The clinical urgency is difficult to overstate. H. pylori infects a large proportion of the global population and is firmly linked to chronic active gastritis, peptic ulcer disease, gastric mucosal atrophy, intestinal metaplasia, and gastric adenocarcinoma. Prospective cohort evidence has connected infection to subsequent gastric cancer development, and meta-analytic data indicate that eradication therapy can reduce cancer risk in appropriate populations. The Kyoto global consensus went so far as to define H. pylori gastritis as an infectious disease that should be diagnosed and treated wherever detected. Yet the therapeutic arsenal is weakening. The Maastricht VI/Florence Consensus Report urges clinicians to tailor regimens to local resistance patterns and prior antibiotic exposure, while the 2024 American College of Gastroenterology guideline now advises against empirical clarithromycin triple therapy unless susceptibility has been confirmed, recommending 14-day optimized bismuth quadruple therapy as the default empirical option when susceptibility is unknown.

Into this resistant landscape, the study’s authors introduce a deliberately small and transparent computational experiment. They selected three targets representing complementary biological functions: H. pylori urease, the nickel-dependent virulence enzyme that allows the bacterium to survive gastric acid by hydrolyzing urea; DNA gyrase subunit A, essential for DNA topology and replication; and shikimate kinase, part of a biosynthetic pathway present in many microorganisms and plants but absent from humans. Crystallographic structures from the Protein Data Bank served for urease (PDB 1E9Y, with its nickel-containing architecture carefully preserved) and shikimate kinase (PDB 1ZUH), while DNA gyrase A was represented by an AlphaFold model because a suitable experimental structure for blind docking was unavailable. Ligands were docked blindly using CB-Dock2, an automated server combining cavity detection with Vina-type scoring, and the resulting poses were inspected in three dimensions and mapped residue-by-residue in two-dimensional diagrams.

The choice of ligands was grounded in prior experimental data rather than computational convenience. Niclosamide, a salicylanilide anthelmintic long used against tapeworms, had already demonstrated striking anti-H. pylori activity in vitro: it inhibited strain 60190 with a minimum inhibitory concentration of 0.25 micrograms per milliliter, remained active at acidic pH, showed partial synergy with metronidazole and proton pump inhibitors, reduced bacterial adhesion and invasion of gastric epithelial cells by roughly 3 log CFU, and produced no detectable resistance after prolonged exposure in a Galleria mellonella infection model. Berberine, a quaternary ammonium alkaloid from plants such as Berberis species, carries clinical credentials of its own. A randomized phase IV trial found that a 14-day quadruple regimen containing berberine achieved eradication rates of 90.1 percent by intention-to-treat and 91.3 percent per protocol, non-inferior to a bismuth-containing regimen, and a later meta-analysis reported that adding berberine to standard triple therapy improved eradication with a risk ratio of 1.22.

The docking results positioned both candidates favorably. Against urease, niclosamide and berberine tied for the most negative score at –8.5 kcal/mol, outperforming levofloxacin’s –7.9 kcal/mol. Berberine again led against DNA gyrase A at –8.5 kcal/mol, ahead of niclosamide at –7.8 and levofloxacin at –7.5 kcal/mol. Shikimate kinase proved the least favorable target overall, with berberine at –7.3, levofloxacin at –7.0, and niclosamide at –6.9 kcal/mol. Across all three targets, berberine posted the most negative descriptive mean score at –8.10 ± 0.69 kcal/mol, compared with –7.73 ± 0.80 for niclosamide and –7.47 ± 0.45 for levofloxacin. The authors are categorical that these Vina-type outputs are ranking scores only, not binding free energies or affinity constants; a favorable docked score can occur even for a compound that fails to bind in solution.

Perhaps the most technically instructive finding emerged from the two-dimensional interaction maps: identical scores can mask entirely different binding modes. At urease, where niclosamide and berberine shared the same –8.5 kcal/mol value, their predicted stabilization patterns diverged sharply. Berberine’s pose relied mainly on pi-alkyl, pi-sigma, and hydrophobic contacts with residues including LEU A:13, ALA A:16, VAL A:33, PHE B:441, and TYR A:32, while niclosamide’s pose combined conventional hydrogen bonds with hydrophobic, pi-related, and van der Waals contacts involving VAL B:473, TYR B:475, PRO B:472, and MET A:12. At DNA gyrase A, berberine’s aromatic contacts with PHE A:549 and TYR A:525 drove its top ranking, whereas levofloxacin interacted with ASP A:91, ARG A:95, and GLN A:98, including one unfavorable positive-positive contact. The authors caution, however, that the isolated DNA gyrase A model cannot reproduce the DNA–enzyme–drug ternary complex on which fluoroquinolone pharmacodynamics actually depends.

The ADMET profiling added a sobering pharmacokinetic dimension. All three compounds showed zero Lipinski rule-of-five violations and identical bioavailability scores of 0.55, with high predicted gastrointestinal absorption, but their developability profiles differ meaningfully. Niclosamide’s consensus logP of 3.78 and topological polar surface area of 49.33 square angstroms predict membrane permeability, yet its poor aqueous solubility makes formulation-oriented work essential. Berberine’s permanent quaternary cationic structure, which was preserved in the docking preparation, is associated with complex, formulation-dependent exposure that demands experimental confirmation. Neither descriptor set, the study emphasizes, constitutes evidence of therapeutic suitability; they flag questions about exposure, solubility, and formulation that only laboratory measurement can resolve.

The methodological self-awareness sets this study apart from much of the computational drug-repurposing literature. The authors explicitly note that no redocking validation, receiver operating characteristic benchmarking, molecular dynamics simulations, or MM/GBSA or MM/PBSA free-energy analyses were performed, meaning pose accuracy and screening enrichment cannot be claimed. Alternative protonation and tautomeric states of the ligands were not systematically enumerated, a constraint that can influence both poses and scores. Thermodynamically, they explain, a docking score approximates none of the free-state, solvation, desolvation, and entropic contributions required for a true binding free energy, and virtual screening workflows carry substantial false-positive rates. With only three ligands and three targets, no inferential statistics were applied; all summaries are purely descriptive of this specific matrix.

The study also situates itself honestly within prior evidence. The docking results neither reproduce niclosamide’s previously observed antibacterial effects nor identify their mechanism; instead they generate a separate, target-oriented hypothesis that urease, gyrase, and shikimate kinase engagement could be tested experimentally. Similarly, the mixed clinical record of berberine—a more recent randomized trial found no clear superiority for berberine-containing triple therapy, and reported MIC values against H. pylori vary widely by strain and method—means the alkaloid should be viewed as a scaffold for mechanistic and formulation-focused testing rather than a docking-supported monotherapy. The shikimate kinase results, with their small score differences, are downgraded to a lower-priority experimental hypothesis despite the target’s theoretical selectivity advantage.

What remains is a deliberately restricted but actionable set of next steps. The authors propose urease inhibition assays as the first priority, followed by broth microdilution or agar-dilution MIC determination against reference and clinical H. pylori strains, checkerboard combination testing with standard antibiotics, cytotoxicity screening on gastric epithelial cell models, and, if justified by results, molecular dynamics and post-docking free-energy analyses. Within this framework, berberine earns the most consistent score pattern, niclosamide remains the clearest repurposing candidate for urease-focused testing given its prior experimental pedigree, and levofloxacin serves its role as a clinically interpretable yardstick. None of these outputs demonstrates binding, enzyme inhibition, bacterial killing, or therapeutic efficacy—but as an exercise in reducing the experimental search space before costly microbiological testing, the study offers a model of how computational prioritization can be conducted with the caveats firmly attached.

Subject of Research: Computational docking and ADMET profiling of niclosamide and berberine against Helicobacter pylori urease, DNA gyrase A, and shikimate kinase

Article Title: Comparative molecular docking and ADMET profiling of niclosamide and berberine against multiple Helicobacter pylori enzymatic targets

Article References: Comparative molecular docking and ADMET profiling of niclosamide and berberine against multiple Helicobacter pylori enzymatic targets. (n.d.). https://doi.org/10.3897/pharmacia.73.e212118

Image Credits: AI Generated

DOI: 10.3897/pharmacia.73.e212118

Keywords: Helicobacter pylori, molecular docking, niclosamide, berberine, levofloxacin, ADMET, drug repurposing, urease, DNA gyrase, shikimate kinase, antimicrobial resistance, in silico screening