Antiviral Potential of Antillogorgia americana and elisabethae Natural Products against nsp16–nsp10 Complex, nsp13, and nsp14 Proteins of SARS-CoV-2: An In Silico Investigation

: Biomolecules of marine origin have many applications in the ﬁeld of biotechnology and medicine, but still hold great potential as bioactive substances against different diseases. The puriﬁca-tion or total synthesis of marine metabolites is expensive, and requires a reliable selection method to reveal their pharmaceutical potential prior to clinical validation. This study aimed to explore the hidden potential of natural products from the gorgonian genus Antillogorgia as anti-SARS-CoV-2 agents, via binding afﬁnity assessments and molecular dynamics (MDs) simulations. The three-dimensional protein structures of the nsp16–nsp10 complex, nsp13, and nsp14 were acquired from the RCSB PDB database. All 165 natural products (NPs) were discovered using the PubChem, ChemSpider, and CMNPD databases. The freeware Autodock Vina was used to conduct the molecular docking procedure, once the proteins and ligands were prepared using BIOVIA discovery studio and Avo-gadro software v1.95. Before running MDs simulations using the CABS-ﬂex 2.0 website, the binding afﬁnity assessments and amino acid interactions were carefully examined. Just twelve NPs were selected, and ﬁve of those NPs interacted optimally with the catalytic amino acids of proteins. To conclude, pseudopterosin A ( − 8.0 kcal/mol), seco -pseudopterosin A ( − 7.2 kcal/mol), sandresolide B ( − 6.2 kcal/mol), elisabatin A ( − 7.0 kcal/mol), and elisapterosin A ( − 10.7 kcal/mol) appeared to be the most promising candidates against the nsp16–nsp10, nsp13, and nsp14 proteins.


Introduction
By July 2023, the COVID-19 pandemic, which continues to have a catastrophic effect on the economy and on humankind as a whole, is expected to have caused over 768 million COVID-19 cases, and almost 7.0 million fatalities globally. Effective pharmaceutical treatments and vaccinations are essential to containing the outbreak. Coronavirus research has received an unusually high level of attention throughout this global crisis, which has sped up the development of vaccines [1,2]. The decrease in mortality and morbidity was made possible by the early adoption of drugs such as dexamethasone and remdesivir [3][4][5][6][7][8].
Remdesivir, however, appears to have a minimal effect on patients' chances of survival, according to available research [9]. To address SARS-CoV-2 infections, research on novel therapies is still important.
The nsp10 is known to be a distinctive protein and, as a complex of nsp16-nsp10, it performs C2 -O methylation at the 5 end of the viral RNA, promoting efficient virus This in silico investigation aimed to assess the effectiveness of metabolites from the Antillogorgia genus against three significant known non-structural enzymes of SARS-CoV-2, with a focus on the two species Americana and Elisabethae . The anti-SARS-CoV-2 capabilities of all 165 NPs with low molecular weights from the two aforementioned species of soft corals were tested. This study serves as an illustration of effective data collecting packaged as a species-specific chemical library. Additionally, this study's docking experiments demonstrated how to use such a library. For the first time, this work discovered several unique NPs that interacted with the concerned catalytic residues of the nsp16-10 complex, nsp13, and nsp14 proteins. The possible toxicity of the NPs was also evaluated during this research, and a detailed analysis of their pharmacokinetic and bioactivity predictions was conducted. The schematic workflow of this work is shown in

Obtaining Ligands
Through an extensive literature search, information on the natural products from Antillogorgia americana and Antillogorgia elisabethae was examined. These two types of soft corals include approximately 165 naturally occurring products (see Table S1). To find all of the ligands' 2D or 3D conformers, a thorough database search was conducted.
Firstly, the SDF chemically formatted 3D conformers of ligands were obtained from the PubChem database [78]. 2D conformers were downloaded in the event that 3D structures were not available. For the ligands that were missing from the PubChem database, a second source, the ChemSpider database, was utilized. The ChemSpider database's ligands were downloaded in Mol chemical format [79]. If the ligands were not found in either the PubChem or Chemspider database, then a third database, CMNPD, was employed as a backup source [80]. In order to advance this work, all 165 ligands were eventually effectively obtained.

Visualization
We used BIOVIA Discovery Studio Visualizer 4.5 [81] to display the protein structures of nsp16-nsp10, nsp13, and nsp14. By doing so, we were able to learn how many ligands, HETATM groups, and amino acid chains were associated with each protein. On the other hand, prior to being used for the ligand preparations, Avogadro software version 1.95 was used to display the ligands [82,83].

Pre-Docking Preparations
Prior to the docking process, it was essential to prepare the proteins and ligands. Using the information compiled throughout the visualization process, BIOVIA Discovery Studio Visualizer was employed to prepare both the target proteins. The amino acid chains A and B were present in all three nonstructural proteins. Upon visualization, all target proteins revealed the presence of ligand groups, HETATM, and water molecules. Thus, unnecessary molecules and amino acid chains were cleaned and removed from all the 3D structures. In this investigation, all the proteins had chain A retained. In addition, the program Swiss-pdb viewer [84] was employed to attempt to repair the damaged amino acid chains, hydrogen bonds were inserted, and energy minimization was accomplished. In order to protect the active site, the ligand groups were only removed following the energy minimization procedure. Additionally, the Swiss-pdb viewer was used to export these structures in the PDB format, which was then converted to the appropriate PDBQT format using Autodock tools, a component of MGL tools [85].
Using the Avogadro program v. 1.95, the 2D and 3D conformers of ligands received in SDF and Mol chemical forms were prepared. The ligands were first converted to 3D by creating 3D coordinates, by opening the 2D structures in Avogadro. All the 2D structures were transformed into 3D conformers in this manner, and were archived. Once all 165 ligands were in 3D format, they were each manually subjected to a process of geometry optimization and energy minimization. The PDB format was used to record all the processed ligands. In order to add Gasteiger charges, all ligands were converted from PDB to PDBQT format, using the auto dock tools.

Molecular Docking
The molecular docking process was carried out using Autodock Vina 1.2.0, the most recent version [86,87]. The Grid box was built and positioned in the best possible manner, using the information on the location of the active site, consisting of vital catalytic amino acids for nsp16-nsp10, nsp13, and nsp14. The PDBQT-formatted nsp16-nsp10, nsp13, and nsp14 protein files were firstly opened individually in the program, and 165 ligands were manually docked. For nsp16-nsp10, nsp13, and nsp14, the cavity volume was 461 Å, 1126 Å, and 1960 Å, respectively. For nsp16-nsp10, the grid box coordinates were set to X = −80 Å, Y = 29 Å, and Z = 39 Å. The grid box for nsp13 was located at X = 15 Å, Y = 17 Å, and Z = −72 Å, but for nsp14, the grid box was maintained at X = 13 Å, Y = −10 Å, and Z = −18 Å. In this study, all ligands were flexible, whereas proteins were rigid. To assure the correctness of the results, the docking operation was carried out three times (refer  to Tables 1-3). To further understand the amino acid interactions, all the protein-ligand complexes in the PDB format were imported by creating 2D and 3D maps illustrating all binding and non-binding residues (see ; BIOVIA Discovery Studio Visualizer 4.5 facilitated this task. Å, and 1960 Å, respectively. For nsp16-nsp10, the grid box coordinates were set to X = −80 Å, Y = 29 Å, and Z = 39 Å. The grid box for nsp13 was located at X = 15 Å, Y = 17 Å, and Z = −72 Å, but for nsp14, the grid box was maintained at X = 13 Å, Y = −10 Å, and Z = −18 Å. In this study, all ligands were flexible, whereas proteins were rigid. To assure the correctness of the results, the docking operation was carried out three times (refer to Tables 1-3). To further understand the amino acid interactions, all the protein-ligand complexes in the PDB format were imported by creating 2D and 3D maps illustrating all binding and nonbinding residues (see ; BIOVIA Discovery Studio Visualizer 4.5 facilitated this task.

Docking Validation
The crystal structure of nsp14 was attached to its inhibitor S-adenosyl-L-homocysteine (SAH) [77]. Thus, the validation of protocol for docking was performed for the protein nsp14 by redocking said inhibitor in the same pocket. The binding affinity score and RMSD value were noted. Additionally, the amino acid interactions were observed and compared, to ensure the correctness of the docking procedure.

Molecular Dynamics Simulation
Using the CABS-flex 2.0 website, MDs simulations were performed for the proteinligand complexes, demonstrating the appropriate amino acid associations [88]. Rapid simulations are used on this platform to adequately assess protein flexibility. The server was loaded with the PDB files for the chosen complexes. The simulation's specifications, including the number of cycles (50), the distance among the trajectory frames (50), the simulation temperature set at default, as well as the random number generator seed, were set to their default values by the web server. The simulation produced 3D models, contact maps, and graphs of the root-mean-square fluctuation. We conducted comparative analyses using these RMSF graphs (in Figure 9).

Toxicity Assessment
ProTox-II and StopTox web servers were employed to assess the possible toxicity of the top leads [89]. In a virtual environment, ProTox-II offers quick forecasts regarding the potential toxicity of small molecules, circumventing the requirement for animal experiments. The toxicity class, toxicity endpoints (hepatotoxicity, carcinogenicity, immunotoxicity, mutagenicity and, lastly, cytotoxicity), and LD50 values for the ligands were all provided by this server (see Table 4). The likelihood of ligands causing acute toxicity was quickly and accurately estimated using the user-friendly application StopTox [90]. Information on acute oral, dermal, and inhalation toxicity, as well as on the likelihood of skin and eye irritation and corrosion and skin sensitization, and other critical information, was all provided by this website (refer to Table 5).

Pharmacokinetics Studies
Using the canonical SMILES of best leads (see Table S12) on the Swiss-ADME server, the drug-like features were examined [91,92]. The first step was to observe and note the five different Lipinski variables (Ro5). The molecular weight (≤500), consensus log p-value (≤5), and hydrogen bond donors and acceptors (≤5) were the criteria. Two additional indicators were included, the topological polar surface area (≤140) and the number of rotatable bonds (≤10), in order to more precisely ascertain the likeliness of the ligands being orally efficacious (in Table 6). The likelihood of GI absorption and water solubility were also noted. A simple mathematical approach was employed to calculate the absorption percentage (AB%) further for an accurate assessment of the GI absorption [93,94] (in Table 7).
Additionally, the Swiss-ADME boiled egg plot was used to assess the cell and bloodbrain-barrier penetrability [95] of the lead metabolites (see Figure 10). Ultimately, the PASSonline web server [96] was used to investigate probable bioactivities (refer to Table 8).

Binding Affinity Analyses
In order for the results to be easily understood, the docking results for all three targets were categorized into three groups. The detailed binding affinities for all the ligands and targets can be found in the supplementary information (Tables S2-S11).

nsp16-nsp10 Docking
For the nsp16-nsp10 complex, the docking scores obtained were classified into three categories: the lower class, moderate class, and higher class. The moderate class was further divided into a lower-moderate and an upper-moderate class.
The lowest docking scores, which ranged from −4.2 to −5.9 kcal/mol, were observed for 22 metabolites (see Table 1). Elisabethadienol, elisabethin D, furanotriene, and ichthyothereol acetate exhibited the highest docking scores of -5.9 kcal/mol in category 1. Nevertheless, cumbiasin C and cumbiasin B showed the lowest binding scores, measuring −4.2 and −4.7 kcal/mol, respectively (refer to Table S2). A total of 66 NPs in category 2 demonstrated lower-moderate affinities for the nsp16-nsp10 complex. Cumbiasin A, CMNPD12139, CMNPD16185, CMNPD17188, and seco-pseudopterosin J showed the best affinities in this category, at −6.9 kcal/mol. Eight NPs, including calamenene, curcuhydroquinone, elisapterosin E, furanogermacrene, germacrene D, isofuranotriene, methoxyamericanolide I, and pseudopterolide achieved the lowest value of −6.0 kcal/mol. In addition, 68 NPs demonstrated considerably better binding than the NPs listed above (see Table S3). Elisabatin B, and pseudopterosin B, D, and M all showed a docking value of −7.9 kcal/mol, closely followed by pseudopterosin C and E at −7.8 kcal/mol. On the other hand, 19 NPs demonstrated the lowest docking scores (refer to Table S4) in the upper-moderate class. Lastly, category 3 is stated, which consists of nine NPs with the best binding affinities (in Table 1). Among all of the NPs in this group, the docking values varied from −8.0 to −8.2 kcal/mol (see Table S5).

nsp13 Docking
By separating the output into the three categories shown below, the docking data for nsp13 were made more comprehensible.
The 36 metabolites in category 1 are those with modest docking scores for the helicase protein (in Table 2). The molecules with the lowest binding values were 12-Acetoxypseudopterolide, β-gorgonene, pseudopterosin methylated aglycone, and calarene. The strongest affinity values in this group, however, were found to be for β-eudesmol, furanotriene, and pseudopterosin G at −5.9 kcal/mol (see Table S6). Furthermore, the 85 metabolites in category 2 had moderate binding affinities, with values between −6.0 and −6.9 kcal/mol. Eight metabolites, including ameristerenol A, caribenol B, CMNPD12140, pseudopterosin J, R, and S, and seco-pseudopterosin B and J demonstrated the best affinity values in this group, whereas 12 metabolites showed the lowest affinity, −6.0 kcal/mol (refer to Table S7). Last but not least, the values for the 44 metabolites with high affinities stated in category 3 ranged from −7.0 kcal/mol to a maximum of −8.2 kcal/mol. The most effective metabolites in this group were CMNPD16187, ileabethoxazole, pseudopterosin T, and seco-pseudopterosin K. Nevertheless, among this set of metabolites, fifteen of them showed the lowest value, which was −7.0 kcal/mol (see Table S8). Table 2. Binding affinities of all the NPs for the nsp13 protein.
A total of 32 metabolites with moderate docking scores, ranging from −6.5 kcal/mol to −7.9 kcal/mol, are displayed in category 1 (in Table 3). Elisabethin F and aurene-16, 18-diol-18-acetate had the maximum value of −7.9 kcal/mol, whereas (−)-β-chamigrene, bicyclogermacrene, and germacrene D displayed the lowest value, of −7.0 kcal/mol, in this group (see Table S9). Category 2, listed in the table above, contains 95 metabolites with high docking scores ranging from −8.0 kcal/mol to −9.9 kcal/mol. Seco-pseudopterosin B came in second place, with a docking score of −9.8 kcal/mol, closely behind ileabethin and pseudopterosin V. Eleven NPs showed the lowest affinities in the group, with a −8.0 kcal/mol value (refer to Table S10). Finally, category 3 refers to the 38 NPs that showed the strongest binding scores, which were −10.0 kcal/mol and higher. Pseudopterosin O, and elisabatin B and C demonstrated the highest score of −11.2 kcal/mol, whereas 14 NPs exhibited the lowest docking score in this group, at −10 kcal/mol (see Table S11). Table 3. Binding affinities of all the metabolites for the nsp14 protein.

Categorization of Ligands-nsp14
Moderate High Very High −6.5 to −7.9 kcal/mol −8.0 to −9.9 kcal/mol −10.0 kcal/mol and above  After being re-docked, the original inhibitor gave a docking score of −7.5 kcal/mol, and the RMSD value for the superimposed crystal structures was calculated as 0.89 Å. The computed RMSD value was below the threshold of 2.0 Å, indicating that the docking operation was carried out with good precision.

Amino Acid Interactions
To find complexes exhibiting optimum amino acid interactions in accordance with the criteria listed below, 495 docked complexes for all three SARS-CoV-2 nonstructural proteins were rigorously examined.

1.
The complex must interact with key catalytic residues 2.
It must interact by forming hydrogen and/or electrostatic bonds 3.
The bond distance for conventional hydrogen must not exceed (Å ≤ 3.00) 4.
The bond distance for electrostatic bonds must not exceed (Å ≤ 5.00) Twelve docked complexes that satisfied the established criteria and, in our opinion, might serve as effective inhibitors were found after a thorough screening approach for the nsp16-nsp10 complex, nsp13, and nsp14 proteins. NPs that solely exhibited hydrophobic interactions (VdW force) and/or no interactions with the concerned residues were eliminated as prospective candidates.

Selected Docked Complexes
Just three docked complexes out of 165 satisfied the selection requirements for the target nsp16-nsp10 protein complex. The important amino acids for the nsp16-nsp10 complex are LYS: 46, ASP: 130, LYS: 170, and GLU: 203. The three NPs that demonstrated desirable residue interactions involving the aforementioned amino acids are shown in Figures 3 and 4

below.
Just three docked complexes out of 165 satisfied the selection requirements for t target nsp16-nsp10 protein complex. The important amino acids for the nsp16-nsp complex are LYS: 46, ASP: 130, LYS: 170, and GLU: 203. The three NPs that demonstrat desirable residue interactions involving the aforementioned amino acids are shown Figures 3 and 4 below.

Selected Docked Complexes
Just three docked complexes out of 165 satisfied the selection requirements for the target nsp16-nsp10 protein complex. The important amino acids for the nsp16-nsp10 complex are LYS: 46, ASP: 130, LYS: 170, and GLU: 203. The three NPs that demonstrated desirable residue interactions involving the aforementioned amino acids are shown in Figures 3 and 4 below.   In Figure 4, Figures 5 and 6 below present the five NPs that showed desirable amino acid interactions involving the three stated residues.      LYS: 288, there were two pi-cation connections (4.00 Å and 4.12 Å). Hydrophobic interactions with ASP: 374, SER: 289, GLN: 404, and ARG: 567 were also observed.
Just four docked complexes, out of 165, met the inhibitory requirements for the target nsp14 protein around six residues: LYS: 288, SER: 289, ASP: 374, GLU: 375, GLN: 404, and ARG: 567. The four NPs that demonstrated advantageous interactions with the residues of importance are shown in Figures 7 and 8 below.   In Figure 8, we can see that elisapterosin A interacted with all three of the critical residues of nsp14. A conventional hydrogen bond (

MDs Simulations
Throughout the simulation (in Figure 9), the three docked nsp16-nsp10 complexes revealed minimal residue variations for the four essential catalytic amino acids.  Comparatively, ASN: 388 and PRO: 429 had the slightly higher RMSF values of 1.015 Å and 1.020 Å, respectively. All the complexes were found to be far below the 3.0 Å criterion. Thus, all twelve NPs formed a stable conformation with the proteins of interest.

Toxicity Evaluation
The toxic properties of the selected natural products were investigated, using the ProTox-II (see Table 4) and StopTox (see Table 5) web servers.  Amphilectosin A  3000  5  Inactive  Inactive  Active  Inactive  Amphilectosin B  3000  5  Inactive  Inactive  Active  Inactive  Ameristerenol A  2842  5  Inactive  Inactive  Inactive  Inactive  Ameristerol A  750  4  Inactive  Inactive  Active  Inactive  Elisapterosin A  50  2  Inactive  Inactive  Active  Inactive  Elisabatin A  220  3  Inactive  Inactive  Active  Inactive  Pseudopterosin A  3000  5  Inactive  Inactive  Active  Inactive  Pseudopterosin R  3000  5  Inactive  Inactive  Active  Inactive  Pseudopterolide  274  3  Inactive  Inactive  Inactive  Active  Pseudopteroxazole  1600  4  Inactive  Inactive  Inactive  Inactive  Sandresolide B  34  2  Inactive  Active  Inactive  Inactive  seco-Pseudopterosin A  3000  5  Inactive  Inactive Active Inactive

ProTox-II Results
The ProTox-II program categorizes ligands into six toxicity classes (a higher number means lower toxicity) on the basis of LD50 scores, which is an estimate of the median fatal dosage at which 50% of the test subjects die upon oral exposure. Moreover, the risk of liver damage, a propensity to induce cancer, adverse immune system effects, and the potential to produce genetic mutations are also evaluated.
Among the twelve NPs, only six NPs were found to be class 5 members (2000 < LD50 ≤ 5000), suggesting their probable non-toxic nature, and were found to be generally harmless when exposed orally. The NPs ameristerol A and pseudopteroxazole were the class 4 (300 < LD50 ≤ 2000) candidates, suggesting a low probability of being toxic when ingested orally. Elisabatin A and pseudopterolide were found to be Class 3 (50 < LD50 ≤ 300), indicating that they could be harmful when consumed. Only two NPs (5 < LD50 ≤ 50) belonged to class 2. They had a significant likelihood of becoming lethal when consumed. Sandresolide B and pseudopterolide, from class 2 and 3, might be a carcinogen and a mutagen, respectively. None of the proposed NPs caused hepatotoxicity. A slight immunotoxicity is desirable, as immunosuppression and immunostimulation to treat SARS-CoV-2 is of paramount importance. To summarize, around half of our recommended NPs may be considered safe to develop into oral drugs.

StopTox Server Output
The estimates of the probability of acute toxicity caused by the NPs were evaluated by the StopTox server.
According to our acute toxicity analysis, the majority of the selected NPs are safe. Just two NPs, elisabatin A and pseudopterolide, showed a probability of toxicity following inhalation and cutaneous exposure. Moreover, pseudopterolide was also found capable of causing oral toxicity. Moreover, both NPs have the potential to produce skin sensitivity. The likelihood of the NPs causing skin and ocular irritation and corrosion is minimal, except for pseudopterolide, which could cause eye irritation.

Drug-Likeness Evaluation
Checking the potential drug-like characteristics of the candidate compounds through in silico research is practically valuable in order to developing them into new therapeutics. Using a set of guidelines known as Lipinski or Pfizer's rule of five, the drug-like characteristics of all the nominated NPs in this study were investigated (see Table 6).

Lipinski Rule
Drug development commonly employs Lipinski's Ro5. This criterion makes it possible to determine the likelihood of biologically active substances possessing the desired chemical and physical characteristics necessary for oral bioavailability and absorption.
All of the NPs that were shortlisted for this study followed the majority of the Lipinski principles. All molecules had molecular weights that were below the threshold of 500 Dalton. Moreover, the restriction of no more than ten rotatable bonds was not breached. The acceptors for hydrogen bonds were likewise discovered to be below ten. None of the candidate NPs went beyond the permitted limit of five hydrogen bond donors. Nonetheless, the consensus log p-value was slightly exceeded by ameristerenol A and pseudopteroxazole. The non-Lipinski parameter for the topological surface area cannot be more than 140 Å, and all of the NPs complied with this requirement as well. At least four of the five requirements must be met in order to pass Pfizer's or Lipinski's rule. As a result, all the NPs seem reasonable in terms of pharmaceutical drug development against viral diseases such as SARS-CoV-2.

Swiss-ADME
All of the selected NPs were put through the scanner on the Swiss-ADME website (see Table 7), which offers vital information regarding the absorption, distribution, metabolism, and excretion of the questioned molecules. Of the twelve NPs, it was determined that seven were easily soluble in water. The moderate solubility of the other four NPs, namely ameristerol A, elisabatin A, pseudopterosin R, and pseudopteroxazole, is probable. Contrarily, only ameristerenol A is weakly soluble, and may not absorb well through the GI tract. All the NPs were shown to have high GI absorption values and good bioavailability scores, however. The highest AB% score was calculated for pseudopteroxazole. Just three of the twelve NPs screened positive for blood-brain barrier penetration (in Figure 10).
The only natural product that the GI tract cannot passively absorb is ameristerenol A. There is a higher probability of Elisabatin A and Sandresolide B being the penetrators of the blood-brain barrier. Pseudopterolide, on the other hand, seems to be virtually partly permeable. The remaining eight NPs, including amphilectosin A and B, ameristerol A, elisapterosin A, pseudopterosin A and R, pseudopteroxazole, and seco-pseudopterosin A may be absorbed passively through the GI tract. This analysis, taken as a whole, shows positive results for some of the NPs we proposed.

Bioactivity Prediction
The filtered ten NPs were screened for their potential bioactivities, to estimate the usefulness of these products as potential anti-SARS-CoV-2 agents.
Except for the three NPs named pseudopterolide, pseudopteroxazole, and sandresolide B, nine of the twelve NPs were anticipated to have anti-viral properties. Pseudopterolide, however, was shown to be a viral entrance inhibitor, and might also be effective in the treatment of severe acute respiratory syndromes. All the NPs had a significant likelihood of affecting the host's immune system (in Table 8). The immune system needs to be strengthened to combat the illness in the early phases of infection. In contrast, the immune system has to be restrained in the later stages, to prevent inflammation, morbidity and, ultimately, the death of the patient [97,98]. Eleven NPs were predicted to be antiinflammatory, except for seco-pseudopterosin A. Thus, the NPs mentioned above would be capable of either increasing or reducing the immune system activity. Depending on the stage of infection, the NPs could be chosen for treatment. Apart from seco-pseudopterosin A, eleven NPs were projected to have anti-inflammatory properties. Hence, the NPs indicated above showed the ability to alter the activity of the immune system. The NPs used for therapy might vary according to the stage of infection.

PASSonline Program Top Ligands
Predicted Relevant Bioactivity Antifungal, Antibiotic, Anti-diabetic, Anti-infective, Anti-viral, Antibacterial, Anti-parasitic, Anticancer, Antioxidant, Beta glucuronidase inhibitor, Histidine kinase inhibitor, 1,3-β-Glucan The only natural product that the GI tract cannot passively absorb is ameristerenol A. There is a higher probability of Elisabatin A and Sandresolide B being the penetrators of the blood-brain barrier. Pseudopterolide, on the other hand, seems to be virtually partly permeable. The remaining eight NPs, including amphilectosin A and B, ameristerol A, elisapterosin A, pseudopterosin A and R, pseudopteroxazole, and seco-pseudopterosin A may be absorbed passively through the GI tract. This analysis, taken as a whole, shows positive results for some of the NPs we proposed.

Bioactivity Prediction
The filtered ten NPs were screened for their potential bioactivities, to estimate the usefulness of these products as potential anti-SARS-CoV-2 agents.
Except for the three NPs named pseudopterolide, pseudopteroxazole, and sandresolide B, nine of the twelve NPs were anticipated to have anti-viral properties. Pseudopterolide, however, was shown to be a viral entrance inhibitor, and might also be effective in the treatment of severe acute respiratory syndromes. All the NPs had a significant likelihood of affecting the host's immune system (in Table 8). The immune system needs to be strengthened to combat the illness in the early phases of infection. In contrast, the immune system has to be restrained in the later stages, to prevent inflammation, morbidity and, ultimately, the death of the patient [97,98]. Eleven NPs were predicted to be anti-inflammatory, except for seco-pseudopterosin A. Thus, the NPs mentioned above would be capable of either increasing or reducing the immune system activity. Depending on the stage of infection, the NPs could be chosen for treatment. Apart from seco-pseudopterosin A, eleven NPs were projected to have anti-inflammatory properties. Hence, the NPs indicated above showed the ability to alter the activity of the immune system. The NPs used for therapy might vary according to the stage of infection. Anti-bacterial, Anti-parasitic, Anti-infective, Anti-viral, Antifungal, Antifungal enhancer, β -Glucuronidase inhibitor, 1,3-β-Glucan synthase inhibitor, Antineoplastic, Anti-inflammatory, Immunosuppressant, Interleukin 6, 10 antagonist, Interferon antagonist, Interferon gamma antagonist, Cytokine release inhibitor, T cell inhibitor, GRP78 expression inhibitor, Expectorant Pseudopterolide Anti-inflammatory, Antioxidant, Antineoplastic, Antifungal, Antifungal enhancer, β-Glucuronidase inhibitor, Transcription factor NF kappa B stimulant, Antibacterial, Anti-parasitic, Histidine kinase inhibitor, Immunosuppressant, Antibiotic, Expectorant, Interleukin 10 antagonist, Severe acute respiratory syndrome treatment, Antineoplastic alkaloid, Interferon gamma antagonist, Cytokine release inhibitor, and a Viral entry inhibitor.

Conclusions
In conclusion, the COVID-19 pandemic has already claimed millions of lives [99], and continues to pose a threat to human health. The non-structural proteins of SARS-CoV-2 undoubtedly play important roles in infection [100][101][102][103]. Thus, these protein targets are important in drug development to tackle the virus. Our study demonstrated the potential of natural compounds from two species of soft coral, A. americana, and A. elisabethae, to combat SARS-CoV-2. In light of our findings, only 12 out of 165 NPs may function as inhibitors of the nonstructural proteins; namely, nsp16-nsp10, nsp13, and nsp14. With the catalytic residues of the target proteins, five NPs showed advantageous residue interactions. These five NPs were pseudopterosin A, seco-pseudopterosin A, sandresolide B, elisabatin A, and elisapterosin A. Fast MDs simulations also showed that all the docked complexes were stable. To validate their anti-viral propensities, in vivo and in vitro research must be conducted. Overall, the majority of the NPs shortlisted passed toxicity and pharmacokinetic assessments. Therefore, these NPs, with minimal laboratory interventions, can be transformed into novel therapeutics to treat SARS-CoV-2 infection.

Conflicts of Interest:
The authors declare no conflict of interest.