Shallot-Derived Organosulfur Compounds Show Computational Promise Against H5N1 Influenza

A team of Indonesian and international researchers has used a battery of computational tools to ask whether the humble shallot, Allium ascalonicum L., could yield new antiviral leads against highly pathogenic avian influenza A (H5N1). Writing in Frontiers in Virology, the group reports that four organosulfur compounds isolated from shallots grown in the Tolaki-Mekongga region of Sulawesi, Indonesia, bind favorably in silico to three key H5N1 viral proteins: the PB2 polymerase subunit, the nucleoprotein (NP), and neuraminidase (NA). One compound in particular, γ-glutamyl-S-propenylcysteine, emerged as the standout candidate, showing stable protein-ligand complexes over 100-nanosecond molecular dynamics simulations and favorable binding free energies that in some cases rivaled or exceeded those of the approved antiviral drugs oseltamivir and favipiravir.

The motivation for the study is grounded in a growing public health concern. H5N1 remains a major threat due to its high pathogenicity and ongoing viral evolution, with the virus capable of rapid genetic change through mutation, reassortment, and recombination. Recent human cases in the United States, including the first severe H5N1 infection linked to backyard poultry exposure in Louisiana, and the identification of distinct viral genotypes such as D1.1 and B3.13 in human cases, underscore the virus’s ability to cross species barriers. Meanwhile, the emergence of resistance to oseltamivir and the potential for resistance to favipiravir, despite its high genetic barrier, highlight the need for alternative therapeutic agents with novel mechanisms of action.

Natural products have long attracted attention in antiviral research because of their structural diversity, multiple mechanisms of action, and generally favorable safety profiles. Plant-derived compounds, including flavonoids, alkaloids, terpenoids, polyphenols, and organosulfur compounds, can interfere with multiple stages of the viral life cycle, from attachment and entry to replication, assembly, and release. This multi-target capacity is particularly valuable against influenza viruses, which can rapidly develop resistance to conventional single-target antiviral agents. Organosulfur compounds, abundant in members of the Allium genus such as garlic, onions, leeks, and chives, are of special interest because they can interact with protein thiol groups, modulate cellular redox homeostasis, and form disulfide bonds with conserved cysteine residues in viral proteins such as hemagglutinin and neuraminidase.

Shallots have a long history of traditional medicinal use in Indonesia, particularly in the Tolaki-Mekongga region of Sulawesi, where they have been used to treat respiratory infections, fever, and influenza-like illnesses. Similar traditional use has been documented in Banyumas, Indonesia. Modern phytochemical analyses have confirmed that shallots contain a complex profile of organosulfur compounds, including various sulfides, thiosulfinates, and sulfoxides, which exert antiviral effects through direct virucidal activity, interference with viral attachment and entry, and modulation of host immune responses. Building on this ethnopharmacological foundation, the research team selected four representative organosulfur compounds for computational evaluation: Propanethiol, Dipropyl disulfide, γ-glutamyl-S-propenylcysteine, and S-n-Propenyl-L-cysteine-S-oxide.

The study employed an integrated in silico pipeline that combined density functional theory (DFT) calculations, quantitative structure-activity relationship (QSAR) screening, pharmacokinetic profiling, protein structure modeling, molecular docking, and molecular dynamics simulations. DFT calculations were performed using the Gaussian16 program at the B3LYP/6-311++g(d,p) level of theory to characterize the electronic properties of each compound, including HOMO-LUMO energy gaps, hardness, softness, chemical potential, electronegativity, and electrophilicity index. The energy gap values ranged from 5.61 to 6.11 electron volts across the four compounds, with Propanethiol exhibiting the largest gap and therefore the greatest chemical stability, while the smaller gaps of the other compounds suggested enhanced reactivity and potentially stronger interactions with biological targets.

QSAR analysis using the Way2Drug PASS Online server predicted moderate potential antiviral activity for Propanethiol (Pa = 0.337) and Dipropyl disulfide (Pa = 0.361), both exceeding the medium-confidence threshold of Pa > 0.3. Pharmacokinetic profiling using pkCSM revealed that all four compounds comply with Lipinski’s rule of five, with molecular weights ranging from 76.164 to 290.341 g/mol and favorable hydrogen bond donor and acceptor counts. ADMET analysis showed acceptable water solubility, moderate tissue distribution, and no predicted Ames toxicity or hepatotoxicity for any of the compounds. Notably, none of the compounds acted as CYP3A4 substrates, suggesting they would not be eliminated by this common metabolic pathway, and all showed maximum tolerated dose values above the low-tolerance threshold of 0.447.

To model the viral targets, the researchers obtained nucleotide sequences for PB2, NP, and NA from the NCBI database, translated them into protein sequences using ExPASy, and built three-dimensional structures with the Phyre2 webserver. Structural validation using ERRAT, Ramachandran plots, and ProSA-web confirmed that all three protein models were of acceptable to high quality, with ERRAT scores ranging from 76.18 to 85.99, Z-scores between -8.57 and -8.92, and 92 to 97 percent of residues in favored Ramachandran regions. These validated structures were then used as receptors for molecular docking with AutoDock Vina, with binding pockets predicted by PrankWeb.

Molecular docking revealed favorable binding conformations across all three targets, with root mean square deviation values below the 2.0 Å threshold for all ligand-receptor pairs. γ-glutamyl-S-propenylcysteine exhibited the most favorable binding energies among the four shallot-derived compounds, with values of -4.9 kcal/mol for PB2, -5.8 kcal/mol for NP, and -5.2 kcal/mol for NA. These values were comparable to those of oseltamivir and favipiravir for NP and NA, although weaker binding was observed against PB2. Notably, γ-glutamyl-S-propenylcysteine showed a more favorable binding energy than favipiravir against NP (-5.8 vs. -5.2 kcal/mol) and than oseltamivir against NA (-5.2 vs. -3.6 kcal/mol). The compound formed multiple hydrogen bonds with key residues in each receptor, including ASN515, HIS362, and PHE328 in PB2; SER176, ARG278, PHE499, PHE349, and SER497 in NP; and THR314, CYS318, TYR321, GLN407, and GLU409 in NA.

Molecular dynamics simulations over 100 nanoseconds using GROMACS 2023.02 confirmed the stability of the protein-ligand complexes. All organosulfur compounds exhibited relatively small structural deviations, with root mean square deviation values remaining below 0.2 nanometers throughout the simulation trajectory. Radius of gyration and solvent accessible surface area analyses indicated that the protein systems became more compact upon ligand binding, particularly when complexed with the larger organosulfur compounds. Hydrogen bond occupancy analysis showed that γ-glutamyl-S-propenylcysteine and S-n-Propenyl-L-cysteine-S-oxide maintained high occupancy rates ranging from 72.313 to 100 percent, while the shorter-chain compounds Propanethiol and Dipropyl disulfide showed negligible hydrogen bond occupancy, relying instead on hydrophobic and van der Waals forces. MM-GBSA calculations further quantified the binding free energies, with γ-glutamyl-S-propenylcysteine showing the most favorable values: -30.52 ± 0.29 kcal/mol for PB2, -22.76 ± 0.12 kcal/mol for NP, and -26.13 ± 0.35 kcal/mol for NA.

The authors emphasize that these computational findings, while encouraging, represent only a preliminary step in the drug discovery process. Experimental validation through in vitro and in vivo studies is required to confirm the biological activity and therapeutic potential of the identified compounds. Nevertheless, the integrated approach demonstrates the value of combining quantum chemical calculations, QSAR screening, pharmacokinetic prediction, molecular docking, and molecular dynamics simulations to prioritize natural product candidates for further development. Among the evaluated compounds, γ-glutamyl-S-propenylcysteine has emerged as the most promising candidate inhibitor based on its consistently favorable docking, molecular dynamics, and MM-GBSA results, suggesting that shallot-derived organosulfur compounds may warrant deeper investigation as potential anti-influenza agents against H5N1.

Subject of Research: In silico evaluation of shallot-derived organosulfur compounds as potential inhibitors of influenza A (H5N1) viral proteins

Article Title: In silico identification of organosulfur compounds from Allium ascalonicum L. as potential inhibitors of influenza A (H5N1): integrated DFT, docking, and molecular dynamics analysis

Article References: Prastika, R. A., Putra, A. P. E., Fadhana, M. A., Setiawan, V. K., Afnani, M. R., Sahadewa, S., Durry, F. D., Khairullah, A. R., Adnyana, I. M. D. M., Singh, A., Rebezov, M., Abdurasulov, A., Jakhmola, V., Sharma, N., Kupaeva, N. V., Siswanto, I., Ansori, A. N. M., & Parikesit, A. A. (2026). In silico identification of organosulfur compounds from Allium ascalonicum L. as potential inhibitors of influenza A (H5N1): integrated DFT, docking, and molecular dynamics analysis. Frontiers in Virology, 6, Article 1857484. https://doi.org/10.3389/fviro.2026.1857484

Image Credits: AI Generated

DOI: 10.3389/fviro.2026.1857484

Keywords: H5N1 influenza, organosulfur compounds, Allium ascalonicum, molecular docking, molecular dynamics, DFT, MM-GBSA, antiviral candidates, natural products, drug discovery, neuraminidase, PB2 polymerase