Virtual screening and molecular dynamics studies of antivenom using flavonoid compounds derived from Andrographis echioides

Abstract

The antivenom potential of Andrographis echioides (A. echioides), a traditional medicinal herb, was investigated using a comprehensive in silico approach to address the limitations of conventional antivenoms against Russell’s viper envenomation. This study aimed to conduct virtual screening and molecular dynamics (MD) studies of flavonoid compounds from A. echioides to explore their potential as antivenom agents. Gas chromatography-mass spectrometry (GC–MS) analysis identified 84 bioactive phytochemicals in the leaf extracts. The selected flavonoid constituents were evaluated for their predicted inhibitory potential against venom-associated enzymes using molecular docking with the iGEMDOCK software. The target proteins included metalloproteinase (PDB ID: 2E3X), serine proteinase (PDB ID: 3S9A), and human pancreatic phospholipase A2 (PDB ID: 6Q42). Among the screened compounds, decanoic acid (− 93.7429 kcal/mol), oxalic acid 6-ethyloct-3-yl isohexyl ester (− 91.6448 kcal/mol), and oxalic acid 6-ethyloct-3-yl hexyl ester (− 85.7934 kcal/mol) exhibited the highest binding affinities to the target protein and compared them with the standard compound. Drug-likeness analysis confirmed the favorable pharmacokinetic properties of compounds. MD simulations spanning 100 ns revealed stable binding interactions, consistent RMSD values, favorable hydrogen bonding patterns, and stable structural dynamics. This is the first study to report the antivenom potential of bioactive compounds from this species using computational methods. These computational insights suggest that flavonoid constituents derived from A. echioides may possess promising binding and predicted inhibitory activities against key venom enzymes, warranting further investigation as potential natural antivenom agents.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-27737-9.

Introduction

Russell’s viper (Daboia russelii) is one of the most medically significant venomous snakes in Southeast Asia, and envenomation often results in serious complications owing to the highly complex nature of its venom¹. Phospholipase A2 (PLA2), a major constituent of snake venom, plays a key role in hemolysis, neurotoxicity, myotoxicity, and edema². The high proteomic variability of venom across geographical regions complicates the development of universally effective antivenoms³. Conventional antivenoms are often limited in their ability to neutralize low-molecular-weight toxins such as PLA2, underscoring the need for alternative therapeutic strategies⁴. Snakebite envenomation is a major global health issue, especially in tropical and subtropical areas, affecting millions of people each year. The World Health Organization (WHO) classifies it as a neglected tropical disease, underscoring the lack of attention and resources dedicated to addressing its impact on public health. This condition often results in severe health outcomes, such as permanent disabilities or fatalities, primarily affecting underserved populations with restricted access to medical care. The WHO’s 2023 classification highlights the urgent need for innovative solutions, such as next-generation toxin-specific antivenoms, to address this burden and improve outcomes for those affected [50]. Despite significant advances in antivenom development, current therapies face critical limitations, including species-specificity, geographical variation in venom composition, high production costs, and limited accessibility in rural areas where snakebites are most prevalent. The need for broad-spectrum, cost-effective alternatives has never been more urgent, particularly given the WHO’s recognition of snakebite envenomation as a priority neglected tropical disease affecting over 5 million people annually.

Many plants have traditionally been used for their antibacterial, antibiotic, and antifungal properties, serving as natural remedies before the advent of modern medicine^5,6. Plants of the genus Andrographis, particularly Andrographis paniculata and A. serpyllifolia, have demonstrated therapeutic value^7,8, however, the antivenom potential of A. echioides has not yet been explored. A. echioides (L.) Nees is a well-known medicinal herb in South Asia, commonly found in the plains and drylands of India, Sri Lanka and China⁹. Traditionally, it has been used to treat ailments such as scorpion and snake bites¹⁰. Flavonoids and isoflavones are crucial bioactive compounds found in various plants and are known for their potential health benefits, including anti-inflammatory, antioxidant, and anticancer activities. Accurate analysis and identification of these compounds in biological fluids are essential for research and medical diagnosis. Although aromatic extracts, such as essential oils, are commonly analyzed, non-aromatic extracts, such as ethanolic or aqueous extracts containing flavonoids and isoflavones, require careful characterization using advanced analytical techniques. One common method for such analyses is GC–MS, which often requires the derivatization of flavonoids and isoflavones to enhance their volatility and detectability. This process involves modifying these compounds with specific reagents, allowing for more precise and efficient analysis. Accurately determining the structure and concentration of flavonoids and isoflavones in both aromatic and non-aromatic extracts aids in understanding their roles in human health and contributes to the development of nutraceuticals and pharmaceutical products^11,12.

In this study, we investigated flavonoids from A. echioides using a comprehensive in silico approach, including GC–MS profiling, molecular docking, ADMET analysis, and molecular dynamics simulations, to identify potential candidates that can target key venom proteins such as metalloproteinases, serine proteinases, and phospholipase A2 (PLA2)¹³. GC–MS detected phytoconstituents in the leaf extract of A. echioides and their binding energies with venom enzyme targets (2E3X, 3S9A, and 6Q42)^2,14. The structural variations among these enzymes, especially when comparing human isoforms to their venom counterparts, offer unique opportunities to design selective inhibitors that can mitigate inflammatory diseases without affecting essential physiological functions. By examining the structural differences and active site configurations between these human enzymes and their venomous analogs, researchers can develop targeted pharmaceuticals that exploit these distinctions, potentially offering new treatments for conditions such as rheumatoid arthritis, asthma, and cardiovascular disease. A comparative study of these enzymes can also provide insights into evolutionary biology and biochemical adaptation strategies in venomous species. The compound 12-methoxy-4-methyl-voachalotine had been previously identified as a bioactive constituent with demonstrated binding affinity to venom proteins, making it suitable as a reference standard for comparative analysis in this study^15,16.

Traditional antivenom discovery methods, which are often labor-intensive and time-consuming, can be significantly accelerated by these computational approaches¹⁷. Virtual screening allows researchers to sift through vast libraries of chemical compounds to identify those that interact with venomous toxins, thereby narrowing down the candidates for further investigation. Virtual screening and molecular dynamics studies have become vital tools in the search for potential compounds that can serve as effective antivenoms for treating envenomation¹⁸. Molecular dynamics simulations provide detailed insights into the interactions between these compounds and venom proteins at the atomic level, aiding in the prediction of their efficacy and stability in venom neutralization. These techniques not only enhance the efficiency of antivenom development but also reduce the reliance on animal testing, aligning with ethical considerations in scientific research. Virtual screening and computational approaches have been extensively employed in drug discovery, with several studies demonstrating their effectiveness in identifying potential therapeutic compounds. Previous computational studies have successfully utilized similar methodologies for various targets, including cardiotonic steroids as Na/K-ATPase inhibitors¹⁹, pharmacophore-based screening for catechol-O-methyltransferase inhibitors in Alzheimer’s disease²⁰, and the identification of antiviral phytochemicals against the SARS-CoV-2 main protease²¹. Building on our previous work, in which we successfully identified the anti-Alzheimer and antimalarial properties of plants (Schleichera oleosa and Cardiospermum halicacabum through docking and dynamic simulations²², we investigated the efficacy of these plants in neutralizing the venom toxins. While previous studies have explored individual Andrographis species for various therapeutic applications, a comprehensive computational evaluation of A. echioides flavonoids, specifically targeting multiple venom enzymes simultaneously, remains unexplored. The aim of this study was to evaluate the predicted anti-venom properties of these phytochemicals using in silico approaches to identify potential lead molecules for further experimental validation. This study addresses three critical research gaps: (1) the lack of systematic screening of A. echioides phytochemicals against venom targets, (2) the absence of integrated ADMET profiling for plant-derived antivenom candidates, and (3) limited molecular dynamics validation of the binding stability of natural compounds against venom enzymes. Our multi-target approach provides a novel framework for identifying broad-spectrum antivenom candidates from traditional medicinal-plant sources.

Materials and methods

Plant identification and extraction

Fresh A. echioides leaves were collected from the Tiruchengode Hills, Namakkal District, Tamil Nadu, and authenticated by Dr. P. Radha, taxonomist. A voucher specimen (A080724133E) was deposited at the Siddha Medicinal Plants Garden (SMPG) in Mettur, Tamil Nadu. Dried leaves (10 g) were powdered and extracted with 20 mL of 99.99% ethanol in an ultrasonic bath KLDUC-5L (Kinglab Instruments, Tamil Nadu, India) for 15 min.

Preparative TLC

Preparative TLC was performed using silica gel plates, as described by Richardson and Harborne²³. Test samples (2 mg/mL) and quercetin reference (1 mg/mL) were used in the experiment. Optimal separation was achieved using a mobile phase comprising n-butanol, acetic acid, and water in a 2:2:6 ratio. The developed plates were observed under UV light at 254 and 366 nm. Flavonoid bands were excised using surgical blades, dissolved in ethanol, filtered through Whatman filter paper, and used for subsequent analysis.

GC–MS analysis

The extracted compounds were analyzed using a Shimadzu GCMS-QP2010 Plus system under standard operating conditions. A 2 µL sample was injected in split mode (1:3 ratio) with helium as the carrier gas. The oven was initially set at 50 °C for 1 min, ramped to 300 °C, and held for 10 min. The compounds were identified by comparing their retention times and mass fragmentation data with those in the NIST database.

Hardware specifications

The computational work was performed using a system equipped with an AMD Ryzen 5 5500U processor integrated with Radeon graphics on a Windows 11 operating system.

Software specifications

Experimental optimization was performed using Design Expert 12, employing the Box–Behnken design within the Response Surface Methodology (RSM) framework. Ligand structures were created using ChemDraw 16.0 software. Molecular docking and interaction analyses were performed using iGEMDOCK V2.1 and Discovery Studio Visualizer V20. Intermolecular interactions were visualized using LigPlot + V2.1. The protein–ligand complexes were prepared using the CHARMM-GUI interface. MD simulations were performed using VMD and NAMD software packages, which provide robust molecular dynamics simulation capabilities for protein–ligand complex analysis.

Target preparation

Three-dimensional protein structures were obtained from the RCSB Protein Data Bank, a comprehensive resource for structural data on biological macromolecules. The selected proteins were prepared using X-ray crystal structures (Fig. 1) with PDB IDs 2E3X, 3S9A, and 6Q42, corresponding to Russell’s viper venom metalloproteinase²⁴, serine proteinase¹³, and human pancreatic phospholipase A2².

Fig. 1.

GC–MS chromatogram of the leaf extract of Andrographis echioides.

Ligand preparation

Ligands identified via GC–MS and the standard compound 12-methoxy-4-methyl-voachalotine were saved in PDB format. Energy minimization was performed using Chem3D 16.0 with the MM2 force field, which optimizes the bond lengths and angles to achieve a more realistic ligand conformation than the MMFF. Minimized ligands were imported into the docking workspace to ensure optimal spatial orientation and binding compatibility with the target protein²⁵.

Molecular docking analysis

Target proteins were prepared using the Discovery Studio Visualizer by eliminating water molecules and unnecessary ligands, followed by the addition of hydrogen atoms, as described by Sharma et al.²⁶. A total of 84 phytocompounds identified in the A. echioides extract via GC–MS were subjected to molecular docking against the protein targets 2E3X and 3S9A. Docking was performed using the default settings of the software, which included a population size of 200 and 70 generations and two predicted binding solutions per ligand. Ligand selection was based on the most favorable conformations and lowest binding-free energy values. The docking software generated comprehensive interaction profiles, highlighting the electrostatic interactions, hydrogen bonding, and van der Waals forces. Binding affinities are expressed in kcal/mol, with the compound exhibiting the lowest energy being considered the most promising inhibitor of the target protein. Compared to other docking tools, it demonstrated greater efficiency and accuracy in predicting ligand–protein interactions.

In silico drug likeness and ADME toxicity

The bioactive constituents identified from A. echioides were assessed for drug-likeness via the SCFBIO web server https://www.scfbio-iitd.res.in, adhering to Lipinski’s Rule of Five: molecular weight ≤ 500 Da, LogP ≤ 5, a maximum of 5 hydrogen bond donors, no more than 10 hydrogen bond acceptors, and a molar refractivity range of 40–130, as described by Hartati et al.²⁷. Compounds meeting these criteria were subjected to further ADME-Tox evaluation using the ADMET-AI platform (admet.ai.greenstonebio.com), following the approach outlined by Swanson et al.²⁸. This online tool enables the efficient prediction of pharmacokinetic and toxicity parameters without requiring software installation or extensive computational expertise. The evaluated parameters included blood–brain barrier permeability, interactions with cytochrome P450 enzymes (CYP1A2, CYP2C19, and CYP2C9), and mutagenic potential, providing a holistic assessment of the viability of the compounds for drug development.

Binding site prediction

BIOVIA Discovery Studio (BDS) offers advanced tools for visualizing molecular structures and simulation results in 3D²⁹. It enables a detailed representation of protein–ligand interactions, including hydrogen bonding and hydrophobic contacts, which are crucial for complex stability and binding affinity³⁰. Additional features include electrostatic potential mapping, surface and volume rendering, and customizable visualization options such as color, transparency, and labeling. These visual outputs can be exported for publication or presentation.

LigPlot analysis

LigPlot was used to generate schematic two-dimensional diagrams of protein–ligand complexes from the standard PDB files. These black-and-white representations highlight the key intermolecular interactions, including hydrogen bonds and hydrophobic contacts, along with their strengths³¹. LigPlot is broadly applicable to various ligands and can depict the interactions between proteins and nucleic acids. A detailed analysis of the interaction patterns between the docked ligands and active site residues was conducted as described by Behera et al.³².

Molecular dynamic simulation

The top-docked conformations of the selected ligands (decanic acid, oxalic acid, dodecyl 3, 5-difluorophenyl ester, and oxalic acid 6-ethyloct-3-yl isobutyl ester) and the reference compound (12-methoxy-4-methyl-voachalotine) were subjected to MD simulations to assess their conformational stability and inhibitory efficacy^33,34. MD simulations were performed using NAMD software, with ligand topology files generated using the CHARMM-GUI interface. The systems were solvated in water, neutralized with K+ and Cl−ions, and prepared using CHARMM-GUI Membrane Builder protocols. These included energy minimization via the steepest descent algorithm (1000 steps), followed by equilibration phases under NVT and NPT ensembles (each for 100 ps), and concluded with a 50 ns production simulation^35,36. The binding energies of both the ligand and standard complexes were calculated to provide insights into their molecular interactions and dynamic stability^37,38. This simulation approach complemented the docking results and allowed for a deeper investigation of ligand-receptor binding mechanisms and their structural dynamics.

Result

To evaluate the antivenom potential of A. echioides flavonoids against Russell’s viper envenomation, we employed a systematic computational approach encompassing phytochemical identification, molecular docking, drug-likeness assessment, and molecular dynamics validation. The following results demonstrate the progression from compound identification to the selection of promising antivenom candidates.

Extraction and isolation

After extraction, flavonoids were spotted onto silica gel or alumina-coated thin-layer chromatography (TLC) plates and developed in a solvent system. TLC separation of the A. echioides extract using silica gel plates in a solvent system of n-butanol, acetic acid, and water in 2:2:6, successfully resolved multiple flavonoid compounds, which appeared as distinct yellow spots under UV light visualization. The separated fractions were recovered and re-dissolved in ethanol for subsequent GC–MS analysis.

GC–MS analysis

GC–MS analysis of the A. echioides leaf extract identified 84 phytoconstituents. This identification was based on a comparison of the fragmentation patterns with reference spectra from the NIST library, allowing for both qualitative and quantitative evaluations. The resulting GC–MS chromatogram (Fig. 1) and the corresponding list of identified compounds arranged by their retention times (RT) are presented in Table 1.

Table 1.

GC–MS detected phytoconstituents in the leaf extract of A. echioides and their binding energy with PLA2 target.

Peak#	R.Time	Area %	Name	Binding energy (kcal/mol)
				2e3x	3s9a	6q42
1.	22.868	0.21	1,2-benzenedicarboxylic acid, diisononyl ester	− 87.0826	− 82.1392	− 81.1611
2.	22.537	0.11	1,3-Propanediol, dodecyl ethyl ether	− 72.9005	− 74.4812	− 75.2654
3.	26.450	0.02	1-bromotriacontane	− 83.8667	− 81.2446	− 69.2657
4.	20.730	0.78	1-Decanol, 2-hexyl-	− 76.8294	− 77.4501	− 86.0679
5.	8.846	0.02	1-Heptanol, 2,4-dimethyl-,	− 59.7913	− 61.7474	− 62.9025
6.	8.761	0.06	1-Hexene, 3,5,5-trimethyl-	− 49.6009	− 48.9835	− 49.3168
7.	13.492	0.06	1-tridecanol	− 73.5048	− 68.4628	− 64.5114
8.	19.153	0.30	2-(4-tert-butylbenzyl)-3-methylbutanal	− 75.2338	− 73.3224	− 68.2696
9.	9.489	0.16	2,3,6,7-tetramethyloctane	− 60.4063	− 58.9925	− 58.014
10.	23.542	0.83	2,5-cyclohexadien-1-one, 2,6-bis(1,1-dimethylethyl) -4-hydroperoxy-4-methyl-	− 79.3982	− 68.2236	− 80.6671
11.	9.215–21.750	1.61	2,6,10-trimethyltridecane	− 65.7714	− 63.11	− 66.6032
12.	8.948	0.01	2-Ethylbutyl isobutyl carbonate	− 89.8255	− 73.6279	− 74.6907
13.	4.298	0.40	2-hydroxy-2-methyl-4-pentanone (diacetone)	− 55.4483	− 57.3911	− 58.5601
14.	11.775	0.01	2-Hydroxyethyl 2,2,2-trifluoroacetate	− 71.3553	− 63.7781	− 69.7239
15.	27.693–30.027	0.14	2-methylhexacosane	− 82.5095	− 76.4913	− 76.0289
16.	21.990	0.20	2-methyltetracosane	− 80.8498	− 73.3732	− 78.6036
17.	8.622	0.02	2-undecen, 4,5-dimethyl-, trans-, erythro-	− 60.0631	− 63.3447	− 62.6874
18.	11.006	0.05	3,4-dimethylcyclohexan-1-ol	− 59.4167	− 57.4452	− 57.8089
19.	11.305–12.151	0.05	3-Hexanone, 2,5-dimethyl-	− 51.3785	− 53.8775	− 58.9425
20.	27.096	0.10	5,5-diethylpentadecane	− 72.7123	− 82.8019	− 74.5073
21.	32.281	0.11	6,6-diethylhoctadecane	− 80.1472	− 75.5752	− 70.2539
22.	18.390	0.18	7-tetradecanol	− 81.2257	− 66.4139	− 79.925
23.	13.327	0.18	11 -methyldodecanol	− 70.2122	− 68.1439	− 65.1344
24.	27.279–29.514	0.34	11 -methyltricosane	− 84.7191	− 80.4095	− 78.6911
25.	19.534	0.49	18-Methyl-nonadecane-1,2-dio, trimethylsilyl ether	− 73.8273	− 69.5519	− 68.221
26.	18.740	0.14	Allyl n-octyl ether	− 64.2714	− 59.4491	− 64.5095
27.	19.282	0.37	Bis(2-ethylhexyl) ether	− 75.0649	− 68.1038	− 81.9889
28.	6.578	0.01	Butane, 2,2-dimethyl-	− 42.4797	− 39.9231	− 40.5199
29.	25.380	1.13	Carbonic acid, (1R)-(-)-menthyl octyl ester	− 75.6659	− 78.8821	− 80.6394
30.	15.987–29.839	1.3	Carbonic acid, eicosyl vinyl ester	− 85.0636	− 82.9729	− 81.2424
31.	22.341–23.170	0.22	Carbonic acid, hexadecyl prop-1-en-2-yl ester	− 85.9619	− 81.9958	− 77.6976
32.	18.632	0.36	Carbonic acid, prop-1-en-2-yl tridecyl ester	− 81.0046	− 73.6088	− 68.3346
33.	30.415	0.84	Cis-2-phenyl-1, 3-dioxolane-4-methyl octadec-9, 12, 15-trienoate	− 75.5655	− 67.8214	− 83.5473
34.	7.198	0.01	Cyclopentanol, 3-methyl-	− 57.9854	− 51.976	− 56.411
35.	14.714	18.2	Decanoic acid	− 93.7429	− 81.4932	− 82.302
36.	19.081	0.30	Docosane, 1-iodo-	− 70.757	− 71.6244	− 63.6955
37.	12.700–13.585	0.36	Dodecane, 1-iodo-	− 58.7806	− 60.0751	− 63.9922
38.	12.972–22.935	1.07	Dodecane, 2,6,10-trimethyl-	− 66.5538	− 63.8595	− 72.645
39.	8.399	0.38	Dodecane, 2,6,11-trimethyl-	− 67.3888	− 64.5738	− 66.159
40.	9.335–18.330	0.59	Dodecane, 4,6-dimethyl-	− 69.1014	− 65.8595	− 67.3178
41.	22.221–27.962	0.19	Dodecyl octyl ether	− 84.2827	− 70.6265	− 66.6359
42.	20.268–27.178	10.8	Dotriacontane	− 87.8307	− 82.0373	− 69.405
43.	24.322–24.400	0.63	Dotriacontane, 1-iodo-	− 81.1009	− 81.9009	− 69.5234
44.	17.700–25.183	4.9	Eicosane	− 75.2369	− 68.7862	− 63.5134
45.	32.469	0.20	Eicosyl isopropyl ether	− 79.2821	− 75.7957	− 69.3181
46.	28.257–30.720	1.74	Glycidyl palmitate	− 78.9345	− 77.2523	− 95.32
47.	16.926–17.857	3.08	Heneicosane	− 76.6316	− 69.9091	− 66.9971
48.	12.805–19.776	3.65	Heptadecane, 2,6,10,15-tetramethyl-	− 78.0261	− 68.9229	− 95.9702
49.	27.799	0.08	Heptadecane, 8-methyl-	− 77.3867	− 77.0576	− 79.2122
50.	11.896	0.04	Heptane, 3,3,5-trimethyl-	− 58.8995	− 54.7226	− 57.6632
51.	31.973–32.775	0.17	Hexacontane	− 97.8277	10.7809	− 46.7831
52.	23.827	0.31	Hexacosyl nonyl ether	− 92.3987	− 84.7914	− 70.5577
53.	16.766–24.771	9.59	Hexadecane, 2,6,10,14-tetramethyl-	− 68.7285	− 65.6759	− 84.9595
54.	22.125	0.05	Hexane, 2,2,3,3-tetramethyl-	− 51.5667	− 53.8786	− 50.5051
55.	4.486	0.19	Hexane, 2,3,4-trimethyl-	− 48.2947	− 50.0765	− 46.9798
56.	3.824	0.05	Hexane, 2,4-dimethyl-	− 48.975	− 46.3225	− 45.3047
57.	26.052	0.10	Isopropyl hexacosyl ether	− 88.1608	− 83.4349	− 74.7061
58.	8.282	0.88	Nonane, 3-methyl-5-propyl-	− 56.5065	− 60.7044	− 67.467
59.	8.142	0.08	Nonane, 5-(2-methylpropyl)-	− 59.4563	− 62.6526	− 68.5533
60.	9.077	0.12	Nonane, 5-butyl-	− 65.9401	− 59.1327	− 69.9444
61.	21.696	0.21	Nonane, 5-methyl-5-propyl-	− 65.6449	− 61.9408	− 64.1646
62.	20.400–23.887	1.7	Octacosane, 1-iodo-	− 84.7095	− 78.8952	− 67.8784
63.	32.694	0.01	Octadecane, 1,1'-[1,3-propanediylbis(oxy)]bis-	− 99.3495	− 81.7921	− 74.9613
64.	20.049	2.02	Octadecanoic acid	− 82.3143	− 75.2985	− 62.5882
65.	16.162–25.849	1.44	Octane, 1,1'-oxybis-	− 66.7207	− 67.6552	− 58.7586
66.	17.253	0.62	Octyl tetradecyl ether	− 85.3473	− 73.6028	− 69.2719
67.	15.080	3.41	Oxalic acid, 2-ethylhexyl hexyl ester	− 84.7603	− 76.4753	− 79.9839
68.	16.380	0.54	Oxalic acid, 2-ethylhexyl isohexyl ester	− 87.1483	− 90.3698	− 84.0206
69.	25.950	0.01	Oxalic acid, 6-ethyloct-3-yl hexyl ester	− 85.7429	− 79.8489	− 85.7934
70.	23.686	0.10	Oxalic acid, 6-ethyloct-3-yl isobutyl ester	− 76.2763	− 81.6475	− 74.0483
71.	22.655	0.02	Oxalic acid, 6-ethyloct-3-yl isohexyl ester	− 79.8832	− 91.6448	− 83.2526
72.	32.176	0.20	Oxalic acid, dodecyl 3,5-difluorophenyl ester	− 82.5002	− 75.4576	− 76.9186
73.	21.840	0.15	Pentadecane, 2,6,10,14-tetramethyl-	− 78.5199	− 69.6418	− 87.0157
74.	31.101	2.83	Phthalic acid, di(2-propylpentyl) ester	− 86.5276	− 88.8146	− 66.3426
75.	12.234	0.02	Prop-2-yn-1-yl 2-methylbutanoate	− 58.9451	− 60.2109	− 57.5712
76.	17.161	0.60	Silane, trichlorooctadecyl-	− 75.7708	− 68.6194	− 69.057
77.	19.704	0.20	Sulfurous acid, 2-ethylhexyl isohexyl ester	− 70.0956	− 75.9197	− 73.986
78.	16.280	0.43	Sulfurous acid, butyl decyl ester	− 75.2366	− 74.0095	− 73.4587
79.	18.502	0.39	Sulfurous acid, decyl 2-ethylhexyl ester	− 86.8329	− 75.613	− 82.2768
80.	15.338	5.26	Tetradecanal	− 69.1835	− 69.955	− 69.4138
81.	24.082–32.917	7.43	Tetrapentacontane	− 97.6213	337.861	− 16.2409
82.	18.850	0.32	Trans-2-Undecen-1-ol	− 63.3504	− 67.2465	− 72.709
83.	18.143–23.747	0.32	Tridecanol, 2-ethyl-2-methyl-	− 75.3846	− 69.7955	− 76.6171
84.	30.270	2.99	Trilinolein	− 115.941	− 107.476	− 111.351
	-	-	Standard 12-methoxy-4-methyl-voachalotine	− 86.1908	− 83.729	− 81.5425

Bold values indicate the binding energy of the standard compound, 12-methoxy-4-methyl-voachalotine.

Ligand preparation

The two-dimensional structure depicted in Fig. 2, created using Chem3D, represents the molecular configuration of the ligands extracted from A. echioides leaves.

Fig. 2.

Two dimensional chemical structures of major bioactive compounds from Andrographis echioides.

Docking analysis

This study identified the most promising drugs by analyzing their interactions with anti-venom targets. Prospective lead compounds were selected based on their strong receptor-binding affinity. iGEMDOCK evaluations indicate that phytocompounds from the leaves extract of A. echioides possess anti-venom potential. Table 1 displays the binding energies between the ligand interactions identified by GC–MS and selected venom targets. A comprehensive docking score summary for the top-performing ligands across all target proteins is presented in Table 2.

Table 2.

Comprehensive docking score summary for top-performing ligands across all target proteins.

Ligand no	Best target	Compound name	Binding energy (Kcal/mol)
			2E3X	3S9A	6Q42
84	2E3X, 3S9A, 6Q42	Trilinolein	− 115.94	− 107.48	− 111.35
63	2E3X	Octadecane, 1,1'-[1,3-propanediylbis(oxy)]bis-	− 99.35	− 81.79	− 74.96
51	2E3X	Hexacontane	− 97.83	− 10.78	− 46.78
81	2E3X	Tetrapentacontane	− 97.62	− 33.79	− 16.24
35	2E3X	Decanoic acid	− 93.74	− 81.49	− 82.30
48	3S9A, 6Q42	Heptadecane, 2,6,10,15- tetramethyl-	− 78.03	− 68.92	− 95.97
46	3S9A, 6Q42	Glycidyl palmitate	− 78.93	− 77.25	− 95.32
71	3S9A	Oxalic acid, 6-ethyloct-3-yl isohexyl ester	− 79.88	− 91.64	− 83.25
68	3S9A	Oxalic acid, 2-ethylhexyl isohexyl ester	− 87.15	− 90.37	− 84.02
69	6Q42	Oxalic acid, 6-ethyloct-3-yl hexyl ester	− 85.74	− 79.85	− 85.79
53	6Q42	Hexadecane, 2,6,10,14-tetramethyl-	− 68.72	− 65.67	− 84.95
Standard	2E3X, 3S9A, 6Q42	12-methoxy- 4-methyl-voachalotine	− 86.19	− 83.73	− 81.54

Top five ligands against 2E3X – venom metalloproteinase

Thirteen compounds were predicted to exhibit stronger binding affinities to 2E3X than the standard. The top five scoring ligands were trilinolein (– 115.94 kcal/mol), octadecane, 1,1'-[1,3-propanediylbis(oxy)]bis- (– 99.35 kcal/mol), hexacontane (– 97.83 kcal/mol), tetrapentacontane (– 97.62 kcal/mol), decanoic acid (– 93.74 kcal/mol) and standard 12-methoxy-4-methyl-voachalotine (− 86.1908 kcal/mol).

Top five ligands against 3S9A – venom serine proteinase

Five compounds were predicted to exhibit stronger binding affinities to 3S9A than the standard. The top five ligands such as trilinolein (– 107.48 kcal/mol), oxalic acid, 6-ethyloct-3-yl isohexyl ester (–91.64 kcal/mol), oxalic acid, 2-ethylhexyl isohexyl ester (–  90.37 kcal/mol), phthalic acid, di(2-propylpentyl) ester (– 88.81 kcal/mol), hexacosyl nonyl ether (– 84.79 kcal/mol) and standard 12-methoxy-4-methyl-voachalotine (− 83.729 kcal/mol).

Top five ligands against 6Q42 – phospholipase A2 (PLA2)

Thirteen compounds were predicted to exhibit stronger binding affinities to 2E3X than the standard. These were trilinolein (– 111.35 kcal/mol), heptadecane, 2,6,10,15-tetramethyl- (– 95.97 kcal/mol), glycidyl palmitate (– 95.32 kcal/mol), pentadecane, 2,6,10,14-tetramethyl- (– 87.02 kcal/mol), 1-decanol, 2-hexyl- (– 86.07 kcal/mol) and standard 12-methoxy-4-methyl-voachalotine (− 81.5425 kcal/mol).

In silico drug likeness and ADME toxicity

Based on the docking scores, the top five ligands were screened for drug-likeness (Table 3) and ADME toxicity profiles (Table 4). After applying Lipinski’s Rule of Five filters, only ligands 35 (2E3X), 71 and 68 (3S9A), and 69 (6Q42) satisfied all drug-likeness parameters, making them the most promising candidates for further development.

Table 3.

Drug likeness properties of the selected top five ligands from A. echiodies.

Ligand No	Dock score rank	Drug likeness
		MR	MW	HBD	HBA	LogP	RO5 (%)
		40–130	˂500	˂5	˂10	< 5
Top ligands for 2E3X
84	1	269.87	878	0	6	17.42	40
63	2	185.34	580	0	2	13.93	40
51	3	279.13	842	0	0	23.65	40
81	4	251.43	758	0	0	21.31	40
35	5	79.365	286	2	4	3.86	100
Top ligands for 3S9A
84	1	269.87	878	0	6	17.42	40
48	2	98.79	296	0	0	7.86	80
46	3	91.01	312	0	3	5.40	80
71	4	88.507	314	2	4	4.50	100
68	5	79.295	286	2	4	3.72	100
Top ligands for 6Q42
84	1	269.87	878	0	6	17.42	40
48	2	98.79	296	0	0	7.86	80
46	3	91.01	312	0	3	5.40	80
69	4	88.577	314	2	4	4.64	100
53	5	94.17	282	0	0	7.47	80
Standard 12-methoxy-4-methyl-voachalotine
STD	-	112.45	411	1	5	2.07	100

Bold values indicate the phytoconstituents from A. echioides that exhibit 100% drug-likeness.

Table 4.

ADMET profile lead hit obtained form the leaves extract of A. echiodies.

Ligand no	Target protein	BBB	CYP1A2 inhibitor	CYP2C19 inhibitor	CYP2C9 inhibitor	CYP2C9 substrate	Mutagenicity
35	2E3X	0.95	0.20	0.16	0.12	0.19	0.07
71	3S9A	0.90	0.04	0.10	0.08	0.22	0.19
68	3S9A	0.96	0.10	0.17	0.11	0.18	0.11
69	6Q42	0.91	0.07	0.09	0.08	0.18	0.12
STD	Standard	0.41	0.004	0.01	0.001	0.02	0.24

ADME toxicity analysis of the ligands with 100% drug likeness revealed that ligands 35, 71, 68, and 69 demonstrated favorable pharmacokinetic and safety profiles compared to the standard compound. Ligand 35 has emerged as a safer and more CNS-permeable alternative to the standard, with excellent BBB penetration and low mutagenicity, although optimization is required to reduce CYP inhibition liability and to improve pharmacokinetics. Similarly, ligands 71 and 68, both targeting 3S9A, demonstrated stronger BBB permeability and lower mutagenic risk than the standard. However, ligand 68 exhibited greater CYP inhibition than ligand 71, suggesting a higher risk of metabolic interaction. Ligand 69, targeting 6Q42, also shows favorable BBB permeability and reduced mutagenicity, but its moderately elevated CYP inhibition and substrate potential indicate a need for further metabolic assessment.

Prediction of binding sites

Supplementary Figures S1–S3 present the hierarchical clustering analysis from iGEMDOCK post-screening, illustrating the interaction patterns between venom targets and phytoconstituents derived from A. echioides extract, alongside standard compounds. Figures 3, 4 and 5 depict the chain representations and 2D interaction diagrams of the lead compounds and standards with the protein structures 2E3X, 3S9A, and 6Q42. The identified binding sites on these protein receptors, characterized by the lowest binding energies, are indicative of potential active sites where ligand binding is most favorable.

Fig. 3.

A 2D and 3D representations of the protein–ligand docking: (A and B) 2E3X with ligand 35, (C and D) 2E3X with standard.

Fig. 4.

A 2D and 3D representations of the protein–ligand docking: (A and B) 3S9A with ligand 71, (C and D) 3S9A with standard.

Fig. 5.

A 2D and 3D representations of the protein–ligand docking: (A and B) 6Q42 with ligand 69, (C and D) 6Q42 with standard.

LigPlot analysis

LigPlot was employed to investigate the hydrogen bonding and hydrophobic interactions of selected compounds that exhibited favorable in silico performance compared to their respective standards (Table 5 and Fig. 6). For the 2E3X protein target, ligand 35 formed two hydrogen bonds with Arg272(A) and Cys291(A) at distances of 3.09 Å and 3.04 Å, respectively, along with 13 hydrophobic contacts. In contrast, the standard ligand (2E3X-std) formed a single, slightly stronger hydrogen bond with Ala122(A) at 2.84 Å and exhibited 11 hydrophobic interactions with the target protein. This suggests that 2E3X-lig 35 offers enhanced specificity through multiple polar interactions, whereas the standard benefits from denser hydrophobic interactions. Similarly, 3S9A-lig 71 formed a hydrogen bond with Thr94(A) at 3.06 Å and 10 hydrophobic contacts, indicating a balanced polar–nonpolar interaction profile, whereas 3S9A-std lacked hydrogen bonds but maintained eight hydrophobic contacts, suggesting less binding specificity. For 6Q42, ligand 69 engaged in 14 hydrophobic interactions without hydrogen bonding, indicating binding stabilization via nonpolar forces alone, whereas the standard formed 15 hydrophobic contacts and one hydrogen bond with Thr120(B) at 2.70 Å, reflecting a slightly stronger and more diverse interaction profile.

Table 5.

LigPlot analysis: protein–ligand interaction studies from A. echioides leaf extract and standard complex.

S. No	Protein complex	Hydrogen bonds	Distance (Å)	Total H-bond	Hydrophobic residues	Hydrophobic bonds	Total bonds
1	2E3X complex ligand 35	Cys291, Arg272	3.04, 3.09	2	Phe12(A), Glu219(A), Asp294(A), Leu206(A), Trp221(A), Arg293(A), Lys10(A), Pro204(A), Glu290(A), Leu257(A), Gly256(A)	11	13
	2E3X complex standard	Ala122	2.84	1	Gln321(A), Asp398(A), Pro399(A), Ile317(A), Asn320(A), Cys315(A), Pro316(A), Asp314(A), Tyr311(A), Gln121(A), Cys120(A)	11	12
2	3S9A complex ligand 71	Thr94	3.04, 3.09	2	Lys101(A), Phe90(A), Cys91(A), Gly98(A), Ala56(A), Asp59(A), Asn97(A), Pro96(A), Phe95A(A), Leu92(A)	10	11
	3S9A complex standard	-	-	-	Ala214(A), Tyr172(A), Val227(A), Gly215(A), Trp173(A), Asp102(A), Leu99(A), His57(A)	8	8
3	6Q42 complex ligand 69	-	-	-	Thr120(A), Leu118(A), Leu31(A), Asn24(A), Gly30(B), Leu31(B), Gly32(B), Tyr25(B), Cys29(B), Gly26(B), Asn24(B), Thr120(B), Cys27(B), Gly33(B)	14	14
	6Q42 complex standard	-	-	-	Gly30(A), Asn24(B), Phe19(B), Lys121(A), Leu31(A), Leu118(A), Asn24(A), Cys27(B), Asp119(A), Thr120(A), Asn23(B), Asn117(A), Leu31(B), Gly33(B), Gly30(B)	15	15

Fig. 6.

Ligplot + representation of the protein–ligand interactions: (A) 2E3X with ligand 35, (B) 2E3X with standard, (C) 3S9A with ligand 71, (D) 3S9A with standard, (E) 6Q42 with ligand 69, (F) 6Q42 with standard.

MD simulations

MD simulations conducted over 100 ns revealed stable protein–ligand interactions across all three complexes (2E3X, 3S9A, and 6Q42). The RMSD plots showed initial fluctuations followed by stabilization, indicating equilibrium and complex stability. The RMSF values confirmed that most protein residues exhibited minimal flexibility, with fluctuations in the loop region. The SASA and radius of gyration (Rg) values remained consistent throughout the simulation, suggesting the compactness and conformational integrity of the complexes. Hydrogen bonding analysis revealed persistent interactions between the ligands and active site residues, supporting the docking results. RMSD, RMSF, SASA, Rg, and hydrogen bonding (Figs. 7, 8 and 9), these dynamic profiles underscored the structural stability and binding persistence of the top-scoring flavonoid compounds during the simulation. Table 6 presents the quantitative analysis by presenting the average RMSD and RMSF values, which are essential metrics for assessing the structural stability and dynamic behavior of protein–ligand interactions during simulations.

Fig. 7.

Molecular dynamics analysis of 2E3X complexes over 100 ns simulation: (A) RMSD showing structural stability, (B) RMSF indicating residue flexibility, (C) Rg measuring protein compactness, (D) SASA tracking solvent exposure, (E) H-bond analysis monitoring binding interactions.

Fig. 8.

Molecular dynamics analysis of 3S9A complexes over 100 ns simulation: (A) RMSD showing structural stability, (B) RMSF indicating residue flexibility, (C) Rg measuring protein compactness, (D) SASA tracking solvent exposure, (E) H-bond analysis monitoring binding interactions.

Fig. 9.

Molecular dynamics analysis of 6Q42 complexes over 100 ns simulation: (A) RMSD showing structural stability, (B) RMSF indicating residue flexibility, (C) Rg measuring protein compactness, (D) SASA tracking solvent exposure, (E) H-bond analysis monitoring binding interactions.

Table 6.

Molecular dynamics simulation parameters for lead compounds from A. echiodies and reference standards over a 100 ns simulation period.

S. no.	Complex	Average ± SD	Average ± SD
		RMSD (Å)	RMSF (Å)
1.	2E3X-ligand 35	0.294 ± 0.081	0.183 ± 0.052
2.	2E3X-standard	0.294 ± 0.071	0.152 ± 0.032
3.	3S9A-ligand 71	0.217 ± 0.061	0.140 ± 0.025
4.	3S9A-standard	0.342 ± 0.085	0.172 ± 0.018
5.	6Q42-ligand 69	0.425 ± 0.139	0.208 ± 0.026
6.	6Q42-standard	0.366 ± 0.098	0.151 ± 0.031

Analysis of the conformational and structural changes in venom target complexes with lead compounds and standards, such as 2E3X-ligand 35 and 2E3X-standard, revealed intriguing insights into their stability and behavior. Figure 7 shows that these metrics over a 100 ns simulation reveal that both complexes exhibit fluctuations in Rg, SASA, and H-bonding. Table 6 shows that both complexes exhibit remarkably similar average RMSD values, 0.294 ± 0.081 Å for 2E3X-ligand 35 and 0.294 ± 0.071 Å for 2E3X-standard, indicating a comparable level of stability. The RMSF values of 0.183 ± 0.052 Å for the ligand and 0.152 ± 0.032 Å for the standard further support this similarity, suggesting that both complexes have nearly identical dynamic behaviors. Time-resolved graphs demonstrated that while the green lines (2E3X-ligand 35) and red lines (2E3X-standard) fluctuated in terms of RMSD, RMSF, Rg, SASA, and H-bonds, these fluctuations were consistent across the 100 ns timeframe. The Rg remains stable at approximately 2.25–2.31 Å, indicating similar compactness, while SASA values between 2100 and 2250 Ų suggest comparable solvent exposure. The hydrogen bond counts, ranging from 400 to 900, reflected similar hydrogen-bonding patterns with some differences in frequency over time. Overall, the data suggest that both complexes maintain structural integrity and demonstrate similar flexibility and solvent accessibility, with only minor differences observed.

A comparative analysis of venom targets, specifically the 3S9A-ligand 71 and 3S9A-standard complexes, offers insights into their structural dynamics. Figure 8 shows that these metrics over a 100 ns simulation reveal that both complexes exhibit fluctuations in Rg, SASA, and H-bonding. Table 6 illustrates that metrics such as RMSD and RMSF, it becomes evident that 3S9A-ligand 71 maintains greater stability with an average RMSD of 0.217 ± 0.061 Å and RMSF of 0.140 ± 0.025 Å. In contrast, the 3S9A-standard showed increased deviation and flexibility, with higher averages of 0.342 ± 0.085 Å for RMSD and 0.172 ± 0.018 Å for RMSF. The Rg ranged from 1.69 to 1.72 Å, indicating similar compactness, while the SASA values between 1100 and 1220 Ų suggest comparable solvent exposure. The hydrogen bond counts fluctuating between 1700 and 1800 further highlight the similar bonding patterns. Overall, these observations suggest that 3S9A-ligand 71 is more structurally stable than the more flexible 3S9A-standard.

A comparative analysis of the 6Q42-ligand 69 and 6Q42-standard complexes revealed intriguing insights into their structural dynamics and stability. Figure 9 illustrates these metrics over time: blue lines represent 6Q42-ligand 69, while black lines represent 6Q42-standard, showing fluctuations in RMSD, RMSF, Rg, SASA, and H-bonds across 100 ns. Over a 100 ns simulation period, both complexes exhibited fluctuations in key metrics, such as Rg, SASA, and H-bonds. In terms of RMSD and RMSF represented in Table 6 suggest that the 6Q42-standard complex displays greater stability, with an average RMSD of 0.366 ± 0.098 Å and RMSF of 0.151 ± 0.031 Å, compared to the 6Q42-ligand 69 values of 0.425 ± 0.139 Å and 0.208 ± 0.026 Å, respectively. The Rg chart shows lines fluctuating around 2.04–2.10 Å, indicating similar compactness with slight variations. The SASA chart displayed traces ranging from 1340 to 1420 to Ų, suggesting comparable solvent exposure with minor differences. The H-bond chart revealed counts between 1560 and 1700, reflecting similar hydrogen bonding patterns with some time-based variations. The traces indicate that 6Q42-standard is more stable, whereas 6Q42-ligand 69 exhibits slightly higher structural deviation and flexibility.

Experimental validation is crucial to confirm in silico findings and ensure that computational predictions are valid under real-world conditions. This bridges the gap between theoretical models and practical applications, enhancing the reliability and credibility of research outcomes³⁹. In vitro and in vivo studies have shown promising outcomes that align with the in-silico predictions previously documented in reports. These results confirm our in-silico findings, suggesting a strong potential for antivenom therapy. Further research is needed to explore its effectiveness and safety for clinical applications. This stepwise approach ensures that only the most viable candidates progress to clinical trials and receive market approval.

Discussion

Andrographis species, particularly A. paniculata and A. serpyllifolia, have been extensively studied for their antivenom properties^7,40. In this study, we explored A. echioides and conducted in silico investigations to assess its potent inhibitors of metalloproteinase, serine proteinase and phospholipase A2 considered to evaluate antivenom agent using its leaf extract. This study represents the first comprehensive in silico evaluation of A. echioides as a source of natural antivenom agents against Russell’s viper envenomation. By employing a comprehensive approach that included GC–MS profiling, multi-target docking, ADMET screening, and 100 ns MD simulations. Trilinolein was identified as the top binder across the three targets, with impressive binding energies of − 115.94, − 107.48, and − 111.35 kcal/mol. However, despite its strong binding affinity, trilinolein lacks desirable drug-like properties. This discrepancy often arises in drug-like properties when a compound exhibits strong target interactions but fails to meet other critical criteria essential for drug development, such as molecular weight, molar refractivity, and lipophilicity. Three lead compounds decanoic acid (ligand 35), oxalic acid, 6-ethyloct-3-yl isohexyl ester (ligand 71), and oxalic acid, 6-ethyloct-3-yl hexyl ester (ligand 69) exhibited strong binding affinity, favorable drug-likeness, and dynamic stability against key venom enzymes. These findings highlight their potential as antivenom drug candidates and underscore the therapeutic value of A. echioides, demonstrating the power of computational approaches in modern venom-based drug discovery.

The outcomes of this study may open new avenues for creating innovative treatments for venomous bites by leveraging the bioactive compounds found in A. echioides. However, additional research, such as laboratory experiments and clinical trials, is crucial to confirm these findings and thoroughly comprehend the therapeutic potential of A. echioides in antivenom development. Among these promising compounds, kaempferol, a naturally occurring flavonol present in fruits, vegetables, and herbs, has shown potential for a wide range of therapeutic uses, including as an antivenom agent. The current research aligns well with earlier reports from computational analyses of natural antivenom compounds, showing promising results. For instance, Ajisebiola et al.⁴¹ identified as a lead hit kaempferol from M. oleifera against metalloproteinases using AutoDock Vina. Similarly, we found from A. echioides leaves which predicted to exhibit a remarkable binding energy against phospholipase A2, metalloproteinase and serine proteinase using iGEMDOCK’s scoring function. Additionally, Borges et al.¹⁵ highlighted 12-methoxy-4-methyl-voachalotine against PLA2, which served as a benchmark in our research. A recent in silico investigation of A. echioides leaf extract as an activator of α-amylase and α-glucosidase has introduced a novel direction for antivenom research⁴². Traditionally linked to diabetes management, these enzymes may offer new therapeutic strategies for venom neutralization through metabolic modulation owing to their role in carbohydrate metabolism. These multifaceted benefits highlight the potential of kaempferol in advancing novel plant-based therapeutic strategies, including innovative antivenom therapies. Moringa oleifera and Vitex negundo are medicinal plants rich in bioactive flavonoids with strong antioxidant and anti-inflammatory properties, making them promising candidates for antivenom therapy.

GC–MS analysis identified numerous flavonoid compounds in both species. Traditionally known for their healing potential, may help counteract the toxic effects of snake venoms. Modern analytical tools, such as GC–MS, have significantly advanced our understanding of flavonoid compounds and their potential applications. These tools allow scientists to interpret the intricate structures of flavonoids, which are vital plant secondary metabolites with diverse biological activities. Flavonoids from Echis ocellatus, a highly venomous snake, exhibit venom-inhibiting properties. By using GC–MS, researchers are able to identify and characterize the specific flavonoid compounds responsible for this inhibition, potentially leading to the development of new antivenom therapies⁴³. The discovery of antivenom compounds with superior binding energies from A. echioides marks a significant breakthrough in the search for effective antivenom agents. This is the first report of such compounds being identified in this plant species, highlighting its previously untapped potential. The identification of these novel antivenom compounds not only expands the botanical sources available for antivenom development but also underscores the importance of exploring diverse plant species to combat venomous bites. The superior binding energies of these compounds suggest that they may offer enhanced efficacy, potentially leading to more effective treatments and improved outcomes for those affected by venomous encounters.

Protein–ligand interaction studies are vital for understanding biological processes and drug development, focusing on hydrogen bonds and hydrophobic interactions⁴⁴. Hydrogen bonds, formed between electronegative atoms, ensure specificity and stability, whereas hydrophobic interactions, driven by entropic gains, cluster nonpolar regions to minimize water exposure. These interactions collectively enhance binding affinity and selectivity, thereby guiding drug design and analysis of biochemical pathways. Protein–ligand interaction studies of the 2E3X, 3S9A, and 6Q42 complexes revealed distinct patterns of hydrogen bonds and hydrophobic interactions. The 2E3X complex (ligand 35) featured two hydrogen bonds at distances of 3.04 Å and 3.09 Å with Cys291 and Arg272, alongside 11 hydrophobic bonds involving residues such as Phe12 and Leu206, totalling 13 interactions. The 3S9A complex (ligand 71) exhibited two hydrogen bonds at 3.04 Å and 3.09 Å with Thr94, paired with 10 hydrophobic bonds, including Phe35A and Leu92A, for a total of 11 interactions. In contrast, the standard lacked hydrogen bonds and instead featured eight hydrophobic bonds. The 6Q42 complex (ligand 69) has no hydrogen bonds but exhibits 14 hydrophobic bonds with residues such as Leu118 and Tyr25, matching its standard’s 15 hydrophobic bonds. These data highlight the varied roles of hydrogen bonds and hydrophobic interactions in the complexes. The 2E3X and 3S9A complexes leverage both interaction types for enhanced specificity and stability, whereas the 6Q42 complex depends heavily on hydrophobic interactions, which may prioritize entropic contributions over directional specificity. These differences are critical for understanding ligand selectivity and optimizing drug design, as the balance between these interactions influences the binding affinity and therapeutic efficacy.

A comparative analysis of the venom target complexes2E3X, 3S9A, and 6Q42 revealed distinct patterns in their structural dynamics and stability over 100 ns simulations. The 2E3X-ligand 35 and 2E3X-standard complexes exhibited nearly identical stability, with equivalent RMSD (0.294 Å) and comparable RMSF values (0.183 Å vs. 0.152 Å), alongside stable Rg (2.25–2.31 Å), SASA (2100–2250 Ų), and hydrogen bond counts (400–900), indicating similar compactness, solvent exposure, and bonding patterns with minor fluctuations. In contrast, the 3S9A-ligand 71 complex demonstrated greater stability (RMSD: 0.217 Å, RMSF: 0.140 Å) than the more flexible 3S9A-standard (RMSD: 0.342 Å, RMSF: 0.172 Å), despite similar Rg (1.69–1.72 Å), SASA (1100–1220 Ų), and hydrogen bond counts (1700–1800). Conversely, the 6Q42-standard complex was more stable (RMSD: 0.366 Å, RMSF: 0.151 Å) than the 6Q42-ligand 69 (RMSD: 0.425 Å, RMSF: 0.208 Å), with both showing comparable Rg (2.04–2.10 Å), SASA (1340–1420 Ų), and hydrogen bond counts (1560–1700). These findings suggest that while all complexes maintain structural integrity, the ligand-bound forms of 3S9A and 6Q42 exhibit contrasting stability profiles compared to their standards, with 3S9A-ligand 71 being more stable and 6Q42-ligand 69 showing greater flexibility, highlighting the nuanced impact of ligand interactions on the dynamics of venom targets.

In silico modeling provides a cost-effective platform for predicting the interactions between plant compounds and venom proteins and serves as a vital precursor to in vivo studies. This innovative approach underscores the potential of plant-derived compounds are effective broad-spectrum antivenom therapies. The evaluation of ADMET properties, such as absorption, distribution, metabolism, excretion, and toxicity, is pivotal in drug development. The favorable ADMET profiles, especially high BBB permeability, suggest potential utility in neurotoxic envenomation. However, in vitro enzyme assays and in vivo neutralization tests are essential to confirm efficacy. Limitations include the use of single protein conformations and lack of whole-venom modeling. Future studies should employ ensemble docking and venom proteomics to enhance predictive accuracy.

The exploration of alternative antivenom compounds is an exciting field of research, particularly given the potential of natural and synthetic compounds to neutralize the effects of venom. Butt et al. (2022) and Parthasarathy et al.⁴⁵ highlighted the efficacy of certain esters, such as 11-Octa-decanoic acid methyl ester and oxalic acid esters, in counteracting scorpion and snake venoms, respectively. These findings align with current research, suggesting that decanoic acid (ligand 35) and oxalic acid esters (ligands 71 and 69) may possess antivenom properties. This correlation opens the door for further investigation into the mechanisms of action and potential applications of these compounds in developing new antivenom therapies. These findings align with those of previous docking studies, reinforcing the potential of ligand 35 as a promising candidate for further development. In silico computational evaluations have validated the bioactivity of lead hits, confirming their efficacy and safety in vitro and in vivo³⁹. Our computational results provide mechanistic insights that support, rather than validate, the traditional medicinal use of A. echioides. These findings establish a theoretical foundation that requires experimental validation to confirm their therapeutic efficacy. This computational study has inherent limitations, including reliance on static protein structures, simplified binding models that may not capture allosteric effects, and lack of experimental validation. The predicted binding affinities may not directly correlate with in vivo efficacy because of factors such as bioavailability, metabolism, and tissue distribution. Additionally, this study focused on individual enzyme targets rather than the complex, multi-component nature of the actual venom.

Conclusion

This study employed an integrated computational approach, including virtual screening, molecular docking, and molecular dynamics simulations, to assess flavonoid compounds from Andrographis echioides as potential antivenom agents. To the best of our knowledge, this is the first report investigating the antivenom potential of bioactive compounds from this species using in silico methods. The results suggest that compounds such as decanoic acid, oxalic acid 6-ethyloct-3-yl isohexyl ester, and oxalic acid 6-ethyloct-3-yl hexyl ester are predicted to exhibit stable and high-affinity interactions with key venom proteins, including PLA2, metalloproteinases, and serine proteases. Our computational results provided mechanistic insights that supported, rather than validated, the traditional medicinal use of A. echioides. These findings established a theoretical foundation that required experimental validation to confirm therapeutic efficacy and guide the development of plant-derived antivenom therapies. Subsequent research efforts should prioritize in vitro enzyme inhibition assays, cytotoxicity evaluations, and in vivo efficacy testing to validate and translate these predictions into potential therapeutic applications.

Abbreviations

ADMET

Absorption, distribution, metabolism, excretion, and toxicity

BBB

Blood–brain barrier

CYP

Cytochrome P450

GC–MS

Gas chromatography–mass spectrometry

TLC

Thin layer chromatography

PDB

Protein data bank

HBD

Hydrogen bond donor

HBA

Hydrogen bond acceptor

Log P

Logarithm of the partition coefficient

MR

Molar refractivity

PLA2

Phospholipase A2

MD Simulation

Molecular dynamics simulation

NAMD

Nanoscale molecular dynamics

RMSD

Root mean square deviation

RMSF

Root mean square fluctuation

RT

Retention time

2D

Two-dimensional

3D

Three-dimensional

H-bond

Hydrogen bond

E

Electrostatic interactions

H

Hydrogen-bonding interactions

Tox

Toxicity

Author contributions

**PJ** : Writing – review and editing, validation, and investigation, **PS** : Writing – original draft, review and editing, visualization, methodology, investigation, data curation, and conceptualization. **SK** : Writing – original draft, review and editing, software, methodology, data curation, conceptualization. All of the authors read and approved the final manuscript.

Funding

This research did not receive any specific grants from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.