Deciphering the Molecular Interaction Process of Gallium Maltolate on SARS-CoV-2 Main and Papain-Like Proteases: A Theoretical Study

Round 1

Reviewer 1 Report

In this manuscript, Taype-Huanca et al. investigated the metal complex Gallium Maltolate (GaM) as a potential antiviral drug against SARS CoV-2’s viral protease PLpro and Mpro. GaM has already been tested to show moderate effectiveness in inhibiting SARS-Cov-2 replication in cell culture. This manuscript provided mechanistic explanation of GaM’s inhibitory effect at a molecular scale through calculating GaM’s reactivity and binding using DFT calculation and molecular docking, respectively. Collectively, these calculations showcase GaM's ability to bind and interact with PLpro and Mpro with affinities comparable to a known inhibitor, N3, thus suggesting GaM as a promising precursor drug. Overall, the calculations were carefully designed and performed, yielding clear results for the most part. The manuscript would benefit from more elaborate discussion of the simulation data and some proofreading. I support the publication of this manuscript with these concerns being fully addressed.

1. My primary concern is related to the molecular docking data of GaM. Previous publication has concluded that N3 peptidyl inhibitor could form 9 hydrogen bonding interactions in the active site of Mpro (PMID 34163906). Compared to N3, GaM forms far fewer hydrogen bonds with Mpro, yet displays similar binding energy. How would the calculation justify this effect? In addition, the authors mentioned that van der Waals forces and hydrophobic interactions also played an important factor for GaM’s binding. To what extent do these non-directional interactions contribute to the stabilization of GaM’s binding as compared to the H-bond? With that being said, this manuscript most definitely requires a better discussion to justify GaM’s binding modality, and elaborate comparisons among the bindings of GaM, N3, and TTT.

2. In Figure 2, the amino acids labels do not align with the Table 3. Additionally, the interactions demonstrated in Figure 2 differ from the molecular interactions discussed by the authors in the main text. These mistakes could significantly undermine the credibility of this manuscript. I strongly suggest that the authors remake this entire figure.

3. The authors mentioned that the H-bond between Cys145 and GaM is critical for GaM’s binding. It would be better if the authors could specify whether this H-bond is between on the backbone of Cys145 or a sulfur H-bond, as the latter is not a canonical H bond therefore needs to be specified. In addition, how does this H-bond on cystine affect it nucleophilicity? As this cystine is critical to most covalent drug targets.

4. In page 2, line 47, the author mentioned the recent development of metal complexes drugs with different center metal ions. The reference, however, only included the ones with Ni(II), Mn(II) and Pd(II). Additional references are needed to elaborate on the recent progress of metal complexes as anti-cancer and anti-viral drugs. Similarly, references for N3 (page 3, line 114) should also be included.

5. The manuscript would benefit from some careful proofreading. Below, I list some issues I picked up, but there are more:

a. Page 2, line 85: missing period sign.

b. Page 3, line 119: Mpro. Pro should be superscript.

c. There are two Table 2. Page 9, line 352: unit of this table is missing.

d. Page 10, line 393: the caption of sub-figure B in Fig. 2 is labeled twice, and sub-figure D is missing.

Author Response

In this manuscript, Taype-Huanca et al. investigated the metal complex Gallium Maltolate (GaM) as a potential antiviral drug against SARS CoV-2’s viral protease PLpro and Mpro. GaM has already been tested to show moderate effectiveness in inhibiting SARS-Cov-2 replication in cell culture. This manuscript provided mechanistic explanation of GaM’s inhibitory effect at a molecular scale through calculating GaM’s reactivity and binding using DFT calculation and molecular docking, respectively. Collectively, these calculations showcase GaM's ability to bind and interact with PLpro and Mpro with affinities comparable to a known inhibitor, N3, thus suggesting GaM as a promising precursor drug. Overall, the calculations were carefully designed and performed, yielding clear results for the most part. The manuscript would benefit from more elaborate discussion of the simulation data and some proofreading. I support the publication of this manuscript with these concerns being fully addressed.

1. My primary concern is related to the molecular docking data of GaM. Previous publication has concluded that N3 peptidyl inhibitor could form 9 hydrogen bonding interactions in the active site of Mpro (PMID 34163906). Compared to N3, GaM forms far fewer hydrogen bonds with Mpro, yet displays similar binding energy. How would the calculation justify this effect? In addition, the authors mentioned that van der Waals forces and hydrophobic interactions also played an important factor for GaM’s binding. To what extent do these non-directional interactions contribute to the stabilization of GaM’s binding as compared to the H-bond? With that being said, this manuscript most definitely requires a better discussion to justify GaM’s binding modality, and elaborate comparisons among the bindings of GaM, N3, and TTT.

Author reply: Dear reviewer. We agree with your comment, in which you mention that N3 forms a large number of hydrogen bonds, however, all with distances longer than 2.5 A. It is mentioned in the literature that there is a relationship between the hydrogen bonds formed by an inhibitor in the active site of a protein and its efficiency to inhibit that protein. Hydrogen bonds contribute significantly to the binding energy and affinity of the inhibitor for the protein. Therefore, a higher number of hydrogen bonds generally correlates with a higher binding affinity and more potent inhibition (https://www.science.org/doi/10.1126/sciadv.1501240, https://doi.org/10.1002/prot.10259). But also the distance of the bond is relevant, the closer the donor is to the proton acceptor, the higher the binding energy. By analyzing the Mpro/GaM complex, two bonds of less than 2.5 A (2.0 and 2.5 A, Fig) can be identified, which may compensate for the smaller number of hydrogen bonds in the compound. Furthermore, it is important to note that, in addition to hydrogen bonds, other factors such as hydrophobic interactions, electrostatic interactions and lower steric hindrance due to the smaller size of GaM also contributes in binding affinity and inhibitor efficiency (https://doi.org/10.1371/journal.pone.0012029, https://doi.org/10.1039/D0CP05714B). Therefore, although GaM does not form a large number of hydrogen bridges, it possesses other types of interactions that favor binding to the Mpro site. It should be noted that optimal interactions depend on a proper balance between hydrogen bonds and other types of intermolecular interactions

2. In Figure 2, the amino acids labels do not align with the Table 3. Additionally, the interactions demonstrated in Figure 2 differ from the molecular interactions discussed by the authors in the main text. These mistakes could significantly undermine the credibility of this manuscript. I strongly suggest that the authors remake this entire figure.

Author reply: We agree with the reviewer. It was an error in placing the order of Figure 2. Consequently, we have made the respective changes marked in yellow in the text of the manuscript. The order of Figure 2 was changed so that it is in accordance with the text and so that the amino acid labels can fit, where Mpro (A-B) and PLpro (C-D) are mentioned first. A complete revision of the text was made so that the molecular interactions discussed are correct and in accordance with Figure 2 and Table 3.

3. The authors mentioned that the H-bond between Cys145 and GaM is critical for GaM’s binding. It would be better if the authors could specify whether this H-bond is between on the backbone of Cys145 or a sulfur H-bond, as the latter is not a canonical H bond therefore needs to be specified. In addition, how does this H-bond on cystine affect it nucleophilicity? As this cystine is critical to most covalent drug targets.

Author reply:  Dear reviewer. I would like to elaborate in detail on the interaction between GaM and Cys145 that we have identified. This specific interaction is located between the carbon of the aromatic ring of the maltolate and the hydrogen of the sulfur atom of Cys145. Importantly, this interaction is characterized as an unconventional hydrogen bridge with the carbon. This finding is supported by previous relevant studies in the scientific literature, which have addressed similar issues. For example, in the reference work (https://doi.org/10.1021/jp907747w), the nature of molecular interactions between specific compounds and amino acid residues in proteins is discussed. Likewise, in another study (https://doi.org/10.1007/s41745-019-00145-5), the importance of ligand-protein interactions in understanding the biological activity of protein complexes is further explored. In addition, the formation and characteristics of hydrogen bonds in different chemical contexts have been extensively investigated, as described in (https://doi.org/10.1016/S0022-328X(98)00661-5). This additional information underlines the relevance and uniqueness of the interaction we have observed between GaM and Cys145, providing a solid basis for our conclusions.

Regarding the effect of this H-bond on the nucleophilicity of Cys145, cysteine residues are known to play a crucial role in the catalytic mechanisms of many enzymes, including covalent drug targets. The nucleophilic sulfhydryl group of cysteine can attack electrophilic centers, forming covalent bonds that inhibit enzymatic activity. If the H-bond affects the backbone of Cys145, it is less likely to significantly affect the nucleophilicity of the sulfhydryl group. However, if the H-bond directly affects the sulfur atom of the cysteine side chain, it could modulate the nucleophilicity of this group by altering its electronic environment and reactivity (https://doi.org/10.1016/j.cbpa.2015.11.004). In our case this interaction is given by the carbon of the aromatic ring and the sulfur hydrogen of Cys145, it is an unconventional hydrogen bridge with the carbon.

For further analysis, we have performed an analysis using the IGM methodology, which does not help to have a better visualization of this interaction.

The promolecular Independent Gradient Model (https://doi.org/10.1002/jcc.26812), with the Multiwfn program (https://doi.org/10.1002/jcc.22885 ). IGM is a new method proposed for visually analyzing intramolecular and intermolecular interactions in chemical systems. It builds upon the existing Independent Gradient Model (IGM) method by replacing the free-state atomic densities used in IGM with atomic densities obtained from promolecular approximation of the molecular density. This gives IGM a more rigorous physical basis. IGM has advantages over other popular visualization methods like Non-covalent Interaction (NCI) plots. It can separately visualize intrafragment and interfragment interactions with smoother and less jagged isosurfaces.

The IGM uses a similar color scale that NCIplotcolor to visualize and characterize different types of molecular interactions based on the δg descriptor and the sign(λ2)ρ function. The δg descriptor measures deviations from the exponential decay of the electron density due to interactions between atoms/fragments. Non-zero values of δg correspond to interaction regions.

The sign(λ2)ρ function, where λ2 is the second largest eigenvalue of the electron density Hessian matrix, is used to distinguish the nature of the interactions:

• sign(λ2)ρ > 0 (blue color) indicates strong attractive interactions like hydrogen bond or covalent bonds.
• sign(λ2)ρ ≈ 0 (green color) indicates weak van der Waals interactions
• sign(λ2)ρ < 0 (red color) indicates repulsive steric clashes

The typical color scale used in IGM is shown below, going from blue (attractive) to red (repulsive) with green as the intermediate.

In the image you can appreciate the interaction that is given by the carbon of the aromatic ring and the hydrogen of the sulfur of the Cys145, it is an unconventional hydrogen bridge with the carbon, the blue coloration in the isosurface can be appreciated.

In addition, in order to perform the studies and see if indeed a bond formation between GaM and Cys145 occurs, it would be necessary to perform QM/MM hybrid method studies, which would be another type of work in relation to what we have done in this one.

4. In page 2, line 47, the author mentioned the recent development of metal complexes drugs with different center metal ions. The reference, however, only included the ones with Ni(II), Mn(II) and Pd(II). Additional references are needed to elaborate on the recent progress of metal complexes as anti-cancer and anti-viral drugs. Similarly, references for N3 (page 3, line 114) should also be included.

Author reply: We added the suggested references in lines 47-52.

5. The manuscript would benefit from some careful proofreading. Below, I list some issues I picked up, but there are more:
6. Page 2, line 85: missing period sign.

Author reply: We agree with the reviewer. Consequently, we have made the respective changes marked in yellow in the text of the manuscript.

1. Page 3, line 119: Mpro. Pro should be superscript.

Author reply: We agree with the reviewer. Consequently, we have made the respective changes marked in yellow in the text of the manuscript

2. There are two Table 2. Page 9, line 352: unit of this table is missing.

Author reply: We agree with the reviewer. Consequently, we have made the respective changes marked in yellow in the text of the manuscript

1. Page 10, line 393: the caption of sub-figure B in Fig. 2 is labeled twice, and sub-figure D is missing.

Author reply: We agree with the reviewer. Consequently, we have made the respective changes marked in yellow in the text of the manuscript

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript is an interesting work dealing with the investigation of the molecular interactions process of Gallium 2 Maltolate (GaM) on SARS-CoV-2.

THE WHOLE WORK IS INTERESTING!

POINTS FOR IMPROVEMENT:  

1. Please avoid using abbreviations in the title or the abstract.

2. Please provide a Nomenclature Section explaing abrreviations and symbols

3. Please, provide qualitative or quantitative evidence of the simulations by comparing the results of this work with experimental data or other colleagues results

4. Please, ellaborate more on how the parameters of the model were selected (e.g. compare the selected parameters  with the parameters of other colleagues)

5. Please, propose applications of this work.

6. Please, provide a brief description of the software used.

7. Please rewrite the manuscript in order to improve the originality as calculated by IThenticate.

Author Response

This manuscript is an interesting work dealing with the investigation of the molecular interactions process of Gallium 2 Maltolate (GaM) on SARS-CoV-2.

THE WHOLE WORK IS INTERESTING!

POINTS FOR IMPROVEMENT:  

1. Please avoid using abbreviations in the title or the abstract.

Author reply: We agree with the reviewer. Consequently, we have made the respective changes marked in yellow in the text of the manuscript.

2. Please provide a Nomenclature Section explaing abrreviations and symbols

Author reply: We agree with the reviewer. Consequently, we have made the respective changes marked in yellow in the text of the manuscript.

3. Please, provide qualitative or quantitative evidence of the simulations by comparing the results of this work with experimental data or other colleagues results.

Author reply: Dear reviewer. Thank you for your inquiry about comparing our simulation results with experimental data or findings of other investigators. We thank you for the opportunity to address this important aspect of our study.

We have compared our simulation results with experimental data from Professor Bernstein of the USA.. Professor Bernstein in June 2020 filed his patent on Gallium maltolate (GaM), which already has been tested in several Phase 1 clinical trials and is also doesn’t consider to cause severe toxicity (even taken at a high daily oral dose for several months). The strong inhibition (half-maximal effective dose, EC50 = 14 μM) obtained by GaM for viral replication was examined in SARS-CoV-2 Vero E6 cells. No cytotoxicity was reported for GaM for uninfected cells at concentrations up to at least 200 μM, through Cell Counting Kit-8 (CCK-8) colorimetric assay (https://patents.justia.com/patent/11213530). The comparison suggests that GaM could be a promising antiviral candidate against SARS-CoV-2, the virus responsible for COVID-19 (https://doi.org/10.1177/2040206620983780 ; https://doi.org/10.2174/1871526522666220127120617). The molecular docking and chemical reactivity results indicate that GaM can effectively bind and potentially inhibit the major protease (Mpro) and papain-like protease (PLpro) of SARS-CoV-2, which are crucial enzymes for viral replication. In addition, we have compared our results with those obtained for the molecules 5-amino-2-methyl-N-[(1R)-1-naphthalen-1-ylethyl] benzamide (TTT) y N-[(5-methyl-1,2-oxazol-3-yl)carbonyl]-L-alanyl-L-valyl-N-{(2S,3E)-5-(benzyloxy)-5-oxo-1-[(3S)-2-oxopyrrolidin-3-yl]pent-3-en-2-yl}-L-leucinamide (N3) in work we have previously published (https://doi.org/10.3390/ph15080986 ; https://doi.org/10.1016/j.biopha.2021.111764 ; https://doi.org/10.3389/fchem.2020.595097 ) and coincide with the parameters and results calculated in these published papers.4. Please, ellaborate more on how the parameters of the model were selected (e.g. compare the selected parameters  with the parameters of other colleagues).

4. Please, ellaborate more on how the parameters of the model were selected (e.g. compare the selected parameters  with the parameters of other colleagues).

Author reply:  Dear reviewer. We sincerely thank you for taking the time to review our work and provide us with your valuable comments. The parameters for the calculations performed, for molecular docking and chemical reactivity calculations, were previously modeled by us in previous publications (https://doi.org/10.3390/ph15080986; https://doi.org/10.1016/j.biopha.2021.11176  ; https://doi.org/10.3389/fchem.2020.595097). All parameters and settings are based on these publications by our group in the area of interest of this manuscript.

To perform the Docking and chemical reactivity calculations, we used the protocol employed in previous articles (https://doi.org/10.3390/ph15080986; https://doi.org/10.1016/j.biopha.2021.11176  ; https://doi.org/10.3389/fchem.2020.595097 ), in which we studied the same SARS-Cov2 protease and predicted the affinity of new drugs to this enzyme. For the Docking case we optimize the structures with a PBE0/Def2-TZVP level of theory, a widely used calculation for metal-organic structures, which delivers stable conformations, and then the Gasteiger charges that are employed in a large number of studies of this type. For the case of reactivity indices, the PBE0/Def2-TZVP level of theory was used with an implicit IEF-PCM water model.

5. Please, propose applications of this work.

Author reply: Dear Reviewer. We sincerely thank you for taking the time to review our work and provide your valuable comments. In order to provide you with a more complete view on the application of the work, we would like to complement the information by mentioning the main applications of our work in the area.

Our research delved into exploring potential antiviral compounds against SARS-CoV-2, the virus responsible for the COVID-19 pandemic. Through extensive molecular docking simulations and analysis, our findings shed light on the promising candidacy of Gallium Maltolate (GaM) as an effective antiviral agent. The molecular docking results, conducted using state-of-the-art computational techniques, unveil GaM's ability to intricately bind to key viral proteins essential for replication, notably the main protease (Mpro) and papain-like protease (PLpro) of SARS-CoV-2. These enzymes play pivotal roles in the viral life cycle, making them prime targets for antiviral intervention. The strong and specific binding affinity exhibited by GaM suggests its potential to inhibit the enzymatic activity of Mpro and PLpro, thereby impeding viral replication and propagation. Our study provides a robust rationale for further experimental validation of GaM's antiviral activity against SARS-CoV-2. The development of GaM as a viable antiviral candidate holds promise in augmenting the arsenal of therapeutic options against the ongoing global health crisis posed by COVID-19.

In addition to identifying GaM as a potential antiviral agent against SARS-CoV-2, our research delved into detailed mechanistic studies using advanced computational techniques. Non-covalent interaction (NCI) analysis and molecular electrostatic potential (MEP) calculations were employed to gain profound insights into the intricate molecular interactions and reactivity patterns of GaM with the target viral proteins. The NCI analysis offers a comprehensive understanding of the non-covalent interactions present within the GaM-SARS-CoV-2 protein complexes. By dissecting the nature and strength of these interactions, we elucidate the key binding motifs and structural features crucial for GaM's binding affinity and inhibitory potential. Furthermore, MEP calculations provide invaluable information regarding the electrostatic properties of GaM and its interactions with the target viral proteins. These calculations elucidate the distribution of charge density and electrostatic potential surfaces, offering insights into the regions of favorable interaction and potential sites for molecular recognition. Beyond characterizing GaM's interactions with SARS-CoV-2 proteins, these computational techniques hold immense potential for broader applications in antiviral drug discovery. By systematically analyzing the mechanisms of action and binding modes of GaM and other potential inhibitors, we can unravel crucial structure-activity relationships and identify key molecular determinants driving antiviral efficacy. This deeper understanding paves the way for rational drug design efforts aimed at optimizing the potency, selectivity, and safety profiles of novel antiviral agents.

This work lays a solid foundation for the exploration of organometallic compounds, particularly gallium-based complexes, as potential therapeutic agents against SARS-CoV-2 and other viral or disease targets. The computational approaches employed can be extended to various applications in drug discovery and development, leveraging the power of in silico methods to accelerate the identification and optimization of promising lead candidates.

6. Please, provide a brief description of the software used.

Author reply: Dear Reviewer. We sincerely thank you for taking the time to review our work and provide your valuable comments. In order to provide you with a more complete overview of the resources and tools employed in the realization of this project, we would like to supplement the information with a detailed list of each software used, as well as a brief description of its function within the research and development process.

Below, you will find further details on the softwares used:

• Gaussian16: This is a computational chemistry software used for performing density functional theory (DFT) calculations to optimize molecular geometries and compute reactivity descriptors like HOMO, LUMO, electronegativity, hardness, etc.
• AutoDock 4.0: A widely used molecular docking suite for predicting the preferred orientation and binding affinity of small molecules (ligands) to their protein targets.
• AutoDockTools (ADT): A graphical user interface for setting up AutoDock docking calculations, including preparing ligand and protein structures, defining the docking grid, etc.
• NCIPLOT: A program for visualizing and analyzing non-covalent interactions between molecules based on the reduced density gradient (RDG) and electron density.
• TAFF: The Topological Analysis of Fukui Function pipeline has been designed in order to “condense” the Fukui function according to it’s topological analysis. This function is defined as the electronic density response when modifying the number of electrons (N), providing information about the reactive sites
• Protein Preparation Wizard (Schrödinger suite): A tool for preparing and refining protein structures for molecular modeling studies, including adding hydrogens, assigning charges, and minimizing energy.
• VMD: Visual molecular dynmamics, is used for visualizing and analyzing molecular structures.

7. Please rewrite the manuscript in order to improve the originality as calculated by IThenticate.

Author reply: Dear Reviewer. We sincerely thank you for taking the time to review our work and provide your valuable comments. The manuscript was completely revised to avoid complications with the IThenticate software.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The author promptly responded to all answers raised. The manuscript is accepted,