Construction of an aerolysin-based multi-epitope vaccine against Aeromonas hydrophila: an in silico machine learning and artificial intelligence-supported approach

Abstract

Aeromonas hydrophila, a gram-negative coccobacillus bacterium, can cause various infections in humans, including septic arthritis, diarrhea (traveler's diarrhea), gastroenteritis, skin and wound infections, meningitis, fulminating septicemia, enterocolitis, peritonitis, and endocarditis. It frequently occurs in aquatic environments and readily contacts humans, leading to high infection rates. This bacterium has exhibited resistance to numerous commercial antibiotics, and no vaccine has yet been developed. Aiming to combat the alarmingly high infection rate, this study utilizes in silico techniques to design a multi-epitope vaccine (MEV) candidate against this bacterium based on its aerolysin toxin, which is the most toxic and highly conserved virulence factor among the Aeromonas species. After retrieval, aerolysin was processed for B-cell and T-cell epitope mapping. Once filtered for toxicity, antigenicity, allergenicity, and solubility, the chosen epitopes were combined with an adjuvant and specific linkers to create a vaccine construct. These linkers and the adjuvant enhance the MEV's ability to elicit robust immune responses. Analyses of the predicted and improved vaccine structure revealed that 75.5%, 19.8%, and 1.3% of its amino acids occupy the most favored, additional allowed, and generously allowed regions, respectively, while its ERRAT score reached nearly 70%. Docking simulations showed the MEV exhibiting the highest interaction and binding energies (-1,023.4 kcal/mol, -923.2 kcal/mol, and -988.3 kcal/mol) with TLR-4, MHC-I, and MHC-II receptors. Further molecular dynamics simulations demonstrated the docked complexes' remarkable stability and maximum interactions, i.e., uniform RMSD, fluctuated RMSF, and lowest binding net energy. In silico models also predict the vaccine will stimulate a variety of immunological pathways following administration. These analyses suggest the vaccine's efficacy in inducing robust immune responses against A. hydrophila. With high solubility and no predicted allergic responses or toxicity, it appears safe for administration in both healthy and A. hydrophila-infected individuals.

Keywords: A. hydrophila; MD simulations; epitopes; immunoinformatics; systems biology; vaccine.

Figure 1

Schematic methodology flow of the study. (A) Data retrieval. (B) B-cell epitope prediction. (C) T-cell epitope prediction. (D) Epitope essentiality filtration. (E) Final MEV construct. (F) Immune simulation. (G) Structure prediction. (H) Stability analysis. (I–K) Molecular docking with MHC-II, TLR4, and MHC-II and simulations. (L) Jcat analysis. (M) In silico cloning.

Figure 2

In the aerolysin structure, the red regions showed B-cell epitopes (represented by the yellow region of the graph), while the yellow regions in the structure represent amino acids that do not contribute to the mentioned epitopes and are represented by the green regions of the graph, respectively.

Figure 3

All these figures include crucial information about the MEV. (A) It is the final sequence of the MEV, where the adjuvant is highlighted in red, EAAAK in green, and GPGPG in yellow, and the nonhighlighted regions are the epitopes. (B) The three-dimensional structure of MEV. (C) The graphical visualization of immune responses against MEV. (D) The Ramachandran plot (for stability analysis) for the MEV.

Figure 4

Binding analysis. The quantity and kind of interactions between the MEV construct’s chains and MHC-I chains are displayed in (A). MHC-I and the MEV construct are shown to be bound and interacting, respectively, in (B), while (C) displays the quantity and kind of interactions between TLR4 and the MEV construct.

Figure 5

This graph shows that only a few codons (deviated red projections) cannot be adapted according to the expression system (increased red lines below the threshold 1.00), while the uniform red line along point 1.00 shows that the maximum of the codon is adapted (A). (B) The vaccine construct cloned into the vector pet28a(+) with the restriction sites Eco53kI and ScaI employed is shown in the red portion. Above is also the translated sequence for the red region.

Figure 6

Vaccine-receptor molecular dynamics simulation analyses. (A) RMSD and (B) RMSF.