Amomum tsao-ko essential oil, a novel anti-COVID-19 Omicron spike protein natural products: A computational study

2.1 Plant materials

Amomum tsao-ko were purchased from Yunnan Province, China and authenticated by Prof. Chunchun Zhang from Zhejiang Chinese Medical University. The dried samples were powdered (0.30 mm) and deposited at 4 °C until use. Ultrapure water was manufactured by Milli-Q IQ 7000 purification system (Merck, Germany). Chromato-graphic grade n-hexane was purchased from Merck (Merck, Germany).

2.2 Extraction procedure

Hydrodistillation was applied for the extraction of A. tsao-ko essential oil (AEO). with a traditional Clevenger apparatus. An aliquot of 200 g material was introduced into the distillation reaction flask with 2.4 L of water soaking for 30 min before extraction and then extracted for 5 h until the essential oil volume was kept constant. The essential oil was collected in the receiving flask with a circulating water-cooling system and transferred to an amber-colored vial, dehydrated with anhydrous sodium sulphate. Finally, the essential oil was kept at 4 °C with N₂ until being used. The experiment was performed in triplicate.

2.3 GC/Q-TOF MS analysis of essential oil

GC/Q-TOF MS analysis was performed on an Agilent quadrupole time of flight mass spectrometer (Agilent Technology Co., LTD, USA) consisting of an Agilent 7890B gas chromatography and an Agilent 7250 time of flight mass spectrometer, using an HP-5MS capillary column (30 m × 250 μm × 0.25 μm film thickness). The operational conditions of GC/Q-TOF MS analysis were preset as follows: helium was used as the carrier gas at 1.0 mL/min flow rate; split ratio 1:20; injection volume 1 μL; injection temperature 230 °C; ion source temperature 230 °C; interface temperature 280 °C; oven temperature progress: 60 °C (holding for 2 min), 60–120 °C (6 °C/min, holding for 2 min), 120–180 °C (5 °C/min), 180–280 °C (30 °C/min); ionization mode used at electronic impact 70 eV; mass range 29–500 m/z; TOF-scan mode at 5 spectra/s; solvent delay 4 min. The mass spectra of essential oil compositions in the NIST14 mass spectral library and the fragmentation patterns of the mass spectra with the data published in the literature were applied for the identification of essential oil compositions.

2.4 Homology modeling of the spike protein

Homology modeling is a highly-accurate method for the prediction and construction of the 3D structure of a protein-based on another protein with a known 3D structure (template). Using the SWISS-MODEL interactive server (Waterhouse et al., 2018), the 3D structures of Omicron mutant spike protein RBD and NTD domains were constructed based on the structures of wild-type SARS-CoV-2 spike protein RBD (PDB ID: 7EKG; 2.63 Å resolution) and NTD (PDB ID: 7BZ5; 1.84 Å resolution) with the sequence identity of 93.30% and 96.31% retrieved from PDB database (https://www.rcsb.org), respectively. The Ramachandran plot and Molprobity scores showed that the built structures were of good quality and suitable for further studies.

2.5 Fpocket analysis

The built structure of RBD_Omicron was analyzed by the Fpocket webserver to predict possible active centres (AC) on the protein surfaces. Fpocket is an open-source pocket detection package based on Voronoi tessellation and alpha-spheres (Schmidtke et al., 2010). The PDB file of the RBD_Omicron was submitted to Fpocket, and the outputs were analyzed using UCSF Chimera software. Binding sites were the distribution of surrounding residues in AC and acted as the active residues.

2.6 Molecular dynamics simulation

Molecular dynamics (MD) simulations were performed using GROMACS 2018.8 on Intel(R) Core (TM) i7-10870H CPU@ 2.20 GHz, using the OPLS-AA/L all-atom force field to treat the protein (Abraham et al., 2015). RBD was then immersed in a pre-equilibrated octahedral box of TIP3P water and the system was neutralized. The system was then minimized using the energy gradient convergence criterion set to 0.01 kcal/mol Ų in four steps involving: (i) an initial 5,000 minimization steps (2,500 with the steepest descent and 2,500 with the conjugate gradient) of only hydrogen atoms, (ii) 20,000 minimization steps (10,000 with the steepest descent and 10,000 with the conjugate gradient) of water and hydrogen atoms, keeping the solute restrained, (iii) 50,000 minimization steps (25,000 with the steepest descent and 25,000 with the conjugate gradient) of protein side chains, water and hydrogen atoms, (iv) 100,000 (50,000 with the steepest descent and 50,000 with the conjugate gradient) of complete minimization. Successively, the water, ions and protein side chains were thermally equilibrated in three steps: (i) 5 ns of NVT equilibration with the Langevin thermostat by gradually heating from 0 K to 300 K, while gradually rescaling solute restraints from a force constant of 10 to 1 kcal/mol Å2, (ii) 5 ns of NPT equilibration at 1 atm with the Berendsen thermostat, gradually rescaling restraints from 1.0 to 0.1 kcal/mol Å2, (iii) 5 ns of NPT equilibration with no restraints. Finally, a production run of 200 ns was performed using a timestep of 2 fs. The SHAKE algorithm was used for those bonds containing hydrogen atoms in conjunction with periodic boundary conditions at constant pressure and temperature, particle mesh Ewald for the treatment of long-range electrostatic interactions, and a cutoff of 10 Å for nonbonded interactions.

2.7 Molecular docking

The computational molecular docking of hACE2 (PDB ID: 6 M17.B, receptor) with Omicron mutant spike protein RBD (ligand) was explored by an information-driven flexible docking approach using HADDOCK and Autodock Vina software. Before docking, the ligand and receptor are dehydrated, hydrogenated, and charged. In docking, the receptor was considered semi-rigid while the ligand was flexible, and the grid matrix of the docking area was sized to ensure complete coverage of the protein molecule. The size of the docking range was set to (40,60,60), and the docking centre was (−32,27,17). The best Autodock Vina score with the suitable energetically favored conformations was used for further analysis. The interaction of AEO compositions with amino acid residues of omicron mutant spike protein RBD was analyzed and represented using ChimeraX.

3.1 Chemical compositions of the essential oil

The chemical compositions of AEO were determined and listed in Fig. 2 and Table 1. And twenty components with a relative area percentage of more than 1% were identified in AEO representing 92.82% of the total AEO constituents. The total percentage of monoterpenes was the most abundant compositions in AEO (77.73%), including monoterpene hydrocarbons (14.57%) of α-Phellandrene, β-Pinene, α-Pinene, 3-Carene and oxygenated monoterpenes (63.16%) of 1,8-Cineole, trans-Citral, trans-2-Decenal, α-Terpineol, cis-Citral, trans-Geraniol, Geraniol acetate, (-)-4-Terpineol, Linalool, (-)-cis-Sabinol. The oxygenated monoterpene compounds possessed higher dipolar moments in essential oils and interacted more vigorously with heating, and they can be extracted more easily than monoterpene hydrocarbons.

Close

Fig. 2

Total ion flow diagram of AEO.

Export to PPT

Due to the rich essential oil components, Chinese herbal aromatic plants not only have a special aroma, but also have the effect of medical care that are widely used in many industries. Essential oil can cause aroma movement and regulate qi², and have the functions of regulating emotions, coordinating human physiological functions, calming and improving human immunity. Chinese herbal aromatic essential oil can improve the therapeutic outcome of metabolic syndrome by interfering with insulin signal transduction, resisting oxidative stress, reducing the inflammatory response and increasing insulin sensitivity. These results suggest that Chinese herbal aromatic essential oil has the advantages of high efficiency, low toxicity, multi-target and multi-level therapy, and has broad research prospects in the future (Tang et al., 2021). As the predominant component in AEO, 1,8-cineole is an oxygenated monoterpene with a fragrance, camphoraceous features and a spicy flavor. Therapeutic applications of 1,8-cineole for the treatment of respiratory diseases, muscle pain, neurosis, rheumatism and kidney stones have been investigated as well (Cai et al., 2021). As an important part of natural essential oil, traditional Chinese medicine essential oil has a wide variety of sources and has high medicinal, industrial and edible values.

3.2 Mapping of RBD_Omicron mutations compared with RBD_WT

Compared with RBD_WT, RBD_Omicron had fifteen additional mutation sites including G339D, S371L, S373P, S375F, K417N, N440K, G446S, S477N, T478K, E484A, Q493R, G496S, Q498R, N501Y, Y505H (Fig. 3C), respectively. Of the fifteen Omicron changes in RBD, nine map to hACE2 binding footprint: K417N, G446S, S477N, E484A, Q493R, G496S, Q498R, N501Y, Y505H, which considerably affects receptor recognition, with N440K and T478K just peripheral (Fig. S1, Dejnirattisai et al., 2022). Additional, mutations occur on G339D, S371L, S373P and S375F, the last three of which are adjacent to a lipid-binding pocket (Toelzer et al., 2020). This pocket has been seen occupied by a lipid similar to linoleic acid in an unusually rigid state of spike protein, where all RBDs are found in a locked-down configuration stabilized by lipid-bridged quaternary interactions between adjacent RBDs, and several were previously shown to modulate ACE2 binding and/or antibody evasion (Harvey et al., 2021). Thus, understanding the underlying molecular basis for the enhanced transmissibility, immune resistance, and virological properties of the Omicron variant may facilitate halting this looming crisis.

Close

Fig. 3

Active centres discovered on RBD using molecular dynamics simulation: (A) The active centre on RBD, (B) RMSF values of RBD on wild-type (WT) and Omicron mutant, (C) amino acid sequence of RBD on wild-type (WT) and Omicron.

Export to PPT

As the Omicron variant showed a considerable number of positively charged amino acids at sites of mutation, the electrostatic potential at the RBD-ACE2 interface and isoelectric point (pI) change of mutation sites on RBD were analyzed. It can be found that the increased pronouncedly electro positivity in the Omicron variant (T478K, Q493R, Q498R and Y505H), and also lost a positively charged amino acid at the interaction interface of K417N (Fig. S2). Combined with existing studies, the largest difference between RBD_WT and RBD_Omicron was found at positions 417 and 493 (Socher et al., 2021). For the Omicron variant that exchanged K417 to N417, a strongly reduced interaction with D30, while Q493 mutated to R493 reduced the interaction with K31, which may be due to electrostatic repulsion between the positively charged R and K. The pI of protein is the pH value at which its surface is completely charged but its net charge is zero. Also, the amino acid pI of a protein can determine the local electric charge of the protein, which is an important basis for finding targets in drug design (Ren et al., 2020a; Xia 2021). A pI value of more than 7 indicates that the protein or amino acid is alkaline, whereas a value less than 7 indicates that it is acidic. At present, there are few studies on pI of RBD mutation of SARS-CoV-2 spike protein (Krebs et al., 2021). In this study, it is considered that the change of pI has a great influence on the mutation of SARS-CoV-2, which is a problem neglected by the scientific community at present. The theoretical pI of the RBD_WT and RBD_Omicron variant were 8.38 and 8.73, indicating that the Omicron variant was expected to have more alkaline pI. As shown in Table 2, the pI of RBD_Omicron mutation sites greater than 7 included N440K, T478K, Q493R, Q498R and Y505H. In addition to steric hindrance, the number of basic amino acids also affects the binding affinity between RBD and hACE2. Studies have found that in the omicron virus multiplication process, supplying excessive basic amino acids is required (Hanai, 2022). Hence, the transmissible strength of COVID-19 variants depended on the mutated basic amino acids that can enhance the binding affinity with ACE-2.

The instability index of less than 40 indicates that the protein structure is stable. In our study, the range remained 20.43–28.51, indicating the great stability of RBD_WT and RBD_Omicron. Aliphatic indexes of 70.31 and 73.21 for RBD_WT and RBD_Omicron indicated that the protein was temperature stable across a wide temperature range. The greater a protein’s aliphatic index, the more thermostable it was. A protein with a low grand average of hydropathicity (GRAVY) value is nonpolar and has a stronger affinity for water, indicating that it is intrinsically hydrophilic. The higher amino acid composition of F, L, I and P in RBD_Omicron, when compared with RBD_WT, suggested that RBD_Omicron included more hydrophobic amino acids. The amino acid composition of RBD_Omicron was low in polar amino acids such as Q, S and E, while was high in nonpolar amino acids such as L, F and P. These results indicated that the Omicron variant was more hydrophobic and more easily combined with hACE2, which played an important role in increasing its transmissibility.

3.3 Molecular dynamics simulation results

The final predicted 3D structure model of RBD_Omicron was shown in Fig. 3A. Although there were many active pocket mining tools, there was no statistically significant difference in structure type and tool performance (Clark et al., 2020). These tools could provide relatively accurate prediction results for protein active pockets regardless of the presence of ligands. Some researchers used Fpocket as an important tool to analyze the protein-ligand binding pocket, screened the active pocket of methionyl-tRNA synthetase, and proved it to be a natural inhibitor of the lead compound with inhibitory ability in subsequent docking experiments (Rowaiye et al., 2022). Similar studies have also reported that Fpocket can be used as a well-known tool to mine active pockets and predict specific action targets (active pockets) in both the structure and protein of COVID-19, accurately predict active pockets and provide the necessary basis for the accurate positioning of amino acids (Gervasoni et al., 2020). All of these are inseparable from the accurate prediction ability of Fpocket. Therefore, Fpocket was applied to detect the pockets on RBD_Omicron surface that were suitable to bind small molecule ligands. It can be found that there were three active centres (AC) on RBD_Omicron surface by Fpocket analysis, considered as ligand binding sites in this protein. Important amino acid residues presented in these three AC were determined as AC1 (orange red) including R355, Y396, P426, D428, F429, T430, F464, S514, F515, E516, L517, AC2 (forest green) including D339, F342, N343, V367, L371, P373, F374, W436, AC3 (cornflower blue) including Y453, L454, F456, R457, D467, S469, E471, I472, Y473, Y491. Since AC1 was the part connected with NTD, the remaining ACs were mainly used for molecular docking with small molecule ligands in subsequent experiments. It is worth noting that among the fifteen mutations of RBD_Omicron, four mutations, namely G339D, S371L, S373P and S375F, were located at AC2, while the remaining eleven mutations were located at or near AC3, which was just on the interaction interface with hACE2.

Topological domain-specific average backbone root mean square deviation (RMSD) value of RBD_Omicron was less than that of RBD_WT as shown in Fig. S3. For RBD_WT, the structure ensembled from simulation (100 ns) deviated from the backbone of RBD_WT by 0.30 Å on average, while the RBD_Omicron possessed averaged RMSD value of 0.29 Å, indicating that the mutations only slightly altered the structure of RBD_Omicron from the RBD_WT. Hence, the mutations in the Omicron variant did not significantly reduce the RBD stability, instead, the hACE2-RBD_Omicron complex was even slightly more stable than hACE2-RBD_WT (Lupala et al., 2022). Molecular dynamics was used to calculate the root mean square fluctuations (RMSF) value of amino acids. RMSF means the degree of freedom of each amino acid, which can prove to some extent which amino acid residues are more active in space. The residue fluctuations were analyzed by calculating the RMSF of RBD_Omicron and RBD_WT (Fig. 3B). According to the average values of RMSF, the RBD_Omicron (0.75 Å) was more rigid than RBD_WT (0.80 Å). The reduction of RMSF was more pronounced at the interfacing residues of RBD, also known as the receptor-binding motif (Yi et al., 2020).

3.4 In silico molecular docking analysis

HADDOCK (high ambiguity-driven protein-protein docking) algorithm was applied to perform the restraint docking of WT, Delta and Omicron variants with hACE2 receptor by defining the interface residues as reported previously (Koukos et al., 2021; Khandia et al., 2022). The docking scores for RBD_WT-hACE2 and RBD_Delta-hACE2 complex as −49.4 kcal/mol and −64.3 kcal/mol, while the docking score for RBD_Omicron-hACE2 complex was −141.4 kcal/mol (Table 3). The data showed that the reported mutations have increased the binding ability of RBD_Omicron to hACE2 than RBD_WT and RBD_Delta, which was similar to the results previously reported (Khan et al., 2021). Furthermore, a comparison of van der waals, electrostatic and desolvation energy values also revealed considerable differences among WT, Delta and Omicron variants. For RBD_WT and RBD_Delta, the van der waals energies were −18.1 kcal/mol and −24.0 kcal/mol, while RBD_Omicron had a van der waals energy of −55.2 kcal/mol. The electrostatic energies for RBD_WT, RBD_Delta and RBD_Omicron were −114.9 kcal/mol, −213.6 kcal/mol and −343.6 kcal/mol, respectively. Both the van der waals energy and electrostatic energy of RBD_Omicron were increased, and higher electrostatic energy was considered the primary reason for enhanced binding (Khan et al., 2022).

The interactions at the interface of the ACE2-RBD complex for WT have been previously reported in the perspectives of both static crystal structures and dynamical conformations (Wang et al., 2020; Lupala et al., 2021). Cryo-EM showed that eleven of the fifteen mutations in RBD_Omicron (K417N, N440K, G446S, S477N, T478K, E484A, Q493R, G496S, Q498R, N501Y, Y505H) were located at ACE2-RBD interface, consequently changing the electrostatics surface changes at the interface. As can be seen from Fig. 4 the salt bridge E484-K31 and K417-D30 of RBD_WT-ACE2 were lost in the Omicron variant, while hydrogen bonds Q493-E35, Q498-K353 and Y505-E37 were enhanced by the Omicron mutations, forming new interactions, such as R498-Y41, R498-Q42 and H505-K353. Also, an additional π-π stacking interaction Y501-Y41 was introduced. The key interactions Y449-D38, Y453-H34, A475-S19, N487-Y83, T500-N330 and T500-D355 in RBD_WT-ACE2 were preserved in the Omicron variant. Cryo-EM structural analysis of the RBD_Omicron-hACE2 complex showed that new salt bridges and hydrogen bonds were formed at R493, S496 and R498 mutations, which may compensate for other Omicron mutations, such as K417N which is known to reduce the affinity between ACE2 and S protein (Mannar et al., 2022). The interactions R493-E35 have previously been shown to be responsible for the anchor locking and correct orientation of the RBD_Omicron when binding to hACE2. Based on docking studies, the Q493R, N501Y, S371L, S373P, S375F, Q498R and T478K mutations contributed significantly to high binding affinity with hACE2. Analysis revealed that mutations at N440K, T478K and N501Y sites may impart ten times more infectivity to WT (Thakur and Ratho, 2022). The differences in binding pattern between the WT and the Omicron variant complexes revealed that the key mutations are responsible for enhanced binding and the consequent transmissibility and infectivity. Dejnirattisai et al., 2022 found Omicron variant caused widespread escape from neutralization by serum obtained following infection by a range of SARS-CoV-2 variants, meaning that previously infected individuals will have little protection from infection with Omicron.

Close

Fig. 4

The interactions between RBD-hACE2 of the Omicron variant.

Export to PPT

Since the outbreak of COVID-19, several large-scale mutations have all been carried out around spike proteins, among which RBD mutation is the most common. The evolution direction of this virus is dependent on transmission from human to human or from animal to human, through the host immune system to complete the screening (Kupferschmidt, 2021). In short, the pandemic’s weakness has led to the evolution of the SARS-CoV-2, which is centred on spike proteins, particularly the RBD part (Sanjuán and Domingo-Calap, 2016). Traditional Chinese medicine has been playing an important role in the prevention and treatment of COVID-19 from the early stage to the present. Studies have shown that Chinese medicine components have good inhibitory capacity on both structural and non-structural proteins of SARS-CoV-2. The study of Chinese medicine components by network pharmacology is one of the most widely used methods to explore the inhibition of SARS-CoV-2 binding, infection and replication from Chinese medicine at present (Ren et al., 2020b). Although this research method has some limitations, it can also provide the best binding skeleton in a specific active centre and provide a good research strategy for drug design and drug screening (Sahlan et al., 2021; Dotolo et al., 2021). Traditional idea of Traditional Chinese medicine is to prevent and control the spread of diseases before they get sick through systematic regulation, which is highly consistent with modern epidemiological views (Zhang et al., 2021).

Based on the theory of Traditional Chinese medicine, the research idea of blocking the binding of viruses to receptors is more valuable than the discovery of inhibitors of key enzymes of viruses. Therefore, this study focuses on the study of the herbal aromatic plant Amomum tsao-ko. It can be found that all components in AEO can interact with RBD and NTD by molecular docking the components with a relative area percentage of more than 1% with RBD_WT, RBD_Delta, RBD_Omicron, NTD_WT and NTD_Omicron. The binding energy between AEO components and RBD_Omicron varied from −7.0 kJ/mol to −4.5 kJ/mol. A more negative value of binding energy implies stronger binding efficacy (Kar et al., 2020). And the binding sites of most AEO components were located at AC2 and AC3 or near the interface. Among them, EO02, 03, 08 and 17 with high relative content and strong binding energy with RBD (Fig. 5 and Table S1). Meanwhile, RBD is the earliest binding region of COVID-19 with the receptor. It was imperative to research spike protein, especially on how to reduce its binding ability with hACE2 in the epidemic prevention of COVID-19. The binding energy of RBD and hACE2 was between −9.8 and −14.0 kcal/mol (COVID-19 spike-host cell receptor GRP78 binding site prediction), which was a relatively strong binding ability. In current studies, the binding energy between no matter the known drug molecules or the drug molecules obtained by virtual screening cannot reach less than −9.8 kcal/mol, that is, the minimum binding energy of RBD and hACE2 (Shah et al., 2020; Cavasotto and Filippo, 2021), indicating that it was difficult to design the inhibition of spike protein to inhibit viral infection. Therefore, this study focused on reducing the ability of the virus to spread in the air and contact with the host, which was consistent with the results of relevant literature and could provide proof for this study to a certain extent (Abdelli et al., 2021; Kumar et al., 2020; Yu et al., 2020; Naeem et al., 2020; Xiong et al., 2020). Virtual screening showed that most components of AEO could interact with AC2 and AC3 of RBD_Omicron, indicating that the interaction between AEO and RBD was a synergistic effect of multiple targets (Fig. 5, Fig. 6 and Fig. S4-12). Spike protein RBD is the most externally exposed part of the SARS-CoV-2 structure. If RBD contact with AEO molecules in the air, the two can interact directly and form a relatively stable complex, thus protecting the host to a certain extent (Wang et al., 2021b). The essential oil is volatile itself and can interact with SARS-CoV-2 spike protein RBD in the air to inhibit the binding ability of RBD and hACE2, thus achieving epidemic prevention effect.

Close

Fig. 5

Molecular interactions of AEO components against RBD_Omicron (A) and the heatmap of the binding energy of AEO to RBD and NTD (B).

Export to PPT

Close

Fig. 6

Surface and maps of AEO components docked into binding sites of RBD_Omicrion different active centres.

Export to PPT