In Silico Identification and Characterization of Antigenic Epitopes of BamA from Edwardsiella

INTRODUCTION

Vaccination has played a major role in controlling various infectious diseases, including influenza, smallpox, diphtheria, polio, tetanus, hepatitis, measles, and HPV. Traditional vaccines based on live attenuated or inactivated pathogens can sometimes pose safety risks and generate undesirable immune responses. Subunit vaccines are gaining increasing attention because they offer several advantages, such as requiring a low dose, having stronger immunogenicity, and causing minimal toxicity. Outer membrane proteins (OMPs) are highly immunogenic among the various bacterial components and represent promising targets for developing novel and effective vaccines. Studies have shown that OMPs can elicit strong humoral and cellular-mediated immune responses.^[1-4] Additionally, certain OMPs offer cross-protection, helping defend against multiple strains and serotypes.^[5]

BamA is an OMP that plays a crucial role in bacteria as the core element of the β-barrel assembly machinery, responsible for inserting and translocating other OMPs. Due to its vital function, BamA remains highly conserved across bacterial species. It is generally a 16-stranded protein and switches its conformation during the assembly and insertion of OMPs.^[6]

The accessibility of genomic and proteomic data enables predicting and identifying pathogenic proteins that may serve as potential vaccine candidates. Reverse vaccinology is a computational approach used to predict and identify antigens and epitopes that may offer protective immunity.^[7] Epitopes that are recognised by the immune system are key targets in vaccine research. Their ability to trigger immune responses makes them valuable for designing effective vaccines. Combined advancements in bioinformatics, recombinant DNA technology, understanding of host genetics, and pathogen genomes are paving the way for rapid vaccine development. Using in silico computational approaches to identify potential targets, vaccines can be developed at a faster rate. Tools in this field rely on statistical and machine learning approaches and are applied to molecular interaction models such as antigen processing, presentation, and binding.

In this paper, we combined various bioinformatics tools to screen, identify, and predict the immunogenicity and binding ability of BamA, a prospective vaccine candidate, an OMP of the bacterium Edwardsiella.

MATERIAL AND METHODS

RESULTS

Amplification of the BamA gene

The full-length gene encoding BamA of E. tarda was successfully amplified using PCR. The amplification yielded a distinct band of ∼1131 bp [Figure 1a].

Close

Figure 1:

(a) Representative image for PCR amplification of the BamA gene of Edwardsiella using respective primers. (b) PCR amplification using vector-specific primers (M13 primer). Lane no. B2, B3, & B9 – Positive colonies of BamA/TamA. PCR: Polymerase chain reaction.

Export to PPT

In silico analysis of the gene

Prediction of Transmembrane beta-barrel

A PRED-TMBB analysis identified the protein as an OMP containing characteristic β-barrel structures, thereby confirming its typical structural features [Figure 2a].

Close

Figure 2:

(a): Predicted 2D structure of BamA generated using PRED-TMBB software, highlighting the presence of β-barrel motifs characteristic of outer membrane proteins. Homology model of the Edwardsiella tarda OMP BamA generated using SWISS-MODEL, (b) Side view, (c) Top view.

Export to PPT

Signal peptide and subcellular localisation

The signal peptide was predicted to span amino acids 1 to 21 with a probability score of 0.96. The subcellular localisation prediction by PSORTb V.3.0 confirmed that BamA of Edwardsiella has a non-cytoplasmic signal peptide, and the protein is localised in the outer membrane, with a score of 9.52.

Homology modelling

Homology modelling using SWISS-MODEL generated about 50 comparable templates, with the top-ranked model (Q6E209.1.A) selected based on the species similarity (E. ictaluri), percentage identity (95.23%), GMQE score (0.90), and Q-MEAN Z-score, predicting a reliable three-dimensional structure of the BamA protein. The predicted model displayed a classical 16-stranded β-barrel structure [Figures 2b and c]. Structural assessment indicated 95.5% of the amino acids were in the favoured region of the Ramachandran plot. ProSA analysis yielded a Z-score of -4.72.

BamA as a potential vaccine candidate

A protein must have certain characteristics to be an efficient vaccine candidate. BamA was predicted as an OMP with an adhesion probability of 0.573. The protein sequence was analysed using Kolasker and Tongaonker antigenicity scale (1990) to identify potential antigenic determinants. Thirteen antigenic sites were determined: MTLALVCALSPLAGQAALLPTR, WGVLPGP, GLGIGAAVAGLYRP, SITLSGYL, GLGVRSY, RIFVTG, AQSLRLQPEVLYRVASQTYLGVGWS, LAQIHYGRSVFSSGPSLSLQY, LASLRYTHY, STYLAL, ETRVLAW, RHGVVAWA, and LPTVGVGYR. Fourteen probable B-cell antigenic epitopes of BamA were identified using the BepiPred 3.0 server, which predicts linear B-cell epitopes based on proprietary algorithms [Table 1]. Twelve peptide sequences were predicted as potential cytotoxic T lymphocyte (CTL) epitopes, based on multiple parameters, including MHC binding affinity, proteasomal C-terminal cleavage, and TAP transport efficiency. These peptides had prediction scores >0.75 [Table 2].

Table 1: List of predicted B-cell epitopes for BamA of E. tarda as determined by Bepipred Linear Epitope Prediction 2.0

Sl. No.	Start	End	Peptide	Length
1	21	48	AALLPTRAQIDEWLGHLGGDDRFDPDKG	28
2	60	67	NPELGLGI	8
3	75	89	YRPDPSDTTSQNSSI	15
4	109	115	SFFDNDR	7
5	126	150	DTPTYYWGQGFRAGADDDGRQKYTA	25
6	154	165	RLQPEVLYRVAS	12
7	177	197	AMHAADVKQTDRLAQIHYGRS	21
8	209	219	YDSRDFVPNPR	11
9	229	241	THYTPETGSDTRF	13
10	266	276	GEFTQGAVPWN	11
11	287	297	MRGYYQGRYRD	11
12	308	314	RQKLSWR	7
13	328	336	PNVRELGSS	9
14	348	352	FAFKP	5

Table 2: List of predicted CTL epitopes for BamA of E. tarda as determined by NetCTL 1.2

Residue number	Peptide sequence	Predicted MHC binding affinity	Rescale binding affinity	C-terminal cleavage affinity	Transport affinity	Prediction score	MHC ligands
29	QIDEWLGHL	0.1300	0.5519	0.9759	1.0740	0.7520	E
67	IGAAVAGLY	0.1224	0.5195	0.9227	2.6680	0.7914	E
86	NSSITLSGY	0.4554	1.9337	0.9054	2.9490	2.2169	E
123	GISDTPTYY	0.2760	1.1717	0.9695	2.8400	1.4591	E
124	ISDTPTYYW	0.1717	0.7292	0.9710	0.7420	0.9119	E
185	QTDRLAQIH	0.3257	1.3827	0.0494	-0.6190	1.3591	E
223	LASLRYTHY	0.3768	1.5998	0.9733	3.0130	1.8964	E
236	GSDTRFDNL	0.1666	0.7074	0.5364	0.6500	0.8204	E
243	NLTTRFSTY	0.2646	1.1235	0.9764	2.8060	1.4102	E
287	MRGYYQGRY	0.1106	0.4697	0.8758	3.1470	0.7584	E
299	NLISSQLEY	0.2442	1.0368	0.8816	3.1820	1.3281	E
338	WLPTVGVGY	0.1721	0.7306	0.9753	3.0340	1.0286	E

E: Epitope; MHC: Major histocompatibility complex.

Molecular docking and validation

Among the twelve predicted CTL epitopes, six, QIDEWLGHL (BTP-01), GISDTPTYY (BTP-02), ISDTPTYYW (BTP-03), QTDRLAQIH (BTP-04), GSDTRFDNL (BTP-05), and MRGYYQGRY (BTP-06) were identified as candidates capable of eliciting both humoral and cell-mediated immunity, as they were also predicted to be B-cell epitopes. These peptides showed strong binding affinity to the MHC molecule (PDB:5HZH), supported by multiple bonding interactions. Figure 3 shows the detailed 3D and 2D interactions between the six peptides and the fish MHC molecule. All six peptides that underwent molecular docking with the MHC molecule exhibited good docking with relevant free binding energy [Table 3].

Close

Figure 3:

Two-dimensional (2D) and three-dimensional (3D) docking results for outer membrane protein BamA peptides of Edwardsiella tarda with the fish major histocompatibility complex (MHC) molecule. (a) QIDEWLGHL–MHC, (b) GISDTPTYY–MHC, (c) ISDTPTYYW–MHC, (d) QTDRLAQIH–MHC, (e) GSDTRFDNL–MHC, and (f) MRGYYQGRY–MHC.

Export to PPT

Table 3: Interactions between OMP- BamA peptides of E. tarda and the fish major histocompatibility complex (MHC) molecule

Sl. No.	Complex	Peptide code	Free binding energy/affinity (kcal/mol)	Hydrogen bonds
1	QIDEWLGHL-MHC	(BTP-01)	-8.6956	10
2	GISDTPTYY-MHC	(BTP-02)	-8.0592	11
3	ISDTPTYYW-MHC	(BTP-03)	-8.1317	10
4	QTDRLAQIH-MHC	(BTP-04)	-7.1180	8
5	GSDTRFDNL-MHC	(BTP-05)	-7.9803	16
6	MRGYYQGRY-MHC	(BTP-06)	-7.6080	10

Source of MHC molecule: Grass Carp (Ctenopharyngodon idella) (PDB: 5H5Z).

DISCUSSION

Developing efficient and safe vaccines remains crucial for preventing and controlling infectious diseases, playing an important role in global health management. Advances in recombinant DNA technology and immunoinformatics have enhanced our ability to identify and validate promising antigens. Among potential targets in bacterial diseases, OMPs are generally strong vaccine candidates because their surface-exposed epitopes make them highly immunogenic. The host immune system can easily recognise regions of these proteins as foreign, facilitating the generation of an immune response.^[24] Multiple studies have targeted bacterial OMPs as vaccine candidates in various pathogens.

Conducting in silico analysis prior to experimental production of subunit vaccines is essential, as it allows the rational identification, screening, and evaluation of potential antigen targets. Such computational approaches help narrow down the most promising candidates, thereby saving time, reducing costs, and increasing the efficiency of subsequent laboratory validation. This study aimed to evaluate the potential of BamA as a vaccine candidate through a comprehensive in silico approach integrating structural analysis and immunoinformatics.

The target protein was confirmed to have a β-barrel transmembrane structure. OMPs are of particular interest in vaccine design, as effective immune recognition generally requires the antigen to either be secreted, surface-exposed, or embedded in the outer membrane of the organism.^[25,26]

Analysis of the signal peptide and sub-cellular localization predicts the protein to be localised in the outer membrane, supporting its accessibility to the host immune system and interaction with immunostimulatory molecules.

A three-dimensional homology model was also generated for BamA, providing insights into the structure and potential biological function of the protein. The selection of the template model is justified by the species and sequence similarity to the target protein. The quality of the model, as indicated by the GMQE and QMEAN scores, was within the acceptable range reported for reliable models. The model predicted a β-barrel structure, typical of the OMPs in Gram-negative bacteria. Validation analysis further strengthened the reliability of the predicted model. The Ramachandran plot assessment demonstrated that the majority of the residues were positioned within favoured regions. The Z-score obtained by ProSA also indicated that the predicted model was well within the range of experimentally determined structures.

Surface-exposed proteins, particularly those located on the outer membrane, are widely recognised as efficient vaccine candidates due to their accessibility to the host immune system. Supporting this, the adhesin probability score for BamA, as determined by Vaxign2, was 0.573, indicating its likelihood of functioning as an adhesin. Furthermore, 13 sites on the protein were predicted to be antigenic.

Since recognizing B-cell epitopes stimulates antibody production, their identification is critical for designing and optimizing pathogen-specific vaccines. This study predicted the target protein to have 14 B-cell epitopes of varying lengths, suggesting their potential to generate an antibody-mediated immune response. Cytotoxic T-lymphocyte (CTL) epitopes were predicted using the NetCTL 1.2 server, which integrates prediction of MHC binding affinity, proteasomal C-terminal cleavage, and TAP transport affinity.^[17] Twelve T-cell epitopes were predicted in BamA. Notably, six of these overlapped with B-cell epitopes, indicating that these regions could induce both the humoral and cell-mediated immune response. These six common epitopes were further used for docking studies. In these studies, the binding of the common epitopes was evaluated with the MHC molecule of bony fishes. Among the six epitopes, BTP-01 and BTP-03 showed the best docking with the lowest binding energy.

The C-ImmSim simulation indicated that the target protein can induce a strong immune response. The early rise of IgM and IgG from day 5 suggests effective B-cell activation. The simultaneous increase in cytokines such as IFN-γ, IL-2, IL-10, IL-12, and TGF-β reflects activation of both humoral and cell-mediated immunity. Notably, the strong IFN-γ response indicates a Th-1-inclined profile, which is beneficial for intracellular pathogens like E. tarda. Overall, the simulation predicts that the vaccine construct can generate a good protective immune response, validating it for further experiments.

A comparable in silico strategy was previously employed to evaluate BamA as a vaccine candidate for E. coli (colibacillosis) in mice, followed by in vivo validation, which resulted in an 80% survival rate relative to the control group.^[27] BamA exhibited high sequence conservation, lacked allergenic properties, demonstrated strong immunogenic responses in mice, and promoted enhanced opsonophagocytic activity, highlighting its potential as a subunit vaccine antigen. Recombinant BamA of A. baumannii elicited strong opsonizing antibody responses and provided significant protection in a murine pneumonia model. The vaccine reduced bacterial load and inflammation, highlighting BamA as a promising candidate against multidrug-resistant A. baumannii infections.^[28]

These studies underscore the translational value of BamA as a broadly immunogenic protein and potential vaccine candidate against E. tarda.

Considering our findings, the structural predictions, antigenicity assessments, and epitope identification and binding strongly support BamA as a promising vaccine candidate against Edwardsiella. Its predicted surface accessibility, presence of multiple B and T cell epitopes, and demonstrated potential to engage both humoral and cell-mediated immune responses provide a good rationale for further exploration. Comparative studies in other pathogens have also validated BamA as an effective antigen. Experimental validation in vivo would be the next crucial step. Additionally, proteomics and transcriptomic approaches can be considered to identify surface proteins involved in host-pathogen interaction. Computational tools reduce the time and cost associated with traditional trial and error vaccine development and enhance the likelihood of success by ensuring that only the most promising vaccine candidates are taken forward to laboratory validation.

CONCLUSION

Although reverse vaccinology has made a significant contribution to human and veterinary vaccines, its application in fish vaccinology remains relatively underexplored. Using in silico analyses provides an efficient and rational strategy for subunit vaccine discovery, enabling antigen identification, screening, and confirmation before labour and cost-intensive laboratory trials. The in silico analyses conducted in this study strongly highlight BamA as a potential vaccine candidate against edwardsiellosis, emphasising the importance of computational approaches in rationalising and strengthening vaccine development. These findings demonstrate the global significance of immunoinformatics in addressing infectious diseases across species and ecosystems, paving the way for sustainable and data-driven solutions to global health challenges.