Structure-Guided Identification of a Potential Small-Molecule Inhibitor Targeting the VacA Toxin of Helicobacter pylori
Key Insights
Researchers identified eight potential small-molecule inhibitors targeting the VacA (search) p55 domain of H. pylori through computational screening of 3,000 ligands from the ZINC15 database.
Lead compound ZINC4004291 (search) (ligand 8) demonstrated a binding affinity of −7.0 kcal/mol and superior stability in 100 ns molecular dynamics simulations compared to the reference molecule ranitidine (−5.0 kcal/mol).
ADMET predictions indicate ligand 8 has no AMES toxicity, no hERG inhibition, no hepatotoxicity, and no skin sensitization, suggesting a favorable preliminary safety profile.
A computational drug discovery study has identified a promising small-molecule candidate predicted to inhibit the vacuolating cytotoxin A (VacA (search)) of Helicobacter pylori (search), a class I carcinogen responsible for over 90% of stomach ulcers and more than 95% of duodenal ulcers worldwide. The research, published in PLOS ONE, employed an integrated in silico pipeline combining molecular docking, molecular dynamics simulations, and ADMET profiling to screen 3,000 compounds against the p55 domain of the VacA toxin.
H. pylori infection prevalence varies dramatically by geography, with recent studies documenting rates ranging from 9.2% in New Zealand to 87.8% in Northern Nigeria. The bacterium has been classified as a class I carcinogen by the World Health Organization due to its well-established role as the primary cause of stomach cancer. Current antibiotic-based treatments — including amoxicillin, clarithromycin, metronidazole, tetracycline, and bismuth — face intrinsic challenges such as antibiotic resistance, potential for reinfection, and considerable treatment costs.
Targeting Virulence Rather Than Bacterial Survival
The study focused on VacA (search), a 140 kDa protoxin secreted by H. pylori that undergoes cleavage into p33 and p55 domains to produce a mature 88 kDa toxin. The p55 domain is responsible for host-cell receptor binding, while the N-terminal p33 domain is associated with pore formation. VacA assembles into hexameric anion-selective channels within biological membranes, inducing cytoplasmic vacuolation in gastric epithelial cells — the phenomenon from which the toxin derives its name.
"Through the inhibition of virulence factors, such as extracytoplasmic molecules, these therapies aim to disrupt the bacteria's ability to induce disease without necessarily eradicating it entirely," the authors note, highlighting the advantage of reduced selection pressure for antibiotic-resistant strains and limited disturbance to the host microbiota.
Computational Screening Pipeline
The crystal structure of the VacA (search) p55 domain (PDB ID: 2QV3), determined at a minimum Bragg spacing of 2.4 Å, served as the docking target. The p55 structure is predominantly alpha-helical with a right-handed parallel twist, featuring a compact globular domain (residues 736–811) at the C-terminus. The refined homology model achieved 98.68% of residues in the favored region of the Ramachandran plot, a QMEAN Z-score of −1.00, and a MolProbity score of 1.32.
Active site prediction using CASTp and PrankWeb identified key binding pocket residues including THR287, ASP315, ALA317, THR318, PHE320, TYR321, LYS322, PRO323, LYS326, TYR375, ASN378, ASN379, ARG380, THR383, CYS384, VAL385, VAL386, ARG387, ASP391, ALA394, CYS395, and ALA398. Three internal tunnels (4.18 Å, 2.39 Å, and 2.36 Å) were identified as potential ligand access pathways.
From 3,000 ligands downloaded from the ZINC15 database, 178 compounds (5.93%) were found to bind the active site with scores ranging from −5.5 to −8.2 kcal/mol. Eight ligands were ultimately selected based on binding affinity and adherence to Lipinski's Rule of Five.
Lead Compound Performance
The eight candidate ligands demonstrated binding affinities ranging from −6.9 to −8.2 kcal/mol. Ligand 8 (ZINC4004291 (search)) exhibited a binding affinity of −7.0 kcal/mol, interacting with residues THR318, ARG380, VAL385, ASP391, ALA394, and CYS395 of the VacA (search) p55 domain. By comparison, the reference molecule ranitidine showed a substantially weaker binding affinity of −5.0 kcal/mol.
During 100 nanosecond molecular dynamics simulations conducted with GROMACS using the CHARMM27 force field and TIP3P water model, the VacA (search)-ligand 8 complex demonstrated superior stability. The ligand 8 RMSD averaged 0.075 nm compared to 2.977 nm for ranitidine. Radius of gyration analysis revealed increased compactness of the VacA-ligand 8 complex after 37.5 ns, with the complex maintaining tighter folding through the end of the simulation. Solvent-accessible surface area measurements corroborated these findings, with the docking complex SASA score falling notably below 205 nm² relative to the control.
Principal component analysis showed the VacA (search)-ligand 8 complex occupied a covariance area of 315.36 nm² versus 714.84 nm² for the ranitidine-VacA control, indicating more constrained and stable conformational dynamics. Gibbs free energy calculations revealed minimal thermodynamic differences between the complexes (18.6 kJ/mol for the lead complex vs. 18.5 kJ/mol for control).
MM/GBSA binding free energy decomposition indicated that van der Waals and electrostatic interactions contributed significantly to total binding energy for the lead compound-VacA (search) complex, suggesting hydrophobic contacts and electrostatic interactions were crucial for stabilizing the ligand within the binding pocket.
Favorable Predicted Safety Profile
ADMET analysis using pkCSM and SwissADME tools predicted that ligand 8 possesses optimal water solubility (−5.576 log mol/L), 91.688% human intestinal absorption, and a steady-state volume of distribution of 0.521 log L/kg. Critically, the compound showed no AMES toxicity, no hERG I or II inhibition, no hepatotoxicity, and no skin sensitization potential. The maximum tolerated dose in humans was predicted at −0.706 log mg/kg/day, with an oral acute toxicity (LD50) in rats of 3.069 mol/kg.
In contrast, ranitidine exhibited positive skin sensitization, indicating potential dermatological concern.
Study Limitations and Next Steps
The authors acknowledge that all findings are predictive in nature and require experimental validation. "Single trajectory simulations were performed; replicate simulations were not conducted due to computational constraints," they note, adding that "functional interpretations derived from the MD results are part of hypothesis-driven observation rather than definitive mechanistic conclusions."
The researchers propose a structured experimental validation roadmap including recombinant VacA (search) binding assays, surface plasmon resonance analysis, VacA-induced vacuolation assays in gastric epithelial cells, oligomerization interference studies, electrophysiological assays to evaluate pore-forming activity disruption, and cytotoxicity testing.
The study represents a computational lead discovery framework targeting the VacA (search) p55 domain, offering hypothesis-generating leads for future experimental validation in the pursuit of anti-virulence therapies against H. pylori infection.