Combining Subtractive Genomics with Computer-Aided Drug Discovery Techniques to Effectively Target S. sputigena in Periodontitis

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1.      Assigning short codes for the compounds would enhance clarity and make the manuscript easier to follow.

2.      It is good to add 3D visualizations of the molecular docking results.

3.      Performing molecular docking with a reference compound targeting the same site would allow for a more robust comparative analysis.

4.      Adding contact interaction images from the dynamics simulation would strengthen the discussion by visually supporting the findings.

Author Response

  Assigning short codes for the compounds would enhance clarity and make the manuscript easier to follow.

Authors: We have used the short codes from the database used and the short codes are:(4bS,8aR)-2,4b,8,8-tetramethyl-7,10-dioxo-5,6,8a,9-tetrahydrophenanthrene-3-carboxylic acid (IMPHY005303)

Pluviatilol (IMPHY006624)

5-[3-(1,3-Benzodioxol-5-yl)-1,3,3a,4,6,6a-hexahydrofuro[3,4-c]furan-6-yl]-1,3-benzodioxole (IMPHY014895)

Gadain (IMPHY004244)

Cubebin (IMPHY001912)

1. It is good to add 3D visualizations of the molecular docking results.

Authors: following the reviewer comment we have added the 3D models of the molecular docking in the Supplementary file (Figure S10-S12).

1. Performing molecular docking with a reference compound targeting the same site would allow for a more robust comparative analysis.

Authors: we thank the reviewer for the comment. We agree with the reviewer that having reference compounds could improve the robustness of the analysis. Unfortunately, in this work, we used homology models and accordingly there are not co-crystalized ligands to use as reference compounds for comparison. We are sorry but there is no chance to have reference ligands for more comprehensive analysis.

1. Adding contact interaction images from the dynamics simulation would strengthen the discussion by visually supporting the findings.

Authors: following the reviewer comment we have added the interaction diagram related to the MD simulation of Cubebin in the Supplementary file (Figure S13A-C).

 

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript entitled “Combining Subtractive Genomics with Computer-aided Drug Discovery Techniques to Effectively Target S. sputigena in Periodontitis” addresses a relevant topic by proposing a combination of subtractive genomics and computational drug discovery to target S. sputigena in periodontitis. While the concept is promising, the manuscript requires significant revisions to improve its clarity, organization, and scientific rigor. Below, I provide my detailed comments to guide the authors in enhancing their work.

 

The writing throughout the manuscript is difficult to follow, and there are numerous typographical and formatting errors. For example, terms like in silico should be italicized consistently. Many expressions are vague, such as “the researchers observed…,” which fails to clarify whether the observations were made by the authors or in previous studies. 

 

In the introduction, the role of various bacteria, such as S. intermedius, A. actinomycetemcomitans, and others, is not clearly explained—are these beneficial, neutral, or pathogenic? This should be addressed to provide better context for the study. Additionally, the introduction should include an explanation of in silico approaches and their relevance to subtractive genomics and drug discovery.

 

Figure 1 is presented as part of the methodology, but it appears to be more of a workflow diagram summarizing the manuscript’s approach. This should be clarified and potentially moved or reframed. Similarly, the methodology section lacks critical details. The PDB codes for proteins used as templates in homology modeling should be included, and information about residue validation (such as the number of residues in allowed regions) should be moved to the results section, along with the Ramachandran plot analysis. The docking procedure is not well explained; it is unclear how many docking simulations were performed per compound. Similarly, the molecular dynamics simulations lack key details, including the number of simulations performed and their duration.

 

The results section requires better organization and presentation. For instance, the caption for Figure 2 is misleading; the Venn diagram does not represent pathways themselves but rather the number of pathways. In Figure 4, a legend is needed to clarify the colors used in the figure. The calculated binding sites are central to the study but are not adequately explained, which hinders the reader’s ability to understand this aspect of the work. Virtual screening results should be introduced with an overall distribution of docking scores before focusing on individual compounds. This would provide a clearer picture of the dataset and enhance the interpretation of the results. Differences in docking scores, such as those in Table 3, should be better analyzed—are these differences statistically significant?

 

The discussion section presents the most significant weaknesses in the manuscript. While the authors emphasize the importance of targeting S. sputigena, they fail to discuss potential side effects, such as the impact on the gut microbiome where S. sputigena is also present. Furthermore, the rationale for selecting only cytosolic proteins is not provided, nor is there any discussion on whether the identified compounds are capable of permeating the bacterial membrane. The authors state that the compounds are intended for buccal infections but analyze them using Lipinski’s rules, which are typically applied to oral drugs. This discrepancy should be addressed, as it raises questions about whether the compounds need to cross the intestinal barrier.

 

There is inconsistency in the software tools used. Calculation of interactions was performed with Discovery Studio, while molecular dynamics simulations were run with Maestro but also used for the interactions in the simulations. The comparability or complementarity of these tools should be discussed. Additionally, the authors should consider re-docking snapshots from the molecular dynamics simulations to provide a dynamic view of the ligand binding process and observe changes in docking scores over time. The discussion also fails to situate the results within the broader context of existing literature. Have similar approaches been used in studies of oral infections? How do these findings compare to those from other studies?

 

The conclusions are repetitive and largely reiterate points from the discussion without providing new insights. This section should be streamlined to highlight the main findings and their implications while acknowledging the limitations of the study.

 

In summary, this manuscript requires major revisions before it can be considered for publication. The writing must be improved for clarity and precision, the methodologies need to be detailed more comprehensively, and the discussion must critically evaluate the findings in the context of the broader literature. These changes are essential to enhance the manuscript’s scientific quality and impact.

Author Response

The manuscript entitled “Combining Subtractive Genomics with Computer-aided Drug Discovery Techniques to Effectively Target S. sputigena in Periodontitis” addresses a relevant topic by proposing a combination of subtractive genomics and computational drug discovery to target S. sputigena in periodontitis. While the concept is promising, the manuscript requires significant revisions to improve its clarity, organization, and scientific rigor. Below, I provide my detailed comments to guide the authors in enhancing their work.

Authors: we thank the reviewer for the evaluation of the work and for the suggestions to improve the manuscript. We have addressed all points raised by the reviewer in the revised version. Furthermore, we provided a tracked version in which you can easily evaluate the modifications made.

 

The writing throughout the manuscript is difficult to follow, and there are numerous typographical and formatting errors. For example, terms like in silico should be italicized consistently. Many expressions are vague, such as “the researchers observed…,” which fails to clarify whether the observations were made by the authors or in previous studies. 

Authors: Authors: we revised the manuscript to improve the readability of the manuscript to avoid confusion. Concerning the terms to italicize as in silico, mdpi police does not allow to italicize this term and for this reason was not italicize.

 

 

In the introduction, the role of various bacteria, such as S. intermedius, A. actinomycetemcomitans, and others, is not clearly explained—are these beneficial, neutral, or pathogenic? This should be addressed to provide better context for the study. Additionally, the introduction should include an explanation of in silico approaches and their relevance to subtractive genomics and drug discovery.

Authors: we thank the reviewer for the careful reading. According to the comments we added some sentences in the introduction to improve the understanding of the role of the considered microorganisms related to their pathogenic characteristics. Furthermore, we explained the computational approach as requested.

 

Figure 1 is presented as part of the methodology, but it appears to be more of a workflow diagram summarizing the manuscript’s approach. This should be clarified and potentially moved or reframed. Similarly, the methodology section lacks critical details. The PDB codes for proteins used as templates in homology modeling should be included, and information about residue validation (such as the number of residues in allowed regions) should be moved to the results section, along with the Ramachandran plot analysis. The docking procedure is not well explained; it is unclear how many docking simulations were performed per compound. Similarly, the molecular dynamics simulations lack key details, including the number of simulations performed and their duration.

Authors: We are glad about the reviewer's comment; the main idea of Figure 1 is to mention the flow of the methodology to obtaining results in a single figure. It was conceived to give a brief explanation of the workflow and methods. We usually use this approach to summarize the type of the work in the introduction Since the homology modeling results with broader results are obtained, all these details, including the sequence alignment, are mentioned in the supplementary figures file from Figure S1 to Figure S9 and they are appropriately cited in the results section as argued by the reviewer. The PDB codes were reported in the results section at the beginning when the type of proteins used as templates were described. Docking details have been added to the materials and methods. We reported the grid dimensions for the docking studies and the number of the simulation as requested (usually in each research in which there is MD simulation we perform two independent runs to confirm the MD results. However, considering that the results for the complex C9LUR0/IMPHY001912 were controversial, we confirmed the lack of the affinity of IMPHY001912 conducting an additional MD run for a total of three independent MD run.

 

 

The results section requires better organization and presentation. For instance, the caption for Figure 2 is misleading; the Venn diagram does not represent pathways themselves but rather the number of pathways. In Figure 4, a legend is needed to clarify the colors used in the figure. The calculated binding sites are central to the study but are not adequately explained, which hinders the reader’s ability to understand this aspect of the work. Virtual screening results should be introduced with an overall distribution of docking scores before focusing on individual compounds. This would provide a clearer picture of the dataset and enhance the interpretation of the results. Differences in docking scores, such as those in Table 3, should be better analyzed—are these differences statistically significant?

Authors: According to the comment in the manuscript, the required edits are made on the captions.

 

The discussion section presents the most significant weaknesses in the manuscript. While the authors emphasize the importance of targeting S. sputigena, they fail to discuss potential side effects, such as the impact on the gut microbiome where S. sputigena is also present. Furthermore, the rationale for selecting only cytosolic proteins is not provided, nor is there any discussion on whether the identified compounds are capable of permeating the bacterial membrane. The authors state that the compounds are intended for buccal infections but analyze them using Lipinski’s rules, which are typically applied to oral drugs. This discrepancy should be addressed, as it raises questions about whether the compounds need to cross the intestinal barrier.

Authors: Based on the above comment, the manuscript content is edited to the respective places mentioned. To improve the characterization of the considered hits we calculated the possibility to cross membrane by using QikProp software and we added these parameters in Table 4.

 

There is inconsistency in the software tools used. Calculation of interactions was performed with Discovery Studio, while molecular dynamics simulations were run with Maestro but also used for the interactions in the simulations. The comparability or complementarity of these tools should be discussed. Additionally, the authors should consider re-docking snapshots from the molecular dynamics simulations to provide a dynamic view of the ligand binding process and observe changes in docking scores over time. The discussion also fails to situate the results within the broader context of existing literature. Have similar approaches been used in studies of oral infections? How do these findings compare to those from other studies?

Authors: we often use different software for docking (GLIDE, GOLD, Autodock, Molegro) as described in several research papers (https://doi.org/10.1039/D0FO01511C; https://doi.org/10.1007/s11030-023-10650-6; https://doi.org/10.3390/v15122291; https://doi.org/10.1080/14786419.2022.2117177; https://doi.org/10.3390/computation10010007). To ensure the complementarity is sufficient to save the docking complex in pdb and after that they can imported in Maestro and used for MD in Desmond. The only thing to do is to prepare the complexes with protein preparation wizard that performs a check of the structure to avoid possible errors or wrong protonation state. Due to the routinary work on the docking complexes, we did not enclose this information in the main text. We apologize for the lack of these details; we have added the procedure at the beginning of the MD simulation paragraph in the Materials and Methods section. Furthermore, to obtain a dynamic view of the binding affinity, we add the calculation of deltaG binding for the entire MD trajectories. As expected, the values are in agreement with the stability of the trajectories. In fact, for the targets C9LRH1 and C9LTU7 we found a good deltaG bind that supports the possibility that IMPHY001912 (cubebin) could strongly interact with the mentioned targets. On the contrary, considering the target C9LUR0 we observed a dramatic decrease of binding affinity consistent with the lack of stability in the binding site (Table 5). Results and methods were added in the respective sections.

To the best of our knowledge the proposed approach by targeting S. sputigena combining subtractive genomics and computer-aided drug discovery is the first and accordingly there is no comparable work in the literature.

 

The conclusions are repetitive and largely reiterate points from the discussion without providing new insights. This section should be streamlined to highlight the main findings and their implications while acknowledging the limitations of the study.

Authors: The conclusion section has been checked, and some parts were rewritten to avoid repetition and to improve the readability.

 

In summary, this manuscript requires major revisions before it can be considered for publication. The writing must be improved for clarity and precision, the methodologies need to be detailed more comprehensively, and the discussion must critically evaluate the findings in the context of the broader literature. These changes are essential to enhance the manuscript’s scientific quality and impact.

Authors: Based on the study objectives and research experience acquired, we have clarified the comments listed above. We hope to have improved the manuscript according to the reviewer’s comments.

Reviewer 3 Report

Comments and Suggestions for Authors

To potentially treat oral infectious caused by S. sputigena, Praveen et al. have utilized computational methods to identify novel targets to advance drug discovery campaigns. The authors have combined computational biology methods and structure-based all-atom approaches and identified 3 unique proteins that can serve as starting points for drug discovery screening campaigns. They have also screened and assessed several natural products and identified several drug-like compounds against these targets.

 

The work has potential scientific significance. I recommend the work can be published after addressing the following comments.

 

Major:

1.    Why did the authors choose to perform homology modeling, given the advancement of cutting-edge highly accurate AlphaFold modeling? Additionally, with each Uniprot ID, AlphaFold models (of high confidence) are very conveniently and readily available for use. Have the authors tried comparing their homology models with existing AlphaFold models to gain more confidence on their models?

2.    It is very important to perform some extent of protein relaxation/equilibration of modeled homology structures before utilizing the structures for structure-based drug design work. Have the authors considered this? 

 

Minor:

1.    Please include the meaning of the various abbreviations used in Figure 1; such as B.A., PCA, RMSD and RMSF, itself in the Figure Caption.

2.    In Section 2.2.3., please include the size of the simulation water box (in Å) used in the study.

3.    In the same section, please briefly include the “default procedure”. Although, it is set to default, it is a common practice to briefly describe the “default NPT procedure”.

4.    What is the “unfavorable bond” between -OH and GLN46 in Figure 5B? Does this mean that there are steric clashes in this compound? Same question for Figure 7A (with GLY81).

5.    Please indicate in Figure 8B, the various protein domains corresponding to the RMSF regions. For example, indicating if the RMSF jumps correspond to loops and if the low RMSF correspond to helices, etc, can help the readers get a sense of how good or bad the RMSFs are. 

6.    Please indicate in all the RMSD plots captions itself, if the RMSDs are of C-alpha or all heavy atoms, etc.

Author Response

To potentially treat oral infectious caused by S. sputigena, Praveen et al. have utilized computational methods to identify novel targets to advance drug discovery campaigns. The authors have combined computational biology methods and structure-based all-atom approaches and identified 3 unique proteins that can serve as starting points for drug discovery screening campaigns. They have also screened and assessed several natural products and identified several drug-like compounds against these targets.

 

The work has potential scientific significance. I recommend the work can be published after addressing the following comments.

Authors: we thank the reviewer for the positive evaluation of the work. We have addressed all points raised by the reviewer in the revised version. Furthermore, we provided a tracked version in which you can easily evaluate the modifications made.

 

 

Major:

1. Why did the authors choose to perform homology modeling, given the advancement of cutting-edge highly accurate AlphaFold modeling? Additionally, with each Uniprot ID, AlphaFold models (of high confidence) are very conveniently and readily available for use. Have the authors tried comparing their homology models with existing AlphaFold models to gain more confidence on their models?

Authors: AlphaFold could offer remarkable accuracy but is not true in each situation. In fact in this case. we opted for homology modelling with Phyre2 due to our expertise and its robust validation process using SAVES and GalaxyRefine, ensuring reliable results. Phyre2’s pipeline, including Ramachandran plots, allowed thorough model validation. We retrieved the AlphaFold models and scrutinized structure checks, but the various residues are not in Ramachandran plot favorable regions, so we followed homology modelling.

2. It is very important to perform some extent of protein relaxation/equilibration of modeled homology structures before utilizing the structures for structure-based drug design work. Have the authors considered this? 

Authors: We did perform relaxation/equilibration of the modeled homology structures before utilizing them for structure-based drug design. The tertiary structures generated by Phyre2 were validated using the SAVES v6.0 server through Ramachandran plots. Initially, the models contained less than 90% of residues in the allowed regions, prompting further refinement through the GalaxyRefine server. Post-refinement, the structures exhibited over 90% of residues in the allowed regions, ensuring their structural stability and suitability for subsequent drug design analyses.

 

Minor:

1. Please include the meaning of the various abbreviations used in Figure 1; such as B.A., PCA, RMSD and RMSF, itself in the Figure Caption.

Authors: we thank the reviewer for the suggestion. We have added the meaning of the abbreviations in Figure 1 caption and in the other captions if necessary.

 

2. In Section 2.2.3., please include the size of the simulation water box (in Å) used in the study.

Authors: the size of the simulation box was added for each complex in section 2.2.3 as requested.

 

3. In the same section, please briefly include the “default procedure”. Although, it is set to default, it is a common practice to briefly describe the “default NPT procedure”.

Authors: the default procedure in Desmond to equilibrate the system before MD consists of several constrained minimizations and MD simulations to gradually relax the system and to the equilibrium. This info was added in the experimental section regarding the MD simulation. The parameters for MD simulation were already reported in the same section.

 

4. What is the “unfavorable bond” between -OH and GLN46 in Figure 5B? Does this mean that there are steric clashes in this compound? Same question for Figure 7A (with GLY81).

Authors: We thank the reviewer for this comment. The reason for the unfavourable interactions are mentioned in the revised version at respective places.

 

 

5. Please indicate in Figure 8B, the various protein domains corresponding to the RMSF regions. For example, indicating if the RMSF jumps correspond to loops and if the low RMSF correspond to helices, etc, can help the readers get a sense of how good or bad the RMSFs are. 

 Authors: We thank the reviewer for this comment. We added sentences regarding the RMSF and the regions that are more flexible. As expected, the regions corresponded to a loop that connected two helices. In fact, often during MD the loop regions are more flexible as well as the head and the tail of the protein. Notably this region is not close to the binding site.

 

6. Please indicate in all the RMSD plots captions itself, if the RMSDs are of C-alpha or all heavy atoms, etc.

Authors: RMSD usually is calculated on C-alpha. We added this detail in the caption of MD Figures.

 

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have significantly enhanced the manuscript in response to the feedback provided, and the improvements are evident throughout the text. The revised version demonstrates a clear and concise presentation of the research, addressing previous concerns effectively. The methodological details are now thorough, the data analysis is robust, and the discussion provides a well-rounded interpretation of the results in the context of existing literature.

Overall, the manuscript has reached a high standard of scientific rigor and clarity, making it suitable for publication in its current form. I commend the authors for their efforts in refining the manuscript. 

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have successfully addressed the concerns raised during the first round of peer-review process. I recommend the article to be published in the current form.