Computational Study Analyzes Plant Compounds Against SARS-CoV-2 Main Protease Target

A computational docking investigation evaluating nine phytochemicals from two medicinal plants associated with traditional Siddha formulations found that structurally diverse natural molecules can engage the catalytic and substrate-recognition regions of the SARS-CoV-2 main protease through distinct polar and hydrophobic mechanisms. The study, conducted by researchers Anbarasan K. and C. Sasikala, applied a standardized in silico docking protocol to compare how alkaloid, flavonoid, polyphenol, sterol, and triterpenoid scaffolds interact with the enzyme known as 3CLpro or Mpro. Across the tested panel, predicted binding free energies ranged from −6.43 kcal/mol to −11.43 kcal/mol, with the glycosylated flavonoid Rutin and the sterols Stigmasterol and β-Sitosterol exhibiting the most favorable predicted computational affinity scores.

The target of the investigation, the SARS-CoV-2 main protease, is an essential viral enzyme responsible for cleaving large polyproteins into functional non-structural proteins required for viral replication and transcription machinery, including enzymes for RNA synthesis. Because the main protease is indispensable for viral maturation and features a conserved catalytic center based on the His41 and Cys145 dyad, it represents a primary focal point for antiviral inhibitor discovery. Additionally, scientific literature notes that the protease modulates host-cell pathways linked to innate immune defence and programmed cell death, and its activity has been suggested to play a role in systemic inflammatory manifestations observed during severe infection. Consequently, identifying small molecules capable of obstructing its active site remains an active avenue in antiviral development.

Natural products and traditional healthcare systems, including India’s Siddha medicine, provide extensive reservoirs of structurally diverse secondary metabolites. During the COVID-19 pandemic, Siddha-based interventions gained public and clinical attention, particularly in Tamil Nadu where select traditional treatments were introduced into designated care facilities. Among these preparations is Sivanethirachooranam, a formulation classified under Gunapadam Mooligai that utilizes the seed of Elaeocarpus sphaericus, locally known as Rudraksha, which is purified, finely powdered, and traditionally combined with honey to manage conditions categorized in Siddha practice as sannipatha thodam, vikkal, pitta-associated dizziness, and specific kapha-related disorders. However, because clinical validation for many traditional botanical formulations is frequently constrained by heterogeneous trial designs and small cohorts, computational molecular modeling provides a preliminary screening method to assess whether individual constituents exhibit direct mechanistic affinity for viral targets.

To establish a controlled comparative framework, the researchers selected nine phytochemicals representing two plant sources: Elaeocarpus sphaericus and Aristolochia indica. Four compounds—Elaeocarpine, Grandisines, Quercetin, and Ellagic acid—were derived from reported phytochemical literature on E. sphaericus, providing alkaloidal and polyphenolic frameworks. The remaining five compounds—β-Sitosterol, Stigmasterol, Friedelin, Cycloeucalenol, and Rutin—were selected based on the documented chemical constituents of A. indica, supplying sterol, triterpenoid, and complex flavonoid glycoside scaffolds. Because earlier literature had already documented individual computational assessments for several of these molecules, the explicit purpose of this investigation was not to assert first-time discovery of binding activity, but to systematically analyze how scaffold rigidity, molecular weight, hydrogen-bonding capacity, and hydrophobic surface area alter binding geometry under identical AutoDock conditions.

Before docking simulations commenced, two-dimensional structures were built using ChemDraw Professional version 12.0 and converted to three-dimensional conformations, followed by geometry optimization and energy minimization via the Merck Molecular Force Field 94 to eliminate unfavorable steric strain. Physicochemical characterization revealed notable structural variation across the ligand set, with molecular weights spanning from 257.33 g/mol for Elaeocarpine to 610.5 g/mol for Rutin. Grandisines, Quercetin, and Ellagic acid exhibited intermediate molecular masses of 276.37 g/mol, 302.23 g/mol, and 302.194 g/mol, respectively. In contrast, the sterol and triterpenoid derivatives displayed higher molecular weights, with Stigmasterol at 412.7 g/mol, β-Sitosterol at 414.7 g/mol, and both Friedelin and Cycloeucalenol at 426.7 g/mol. Conformational flexibility also diverged sharply: Elaeocarpine, Grandisines, Ellagic acid, and Friedelin presented zero rotatable bonds, reflecting rigid structures, whereas Stigmasterol and Cycloeucalenol possessed five rotatable bonds, and β-Sitosterol and Rutin each contained six.

The structural model chosen for the viral target was the experimentally determined crystal structure of SARS-CoV-2 3CLpro complexed with inhibitor N3, retrieved from the RCSB Protein Data Bank under accession code PDB ID: 6LU7, solved at an X-ray crystallographic resolution of 2.16 Å with an R value of approximately 0.202. Using AutoDock Tools, water molecules and the co-crystallized N3 ligand were stripped, polar hydrogen atoms were added, Kollman united-atom charges were applied, and solvation parameters were established. Validation through the MolProbity server confirmed high stereochemical quality, placing 97.05% of protein residues within favored regions of the Ramachandran plot with zero residues in disallowed regions, alongside an overall Ramachandran Z-score of −1.92 ± 0.41 and 87.17% of side-chain rotamers residing in favored conformations. The docking grid was constructed with AutoGrid, centered at coordinates x = −9.732 Å, y = 11.403 Å, and z = 68.925 Å with a 60 × 60 × 60 point grid box and 0.375 Å spacing, encompassing His41, Cys145, and the broader substrate-binding cleft.

Simulations utilized the Lamarckian Genetic Algorithm coupled with Solis and Wets local searching across two independent runs per compound, capping energy evaluations at 250,000 per run with an initial population of 150 individuals, a translational step size of 0.2 Å, and rotational and torsional step sizes of 5°. When the resulting poses were ranked by calculated binding free energy, Rutin demonstrated the strongest predicted binding affinity at −11.43 kcal/mol, paired with a calculated inhibition constant of 4.19 nM, an electrostatic energy contribution of −0.03 kcal/mol, an intermolecular energy of −6.73 kcal/mol, and a total interaction surface of 840.57 Ų. In comparison, Stigmasterol achieved −9.66 kcal/mol with a calculated Ki of 82.60 mM, an intermolecular energy of −10.11 kcal/mol, and an interaction surface of 811.29 Ų, while β-Sitosterol scored −9.09 kcal/mol with a predicted Ki of 216.19 nM, an intermolecular energy of −10.37 kcal/mol, and the largest interaction surface area among all ligands at 946.79 Ų.

The triterpenoids Friedelin and Cycloeucalenol produced docking free energies of −8.72 kcal/mol and −8.68 kcal/mol, accompanied by predicted inhibition constants of 406.72 nM and 318.21 nM, intermolecular energies of −8.72 kcal/mol and −10.60 kcal/mol, and surface contact areas of 866.97 Ų and 946.67 Ų, respectively. Meanwhile, the smaller phenolic and alkaloidal molecules displayed moderate binding energies: Ellagic acid scored −7.04 kcal/mol with a Ki of 6.89 μM; Grandisines reached −6.89 kcal/mol with a Ki of 8.89 μM; Elaeocarpine yielded −6.86 kcal/mol with a Ki of 9.44 μM; and Quercetin showed −6.43 kcal/mol with a Ki of 19.38 μM. These smaller scaffolds generated comparatively compact total interaction surfaces, measuring 592.17 Ų for Ellagic acid, 635.61 Ų for Elaeocarpine, 642.93 Ų for Quercetin, and 679.77 Ų for Grandisines, alongside modest electrostatic contributions between −0.01 and −0.33 kcal/mol.

Detailed examination of the simulated binding poses revealed distinct amino-acid engagement profiles across the catalytic site and neighboring substrate-recognition residues, which include Leu27, Gly143, His163, His164, Met165, Glu166, Pro168, and His172. Stigmasterol engaged the highest number of predefined core residues, forming seven contacts across Phe140, Asn142, Gly143, Cys145, His163, Met165, Glu166, Pro168, His172, Gln189, and Gln192, directly interacting with catalytic residue Cys145 although lacking contact with His41. Both β-Sitosterol and Grandisines engaged six core residues; β-Sitosterol contacted Leu27, His41, Met49, Asn142, Ser144, Cys145, Met165, Glu166, Pro168, and Gln189, whereas Grandisines contacted His41, Met49, Tyr54, Gly143, Cys145, His163, Met165, Glu166, and Gln189, allowing both compounds to contact both catalytic dyad partners simultaneously.

Rutin, Cycloeucalenol, and Quercetin each established contacts with five core residues. Rutin engaged Thr25, Leu27, His41, Phe140, Asn142, Ser144, Cys145, His163, Met165, and Gln189, successfully spanning both His41 and Cys145 within the active site. Cycloeucalenol contacted Thr25, Leu27, His41, Met49, Asn142, Cys145, Met165, Pro168, and Gln192, also engaging both catalytic residues. Conversely, Quercetin contacted His41, Phe140, Asn142, His163, Met165, Glu166, His172, and Gln189, interacting with His41 while omitting contact with Cys145. For the remaining molecules, Elaeocarpine and Ellagic acid each engaged four core residues, with both ligands contacting both His41 and Cys145, while Friedelin established three core residue interactions encompassing Thr25, His41, Asn142, Cys145, Met165, and Gln189.

The study highlights how distinct chemical architectures achieve affinity through complementary energetic mechanisms. In the comparison between Quercetin and its glycoside Rutin, the addition of a disaccharide moiety expands Rutin’s polar functionality to ten hydrogen-bond donors and sixteen acceptors, allowing extensive multipoint polar interactions that yield a substantially more negative docking score than Quercetin’s compact aglycone core. In sharp contrast, the sterols Stigmasterol and β-Sitosterol possess only one hydrogen-bond donor and one acceptor, relying instead on hydrophobic packing, van der Waals forces, and large non-polar contact surfaces to achieve favorable docking energetics within the protease cavity.

The authors emphasize critical methodological limitations inherent to docking simulations. Computational docking scores and AutoDock-calculated inhibition constants represent static energetic estimates rather than direct measurements of biochemical enzyme inhibition or antiviral efficacy. Furthermore, high predicted target affinity does not account for pharmacokinetic hurdles; Rutin’s molecular weight and high hydrogen-bonding count exceed Lipinski Rule-of-Five thresholds, presenting challenges for oral bioavailability and membrane permeability. Because the investigation evaluated a rigid crystal receptor without dynamic simulation, the authors conclude that molecular dynamics trajectories, MM-PBSA or MM-GBSA binding free-energy calculations, in vitro enzymatic inhibition assays, ADME profiling, and cell-based viral replication studies remain necessary to determine whether any of these prioritized botanical scaffolds possess genuine therapeutic utility.

Subject of Research: Structure-based molecular docking of plant-derived phytochemicals against the SARS-CoV-2 main protease

Article Title: Structure-based docking analysis of siddha formulation derived phytochemicals targeting the SARS-CoV-2 Main protease (3CLpro): Insights into plant-based antiviral drug discovery

Article References: K., A., & Sasikala, C. (2026). Structure-based docking analysis of siddha formulation derived phytochemicals targeting the SARS-CoV-2 Main protease (3CLpro): Insights into plant-based antiviral drug discovery. Results in Chemistry, 32, Article 103984. https://doi.org/10.1016/j.rechem.2026.103984

Image Credits: AI Generated

DOI: 10.1016/j.rechem.2026.103984

Keywords: SARS-CoV-2, 3CLpro, Molecular Docking, Phytochemicals, Siddha Medicine, Drug Discovery, Rutin, Antiviral Research, Structure-based, docking, analysis, siddha