The complete EP and RdrRP sequences of the Nigerian strain of CCHFV were retrieved from Uniprot with IDs Q8JSZ3 and Q6TQR6, respectively, and saved.

MHC (HLA) class I allele prediction was performed using the NetMHC 4.0 online server (Lundegaard et al., 2008). While MHC class II allele prediction was conducted via the RANKPEP server (Reche et al., 2004). Linear B-cell epitopes were identified using both the Emini Surface Accessibility method (Emini et al., 1985) and the BepiPred Linear Epitope Prediction tool (Ponomarenko and Bourne, 2007). The antigenic potential of the selected epitopes was evaluated through the VaxiJen v2.0 server (Doytchinova and Flower, 2008), using “virus” as the target organism and a threshold of ≥0.4.

The allergenicity of each epitope was assessed using AllergenFP v.1.0 (Buus et al., 2003) and AllerTOP v. 2.0 (Dimitrov et al., 2014). Only epitopes that were both antigenic and non-allergenic were selected for inclusion in the vaccine construct. Overlapping epitopes were further analyzed for population coverage of HLA class I (HLA-Ia, HLA-Ib, and HLA-Ic) and class II (HLA-II DP, DQ, and DR) alleles worldwide using the IEDB population coverage tool (Bui et al., 2006). Finally, all candidate epitopes were aligned against human proteins using the BLAST tool on the NCBI platform to ensure non-homology.

Based on the previously obtained results, the final vaccine construct was composed of two epitopes derived from the envelopment polyprotein (EP) and one epitope from the RNA-directed RNA polymerase (RdRP). The epitopes were connected to two immunostimulatory adjuvants, the mycobacterial lipoprotein LprG and the synthetic peptide of human beta-defensin-3 (HBD-3), to enhance the immunogenic potential of the construct.

The different components of the vaccine were linked using KK linkers to ensure structural flexibility and maintain the independent immunogenic activity of each segment. Subsequently, the physicochemical properties of the designed vaccine, including molecular weight, theoretical pI, instability index, aliphatic index, and grand average of hydropathicity (GRAVY), were evaluated using the Expasy ProtParam tool (Gasteiger et al., 2005). In addition, the antigenicity and allergenicity of the construct were further assessed using the Vaxijen v2.0, ANTIGENpro (Gasteiger et al., 2005), AllergenFP v.1.0, and AllerTOP v. 2.0 servers, respectively.

The SOPMA server (Geourjon and Deleage, 1995) was used for the secondary structure prediction of the final vaccine. The 3D structure was predicted using GalaxyTBM (Ko et al., 2012). This server predicted 10 models of the vaccine’s 3D structure, and the best model was selected based on the ERRAT evaluation (Colovos and Yeates, 1993). The Galaxy Refine server refined this model (Ko et al., 2012), and finally, ProSA-web (Wiederstein and Sippl, 2007) and Zlab validated it (Anderson et al., 2005). An effective vaccine should include B-cell epitopes capable of eliciting robust immune responses from B lymphocytes. Therefore, the discontinuous (conformational) B-cell epitopes of the vaccine construct were predicted using the ElliPro server. (Ponomarenko et al., 2008).

The three-dimensional structures of human Toll-like receptors (TLRs) 2 and 3 were retrieved from the Protein Data Bank (PDB) (www.rcsb.org) using the PDB IDs 6NIG, 5GS0, and 3W3G. All associated ligands and water molecules were removed from the structures prior to docking. Protein–protein docking between the vaccine structure (as the ligand) and each TLR (as the receptor) was carried out using the HDOCK online server (Remmert et al., 2012; Yan et al., 2020; Yan et al., 2017). The best docked complexes, determined by maximal interaction scores, were visualized and analyzed using Discovery Studio 4.5 software. NMA of the docked complexes was performed using the iMODS server (López-Blanco et al., 2014). The structural stability of each complex was evaluated through main-chain deformability plots, B-factor values, eigenvalue scores, and covariance matrix analyses.

The Genecorner tool was used to convert the amino acid sequence of the designed vaccine into its corresponding DNA sequence. The DNA sequence was further analyzed for codon optimization using the GenScript server. To evaluate the mRNA secondary structure following transcription, the DNA sequence was transcribed into RNA using the DNA↔RNA↔Protein tool, and the RNAfold server (Mathews et al., 2004) was employed to predict the stability of the mRNA structure. For cloning purposes, the optimized DNA sequence was inserted into the pBAD24 expression vector using ClaI and EcoRV restriction sites with the help of CLC Sequence Viewer v8.0 (Bio-Qiagen, 2016). Additionally, the expression of the vaccine gene in E. coli was modeled using NotI and BamHI restriction sites within the PMT-puro vector system.

The C-ImmSim server was utilized to appraise the immunogenicity profiles and the immune response of the human body against the vaccine. To stimulate the immune response in humans, the simulation step characteristics, random seed, and simulation volume were set to 105, 12345, and 10 μL, respectively (Rapin et al., 2010).

The NetMHC 4.0 server was employed to predict HLA class I-binding epitopes from the envelopment polyprotein (EP) and the RNA-directed RNA polymerase (RdRP) proteins. For the identification of HLA class II-binding epitopes, the RANKPEP server was used. Linear B-cell epitopes were predicted using the IEDB database. All selected peptides were evaluated for their antigenicity using the VaxiJen v2.0 server, and their allergenicity was assessed via AllergenFP v1.0 and AllerTOP v2.0. Based on these analyses, three immunodominant epitopes—two derived from the EP and one from the RdRP protein—were selected for inclusion in the final vaccine construct. These epitopes were predicted to be antigenic and non-allergenic (Table 1), and they demonstrated 100% global HLA allele coverage (Supplementary Figure S1). In this study, the number of selected epitopes was intentionally limited to three, with each one overlapping with MHC class I, MHC class II, and linear B-cell epitopes. In contrast, previous studies have utilized a higher number of epitopes (e.g., 16 in Khan et al., and 9 in Qamar et al.).

To enhance the immunogenicity of the vaccine construct, the selected epitopes were conjugated with two adjuvants, LprG and HBD-3 (Figure 2A). LprG, a well-known TLR2 agonist, stimulates the innate immune response (Khan et al., 2021). HBD-3, the second adjuvant, has been widely used in the design of viral vaccines (Leikina et al., 2005; Srivastava et al., 2018). This peptide forms a protective barrier of immobilized surface proteins, effectively preventing viral fusion (Joly et al., 2005; Judge et al., 2015). It also interacts with chemokine receptors CCR2 and CCR6, promoting the chemotaxis of immature dendritic cells, T cells, and monocytes (Funderburg et al., 2007; Judge et al., 2015). Additionally, HBD-3 may activate antigen-presenting cells (APCs) via TLR1 and TLR2, and induce the expression of interleukin-22 (IL-22), transforming growth factor-α (TGF-α), and IFN-γ (Ferris et al., 2013; Sørensen et al., 2003; Wolk et al., 2004; Yazdani et al., 2020c). The physicochemical and immunological properties of the designed multi-epitope vaccine are summarized in Table 2. The construct exhibited high stability, with predicted half-lives of approximately 10 h in E. coli, 20 h in yeast, and 30 h in mammalian reticulocytes. Antigenicity predictions by VaxiJen v2.0 and ANTIGENpro confirmed the vaccine as a strongly antigenic protein. Allergenicity analysis using AllergenFP v1.0 and AllerTOP v2.0 verified that the construct is non-allergenic. Furthermore, BLAST analysis against the human proteome showed no significant sequence similarity, except in the HBD-3 adjuvant region (Supplementary Figure S2). Notably, the limitation to three epitopes in this vaccine construct results in a lower molecular weight compared to the vaccine designed by Khan et al. (Khan et al., 2021), potentially facilitating more efficient cellular uptake and in vivo delivery.

The SOPMA server was employed to predict the secondary structure of the multi-epitope peptide vaccine. The analysis revealed that the construct consists of approximately 21.3% α-helices, 18.58% extended strands, 6.62% β-turns, and 53.44% random coils. The 3D structure of the vaccine was generated using the GalaxyTBM server, which produced 10 structural models. The optimal model was selected based on ERRAT quality assessment (Supplementary Figure S3) and was subsequently refined using the GalaxyRefine server, which produced five refined models. These models were evaluated using the ProSA and ZLab servers. Model 3 exhibited the highest structural quality among the refined models. ProSA analysis returned a Z-score of −5.53, which falls within the range of native proteins of comparable size, confirming the model’s acceptable global quality (Supplementary Figure S3). Ramachandran plot analysis via ZLab showed that 98.79% of residues were located in the most favored regions, 0.906% were in additional allowed regions, and only 0.302% were in disallowed regions. These results confirm the reliability and accuracy of the final 3D model, which can be utilized for subsequent structural and functional analyses (Figure 2B). Prediction of conformational (discontinuous) B-cell epitopes using the Ellipro server identified 108 amino acid residues as part of potential conformational epitopes within the vaccine structure. The epitope scores ranged from 0.61 to 0.711, indicating their probable surface accessibility and immunogenic potential (Table 3; Figure 3).

Molecular docking between the 3D model of the CCHFV vaccine and TLRs 2, 3, and 8 was carried out using the HDOCK server. The complex demonstrating the most favorable and extensive receptor–ligand interactions among the generated docking models was selected as the optimal docking configuration. These findings suggest that the vaccine construct is successfully docked into the active binding pockets of TLR2, TLR3, and TLR8 (Figures 4–6). To further evaluate the stability and conformational flexibility of the docked complexes, NMA was performed using the iMODS server. The results, visualized in Figures 4–6, provide insight into the dynamic behavior of the vaccine–receptor complexes. The hinge regions, identified as the most flexible parts of the complex, indicate potential sites of conformational motion. The B-factor plots represent atomic fluctuations, while eigenvalues reflect the energy required for structural deformation—with lower values indicating greater structural flexibility. The covariance matrix shows correlated (red), uncorrelated (white), and anti-correlated (blue) motions between residue pairs. Additionally, the elastic network model displays the strength and connectivity of interatomic interactions via spring representations. Overall, NMA results support that interactions between the CCHFV vaccine and TLRs 2, 3, and 8 are structurally stable. Compared to previous CCHFV vaccine designs, the incorporation of three TLR targets suggests enhanced immunogenicity and a stronger potential for immune system activation.

The codon optimization and expression potential of the vaccine construct in E. coli were assessed using the GenScript and GeneCorner online tools. The optimized codon sequence contained 1,044 nucleotides and exhibited a codon adaptation index (CAI) of 1.0, an average GC content of 62.26%, and a codon frequency distribution (CFD) of 0%, all indicating excellent compatibility and potential for high-level expression in E. coli. For in silico cloning, the optimized DNA sequence was inserted into the pBAD24 expression vector using ClaI and EcoRV restriction sites via CLC Sequence Viewer v8.0. The final recombinant plasmid had a total length of 5,209 base pairs (bp) (Figure 7A). In addition, a second cloning approach involved inserting the vaccine sequence into the pMT-puro vector using NotI and BamHI restriction sites, resulting in a final construct of 5,380 bp (Figure 7B). This study uniquely evaluates and demonstrates the insertion of the vaccine gene into multiple expression vectors, a step not addressed in previous CCHFV vaccine design studies. Furthermore, the minimum free energy (ΔG) of the mRNA secondary structure was calculated using the RNAfold web server. The predicted ΔG was −409.10 kcal/mol, suggesting a stable mRNA structure conducive to efficient translation. The secondary structure of the mRNA is presented in Supplementary Figure S5.

The C-ImmSim server was employed to simulate the dynamic immune response of the human immune system following administration of the designed vaccine (Figure 8A). The simulation demonstrated a strong humoral immune response, with the combined levels of IgM and IgG reaching a peak concentration of approximately 10,000 antigenic counts/mL, predominantly driven by IgG1 and IgG2 isotypes. Additionally, cellular immune responses were significantly activated, as evidenced by the elevated levels of CTLs and helper T cells (Th cells). Dendritic cells (DCs), macrophages, and Natural killer cells (NKs) maintained consistent and elevated activity throughout the simulated immunization period. Notably, the level of IFN-γ surpassed 400,000 ng/mL, indicating a robust Th1-mediated immune response. These results suggest that the designed vaccine has the potential to induce a strong and comprehensive immune reaction, encompassing both the humoral and cellular arms of the immune system (Figure 8).