In-silico screening of bioactive compounds from nut grass (Cyperus rotundus) to inhibit the α-bungarotoxin from Kraits (Bungarus spp.) venom

ABSTRACT

Kraits (Bungarus spp.) are the snakes that commonly cause fatal bites in Tropical Asia due to their neurotoxic venom. The complex process of antivenom production contributes to the lack of available antivenom and amplifies the burden of snakebite in that region. Thus, the victims seek traditional medicine such as root extract from nut grass (Cyperus rotundus). In the present study, we utilized bioinformatics methods to analyze the potential inhibitory activity of bioactive compounds from nut grass (Cyperus rotundus) against kraits’ α-bungarotoxin. Twenty-one bioactive compounds were screened using ADMET, membrane permeability prediction, molecular docking and molecular dynamics simulations. Among these bioactive compounds, we found that alpha-cyperone, cyperol, cyperusol, rotundine A, rotundine B and 4,6,3’,4’-Tetramethoxyaurone had inhibitory activity by presenting a good ability to pass phospholipid bilayer and could bind stably into active sites of α-bungarotoxin with minimum energy required. Importantly, the neutralization property of these bioactive compounds resides in their combinational function, underscoring the crucial significance of their combined action rather than individual contributions. In conclusion, this study identifies a combination of six bioactive compounds from C. rotundus with a potential as an alternative antivenom based on in-silico screening and provides essential data for the development of a plant-based antivenom.

KEYWORDS:

• α-bungarotoxin
• antivenom alternative
• Cyperus rotundus
• snakebite
• terpenes
• traditional medicine

Background

Snakebite envenoming is one of the diseases categorized by WHO as a neglected tropical disease, causing 2.7 million victims annually, and among them, 138,000 people lose their life [1–3]. Most snakebite envenoming cases are taken place in Asia, with 74% of total cases worldwide [2]. This high incidence of snakebite envenoming is correlated with the high diversity of venomous snakes in this region [4,5].

Kraits (Bungarus spp.) are snakes belonging to the family Elapidae, consisting of 16 species and distributed in tropical Asia [6]. Bungarus spp. are considered medically important snakes due to their neurotoxic venom, thanks to phospholipase A2 (PLA2) and three-finger toxin (3FTX) proteins [7–10]. Among 3FTX proteins, α-bungarotoxin is a major neurotoxic protein consisting of 74 amino acid residues with a molecular weight of 7800–8000  Da [11]. This protein was initially found in B. multicinctus but then in other Bungarus species [8,11–14]. α-bungarotoxin acts as a postsynaptic neurotoxin by binding into neuronal nicotinic acetylcholine receptor (nAChR), causing paralysis and death [15,16]. Despite the venom’s danger, antivenom’s availability is limited by several difficulties, such as antivenom production and supply and a distant health facility [17,18]. Consequently, the snakebite victims seek traditional medicines as an alternative [18,19].

The use of traditional medicines to treat snakebite envenomation is common among rural communities as they lack access to medical facilities [18–21]. In South Asia, the traditional healer uses plant-based medicine such as bulb powder from Cyperus rotundus to treat victims bitten by Bungarus spp [20,22–24]. However, previous studies focused on the practical application of plants instead of their molecular interaction. Whereas the medicinal properties of traditional medicines resulted from the interaction between bioactive compounds and snake venom [25–27]. Therefore, the present work is designed to investigate the bioactive compounds from C. rotundus, which has the potential to inhibit α-bungarotoxin from Bungarus spp. venom using bioinformatics methods.

Materials and methods

Data mining

Twenty-one bioactive compounds of Cyperus rotundus were retrieved from previous study [28] . The canonical SMILES and the 3D structure of the bioactive compounds were collected from PubChem (https://pubchem.ncbi.nlm.nih.gov/). The 3D structure of α-bungarotoxin from B. multicinctus and the control ligand were downloaded from the protein data bank (https://www.rcsb.org/) with accession number 1HAA [29].

Drug-likeness and pharmacokinetics prediction

We predicted the drug-likeness and pharmacokinetics using SwissADME (http://www.swissadme.ch/) [30]. The predictions were carried out to understand the physicochemical and metabolism properties of the bioactive compounds. The bioactive compounds that did not violate Lipinski’s rule were considered. All bioactive compounds with consensus Log Po/w higher than three proceeded to the molecular docking analysis.

Toxicity evaluation and membrane permeability prediction

The toxicity of each compound was evaluated using Pro-Tox II (https://tox-new.charite.de/protox_II/) webserver [31,32]. The compound’s potential to induce hepatotoxicity, carcinogenicity, immunotoxicity, mutagenicity, and cytotoxicity was also analyzed using Pro-Tox II. Membrane permeability prediction was carried out using PerMM Server (https://permm.phar.umich.edu/server) [33]. We set the experimental temperature to 310 K and pH 7.4. Drag optimization and BLM permeability estimation were chosen for the calculation methods.

Molecular docking

Prior to the molecular docking, the selected bioactive compounds were prepared by minimizing the energy using open babel tool in PyRx v0.8. The bioactive compounds then converted to pdbqt ligand format using PyRx v0.8 [34]. The protein (α-bungarotoxin) was prepared by removing the water molecules and native ligands using Discovery Studio 2019 [35]. The PyRx v0.8 software comes prepackaged with Autodock Vina. Molecular docking was performed using a specific docking method using Autodock Vina. The α-bungarotoxin was set as rigid receptor and the bioactive compounds were set as ligands [34]. The specific search area for molecular docking is listed in Table 1. The interactions between bioactive compounds and the protein were then visualized using Discovery Studio 2019 [35].

Table 1. Vina search space used in this study.

Grid	Axis
	X	Y	Z
Centre	8.2022	2.8887	1.4352
Dimensions (Angstrom)	20.2814	18.1712	13.4120

Molecular dynamics simulation

Molecular dynamics simulation was conducted using GROMACS [36] at the WebGro server (https://simlab.uams.edu/ProteinWithLigand/index.html) (UAMS) [37]. We used GROMOS96 43a1 forcefield [38,39] and cubic box type for input parameters. The temperature for molecular dynamics was adjusted to 310 K, and the simulation time was set to 20 ns.

Results and discussion

Drug-likeness, pharmacokinetics, toxicity and membrane permeability of the bioactive compounds

We assessed 21 bioactive compounds from C. rotundus and found 19 bioactive compounds met these Lipinski’s rule for drug-like characteristics (Table 2). This assessment was based on physicochemical characteristics as follows: their molecular weight must not exceed 500 Da, H-bond donors ≤5, H-bond acceptors ≤10 and log p ≤ 5 [40]. The bioactive compounds that meet these criteria could enter and circulate throughout the body and reach the protein target [41].

Table 2. The 19 bioactive compounds that had drug-like characteristics, sorted by their log P_o/w..

Bioactive compound	PubChem Id	Formula	Molecular Weight (g/mol)	No. of H-Bond Donors	No. of H-Bond Acceptors	Lipophilicity (Consensus Log Po/w)
Alpha-Cyperone	6452086	C15H22O	218.33	0	1	3.69
Cyperol	14076601	C15H24O	220.35	1	1	3.49
Cyperusol B2	11310916	C15H24O2	236.35	1	2	3.15
Rotundine B	21603505	C15H23NO	233.35	1	2	3.07
Rotundine A	10728239	C15H21NO	231.33	0	2	3.04
4,6,3’,4’ Tetramethoxyaurone	11186921	C19H18O6	342.34	0	6	3.01
4,7-Dimethyl −1-tetralone	5316902	C12H14O	174.24	0	1	2.85
Beta rotunol	5321005	C15H22O2	234.33	1	2	2.77
6,3‘, 4’- Trihydroxy-4-methoxy-5-methylaurone	42607777	C17H14O6	314.29	3	6	2.25
Cyperusol D	11311311	C15H24O3	252.35	3	2	2.16
Apigenin	5280443	C15H10O5	270.24	3	5	2.11
Scaberin	42608130	C19H18O7	358.34	1	7	1.92
Cyperusol A1	10083471	C15H22O3	250.33	2	3	1.9
Sulfuretin	5281295	C15H10O5	270.24	3	5	1.9
Luteolin	5280445	C15H10O6	286.24	4	6	1.73
Aureusidin	5281220	C15H10O6	286.24	4	6	1.72
p-Coumaric acid	637542	C9H8O3	164.16	2	3	1.26
Quercetin	5280343	C15H10O7	302.24	5	7	1.23
4-Hydroxybenzoic acid	135	C7H6O3	138.12	2	3	1.05

We also evaluated the bioactive compounds using their lipophilicity value. From 19 bioactive compounds, we found 6 of the bioactive compounds have consensus Log Po/w greater than 3 (Table 2). The lipophilicity value was essential to understand the ability of bioactive compounds to cross the phospholipid bilayer [42,43]. The ability to cross the phospholipid bilayer enabled the bioactive compound to pass into the lymphatic system [44]. As snake venom enters the body via the lymphatic system [45], the bioactive compounds must be able to cross the epithelial phospholipid layer and then enter the lymphatic system to block the specific protein venom.

After selecting bioactive compounds, we evaluated their pharmacokinetics, with a specific focus on gastrointestinal metabolism and blood–brain barrier (BBB). Our findings revealed that all six bioactive compounds exhibited a high rate of metabolism in the gastrointestinal tract (Table 3). Consequently, the bioavailability of these bioactive compounds is likely to be low [46]. Furthermore, it is noteworthy that all bioactive compounds could cross the blood–brain barrier. These results add a layer of complexity to the factors influencing the pharmacokinetic profile of these compounds. While the high metabolism in the gastrointestinal tract underscores the need for optimizing dosage and formulation strategies, the ability to cross the BBB introduces an additional aspect to the dosage optimization process [47].

Table 3. Pharmacokinetics and metabolism prediction of bioactive compounds.

No	Bioactive compound	Blood-brain barrier permeability	Gastrointestinal tract absorbance
1	Alpha-Cyperone	Yes	High
2	Cyperol	Yes	High
3	Cyperusol B2	Yes	High
4	Rotundine B	Yes	High
5	Rotundine A	Yes	High
6	4,6,3’,4’ Tetramethoxyaurone	Yes	High

The toxicity of six potential active compounds was evaluated using the Pro-Tox server. Based on toxicity class, Alpha Cyperone, Cyperusol B2, Rotundine B, Rotundine A, and 4,6,3‘,4’ Tetramethoxyaurone has toxicity class 4 with an LD50 of 300–2000 mg/kg. Meanwhile, Cyperol has toxicity class 5 with an LD50 of 3000 mg/kg (Figure 1(a)). The toxicity class, determined by the Globally Harmonised System (GHS) and ranging from class I to VI based on LD50 values in mice with the highest class (Class VI), signifies safety when taken orally. In our case, most bioactive compounds fall into toxicity class 4, which is harmful if swallowed. One compound namely Cyperol, fell into toxicity class 5 represent the compound may be harmful if swallowed [31]. The prediction of oral toxicity could help in-vivo studies to determine the safe dosage. 4,6,3‘,4’ Tetramethoxyaurone has potential carcinogenic, immunogenic and mutagenic properties. In comparison, the other compounds were relatively safe (Figure 1(b)). Therefore, the use of 4,6,3‘,4’ Tetramethoxyaurone in experimental studies needs to consider and focus on safe doses.

Figure 1. The toxicity evaluation and membrane permeability prediction of six potential compounds. a) Toxicity classification based on toxicity class and LD50. b) The probability of inducing several kinds of toxicity. c) The membrane permeability prediction of six bioactive compounds (top) with the energy required to pass the phospholipid membrane (bottom).

Further membrane permeability simulation was conducted to confirm the six bioactive compounds’ permeability through the phospholipid membrane. Our results showed that five bioactive compounds from C. rotundus could pass over the phospholipid membrane with the minimum transfer energy required (Figure 1(c)). One of the bioactive compounds, namely 4,6,3’,4’-Tetramethoxyaurone, required relatively high transfer energy to cross the phospholipid bilayer, indicating that this compound was slightly lipophilic. The low lipophilicity of 4,6,3’,4’-Tetramethoxyaurone was caused by the high number of H-bond acceptor atoms in this compound compared to other compounds (Table 2). This result was unsurprising as the previous study suggested that the number of H-bond acceptors was the determinator for lipophilicity calculation [40].

Several parameters were available to predict the drug-likeness of bioactive compounds. Among that, the Lipinski parameter was broadly used for drug-likeness due to its simplicity, fast and feasibility [48]. With the help of the Lipinski parameter and membrane permeability simulation, we obtained six bioactive compounds from C. rotundus that could potentially enter the lymphatic system.

Interactions between bioactive compounds and α-bungarotoxin

The beta-hairpin structures of three-finger region of α-bungarotoxin are crucial for binding to nAChRs [49]. The six bioactive compounds of C. rotundus were able to bind to at least one of three finger regions of α-bungarotoxin. The binding affinities from these interactions ranged from −4.8 to −6.0 kcal/mol. Notably, Alpha-cyperone showed the highest interaction with nine bonds, while Rotundine A had the least interaction, with only one hydrophobic bond at Lys70 (Figure 2; Table 4). These results indicate that individual bioactive compound interacted with different sites of the α-bungarotoxin and with different binding affinity. Despite this, it is noteworthy that the interaction of one bioactive compound does not preclude the remaining bioactive compounds from engaging in interactions. This underscores the complexity of multiple interactions within a system and highlights the need for a comprehensive understanding of the collective impact of bioactive compounds on the target protein, α-bungarotoxin.

Figure 2. The visualization of molecular docking between α-bungarotoxin (cyan) with a) Alpha-cyperone, b) Cyperol, c) Cyperusol B2, d) Rotundine a, e) Rotundine b, f) 4,6,3’,4’-tetramethoxyaurone, g) Anti-α-bungarotoxin peptide. Dashed lines represent interaction type, with the dark and light green lines representing hydrogen bonds and the pink line representing hydrophobic interaction.

Table 4. Detailed interactions between α-bungarotoxin with bioactive compounds and inhibitor control.

Protein	Bioactive compound	Binding affinity (Kcal/mol)	Interaction	Amino acid residues
α-bungarotoxin	anti-α-bungarotoxin peptide	−3.8	Hydrogen bond	THR6 PRO10 ILE11 SER35 ARG36 LYS38
			Hydrophobic	–
	Alpha-cyperone	−5.1	Hydrogen bond	ARG36
			Hydrophobic	LYS38 VAL40 HIS68 PRO69 LYS70
	Cyperol	−5.5	Hydrogen bond	–
			Hydrophobic	LYS38 VAL40 HIS68 PRO69 LYS70
	Cyperusol B2	−4.8	Hydrogen bond	ASP30 GLY37 HIS68
			Hydrophobic	LYS38 LYS70
	Rotundine A	−5.0	Hydrogen bond	–
			Hydrophobic	LYS70
	Rotundine B	−5.1	Hydrogen bond	ASP30 GLY37
			Hydrophobic	LYS70
	Tetramethoxyaurone	−6.0	Hydrogen bond	THR6 SER35 ARG36 VAL40 HIS68
			Hydrophobic	VAL40

On the other hand, the anti-α-bungarotoxin peptide could interact with all three finger regions of α-bungarotoxin specifically at Thr6, Pro10, Ile11 (finger 1) Ser35, Arg36, Lys38 (finger 2) and His68 (finger 3) with a binding affinity of −3.8 kcal/mol. All molecular interactions that were responsible for binding the peptide with the α-bungarotoxin were hydrogen bond (Figure 2; Table 4). Despite having the highest hydrogen bond, the peptide showed a relatively high binding affinity with α-bungarotoxin. This instance could happen with the presence of mixed strong and weak attachments of the hydrogen bond [50].

The key to toxin inhibition is the ability to bind into all three finger regions of the α-bungarotoxin. The previous study showed that a peptide that mimics the native receptor of α-bungarotoxin could neutralize its toxicity [29]. The peptide interacted with all three finger region structures of α-bungarotoxin thanks to its structural similarity with the native ligand. In contrast, our study used smaller compounds in which individual compounds could only interact with one of the finger regions from α-bungarotoxin. However, when combined, all bioactive compounds from C. rotundus could bind with all three finger regions of α-bungarotoxin and possibly could co-interact with each other and further strengthen the conformation. Combination of all bioactive compounds had a similar interaction with the peptide from the previous study [29], indicating that bioactive compounds from C. rotundus had the potential to inhibit the α-bungarotoxin.

Several studies have utilized bioactive compounds from plant extracts to inhibit the toxic effect of snake venom. These studies included some terpenes products such as ar-turmerone from Curcuma longa [25] and (E)-17-ethyliden-labd-12-ene-15,16-dial from Curcuma zedoaria [26]. The latter showed the antagonist ability with king cobra (Ophiophagus hannah) venom, preventing the blockade of the neuromuscular junction and thus neutralizing the lethal effect of the venom [26]. By using similar compounds within the terpenes class, this study could have promising results when tested in-vitro and in-vivo.

Molecular dynamics simulation

The root-mean-square deviations (RMSD) were calculated to determine the stability of ligand-protein complexes [51,52]. The results showed that one specific combination, α-bungarotoxin with Alpha-cyperone and Cyperol, was the most stable, with very slight fluctuations of 0.23 nm. On the other hand, the combination of α-bungarotoxin with Rotundine A showed more fluctuations, reaching 0.38 nm with an average of 0.43 nm (Figure 3(a)). The individual stability of ligand-protein complexes could be determined by the RMS fluctuation (RMSF) value [53]. The RMSF showed a relatively stable value with an average of 0.15 nm with only several peaks at amino acid residues 30–37 and 70–74 (Figure 3(e)).

Figure 3. The RMSD (a), Radius of gyration (b), Number of Hydrogen Bond (c), Surface area (d), and RMSF (e) Calculation for α-bungarotoxin and bioactive compound complexes after 20 nanoseconds of simulation. αbtx-acypn: α-bungarotoxin- alpha-cyperone, αbtx-cyprl: α-bungarotoxin-cyperol, αbtx-cypB2: α-bungarotoxin-cyperusol B2, αbtx-rtdA: α-bungarotoxin-rotundine A, αbtx-rtdB: α-bungarotoxin-rotundine B, αbtx-ttrx: α-bungarotoxin-4,6,3’,4’-tetramethoxyaurone.

The molecular dynamics simulation also calculated several factors related to the compactness of the ligand-protein binding, such as radius of gyration (Rg), solvent-accessible surface area (SASA) and the number of hydrogen bonds [53–55]. All complexes showed a relatively low average Rg value that ranged from 1.20 to 1.24 nm (Figure 3(b)). Likewise, the SASA value of the complexes was similar, with an average value of 47.99, 48.33, 48.53, 48.51, 47.94 and 47.99 for alpha-cyperone, cyperol, cyperusol, rotundine A, rotundine B and 4,6,3’,4’-Tetramethoxyaurone complexes, respectively (Figure 3(d)). These results indicated that the binding of bioactive compounds with α-bungarotoxin was compact.

A similar hydrogen bond pattern was found in all ligand-protein complexes during 20 ns molecular dynamics simulation. The number of hydrogen bonds that formed was in the range of 24–53 (Figure 3(c)). The relatively high number of hydrogen bonds suggests that the ligand-protein complexes have the best conformation when docked [56,57]. Overall, our findings suggest that these combinations of substances with the protein were stable and compact.

Conclusion

In summary, six out of the 21 bioactive compounds, namely alpha-cyperone, cyperol, cyperusol, rotundine A, rotundine B and 4,6,3’,4’-Tetramethoxyaurone from C. rotundus, demonstrated inhibitory potential against α-bungarotoxin from Bungarus spp. Notably, their collective action revealed neutralization potential, emphasizing the synergy when these compounds are combined. These findings provide initial data for an alternative antivenom against neurotoxicity from Bungarus spp. venom. However, further studies are needed to assess the inhibitory mechanism and dosage of bioactive compounds using in-vitro and in-vivo methods.

Additionally, employing bioinformatics to screen plant-based antivenom alternatives provides insight into the future of snakebite management. This approach encourages a shift in the exploration of effective and accessible treatments, offering a potential solution to the wider challenges of antivenom scarcity.

Abbreviations

3FTX:	=	three-finger toxin
PLA2:	=	phospholipase A2
nAChR:	=	neuronal nicotinic acetylcholine receptor
RMSD:	=	root-mean-square deviations
Rg:	=	radius of gyration
RMSF:	=	RMS fluctuation
SASA:	=	three-finger toxin

Authors’ contributions

RG: Conceptualisation, Formal analysis, Investigation, Writing – Original Draft. MHW: Formal analysis, Methodology, Writing – Original Draft. SR: Supervision, Writing – review & editing. MR: Supervision, Writing – review & editing. NW: Conceptualisation, Resources, Supervision, Writing – review & editing.

Availability of data and materials

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

Acknowledgments

We are grateful to Feri Eko Hermanto for assisting molecular dynamics simulation and providing the early review of this manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).