The amino acid sequences of the main Brucella antigens, specifically OMP16 (AEF59023.1), OMP19 (AAB06277.1), L7/L12 (AAL51929.1), and OMP25 (AFJ79953.1), were retrieved from the NCBI database and are detailed in Table 1. To determine whether these four proteins have antigenic properties, we employed the VaxiJen server v2.0, setting a threshold value at 0.5 (25).

To predict Cytotoxic T Lymphocyte (CTL) epitopes for the primary Brucella antigen, we utilized the IEDB MHC I server, accessible (26).¹ The specific parameters we used were as follows: Prediction Method: IEDB recommended 2020.09 (NetMHCpan EL 4.1); Source species of MHC: Human; HLA allele reference set, and Length set to 9 and 10. We selected epitopes with a percentile rank of less than 0.5 for further analysis. We also used Class I Immunogenicity to predict immunogenicity and chose CTL epitopes with a percentile rank of less than 0.5 and an immunoscore greater than 0 for further analysis. To determine the antigenicity of these screened epitopes, we employed the VaxiJen server v2.0, setting a threshold at 0.5 (25). The epitopes that passed these criteria were considered as immunodominant CTL epitopes for the construction of MEV-Fc.

To predict Helper T Lymphocyte (HTL) epitopes for the dominant Brucella antigen, we utilized the IEDB MHCII server 2, which can be accessed (27).² The specific parameters we used were as follows: Prediction Method: IEDB recommended 2.22; MHC source species: Human; MHC allele set selected as all reference. We set the epitope length to 15. The predicted epitopes were ranked based on percentile rank, and we selected those with a percentile rank of less than 0.5. To predict their antigenicity, we used the VaxiJen server v2.0 with a threshold set at 0.5 (25). We chose epitopes with an antigenicity value greater than 0.5, and we predicted their inducibility of IFN-γ using an epitope server (28) specifically designed for IFN-γ. Ultimately, the epitope that was predicted to induce IFN-γ was selected as the HTL immunodominant epitope for the construction of MEV-Fc.

To predict linear B-cell epitopes for the dominant Brucella antigen, we used the ABCpred server, which is accessible,³ with the following parameters: epitope length = 16 and screening threshold = 0.51. We employed thresholds ranging from +0.1 to +1.0 (29). The highest ABCpred score among these thresholds was selected as the immunodominant B-cell epitope for MEV (Multi-Epitope Vaccine) construction.

To create the MEV, we used an appropriate linker to connect the individual epitopes. This linker prevents these epitopes from interacting with each other. The MEV was formed by connecting the predicted and screened CTL epitopes, an HTL epitope, and a linear B-cell epitope using the GPGPG linker.

To further enhance its ability to trigger an immune response and ensure the persistence of the helper T cell 1 (TH1) response (30), we added the hBD-3 sequence (PDB ID: 1KJ6) and the PADRE sequence to the N-terminal end of the MEV, using EAAAK linkers. Finally, to make the peptide more effective at triggering an immune response and to make it last longer in the body, we added the sequence of the mouse immunoglobulin IgG Fc fragment (P01868-1) by connecting it to the C-terminus of the MEV with KK linkers. This resulted in the complete sequence of the vaccine construct.

We evaluated the properties of MEV-Fc using bioinformatic analysis software. Firstly, we analyzed its physical and chemical characteristics using the ProtParam server, accessible.⁴ Next, we conducted antimicrobial and immunogenicity analyses using the VaxiJen v2.0 server. We specifically utilized the IEDB I class immunogenicity module, with VaxiJen predictions showing an accuracy range of 70 to 89% (25). Subsequently, we predicted solubility using the SOLpro server, available,⁵ which demonstrated an overall accuracy of 74.15% (31). Finally, we performed sensitivity and toxicity analyses using AlergenFP v1.0, accessible (32),⁶ and ToxinPred, available (33).⁷

We predicted the secondary structure of MEV-Fc using the SOPMA secondary structure prediction tool. To forecast the tertiary structure of MEV-Fc, we utilized the Robetta server (34). To assess the quality of the constructed models, we employed the Prosa Web Server (35), the Ramachandran Plot, and the PROCHECK Server (36), which can be accessed.⁸ These tools help evaluate the correctness and quality of the protein’s three-dimensional structure. We further assessed the uncertainty of the tertiary structure through ERRAT, available.⁹ ERRAT examines the overall quality factor of non-bonded atomic interactions, with higher values indicating better quality (37).

B-cell epitopes are particular sites on antigens where B-cell antibodies can bind and trigger immune responses (38). In vaccine development, these epitopes play a crucial role in generating an effective immune response. To predict B-cell epitopes for MEV-Fc, we utilized the ElliPro server with default parameters (39).

Molecular docking is a fundamental and promising method for studying how peptides interact with human immune receptors like TLR2 and TLR4 (40). It offers a quick and cost-effective way to understand the detailed interactions at the atomic level within the structure of the antibody–antigen complex and its interface.

Here’s how we conducted the molecular docking analysis: We obtained the PDB files of TLR2 (PDB ID: 2Z7X) and TLR4 (PDB ID: 2Z63) from the NCBI Molecular Modeling Database (MMDB), which is accessible.¹⁰ Next, we performed ligand-receptor docking analysis using the ClusPro 2.0 online server, which you can find (41).¹¹ We then examined the interaction surfaces of the resulting complexes using PDBE Pisa, available (42).¹² Finally, we visualized the complexes using PyMOL software and analyzed them using Ligplot+ software.

We performed molecular dynamics simulations of the MEV-Fc-TLR2 and MEV-Fc-TLR4 complexes using the iMODSweb server. This server, which is accessible,¹³ allows us to explore possible trajectories between two conformations and enables interactive analysis of the resulting structures, animations, and trajectories in three dimensions. Importantly, it can effectively simulate and explore even large molecules (43).

We used the C-ImmSim server, which you can access,¹⁴ to predict the ability of MEV-Fc to stimulate the production of specific antibodies and various cytokines by immune cells (44). This server is capable of predicting the immune response of both B lymphocytes and T lymphocytes, including Th1 and Th2 lymphocytes.

In our simulation, we considered the recommended minimum interval between the initial and subsequent doses of most vaccines, which is 28 days (45). Therefore, we configured the simulation parameters as follows: three injections, with each injection spaced 28 days apart; a random seed value set to 12,345; a simulation volume of 50; and a total of 1,050 simulation steps. We kept the remaining parameters at their default values.

To obtain an optimized nucleotide sequence, we used the Java Codon Adaptation Tool (JCat), which is available.¹⁵ This tool quickly generates optimized codon sequences tailored to the chosen expression host, thereby enhancing the production yield of heterologous proteins (46). In this case, we selected E. coli (strain K12) as the preferred bacterial strain. The ideal Codon Adaptation Index (CAI) value is 1, and the GC content percentage should ideally fall within the range of 30 to 70%.

After obtaining the optimized gene sequence, we cloned it into the expression vector pET28a(+) using Hind III and BamHI digestion sites. Subsequently, we analyzed the cloned sequence using Snapgene software.

A set of four proteins was retrieved from the NCBI server, and subsequently, their antigenicity was predicted using VaxiJen (Table 1). Utilizing the IEDB MHCI and MHC II servers, we predicted a total of 308 CTL epitopes (Supplementary Table 1) and 89 HTL epitopes (Supplementary Table 2) with a percentile rank of 0.5. Among the 308 CTL epitopes, 57 epitopes with immunogenicity >0 and an antigenicity score > 1 were selected (Supplementary Table 3). Additionally, from the 89 HTL epitopes, 28 epitopes exhibiting positive induction of IFN-γ and an antigenicity score > 0 were selected (Supplementary Table 4). The ABCpred server was employed to predict 68 linear B cell epitopes (Supplementary Table 5). Ultimately, 18 epitopes were chosen as candidates for vaccine construction (Table 2), including eight epitopes with the highest immunogenicity or highest antigenicity score, six HTL epitopes with the highest antigenicity score or highest IFN-γ scores, and four B cell epitopes with the highest ABCpred prediction scores. The predicted and screened CTL epitopes, HTL epitopes, and linear B-cell epitopes were linked together using GPGPG linkers. The hBD-3 sequences and PADRE sequences were connected via EAAAK linkers, while the fc (PDB ID: 1KJ6) sequences were linked using KK linkers. The resulting peptide molecule was named MEV-Fc, and its schematic diagram and amino acid sequence are presented in Figure 2.

MEV-Fc, a protein we studied, has some important characteristics. It is composed of 600 amino acids and has a molecular weight of about 61.2 kilodaltons. Its isoelectric point (PI) is approximately 8.89, indicating its charge at a specific pH. It is relatively stable with an instability index of 24.06 and Aliphatic index of 57.35. In terms of its lifespan, MEV-Fc is estimated to last about 30 h in mammalian reticulocytes in vitro. It has a slight preference for water, making it slightly hydrophilic with a GRAVY value of −0.500 (Table 3). Moreover, it is highly likely to be soluble during production, with a probability of 0.976819.

Regarding its potential as an antigen, predictions suggest it has a high likelihood of triggering an immune response with an immunogenicity score of 4.18 and good antigenicity (1.0342). Importantly, MEV-Fc is considered non-sensitizing and non-toxic based on prediction, indicating its safety profile.

In summary, MEV-Fc possesses favorable characteristics, including stability, solubility, and the potential to elicit an immune response, making it a strong candidate for a peptide vaccine. Additionally, it is considered safe for use.

The results of the secondary structure prediction (Figure 3A) show that MEV-Fc’s conformation consists of 15.33% alpha helix (Hh), 19.17% extended strand (Ee), 7.17% beta-turn (Tt), and 58.33% random coil (Cc). To predict the tertiary structure of MEV-Fc, we used the Robetta server, and the resulting model was visualized using the Pymol software package (Figure 3B). To assess the model’s quality, we conducted ProSA-web analysis, which yielded a Z-Score of −7.88 (Figure 4A). This Z-Score falls within the normal range for native proteins of similar size, confirming the credibility of the model. Additionally, we generated a Ramachandran plot using the PROCHECK server to comprehensively evaluate the overall quality of the structural arrangement (Figure 4B). The analysis showed that 87.1% of residues were in the most favored regions, 10.2% in additional allowed regions, 1.2% in generously allowed regions, and 1.6% in disallowed regions.

In conclusion, the ERRAT server indicated an overall quality factor of 86.531, surpassing the generally accepted threshold of 50 for a good model. Therefore, we can confidently consider the tertiary structure of MEV-Fc as a reliable and accurate model.

Using the ElliPro server, we identified a comprehensive set of 311 residues that make up discontinuous B cell epitopes (Supplementary Table 6). These epitopes had scores ranging from 0.524 to 0.854. Additionally, our prediction revealed 18 consecutive B cell epitopes (Supplementary Table 7), which varied in size from 4 to 42 residues, with scores ranging from 0.512 to 0.832.

We used a powerful tool called ClusPro 2.0 to understand how our vaccine interacts with TLR2 and TLR4, two important proteins. When we examined how the vaccine connects with TLR2, we got 10 different models. We found that one model stood out, showing 11 hydrogen bonds and 5 salt-bridge interactions (Figure 5; Supplementary Table 8). The strength of the bond, called binding free energy, was −10.4 kcal/mol, indicating a strong connection between MEV-Fc and TLR2. To better understand this interaction, we used Ligplot+ to create a clear 2D picture of how they bond (Figure 6).

In a similar way, we looked at how our vaccine interacts with TLR4. We also got 10 different models. One model showed an impressive 27 hydrogen bonds and 5 salt-bridge interactions (Figure 7; Supplementary Table 9). The strength of this bond, in terms of binding free energy, was even higher at −18.8 kcal/mol, suggesting a very strong link between MEV-Fc and TLR4. We used Ligplot+ again to create a clear 2D picture of this complex bond (Figure 8).

We used a tool called IModS to simulate how the molecules in our vaccine move and behave. In Figures 9 A-F, we show the results of this simulation. In Figure A, we can see how our vaccine (MEV-Fc) behaves when it is connected to TLR2. We used a method called normal mode analysis (NMA) to understand how flexible the proteins are. Figure B shows specific values related to this analysis for the MEV-Fc-TLR2 complex. In Figure C, variance plots showed a cumulative or individual variance of MEV-Fc-TLR2 complex with green or purple, respectively. In Figure D, we compare the actual movement (B factor) with what we predicted through NMA. Figure E displays a map that highlights how different parts of the complex move together. The red areas show significant motions happening together. Lastly, Figure F provides a diagram showing how the parts of the docked protein complexes are interconnected, like springs with varying stiffness. Darker areas indicate stiffer connections. Similar results were obtained for TLR4 (Supplementary Figure 1).

We used a tool called C-ImmSim to predict how the immune system in mice would respond to the MEV-Fc vaccine. Here are the key findings (Figure 10): (1) Antibody Levels: The results showed that the levels of IgM and IgG antibodies increased progressively after three rounds of vaccine injections (Figure 10A). This suggests that the vaccine triggered an antibody response in mice. (2) B-Cells: B-cells, which are responsible for producing antibodies, became more active with each vaccine injection, reaching their highest level after the final injection (Figure 10B). (3) Helper T Cells (TH-Cells): Both the total and memory populations of helper T cells (TH-cells) increased significantly (Figure 10C). The active TH-cell population also expanded robustly, reaching its peak after the third immunization (Figure 10D). (4) Cytotoxic T Cells (TC Cells): The count of active cytotoxic T lymphocytes (TC cells) increased gradually after each immunization injection (Figure 10E). (5) Cytokines: There was a notable increase in the levels of interferon-γ and IL-2, which are immune system signaling molecules, in response to antigen stimulation (Figure 10F). These findings suggest that the MEV-Fc vaccine could elicit a strong and diverse immune response in mice, involving antibodies, B-cells, helper T cells, cytotoxic T cells, and cytokines.

Codon optimization is a technique used to improve the way a particular genetic sequence is expressed by enhancing translational efficiency (47). In this context, we optimized the codon usage in the MEV-Fc construct sequence using a tool called JCat. This optimization resulted in a Codon Adaptation Index (CAI) value of 1, and the GC content was measured at 56.99%, which falls within the desirable range. These findings suggest that the optimized sequence is likely to be efficiently expressed in the E. coli expression system. You can see the result in (Figure 11), which shows the expression vector pET 28a(+) with the inserted fragment of the multi-epitope vaccines.