Essential Oils as Antiviral Agents, Potential of Essential Oils to Treat SARS-CoV-2 Infection: An In-Silico Investigation

Essential oils have shown promise as antiviral agents against several pathogenic viruses. In this work we hypothesized that essential oil components may interact with key protein targets of the 2019 severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). A molecular docking analysis was carried out using 171 essential oil components with SARS-CoV-2 main protease (SARS-CoV-2 Mpro), SARS-CoV-2 endoribonucleoase (SARS-CoV-2 Nsp15/NendoU), SARS-CoV-2 ADP-ribose-1″-phosphatase (SARS-CoV-2 ADRP), SARS-CoV-2 RNA-dependent RNA polymerase (SARS-CoV-2 RdRp), the binding domain of the SARS-CoV-2 spike protein (SARS-CoV-2 rS), and human angiotensin−converting enzyme (hACE2). The compound with the best normalized docking score to SARS-CoV-2 Mpro was the sesquiterpene hydrocarbon (E)-β-farnesene. The best docking ligands for SARS−CoV Nsp15/NendoU were (E,E)-α-farnesene, (E)-β-farnesene, and (E,E)−farnesol. (E,E)−Farnesol showed the most exothermic docking to SARS-CoV-2 ADRP. Unfortunately, the docking energies of (E,E)−α-farnesene, (E)-β-farnesene, and (E,E)−farnesol with SARS-CoV-2 targets were relatively weak compared to docking energies with other proteins and are, therefore, unlikely to interact with the virus targets. However, essential oil components may act synergistically, essential oils may potentiate other antiviral agents, or they may provide some relief of COVID-19 symptoms.


Introduction
The 2019 severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a newly emerging respiratory illness. The epidemic started in December 2019 in Wuhan, China, and has rapidly spread throughout China and the world and is now a global pandemic. SARS-CoV-2 can be efficiently transmitted among humans and has shown a high degree of morbidity and mortality [1,2]. As of April 20, 2020, the worldwide number of infected individuals was 2,544,792, with as many as 175,694 deaths [3]. There are currently no approved vaccines available for the prevention of SARS-CoV-2 infection and only just recently, remdesivir has received "emergency use authorization" for treatment of COVID-19 in the United States; therefore, there is an urgent demand for potential chemotherapeutic agents to treat this disease. Essential oils have been screened against several pathogenic viruses (Table 1), including influenza and other respiratory viral infections. Influenza is an infectious respiratory disease caused by one of three types of influenza viruses, type A, type B, or type C [4]. The most significant in terms of human morbidity and mortality is influenza virus type A, which is found in several bird and mammal species [5]. Several different serotypes of influenza type A have caused global flu pandemics [6]: H1N1, which caused the Spanish flu in 1918 (40-50 million deaths worldwide) [7] and the swine flu in 2009 [8]; the Asian flu of 1957-1958 (ca. 1.5 million deaths worldwide) was caused by influenza A H2N2 [8]; serotype H3N2 caused the Hong Kong flu in 1968 [9]; and H5N1, which caused the bird flu in 2004 [10]. Influenza virus type B, however, is largely confined to human hosts [11].
One study evaluated the in vitro antiviral effect against influenza type A (H1N1) of commercial essential oils that included cinnamon (Cinnamomum zeylanicum), bergamot (Citrus bergamia), lemongrass (Cymbopogon flexuosus), thyme (Thymus vulgaris), and lavender (Lavandula angustifolia). The oils were tested in the liquid phase at a concentration of 0.3% and in the vapor phase. The oils of cinnamon, bergamot, thyme, and lemongrass displayed 100% inhibition of H1N1 in the liquid phase, while the inhibition for lavender essential oil was 85%. However, in the vapor phase, 100% inhibition was observed only for cinnamon leaf essential oil after 30 min of exposure. The bergamot, lemongrass, thyme, and lavender essential oils displayed inhibition rates of 95%, 90%, 70%, and 80%, respectively [12].
In addition to essential oils, several individual essential oil components have been screened for antiviral activity (Table 2).  Because of the activities of several essential oils and essential oil components against human pathogenic viruses, we hypothesized that essential oil components may be potentially useful as antiviral agents against SARS-CoV-2. In this work, we carried out a molecular docking analysis of the major components of essential oils that exhibit antiviral activity (Tables 1 and 2) with known SARS-CoV-2 protein targets.
RNA-dependent RNA polymerase catalyzes RNA replication from an RNA template and is an essential enzyme in RNA viruses. Because these enzymes are crucial in viral replication, they are viable targets in antiviral chemotherapy [112]. Molecular docking of essential oil components with SARS-CoV-2 RdRp showed only weak docking with this enzyme target ( Table 3). The ligand with the best docking score was (E,E)-farnesol, with DS norm = −89.6 kJ/mol.
The docking results of the essential oil components with the six randomly selected proteins indicate the best docking ligands to SARS-CoV-2 targets (i.e., (E,E)-α-farnesene, (E)-β-farnesene, and (E,E)-farnesol) have better docking energies with other proteins. These three sesquiterpenes have docking energies of −129.8, −122.7, and −133.0 kJ/mol with TcAChE, respectively, and −131.8, −131.8, and −135.6 kJ/mol, respectively, with BaNadD. Indeed, most of the essential oil ligands have better docking properties with one or more of the random proteins compared to the SARS-CoV-2 proteins.
Based on the docking energies of essential oil components with key protein targets of SARS-CoV-2, the individual essential oil components cannot be considered viable chemotherapeutic agents for interaction with the SARS-CoV-2 target proteins. However, essential oils are complex mixtures of compounds and several essential oil components may act synergistically to inhibit the virus. Astani and co-workers have shown, for example, that the antiviral activity (HSV-1) of Eucalyptus oil is much greater than the major component 1,8-cineole, and that tea tree oil has a greater antiviral activity than its components terpinen-4-ol, γ-terpinene, and α-terpinene [52].

Bibliographic Search Criteria
The bibliographic research was performed using the databases Google Scholar, Pubmed, Science Direct, Medline, and Scopus. The keywords applied were "antiviral activity" and "essential oils", "antiviral activity" and "volatile compounds", and "essential oils" and "respiratory diseases".

Ligand Selection
The major components (>5%) of essential oils and pure essential oil components that have been screened against human pathogenic viruses were selected. In the case where enantiomers are known to be natural products, both structures were selected. A total of 171 essential oil components were used in the virtual screening.

Molecular Docking
Each ligand structure was prepared using Spartan '18 v. 1.4.4 (Wavefunction, Inc., Irvine, CA, USA). The lowest-energy conformations of the ligands were determined and used as starting structures in the molecular docking. This is particularly important to include all potential conformations in medium-sized rings where interconversion between conformations may be hindered (e.g., bicyclogermacrene, costunolide, curdione, germacrene D, germacrone, and α-humulene). A total of six protein targets of SARS-CoV-2 from the Protein Data Bank (PDB), represented by a total of 17 structures, were used in the molecular docking, including SARS-CoV-2 main protease (PDB: 5R7Z, 5R80, 5R81, 5R82, 5R83, 5R84, 6LU7, 6M03, and 6Y84), SARS-CoV-2 endoribonuclease (PDB: 6VWW), SARS-CoV-2 ADP ribose phosphatase (PDB: 6W01 and 6W02), SARS-CoV-2 RNA-dependent RNA polymerase (PDB: 6M71), SARS-CoV-2 spike protein binding domain (PDB: 6M0J, 6VX1, 6VW1, and 6M17), and the human angiotensin-converting enzyme (PDB: 6M0J, 6VX1, 6VW1, and 6M17). Molecular docking was carried out using Molegro Virtual Docker v. 6.0.1 (Aarhus, Denmark) as previously reported [128,129]. Briefly, a 15-Å radius sphere centered on the binding sites of each protein structure in order to permit each ligand to search. In the case of the spike protein and human ACE2, the docking sphere was located at the interface between the spike protein and ACE2. In one case, ACE2 was removed and docking was carried out with the spike protein, and in the other case, the spike protein was removed and docking was carried out with ACE2. Standard protonation states of each protein, based on neutral pH, were used, and charges were assigned based on standard templates as part of the Molegro Virtual Docker program. Each protein was used as a rigid model without protein relaxation. Flexible-ligand models were used in the docking optimizations. Different orientations of the ligands were searched and ranked based on their "rerank" energy scores. A minimum of 100 runs for each ligand was carried out. In analyzing the docking scores, we accounted for the recognized bias due to molecular weight [130][131][132] using the scheme: DS norm = 7.2 × E dock /MW 1 / 3 , where DS norm is the normalized docking score, E dock is the MolDock re-rank score, MW is the molecular weight, and 7.2 is a scaling constant to ensure the average DS norm values are comparable to those of E dock [128]. The best docking results are summarized in Table 1.