In Silico Approach: Anti-Tuberculosis Activity of Caespitate in the H37Rv Strain

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The author had investigated Caespitate, a compound from Helichrysum caespititium, for its anti-tubercular potential against Mtb. Molecular docking reveals interactions with key Mtb enzymes involved in cell wall formation such as InhA, MabA, UGM, PanK. Conformational analysis favors a specific structure (CS) for binding PanK and UGM, maintaining stability via intramolecular hydrogen bonds. Pharmacokinetic evaluation suggests favorable absorption and bioavailability, indicating potential as an anti-tuberculosis agent.

Here are some comments for improvement:

It would be great if the scheme-1 is provided with legends so that readers will easily understand it, including the explaining what red, blue, and green dots is with lines.

Author Response

Response to the reviewers 

We express our gratitude to the reviewers for their diligent revisions and insightful comments, which greatly contributed to enhancing this work. The revised manuscript incorporates additional details and modifications, all of which are highlighted in yellow. Below, we provide responses to each comment. 

 

Reviewer 1 

The author had investigated Caespitate, a compound from Helichrysum caespititium, for its anti-tubercular potential against Mtb. Molecular docking reveals interactions with key Mtb enzymes involved in cell wall formation such as InhA, MabA, UGM, PanK. Conformational analysis favors a specific structure (CS) for binding PanK and UGM, maintaining stability via intramolecular hydrogen bonds. Pharmacokinetic evaluation suggests favorable absorption and bioavailability, indicating potential as an anti-tuberculosis agent. 

 

Comment 1 

It would be great if the scheme-1 is provided with legends so that readers will easily understand it, including the explaining what red, blue, and green dots is with lines. 

 

Answer to Comment 1 

In Scheme 1, the lines with dot have been changed to arrows. The caption was supplemented by emphasizing the meaning of the red, blue, green, and yellow arrows. It can be seen in the attachment, and in page 3 of the new version of the manuscript. 

 
 

Scheme 1. Chemical reactions catalysed by the four enzymes proposed with distinct localization: those involving the InhA and MabA enzymes take place within the cell wall of Mtb, specifically in the mycolic acids (red and blue arrows), while the reaction catalysed by UGM is situated in the arabinogalactans (green arrow). Additionally, the reaction mediated by PanK contributes to the bacterium's growth process (yellow arrow). 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

 

The work of Moreno-Ceballos et al. shows an in silico approach to assess a potential anti-tuberculosis drug, caespitate, which was found in plant used in traditional South African medicine. The manuscript is clearly written, the results are clear, however I would like the authors to reflect on the following questions/comments taken as minor revisions:

1. The target enzymes have been selected due to their physiological role or absence in humans, but that selection may not be sufficient. What is also important is the molecular structure of the enzyme active site and the reaction mechanism. The authors could check for example sequence similarity or RMSD values between the selected enzymes (or their active site cavities) to ensure that the targets are structurally different.

2. Why were water molecules removed from the enzymes? Water molecules could provide some crucial interactions for the binding. On the other hand, the description of the docking procedure is unclear, as the authors don’t say anything about water molecules when using Glide software.

3. The authors seem to suggest that the lowest binding energy obtained from both Vina and Glide indicate the best target protein from their set, which may not be the case. Scoring functions used in docking software may favor compounds able to form the highest number of interactions, and thus they are not comparable for a set of target proteins. This means that the scoring function can be used to compare the binding of a set of ligands to one target protein and thus selecting the “best binder”, but not the other way around – it will not indicate the target protein to which the compound binds “best”.

Author Response

Response to the reviewers

We express our gratitude to the reviewers for their diligent revisions and insightful comments, which greatly contributed to enhancing this work. The revised manuscript incorporates additional details and modifications, all of which are highlighted in yellow. Below, we respond to each comment.

 

Reviewer 2

The work of Moreno-Ceballos et al. shows an in silico approach to assess a potential anti-tuberculosis drug, caespitate, which was found in plant used in traditional South African medicine. The manuscript is clearly written, the results are clear, however I would like the authors to reflect on the following questions/comments taken as minor revisions.

Comment 1

The target enzymes have been selected due to their physiological role or absence in humans, but that selection may not be sufficient. What is also important is the molecular structure of the enzyme active site and the reaction mechanism. The authors could check for example sequence similarity or RMSD values between the selected enzymes (or their active site cavities) to ensure that the targets are structurally different.

Answer to Comment 1

Multiple alignment of sequence in Clustal Omega (https://www.ebi.ac.uk/jdispatcher/msa/clustalo) and, pairwise alignment of structure in PDBeFold tool (https://www.ebi.ac.uk/msd-srv/ssm) of the chosen enzymes were performed. The results show no relevant similarity or functional or evolutionary relationship between them.

The sequence alignment output with the binding sites highlighted with different colors is presented below and on page 3 of Supplementary Information, Figure S1.

 

 

 

 

 

 

 

 

CLUSTAL O(1.2.4) multiple sequence alignment


sp|P9WPA7|COAA_MYCTU ------------------------------------------------------MSR---         3
sp|P9WIQ1|GLF_MYCTU --------MQPMTARFDLFVVGS---GFFGLTIAERVATQLDKRVLVLERRP----HIGG 45
sp|P9WGT3|MABA_MYCTU MTATATEGAKPPFVSRSVLVTGGN--RGIGLAIAQRLAADGHKVAVTH--------R-GS     49
sp|P9WGR1|INHA_MYCTU MTG--------LLDGKRILVSGIITDSSIAFHIARVAQEQGAQLVLTGFDRLRLIQRITD         52
 :

sp|P9WPA7|COAA_MYCTU --LSEPSP--YVEF-----------DRRQWRALRMSTPLALTEEE---------LVGLRG      39
sp|P9WIQ1|GLF_MYCTU NAYSEAEPQTGIEVHKYGAHLFHTSNKRVWDYVRQFTDFTDYRHRVFAMHNGQAYQFPMG         105
sp|P9WGT3|MABA_MYCTU G-APKGLF--GVECDVTDSDAVDRAFTAVEEH-------QGPVEV---------LVSNAG           90
sp|P9WGR1|INHA_MYCTU R-LPAKAP--LLELDVQNEEHLASLAGRVTEAIGAGNKLDGVVHS-------IGFMPQTG      102
 :* . *

sp|P9WPA7|COAA_MYCTU LGEQIDLLEVEEVYLPLARLIHLQVAARQRLF------AATAEFLGEPQQNPDRPVPFII      93
sp|P9WIQ1|GLF_MYCTU LGLVSQFFGKYFTPEQARQLIAEQ-AAEIDTADAQNLEEKAISLIGRP-----LYEAFVK        159
sp|P9WGT3|MABA_MYCTU LSAD-AFLMRM-TEEKFEKVINANLTGAFRVA-QRASRSMQRNKFGRM-----IFIGSVS    142
sp|P9WGR1|INHA_MYCTU MGIN-PFFDA-----PYADV-----SKGIHIS-AYSYASMAKALLPIM-----NPGGSIV  145
 :. :: : : : :

sp|P9WPA7|COAA_MYCTU GVAGSVAVGKSTTARVLQALLARWDHHPRVDLVTTDGFLYPNAELQRRNLMHRKGFPESY      153
sp|P9WIQ1|GLF_MYCTU G-----------------YTAKQWQTDPKELPA---ANITR--------LPVRYTFDNRY           191
sp|P9WGT3|MABA_MYCTU G-----SWG---------IGNQANYAASKAGVIGMARSIAR--------ELSKANVTANV  180
sp|P9WGR1|INHA_MYCTU GMDFDPSRA---------MPAYNWMTVAKSALESVNRFVAR--------EAGKYGVRSNL         188
 * : : : .

sp|P9WPA7|COAA_MYCTU -----NRRALMRFV-------------------------TSVKSGSDYACAPVYSHLHYD         183
sp|P9WIQ1|GLF_MYCTU FS----------DTYEGLPTD--------GYTAWLQNMAADHRI---------EVRLNTD         224
sp|P9WGT3|MABA_MYCTU VAPGYIDTDMTRAL-----DE----RI---QQGALQFIPAK-RVGTPAEVAGVVSFLASE        227
sp|P9WGR1|INHA_MYCTU VAAGPIRTLAMSAIVGGALGEEAGAQIQLLEEGWDQRAPIGWNMKDATPVAKTVCALLSD        248
 . * :

sp|P9WPA7|COAA_MYCTU IIPGAEQVVRHPDILILEGLNVLQTGPTLMVSDLFDFSLYVDARIEDI------EQWYVS      237
sp|P9WIQ1|GLF_MYCTU WFDVRGQLRPG-----------SPAAPVVYTGPLDRYFDYAEGRLGWRTLDFEVEVLPIG         273
sp|P9WGT3|MABA_MYCTU DASYI-----------------------------SGAVIPVDGGMGMGH-----------     247
sp|P9WGR1|INHA_MYCTU WLPAT-----------------------------TGDIIYADGGAHTQLL----------      269
 .:.

sp|P9WPA7|COAA_MYCTU RFLAMRT---TAFADPESHFHHYAAFSDSQAV-----VAAREI----------WRTINR-  278
sp|P9WIQ1|GLF_MYCTU DFQGTAVMNYNDLDVPYTRIHEFRHFHPERDYPTDKTVIMREYSRFAEDDDEPYYPINTE 333
sp|P9WGT3|MABA_MYCTU ------------------------------------------------------------           247
sp|P9WGR1|INHA_MYCTU ------------------------------------------------------------ 269
 

sp|P9WPA7|COAA_MYCTU -------------------------------------------PNLVENILPTRPRATLV           295
sp|P9WIQ1|GLF_MYCTU ADRALLATYRARAKSETASSKVLFGGRLGTYQYLDMHMAIASALNMYDNVLAPHLRDGVP         393
sp|P9WGT3|MABA_MYCTU ------------------------------------------------------------           247
sp|P9WGR1|INHA_MYCTU ------------------------------------------------------------ 269
 

sp|P9WPA7|COAA_MYCTU LRKDADHSINRLRLRKL      312
sp|P9WIQ1|GLF_MYCTU LLQDGA-----------     399
sp|P9WGT3|MABA_MYCTU -----------------       247
sp|P9WGR1|INHA_MYCTU -----------------        269

 

The analysis of the 3D structures revealed that there is no alignment of the α- and β-secondary structures in all enzymes except MabA and InhA (RMSD values between 2.06 and 5.18 Å). The pairwise structure alignment results are shown in the next figure, which can be found in Supplementary Information on page 4, Figure S2.

 

 

 

The four proteins are represented in grey ribbons and the active sites in surface rendering, InhA orange, MabA blue UGM iceblue, and PanK yellow. A. InhA-MabA, B. InhA-UGM, C. InhA-PanK D. PanK-MabA, E. PanK-InhA, F. UGM-MabA.

Sequence and structural alignment will be discussed in the manuscript on page 4 as follows:

In addition to selecting therapeutic targets for their important physiological function or absence in humans, multiple sequence alignment was performed using Clustal Omega tool (https://www.ebi.ac.uk/jdispatcher/msa/clustalo). This multiple alignment was performed between InhA, MabA, UGM, and PanK proteins. The results showed that there is no relevant similarity or functional or evolutionary relationship between the selected proteins, see Figure S1.

On the other hand, a structural alignment was carried out with the PDBeFold tool (https://www.ebi.ac.uk/msd-srv/ssm/), which allows the identification of the similarity of secondary structures between proteins. This structural alignment was performed between the four proteins. To ensure that the selected targets were structurally distinct, a structural alignment was performed obtaining RMSD values between 2.06 and 5.18 Å. However, analysis of the 3D structures revealed that there is no alignment of the alpha- and beta-folded secondary structures in all enzymes except MabA and InhA, see Figure S2.

 

 

Comment 2

Why were water molecules removed from the enzymes? Water molecules could provide some crucial interactions for the binding. On the other hand, the description of the docking procedure is unclear, as the authors don’t say anything about water molecules when using Glide software.

Answer to Comment 2

The conservation of water molecules in the protein structure depends on the location and origin of these molecules. Sometimes they can be found inside a protein binding pocket, and they definitively must be conserved to model the interactions between them and organic cofactors, ligands, or a metal. No one of the proteins studied in this work presents this condition.

The water molecules from the solvent used in the experimental procedure (crystallization solvent) are not important in the simulation and they must be removed [1-4].

The next lines will be added to clarify the Glide software methodology in the manuscript, see page 4.

The structures of the ligands and the enzymes were prepared using the program default steps. Water molecules and non-standard residues were removed using the Protein Preparation Wizard and the corresponding partial charges were assigned

 

Comment 3

The authors seem to suggest that the lowest binding energy obtained from both Vina and Glide indicate the best target protein from their set, which may not be the case. Scoring functions used in docking software may favor compounds able to form the highest number of interactions, and thus they are not comparable for a set of target proteins. This means that the scoring function can be used to compare the binding of a set of ligands to one target protein and thus select the “best binder”, but not the other way around – it will not indicate the target protein to which the compound binds “best”.

 

Answer to Comment 3

This comment is very valuable. We agree that the scoring functions are designed to compare affinity among several ligands and one protein. However, in this study, we compare from semiflexible docking, the affinity of two conformers of the ligand (caespitate structure from gas phase calculation, caespitate structure from solvation phase with DMSO solvent calculation) and ligand form crystal structure for each protein. Additionally, we show a tendency of these affinities from the four proteins between Glide and Vina software.

 

The fundamental purpose of this study is to gain an initial and broad insight into the mechanism of action of caespitate by examining interactions with residues and assessing the affinity of the compound for specific proteins identified in strain H37Rv. Caespitate has demonstrated biological activity with this strain, indicating that complexes exhibiting optimal protein-ligand interaction could be selected as experimental candidates for validation [5]. Although this approach is not as conventional as standard docking methods, it offers a broader perspective on the investigated compound's mechanism of action and pharmacological potential. In this way, we are interested in the design of a selective drug based on caespitate structure directed towards specific targets of the H37Rv strain.

 

 

References

[1] Huang, W.J.; Blinov, N.; Wishart, D.S.; Kovalenko, A. Role of Water in Ligand Binding to Maltose-Binding Protein: Insight from a New Docking Protocol Based on the 3D-RISM-KH Molecular Theory of Solvation. J. Chem. Inf. Model. 2015, 55, 2, 317–328. https://doi.org/10.1021/ci500520q   

[2] Wojciechowski, M. Simplified AutoDock force field for hydrated binding sites. J. Mol. Graph. Model. 2017, 78,74-80. https://doi.org/10.1016/j.jmgm.2017.09.016.

[3] Ross, G.A.; Morris, G.M.; Biggin, P.C. Rapid and Accurate Prediction and Scoring of Water Molecules in Protein Binding Sites. PLOS ONE. 2012, 7(3): e32036. https://doi.org/10.1371/journal.pone.0032036

[4] Forli, S.; Huey, R.; Pique, M. et al. Computational protein-ligand docking and virtual drug screening with the AutoDock suite. Nat. Protoc. 2016, 11, 905–919. https://doi.org/10.1038/nprot.2016.051

 [5] Tang, Y.; Zhu, W.; Chen, K.; Jiang, H. New technologies in computer-aided drug design: toward target identification and new chemical entity discovery. Drug Discov. Today Technol. 2006, 3(3):307-313. https://doi.org/10.1016/j.ddtec.2006.09.004

 

 

 

Sincerely,

 

 

Dra. Norma A. Caballero

Laboratorio de Química y Biología Computacional

Facultad de Ciencias Biológicas

Benemérita Universidad Autónoma de Puebla

Puebla, México.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

In the manuscript "In Silico Approach: Anti-tuberculosis Activity of Caespitate in H37Rv Strain”, Moreno-Ceballos et al proposed inhibition modes of InhA, MabA, UGM, enzymes involved in the formation of bacterial cell wall, and PanK in cell growth, by caespitate through computational analysis. Molecular docking calculations, MM/GBSA analysis, and ADME parameters evaluation were performed.

The current manuscript should be revised significantly for publication.

Given InhA, MabA, UGM, and PanK as targets, the author must compare their structures and inhibition parameters to the human counterparts for specificity.

Although docking was carried out, whether their binding is located within the active site must be clearly sketched. The author also needs to compare the enzyme complex to current model.

Although ADME parameters were listed, this reviewer needs two common important factors, mutagenicity and hERG factor for evaluating druggable potential.

Author Response

Response to the reviewers

We express our gratitude to the reviewers for their diligent revisions and insightful comments, which greatly contributed to enhancing this work. The revised manuscript incorporates additional details and modifications, all of which are highlighted in yellow. Below, we respond to each comment.

 

Reviewer 3

In the manuscript "In Silico Approach: Anti-tuberculosis Activity of Caespitate in H37Rv Strain”, Moreno-Ceballos et al proposed inhibition modes of InhA, MabA, UGM, enzymes involved in the formation of the bacterial cell wall, and PanK in cell growth, by caespitate through computational analysis. Molecular docking calculations, MM/GBSA analysis, and ADME parameters evaluation were performed.

 

Comment 1

Given InhA, MabA, UGM, and PanK as targets, the author must compare their structures and inhibition parameters to the human counterparts for specificity.

Answer to Comment 1

A comprehensive sequence similarity search was performed using the UniProtKB, UniProtKB/Swiss-Prot, UniProtKB/Swiss-Prot (SV), UniProtKB/TrEMBL, UniProtKB/RefProt/Swiss-Prot, UniProtKB/COVID-19 and ChEMBL databases in FASTA tool (https://www.ebi.ac.uk/jdispatcher/sss/fasta).

For this analysis, BLOSUM 80 punctuation matrix with a gap penalty value of -12, and BLOSUM 50 punctuation matrix with default parameters were used. No evidence was found for the expression of InhA, MabA, and PanK enzymes in the human organism, using BLOSUM 80 punctuation matrix, which currently precludes comparison with their human counterparts. However, in the search with BLOSUM 50 punctuation matrix, for the UGM enzyme one enzyme in a human organism was found (MAO-A), showing 30% identity, which is not very relevant because this identity is not found in the active site of both proteins. Moreover, both proteins belong to different families.

 

The web links to the results are placed below:

1. InhA

BLOSUM 80

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240518-034444-0243-89197384-p1m

 

BLOSUM 50

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240523-043244-0243-75796772-p1m

 

2. MabA

BLOSUM80

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240518-034330-0570-98373176-p1m

BLOSUM50

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240523-042611-0970-37732427-p1m

3. PanK

BLOSUM80

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240518-034211-0575-46354390-p1m

BLOSUM50

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240523-041653-0482-8295145-p1m

4. UGM

BLUSOM80

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240517-184051-0466-35380967-p1m

BLUSOM50

https://www.ebi.ac.uk/jdispatcher/sss/fasta/summary?jobId=fasta-I20240523-013841-0332-53762106-p1m

Comment 2

Although docking was carried out, whether their binding is located within the active site must be clearly sketched. The author also needs to compare the enzyme complex to current model.

Answer to Comment 2

Images of the active site of each of the proteins are presented below and have already been added to the manuscript, see page 10.

 

Figure 6. 3D graphical representation of the initial complex models with the complexes obtained in Vina and Glide. A. UGM enzyme, B. PanK enzyme, C. MabA enzyme, D. InhA enzyme.

 

Comment 3

Although ADME parameters were listed, this reviewer needs two common important factors, mutagenicity and hERG factor for evaluating druggable potential.

Answer to Comment 3

The two important common factors proposed by reviewer 3 have already been incorporated into the manuscript [55]. According to the results, caespitate exhibits non-toxicity in the Ames test, therefore may not cause DNA changes due to mutagenicity. The values presented in Table 6 are within acceptable ranges for inhibition of the inwardly rectifying potassium channel (HERG).

Next paragraphs are added to the Manuscript:

Materials and Methods section, 2.3. Analysis of pharmacokinetic parameters (ADME and Ames), page 5.

On the other hand, the Ames test was carried out with AdmeSAR 2.0 [55] prediction software, which provides information about drug toxicity. SMILES of completed ligands were entered into the program to assess their toxicity, specifically concerning the Ames test [56] prediction to verify whether the compound could induce DNA changes.

In the Results section, 3.3 ADME and Ames, page 14:

Finally, the Ames toxicity shows that caespitate is a non-toxic compound, suggesting that this molecule may not induce mutagenicity. In addition, its values (see Table 6) are within acceptable ranges for inhibition of the inwardly rectifying potassium channel (hERG), which is associated with possible adverse cardiovascular effects.

 

Table 6. Evaluation of the ADME and drug-like properties of caespitate conformers using QikProp and AdmetSAR.

Pharmacological property	CG	CS	Pharmacological property	CG	CS
MWa	322.4	322.4	QPlogBBf	−1.2	−1.3
Donors HBb	1.0	1.0	QPlogSg	−4.1	−4.0
Acceptors HBc	4.3	4.3	Rule Of Five	0.0	0.0
SASAd	595.4	589.0	#metabh	7.0	7.0
QPlogPo/we	3.2	3.2	%Human Oral Absortioni	94.8	93.9
QPloghERGj	−4.5	−4.7	AMES Toxicityk	Non-toxic AMES/0.58	Non-toxic AMES/0.58

^aMolecular weight (acceptable range from <500); ^bHydrogen bond donor (acceptable range from  ≤ 5); ^cHydrogen bond acceptor (acceptable range from ≤ 10); ^dTotal solvent accessible surface area in square angstroms using a probe with a 1.4 radius (acceptable range from 300 to 1000); ^ePredicted octanol/water partition coefficient (acceptable range from −2 to 6.5); ^fPredicted blood-brain partition coefficient (acceptable range from −3 to 1.2); ^gPredicted aqueous solubility, S in mol/dm^-3 (acceptable range from −6.5 to 0.5); ^hNumber of likely metabolic reactions (acceptable range from 1 to 8); ⁱPredicted human oral absorption on 0 to 100% scale (<25% is poor and >80% is high). ^jPredicted IC₅₀ value for blockage of hERG K⁺ channels (concern below –5). ^kThe predictive property can be classified as “toxic AMES” or “non-toxic AMES”. Values ranging from 0 to 1 quantify the certainty of the prediction, providing a measure of confidence in the assessment.

 

The next references were incorporated:

[55] Cheng F.; Li W.; Zhou Y.; Shen J.; Wu Z.; Liu G.; Lee P.W.; Tang Y. admetSAR: a comprehensive source and free tool for assessment of chemical ADMET properties. J Chem Inf Model. 2012; 52(11):3099-105. https://doi.org/10.1021/CI300367A.

[56] Modi, S.; Li J.; Malcomber, S.; Moore, C.; Scott, A.; White, A.; Carmichael, P. Integrated in silico approaches for the prediction of Ames test mutagenicity. J Comput Aided Mol Des. 2012 Sep;26(9):1017-33. https://doi.org/10.1007/s10822-012-9595-5.

 

 

Sincerely,

 

Dra. Norma A. Caballero

Laboratorio de Química y Biología Computacional

Facultad de Ciencias Biológicas

Benemérita Universidad Autónoma de Puebla

Puebla, México.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed most issues. This manuscript is acceptable for publication.