Brucellosis novel multi-epitope vaccine design based on in silico analysis focusing on Brucella abortus

Abstract

Brucella is a common kind of bacteria that has the ability to live within cells and may cause diseases that can be transmitted between animals and humans. Current medical therapy struggles to effectively eradicate Brucella. Thus, it is necessary to develop a multi-epitope vaccine (MEV) in order to effectively prevent Brucella infection. To achieve this objective, we used the reverse vaccinology methodology based on omp19 and Bacterial surface antigen (D15). After conducting our research, we successfully identified 2 cytotoxic T lymphocyte (CTL) epitopes, 2 helper T lymphocyte (HTL) epitopes, and 2 linear B cell epitopes from Omp19 and Bacterial surface antigen (D15). These epitopes will be further examined in our study.

In order to maintain the proper folding of the protein, we connected GGGS and EAAAK consecutively. Adjuvants are added to the N-terminal of the vaccine peptide to boost its immunogenicity. In order to assess the immunity, stability, protection, and practicality of the final MEV, a construct consisting of 387 amino acids was created by connecting linkers and adjuvants. Furthermore, molecular docking and simulations using molecular dynamics were conducted to confirm the binding strength and durability of the MEV-TLR5. Subsequently, codon adaptation and in silico cloning analyses were conducted to determine the potential codons for expressing the MEV. The findings indicated that the MEV exhibited a significant level of immunogenicity. This work has collectively established a theoretical foundation for the development of a vaccine against Brucella.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12865-025-00728-1.

Introduction

Brucella is a minute, spherical, gram-negative bacterium that lacks buds, flagella, and pods. Brucella is classified into several different categories (B. melitensis, B. abortus, B. suis) and has been well-documented in earlier research [1]. Brucella abortus mostly affects cattle, and although horses may sometimes get the infection, transfer from horses to humans is very uncommon [2]. Brucellosis is a disease that affects ruminant animals and is triggered by the bacteria Brucella abortus. This disease is found globally [3, 4]. The primary clinical symptoms are fever, fatigue, joint discomfort, and muscular soreness [5]. The primary modes of spread are the digestive system, skin, mucosal membranes, blood bodily fluids, and aerosols [6, 7]. The prevalence of human brucellosis among Asian people has seen a significant surge in recent years [8]. Nevertheless, the diagnosis of brucellosis remains challenging due to its non-specific clinical characteristics, the sluggish growth rate in cultures of blood, and the intricacy of serodiagnosis [9]. Diagnosing brucellosis is challenging because Brucella’s delayed proliferation in blood cultures hinders timely detection. Serodiagnosis is complex, hindered by cross-reactivity with different bacteria and fluctuating antibody kinetics, resulting in possible false positives or negatives. Furthermore, this illness has a propensity to progress into a chronic condition, therefore impacting numerous organs simultaneously. Hence, it is essential to devise a novel strategy for the early identification and treatment of brucellosis [10, 11].

Infection with Brucella species triggers an initial inflammatory response, serving as the primary defense against bacterial growth. This early innate immune reaction effectively reduces the bacterial population [11]. The natural immune system then activates cellular immunity, involving CD4 + and CD8 + T lymphocytes, macrophages, dendritic cells (DCs), and pro-inflammatory cytokines like interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α), all contributing to protection [12]. Despite the host’s robust immunological reaction, Brucella abortus is able to persist within macrophages and DCs by expressing several virulence markers. This enables the bacteria to establish a replicative niche and evade immune-mediated elimination [13]. Several virulence factors have been shown to have a role in intracellular replication and evading the immune system. Each of these variables may serve as an immunogenic factor [12, 13].

Vaccines provide an optimal method for preventing brucellosis. Regrettably, a vaccination for human brucellosis is currently unavailable. All animal vaccinations now on the market use live attenuated variants of Brucella, which have the ability to cause abortions in animals that are pregnant and have the potential to infect humans. Thus, there is a pressing need for a highly effective Brucella vaccination in human [14]. Presently, vaccine research employs several techniques such as reverse genetically engineered live-attenuated vaccines, vector-based vaccination, vaccines made from DNA, and multi-epitope vaccines (MEV). Specifically, reverse vaccinology (RV) has shown significant efficacy [14, 15]. Alternatively, the RV technique is used for vaccination prediction, which has the potential to enhance both its safety and effectiveness [15]. The presence of surface antigens such as outer membrane proteins (omp) and Bacterial surface antigen (D15) has been shown and has undergone significant conservation [16]. The process (presence of surface antigens) disturbs the cellular pathways and triggers a reaction from the host’s immune system by releasing substances that have an impact. It also enhances the reproduction of Brucella in the host’s cells and leads to a long-lasting infection [17, 18]. The D15 antigen, or D-15-Ag, has been shown to serve as a target for immune protection and is conserved throughout many strains of Brucella bacteria. The D15 antigen is an essential functional protein in Brucella abortus and constitutes a fundamental component of the bacterium’s surface [19]. Research indicates that the D15 antigen exhibits significant immunogenicity in laboratory infection models, rendering it a promising option for vaccine development [20]. Both proteins are ultimately advantageous for the development of MEVs targeting Brucella abortus.

Recent advancements in genomics have led to the development of DNA vaccines that have the capacity to stimulate both humoral and cellular immune responses. These vaccines also have the ability to extend the duration of antigen expression [21]. Utilizing epitopes in the formulation of this particular vaccination is a novel approach in the creation of multi-epitope DNA vaccines [22, 23]. This technique included making a deliberate choice of antigenic factors that are closely associated with the ability to induce an immune response. A schematic of the advantages of using reverse vaccinology to construct multi-epitope vaccines is shown in Fig. 1.

Fig. 1.

Schematic of the advantages of using reverse vaccinology to construct multi-epitope vaccines

This manuscript is explicitly focused on developing a human brucellosis vaccine. This is a critical distinction because, as stated, existing animal vaccines are live-attenuated strains that pose significant risks to human health, including the potential to cause disease in humans and abortions in pregnant animals. Therefore, the justification for this work is the urgent, unmet public health need for a safe and effective human vaccine to protect individuals at risk of exposure, such as farmers, veterinarians, and laboratory workers, and to contribute to global disease control efforts by reducing the human burden of this debilitating zoonosis. This work used bioinformatics algorithms to determine antigenic factors from B. abortus and then created a multi-epitope recombinant DNA vaccine. The study investigated the humoral, cell-mediated, and protective immunity elicited by this multi-epitope DNA vaccination in BALB/c mice.

Materials and methods

Selection of target proteins

Vaxign is a technology that predicts and analyzes vaccination targets using reverse vaccinology. The work utilized the Vaxign dataset [http://www.violinet.org/vaxign/] to find 30 genome-wide sequences of Brucella sp. The target genes in the Vaxign dataset are chosen using a reverse vaccinology methodology, utilizing bioinformatics criteria including subcellular localization (e.g., surface-exposed, secreted), absence of transmembrane helices, adhesin probability, conservation across pathogenic strains, and dissimilarity to host proteins to prevent autoimmunity. Protein selection for de novo genome-wide prediction of novel vaccination candidates was not exclusively based on prior research, avoiding restriction to previously identified or examined antigens. The Vaxign dataset is crucial as it provides a systematic, high-throughput computational analysis of complete pathogen proteomes to uncover possible vaccine targets that conventional approaches may overlook, hence expediting vaccine development.

The genome of B. abortus 2308 was chosen, and 20 genes were identified based on their localization in the outer membrane, absence of homology with host proteins. The requirements were at least one transmembrane helix (≤ 1), a likelihood of adhesion above 0.51, and an absence of homology to molecules found in humans or mice. In order to reduce the likelihood of a host cell interacting with the vaccination, the vaccine candidate must not exhibit homology with human proteins [22, 23]. To achieve this objective, protein sequences were examined using the BLASTp web server in the NCBI database, comparing them to the host proteome of Homo sapiens. If the BLASTp web service confirms the protein specificity, Brucella sp. proteins will be utilized for further investigation. The UNIPORT database was used to examine the existence of various proteins in the Brucella sp. proteome.

Prediction of potential vaccine candidate from target proteins

The ProtParam program, available at http://web.expasy.org/protparam/, was used to examine the physicochemical characteristics of proteins. The molecular equation, the general average of the water solubility (GRAVY), the instability index, and the number of amino acids were all included. The antigenic properties of the selected antigens were analyzed using VaxiJen 2.0, with a minimum score of 0.5 [23]. The website http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html provides further information on VaxiJen 2.0. The VaxiJen 2.0 database offered predictions of possible proteins that relied not on sequence alignment but on physicochemical properties. The molecule sequences were automatically converted using cross-covariance (ACC) into a homogeneous vector representing the principal amino acid characteristics. An antigenic threshold of 0.5 was applied to each bacterial molecule. In addition, we compared the hydrophilic properties and durability of the selected protein. Furthermore, we assessed its allergenicity using the online platform AllergenFP v.1.0 [24]. The CELLO program, available at http://cello.life.nctu.edu.tw/, was employed to accurately establish the location of molecules based on a Localization Reliability score of ≥ 1.5. This software can identify if the protein is located on the outer or inner membrane and determine whether it is cytoplasmic or periplasmic [25]. The SOLpro program utilizes a sequence-based forecasting methodology to forecast protein solubility through a two-step SVM approach. Thus, SOLpro was used to forecast the soluble state of the molecules [23–25].

Forecasting of T-cell epitopes

Cytotoxic T-lymphocytes (CTLs) are crucial in the host’s immune reaction. These cells can promptly interact with and eradicate the harmful cells within the immunity system. The NetCTL v1.2 website, which may be found at http://www.cbs.dtu.dk/services/NetCTL/, possesses remarkable predictive powers. The technology utilizes artificially generated neural networks and an array of weights to predict 9-mer CTL epitopes. This tool is designed to anticipate the CTL epitopes with high combination scores among the selected sequences of proteins. A threshold score of 0.90 was employed to get a sensitivity value of 0.74 and an accuracy of 0.98. The MHC-I binding variants for each CTL epitope were found utilizing the MHC-I binding tool provided on the IEDB webpage (http://tools.iedb.org/mhci/) utilizing the consensus technique. We utilized the consensus approach to forecast the binding genotypes of the proteins, according to a percentile score threshold of 2 or below in order to maintain consistency. The antigenic properties, allergic profile, toxicity prediction, and immunogenicity of each CTL epitope were evaluated using specific online tools: the VaxiJen v2.0 online for antigenic properties, the AllerTOP v2.0 webpage for allergenic characteristics, the ToxinPred server for toxicity prediction, and the IEDB Class I Immunogenicity tool for immunogenicity.

Evaluation of helper T-Lymphocyte (HTL) epitope

Adaptive immune responses consist of helper T lymphocytes (HTLs) that may recognize foreign antigens and stimulate B-cells and CTLs to eradicate pathogenic pathogens. The MHC-II binding program from the IEDB resource (http://tools.iedb.org/mhcii/) was used to identify 15-mer HTL epitopes within the selected sequences of protein. We utilized the consensus approach to forecast the binding genotypes of the proteins, according to a percentile score threshold of 2 or below in order to maintain consistency.

Linear B cell lymphocytes (LBL)

B-cell epitopes are essential for augmenting humoral or antibody-mediated immune responses. In order to predict the epitopes of linear B cell lymphocytes (LBL), we used the IEDB website (http://tools.iedb.org/main/bcell/) with the default configurations. The epitopes were assessed utilizing the VaxiJen v2.0 and AllerTop v2.0 systems. The length of projected B-cell epitopes surpassing six amino acids was ascertained.

Exploration of major histocompatibility complex (MHC) clusters

The Major Histocompatibility Complex (MHC) genomic area in most mammals displays notable variability. The human MHC genomic area, often known as HLA, has a significant level of genetic variation, including several alleles. The potential distinct specificity of the MHC allele is yet unknown in most cases. Cluster analysis of MHC variants may be utilized to find MHC molecules with similar binding specificities. The study utilized the MHCcluster 2.0 webpage (http://www.cbs.dtu.dk/services/MHCcluster/) to generate phylogenetic tree-based representations and user-friendly heatmaps of the functional cluster among MHC variants. For this reason, the settings utilized were the default ones. The MHC class I clustering study utilized the NetMHCpan-2.8 technique, which employed a prevalent and described HLA module. The suitable DRB allele modules were selected for MHC class II clustering studies. For our study, we utilized MHCcluster 2.0 to functionally cluster MHC variants, aiming to identify molecules with similar binding specificities. Our selection of MHC alleles was guided by the default settings of the NetMHCpan-2.8 technique for MHC Class I and suitable DRB allele modules for MHC Class II. This ensured we included prevalent and well-described HLA modules, covering common human MHC alleles that are crucial for broad population vaccine coverage and for which robust peptide binding data exist. Our approach thus leveraged established computational tools and extensive datasets to systematically analyze MHC-peptide interactions.

The development of a multi-epitope vaccination

The vaccine was created by mixing the chosen LBL epitopes with a suitable adjuvant and linking them using suitable linkers. LBL epitopes, or Linear B Lymphocyte epitopes, are distinct, contiguous sequences of amino acids within an antigen immediately identified by B cell receptors and antibodies. In contrast to conformational epitopes, their identification relies solely on the primary amino acid sequence rather than the protein’s overall three-dimensional folded shape. The adjuvant used in this investigation was a Toll-like receptor 5 (TLR5) agonist because extramembrane molecules can recognize TLR5. Adjuvants are essential in overcoming the constraints of translation and biosynthesis. The immunological efficacy of the vaccine candidate was evaluated by assessing the RS09/Hsp70 adjuvant, a synthetic agonist of Toll-like receptor-5. This process enhances the activation of both the innate and adaptive functions of the immune system’s function. The receptors for toll-like molecules and antigen-presenting cells (APCs) initiate natural immunity. The EAAAK linker enabled the linkage between multi-epitope peptides and adjuvants. In contrast, the selected distinct proteins were linked together utilizing GGGS linkers.

Structural analysis of the vaccine

The fundamental characteristics of an amino acid are delineated by its physiochemistry. The ProtParam website (https://web.expasy.org/protparam/) was employed to predict the physicochemical properties of the vaccine in order to get a comprehensive understanding of its essential role. The antigenic properties were evaluated utilizing sophisticated tools consisting of VaxiJen v2.0 (http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.html) and SOLpro (https://protein-sol.manchester.ac.uk/). The SOPMA website (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_sopma.html) and the PSIPRED v4.0 web page (http://bioinf.cs.ucl.ac.uk/psipred/) employ default configurations to detect the two-dimensional (2D) structural components of the construct, such as the α-helix, β-turn, and random coils. The predictive accuracy of SOPMA surpasses 80%. To better comprehend the vaccine’s compositional excellence, we acquired and assessed 2D structural data.

Homology modeling to verify the accuracy of three-dimensional structures

The vaccine composition’s tertiary structure (3D) system was generated utilizing the homology modeling application available on the SWISS-MODEL (https://swissmodel.expasy.org/). The GalaxyRefine module, part of the GalaxyWEB technology (http://galaxy.seoklab.org/), enhanced the 3D vaccination model. The result produces five modified models, each including diverse attributes such as rmsd, GDT-HA, clash rating, poor rotamers, MolProbity, and Rama-favored. The ProSA-web server, available at https://servicesn.mbi.ucla.edu/PROCHECK/, authenticated the projected models. This website computes the Z-score and evaluates the stereochemical stability of each protein structure by analyzing the geometry of individual residues and the whole structure geometry. The Ramachandran plot underwent further analysis utilizing the Procheck website server to evaluate the general quality of the enhanced 3D vaccine structure. The Ramachandran plot is a graphical depiction that showcases the abundance of amino acid residues throughout various places, including the most favored, disallowed, lavishly allowed, and exceptionally allowed areas. This picture is derived from the dihedral angles psi (β) and phi (ι) linked to each amino acid. A model is considered adequate when more than 90% of residues are situated inside the most desired regions.

Prediction of conformational B-cell epitopes

The ElliPro software from the IEDB website (http://tools.iedb.org/ellipro/) was utilized to predict the discontinuous or structural B-cell epitopes in the vaccine that was created. The default values were preserved, including a maximum range of 6 ± and a minimal value of 0.5. The results originate from an altered version of Thornton’s methodology, which integrates residue clustering techniques. The prediction considers the residual protein index (PI), neighbor peptide clustering, and peptide conformation.

Computational modeling of the interaction between the TLR-5 receptor and the vaccine construct

The TLR5 molecules with the Protein Data Bank NP_003259.2 (https://www.ncbi.nlm.nih.gov/protein/NP_003259.2) were acquired from the NCBI Data Bank. The Discovery Studio program removed the heteroatoms and the B, C, and D chains. The peptide was subjected to energy reduction utilizing the GROMOS 43B1 field of force in the Swiss PDB Viewer tool. Afterwards, the vaccine structure and the TLR5 complexes were subjected to docking research utilizing the HADDOCK software platform (https://rascar.science.uu.nl/haddock2.4/). This program is an online application created explicitly for the computerized docking of protein-protein or peptide-protein interactions. For our HADDOCK docking simulations, we identified the active binding sites of the proteins using a combination of computational prediction and established biological knowledge. Specifically, we leveraged tools like CPORT, which predicts protein-protein interaction interfaces based on a consensus of various algorithms, to define “active” residues. These active residues represent the key amino acids predicted to be directly involved in the interaction interface. Furthermore, we incorporated “passive” residues, defined as solvent-accessible surface neighbors of the active residues, to account for the broader potential interaction surface. This information, provided as Ambiguous Interaction Restraints (AIRs) to HADDOCK, effectively guided the docking process by biasing the exploration towards these predicted interaction hotspots, thereby enhancing the accuracy and efficiency of identifying plausible binding poses for our vaccine construct. The current docking technique evaluates several candidate complexes with favorable surface mutually beneficial relationships. The program produces a succinct compilation of probable complexes using clustering characteristics.

Molecular dynamics simulation evaluation

The iMODS modeling website, accessed at https://imods.iqfr.csic.es/, provides an open-source platform that simplifies performing internal average mode assessment (NMA) on the given protein. This study aims to assess the protein’s deformation, stability, and mobility. One may use Normal Mode Analysis (NMA) or simulate probable trajectories connecting two conformations, allowing the display of possibilities in three dimensions, even in many biomolecular mixtures. This work used the iMODS technique to evaluate the fluctuating stability of the TLR5-vaccine complexes.

Utilizing codon adaptability and in silico cloning techniques

Efficiently optimizing codons is essential for efficiently expressing a foreign gene in a host organism. Consequently, the structure was made ready for codon modification utilizing the JCat software platform, which may be accessed at http://jcat.de/. The Lactococcus lactis (L. lactis) strain was employed as the host organism in this investigation. The whole process was implemented with a rigorous focus on three essential criteria:

• The act of evading areas where restriction enzymes cleave DNA.

• The act of evading binding locations on ribosomes in prokaryotes.

• The action of preventing the termination of rho-independent transcription.

The evaluation of the modified sequence included using the codon adaptation index (CAI) value and the guanine-cytosine (GC) ratio. Utilizing computational approaches, the tailored nucleotide arrangement was cloned into the pNZ8121 expression plasmid. The full in silico cloning method utilized the SnapGene v4.2 software. The reliability of thermodynamics and the effectiveness of translating expressed mRNA sequences were evaluated utilizing the RNAfold webpage (http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi). The bioinformatics studies of this research were done only with BALB/c mice.

Results

A possible vaccination candidate has been identified

The Vaxign algorithm initially selected 50 proteins using a reverse vaccinology technique, emphasizing attributes essential for vaccine production. Essential criteria encompassed subcellular localization, emphasizing surface-exposed or secreted proteins available to the host immune system and excluding many transmembrane helices to facilitate effective recombinant production. Furthermore, proteins with a high adhesin likelihood were preferred due to their significance in infection. To ensure broad-spectrum efficacy, conservation across pathogenic Brucella strains was required. Ultimately, proteins exhibiting little resemblance to host proteins were selected to reduce autoimmune hazards. Out of these 50 proteins, 33 were selected based on these criteria. This additional selection likely included predictions of virulence factors generated by the Vaxign system, pinpointing proteins essential for pathogenicity. The selection likely considered the critical role of the proteins in bacterial survival, rendering them optimal targets. Moreover, applying antigenicity prediction scores, derived from machine learning methods like Vaxign-ML, would have been pivotal. This facilitated the incorporation of innovative vaccine candidates based on their immunogenic potential, irrespective of their prior classification as direct virulence factors.

The identified proteins were categorized as extracellular, outer membrane, periplasmic, inner membrane, and cytoplasmic proteins based on the results obtained from the CELLO programs. The specificity of the selected proteins for Brucella sp. was confirmed using BLASTp protein analysis, and the homology with human proteins was done. Table 1 displays the results of the preliminary examination of 8 proteins from the Brucella species. The proteins omp19 and Bacterial surface antigen (D15) were identified as the most suitable for immunogenicity due to their high solubility (Fig. 2A), flexibility, and antigenic values of 0.65 and 0.60, respectively, as determined in the first screening (Table 1).

Table 1.

Initial screening of Brucella sp. proteins

UniProt ID	Protein Names	Organism	Length	Subcellular Location	Score	SOLpro	Antigenicity score
A0A0D4D953	19 kDa outer membrane protein	Brucella abortus	177 AA	Extracellular	1.290	0.644	0.65
				Cytoplasmic	1.014
				InnerMembrane	0.961
				Periplasmic	0.873
				OuterMembrane	0.862
A0A0F6ANH2	Bacterial surface antigen (D15)	Brucella abortus	653 AA	OuterMembrane	4.096	0.505	0.60
				Extracellular	0.395
				Periplasmic	0.314
				Cytoplasmic	0.110
				InnerMembrane	0.085
Q2YJB2	Heme transporter BhuA	Brucella abortus	661 AA	OuterMembrane	4.632	0.277	0.60
				Extracellular	0.255
				Periplasmic	0.073
				InnerMembrane	0.028
				Cytoplasmic	0.012
A0A0F6ANF8	UPF0324 membrane protein	Brucella abortus	336 AA	InnerMembrane	4.912	0.323	0.59
				Cytoplasmic	0.034
				Extracellular	0.020
				Periplasmic	0.020
				OuterMembrane	0.013
A0A0F6ANE1	ABC transporter, ATP-binding protein	Brucella abortus	546 AA	Cytoplasmic	3.106	0.285	0.55
				InnerMembrane	1.371
				Periplasmic	0.461
				OuterMembrane	0.048
				Extracellular	0.013
A0A0F6ANG8	Ubiquinol oxidase subunit 2	Brucella abortus	337 AA	InnerMembrane	2.964	0.323	0.54
				Periplasmic	1.569
				Cytoplasmic	0.335
				OuterMembrane	0.094
				Extracellular	0.039
A0A0F6ANE2	Binding-protein-dependent transport systems inner membrane component	Brucella abortus	379 AA	InnerMembrane	4.959	0.188	0.51
				Cytoplasmic	0.015
				Periplasmic	0.012
				OuterMembrane	0.011
				Extracellular	0.003
A0A0F6ANE3	Binding-protein-dependent transport systems inner membrane component	Brucella abortus	365 AA	InnerMembrane	4.832	0.260	0.35
				Cytoplasmic	0.065
				OuterMembrane	0.063
				Periplasmic	0.033
				Extracellular	0.007

The bolded portion was selected for further screening

Fig. 2.

A Solubility of Brucella abortus selected proteins. omp19 and Bacterial surface antigen (D15) proteins obtained a solubility score higher than 0.45 and were selected for further screening. The red line indicates the threshold (0.45). B Schematic sequence of the Constructed vaccine

Evaluation of CTL, HTL, and LBL epitopes

An extensive investigation was performed on the selected proteins to discover more than 300 potential CTL, HTL and LBL epitopes that were immunologically active, non-allergenic, non-toxic, and efficient within the specified range of selection. After a comprehensive screening approach, we have found a collection of 169 CTL epitopes from each protein that exhibits the potential for developing a multi-epitope vaccine (Supplementary material 1).

Potential HTL epitopes were discovered out of the 645 selected epitopes. Based on the specified range of selection, these epitopes were immunological, non-allergenic, and non-toxic. However, the specific data has yet to be available for display. The ability of the selected HTL epitopes to induce cytokines was evaluated, resulting in the discovery of potential epitopes from the proteins for the development of a multi-epitope vaccine (Supplementary material 2).

The selected proteins yielded potential LBL epitopes; all confirmed to be non-allergenic, immunological, and non-toxic (Supplementary material 3). The selection of potential LBL epitopes from each protein was based on many parameters, such as their non-allergenic properties, high antigenicity, toxicity features, and probability ratings. Subsequently, these specific antigenic determinants were selected to develop a vaccine that contains many epitopes. Table 2 presents a collection of the CTL, HTL, and LBL epitopes chosen to create vaccines.

Table 2.

Final overlapped LBL, CTL, and HTL epitopes of the two target proteins

Protein	Epitopes		Immunogenicity THRESHOLD	Antigenicity	Allergenicity	Toxicity	Epitope conservancy hit (%)
omp19	CTL	KIATPQTKY	0.6542	0.9770	non-allergen	non-toxin	100
		SPPPPPAPV	0.0532	0.9784	non-allergen	non-toxin	100
		MSAQSGTQV	0.7213	0.9118	non-allergen	non-toxin	100
		PNAPSTDMS	0.0493	0.0652	non-allergen	non-toxin	100
	HTL	ASLLSLAAAGIVLAG	---	1.1253	non-allergen	non-toxin	100
		LYDANGGTVASLYSS	---	1.1396	non-allergen	non-toxin	100
		LVLYDANGGTVASLY	---	1.0372	non-allergen	non-toxin	100
	LBL	SSRLGNLDNVSPPPPPAPVNAVPAGTVQKGNLDSPTQ FPNAPSTDMSAQSGTQVASLPPASAPDLTP	0.5	0.6178	non-allergen	non-toxin	100
		SASLGG	0.5	1.8193	non-allergen	non-toxin	100
		QTKYGQGYRAGPLRCPGELANLASWAV	0.5	1.2937	non-allergen	non-toxin	100
		SSGQGRFDGQTTGGQA	0.5	1.9213	non-allergen	non-toxin	100
Bacterial surface antigen (D15)	CTL	IIDPKTYSV	0.9761	1.6015	non-allergen	non-toxin	100
		QEAKSGTIL	0.8144	1.4291	non-allergen	non-toxin	100
		LTAHFSSFA	0.1805	1.8145	non-allergen	non-toxin	100
		LLAKARGDY	2.9350	1.2565	non-allergen	non-toxin	100
		ALDPGRKAY	2.9210	1.1017	non-allergen	non-toxin	100
		SDELSGALY	2.4210	0.7574	non-allergen	non-toxin	100
		FTQIFSDEL	0.7760	0.8823	non-allergen	non-toxin	100
		KPDPSSGFY	2.6920	0.9910	non-allergen	non-toxin	100
		AVAFVDAGY	3.1940	1.0146	non-allergen	non-toxin	100
		SGDPNYGFY	2.6670	1.5970	non-allergen	non-toxin	100
	HTL	RGDYRRILSALYGEG	---	1.7797	non-allergen	non-toxin	100
		GDYRRILSALYGEGR	---	0.6380	non-allergen	non-toxin	100
	LBL	DGREANDIPPDTEIPN	0.5	0.7231	non-allergen	non-toxin	100
		IAPPPGNRRDKVQTPEEAGFAPGQEAKSG	0.5	0.7271	non-allergen	non-toxin	100
		KNADGKEADLKS	0.5	2.9986	non-allergen	non-toxin	100
		VSDADKPASGSAG	0.5	1.1388	non-allergen	non-toxin	100
		TARMDPQF	0.5	1.7069	non-allergen	non-toxin	100
		LKPGQEYDPDDIEN	0.5	0.5070	non-allergen	non-toxin	100
		YSTIDGFGV	0.5	0.5842	non-allergen	non-toxin	100
		SGIGGSQDNSFDPKNYTY	0.5	0.9504	non-allergen	non-toxin	100
		VYTPDT	0.5	0.5147	non-allergen	non-toxin	100
Threshold			More than 0.5				More than 80%

The bolded portion was selected for further screening

Progress in vaccine formulation

The vaccine was created by integrating the most advantageous LBL, CTL, and HTL epitopes using GGGS linkers. The RS09/Hsp70 adjuvant was chemically bonded to the front of the vaccination utilizing the EAAAK linker. The present study involved the development of a multi-epitope vaccination. For this reason, two CTL epitopes, two HTL epitopes, and two LBL epitopes were chosen from the Brucella abortus protein sequences. The selection criteria encompassed an antigenicity score, immunogenicity threshold over 0.5, and an epitope conservancy hit percentage surpassing 80%. A vaccine construct consisting of 387 amino acid residues was generated. Table 3 displays the chosen LBL, CTL, and HTL epitopes for the final creation of the vaccine. Furthermore, Fig. 2B displays the schematic representation of the placement of epitopes and the production of the vaccine.

Table 3.

The physicochemical characteristics of the vaccine constructs produced in this study

Red color: Linker

MHC cluster evaluation

The MHCcluster v2.0 website was employed to group the MHCI and MHCII alleles that interact with the epitopes produced from the selected structural proteins. The study incorporated HLA-A01:01, HLA-A02:01, HLA-A03:01, HLA-A24:02, HLA-A26:01, HLA-B07:02, HLA-B08:01, HLA-B27:05, HLA-B39:01, HLA-B40:01, HLA-B58:01, HLA-B15:01 alleles from HLA supertype representative primary histocompatibility complex (MHC) class. The study incorporated DRB1_0101, DRB1_0301, DRB1_0401, DRB1_0701, DRB1_0802, DRB1_0803, DRB1_0901, DRB1_1101, DRB1_1302, DRB1_1402, DRB1_1501, DRB3_0101, DRB4_0101, DRB5_0101 alleles from MHC class II. The clustering outcomes for the MHCI and MHCII alleles are depicted in Fig. 3. In addition, Fig. 3A and C present an intricate tree map that demonstrates the cluster study of MHCI and MHCII. The red areas on the heatmap (Fig. 3B and D) indicate the more pronounced interactions among the clustered HLA alleles, whereas the red regions represent less pronounced interactions.

Fig. 3.

The results are obtained from the MHC clustering investigation. The included visualizations are the MHC class-I cluster assessment tree map (A), the advanced heatmap (B), the tree map (C), and the advanced heatmap (D) of the MHC class-II clustering analysis

A comprehensive assessment of physicochemical characteristics

An assessment was conducted to evaluate the physicochemical properties of the formulated construct. The vaccine was determined to possess a chemical formula of C₁₆₆₈H₂₆₅₁N₅₀₃O₅₆₈S₃. The molecular weight of the vaccine design was determined to be 38935.61Da, suggesting an average weight. The vaccine had an acidic character, as shown by its theoretical PI value 6.16. Furthermore, the aliphatic index was determined to be 59.35, while the instability score was calculated to be 52.29. The hydrophilic character of the construct was indicated by the grand average of hydropathicity (GRAVY) score of −0.641. The total number of negatively charged residues (Asp + Glu) was 39, while the total number of positively charged residues (Arg + Lys) was 37. Multiple systems were used to evaluate the allergenicity and antigenic properties of the developed vaccine. The antigenicity scores obtained from the Vaxijen2.0 servers were 1.0422. Crucially, the vaccination was shown to be both antigenic and non-allergenic in all the servers. In addition, the solubility of the vaccine design was assessed using Protein-Sol servers, yielding a score of 0.553. The prediction findings showed that the vaccine design did not exhibit any TM helices and signal peptides (Table 4).

Table 4.

Antigenic properties, allergenicity, and physicochemical properties of the final constructed vaccine

Row	Characteristics	Finding	Remark
1	Number of amino acids	387	Suitable
2	Molecular weight (Da)	38935.61Da	Average
3	Theoretical pI	6.16	Acidic
4	Chemical formula	C1668H2651N503O568S3	---
5	Instability index of vaccine	52.29	Stable
6	Aliphatic index of vaccine	59.35	Thermostable
7	GRAVY	−0.641	Hydrophilic
8	Antigenicity	1.0422	Antigenic
9	Immunogenicity	Positive	Immunogenic
10	Allergenicity	No	Non-allergen
11	Solubility	0.553	Soluble
12	Alpha helix (Hh)	78	19.95%
13	Extended strand (Ee)	53	13.55%
14	310 helix (Gg)	0	0.00%
15	Pi helix (Ii)	0	0.00%
16	Beta bridge (Bb)	0	0.00%
17	Beta turn (Tt)	40	10.23%
18	Bend region (Ss)	0	0.00%
19	Random coil (Cc)	220	56.27%
20	Ambiguous states	0	0.00%

Secondary structure prediction

Two servers were used to evaluate the secondary structural characteristics of the manufactured vaccine, such as random coils, alpha-helix, and beta-turn. The SOPMA website projected that the final vaccine build would consist of 19.95% α-helix, 10.23% β-strand, 13.55% Extended strand, and 56.27% random coils (Table 4). Also, the PSIPRED server predicted the final structure of the vaccine (Supporting Information 4).

Modeling homology, refining 3D structure, and validating

The 3D structure of the final vaccine was generated using the SWISS-MODEL platform. The server produced 5 models for the provided build. Subsequently, the models were evaluated using the ProSA websites. In our current investigation, the initial crude model 4, which underwent pre-refinement, had the highest Z-score of −3.22. Models 1, 2, 3, 4 and 5 had Z-Score of −1.51, −1.79, −1.88, −3.22 and − 1.91, respectively as shown in Fig. 4. Additionally, Models 1, 2, 3, 4 and 5 had 83.54%, 94.07%, 98.32%, 91.06% and 97.47% of the residues (Ramachandran Favored) were in the most preferred locations. Of these, models 3 and 5 with Ramachandran Favored above 95% were further analyzed. Table 5 shows that model 3 did not have a bad band (0/954) and its angle was bad (10/1284). While model 5 had 4 bad bands (4/647) and bad angle (39/872).

Fig. 4.

Modeling homology and validating of 5 predicted vaccine models based on ProSA and SWISS-MODEL platform

Table 5.

The molprobity results of examining the best vaccine models and choosing a better construct

MolProbity Results	Model 3		Model 5
MolProbity Score	0.68	---	1.83	---
Clash Score	0.54	---	16.77	---
Ramachandran Favoured	98.32%	---	97.47%	---
Ramachandran Outliers	0.84%	---	1.27%	---
Rotamer Outliers	0.00%	---	0.00%	---
C-Beta Deviations	0	---	6	A73 ALA, A72 HIS, A71 ALA, A41 ILE, A32 ILE, A75 GLU
Bad Bonds	0/954	---	4/647	A41 ILE-A42 ALA, A41 ILE, A72 HIS, A71 ALA
Bad Angles	10/1284	A65 THR, A5 HIS, (A117 TYR-A118 ARG), A25 ASP, A117 TYR, A56 HIS, A72 HIS, (A3 PRO-A4 PRO)	39/872	A44 HIS, A32 ILE, A34 VAL, A71 ALA, (A72 HIS-A73 ALA), (A74 GLU-A75 GLU), A69 ASN, A72 HIS, A73 ALA, A75 GLU, (A33 GLN-A34 VAL), A74 GLU, (A73 ALA-A74 GLU), (A12 ILE-A13 PRO), (A58 THR-A59 ALA), A43 ALA, A65 THR, A14 THR, (A32 ILE-A33 GLN), A77 ARG, (A41 ILE-A42 ALA), A41 ILE, (A78 LYS-A79 ARG), (A43 ALA-A44 HIS), A19 THR
Twisted Non-Proline	1/5	(A12 ILE-A13 PRO)	1/78	(A75 GLU-A76 ASP)

Model 3 was uploaded to the GalaxyRefine platform for refining. Additionally, this service generates five models for the provided basic model. After refining, it was observed that all of the models had enhanced Rama-favored areas compared to the first presented crude model. Of all the developed models, model 3 was identified as the most optimal refined model. The GalaxyRefine server (Table 6) demonstrated favorable GDT-HA (0.9174), RMSD (0.474), MolProbity (0.980), clash score (2.1), unfavorable rotamers (0.0), and Rama-favored (99.2%) scores. The enhanced models underwent further validation via the ProSA websites. This study observed that model 3 had the highest Z-score (−1.93) and successfully identified 99.2% of the amino acid residues in the most preferred parts of the Random Forest plot. Therefore, in the current investigation, model 3 has been chosen for further research.

Table 6.

Structural information of the top 5 vaccine models

Model	GDT-HA	RMSD	MolProbity	Clash score	Poor rotamers		Rama favored
Initial	1.0000	0.000	0.812	1.1	0.0		98.3
MODEL 1	0.9298	0.486	1.118	3.1	1.0		99.2
MODEL 2	0.9318	0.495	1.200	4.2	0.0		99.2
MODEL 3	0.9174	0.474	0.980	2.1	0.0		99.2
MODEL 4	0.9256	0.498	1.105	3.1	0.0		98.3
MODEL 5	0.9277	0.470	1.118	3.1	1.0		99.2

Disulfide mutation analysis using the Disulfide by Design v2.12 online platform
Res1 Chain	Res1 Seq #	Res1 AA	Res2 Chain	Res2 Seq #	Res2 AA	Chi3	Energy	Sum B-Factors
A	39	ARG	A	44	HIS	−71.83	2.75	0.7
A	59	ALA	A	69	ASN	−83.37	3.77	0.77
A	61	LYS	A	65	THR	−94.18	2.23	0.83
A	80	ARG	A	84	ASP	−71.24	1.53	0.89

In disulfide engineering, 9 amino acid residues were identified as appropriate for inducing disulfide mutation using the Disulfide by Design v2.12 online platform (Table 6). Upon careful assessment of the Chi3 angle and the energy score, a final selection was made of 1 pairings of amino acids that meet the specified parameters. Specifically, the Chi3 angle should fall within the range of − 87 and + 97°, while the energy score should not surpass 2.2 kcal/mol. The study generated two mutations in the residue pairs ARG80- ASP84. The corresponding Chi3 angles were determined to be −71.24, with corresponding energy scores of 1.53 kcal/mol, respectively.

Furthermore, an examination was conducted on the proposed vaccine with regards to its third structure. Based on the findings of the SWISS-MODEL analysis, it was determined that the vaccine structure did not have a bad band. In the B-factor of the vaccine structure, the QMEAN local scores were found to be 0.40 ± 0.08.

Screening for conformational B-Cell epitopes

Three conformational B-cell epitopes were found in the vaccine construct sequence utilizing the ElliPro program (Fig. 5A and B). The estimated discontinuous and linear epitope ranges are also shown in Supporting Information 5. Furthermore, the anticipated conformational B-cell epitope score exhibited values between 0.631 and 0.812.

Fig. 5.

A The 3D representations of predicted conformational B-Cell Epitope. B Performing molecular docking between the structure of the vaccination and the TLR5 receptor. The highest-ranking cluster was chosen as the ultimate structures. C The vaccine’s interacting residues are shown

Molecular docking of the TLR5 and vaccine complex

Facilitating significant interaction between immune system cells and the vaccine formulation is crucial for attaining a strong and reliable immunological response. HADDOCK clustered 38 structures, which represents 19% of the water-refined models HADDOCK generated. The statistics of the top 7 clusters are shown in Table 7. The top cluster is the most reliable according to HADDOCK. Its Z-score indicates how many standard deviations from the average this cluster is located in terms of score (the more negative the better).

Table 7.

Cluster scores of docked vaccine TLR5 and vaccine complex

Property	Cluster 4	Cluster 2	Cluster 5	Cluster 7	Cluster 6	Cluster 1	Cluster 3
HADDOCK score	156.5 +/- 14.8	160.3 +/- 25.9	167.2 +/- 30.8	193.1 +/- 21.6	201.4 +/- 42.1	209.9 +/- 23.1	226.9 +/- 12.4
Cluster size	5	6	5	4	4	8	6
RMSD from the overall lowest-energy structure	11.2 +/- 0.4	14.5 +/- 0.3	13.4 +/- 0.1	19.5 +/- 0.3	20.6 +/- 0.1	16.2 +/- 0.2	18.8 +/- 0.4
Van der Waals energy	−67.3 +/- 5.0	−76.1 +/- 9.2	−66.3 +/- 12.4	−53.1 +/- 8.3	−77.2 +/- 11.2	−61.0 +/- 5.7	−62.8 +/- 8.0
Electrostatic energy	−518.2 +/- 44.4	−525.5 +/- 57.6	−611.6 +/- 42.4	−657.8 +/- 41.8	−501.9 +/- 86.5	−433.2 +/- 36.9	−278.7 +/- 88.8
Desolvation energy	14.5 +/- 6.6	−4.0 +/- 4.6	13.1 +/- 5.1	9.9 +/- 5.5	−22.4 +/- 4.4	4.1 +/- 7.3	−15.7 +/- 4.8
Restraints violation energy	3130.0 +/- 161.4	3454.4 +/- 105.8	3426.7 +/- 171.1	3678.1 +/- 118.8	4014.0 +/- 281.0	3534.1 +/- 207.0	3611.0 +/- 128.5
Buried Surface Area	3138.8 +/- 65.4	3571.3 +/- 234.7	3401.2 +/- 112.3	3065.9 +/- 336.5	3352.0 +/- 187.0	2784.1 +/- 220.7	2755.9 +/- 221.3
Z-Score	−1.3	−1.1	−0.8	0.2	0.5	0.9	1.6

The Electrostatic energy values of the 7 most significant clusters were − 518.2 +/- 44.4, −525.5 +/- 57.6, −611.6 +/- 42.4, −657.8 +/- 41.8, −501.9 +/- 86.5, −433.2 +/- 36.9 and − 278.7 +/- 88.8 kcal/mol. The cluster 4 was selected because to its higher HADDOCK score (156.5 +/- 14.8) in comparison to the other clusters. Furthermore, Fig. 5B displayed the top molecular docking models of cluster 4. Figure 5C depicts the alignment of the proposed vaccination model with the TLR-5 receptor in three dimensions. A graphical representation of the results is also provided at the Fig. 6A and B.

Fig. 6.

The statistics of the top 7 clusters of TLR5/Vaccine interaction. A Diagrams related to electrostatic energy, HADDOCK score, de-solvation energy and Restraints violation energy individually. B Graphs related to average electrostatic energy, Van der Waals energy and Restraints violation energy. According to the results, Cluster 4 had the best molecular docking result between the vaccine structure and the TLR5 receptor

Normal-mode immune simulation analysis

The vaccine complexes were subjected to a normal-mode examination utilizing iMODS approaches (https://imods.iqf.csic.es/). This evaluation included assessing several descriptors to understand the stability and physical movement of the atoms inside the vaccine components. The regions of protein deformability are shown by the peaks illustrated in Fig. 7A. The graph depicting the B factor offers a visual illustration and comparison of the NMA and PDB attributes of the complexes, as shown in Fig. 7B. The eigenvalue graph shown in Fig. 7C indicates that the given complex has an eigenvalue of 7.552102e-04. Figure 7D illustrates the variability across individuals in the color purple, while the color green reflects the total variation. The elastic maps shown in Fig. 7E demonstrate the interatomic linkages within the complexes, where a darker shade of grey signifies a greater level of rigidity. Figure 7F depicts the covariance map, where the red color indicates the existence of correlated mobility among pairs of residues, the white color represents uncorrelated motion, and the blue color indicates anticorrelated motion.

Fig. 7.

Molecular dynamics simulation of a multi-epitope vaccine. A The experimental B-factor is taken from the corresponding PDB field and the calculated from NMA is obtained by multiplying the NMA mobility by (8pi^2). Be aware that many PDB files of averaged NMR models contain no B-factors (actually, the B-factor column gives an averaged RMSD); B The main-chain deformability is a measure of the capability of a given molecule to deform at each of its residues. The location of the chain ‘hinges’ can be derived from high deformability regions; C The eigenvalue associated to each normal mode represents the motion stiffness. Its value is directly related to the energy required to deform the structure. The lower the eigenvalue, the easier the deformation; D The variance associated to each normal mode is inversely related to the eigenvalue. Colored bars show the individual (red) and cumulative (green) variances; E complex elastic network, Covariance matrix indicates coupling between pairs of residues, i.e. whether they experience correlated (red), uncorrelated (white) or anti-correlated (blue) motions. F The elastic network model defines which pairs of atoms are connected by springs. Each dot in the graph represents one spring between the corresponding pair of atoms. Dots are colored according to their stiffness; the darker grays indicate stiffer springs and vice versa

Codon adaptation and in silico cloning

In order to enhance the effectiveness of vaccine translation, we utilized the JCat software to alter the codons obtained from the L. lactis strain. The protein incorporated into the vaccine formulation yielded a total of 391 amino acids in its nucleotide sequences (1173 base sequence). In addition, the altered nucleotide sequence exhibits a GC content of 66.07% and a CAI value of 0.89, as seen in Fig. 8A. The modified sequence was inserted into the pNZ8121 vector by employing the XhoI and BamHI restriction regions as the sites for cleavage initiation and termination, respectively (Fig. 8B). The pNZ8121 cloning plasmid has been employed to include the improved vaccine design via the utilization of the SnapGene software. The RNA fold database was employed to forecast the secondary configuration of mRNA. The thermodynamic stability of the mRNA structure may be determined by the lowest free energy, which is −556.50 kcal/mol. Furthermore, the positional entropy is shown for each location, as seen in Fig. 8C. The free energy of the thermodynamic ensemble was − 572.39 kcal/mol. The minimum free energy structure (MFE) structure in the ensemble was 0.00%. The ensemble diversity was 261.03. Furthermore, it was noted that the first 12 nucleotides of the mRNA secondary structure lacked pseudoknots or long stable hairpins. This characteristic enhances the mRNA’s ability to initiate translation effectively inside its framework. Two mRNA structures were predicted utilizing the MFE and Centroid methods for plain structure drawing, as seen in Fig. 8D.

Fig. 8.

A The process of optimizing codons in vaccine constructions targeting Lactococcus lactis. B Vaccine cloning using computational methods. The blue region inside the pNZ8121 expression vector represents the multi-epitope vaccination insert. C The mRNA entropy of the vaccine structure indicates the mRNA stability. D Different mRNA structures predicted from the vaccine structure

Discussion

Brucellosis is a resurging zoonotic infection characterized by intricate clinical manifestations and significant fatality rates [22–24]. The vaccination remains the primary method for preventing brucellosis, however, there is currently no effective Brucella vaccine available [22–25]. The majority of disease-causing bacteria start their pathogenic activities on mucosal surfaces [26]. Thus, if the colonization and infiltration of pathogenic microbes were to cease at this juncture, infection would not ensue. Hence, it is essential to create a vaccination that enhances both cellular and mucosal defense mechanisms [22–26]. Current research suggests that lactic acid bacteria (LAB) are very successful for producing mucosal vaccines due to their ability to induce intestinal immunity when employed as a live transport vector for proteins [22–26]. The objective of this study is to propose L. lactis as a useful, non-harmful treatment for mucosal immunotherapy [22–26]. Additionally, we aim to produce immune-stimulating B-cell + CTL + HTL epitopes from omp19 and Bacterial surface antigen (D15) protein using L. lactis as an alternative to the E. coli manufacturing system due to its advantages. Unlike live attenuated varieties of Brucella, which have the potential to revert back to their original virulent structure, selecting specific antigens of bacteria that only include the desired immunogenic epitopes for future subunit-based vaccines, rather than using the entire bacterium, may be the optimal approach for developing innovative vaccination strategies [27, 28]. Research on reverse vaccination is a cutting-edge field that explores possible antigenic alternatives [27, 28]. This report includes a genetic study aimed at identifying a possible vaccine for B. abortus isolates. The B. abortus genome is utilized solely for the identification of potential vaccine candidates [29]. In 2000, a study conducted to identify prospective vaccine candidates for meningococcus serogroup B (Neisseria meningitides) was the first use of practical reverse vaccinology. In 2006, a vaccine for meningococcus serogroup B was identified. Promising results were shown in human clinical trials in 2011 [22, 23, 29].

Recently, many new research studies have been published that support the use of the reverse vaccinology top-down technique to identify dominant epitopes from the entire proteome of viruses and bacteria [30]. Our work included the unique design of MEV, followed by an evaluation of its immunogenicity. Proteins that exhibit a high level of antigenic properties might be considered as potential targets for the development of MEVs [31]. Consequently, using the most recent study findings, we conducted an examination of the principal proteins involved with Brucella. Subsequently, we selected Omp19 and Bacterial surface antigen (D15) due to their notable antigenicity, stability, and hydrophilic properties for the construction of the MEV. Currently, there are no pertinent studies on the development of MEVs based on these potential proteins [32]. The signal peptide had an effect on the initiation of protein translation, and the distinct main components of the signal peptides also impacted the folding and transportation of the proteins. Consequently, we eliminated the signal peptide sequences.

The goal of anticipating HTL epitopes and CTL epitopes is to identify the concise peptide sequence within an antigen that activates CD4 + or CD8 + T lymphocytes in a living organism [33]. Moreover, the identification of B cell epitopes in pathogenic organisms expands the range of antigenic proteins and enhances the immunogenicity of proteins [34]. Our finding of an extensive array of CTL, HTL, and both linear and conformational B-cell epitopes corroborates other studies highlighting the significance of a multi-epitope strategy for successful Brucella vaccines. The significant affinity noted in our HLA-T-cell epitope complexes supports findings from earlier studies that associate strong MHC binding with vigorous cellular immune responses to intracellular infections. Nonetheless, our research presents original additions. Although prior endeavors have recognized some significant antigenic proteins, our comprehensive genomic prediction, especially with Vaxign, may have revealed novel epitopes on proteins not conventionally investigated in wet lab research. This may encompass epitopes on “non-virulent” proteins that provoke a protective immune response, broadening the possible array of vaccine candidates beyond traditionally acknowledged virulence factors. Variations in individual epitope sequences or their anticipated immunogenicity may be ascribed to discrepancies in computational algorithms or the particular Brucella strains examined, necessitating additional experimental confirmation.

During the first phases of vaccine development, it is important to take into account certain characteristics of the MEV [35, 36]. The AAY linker sequence (Alanine-Alanine-Tyrosine) is strategically incorporated at the end of the vaccine construct for several critical functional reasons in multi-epitope vaccine design. Primarily, AAY is known to act as a proteasomal cleavage site in mammalian cells. This property is crucial for efficient antigen processing and presentation by the host’s immune system. By positioning an AAY linker, it facilitates the proteolytic breakdown of the larger chimeric vaccine protein into individual epitopes within antigen-presenting cells (APCs). This ensures that each T-cell epitope is properly liberated and presented on MHC molecules, thereby maximizing the induction of specific T-cell responses (both CD4 + helper T cells and CD8 + cytotoxic T lymphocytes). Furthermore, the AAY linker also contributes to the stability and rigidity of the overall vaccine construct, which is important for maintaining its structural integrity and functional efficacy before and during immune processing, preventing unwanted conformational changes that could compromise epitope presentation or lead to the formation of neo-epitopes [35, 36]. In reverse vaccinology, the protein’s molecular weight employed for vaccine design should ideally be below 110 KD [37]. The findings of this experiment demonstrated that the MEV, consisting of 387 amino acids and having a molecular weight of 38 KD, was a soluble protein. Furthermore, it had an antigenicity value of 1.0422, beyond the established threshold, and showed no signs of allergenicity. The findings demonstrated that the MEV exhibited favorable stability, hydrophilicity, antigenicity, solubility, and non-allergenicity. Subsequently, the secondary structure prediction indicated that the β-turn constituted 10.23% while the random coil contributed 56.27%. The significant presence of β-turn and random coil structures in the MEV indicates that the protein is prone to forming antigenic epitopes. Next, we use computational methods to forecast the three-dimensional arrangements of proteins and the minimum energy value (MEV). Ultimately, the ProSA-web and the structure evaluation service of SWISS-MODEL were utilized to confirm the high quality of the tertiary structure of the MEV. The findings shown that the tertiary conformation of MEV exhibits a remarkable level of precision and a strong correlation coefficient with the authentic structure. Overall, the tertiary structure of the MEV showed a significant presence of β-turn and random coil, aligning with the secondary structure predictions. This suggests that the MEV has strong antigenic properties.

TLR5 functions as a protein receptor with a cytoplasmic signaling component. It preferentially detects chemicals emitted by injured or ischemic organs, which are known as pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) [38]. TLR5 specifically detects lipopolysaccharides (LPS) found in gram-negative bacteria and functions as a receptor for precise binding with the MEV [39, 40]. By means of docking, it becomes evident that there exists a robust contact force between molecules via the atomic interaction surface of TLR5-MEV [41]. The findings indicate that the vaccine has the ability to establish durable interactions with immunological receptors, enabling its distribution throughout the whole host organism.

Research was done to investigate and analyze the immune system’s reaction profile and immunogenicity by an immunological simulation. The MEV plays a significant role in cell-mediated immunity by augmenting the bactericidal reactions of macrophages, facilitating the development of Th1 cells, triggering B cell antibody class changing, intensifying the cytotoxic reactions of NK cells, and stimulating the presentation of antigens to T cells. Elevated quantities of immune system cells and cytokines suggest that the vaccination has significant promise in stimulating a robust immunological response. In order to conduct a more in-depth examination of the structural integrity of the docked TLR5-MEV, a molecular dynamics (MD) simulation of the complex was carried out. The TLR5-MEV conformational changes and stability were evaluated by computing the RMSD, RMSF, and ROG, using the initial structure as a reference. All the findings indicated that the TLR5-MEV exhibits robust stability.

Reverse vaccinology is a logical vaccine design strategy that is successful [42]. However, it does not guarantee efficiency against all strains in terms of the subsequent immune response [41, 42]. Consequently, we selected B. abortus for future investigation. The current investigation included the formulation of a multi-epitope vaccine (MEV) that includes cytotoxic T lymphocyte (CTL) epitopes, helper T lymphocyte (HTL) epitopes, and B cell epitopes. This MEV was paired with suitable adjuvants to augment the immune system’s reaction in animal models. The findings demonstrated that this particular MEV has exceptional qualities and is a very good choice for combating B. abortus. In this study a single vaccine construct was designed in total, with “5 predicted vaccine models” referring to the multiple structural predictions generated for this one sequence using ProSA and SWISS-MODEL. This single construct was finalized for downstream applications based on a comprehensive evaluation of its predicted 3D structure, utilizing parameters such as a favorable ProSA Z-score (indicating overall model quality), high SWISS-MODEL GMQE and QMEANDisCo scores (reflecting expected accuracy and local/global quality), a significant percentage of residues in the favored regions of the Ramachandran plot (for structural stability), a high ERRAT score (assessing non-bonded interactions), and a robust Verify3D score (evaluating sequence-structure compatibility). These metrics collectively ensured the selection of the most structurally sound and reliable model for the designed vaccine. The modeling and evaluation of five variants of the same vaccine construct aimed to examine structural variability and enhance the ideal projected 3D structure. Given that computational protein structure prediction lacks perfect accuracy, generating multiple models enabled us to address inherent prediction uncertainties and ascertain the most stable and reliable fold by evaluating quality metrics such as ProSA Z-score, SWISS-MODEL GMQE/QMEANDisCo, Ramachandran plot percentages, ERRAT, and Verify3D scores. This meticulous methodology enhances confidence in the structural integrity of the chosen model, guaranteeing a solid foundation for essential downstream applications like molecular docking simulations.

Conclusion

Brucella is a kind of bacteria that has a spherical shape and stains pink when tested using a certain staining method. It has the ability to cause illnesses that may be transmitted from animals to humans. The current study on reverse vaccinology has the potential to provide novel insights into the development of vaccines targeting B. abortus. The immunogenicity of the MEV was enhanced by fusing dominant epitopes utilizing linkers and adjuvant. The physiochemical characteristics, antigenicity, allergenicity, solubility, and tertiary structure studies of MEV were determined to be very acceptable. Furthermore, the docked complexes seen throughout the simulation demonstrated a robust and enduring binding relationship between MEV and TLR5.

Authors’ contributions

Conceptualization, A.D., H.GH.; methodology, A.D.; software, R.A. and A.D.; All authors reviewed the manuscript.

Funding

This research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the Islamic Azad University of Shahrekord Branch in Iran (IR.IAU.SHK.REC.1400).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.