Computational Vaccine Design Targets Toxoplasma’s Dense Granule Arsenal

Toxoplasma gondii is one of the most successful parasites on the planet, silently infecting roughly a quarter of the human population while also taking a heavy toll on livestock. A new computational study published in Veterinary Medicine and Science has now harnessed the parasite’s own molecular machinery to propose a fresh line of defence: a multi-epitope vaccine candidate assembled from recently discovered dense granule antigens, the proteins the parasite deploys to hijack host cells from within. The lead construct, named Rpf-Toxo, emerged from a rigorous immunoinformatic screening pipeline as the most stable and structurally promising of four designs, and now awaits the wet-lab validation that will determine whether the computer’s optimism survives contact with biology.

The rationale for targeting dense granule antigens, or GRAs, rests on their central role in the parasite’s survival strategy. Once T. gondii invades a host cell, it seals itself inside a parasitophorous vacuole and secretes more than 70 GRA proteins that decorate the vacuole membrane, the intravacuolar network and even the host cell nucleus. These effectors allow the parasite to manipulate host cell function and evade destruction. The research team focused on six GRAs selected for their recent discovery and documented roles in virulence or immunogenicity: GRA15, a well-characterised immunogen that activates the NF-κB pathway; GRA60, a virulence factor acting alongside ROP18; GRA76, identified in 2024 as a growth factor; GRA83, which provokes innate immune responses; and two novel effectors, GRA-α and GRA-β, shown to suppress parasite replication and cyst formation.

No approved vaccine exists for human toxoplasmosis. The only licensed product, the live-attenuated Toxovax used in sheep, cannot be given to people because of its residual pathogenic potential. Existing drug therapies for the acute tachyzoite stage often carry side effects and do not eliminate the dormant tissue cysts that cause lifelong infection. This therapeutic gap has pushed researchers toward reverse vaccinology, the computational discipline that screens pathogen genomes for the most promising antigenic fragments before any experiment begins. By filtering candidates in silico first, laboratories can save years of work, reduce costs and avoid animal testing of constructs destined to fail.

The epitope-mapping stage was deliberately stringent. The team retrieved the amino acid sequences of the six GRAs from the ToxoDB database and predicted linear B-cell epitopes using three independent web servers, retaining only fragments identified by at least two tools. Each surviving epitope was then scored for antigenicity with VaxiJen, screened for allergenicity with AllergenFP and checked for water solubility. In parallel, the researchers predicted cytotoxic T-lymphocyte epitopes of nine or ten residues bound by MHC class I molecules, and helper T-lymphocyte epitopes of fifteen residues bound by MHC class II, selecting the strongest allele-epitope combinations by percentile rank. CTL candidates underwent additional testing for immunogenicity and interferon-gamma induction potential, while HTL candidates were screened for toxicity as well.

From this filtering emerged eighteen final epitopes: six B-cell, six CTL and six HTL segments drawn from the six GRAs. The team then assembled four vaccine constructs, each combining the same epitope payload with a different immunostimulatory adjuvant. RS09-Toxo carried the synthetic TLR-4 agonist peptide RS-09; RSFN-Toxo was double-adjuvanted with RS-09 and human interferon-gamma; 50S-Toxo used a Mycobacterium tuberculosis ribosomal protein; and Rpf-Toxo employed the M. tuberculosis resuscitation-promoting factor E, RpfE. Flexible glycine-serine linkers, proteasomal cleavage sites and the rigid EAAAK spacer held the domains apart, and a C-terminal histidine tag enabled purification. The adjuvant choices were not arbitrary: all favour the Th1-type immunity that is critical for clearing intracellular pathogens such as T. gondii.

Biochemical profiling separated the field quickly. All four constructs scored highly for antigenicity, with values ranging from 0.9148 for Rpf-Toxo to 2.1188 for RS09-Toxo, and none was predicted to be allergenic. However, the ProtParam analysis revealed a critical weakness in the two most antigenic designs: RS09-Toxo and RSFN-Toxo carried instability indices of 60.23 and 50.51, well above the threshold of 40 that marks a protein as unstable in laboratory conditions. Only 50S-Toxo, with an instability index of 24.83, and Rpf-Toxo, at 35.41, passed the stability test, and both showed respectable aliphatic indices suggesting moderate thermal tolerance. All four constructs were hydrophilic, as indicated by negative GRAVY scores, and all exceeded the solubility threshold of the Protein-Sol predictor.

With Rpf-Toxo selected as the lead candidate, the team turned to structural biology. Homology modelling with I-TASSER produced a three-dimensional model that was then refined with the GalaxyRefine server, yielding improved geometry: 86.1 percent of residues fell into favoured Ramachandran regions, the MolProbity score dropped to 2.145 and the clash score fell to 9.6. The researchers candidly note that this figure remains below the 90 percent expected of high-resolution crystal structures, a limitation they attribute to the flexible spacers inherent in multi-epitope constructs. Molecular docking against the human TLR-4 receptor, the innate immune sensor central to Th1-polarised responses against the parasite, produced a stable predicted complex with a binding energy of −20.6 kcal/mol, although the authors caution that the corresponding dissociation constant of 7.3 × 10⁻¹⁶ M likely reflects overestimation by the PRODIGY algorithm rather than a physically realistic affinity.

Immune simulation with the C-ImmSim server, modelling a three-dose schedule at days 1, 84 and 186, painted an encouraging picture. Combined IgG and IgM titres peaked above 170,000 at day 65, memory B-cells reached 500 cells per cubic millimetre and settled at a stable set-point of around 200, and the helper T-cell population included roughly 1,800 memory cells that persisted through day 350. Interferon-gamma levels climbed to approximately 380,000 ng/mL, a signature consistent with the Th1-skewed response needed to eliminate intracellular tachyzoites, and natural killer cell activity peaked near 380 cells per cubic millimetre. Six conformational B-cell epitopes were also predicted on the folded surface of the construct, with ElliPro scores between 0.515 and 0.859, suggesting multiple antibody-accessible sites.

Safety and manufacturability checks completed the computational dossier. A BLASTp search against the human proteome found no significant homology, reducing the risk of autoimmune cross-reactivity, and the sequence was reverse-translated and codon-optimised for expression in E. coli K12, raising the codon adaptation index toward 1.00 and adjusting GC content to 65.63 percent. Restriction sites for Eco53kI and EcoRV, a Shine-Dalgarno ribosome-binding sequence and a stop codon were added to enable directional cloning into the pET-28a(+) vector for recombinant production.

The authors are emphatic that every result remains theoretical until confirmed experimentally. They list a series of limitations, including algorithmic bias in epitope prediction, the simplified nature of immune simulations, the absence of molecular dynamics analysis of the vaccine-receptor complex, and the lack of conservation analysis across genetically diverse T. gondii strains, which will be essential for assessing broad cross-protection. They also flag the inherent risks of multi-epitope design, such as epitope competition or immune masking, which flexible linkers can mitigate but not guarantee against. The proposed next steps include molecular dynamics simulations of the TLR-4 complex, advanced delivery platforms such as liposomes and nanoparticles, and challenge studies in murine models. If those experiments succeed, Rpf-Toxo could represent a meaningful step toward a long-sought human vaccine against one of medicine’s most persistent parasites.

Subject of Research: Immunoinformatic design of a multi-epitope vaccine candidate against Toxoplasma gondii based on dense granule antigens

Article Title: Rpf‐Toxo: A Preliminary Computationally Designed Dense Granule Antigen‐Based Multi‐Epitope Vaccine Against Toxoplasma gondii

Article References: Safari, M. H., Hosseini, S. A., Ghiabi, S., Ghiabi, S., Siamian, D., Majidiani, H., Irannejad, H., & Asghari, A. (2026). Rpf‐Toxo: A Preliminary Computationally Designed Dense Granule Antigen‐Based Multi‐Epitope Vaccine Against Toxoplasma gondii. Veterinary Medicine and Science, 12(5), Article e71189. https://doi.org/10.1002/vms3.71189

Image Credits: AI Generated

DOI: 10.1002/vms3.71189

Keywords: Toxoplasma gondii, toxoplasmosis, multi-epitope vaccine, dense granule antigens, reverse vaccinology, immunoinformatics, TLR-4, RpfE adjuvant, molecular docking, immune simulation, epitope prediction, vaccine design