The protein sequence of SpA was obtained from the NCBI database https://www.ncbi.nlm.nih.gov/ bearing ID: WP_033567379. To ensure comprehensive coverage and variability analysis, the BLASTP tool was used to extract approximately 250 homologous sequences. Multiple sequence alignment (MSA) was conducted utilizing the COBALT tool available at NCBI https://www.ncbi.nlm.nih.gov/tools/cobalt/re_cobalt.cgi (Papadopoulos and Agarwala, 2007) to identify conserved and variable regions within the protein, which is crucial for epitope prediction. The ProtParam tool https://web.expasy.org/protparam/ was utilized to examine the physicochemical properties of SpA (Wilkins et al., 1999). The physicochemical properties help in understanding the stability and solubility of the protein, which are important for vaccine design.

Epitopes were selected based on several criteria:

Predicted epitopes (CTL & HTL) were subjected to molecular docking studies to assess their binding affinities with HLA molecules. ClusPro tool https://cluspro.org/help.phpb was used for these studies (Vajda et al., 2017). The binding interactions were visualized and the interaction patterns were analyzed using Pymol to ensure strong and specific binding.

The selected T&B-cell epitopes were concatenated into a single multi-epitope chain. Linker sequences were used to join the epitopes to enhance their presentation and processing by the immune system. Furthermore, the vaccine construct was supplemented with an adjuvant and PADRE sequence to bolster its immunogenicity and facilitate epitope presentation. The physicochemical properties of the finalized vaccine construct was also noted. The I-Tasser server https://zhanggroup.org/I-TASSER/ (Zhou et al., 2022) was employed to initially model the multi-epitope chain, followed by refinement utilizing the Galaxy refine server https://galaxy.seoklab.org/cgi-bin/submit.cgi?type=REFINE (Heo et al., 2013). The refined vaccine’s 3D model underwent validation using the ProSA-web tool https://bio.tools/prosa-web (Wiederstein and Sippl, 2007). In addition, the Ramachandran plot was also investigated using MolProbity https://phenix-online.org/documentation/reference/molprobity_tool.html and the structural robustness of the modelled vaccine was examined using CABS-flex 2.0 https://bio.tools/cabs-flex_2.0 (Kuriata et al., 2018). Molecular docking with the Cluspro server https://cluspro.org/help.php was performed to determine the ability of the multi-epitope construct to trigger innate immune activation via TLR-4.

In order to improve the structural stability of the vaccine structure numerous disulfide bonds were incorporated into the refined vaccine construct utilizing the design v2.0 web server (Xie et al., 2018).

The vaccine sequence was also investigated foe the ability to induce humoral and adaptive immune responses. We employed in silico immune simulation assay for the same utilizing C-immsim server (Stolfi et al., 2022).

Multiple sequence alignments revealed some mutations in the Spa protein. These mutations were noted for emitting the epitopes lying within these regions ( Figure 1 ). The protein’s properties unveiled its molecular weight at 45528.47, theoretical pI at 5.65, with 56 +vely charged residues and 62-vely charged residues. The projected ½ life spanned 30 hrs (mammalian reticulocytes, in-vitro), greater than 20 hours (yeast, in-vivo), and greater than 10 hours (E. coli, in-vivo). The Aliphatic and instability index stood at 69.47, (II) and 50.49 respectively. The GRAVY was measured at -0.991. The identified mutations could potentially impact the immunogenic properties of SpA by altering its epitope structure or antigen presentation. These changes may influence vaccine efficacy, necessitating careful consideration during epitope selection and structural modeling. The molecular weight and pI of the SpA are crucial factors that can influence its immunogenic properties. The molecular weight of a protein impacts its processing and presentation by antigen-presenting cells, as smaller peptides may be more readily processed, while larger proteins may require additional processing steps. The theoretical pI, on the other hand, reflects the protein’s charge at physiological pH, affecting its solubility, stability, and binding interactions with immune receptors and antibodies. Proteins with pI values close to the physiological pH are often more soluble and can interact more effectively with immune components, potentially enhancing their immunogenicity. These properties, therefore, underscore the significance of SpA’s structural and physicochemical characteristics in the design of effective epitope-based vaccines.

The Tepitool server from IEDB was used to forecast HLA-Class I & II epitopes. Tepitool predicts how peptides bind to HLA molecules, with HLA class II molecules accommodating longer peptides due to their expanded binding groove, in contrast to the confined groove in HLA class I molecules. Despite this difference, the core binding region of class II molecules is typically around 9 amino acids in length. Flanking amino acids enhances the stability of the binding without directly interacting with the HLA binding groove. To maximize the coverage of potential binding cores while minimizing redundancy, we selected 15mer peptides overlapping by 10 amino acids. This approach ensured that all possible 9mer binding cores were represented, with a minimum of one flanking residue on each side for stability. This approach eliminates the selection of overlapping peptides that share identical binding cores (Paul et al., 2016). We followed the IEDB recommended method for predicting and selecting the most ideal peptides, which is the default prediction method in Tepitool. While 15mer peptides were selected for optimal CD4+ T cell recognition, their larger size may lead to overlap with adjacent sequences, potentially reducing specificity and increasing cross-reactivity. Additionally, larger peptides may face processing limitations that could impact overall immunogenicity. However, if we do carry out blast analysis against the human genome, this will reduce the probability of cross reactivity. For HLA Class I epitopes, 9 mers were selected. Tepitool selects the best prediction approach available for a given MHC molecule based on predictor availability and historical performance. Peptides with a percentile rank of less than 0.5 were only selected for further screening. A total of 71 epitopes targeting HLA Class I alleles and 33 promiscuous epitopes targeting HLA Class II alleles were predicted. In addition, BepiPred 2.0 predicted nearly 12 B cell epitopes ( Supplementary Figure 2 ). The IEDB database-based tools have earlier been employed by numerous research groups to predict promising epitopes targeting viral (Chauhan et al., 2018; Chauhan and Singh, 2020) bacterial (Ghafouri et al., 2024), parasitic (Chauhan et al., 2023) and other metabolic disorders (Banesh et al., 2024). This underscores the reliability of the predicted epitopes for vaccine applications.

The selected epitopes underwent a comprehensive filtering process before molecular docking analysis to ensure their suitability for vaccine development. The filters applied and their relevance are discussed below:

After applying the above filters, we identified 10 HTL and 7 CTL epitopes. These epitopes demonstrated high binding affinity ( Figure 3 ), broad population coverage, and immunogenicity, and were free from glycosylation sites. For the identification of the B cell epitope, we employed the BepiPred 2.0 tool with a threshold of 0.5. We selected epitopes that were above this threshold, greater than 10 amino acids in length, conserved, immunogenic, and not within glycosylation sites. Based on these criteria, 3 B cell epitopes were selected ( Table 1 ). The above filters led to the identification of the most promising epitopes within SpA protein. Similar kinds of filters have earlier been employed by several research groups in their immuno-informatics study to screen out epitopes for high success probabilities (Chauhan et al., 2021). The physicochemical properties of the finally selected epitopes are represented in Supplementary Table 1 .

The vaccine construct was designed by integrating the optimal T & B cell epitopes to induce a strong immune response. The epitopes selected from previous screening stages (10 HTL, 7 CTL, and 3 B cell epitopes) were used as building blocks for these constructs. The vaccine had molecular weight of 53667.56 Daltons, theoretical PI: 5.75, extinction coefficient 15930, estimated half life 30 hours, instability index 32.26, aliphatic index 63.64 and GRAVY -0.796 as predicted by protparam tool. The incorporation of adjuvants and synthetic linkers was aimed at enhancing immunogenicity and ensuring the proper folding of the vaccine constructs. The addition of the 50S ribosomal protein L7/L12 fortified the vaccine construct to enhance the immune response, capitalizing on its robust immunostimulatory capabilities ( Figure 4A ). The initial vaccine constructs were modelled using the I-Tasser server ( Figure 4B ). This tool provided a preliminary 3D structure of the multi-epitope vaccines by predicting the folding and spatial arrangement of the included epitopes. The top 10 templates used by the I-Tasser server for modelling the vaccine construct were 1rquA, 7wkkB, 1rqv, 2nbiA, 1rqv, 8wxbY, 1rquA, 8p0vL, 5h7c and 1dd3A. The model generated had a C-score of -0.97, with estimated TM and RMSD scores as 0.59 ± 0.14 and 9.7 ± 4.6Å respectively. Refinement of the preliminary structure was conducted using the online available Galaxy Refine server. This step was essential to enhance the integrity of the vaccine model by refining both local and global structure, thereby enhancing the overall stability and accuracy of the model. The refinement process included side-chain repacking and overall energy minimization, leading to more reliable and realistic vaccine structures. The refined vaccine models were validated using Ramachandran plot analysis using MolProbilty. The overall amino acids lying in the favored region were 98.5% ( Supplementary Figure 3 ). This method assesses the stereochemical characteristics of the protein structures by assessing the φ (phi) and ψ (psi) dihedral angles of the aa residues. It was observed that a predominant number of residues in the refined vaccine constructs were situated within the favored regions of the Ramachandran plot (about 98.5%), indicating a high-quality and reliable structure. The results obtained by the ProSa-web tool also demonstrated the considerable structural integrity of the vaccine construct ( Figures 4C, D ). Finally, employing the CABSflex 2.0 server, the vaccine’s flexible and rigid domains were delineated by molecular motion analysis, within the vaccine construct. Evaluation of the final 3D structures (ten) demonstrated minor changes at the beginning locations, where the adjuvant was connected. Conversely, notable fluctuations were detected close to the terminus. This was the region where a succession of epitopes was affixed using linkers, delineated in dark brown and red sections ( Figure 5A ). Across all 10 models, the contact map illustrated the patterns of residue-residue interactions ( Figure 5B ). According to the RMSF plot, residue fluctuations varied from 0 Å to 7 Å, with more pronounced variability towards the latter part of the sequence, mainly comprising the epitopes, suggesting a high degree of flexibility in these regions ( Figure 5C ).

In order to explore the interaction between the vaccine model and TLR-4, molecular docking analysis was performed, aiming to assess its binding affinity and potential immunostimulatory effects. The ClusPro server was employed for the same. It provided a detailed understanding of the binding interactions, depicting binding energies and the interacted amino acids. The vaccine construct exhibited a strong binding affinity to TLR-4, as indicated by the docking score, which revealed stability with significant interactions. Visual analysis of the docking complex demonstrated multiple H-bond interactions between the vaccine and TLR-4. Key residues of TLR-4 involved in binding with the vaccine were identified, ensuring the stability and specificity of the interaction ( Figure 6 ). The robust binding conformation of the vaccine construct in the TLR-4 binding pocket hints at effective interaction with the receptor, potentially bolstering immune response through TLR-4-mediated signaling pathways. This strong and stable binding interaction with TLR-4 is crucial for activating innate and adaptive immune responses. The presence of an adjuvant in the vaccine construct may significantly contribute to the enhanced immunogenicity, acting as a potent immunostimulant and enhancing interaction with TLR-4. Overall, the stable and specific interactions observed in the docking studies highlight the suitability of the construct for further experimental validation as an effective vaccine against MRSA. These findings support the potential of the proposed vaccine construct to engage TLR-4 and promote an enhanced immune response through the activation of TLR-4-dependent pathways.

In order to enhance the structural integrity of the vaccine construct, it was subjected to disulfide engineering. Disulfide by Design 2.12 was used for the same. We uploaded the refined model of the vaccine construct to the server and used for residue pair discovery before being used for disulfide engineering. Following that, 18 latent amino acid pairs were shortlisted ( Figure 7 ).

The immune simulation was carried out at random seed, simulation volume 10, and number of steps 100, 3 injections at 1-, 24- and 48-weeks interval. The vaccine sequence was observed to elicit varied immune cells revealing its potential immune system activation capability ( Figure 8 ).