Explore on screening COX-2 inhibitors from the essential oil of Solanum lyratum Thunb. By molecular docking and molecular dynamics simulation

This study aimed to investigate Solanum lyratum Thunb. with respect to the potential ingredients with anti-inflammatory activity from its essential oil by silico study. To this regard, the essential oil of Solanum lyratum Thunb. was extracted by hydrodistillation. 25 compounds were identified by GC-MS. Using virtual screening, molecular docking and molecular dynamics simulation of the 25 identified compounds, the ones showing anti-inflammatory activity on COX-2 were identified. According to the drug-like principle and the prediction of ADEMT properties, the six compounds of Spathulenol, Cedrol, Juniper camphor, Santalol, Nootkatone and 7,9-Di-tert-butyl-1-oxaspiro[4.5]deca-6,9-diene-2,8-dione were identified and then studied for molecular docking, and based on which the top two compounds of binding free energy were studied by the molecular dynamics simulation. The molecular docking data indicated that the binding free energies of Spathulenol, Cedrol, Juniper camphor, Santalol, Nootkatone and 7,9-Di-tert-butyl-1-oxaspiro[4.5]deca-6,9-diene-2,8-dione to COX-2 protein were −5.65, −7.19, −6.35, −4.94, −5.82 and −5.14 kcal/mol, respectively. The findings showed the steady interactions of hydrogen bonds and hydrophobic bonds between both the top two compounds of binding free energy and the active site residues of COX-2 (4M11) throughout the simulation via hydrogen bonds and hydrophobic bonds. The very study shall be supportive for in vitro and in vivo studies in developing drug products using the lead bioactive ingredients for anti-inflammatory in the future.


Introduction
Cyclooxygenase (COX) is an enzyme playing critical part in generating prostaglandins and leukotrienes from arachidonic acid.Selective inhibition of COX can be used to treat many diseases, such as antipyretic, analgesic and anti-inflammatory [1][2][3][4][5].As an inducible enzyme, the expression of COX is rarely found in normal cells, but high expression levels are always found in the course of inflammation.It mediates the expression of pro-inflammatory mediators and cytokines, thereby playing a vital part in the pathological process of inflammation [6][7][8].Therefore, developing medicines that can effectively inhibit COX-2 will be of great significance for treating inflammatory diseases.
The search for effective compounds from natural drugs has gained more popularity in recent years.Solanum lyratum Thunb., which belongs to the Solanaceae family, has been utilized as a Chinese medicinal material more than 2000 years.Studies have confirmed that Solanum lyratum Thunb.contains rich chemical components, such as flavonoids, terpenoids, organic acids, saponins, essential oils, and has a extensive range of pharmacological effects, including immune regulation, antibacterial, anti-inflammatory, anticancer, antioxidant and so on, and is clinically used for treating rheumatoid arthritis, hepatitis, cancer, malaria, cold and other diseases [9][10][11][12][13][14].The essential oil of Solanum lyratum Thunb. is a very important class of compounds, mainly including monoterpenes, sesquiterpenes and their oxides.Modern studies have found that essential oil usually has better anti-inflammatory and antioxidant effects.However, the actually effective chemical components contributing to the better anti-inflammatory effects of the essential oil remain unknown.Herein, using virtual screening technology, the interactions between identified chemical components of the essential oil from Solanum lyratum Thunb.and COX-2 were simulated and calculated.Screening anti-inflammatory lead compounds to improve the efficiency of new drug discovery [15][16][17][18].
Herein, the essential oil of Solanum lyratum Thunb.was prepared to identify the chemical components with anti-inflammatory effects and also to study the underlying mechanism.The essential oil of Solanum lyratum Thunb.was prepared by hydrodistillation, then the compounds in the essential oil were analyzed and identified using GC-MS technique, and the chemochemical characteristics of the identified compounds were investigated in silico study.Finally, the compounds that meet the principle of drug-like and ADMET characteristics were used for molecular docking and molecular dynamics simulation to screen the lead compounds that can effectively inhibit COX-2, providing important theoretical basis for in vivo and in vitro experimental studies in the future.

Plant materials
The grass of Solanum lyratum Thunb.was sourced from Zhangshu, (Jiangxi province, China) in June 2023, and has been identified by a professor from Jiangxi University of Chinese Medicine in accordance with the Pharmacopoeia of the People's Republic of China.The voucher specimen (XF20220306002) was deposited also at the Jiangxi University of Chinese Medicine.After being completely dried, the collected grass of Solanum lyratum Thunb.were ground into powders, and then the powders were sifted using 40 mesh sieve, and finally the sieved materials were packaged in well-closed PE bags before further use.

Essential oil Extractions
The powder (100g) of the dried grass of Solanum lyratum Thunb.was subjected to 4-h hydrodistillation using a clevenger apparatus to prepare the essential oil, afterward, the resultant was dried with addition of anhydrous sodium sulfate, then sealed in dark vials, and stored at 4 • C before further use.

Gas chromatography-mass spectrometry (GC-MS)
The sample were analyzed by Agilent 7890A GC system and 5975C MS system using a column packed with HP-5 MS (5 % phenylmethylpolysiloxane, 30 m × 0.25 mm × 0.25 μm; Agilent, America).For the test conditions, the oven temperature was initially maintained at 50 • C for 2 min and then ramped to 250 • C by 4 • C/min, and finally maintained at 250 • C for 15 min; the injector port temperature was 250 • C; the transfer interface and source temperature was 250 • C; the flow rate of carrier gas (He) was 1 mL/min; the electron ionization source was 70 ev; the scanning was performed at a rate was 0.5 s (cycle time: 0.2 s) within50-400 amu.The various compounds were identified through a comparison of between the mass spectra of detected compounds and the NIST Mass Spectral library (NIST 08).The relative contents (%) of the volatile components were calculated by the area normalization method.

Compounds database
The compounds of the essential oil from the dry above-ground plant of Solanum lyratum Thunb.were prepared with the 3D SDF files downloaded from the PubChem database [19,20].

Drug-likeness properties
The predictions of drug-like properties for the identified compounds were performed with the SwissADME online database.By H. Xiao et al. entering the SMILES notations sourced from the PubChem database for the identified compounds into SwissADME, the drug-like properties of important molecular properties were calculated.Then in accordance with the criteria followed in drug design, screening with respect to drug likeliness was performed for the compounds.The major criteria considered in this study were the Lipinski rule, Veber's Law, molecular weight, number of hydrogen donors, number of hydrogen acceptors, number of rotatable bonds, and so on [21][22][23][24][25][26].

Bio-activity scores prediction
The calculation of bio-activity scores for the compounds were completed by online tool-Molinspiration.The important parameters include GPCR ligands, protease inhibitors, kinase inhibitors, ion channel modulators, nuclear receptors, and enzyme inhibitors.Based on this, the compounds with positive enzyme inhibitor scores were selected for subsequent studies.The CLogP value of a compound is an important parameter to measure its hydrophilicity [27,28].Herein, the calculation of CLogP values for the selected compounds was completed using the ChemBioDraw Ultra 14.0 software.

Toxicity prediction
As an important parameter for evaluation of drug safety, toxicity is always of high concern.In this study, ProTox-II webserver was searched for the toxicity of the selected compounds, and then the compounds were classified regarding their toxicity according to the LD50 data [29].

Protein preparation
The high-resolution (2.45 Å) X-ray crystal structure of the chosen enzyme target of COX-2 protein (PDB ID: 4M11) was sourced from RCSB PDB database.The chosen enzyme targets for molecular docking were prepared by removing hetero molecules and water molecules using Pymol 1.8.The A chain of the target protein was selected for molecular docking study.The prepared proteins were saved as PDBQT format after adding hydrogen atoms and charges [30][31][32].

Ligand preparation
The selected compounds in the essential oil, which conforms to the drug-like criterions, positive enzyme inhibitor scores and toxicity classes 4 or above, were treated with PMV 1.5.6 and saved as PDBQT for molecular docking.

Molecular docking
Using PMV 1.5.6, the selected compounds were investigated for their molecular interactions to the target protein (COX-2) were studied using PMV 1.5.6.In the process of molecular docking, the ligands and target proteins were defined as flexible and rigid molecules, respectively.Herein, the method adopted was blind docking.The grid boxes covers the entire surface of the proteins to ensure docking with the selected ligands.The docking was performed for 50 runs, and the molecular interactions were visualized with Pymol 1.8.For the molecular dynamics simulation of the ligands, the lowest energy conformation were used as input [33,34].

Molecular dynamics simulation
The molecular dynamics simulation can elucidate the conformation change of the ligand-receptor complexes under simulation conditions.In this study, the simulation was performed by GROMACS 2020.6 package implemented with CHARMM36 force field for the complexes of the top two compounds of binding free energy (Cedrol and Juniper camphor) with target protein (4M11).First, the topology files of receptor and ligand were prepared respectively, and the topology file of ligand-receptor complex was generated.Then define the unit cell and add solvent and ions to the unit cell to make the whole system neutral; the simulation was carried out after optimizing the system to minimize the system energy and balancing the ligand-receptor complex.After the simulation, the ligandreceptor interaction and ligand dynamics were analyzed [35,36].

Analysis of chemical composition
The yield of essential oil from Solanum lyratum Thunb.was (0.84 ± 0.25)% (w/w).The TIC chromatogram of the essential oil is presented in Fig. 1, and its chemical composition by HD was determined and provided in Table 1, from which, it is known that there are 25 chemical components were identified from the essential oil by GC-MS, including mainly sesquiterpenes, oxidized sesquiterpenes, monoterpenes and fatty acid.

ADMET/drug-likeness properties and bio-activity score prediction of selected ligands
Using Swiss ADME, the prediction of drug-likeness properties was performed for the 25 identified chemical compounds.The method was able to remove compounds without obvious drug-likeness properties.The rules should be followed during the druglikeness properties screening process: Lipinski rule, Veber rule, MW < 500, number of hydrogen donors (≤5), hydrogen acceptors (≤10) and rotatable bonds (≤10), as well as total polar surface area ≤140 Å 2 .The calculation of CLogP values for the compounds were performed with ChemBioDraw Ultra 14.0 software, and it is negatively correlated with the water permeability.The compounds with a CLogP value of <5 were selected for subsequent analyses.Therefore, according to the results of ADME and CLogP value, 10 compounds out of 25 compounds conform to the drug-likeness criterions.The physical-chemical properties are presented in Table 2, and the ADME properties are presented in Table 3.
Using the ProTox-II webserver, the prediction of toxicity was performed for the selected compounds.The compounds of LD50 category IV/V/VI, which were not hepatotoxic, carcinogenic, immunotoxic, mutagenic and cytotoxic, were selected for subsequent molecular docking analyses.The results showed that among the 10 compounds, except Caryophyllene oxide and 4-Bromo-1naphthylamine, the other 8 compounds were less toxic.The toxicity prediction results of the compounds are presented in Table 4.
In general, it is assumed that a compound is considered as active, moderately active, and inactive for a bioactivity score of >0, − 0.5-0, and < − 0.5.In this study, the online website Molinspiration was used to score the biological activities of compounds, and compounds with a positive enzyme inhibitor score were selected for subsequent studies.The results showed that 6 compounds had enzyme inhibitor scores greater than zero, namely Spathulenol, Cedrol, Juniper camphor, Santalol, Nootkatone and 7,9-Di-tert-butyl-1-oxaspiro [4.5]deca-6,9-diene-2,8-dione.Therefore, they were selected for subsequent molecular docking studies.The bio-activity Table 3 ADME properties and the drug-likeness of 10 compounds conform to the drug-likeness criterions.

Table 4
The toxicity prediction results of the selected compounds.scores of the selected compounds were shown in Table 5.
The interaction between ligand and protein mainly includes hydrogen bond interactions, hydrophobic bond interactions, Vander Waal interactions and so on.The interaction between the compound and COX-2 protein was shown in Fig. 2. From Fig. 2A-F, it can be seen that Spathulenol (2A) had hydrogen bonding with GLN-372, LYS-532 and had hydrophobic bonding with PHE-371 and LYS-532;

Molecular dynamics simulation
For Cedrol and Juniper camphor, which showed top two binding free energy docked ligand-receptor complexes, Cedrol and Juniper camphor, the molecular dynamics simulations were performed.Fig. 3 showed the RMSD and RMSF between the COX-2 protein (4M11) receptor with Cedrol (Fig. 3A) and Juniper camphor (Fig. 3B), respectively.

Root mean square deviation (RMSD)
The parameter of RMSD represents the stability of conformation for the ligand-receptor complex under the conditions during simulation.The ligand-receptor complex's 100-ns trajectory can be seen in Fig. 3A.As seen in the figure, the molecular dynamics simulation does not result in large variations of the complex, and the average RMSD of all ligand-receptor complexes was around 0.2 nm.The stability of the ligand-receptor complexes was demonstrated by the absence of significant deviations on the trajectory diagram within 100 ns.

Root mean square fluctuation (RMSF)
Molecular dynamics simulations reveal the flexibility of proteins through RMSF.The protein's flexibility decreases after the drug binds to it, which leads to its role as a stabilizing agent.Fig. 3B shows the RMSF of the 500 residues trajectory of the ligand-receptor complex.Fig. 3B shows that the entire sequence segment of the protein has a low RMSF value.The fluctuation of most amino acid residues fluctuate were within 0.25-1.25 nm (0.5-2.0 nm of fluctuation at 160-190 locations), this indicating that the compound can  form a stable complex with the key amino acids of the protein, which is conducive to the interaction of ligand and receptor.

Conclusion
The present silico study aims to screen out phytochemical as lead compounds from the essential oil of Solanum lyratum Thunb., which may be potential potent ingredients in inhibiting COX-2 protein (4M11).In this study, 25 phytochemical compounds were identified from the essential oil of Solanum lyratum Thunb.by GC-MS.For these compounds, the drug-like properties were predicted and then analyzed in accordance with Lipinski's rules, Veber's rules, toxicity rules and bioactivity rules for oral drugs.As a result, Spathulenol, Cedrol, Juniper camphor, Santalol, Nootkatone and 7,9-Di-tert-butyl-1-oxaspiro [4.5]deca-6,9-diene-2,8-dione were identified as potential lead phytochemical compounds for treatment of inflammatory diseases; and these compounds were also proved to show strong binding affinities with 4M11 by the molecular docking analyses.Furthermore, the docked complexes of these compounds were demonstrated to be stable under the conditions during MD simulation.The present work lays down the basis for these theoretical studies and offers insight into the understanding of in vitro and in vivo tests, in the future.

H
.Xiao et al.

Table 1
Chemical composition of essential oil from Solanum lyratum Thunb.

Table 2
Physical-chemical properties of 10 compounds conform to the drug-likeness criterions.

Table 5
Bio-activity scores of the selected compounds.