to assist with the preparation of this manuscript.

CRediT authorship contribution statement

T.L.N

Conceptualization, Methodology, Formal analysis, Validation, Visualization, Writing – original draft, Writing – review & editing, Supervision. All authors have read and agreed to the published version of the manuscript.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Electronic Supplementary Material

Below is the link to the electronic supplementary material

Supplementary Material 1

Supplementary Material 2

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Data Sheets

Supplementary Information

Data availability

All data generated and analyzed during this study are included in this article.

Author Contribution

T.L.N : Conceptualization, Methodology, Formal analysis, Validation, Visualization, Writing – original draft, Writing – review & editing, Supervision. All authors have read and agreed to the published version of the manuscript.

References

1. Global tuberculosis report 2020. (n.d.). Retrieved 8 January 2024, from https://www.who.int/publications/i/item/9789240013131

2. LoBue, P. A., Enarson, D. A., & Thoen, C. O. (n.d.). Tuberculosis in humans and animals: an overview [Serialised article. Tuberculosis: a re-emerging disease in animals and humans. Number 1 in the series].

3. Buddle, B. M., Skinner, M. A., & Chambers, M. A. (2000). Immunological approaches to the control of tuberculosis in wildlife reservoirs. Veterinary Immunology and Immunopathology, 74(1–2), 1–16. https://doi.org/10.1016/S0165-2427(00)00163-X

4. Lobo, N., Brooks, N. A., Zlotta, A. R., Cirillo, J. D., Boorjian, S., Black, P. C., … Kamat, A. M. (2021). 100 years of Bacillus Calmette–Guérin immunotherapy: from cattle to COVID-19. Nature Reviews Urology 2021 18:10, 18(10), 611–622. https://doi.org/10.1038/s41585-021-00481-1

5. Calmette, A., & Guérin, C. (1911). Recherches expérimentales sur la défense de l’organisme contre l’infection tuberculeuse. Ann Inst Pasteur, 25, 625–641.

6. Waters, W. R., Palmer, M. V., Buddle, B. M., & Vordermeier, H. M. (2012). Bovine tuberculosis vaccine research: Historical perspectives and recent advances. Vaccine, 30(16), 2611–2622. https://doi.org/10.1016/J.VACCINE.2012.02.018

7. Vordermeier, H. M., Jones, G. J., Buddle, B. M., Hewinson, R. G., & Villarreal-Ramos, B. (2016). Bovine Tuberculosis in Cattle: Vaccines, DIVA Tests, and Host Biomarker Discovery*. https://doi.org/10.1146/annurev-animal-021815-111311, 4, 87–109. https://doi.org/10.1146/ANNUREV-ANIMAL-021815-111311

8. Buddle, B. M., Parlane, N. A., Keen, D. L., Aldwell, F. E., Pollock, J. M., Lightbody, K., & Andersen, P. (1999). Differentiation between Mycobacterium bovis BCG-Vaccinated and M. bovis-Infected Cattle by Using Recombinant Mycobacterial Antigens. Clinical and Diagnostic Laboratory Immunology, 6(1), 1. https://doi.org/10.1128/CDLI.6.1.1-5.1999

9. Vordermeier, H. M., Cockle, P. C., Whelan, A., Rhodes, S., Palmer, N., Bakker, D., & Hewinson, R. G. (1999). Development of Diagnostic Reagents To Differentiate between Mycobacterium bovis BCG Vaccination and M. bovis Infection in Cattle. Clinical and Diagnostic Laboratory Immunology, 6(5), 675. https://doi.org/10.1128/CDLI.6.5.675-682.1999

10. Asadian, M., Hassanzadeh, S. M., Safarchi, A., & Douraghi, M. (2022). Genomic characteristics of two most widely used BCG vaccine strains: Danish 1331 and Pasteur 1173P2. BMC Genomics, 23(1), 1–13. https://doi.org/10.1186/S12864-022-08826-9/TABLES/3

11. Aguilo, N., Gonzalo-Asensio, J., Alvarez-Arguedas, S., Marinova, D., Gomez, A. B., Uranga, S., … Martin, C. (2017). Reactogenicity to major tuberculosis antigens absent in BCG is linked to improved protection against Mycobacterium tuberculosis. Nature Communications, 8. https://doi.org/10.1038/NCOMMS16085

12. Rappuoli, R. (2000). Reverse vaccinology. Current Opinion in Microbiology, 3(5), 445–450. https://doi.org/10.1016/S1369-5274(00)00119-3

13. Bellini, C., & Horváti, K. (2020). Recent Advances in the Development of Protein- and Peptide-Based Subunit Vaccines against Tuberculosis. Cells, 9(12). https://doi.org/10.3390/CELLS9122673

14. Qamar, M. T. U., Shokat, Z., Muneer, I., Ashfaq, U. A., Javed, H., Anwar, F., … Saari, N. (2020). Multiepitope-Based Subunit Vaccine Design and Evaluation against Respiratory Syncytial Virus Using Reverse Vaccinology Approach. Vaccines, 8(2), 1–27. https://doi.org/10.3390/VACCINES8020288

15. Nguyen, T. L., Lee, Y., & Kim, H. (2022). Immunogenic Epitope-Based Vaccine Prediction from Surface Glycoprotein of MERS-CoV by Deploying Immunoinformatics Approach. International Journal of Peptide Research and Therapeutics, 28(3), 1–11. https://doi.org/10.1007/S10989-022-10382-5/FIGURES/6

16. Nguyen, T. L., & Kim, H. (2024). Immunoinformatics and computational approaches driven designing a novel vaccine candidate against Powassan virus. Scientific Reports 2024 14:1, 14(1), 1–15. https://doi.org/10.1038/s41598-024-56554-9

17. Nguyen, T. L., Samuel Leon Magdaleno, J., Rajjak Shaikh, A., Choowongkomon, K., Li, V., Lee, Y., & Kim, H. (2023). Designing a multi-epitope candidate vaccine by employing immunoinformatics approaches to control African swine fever spread. Journal of Biomolecular Structure and Dynamics, 41(19), 10214–10229. https://doi.org/10.1080/07391102.2022.2153922

18. Nain, Z., Karim, M. M., Sen, M. K., & Adhikari, U. K. (2020). Structural basis and designing of peptide vaccine using PE-PGRS family protein of Mycobacterium ulcerans—An integrated vaccinomics approach. Molecular Immunology, 120, 146–163. https://doi.org/10.1016/J.MOLIMM.2020.02.009

19. Carlos, P., Roupie, V., Holbert, S., Ascencio, F., Huygen, K., Gomez-Anduro, G., … Angulo, C. (2015). In silico epitope analysis of unique and membrane associated proteins from Mycobacterium avium subsp. paratuberculosis for immunogenicity and vaccine evaluation. Journal of Theoretical Biology, 384, 1–9. https://doi.org/10.1016/J.JTBI.2015.08.003

20. Bibi, S., Ullah, I., Zhu, B., Adnan, M., Liaqat, R., Kong, W. B., & Niu, S. (2021). In silico analysis of epitope-based vaccine candidate against tuberculosis using reverse vaccinology. Scientific Reports 2021 11:1, 11(1), 1–16. https://doi.org/10.1038/s41598-020-80899-6

21. Rappuoli, R. (2000). Reverse vaccinology. Current Opinion in Microbiology, 3(5), 445–450. https://doi.org/10.1016/S1369-5274(00)00119-3

22. Dolgin, E., & Ledford, H. (2023). mRNA COVID vaccines saved lives and won a Nobel — what’s next for the technology? Nature. https://doi.org/10.1038/D41586-023-03119-X

23. Paliwal, D., Thom, M., Hussein, A., Ravishankar, D., Wilkes, A., Charleston, B., & Jones, I. M. (2022). Towards Reverse Vaccinology for Bovine TB: High Throughput Expression of Full Length Recombinant Mycobacterium bovis Proteins. Frontiers in Molecular Biosciences, 9. https://doi.org/10.3389/FMOLB.2022.889667

24. Monterrubio-López, G. P., González-Y-Merchand, J. A., & Ribas-Aparicio, R. M. (2015). Identification of Novel Potential Vaccine Candidates against Tuberculosis Based on Reverse Vaccinology. BioMed Research International, 2015. https://doi.org/10.1155/2015/483150

25. Dhanda, S. K., Vir, P., Singla, D., Gupta, S., Kumar, S., & Raghava, G. P. S. (2016). A Web-Based Platform for Designing Vaccines against Existing and Emerging Strains of Mycobacterium tuberculosis. PLOS ONE, 11(4), e0153771. https://doi.org/10.1371/JOURNAL.PONE.0153771

26. Smith, S. G. J., Mahon, V., Lambert, M. A., & Fagan, R. P. (2007). A molecular Swiss army knife: OmpA structure, function and expression. FEMS Microbiology Letters, 273(1), 1–11. https://doi.org/10.1111/J.1574-6968.2007.00778.X

27. Pym, A. S., Brodin, P., Brosch, R., Huerre, M., & Cole, S. T. (2002). Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti. Molecular Microbiology, 46(3), 709–717. https://doi.org/10.1046/J.1365-2958.2002.03237.X

28. Wang, C., Lu, J., Du, W., Wang, G., Li, X., Shen, X., … Xu, M. (2019). Ag85b/ESAT6-CFP10 adjuvanted with aluminum/poly-IC effectively protects guinea pigs from latent mycobacterium tuberculosis infection. Vaccine, 37(32), 4477–4484. https://doi.org/10.1016/J.VACCINE.2019.06.078

29. Yang, E., Lu, Y., Xu, Y., Liang, Q., Wang, C., Wang, H., & Shen, H. (2014). Recombinant BCG coexpressing Ag85B, ESAT-6 and Rv3620c elicits specific Th1 immune responses in C57BL/6 mice. Microbial Pathogenesis, 69–70(1), 53–59. https://doi.org/10.1016/J.MICPATH.2014.03.011

30. Hoang, T., Aagaard, C., Dietrich, J., Cassidy, J. P., Dolganov, G., Schoolnik, G. K., … Andersen, P. (2013). ESAT-6 (EsxA) and TB10.4 (EsxH) Based Vaccines for Pre- and Post-Exposure Tuberculosis Vaccination. PLoS ONE, 8(12). https://doi.org/10.1371/JOURNAL.PONE.0080579

31. Skjøt, R. L. V., Oettinger, T., Rosenkrands, I., Ravn, P., Brock, I., Jacobsen, S., & Andersen, P. (2000). Comparative Evaluation of Low-Molecular-Mass Proteins from Mycobacterium tuberculosis Identifies Members of the ESAT-6 Family as Immunodominant T-Cell Antigens. Infection and Immunity, 68(1), 214. https://doi.org/10.1128/IAI.68.1.214-220.2000

32. Li, W., Li, M., Deng, G., Zhao, L., Liu, X., & Wang, Y. (2015). Prime-boost vaccination with Bacillus Calmette Guerin and a recombinant adenovirus co-expressing CFP10, ESAT6, Ag85A and Ag85B of Mycobacterium tuberculosis induces robust antigen-specific immune responses in mice. Molecular medicine reports, 12(2), 3073–3080. https://doi.org/10.3892/mmr.2015.3770

33. Saha, S., & Raghava, G. P. S. (2006). Prediction of continuous B-cell epitopes in an antigen using recurrent neural network. Proteins: Structure, Function, and Bioinformatics, 65(1), 40–48. https://doi.org/10.1002/PROT.21078

34. Karlin, S., & Brocchieri, L. (1998). Heat shock protein 70 family: Multiple sequence comparisons, function, and evolution. Journal of Molecular Evolution, 47(5), 565–577. https://doi.org/10.1007/PL00006413/METRICS

35. Peake, P., Basten, A., & Britton, W. J. (1991). Characterization of the functional properties of the 70-kDa protein of Mycobacterium bovis. Journal of Biological Chemistry, 266(31), 20828–20832. https://doi.org/10.1016/S0021-9258(18)54784-0

36. Bulut, Y., Michelsen, K. S., Hayrapetian, L., Naiki, Y., Spallek, R., Singh, M., & Arditi, M. (2005). Mycobacterium Tuberculosis Heat Shock Proteins Use Diverse Toll-like Receptor Pathways to Activate Pro-inflammatory Signals. Journal of Biological Chemistry, 280(22), 20961–20967. https://doi.org/10.1074/JBC.M411379200

37. Fang, H., Wu, Y., Huang, X., Wang, W., Ang, B., Cao, X., & Wan, T. (2011). Toll-like Receptor 4 (TLR4) Is Essential for Hsp70-like Protein 1 (HSP70L1) to Activate Dendritic Cells and Induce Th1 Response. The Journal of Biological Chemistry, 286(35), 30393. https://doi.org/10.1074/JBC.M111.266528

38. Magnan, C. N., Randall, A., & Baldi, P. (2009). SOLpro: accurate sequence-based prediction of protein solubility. Bioinformatics, 25(17), 2200–2207. https://doi.org/10.1093/BIOINFORMATICS/BTP386

39. The Proteomics Protocols Handbook. (2005). The Proteomics Protocols Handbook. https://doi.org/10.1385/1592598900

40. Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., … Hassabis, D. (2021). Highly accurate protein structure prediction with AlphaFold. Nature, 596(7873), 583. https://doi.org/10.1038/S41586-021-03819-2

41. Heo, L., Park, H., & Seok, C. (2013). GalaxyRefine: protein structure refinement driven by side-chain repacking. Nucleic Acids Research, 41(W1), W384–W388. https://doi.org/10.1093/NAR/GKT458

42. Wang, Y., Zhou, X., Lin, J., Yin, F., Xu, L., Huang, Y., … Zhao, D. (2011). Effects of Mycobacterium bovis on monocyte-derived macrophages from bovine tuberculosis infection and healthy cattle. FEMS Microbiology Letters, 321(1), 30–36. https://doi.org/10.1111/J.1574-6968.2011.02304.X

43. Fremond, C. M. C., Nicolle, D. M. M., Torres, D. S., & Quesniaux, V. F. J. (2003). Control of Mycobacterium bovis BCG infection with increased inflammation in TLR4-deficient mice. Microbes and Infection, 5(12), 1070–1081. https://doi.org/10.1016/J.MICINF.2003.06.001

44. Vabulas, R. M., Ahmad-Nejad, P., Ghose, S., Kirschning, C. J., Issels, R. D., & Wagner, H. (2002). HSP70 as Endogenous Stimulus of the Toll/Interleukin-1 Receptor Signal Pathway. Journal of Biological Chemistry, 277(17), 15107–15112. https://doi.org/10.1074/JBC.M111204200

45. Asea, A., Rehli, M., Kabingu, E., Boch, J. A., Baré, O., Auron, P. E., … Calderwood, S. K. (2002). Novel Signal Transduction Pathway Utilized by Extracellular HSP70 ROLE OF Toll-LIKE RECEPTOR (TLR) 2 AND TLR4*. Journal of Biological Chemistry, 277, 15028–15034. https://doi.org/10.1074/jbc.M200497200

46. Kozakov, D., Hall, D. R., Xia, B., Porter, K. A., Padhorny, D., Yueh, C., … Vajda, S. (2017). The ClusPro web server for protein-protein docking. Nature protocols, 12(2), 255. https://doi.org/10.1038/NPROT.2016.169

47. Laskowski, R. A., Jabłońska, J., Pravda, L., Vařeková, R. S., & Thornton, J. M. (2018). PDBsum: Structural summaries of PDB entries. Protein Science : A Publication of the Protein Society, 27(1), 129. https://doi.org/10.1002/PRO.3289

48. Abraham, M. J., Murtola, T., Schulz, R., Páll, S., Smith, J. C., Hess, B., & Lindah, E. (2015). GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX, 1–2, 19–25. https://doi.org/10.1016/J.SOFTX.2015.06.001

49. Vangone, A., Spinelli, R., Scarano, V., Cavallo, L., & Oliva, R. (2011). COCOMAPS: a web application to analyze and visualize contacts at the interface of biomolecular complexes. Bioinformatics, 27(20), 2915–2916. https://doi.org/10.1093/BIOINFORMATICS/BTR484

50. Dimitrov, I., Bangov, I., Flower, D. R., & Doytchinova, I. (2014). AllerTOP v.2 - A server for in silico prediction of allergens. Journal of Molecular Modeling, 20(6), 1–6. https://doi.org/10.1007/S00894-014-2278-5/TABLES/3

51. Morozov, V., Rodrigues, C. H. M., & Ascher, D. B. (2023). CSM-Toxin: A Web-Server for Predicting Protein Toxicity. Pharmaceutics, 15(2), 431. https://doi.org/10.3390/PHARMACEUTICS15020431/S1

52. Doytchinova, I. A., & Flower, D. R. (2007). VaxiJen: A server for prediction of protective antigens, tumour antigens and subunit vaccines. BMC Bioinformatics, 8(1), 1–7. https://doi.org/10.1186/1471-2105-8-4/TABLES/2

Table legends

Table 1. Physiochemical properties and overall traits of the vaccine construct.

Yes