Molecular Docking in silico Analysis of Brazilian Essential Oils Against Host Targets and SARS-CoV-2 Proteins

The inhibitory activity of thirty-one sesquiterpenes identified from Brazilian essential oils (Copaifera langsdorffii Desf., Croton cajucara Benth. and Siparuna guianensis Aublet.) were analyzed by in silico molecular docking. The compounds were characterized by gas chromatography-mass spectrometry (GC-MS) and gas chromatography with flame-ionization detection (GC-FID), and then, applied against severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) proteins and human angiotensin-converting enzyme 2 (hACE2). Applying molecular docking and AutoDock Vina software, a total of 496 individual interactions were detected for sesquiterpenes along with SARS-CoV-2 proteins and hACE2 human angiotensin converting enzyme-2 protein. The findings showed considerable binding affinity of sesquiterpenes with the tested macromolecules. In that, β-selinene from C. langsdorffii displayed the best energy (-7.2 kcal mol) and showed strong interactions with the amino acids of the SARS-CoV-2 M-Pro protein. Spathulenol from C. cajucara strongly interacted with human ACE2, with a binding energy of -7.1 kcal mol. Meanwhile, γ-eudesmol from S. guianensis presented the lowest binding energy (-7.5 kcal mol) by interacting with the SARS-CoV-2 M-Pro complex. Additionally, measurements were performed aiming to evaluate the best sesquiterpenes binding interactions with the main proteins and its homologue files. According to results, these Brazilian essential oils hold antiviral potential being a rich source for further in vitro and in vivo studies focusing on herbal therapeutic adjuvants against SARS-CoV-2 infections.


Introduction
Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) type has been spreading throughout the world since the end of 2019, when the first infections were reported from China. 1,2 By mid-December of 2021, the World Health Organization reported over 271 million infected people worldwide and about 5.31 million deaths due to the complications of the COVID-19 (coronavirus disease 2019) infection. 3 SARS-CoV-2 is a virus from the Coronaviridae family, belonging to the β-coronaviruses lineage B, and represents one of the seven coronaviruses able to infect humans. 2 It is a single-stranded ribonucleic acid (RNA) genome virus, encapsulated by a membrane envelope encrusted with transmembrane spike glycoproteins (S proteins), granting its crown-like morphology. 4 Several studies have tried to identify SARS-CoV-2 fusion inhibitors that could mediate the membrane-fusion process associated with the human angiotensin-converting enzyme 2 (hACE2), an exopeptidase expressed on epithelial cells present in most tissues, such as lungs, kidneys, and heart, acting as the primary receptor in human cell infection. 2,4,5 Other virus proteins associated with this infection are SARS-CoV-2 M-Pro, SARS-CoV-2 Nsp15/ NendoU, SARS-CoV-2 RdRP, and SARS-CoV-2 Spike S, which are responsible for the viral genome replication and generate new viruses inside host cells. 1,4,6 SARS-CoV-2 presents higher receptor-binding capacity rather than SARS-CoV, contributing to its effective transmissibility and infectivity. 4,7 Therefore, to search the effectiveness of fusion inhibitor compounds against this severe publichealth threat become mandatory in the pharmacological science field. Since many herbals show antiviral activity, the phytochemicals effectiveness against SARS-CoV-2 has become of crucial interest. 6,8 Brazil hosts 15 to 20% of the world's biological diversity, with a vast potential to discover new biocompounds from natural sources. 9 Many of the herbal extracts exhibit biological properties due to their complex mixture of compounds produced for plant protection, 10 such as flavonoids, 11 resveratrol, 12 betulinic acid, indigo, aloeemodin, luteolin, and quinone-methide triterpenoids. 13 These compounds act as immunomodulators, suppressing inflammatory reactions, which are important aspects related to the elevated morbidity and mortality of the SARS-CoV-2 infection. 4 From this perspective, the ethanol extract of Torreya nucifera and its isolated diterpenoids and flavonoids exhibited good SARS-CoV 3CL(pro) inhibitory activity 13 and essential oils also presented potential against SARS-CoV-2 proteins, such as garlic oil (Allium sativum L.). 5 Focusing on new findings, the essential oil of copaiba (Copaifera langsdorffii Desf.) has already demonstrated antibacterial, antifungal, anti-inflammatory, anti-leishmania, and anti-cancer activities 10 which have been linked to the bioactive action of nonpolar diterpenes and sesquiterpenes. 10 Meanwhile, sacaca (Croton cajucara Benth.) 14 and negramina (Siparuna guianensis Aublet.) 15,16 essential oils present bioactive mono-and sesquiterpenes. To the best of our knowledge, there are no theoretical or experimental studies by assaying these essential oils against SARS-CoV-2. So, in the current study molecular docking in silico analysis were applied on protein-ligand interaction models, analyzing target protein groups such as: angiotensin-converting enzyme (hACE2), coronavirus main proteinase (SARS-CoV-2 MPro), SARS-CoV-2 endoribonuclease (SARS-CoV-2 Nsp15/NendoU), RNA-dependent RNA polymerase (SARS-CoV-2 RdRP), and Spike S protein.
In addition, the potential application of these essential oils as adjuvants against SARS-CoV-2 are herein discussed for further in vitro and in vivo studies.

Experimental
Herbal and essential oils extraction Samples of Copaifera langsdorffii Desf. (oil resin), Croton cajucara Benth. (leaves), and Siparuna guianensis Aublet. (leaves), were collected in the Tupé Sustainable Development Reserve ("Reserva de Desenvolvimento Sustentável do Tupé-RDS Tupé") at the Comunidade Colônia Central, located about 25 km from the city of Manaus, Amazonas state, Brazil. The essential oils were obtained by hydrodistillation in a Clevenger apparatus at 100 °C for three hours using a 1:4 ratio for sample:ultrapure water. The extracted materials were dried on anhydrous sodium sulfate (Sigma-Aldrich, St. Louis, USA), filtered through a 0.22 μm cellulose membrane, and stored in borosilicate glass vials at −20 °C until further analysis. 10

Characterization of the essential oils
The identification of the copaiba (Copaifera langsdorffii Desf.), sacaca (Croton cajucara Benth.), and negramina (Siparuna guianensis Aublet.) essential oil compounds followed gas chromatography-mass spectrometry (GC-MS) and gas chromatography with flame-ionization detection (GC-FID) methods. The GC-MS analysis occurred into HP 6890 gas chromatograph with split/splitless injection port, combined with an HP-5MS cross-linked fused silica capillary column (30 m × 0.25 mm × 0.25 μm) and HP-5975 mass selective detector (Hewlett-Packard Company, Wilmington, USA). The carrier gas applied was helium. The injection of essential oils corresponded to a volume of 1 μL and a 1:25 split ratio. The chromatographic parameters were: oven initial temperature, 60 °C; ramp rate, 5 °C min −1 ; oven final temperature, 250 °C; injector, 220 °C; detector, 250 °C. The electron ionization system was set at 70 eV. Data acquisition and integration were carried out through the MSD ChemStation software. The essential oils' components identification applied the comparison of their fragmentation pattern with the National Institute of Standards and Technology (NIST-05) mass spectral library data, the Kovats index, retention index from co-injection of n-alkane standard solutions (C 8 -C 20 ) (Sigma-Aldrich, St. Louis, USA), under the same chromatographic conditions. 10 The analyte concentrations were calculated according to their peak areas in the chromatogram.
The GC-FID analysis employed a PR2100 GC-FID instrument with split/splitless injection port (Alpha MOS, Toulouse, France), combined with a fused silica capillary column (30 m × 0.25 mm × 0.25 μm) coated with cross-linked 5% phenyl polysilphenylene-siloxane (SGE Analytical Science Pty Ltd, Victoria, Australia). The chromatographic parameters correspond to those described before to GC-MS. Data acquisition and integration were carried out using Winilab software. β-Caryophyllene, α-humulene, and isoshyobunone were selected as the standard for the quantification of the main components present in the essential oils. 10

Ligand file selection and preparation
In the search for a relevant ligand-protein interaction, the molecular docking evaluated the main components of the assayed essential oils, considering compounds with concentrations higher than 2%. A total of 31 main sesquiterpenes were identified and then applied against the human pathogenic protein of the SARS-CoV-2 virus, such as SARS-CoV-2 Main Protease, Nsp15/NendoU, RdRp, Spike S, and also the protein hACE2, were considered for the docking in silico analysis. The compound's chemical structures were downloaded from NIH-National Library of Medicine (PubChem) 17 in SDF format of the 3-D conformation of all sesquiterpene type-compounds and then converted to a protein data bank (PDB) file by applying OpenBabel 18 software. Afterward, AutoDock Tools 4.2 19 was used to detect the compound roots and structural torsion. Finally, files were saved separately in a PDBQT format file (protein data bank, partial charge "Q" and atom type "T") for further analysis.

Protein preparation
Since the assayed essential oil compounds interacted with SARS-CoV-2 through its target proteins and also hACE2 protein, Swiss-model server [20][21][22] (a fully automated protein structure homology-modeling server) was then used to find the most suitable templates for the selected target proteins. For this purpose, this server performed a basic local alignment search (BLAST) in order to find regions of similarity between biological sequences, and the best matches. Therefore, a structure comparison becomes possible by using QMEAN analysis. The minimal acceptable resolution for the proteins files was considered as 2.0 Å. Protein structures were downloaded from RCSB-Protein Data Bank 23 after template homology studies. Host proteins were split into five main groups aiming to analyze the molecule's interactions separated by group (Table 1).
Posteriorly, AutoDock Tools (The Scripps Research Institute) properly checked all chemical particularities of each molecule involved in the molecular interaction. 19 The target macromolecules were dehydrated and hydrogens were added afterward only on the polar atoms. Also, Kollman charges were added to the target macromolecules Table 1. Protein Data Bank (PDB) files separated in five groups by homologues of each studied crystal structure hACE2 group: 6M17 is a complex composed by the interactions of the angiotensin-converting enzyme 2 (chains B and D), SARS-CoV-2 receptor binding domain (chains E and F) and sodium-dependent neutral amino acid transporter B(0)AT1 (chains A and C).1R42 is the native human angiotensin-converting enzyme-related carboxypeptidase -hACE2 (chain A). SARS-CoV-2 M-pro group: 5R7Z is the 3C-like proteinase, a crystal structure of the SARS-CoV-2 main protease in complex with Z1220452176; 5R80 is the 3C-like proteinase, a crystal structure of the SARS-CoV-2 main protease in complex with Z18197050; 5R81 is the 3C-like proteinase, a crystal structure of the SARS-CoV-2 main protease in complex with Z1367324110; 5R82 is the 3C-like proteinase, a crystal structure of the SARS-CoV-2 main protease in complex with Z219104216; 5R83 is the 3C-like proteinase, a crystal structure of the SARS-CoV-2 main protease in complex with Z44592329; 5R84 is the 3C-like proteinase, a crystal structure of the SARS-CoV-2 main protease in complex with Z31792168; 6LU7 is a crystal structure of the SARS-CoV-2 main protease in complex with an inhibitor N3; 6M03 is a crystal structure of the SARS-CoV-2 main protease in apo form; 6Y84 the SARS-CoV-2 main protease with unliganded active site (2019-nCoV, coronavirus disease 2019, COVID-19). SARS-CoV-2 Nsp15/NendoU group: 6VWW is the crystal structure of NSP-15 endoribonuclease from SARS-CoV-2; 6W01 is a 1.9 Å crystal structure of NSP15 endoribonuclease from SARS-CoV-2 in complex with citrate; 6W02 is a crystal structure of ADP ribose phosphatase of NSP3 from SARS CoV-2 in complex with ADP ribose. SARS-CoV-2 RdRp group: 6M71 is a SARS-CoV-2 RNA-dependent RNA polymerase in complex with cofactors. Spike S group: 6M0J is a crystal structure of the SARS-CoV-2 spike receptor-binding domain bound with hACE2.
and Gasteiger charges were computed, as a common preparation required for further docking on these software.

Molecular docking simulation
A blind docking was initially performed by AutoDock Vina 24 to locate alternate binding sites enclosing the whole macromolecules with a very extended grid (60 Å × 60 Å × 60 Å). So, procedures were achieved for each macromolecule to search the whole interaction sites possibilities by applying exhaustiveness values of 8, 16, 24, 32, 64, 128, and 500. The analysis of an output file composed of 496 individual interaction files, regarding the 31 herbal chemical compounds and 15 host proteins, which were considered among the 20 best VINA score bindings. The best binding energy modes were chosen to rerun separately in their specific binding sites to get precisely the best scores of specific regions to its possible binding poses. In this way, a detailed study was conducted with a smaller grid (44 Å × 32 Å × 48 Å) centered on the groove, with exhaustiveness values of 64 and 3 repetitions for each selected molecule, printing up to 20 modes and results within -5.0 kcal mol -1 from the lowest energy pose. Hence, the main molecules with the lowest energy binding level of each group were selected for visual presentation. Also, the energy range was set to 4 for all VINA's files. Due to the huge amount of data involved in this processing thread, the dockME LITE Alpha 25 software was applied. This open source software is just an automated computational tool that makes an interface with Autodock VINA by activating it through terminal command lines automatically, screening every macromolecule contained in the folder against every ligand prepared in another folder. 25 In addition, PyMol software 26 was also used for the best binding sites visualization. Meanwhile, the APBS (Adaptive Poisson-Boltzmann Solver) model was then assessed for electrostatics and solvation complementary analysis due to its accuracy to determine the electrostatic energy under polar solvent conditions. For the automatic generation of 2D ligand-protein interaction, LigPlot+ software 27 was applied to depict hydrophobic bonds, hydrogen bonds, and their bond lengths in each docking pose in the form of graphical representation.

Results and Discussion
Essential oils composition GC-MS and GC-FID analysis showed a total of 31 main sesquiterpenes from Copaifera langsdorffii Desf., Croton cajucara Benth., and Siparuna guianensis Aublet. which contents higher than 2% corresponded to over 78% of the found sesquiterpenes ( Table 2).
Facing COVID-19 disease, as utmost need, previous reports have shown molecular docking approaches to search natural compounds against SARS-CoV-2 infections. 1,27 From this perspective, concerning Copaifera langsdorffii Desf., Croton cajucara Benth., and Siparuna guianensis Aublet., the best ligand-protein binding energy to perform molecular docking were obtained according to Table 3.
According to Table 3, the finds showed that the assessed sesquiterpenes possessed negative values of the binding Vol. 00, No. 00, 2022  affinity playing an important role in the inhibition of the macromolecules. Indeed, previous studies predict that compounds with a binding affinity lower to -6.5 kcal mol -1 or less acting on active sites have shown effective proteins inhibition under significant stability of the ligand-protein complex. 32,33 It was also evidenced that the assessed plants have shown a good quantity of sesquiterpenes with considerably acceptable negative binding affinities with the selected targets of SARS-CoV-2 proteins. Moreover, sesquiterpenes of S. guianensis showed significant single or multiple interactions with some SARS-CoV-2 proteins. Otherwise, weak or no relevant interaction with any SARS-CoV-2 proteins was observed for α-bergamotene and germacrene D (C. langsdorffii), as well as linalool and germacrene D (C. cajucara). Meanwhile, higher binding affinities were obtained between SARS-CoV-2 M-Pro with β-selinene (-7.2 kcal mol -1 ) or γ-eudesmol (-7.5 kcal mol -1 ), from C. langsdorffii and S. guianensis, respectively. SARS-CoV-2 M-Pro corresponds to the primary enzyme responsible for processing the polyproteins translated from the viral RNA. 4,29 Sesquiterpenes from C. cajucara presented a higher binding affinity (-7.1 kcal mol -1 ) to spathulenol-hACE2, which is an important receptor for viral entrance in human cells. 2,4,5 For SARS-CoV-2 RdRP, sesquiterpenes from C. langsdorffii and S. guianensis did not present significant binding energy. On the other hand, bicyclogermacrene (C. cajucara) were able to interact with this protein, which is a target polymerase responsible for the viral genome replication within the host cells. 4 At this point, it is important that these inhibition effects on the SARS-CoV-2 main target proteins are encouraging to consider the Brazilian essential oils (C. langsdorffii, C. cajucara, and S. guianensis) or its isolated compounds to further studies aiming at to formulate adjuvants against SARS-CoV-2.
It is noteworthy that some compounds may present an inhibitory effect on several proteins simultaneously, and the analyzed sesquiterpenes also displayed equal binding energy for the same protein. These data suggest that the evaluated components of the Brazilian essential oils can be used in natura in further in vitro and in vivo studies as options to be considered to treat or prevent synergistically the coronavirus, as well as in form of isolated sesquiterpenes since some of them acted directly in one or multiple SARS-CoV-2 proteins. It is also possible to confirm that all the studied herbal oils have presented meaningful binding energy levels in its solvation process with the target proteins that they were docked, with an energy level lower than -5.4 kcal mol -1 (Figure 1a), also noticed when this evaluation is conducted isolated by group (Figure 1b), having the highest meaningful binding energy level result of -5.3 kcal mol -1 , therefore, corroborating with Figure 1.
The presented binding energy levels are also observed when data were evaluated in a Gaussian (or Normal distribution): (1) where µ is the expected value of a finite case, calculated as ; and s is the standard deviation ( Figure 2).
Those results corroborate with Silva et al. 1 that performed molecular docking of essential oils compounds against SARS-CoV-2 targets and suggested that these substances may act synergistically, amplifying other antiviral agents, and relieving some disease symptoms 1 (Table 4). Indeed, it is also possible to analyze through the normal distribution above described from which more than 90% of the evaluated data presented a binding energy level lower than -5.0 kcal mol -1 . This find suggests that the assayed sesquiterpenes applied through an appropriate pharmacological approach may reach the expected synergic effect against the main proteins of SARS-CoV-2.
The assayed sesquiterpenes from these Brazilian essential oils showed meaningful docking values indicating that these natural compounds can be directed to the human protein and decrease its binding with the protein virus, and now becoming promising candidates to inhibit both hACE2 isolated (PDB: 1R42) and 2019-nCoV RDB/hACE2-B0AT1 complex (PDB: 6M17). It was possible to predict that there are five main amino acid hydrophobic interactions occurring in the peptidase M2 at domain 3 closest to the disulfide bond of the macromolecule. Moreover, APBS solvation analysis showed strong binding in the very electronegative site on the electrostatic surface of this complex, showing a higher affinity level at these physical conditions.
From nine macromolecules related to SARS-CoV Main Protease, the best energy binding levels were taken from the PDB: 6Y84 with the β-selinene as a ligand, reaching the binding energy level of -7.2 kcal mol -1 on its interaction (Figure 3b and Table 3). So, among ten other compounds from copaiba oil, β-selinene was chosen due to its lowest binding energy level. However, for this protein, it is important to point out that all tested sesquiterpenes showed promising results with a very narrow difference range among its binding values. Concerning the Main Protease group, PDB: 6Y84, β-selinene showed the best binding result, while for PDB: 5R84 it was found as the weakest result. APBS solvation data showed that this molecule binds with an isocontour on electrostatic potential surface. Additionally, β-selinene is bound with seven amino acids making hydrophobic bonds from the PDB: 6Y84 macromolecule: Lis5, Phe291, Glu288, Trp207, Leu282, Phe3, and Arg4. It is also relevant to point out that among the other eight host molecules analyzed, copaiba oil presented the best energy binding levels to 6Y84macromolecule. Taken all results together, sesquiterpenes from copaiba essential oil are the best inhibitor candidates for further in vitro or in vivo studies against SARS-CoV Main Protease (M-Pro), being in accordance with other structures of M-Pro from SARS-CoV-2. 34 Concerning the docking of copaiba essential oil SARS-CoV-2 Nsp15/NendoU protein, the lowest energy binding levels corresponded to α-selinene and caryophyllene oxide ligands with PDB: 6W01 macromolecule (-6.9 kcal mol -1 ) (Figures 3c and 3d). β-Bisabolene showed a weaker binding affinity with PDB: 6W02 (-4.4 kcal mol -1 ) and α-selinene was hydrophobically bound with 12 amino acids from the PDB: 6W01: Asp297, Val295, Thr275, Ser274, Ser198, Lys90, Asn200, Tyr279, Leu252, Arg199, Ile296, and Lys277.
Caryophyllene oxide ligand registered interaction with seven amino acids from the crystal structure of NSP15 endoribonuclease from SARS-CoV-2 (PDB ID: 6W01), which are: Asp273, Arg199, Leu252, Lys277, Asn200, Lys71 (H-bound of 2.96 Å), and Tyr279. Both chains (A and B) are related to the uridylate-specific endoribonuclease. 35 In addition, a hydrogen bond at 2.96 Å was verified between the caryophyllene oxide and the Lys71 segment of the SARS-CoV-2 Nsp15/NendoU protein, suggesting strong inhibition. APBS solvation indicated binding at neutral electrostatic sites in both molecules. Thus, sesquiterpenes from copaiba oil can perform a significant role in the inhibition of these essential macromolecules for viral replication. 36 Additionally, α-humulene has shown the best docking score with PDB: 6M71 (affinity energy of -6.4 kcal mol -1 ) (Figure 3e; Table 4) while β-bisabolene was the worst to the same host macromolecule (-4.9 kcal mol -1 ). Thus, α-humulene interacted with seven hydrophobic residues (Asp218, Asn209, Tyr217, Arg33, Thr123, Phe35, and Lys50) and one hydrogen bond with the Arg 116 at 3.09 Å. APBS solvation screening shows that the molecule's site is very electrostatically negative.
Regarding the SARS-CoV-2 spike receptor (Figure 3f), β-bisabolene registered the lowest energy binding with PDB:6M0J (-6.8 kcal mol -1 ) in contrast with germacrene D (-5.6 kcal mol -1 ). This compound held a very narrow energy difference by comparing it with the other copaiba essential oil compounds. The binding mechanism was hydrophobic interactions with Ala413, Met366, Thr434, Glu430, Pro415, Glu435, Ile291, Asn290, and Phe438 at related hACE2 receptor. β-Bisabolene bonded at domain 5 of the hACE2/Spike S complex, containing an active site with a disulfide bond and zinc-binding site. Therefore, it is a significant indication of inhibition, particularly since hACE2 is an important human target for inhibition to avoid engagement with the SARS-CoV-2 viral spike (S) protein. 32,37 APBS solvation results showed β-bisabolene paired with a very electrostatically electronegative site on the surface complex.
The sesquiterpenes are identified from the essential oil leaves of Croton cajucara Benth. showed six main interactions with target SARS-CoV-2 proteins, according to the molecular docking analysis (Figure 4).
Among the analyzed compounds, spathulenol showed the best binding sites and also the highest energy binding levels observed on electrostatic solvation with hACE2. For the other compounds: (i) (E)-nerolidol the highest energy binding levels were detected with SARS-CoV-2 M-Pro; (ii) caryophyllene oxide with SARS-CoV-2 Nsp15/ NendoU; (iii) bicyclogermacrene with SARS-CoV RdRp; and (iv) β-bourbonene and bicyclogermacrene with Spike S macromolecules.
Among C. cajucara compounds, spathulenol showed the lowest binding energy (-7.1 kcal mol -1 ) with ACE2 (PDB: 1R42). Otherwise, the main compound linalool was not satisfactory (-4.9 kcal mol -1 ). Thus, if compared with all SARS-CoV-2 proteins docked, the best docking scores was obtained from this interaction. Indeed, six spathulenol-PDB:1R42 hydrophobic interactions were found at chain A related to the hACE2 complex (Pro612, Tyr158, Pro490, Leu162, Val491, and Tyr613). Strongest binding determined by APBS was found in a very electronegative site on the electrostatic surface of this complex. Therefore, spathulenol presented a promise inhibitory effect on the hACE2 receptors. From nine macromolecules related to the SARS-CoV Main Protease, the best energy binding levels was taken to PDB6Y84-(E)-nerolidol complex (-7.0 kcal mol -1 ) (Figure 4b).
Linalool-PDB5R7Z complex showed weak binding affinity (-4.4 kcal mol -1 ), corroborating with Aanouz et al. 38 studies. Meanwhile, (E)-nerolidol ligand bound with 6 amino acids on hydrophobic site composed by Lys137, Arg4, Phe3, Phe291, Trp207, and Glu288. There were also three H-bond Glu290, Gln127, and Lys5, with distances of 2.78, 2.76, and 2.94 Å, respectively. This favorable interaction with human hACE2 occurred in the receptorbinding motif (RBM) at the domain 4 sites. 39 The evaluation of the hydrophobic binding residues is as much important as the evaluation of the hydrogen bonds in order to understand the stability of the complex formed. 40 In this way, the presence of hydrogen bond interaction explains a stable interaction between the sesquiterpenes and the studied protein. 38 These bindings happened at the domain I and II of the main proteases. 41 According to Figure 4, it is possible to determine that the sesquiterpenes from C. cajucara were binding on the slightly electropositive electrostatic surface.
PDB6W01 was selected between three main SARS-CoV-2 Nsp/NendoU macromolecules sites (Figure 4c). Caryophyllene oxide showed the lowest binding affinity (-6.9 kcal mol -1 ), while linalool was the highest (-4.4 kcal mol -1 ). Caryophyllene oxide ligand bound with 5 amino acids on hydrophobic bonds: Pro206, Phe204, Ala256, Lys260, and Leu215 at the domain II and IV in the chain A of the uridylate-specific endoribonuclease. 42 This molecule linked on a neutral electrostatic site of the molecule, with a slightly negative potential surface. Positive charges are attracted to regions of high electron density (negative potential) and repelled in regions of low electron density (positive potential). 37 Therefore, the studied surface corresponded to a balanced potential density. For the SARS-Cov-2 RNA-dependent RNA polymerase (PDB6M71), the best interaction was observed to the bicyclogermacrene (-6.7 kcal mol -1 ) (Figure 4d). Linalool showed weak interaction with SARS-Cov-2 protein. 38 Virtual screening presented potential inhibitions of the protein with most of the molecules through of hydrophobic site composed by Tyr217, Arg116, Val71, Thr123, Arg33, Phe35, Asp208, and Asn209. APBS solvation screening showed that the molecule's site surface is very electronegative, an indication that the molecule of bicyclogermacrene is attracted to this region due to its high electron density. All bindings happened at chain A, which is related to the NSP-12, corresponding to the domains I, II, and III of the macromolecules. 42 At last, β-bourbonene and bicyclogermacrene presented the lowest energy binding levels (-6.7 kcal mol -1 ) with the PDB6M0J, a crystal structure of SARS-CoV-2 Spike S receptor-binding domain bound with hACE2 (Figures 4e  and 4f). Hydrophobic interactions were responsible for producing β-bourborene-PDB:6M0J complexes with Ser47, Ser44, Asp350, Phe40, Thr347, Ala348, and Trp349. Bicyclogermacrene-PDB6M0J complexed with Phe40, Phe390, Trp69, Asn394, Leu391, and Arg393. Both interactions were related to the hACE2 domain. The sesquiterpenes were fitted onto the electrostatically electronegative domain on the protein surface corresponding to the domain 1 and 5 receptor-binding site with neither zinc channel. 43 At this position, there is a complete loss of human hACE2 binding in vitro. 44 Thus, these sesquiterpenes may inhibit the main host-virus protein complex.
Finally, as the results of the main interactions of Siparuna guianensis Aublet. essential oil, five ligandmacromolecule complexes were considered from the molecular docking, according to Figure 5.
Thus, this promising result indicates the potential application of γ-eudesmol in different ways to prevent or treat the severe acute respiratory infections caused by SARS-CoV-2. The lowest binding energy was related to γ-eudesmol-SARS-CoV Main Protease complex, suggesting this route as the preferable one for treating the viral infection. About hACE2, γ-eudesmol interacted by hydrophobic and hydrogen bonds with the amino acid at a very electronegative site on the electrostatic surface of this complex. 41 The main interaction of the γ-eudesmol with ACE2 occurred in the RBM domain at Zn binding active site, proving an opportunity to uncover a cross-reactive epitope of protein. 45 In addition, γ-eudesmol also interacted with peptidase M2 at the region responsible for glycosylation reactions, slightly inhibiting interaction with SARS-CoV Spike glycoprotein. In another way, SARS-CoV Main Protease had an excellent virtual screening with γ-eudesmol, showing hydrophobic bonds at the electrostatic site of greater electronegativity of the molecule. SARS-CoV-2 Nsp15/NendoU formed a complex with the sesquiterpene by hydrophobic and strong hydrogen bonds predominantly in the site with an isocontour of the electrostatic potential surface at a sheet of the uridylate-specific endoribonuclease, two domains (I and IV) and a helix. 36 SARS-Cov-2 RNA-dependent RNA polymerase interacted with γ-eudesmol, spathulenol, and α-muurolene showing the same energy value (PDB6M71 = -6.3 kcal mol -1 ) (Figures 5d, 5e, 5f). These complexes displayed a lower binding affinity compared with other proteins. The complexes were formed by hydrophobic and hydrogen bonds at a different electrostatic potential surface of the macromolecule (electronegative, neutral, and electropositive for γ-eudesmol, spathulenol, and α-muurolene, respectively). The interaction occurred according to: (i) γ-eudesmol: in the domain 1 at Nsp8 interaction site for; (ii) spathulenol: in domain 1 as a putative inhibitor binding site and conserved polymerase motif D, and also in domain 3; and (iii) α-muurolene: domain 1 conserving the polymerase motif F and as a putative inhibitor binding site, and also in domain 4 (Nsp7 interaction site).
For the other twelve tested compounds, t-muurolol had the lowest energy binding level interacting with the crystal structure of the SARS-CoV-2 Spike S receptor-binding domain bound with hACE2 (Figure 5g). The weakest energy level (-5.8 kcal mol -1 ) corresponded to the protein complex formed with elemol and with iso-shyobunone. The sesquiterpene t-muurolol fits by hydrophobic and hydrogen bonds in a very electrostatically electronegative site on the complex surface (Arg393 (2.78 Å), Asp350, Phe40, Gly352, Phe390, Trp69, and Asn394). In addition, this compound was bound at the receptor-binding site in domain 5, which partially inhibits the human hACE2 binding in vitro. 32,40

Conclusions
The present study provided a potential approach to the application of 31 sesquiterpenes identified from the essential oils of Copaifera langsdorffii Desf., Croton cajucara Benth., and Siparuna guianensis Aublet., which were characterized by using GC-MS and GC-FID techniques. These compounds assayed by the molecular docking applying the AutoDock Vina software showed to be promising candidates to further studies against coronavirus (SARS-CoV-2). Indeed, all tested compounds have shown good potential to inhibit both viral proteins and hACE2 protein, blocking the host receptor and also attacking the macromolecules of SARS-CoV-2 M-Pro, SARS-CoV-2 Nsp15/NendoU, SARS-CoV RdRp, and Spike S protein groups, that represent the major proteins of SARS-CoV-2, responsible to originate the disease COVID-19. Molecular docking attested the spontaneous interaction of all investigated compounds with active sites of viral and human proteins, by strongly binding to their amino acids. Analyzing the anchorage simulation, sesquiterpenes showed a high binding affinity with only one or several target proteins at the same time. Therefore, the present study reinforces that natural oils should be considered as candidates to further studies, considering that the SARS-CoV-2 pandemic is not over. Thus, there is a social and scientific utmost need for further in vitro and in vivo studies focusing on discovering therapeutic adjuvants against SARS-CoV-2 infection.
Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). The authors also would like to thank Dr G. M. Seabra (Department of Medicinal Chemistry and Center for Natural Products, Drug Discovery and Development (CNPD3), School of Pharmacy, University of Florida, Gainesville, FL, USA) for the AutoDock Vina support.