Junin virus (JUNV) is a causative agent of Argentine haemorrhagic fever (AHF) in humans (especially in farmers) with no long-term approved antiviral therapeutics. We developed a multi-epitope vaccine against JUNV using an immunoinformatics approach comprising sequence retrieval, epitope prediction, vaccine design and analysis, molecular docking, molecular dynamics (MD) simulations, immune simulation, and in silico cloning. Moreover, we evaluated six Cytotoxic T-Lymphocytes (CTL), four Helper T-Lymphocytes (HTL), and four Linear B-Lymphocytes (LBL) potential epitopes with high antigenicity, no allergenicity, outside membrane, and no toxicity. The vaccine was linked by β-defensin 3 and a PADRE sequence at both termini and multiple structural linkers (M, EAAAK, TAT, SSL, GGGS, GGGGS, KK) to ensure stability and proper epitope spacing. The designed vaccine showed a worldwide population coverage (98.83%), antigenic (0.8295), non-allergenic, outside membrane, and non-toxic. Additionally, structural assessment showed an ERRAT score of 94.758%, with 93.8% residues in the favored region of the Ramachandran plot and a ProSA Z-score of −6.3. Additionally, molecular docking reveals the vaccine’s favorable predicted binding affinity score with human TLR2 (−194.2) than human TLR4 (−125.5 ± 2.6), whereas MD simulation with 100 ns indicates stable behavior both structural dynamic complexes. Finally, immune simulation predicted enhanced humoral and cellular responses, with peak IFN-γ and IL-4 concentrations of approximately 400,000 and 120,000 ng/mL, respectively, while in silico cloning indicated favorable predicted expression feasibility in Escherichia coli expression of the vaccine. Furthermore, the vaccine shows potential immunity against JUNV, whether in vitro or in vivo studies are necessary to confirm its efficacy.
Dengue virus (DENV) comprises four antigenically distinct serotypes, complicating the development of broadly active antiviral agents. This study developed a multi-target computational framework to prioritize phytochemicals with favourable and relatively consistent predicted interactions across DENV-1 to DENV-4. A library of 200 phytochemicals was screened against 10 viral proteins across all four serotypes, yielding 40 target–serotype combinations and 8000 molecular-docking predictions. Candidate prioritisation integrated Lipinski drug-likeness filtering, docking-score enrichment, density functional theory-based electronic characterisation, and a cross-serotype consistency metric, SUniversal. Nine compounds representing 10 protein–ligand complexes were subsequently evaluated by triplicate 100 ns molecular dynamics simulations, MM-GBSA end-point energetic analysis, and multi-platform ADMET and toxicity prediction. Among the retained complexes, the library ligand annotated as silymarin in complex with DENV-2 NS3 helicase produced the most favourable MM-GBSA estimate ( − 38.43 ± 1.16 kcal mol−1), followed by chaparrinone–NS3/DENV-4 ( − 31.48 ± 4.38 kcal mol−1) and 4,7-DMDF–capsid/DENV-3 ( − 27.37 ± 1.63 kcal mol−1). Docking and MM-GBSA ranks showed no clear monotonic association across the retained complexes (ρ = 0.29, P = 0.45), while sanguinarine, despite sharing the most favourable docking score among the simulated candidates, reproducibly departed from its initial binding region. Receptor-model choice also substantially affected predicted dynamics and energetics for the DENV-4 NS3–isolicoflavonol system. These findings indicate that docking strength, cross-serotype consistency, dynamic persistence, and predicted safety provide complementary information during candidate prioritisation. The study therefore provides a reproducible in silico framework and a set of experimentally testable phytochemical candidates for cross-serotype DENV antiviral discovery, while emphasising that the predictions require biochemical and antiviral validation.
