Effect of guggulsterone, a sterol identified in Commiphora gileadensis (Becham), on the dengue virus enzymes: Pharmacokinetics, molecular docking and molecular dynamics simulations studies

2.1 Natural compound selection and preparation

The antiviral activity of guggulsterone has been documented against different viruses, including herpes simplex virus type 2 and SARS-CoV2 (Bouslama et al., 2019; Mishra et al., 2021). In this study, guggulsterone was selected for in silico evaluation of the antiviral activity on four dengue proteins, the structures of guggulsterone and a reference antiviral drug (ribavirin) were obtained from the PubChem database (Kim et al., 2016). The compounds (guggulsterone and ribavirin) were prepared for docking using LigPrep of Maestro (Schrodinger maestro 2020–3). Ligands energy was minimized with the OPLS3e force field, then protonation and tautomeric states were generated at pH 7 ± 2 using Epik (Greenwood et al., 2010).

2.2 Protein models preparation

Four proteins from the dengue virus were selected as a receptor for docking experiments, and their 3D structures were obtained from the RCSB PDB database (Berman et al., 2002). These proteins include: NS5 RNA dependent RNA polymerase domain (PDB id: 2J7W), dengue methyltransferase (PDB id: 1L9K), NS3 protease-helicase (PDB id: 2VBC) and dengue virus type 2 envelope glycoprotein (PDB id: 1OK8) (Nag and Chowdhury, 2020). Prior docking proteins structures were subjected to protonation step, removal of water atoms beyond 3 Å of heteroatoms, and energy minimization step achieved by OPLS3e at pH 7 ± 2. Maestro interface (Protein preparation, SiteMap and Receptor Grid generation) were utilised for active site prediction and protein preparation.

2.3 Molecular docking

Prior docking procedure the docking protocol was validated by redocking the co-crystalized ligand of the dengue methyl transferase (1L9K), then RMSD was calculated for superposition of ligands using Maestro. After preparing the selected proteins and ligands, molecular docking was used to study the interaction of guggulsterone and dengue target proteins using Precision (SP) docking of Maestro interface, docking score, the energy of van der Waals (evdW), and glide hydrogen bond was calculated.

The protein was set as rigid while ligand was set as flexible, the van der Waals radii scaling factor was set at 0.8 and partial charge cutoff at 0.15. Post docking energy minimization step was done, and only 55 poses per ligand were included. After docking, the PLIP webserver and Maestro were used to examine and visualize residues of the protein associated with ligand interaction from the docking process.

2.4 Molecular Dynamics simulation

Molecular Dynamics (MD) simulation was proven to be a valuable method for analysing the physical foundations of biomolecule structure and functional properties (Khan et al., 2022). In this study, the stability of the complex of guggulsterone and the four selected proteins have been studied by MD simulation for 50 ns. The simulation process was achieved using the Desmond package in the Schrödinger Maestro interface (Release, 2020). The system was prepared by using TIP4P water model, and orthorhombic box shape was assigned with distances 10 Å and force filed OPLS4 (Lu et al., 2021). Then ions were neutralized by additions of required charges. During MD simulation at each 100 ps interval, snapshots were recorded, the nose-hover thermostat method was specified with a relaxation time of 1 ps and 2 fs time step. The system was minimized at 2000 iterations, NPT temperature was at 300 °K, then Maestro was used for visualization of trajectories. Then Root Mean Square Deviation (RMSD) and Root Mean Square Fluctuation (RMSF) were generated.

2.5 Calculation of binding free energy

The relative binding free energy (ΔG_bind) was calculated for the complex of protein and ligands during MD simulation. The ΔG_bind of the trajectories generated from simulation were calculated by Schrodinger Prime MM/GBSA using thermal_mmgbsa.py script with step size 10 (around 100 conformations were used as input for prime) (Alnajjar et al., 2020).

2.6 ADME and toxicity profiles prediction

In order to select a Drug-like molecule, absorption, tissue distribution, metabolic effects (ADME), and toxicity profiles should present favourable parameters and fill with Lipinski's rule of 5 (Lipinski et al., 1997). Physicochemical and pharmacokinetic properties of guggulsterone were estimated using Swiss ADME (Daina et al., 2014). The ProTox-II web server (Banerjee et al., 2018) was also used to explore hepatotoxicity, immunotoxicity, carcinogenicity, cytotoxicity and mutagenicity. Values were estimated after comparing the properties of a selected molecule with those of 95% of known recognized medicines (Saxena et al., 2019).

3.1 Validation of the docking protocol and active site prediction

For validation of docking protocol, the co-crystalized ligand (S-adenosyl-l-homocysteine) of dengue methyl transferase (1L9K) was removed and redocked in the protein binding site, the redocked ligand gave an RMSD of 0.6407 Å, which indicates the validity of the used docking protocol (Jamal et al., 2017). The superposition of co-crystallized ligands and redocked is shown in Fig. 1. Schrodinger SiteMap was used to assess the druggability of binding pockets. Two proteins were of highly druggability (druggability score [D score] > 1.0) (2VBC: 1.105 and 2J7W: 1.026), while only one (1OK8: 0.663) was poorly druggable (Yazdani et al., 2021). The active site of dengue methyl transferase (1L9K) was estimated according to the site of the co-crystalized ligand.

Close

Fig. 1

Superposition of co-crystalized (green) and redocked (yellow) ligands.

Export to PPT

3.2 Molecular docking analysis

Molecular docking studies were done to examine the best conformation of guggulsterone (ligand) on selected dengue virus target proteins, and the binding affinity scores are shown in Table 1. The docking results revealed that guggulsterone has a binding affinity ranging from −5.485 to −3.422 kcal/mol. The in silico activity of guggulsterone on dengue target proteins is aligned with a previous in vitro study in which they showed the activity of guggulsterone on the dengue virus (Chen et al., 2021).

The guggulsterone showed the best binding affinity to NS5 RNA dependent RNA polymerase (RdRp) domain complex (PDB id: 2J7W), the docking score was −5.5 kcal/mol, Van der Waals energy (evdw) −27.5 kcal/mol, and glide energy −26.64 kcal/mol. The docking score of guggulsterone was near to the score of reference antiviral drug ribavirin (−6.2 kcal/mol), which is known as an inhibitor to dengue inosine 5′-monophosphate dehydrogenase inhibiting polymerases enzymes (Benarroch et al., 2004). We also found that TRP447 is associated with stabilization of docking reaction via formation of one strong hydrogen bond with small distance (2 Å) of hydrogen to acceptor atom (H-A) and between donor and acceptor atom (D-A) 2.8 Å (Table 2 and Fig. 2A and B). The previously published paper revealed that the TRP447 is located at the catalytic domain of dengue RdRp domain (Yap et al., 2007); the inhibition of such residue could block the enzyme activity. Also, as shown in Table 3, four hydrophobic interactions were formed as a result of guggulsterone binding to RdRp active site, TRP302, PHE354, VAL358, and GLN602, residues from 341 to 361 were located at the conserved region of βNLS region of NS5, which is associated with NS3 helicase interaction (Yap et al., 2007).

Close

Fig. 2

Interaction of the guggulsterone compound with: dengue NS5 RNA dependent RNA polymerase: A. 3D interaction during molecular docking and B. 2D interaction of ligand during MD simulation; dengue methyl-transferase: C. 3D interaction during molecular docking and D. 2D interaction of ligand during MD simulation; dengue NS3 protease-helicase: E. 3D interaction during molecular docking and F. 2D interaction of ligand during MD simulation; and Dengue virus type 2 envelope glycoprotein: G. 3D interaction during molecular docking and H. 2D interaction of ligand during MD simulation.

Export to PPT

The interaction of dengue methyltransferase (PDB id: 1L9K) and guggulsterone produced −4.724 kcal/mol docking energy, one hydrogen bond was formed with LYS181 with a small distance of 2.33 Å, and three hydrophobic interactions were formed with LYS105, ASP146, and ILE147. This finding agree with Geoffrey et al. (2020) who found various compounds targeting the dengue virus protein and displayed hydrogen bonds with LYS181, ASP146, and LYS105 (Fig. 2C and D).

The NS3 protease-helicase (PDB id: 2VBC) and guggulsterone have been shown −4.1 docking score (Table 1), which may be due to one hydrophobic interaction (Fig. 2E and F) with the residue VAL544 with a distance of 3.66 Å (Tables 3).

The interaction of dengue virus type 2 envelope glycoprotein (PDB id: 1OK8) and guggulsterone generate a docking score of −3.4 kcal/mol, glide energy of −25.1 kcal/mol, and Van der Waals energy (evdw) −24.1 kcal/mol suggesting its potency to interact with dengue virus by blocking major the envelope glycoprotein. Two hydrogen bonds were formed with SER7 and ASN355; their distances were 3.2 Å and 2.06 Å, respectively (Tables 1 and 2). Additionally, three hydrophobic interactions were generated with dengue major envelope protein E residues VAL24, VAL31, and ASN355; their distances were 3.71 Å, 3.47 Å, and 3.46 Å respectively (Table 3 and Fig. 2G and H). Geoffrey et al. (2020) reported a good binding affinity of piperine to different proteins of dengue virus, including major envelope protein E; and they found that Ile4A, Arg9A, Val24A, Val31A, Lys284A were associated with this interaction. In agreement with these findings, it was found that guggulsterone strongly interacted with VAL24.

3.3 Molecular dynamic (MD) simulation

Since we set the protein rigid during docking, MD simulation will compute atom movement during a period of time by applying Newton’s equation of motion (Adcock and McCammon, 2006). The stability of the protein–ligand complex was assessed with MD simulation, therefore the four docking complexes were run at 50 ns simulation time. As presented in Fig. 3A, the root mean square deviation (RMSD) plot showed that the guggulsterone is not stable with dengue NS5 RNA dependent RNA polymerase (2J7W) and protein backbones fluctuated from the beginning of simulation until the end. Most of the observed interactions lasted <20% of simulation time except GLN580, in which one hydrogen bond last for 50% (Fig. 4A); which might be the reason for the instability of the complex.

Close

Fig. 3

Root mean square deviations (RMSDs) trajectories of the dengue protein residues (A = 2J7W, B = 1L9K, C = 2VBC, D = 1OK8) and guggulsterone complexes run at 50 ns, the protein backbones carbon alpha atoms showed in in blue color and the ligand in red color.

Export to PPT

Close

Fig. 4

The histogram shows the percentage of bonds lasting during MD simulation of the complex of Dengue protein residues (A = 2J7W, B = 1L9K, C = 2VBC and D = 1OK8) and guggulsterone. The green color indicated the H-bonds, the blue color showed the water bridges, and the violet color demonstrated the hydrophobic interaction.

Export to PPT

Fig. 3B show that the protein backbones of dengue methyltransferase (PDB id: 1L9K) is more flexible, most of the flexibility was observed at a region between 2 and 8, 13–20, and 27–45 ns of trajectories. These trajectories are part of the flexible region β2 and β3 which are part of the AdoMet-binding pocket (Zhao et al., 2015). Instead of high protein flexibility, the guggulsterone is still stable in protein backbones from the beginning until 50 ns of simulation time. A stable hydrogen bonds with protein residues (PHE133) in more than 60 % of simulation time, and the presence of water bridges in more than 50 % of simulation time, with GLY 148 (Fig. 4B) were also noticed.

Fig. 3C showed the complex of the NS3 protease-helicase (PDB id: 2VBC) and guggulsterone, the ligand RMSD fluctuation observed from 1 ns until the end of the simulation, the ligand separated from protein backbones in the range of 0.4 to 1.8 Å, which fallen in the permissible range of 1–3 Å (Choudhary et al., 2020), and indicate the stability of the complex.

According to Fig. 4C, a water bridges with ASP 603 (more than 60 % of simulation time) and a hydrophobic interaction with VAL 544 were observed.

The RMSD plot in Fig. 3D show the complex of dengue virus type 2 envelope glycoprotein (PDB id: 1OK8) and guggulsterone. The stability of the complex was the best during MD simulation; but the docking score was the lowest one (−3.4 kcal/mol); this could be due to the considering protein as rigid during docking, which may cause failure in assessing binding affinity in some proteins (Radwan and Mahrous, 2020). The complex remains stable (Fig. 3D) and aligned during the whole simulation time (50 ns); a slight deviation within the acceptable range is observed from 10 to 20 ns. This stability could be due to the presence of stable hydrogen bonds with protein residues (ASP10, VAL24, ASN355) in more than 80 % of simulation time, and the presence of water bridges in more than 60 %, with SER29, CYS30, and ILE357 (Fig. 4D).

As shown in Fig. 5 (A–D), the RMSF showed the fluctuation interacting residues during MD simulation. The interaction of guggulsterone and dengue virus NS5 RNA dependent RNA polymerase domain ((PDB id: 2J7W) showed 80 % (25) of interacting residues were at stable region (Fig. 5A), the interaction of the ligand occurred with 22 (100 %) stable residues of dengue methyltransferase (PDB id: 1L9K), 20 (100 %) stable residues of NS3 protease-helicase ((PDB id: 2VBC) and 24 (100 %) residues of dengue virus type 2 envelope glycoprotein ((PDB id: 1OK8) (Fig. 5B–D).

Close

Fig. 5

The RMSF lasting during MD simulation of the complex of dengue protein residues (A = 2J7W, B = 1L9K, C = 2VBC and D = 1OK8) and guggulsterone. The green coloured lines indicate the interacting residues.

Export to PPT

3.4 Calculation of binding free energy

MM/GBSA analysis was used for the estimation of relative binding energies of ligands binding to protein's active sites, usually more negative values indicate better stability (Dash et al., 2019). As shown in Table 4, the dengue virus type 2 envelope glycoprotein ((PDB id: 1OK8) showed the lowest ΔG_bind (−53.24) in a good range −59.0159 to −47.1885 kcal/mol, this finding is aligned with MD simulation in which the ligand was the most stable in protein backbones. While the complex of guggulsterone and dengue virus NS5 RNA dependent RNA polymerase domain ((PDB id: 2J7W) showed higher fluctuation during MD simulation and thermodynamic energy calculation (high standard deviation 7.26).

3.5 ADME and toxicity profiles prediction

As presented in Table 5, the pharmacokinetics, drug-likeness and toxicity profiles of guggulsterone were estimated in silico. Results showed that guggulsterone follows Lipinski’s rule of five (molecular weight (MW) < 500 Da, number of hydrogen bond donors (HBDs) < 5, LogP < 5, and number of hydrogen bond acceptors (HBAs) < 10). Also, no violation of Veber and collaborators rules (number of rotatable bonds (NBR) < 10 and polar surface area (PSA) < 140 Å²) was noticed (TPSA value of 34.14 Å²), which may facilitate the cross of guggulsterone toward cell membranes (Veber et al., 2002).

Table 5 Physicochemical properties, pharmacokinetics, druglikeness, bioavailability and toxicity parameters of guggulesterone.

Physicochemical properties
Formula	C21H28O2
Molecular weight	312.45 g/mol
Num. heavy atoms	23
Num. arom. heavy atoms	0
Fraction Csp3	0.71
Num. rotatable bonds	0
Num. H-bond acceptors	2
Num. H-bond donors	0
Molar Refractivity	93.54
TPSA	34.14 Å2
Pharmacokinetics
GI absorption	High
BBB permeant	Yes
P-gp substrate	No
CYP1A2 inhibitor	No
CYP2C19 inhibitor	Yes
CYP2C9 inhibitor	Yes
CYP2D6 inhibitor	No
CYP3A4 inhibitor	No
Log Kp (skin permeation)	−5.41 cm/s
Druglikeness
Lipinski	Yes; 0 violation
Ghose	Yes
Veber	Yes
Egan	Yes
Muegge	Yes
Bioavailability Score	0.55
Bioavailability (Radar plot)
LIPO: Lipophilicity FLEX: Flexibility POLAR: Polarity INSAT: Insaturation INSOLU: Insolubility SIZE: Size
Oral toxicity prediction (Probability)
Hepatotoxicity	Inactive (0.67)
Immunotoxicity	Active (0.98)
Carcinogenicity	Active (0.56)
Cytotoxicity	Inactive (0.79)
Mutagenicity	Inactive (0.99)

The blood–brain barrier (BBB) permeability and glycoprotein P (P-gp) substrate were used to predict distribution. The BBB is a structure separating the central nervous system (CNS) from peripheral tissue, controlling the passage of nutrients and materials to maintain homeostasis (Domínguez-Villa et al., 2021). It appeared from this investigation that guggulsterone was highly absorbed in the gastrointestinal (GI), cannot be a substrate of P-gp, and passed the blood–brain barrier (BBB).

Based on the bioavailability index, radar plot provides a graphical insight about the drug likeness of a tested molecule (Ali and Shabeer, 2021). The pink area is the acceptable range for oral bioavailability of each parameter (Mendie and Hemalatha, 2022). As shown in Table 5, the radar plot of guggulsterone falls in the pink region, indicating that it may be regarded as drug-like. Furthermore, FLEX and POLAR are two characteristics that define a compound's bioavailability. FLEX is defined by rotatable bonds (rotatable bonds greater than ten indicate limited oral bioavailability), whereas polarity is determined by topological polar surface (TPSA). The result showed that the number of rotatable bonds was 0 and TPSA 34.14 Å².

Metabolism was also estimated by determining the effect of guggulsterone on inhibition of the main cytochromes (CYP1A2, CYP2C19, CYP2C9, CYP2D6 and CYP3A4) among the P450 superfamily. CYP2C19 is involved in the detoxification of potential carcinogens and CYP2C9 is the major enzyme implicated in drugs metabolism. According to Manikandan and Nagini (2018), the inactivation of CYPs may increase toxicity or decrease the efficiency of a potential drug. Guggulsterone was predicted to likely inhibit CYP2C19 and CYP2C9 and cannot inhibit CYP1A2, CYP2D6, and CYP3A4.

The liver is involved in the biotransformation of a variety of xenobiotics and drugs. Our results revealed that guggulsterone doesn’t induce hepatotoxicity, cytotoxicity, and mutagenicity but may cause immunotoxicity and carcinogenicity.