Dengue is a virus that is transmitted by mosquitoes and is the primary cause of arthropod-borne viral diseases around the globe. Dengue virus (DENV), which is mostly spread by Aedes aegypti and Aedes albopictus, can cause a variety of illnesses, from self-limiting dengue fever to potentially fatal diseases including Dengue Hemorrhagic Fever (DHF) and Dengue Shock Syndrome (DSS). There are currently no specific FDA-approved antiviral drugs available for clinical use, and current vaccines have issues with efficacy across all four serotypes and the risk of antibody-dependent enhancement, despite the high global burden of the virus and its increasing geographic expansion due to urbanization and climate change^1,2.

Figure 1. Global distribution of dengue virus serotypes (DENV1–4) based on WHO data (1960–2025). The pie charts represent the relative proportion of dengue virus serotypes across various geographic regions, highlighting the simultaneous circulation of multiple serotypes. (Source: World Health Organization (WHO). (2025). Global Dengue Surveillance: Genomic Data Dashboard. Retrieved June 27, 2025, from https://worldhealthorg.shinyapps.io/dengue_global/)³

The pie charts in figure 1 represent the relative proportion of dengue virus serotypes across various geographic regions, highlighting the simultaneous circulation of multiple serotypes. DENV 1 accounts for the highest proportion (40.37%), followed by DENV2 (32.32%), DENV3 (18.10%), and DENV4 (9.20%). Co-circulation is most intense in Southeast Asia and South America, underscoring the risk of secondary infections and antibody-dependent enhancement (ADE). In India, DENV2 is the most prevalent (33.72%), followed by DENV3 (29.43%), DENV1 (27.38%), and DENV4 (9.47%)³.

Dengue has become prevalent worldwide in both endemic and epidemic transmission patterns. The Indian subcontinent, because of its favorable environment, has numerous records of dengue outbreaks that involve all serotypes². Although all major areas of the globe are impacted by rising dengue rates, around 75% of individuals infected with dengue reside in the Asia-Pacific region⁴. The World Health Organization identified dengue as one of the “top ten threats to global health in 2019”. Dengue is present globally and appears endemically, sporadically, or as epidemics⁵.

Seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) and three structural proteins (Capsid, Membrane, and Envelope) are encoded by the DENV genome. While non-structural proteins like the NS2B/NS3 protease and NS5 polymerase are crucial for RNA replication and polyprotein processing, respectively, structural proteins are necessary for viral assembly and entrance. The NS1 protein is also essential for vascular leakage and immune evasion. A viable approach to achieving broad-spectrum efficacy and lowering the risk of drug resistance is the development of therapies that can concurrently target several stages of this intricate viral life cycle^2,6.

Natural products have long been a rich supply of antiviral scaffolds with good safety profiles, especially secondary metabolites including flavonoids, alkaloids, and terpenoids. Many of these substances have innate anti-inflammatory and antioxidant qualities, which are essential for controlling the oxidative stress and cytokine storm linked to severe dengue. Since many of these compounds have polypharmacological actions that can disrupt both viral replication and host-mediated pathogenesis, using a library of natural products provides an advantageous starting point for drug discovery^7,8. To find effective inhibitors against a panel of six crucial DENV targets-the capsid, membrane, envelope, NS2B/NS3 protease, NS5 polymerase, and NS1 protein, we used an integrated computational and experimental strategy in this investigation.

1.    METHODS

1.1.   Computational Screening of Phytocompounds

1.1.1.   Pre-processing of the structures

The Schrödinger Suite 2022-3 version (Schrödinger, LLC, New York, NY, 2022) was used to conduct HTVS of the compound library of natural products (Cat. No.: HY-L021) acquired from the MedChemExpress (MCE)^9,10. The three-dimensional structures of the most common and validated inhibitory targets of the DENV were taken from RCSB PDB database¹¹. The structures of lead compounds were taken from the PubChem database. Table 1 includes the proteins that were selected based on the literature and structural parameters. The pre-processing of the structures involved the assignment of bond orders, the addition of disulfide bonds and missing hydrogens, the use of the Prime module to fill in loop gaps, the removal of water molecules, ions, metals, and other co-crystallized ligands, and hydrogen bond optimization to optimize charge-charge interaction and hydrogen bonding via the Maestro workspace of the suite⁹.

MCE offers a special assortment of the natural compounds, including flavonoids, alkaloids, quinones and terpenoids which are known to inhibit NS3 protease, NS5 polymerase, and envelope-mediated entry, making this library ideal for mechanistic screening. The drug library was prepared for HTVS using Schrödinger's LigPrep module. Using the force field OPLS_2005 (Optimized Potentials for Liquid Simulations), the energy of each pharmacological compound was minimized at pH 7.0 ± 2⁹. The screened compounds were also compared with the list of antiviral phytocompounds from Dr. Duke’s phytochemical and ethnobotanical database to filter molecules with antiviral biological activity¹².

Table 1. Dengue virus drug target proteins retrieved from the RCSB PDB.

1.1.2.  Binding site prediction and grid generation

Using Schrödinger Glide's receptor grid generation module, a grid surrounding the co-crystallized ligands was defined with a partial charge cut-off of 0.25 and a scaling factor of 1. Using Schrödinger Glide's SiteMap tool, the binding site residues for the protein which does not have a co-crystallized ligand were predicted^18,19.

1.1.3.  HTVS of compound libraries against DENV structural proteins

Using the Schrödinger glide module, ligand-flexible docking was carried out using the HTVS precision mode and the SP (standard precision) mode using the standard inhibitors for comparison. Because glide docking approaches produce a docking score that is tied to the free energy of attaching a ligand to a receptor, the more negative the docking score, the better and larger the binding affinity^19,20.

1.1.4.  MM/GBSA Free Energy of Binding Calculations

Molecular mechanics with generalised Born and surface area solvation (MM/GBSA) uses generalized Born (GB) and surface area (SA) terms to account for solvent effects. Further, this binding free energy (dGbind) calculation was also used to assess the strength of the inhibitor and the protein interaction as modelled. The following formula is used to calculate the binding-free energy of a protein and ligand:

where ΔG_protein and ΔG_ligand are the free energies of the protein and ligand in the solvent, respectively, and ΔG_Bind is the total free energy of the complex. The prime module of the suite was used for these calculations using the OPLS4 force field and VSGB 2.0 solvation model. These analyses were done on all the top five hits against all the DENV target proteins^9,18-20.

1.1.5.  Molecular Dynamics Simulation

The Desmond product of the Schrödinger suite was employed to perform the simulations for the docked complexes at 100 ns. The docked complexes were placed in an orthorhombic shaped box at a buffer distance of 16 Å solvated with SPC water model. To neutralize the systems, more Na+/Cl− ions were added and pressure and temperature were set at 1.01325 bar and 310 K, respectively^18-20. The MD simulation trajectories were analyzed using simulation interaction diagram tool of Desmond module⁹. 

1.2.   In vitro analysis of the top hit compounds

1.2.1.  Chemicals and Reagents

The compounds screened via computational analyses were procured and purchased to evaluate the anti-dengue potential against viral cell lines. The compounds diosmetin (Cat No.: D7321) and vitexin (Cat No.: 49513) were purchased from Merck Sigma-Aldrich. All other chemicals and the reagents used were of analytical grade and purchased from Merck Sigma-Aldrich and Himedia, India. All the compounds were dissolved in DMSO, aliquoted to yield 50 mg/mL stocks and stored at -80ºC until further use.

1.2.2.  Anti-inflammatory activity

For the ROS scavenging assay, HepG2 cells were seeded in 6-well plates at a density of 1 × 10⁶ cells/well and treated with different concentrations of NP4 and NP5 compounds. The positive control group was exposed to 500 µM hydrogen peroxide (H₂O₂) for 1 hour to induce oxidative stress. Following treatment, cells were washed twice with PBS and incubated with 10 µM DCFH-DA (Dichlorofluorescin diacetate) for 30 minutes at 37°C. DCFH-DA is a non-fluorescent compound that, once inside the cells, is hydrolyzed to DCFH and then oxidized by ROS, resulting in the fluorescent DCF product. The fluorescence intensity of DCF was measured using a flow cytometer (BD FACSCalibur) at an excitation wavelength of 488 nm and emission at 525 nm. The percentage of cells showing fluorescence was used as an indicator of ROS levels in each group. Data were analyzed using FlowJo software, and the results were compared between the control, H₂O₂-treated, and NP4 and NP5-treated groups to assess the ROS scavenging activity of the compounds^21-23.

1.2.3.  Cytotoxicity Assay

To determine the non-toxic dose range of phytocompounds, the MTT assay was performed using Vero E6 cells that were seededat a density of 5 × 10⁴ cells/well in 100 µL complete DMEM and incubated at 37°C for 24 hours²³. Test compounds were added in serial dilutions ranging from 3.13 to 200 µg/mL. After treatment, 20 µL of 5 mg/mL 3-(4, 5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2 H-tetrazolium bromide (MTT) solution was added to each well and incubated for 3 hours at 37°C. The formazan crystals formed were solubilized in 100 µL of DMSO, and absorbance was read at 570 nm and 490 nm using a microplate reader (Multiscan microplate spectrophotometer, Model No. 1530, Multisky, Thermo Fisher Scientific, USA). Quercetin was used as the positive control. Dose–response curves were constructed to determine CC₅₀ (concentration reducing cell viability by 50%) and MNTD₈₀ (maximum non-toxic dose maintaining ≥80% viability) using nonlinear regression in GraphPad Prism^24-26.

1.2.4.  Plaque Reduction Neutralization Test (PRNT)

The antiviral efficacy of NP4, NP5, and quercetin against DENV-2 was determined by PRNT as adapted from Cheng et al. (2002) and Jayasekara et al. (2024)^25,27. Cells seeded at 2 × 10⁵ cells/ml were pre-treated with varying concentrations of the compounds. After pretreatment, compound solutions were removed, cells were washed with PBS, and infected with DENV-2 suspension containing approximately 90–100 PFU/well. Virus adsorption was allowed for 1 hour at 37°C, followed by overlaying cells with medium containing 1% methylcellulose or carboxymethyl cellulose to restrict viral spread. After a 5-day incubation, the cells were fixed with 10% formalin, followed by staining with 1% crystal violet. Plaques were counted manually, and percentage plaque inhibition compared to virus controls was calculated. The IC₅₀ (compound concentration achieving 50% plaque reduction) was calculated by nonlinear regression using GraphPad Prism^25-27.

1.2.5.  In vitro Cytopathic Effect (CPE)-based Antiviral Assay

The antiviral potential of phytocompounds was evaluated using a CPE-based in vitro antiviral assay. Vero E6 cells were seeded into 96-well plates at a density of 1.5 × 10⁴ cells/well and incubated at 37°C with 5% CO₂ for 24 hours. After virus adsorption, inoculum was removed, and cells were washed with serum-free medium. The test compounds were added at their predetermined MNTD concentrations (based on MTT assay results) to the respective wells. The plate was further incubated for 5 days, with daily microscopic evaluation of CPE. The degree of viral inhibition was assessed using the standard grading scale described as : ‘++++’ = complete inhibition of CPE, ‘+++’ = 75% inhibition, ‘++’ = 50% inhibition, ‘+’ = <50% inhibition, ‘–‘ = no inhibition. Each experimental condition was tested in triplicates^24,25,27,28.

To validate the antiviral effect observed microscopically, MTT assays were performed in parallel to quantify cell viability. Cells from the above treatment groups (cells alone, DENV-infected cells, and infected cells treated with test compounds) were incubated with MTT reagent for 3 hours. The resulting formazan crystals were dissolved in DMSO, and absorbance was measured at 570 nm using a microplate reader. The percentage cell viability was calculated, and results were statistically compared to the virus control group using an unpaired Student’s t-test. Results were expressed as mean ± SD. Differences were considered statistically significant at p < 0.05. All statistical analyses were performed using GraphPad Prism^24,28.

RESULTS and DISCUSSION

1.3.   HTVS Analysis

A graphical representation of the docking scores obtained from HTVS is given in figure 2. The compounds were filtered considering -8 kcal/mol as the cut off value and the intermolecular interactions of these complexes were examined. The compounds were then filtered again based on receptor-ligand binding complementarity, standard inhibitors and the compounds that are the subject of ongoing study/trials. Upon thorough analysis, the best five hit compounds were considered for further analysis.

Figure 2. A graphical representation of the docking scores (Kcal/mol) of the ligands docked to A. capsid protein are represented in orange, B. membrane protein in red, C. envelope protein in purple, D. NS2B/NS3 protein in green, E. NS5 protein in blue and F. NS1 protein in brown.

1.4.   Specific Molecular Docking Analysis

The docking scores of the top hits labelled as NP1-NP5 docked to respective proteins are represented in figure 3. Table 2 summarizes the molecular docking details of these docked complexes.

Figure 3. A graphical representation of the binding energy scores of the top five NP hits bound to the target proteins. Each bar of the graph represents colour coded compounds.

Ligands that bind firmly to target proteins are selected as lead compounds. Interactions include hydrophobic contacts, hydrogen bonds, π-stacking, salt bridges, amide stacking, and cation–π interactions. The ligand interaction diagram shows different coloured binding site residues of the proteins (Figure 4 (A-F)).

Table 2. Specific molecular docking details of the compounds, NP1-NP5, relative to a standard antiviral DENV compound.

Figure 4. Two-dimensional protein-ligand interactions of compounds, NP1-NP5 bound to the DENV proteins in comparison with the standard inhibitor. Green coloured discs represent hydrophobic residues, cyan denotes polar residue, blue is positively charged residue and red is negatively charged. The grey coloured atom background denotes per-atom solvent accessible surface area (SASA).

Based on the above data, for the capsid protein, NP4 exhibited the strongest binding affinity, outperforming the standard, while NP5 formed 4 hydrogen bonds, suggesting a more stable interaction. For the membrane protein, NP5 emerged as the top-scoring phytocompound, forming 4 H-bonds and interacting via 1 salt bridge and 1 π-π stacking. NP4 also showed strong binding (-10.04 kcal/mol) with rich interactions including 1 salt bridge, 1 π-π stacking, and 2 π-cation bonds. Both compounds approached the standard score (-11.01 kcal/mol), indicating significant inhibitory potential. For the envelope protein, NP4 showed the least affinity while NP5 shared 5 H-bonds and a salt bridge, indicating these two compounds could effectively disrupt envelope protein-mediated viral entry. For NS2B/NS3 Protease, NP4 was the most favourable compound, better than the standard. For the NS5 protein, although NP2 had the highest score, NP4 showed s strong docking affinity with 2 salt bridges and NP5 showed 4 H-bonds indicating effective target engagement. NP5 demonstrated excellent binding, forming the highest number of hydrogen bonds along with 2 π-cation and 1 π-π stacking interactions, closely matching the standard (–9.9 kcal/mol). NP4 also showed favourable results with 2 hydrogen bonds, 1 salt bridge, and 1 π-π stacking, supporting their potential in targeting NS1-associated immune evasion. Across all six dengue virus targets, NP4 and NP5 consistently showed high docking scores and key molecular interactions, sometimes outperforming the standard drug. Their multi-target potential highlights them as lead candidates for further anlaysis and in vitro antiviral assays.

1.5.   MM/GBSA Analysis

A greater negative value denotes stronger binding as the MM-GBSA binding energies are approximations of the free energies of binding. It was also possible to determine the importance of the dGBind Coulomb (Coulomb energy), dGBind Hbond (Hydrogen bonding correction), dGBind Lipo (Lipophilicity energy), and dGBind vdW (Van der Waals energy). The information of the top hit compounds considered for further analysis is summarized in table 3.

Figure 5. Assessment of the binding free energy (orange bars) and other relative free energies (kcal/mol) such as coulomb energy (red), hydrogen bonding energy (dark blue), lipophilicity energy (dark green) and Van der Waals energy (dark purple) obtained by Prime MM-GBSA for the complexes of DENV target proteins to NP1-NP5 in comparison with the standard.

The MM/GBSA binding free energy analysis across six key dengue virus proteins revealed that compounds NP4 and NP5 consistently exhibited the most favourable binding affinities (Figure 5). Their interactions were predominantly stabilized by the other interactions, indicating strong and specific binding within the active sites of the viral proteins

Table 3. Information of the top Natural compounds screened for further inter-molecular interactions.