Computational Screening of Plant-Derived Natural Products against SARS-CoV-2 Variants

Round 1

Reviewer 1 Report

The study “Computational Screening of Plant-Derived Natural Products 2 against SARS-CoV2 Variants” is well-designed and conducted the study. The authors identified polyphenols as computational hits after combining multiple computational approaches. In my opinion, the paper is suitable for publication in Future Pharmacology after addressing the following minor points:

 

1. Several polyphenols have already shown promising activity as antivirals. Expand the discussion of the results of this paper. See, for instance: https://doi.org/10.3390/foods10102277 and

https://doi.org/10.1016/j.scitotenv.2021.149719

2. The computational hits identified in this work are polyphenols that have multiple bioactivities activities due to free radicals processes and are not specific. These compounds are computational hits in many natural products-based virtual screenings. Discuss this fact in the paper. For example, see discussions in paper https://doi.org/10.1002/minf.202000171

3. There is a large volume of virtual screening of compound databases against SARS-CoV2 targets. Elaborate more on the novelty of this work vs. what has been published. As exanples, see the review papers: http://dx.doi.org/10.1039/d0cs01065k,  https://doi.org/10.1002/med.21862 

4. Page 2, lines 82-84: the sentence “significantly inhibited the main protease of SARS-CoV-2 as demonstrated by molecular docking, molecular dynamic simulations and quantum computational studies” is not correct. Computational studies do not demonstrate experimental results. See discussion in: https://doi.org/10.12688/f1000research.52676.1 

5. Page 6, line 119: add reference to SMILES notation: https://doi.org/10.1021/ci00057a005

Author Response

We are very thankful for the valuable and positive comments on our manuscript from the reviewer that are important to improve this paper. The critical and informative remarks of the reviewers resulted in an enhanced revised manuscript. The comments are reviewed carefully and a response to each individual comment is provided in the following. We hope this paper will be accepted after this revision.

Comment 1: Several polyphenols have already shown promising activity as antivirals. Expand the discussion of the results of this paper. See, for instance: https://doi.org/10.3390/foods10102277 and https://doi.org/10.1016/j.scitotenv.2021.149719

Response 1: We are grateful to the learned reviewer for making the constructive recommendation. We have now added some discussion about polyphenols' antiviral activity. It will improve the quality of our work and help the readers of our journal. The new section has been incorporated in the updated manuscript at the proper location and is also provided here.

“Polyphenols, particularly flavonoids, have been extensively studied for their antiviral ac-tivity against hepatitis viruses, dengue viruses, Epstein-Barr viruses, herpes viruses, in-fluenza viruses, HIV, rotaviruses, and coronaviruses. Although the precise mechanism of action is uncertain, they reduced viral infection in host cells by preventing virus entry or decreasing virus multiplica-tion. Several studies have shown that tea polyphenols have antiviral properties by inhibiting DNA viruses such as herpesviruses, papillomaviruses, poxviruses, and HIV-1. In a dose-dependent way, resveratrol, ferulic acid, and gallic acid greatly suppressed the expression of Epstein-Barr Virus lytic genes. Furthermore, resvera-trol has been shown to inhibit respiratory viruses such as rhinoviruses and syncytial vi-ruses, as well as Varicella-zoster, a virus that causes fever and a vesicular rash. Further-more, Chojnacka et al. summarised a study based on the analysis of several studies and concluded that ellagic acid, myricetin, kaempferol, and quercetin exhibited anti-influenza activity, whereas cyanidin-3 rutinoside, cyanidin-3-glucoside, rutin, and gallic acid were found to show inhibitory activity against the H1N1 influenza virus via inhibiting viral at-tachment or by influencing viral entrance inhibition into host cells”.

Comment 2: The computational hits identified in this work are polyphenols that have multiple bioactivities activities due to free radicals processes and are not specific. These compounds are computational hits in many natural products-based virtual screenings. Discuss this fact in the paper. For example, see discussions in paper https://doi.org/10.1002/minf.202000171

Response 1: We are very thankful to the learned reviewer for making a constructive suggestion. We now added some discussion based on the computational studies of polyphenols. It will improve the quality of our work and help the readers of our journal. The new section has been incorporated in the updated manuscript at the proper location and is also provided here.

“Additionally, because computational approaches are straightforward, quick, and eco-nomical, they are a useful tool for the discovery and development of drugs based on bioac-tive natural products. To provide 2D and 3D molecular profiles of natural products, they construct pharmacophore models, molecular interaction fields, docking, and simulations of complexes. Because experimental procedures are time-consuming, expensive, and replete with errors when a suitable protein target is lacking, computational tools can be used to enhance the pharmacological efficacy and safety of natural products as well as their further development. Polyphenols are a diverse class of natural compounds that serve a variety of biological functions. Their anti-inflammatory and antioxidant effects are well established. They've been investigated for antiviral properties, including the recently found SARS-CoV2, which has wreaked havoc due to its high infectivity and mortality rate. Many research organisations found different therapeutic compounds from medicinal plants in a relatively short period of time, attributable to computational strategies in re-search. The majority of research has concentrated on polyphenols because of their antiviral char-acteristics, which have been widely documented in prior works. Furthermore, substantial research has been conducted in the last two years on natural products that demonstrate high anti-SARS-CoV2 activity, but no report on natural products as inhibitory agents of various SARS-CoV2 variants is known. As a result of these findings, we conducted a thorough cross-analysis of previously published pharmacological and computational re-search including the polyphenols identified in our study as anti-S-glycoprotein of SARS-CoV2 variants”.

Comment 3: There is a large volume of virtual screening of compound databases against SARS-CoV2 targets. Elaborate more on the novelty of this work vs. what has been published. As examples, see the review papers: http://dx.doi.org/10.1039/d0cs01065k, https://doi.org/10.1002/med.21862

Response 1: Yes, we agree with learned reviewer and thankful for providing constructive suggestion. We now added some discussion to elaborate the novelty of our work. The added section has been included in the revised manuscript at the appropriate place and is also provided here as below.

“The majority of research has concentrated on polyphenols because of their antiviral char-acteristics, which have been widely documented in prior works. Furthermore, substantial research has been conducted in the last two years on natural products that demonstrate high anti-SARS-CoV2 activity, but no report on natural products as inhibitory agents of various SARS-CoV2 variants is known. As a result of these findings, we conducted a thorough cross-analysis of previously published pharmacological and computational re-search including the polyphenols identified in our study as anti-S-glycoprotein of SARS-CoV2 variants”.

Comment 4: Page 2, lines 82-84: the sentence “significantly inhibited the main protease of SARS-CoV-2 as demonstrated by molecular docking, molecular dynamic simulations and quantum computational studies” is not correct. Computational studies do not demonstrate experimental results. See discussion in: https://doi.org/10.12688/f1000research.52676.1

Response 1: We agree with the knowledgeable reviewer that computational analyses do not show experimental results. We have now revised the sentence in accordance with the recommendation, and the correction has been incorporated into the text at the proper location in the revised manuscript.

“Moreover, diosmetin, an important flavonoid of citrus lemon, significantly acts as predicting inhibitor of SARS-CoV-2 main protease as demonstrated by molecular docking, molecular dynamic simulations and quantum computational studies.”

Comment 5: Page 6, line 119: add reference to SMILES notation: https://doi.org/10.1021/ci00057a005

Response 1: Following the learned reviewer’s suggestion, the SMILES notations have been added in table no. 3 before each of ligand names in the text of revised manuscript

Author Response File: Author Response.docx

Reviewer 2 Report

The theoretical calculation of the interaction parameters of natural compounds with the coronavirus spike protein is an interesting and important task, since contributes to the development of drug screening methodology at the preliminary stage of preclinical trials.

Generally, the manuscript is appropriate for publication in the Future Pharmacology, nevertheless I have some questions and comments on the methodology, which should be taken into consideration and appropriate corrections/additions should be made.

Section 2.3, lines 131-132. The pdb-structures that were used in the work have missing amino acids. Did the authors build the missing parts of the protein structures or simulate the structures with gaps?

Section 2.3, line 139. It should be described in more detail how the binding site of the studied ligands in the S-protein was chosen. What are the native ligands of the spike protein? There are no ligands in the used pdb-structures. A figure is needed, which will indicate the location of the site in the receptor molecule.

Section number 2.3. repeated twice (lines 128 and 142).

Section 2.4, line 159. It is unlikely that the box size is 1 nm. Do you mean 1 nm from the protein surface? Formulate more accurately.

Section 2.4, lines 157 and 161. Write clearly the length of equilibration and simulation.

Section 2.4. Were counter-ions added to neutralize the system?

Table 4. The abbreviations used in the table should be decrypted in the footer. The gyration radius has different abbreviations: Rgyr in Table 4 and Rg in paragraph 3.3.2.

Table 4. Parameter IntraHB is not discussed anywhere in the text. It is then necessary to remove it from the table or describe the obtained values in the text.

Figure 4. Different CoV-19 variants have substantially different numbers of amino acids in the figure. What is it connected with? For example, the pdb file-for the Omicron variant (7qo7) contains 3 subunits of approximately 1100 residues in each, that is, a total of approximately 3300 residues, however only about 2250 residues are shown in Figure 4A. Why?

Figure 4. If the missing amino acids were not built in the models, then these gaps should be somehow indicated in the figure.

Figure 5. The figure capture needs to be corrected, since when reading the capture, it is not immediately clear what characteristics are given. The caption to the figure says that these are the structural characteristics of the protein-ligand complexes, but apparently the talk is about the characteristics of the ligand inside the complex, right? The abbreviation MolSA is not deciphered anywhere, this characteristic is not discussed anywhere in the text.

Figure 6. The abbreviation ESP is not deciphered anywhere.

Author Response

We are very thankful for the valuable and positive comments on our manuscript from the reviewer that are important to improve this paper. The critical and informative remarks of the reviewers resulted in an enhanced revised manuscript. The comments are reviewed carefully and a response to each individual comment is provided in the following. We hope this paper will be accepted after this revision.

Comment-1: Section 2.3, lines 131-132. The pdb-structures that were used in the work have missing amino acids. Did the authors build the missing parts of the protein structures or simulate the structures with gaps?

Response 1: Thank you to the knowledgeable reviewer for bringing this to our attention. We concur with the observation that PDB structures of proteins under investigation have missing amino acids. For docking investigations, we employed the trimer structure of S-glycoprotein in all variants studied here in our study. After docking, we created a model with solely active residues and then ran a simulation of each complexes.

Comment-2: Section 2.3, line 139. It should be described in more detail how the binding site of the studied ligands in the S-protein was chosen. What are the native ligands of the spike protein? There are no ligands in the used pdb-structures. A figure is needed, which will indicate the location of the site in the receptor molecule.

Response 2: We agree with this point on the native (crystallographic) ligand of protein structure, which was highlighted by the honourable reviewer. Because no native (crystallographic) ligand was detected in the delta variant pdb structure, we constructed the binding site using Goodford's grid technique via the sitemap tool, which determines the active site for protein. Following that, we determined the most popular grid dimension for our further work. We can see this in the following reference

[Goodford, P. J. J. Med. Chem. 1984, 27, 557].

Comment-3: Section number 2.3. repeated twice (lines 128 and 142).

Response 3: Yes, there was a typo that has been corrected in the revised manuscript as per the reviewer's suggestion.

Comment-4: Section 2.4, line 159. It is unlikely that the box size is 1 nm. Do you mean 1 nm from the protein surface? Formulate more accurately.

Response 4: Thank you to the knowledgeable reviewer to point out the typo. It has now been changed to a phrase in the revised manuscript. The revised sentence is as follows:

“Following that, the complexes were neutralised by introducing Na+/Cl- ions to a transferable intermolecular potential 3P (TIP3P) water model with a box of 10.0 dimensions and a periodic boundary condition (PBC).”.

Comment-5: Section 2.4, lines 157 and 161. Write clearly the length of equilibration and simulation.

Response 5: Following the erudite reviewer's suggestion, the two sentences in the revised text have been properly verified, corrected, and revised as follows

At line 157, the modified sentence is

“As previously described, the simulation was performed with the OPLS-3e force field at 1000 trajectory frames of 100 ns time.”

At line 161, the modified sentence is

“The final step of simulation was conducted at constant temperature 300 K along with isobaric isothermal ensemble (NPT equilibrium) by employing Nose-Hoover chain thermostat at 300 K temperature 1 bar pressure.”

Comment-6: Section 2.4. Were counter-ions added to neutralize the system?

Response 6: Yes, we did add counter-ions, such as Na+/Cl-, to neutralise the system. We have now included this sentence to our revised manuscript in response to a learned reviewer's comment.

Comment-7: Table 4. The abbreviations used in the table should be decrypted in the footer. The gyration radius has different abbreviations: Rgyr in Table 4 and Rg in paragraph 3.3.2.

Response 7: To ensure the uniformity of the manuscript, we, the writers, thoroughly checked the abbreviation of the gyration radius. The corrected abbreviation (Rg) has been inserted to table 4, paragraph 3.3.2, as well as in the revised manuscript where necessary throughout the text.

Comment-8: Table 4. Parameter IntraHB is not discussed anywhere in the text. It is then necessary to remove it from the table or describe the obtained values in the text.

Response 8: Thank you for this constructive advice, learned reviewer. We authors agree that the IntraHB phrase was not addressed in the text. We removed the IntraHB from the table in the revised manuscript, as suggested.

Comment-9: Figure 4. Different CoV-19 variants have substantially different numbers of amino acids in the figure. What is it connected with? For example, the pdb file-for the Omicron variant (7qo7) contains 3 subunits of approximately 1100 residues in each, that is, a total of approximately 3300 residues, however only about 2250 residues are shown in Figure 4A. Why?

Response 9: Yes, we agree with the knowledgeable reviewer. The spike glycoproteins of all coronavirus variants (alpha, beta, gamma, delta, and omicron) include three subunits. First, we performed a molecular docking analysis on the spike glycoproteins' complete protein (three subunits) structures, such as the omicron version (PDB: 7qo7), which had 3300 amino acid residues. We discovered that a region of residues was inactive and had no interaction with the ligand after evaluating the docking results. To prevent a lengthy simulation run, we created a model by deleting inactive residues from all spike proteins, such as the omicron form, which includes 2250 active residues while interacting with the ligand. As a result, in Figure 4, we show the complex structure with contacts with active sites of spike proteins of all types. This pattern for simulation analysis is also described in the following research, where the scientists used active residues regardless of protein total size for further study. [https://doi.org/10.1039/D0NJ02844D]

A modified line (regarding this) has been added in the caption of figure 4 in revised manuscript

Comment-10: Figure 4. If the missing amino acids were not built in the models, then these gaps should be somehow indicated in the figure.

Response 10: Thanks to learned reviewer for the constructive suggestion. This correction has been included in the figure 4 in the revised manuscript.

Comment-11: Figure 5. The figure capture needs to be corrected since when reading the capture, it is not immediately clear what characteristics are given. The caption to the figure says that these are the structural characteristics of the protein-ligand complexes, but apparently the talk is about the characteristics of the ligand inside the complex, right? The abbreviation MolSA is not deciphered anywhere, this characteristic is not discussed anywhere in the text.

Response 11: We checked the caption of the figure as suggested by the learned reviewer. It has now been changed and corrected in the revised manuscript. We have not discussed the MolSA parameter anywhere in the paper; hence, we have removed it from Figure 5. The revised caption, as shown in the manuscript, is shown below.

Ligand properties during the complex formation at 100 ns simulation trajectory: (A) Rutin-omicron complex (B) Hesperidin-alpha complex (C) EGCG-beta complex (D) Rosmarinic acid-gamma complex (E) Rutin-delta complex.

Comment-12: Figure 6. The abbreviation ESP is not deciphered anywhere.

Response 12: We agree with the learned reviewer. It was a typo error, using MEP (molecular electrostatic potential) instead of ESP. It has now been reviewed and modified in the revised manuscript's text.

NOTE: Additional comment

Comment: English language and style are minor spell check required

Response: We have thoroughly reviewed the manuscript's text. Many errors were found that have been corrected in the updated manuscript. The changes to the english language are marked in green in the revised version of manuscript.

Author Response File: Author Response.docx