|Year : 2022 | Volume
| Issue : 4 | Page : 333-338
Drug repurposing against angiotensin-converting enzyme-related carboxypeptidase (ACE2) through computational approach
Golnaz Vaseghi1, Ali Golestaneh2, Leila Jafari3, Fahimeh Ghasemi4
1 Isfahan Cardiovascular Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran
2 Applied Physiology Research Center, Cardiovascular Research Institute, School of Pharmacy and Pharmaceutical Sciences, Isfahan University of Medical Sciences, Isfahan, Iran
3 Department of Bioinformatics and Systems Biology, School of Advanced Technologies in Medicine, Isfahan University of Medical Sciences, Isfahan, Iran
4 Medical Image and Signal Processing Research Center, Isfahan University of Medical Sciences, Isfahan, Iran, Iran
|Date of Submission||07-Sep-2020|
|Date of Decision||29-Oct-2021|
|Date of Acceptance||03-Feb-2022|
|Date of Web Publication||10-Nov-2022|
Department of Bioinformatics and Systems Biology, School of Advanced Technologies in Medicine, Isfahan University of Medical Sciences, Isfahan
Source of Support: None, Conflict of Interest: None
Ongoing novel coronavirus (COVID-19) with high mortality is an infectious disease in the world which epidemic in 2019 with human-human transmission. According to the literature, S-protein is one of the main proteins of COVID-19 that bind to the human cell receptor angiotensin-converting enzyme 2 (ACE2). In this study, it was attempted to identify the main effective drugs approved that may be repurposed to the binding site of ACE2. High throughput virtual screening based on the docking study was performed to know which one of the small-molecules had a potential interaction with ACE2 structure. Forasmuch as investigating and identifying the best ACE2 inhibitors among more than 3,500 small-molecules is time-consuming, supercomputer was utilized to apply docking-based virtual screening. Outputs of the proposed computational model revealed that vincristine, vinbelastin and bisoctrizole can significantly bind to ACE2 and may interface with its normal activity.
Keywords: Angiotensin-converting enzyme 2, computer simulation, coronavirus disease 19, drug repurposing, high-throughput virtual screening
|How to cite this article:|
Vaseghi G, Golestaneh A, Jafari L, Ghasemi F. Drug repurposing against angiotensin-converting enzyme-related carboxypeptidase (ACE2) through computational approach. J Med Signals Sens 2022;12:333-8
|How to cite this URL:|
Vaseghi G, Golestaneh A, Jafari L, Ghasemi F. Drug repurposing against angiotensin-converting enzyme-related carboxypeptidase (ACE2) through computational approach. J Med Signals Sens [serial online] 2022 [cited 2022 Dec 8];12:333-8. Available from: https://www.jmssjournal.net/text.asp?2022/12/4/333/360847
| Introduction|| |
In December 2019, an unknown respiratory disease emerged in Wuhan, Hubei province, China, which is now known as coronavirus disease 19 (COVID-19). A novel coronavirus is closely related to severe acute respiratory syndrome coronavirus (SARS-CoV). SARS-CoV-2 binds to human angiotensin-converting enzyme 2 (ACE2) which may lead to severe pneumonia and lung fibrosis in patients. ACE2 is a cell membrane enzyme which expresses in the outer surface of cells, mainly in the lungs, and converts angiotensin II into angiotensin 1–7. It has been shown that exogenous Ang-(1–7) and upregulation of the ACE2 may protect against lung fibrosis by blocking the MAPK/NF-κB pathway. On the other hand, SARS-CoV-2 spike protein entrance site is different from the active site of ACE enzyme. ACE2 is a receptor to help SARS-S to entry to human cell, using the cellular serine protease TMPRSS2 for S protein priming.
Few numbers of articles have been published about ACE2 using computational approach. For example, identification of exact amino acid residues in the place of interaction of S-protein with ACE2 was investigated to develop antiviral inhibitor by Zhang et al. in 2005. Pharmacophore model and virtual screening was the other model that was published by Rella et al. in 2005. Simulation S-protein in complex with ACE2 and their interaction with four host species-specific receptors was another computational manuscript that was published by Zhang et al. in 2007, and changing conformational active site and reducing level of ligand binding was the other model considered by Lokeshwari et al. in 2015.
Besides, numerous computational methods have been proposed to find the best lead compound as a de novo drug candidate for other homo-sapience targets since 1980. As evidenced by the recent publications in drug discovery, the search for finding drug-like compounds with desired biological activities in the large libraries of chemical compounds such as ZINC, called high-throughput virtual screening (HTVS), is a really risky, time-consuming procedure as well as low success percentages. Drug repurposing or repositioning is one of the proposed virtual screening approaches which refers to rediscover a narrow the list of drug candidates which already passed safety tests in clinical trials and a great chance of taking the desired activity via computational approaches. Various computational approaches have been used in drug repositioning which can be categorized into three main groups: (i) ligand-based models, i.e., machine learning-based models, (ii) target-based screening such as docking-based method, and (iii) network-based methods.
Since March 2020, various manuscripts were published based on mentioned approaches to find out the best appropriate drugs among existed ones. Some of the main of them are reviewed in the following text. Basu et al. focused on some of the natural products and their effects on the ACE2. In another study, it was tried to find compounds which can bind to spike, ACE2, and the ACE2:spike complex with good affinity. Joshi et al. focused on the main effective inhibitors of ACE2 among natural compounds. In another study, the current efforts of exploiting ACE2 as a therapeutic target were reviewed via Jia et al.
Machine learning-based models depend on the known drug-target interaction. Shallow learning methods such as k-nearest neighbor and deep learning model, for instance, convolutional neural network, have been the main techniques which exploited to achieve drug repurposing prediction.,, Target-based virtual screening, most of the time, is based on the docking studies with supercomputer, and the best top-ranked drugs can be tested via molecular dynamic simulation or experimental test. Moreover, some of the scientists, efforts to find out drug-disease associations and drug-target interactions simultaneously with network-based methods.
In this study, it was attempted to figure out the best approved drugs to inhibit ACE2 binding site based on the drug re-purposing methodology. Hence, more than 3500 existed drugs in DrugBank website were downloaded and drug-target interaction for each drug was simulated via high-throughput docking virtual screening. Finally, the best identified effective drugs were extracted and grouped according to their categories [Figure 1]. We hypothesized that binding proposed drug to the ACE2 may exacerbate pulmonary fibrosis.
| Materials and Methods|| |
Receptor and ligand preparation
ACE2 receptor structure was downloaded from Protein Data Bank (PDB) website (PDB ID: 1r4l, resolution 3 Å) and its three-dimensional conformation was refined with AutoDock4 software, such as add hydrogen and remove nonpolar hydrogen and add total Kollman charge. ACE2 binding site was chosen around ZN (x = 37, y = 5, z = 25, with search box 60 × 60 × 60) set its binding site
Besides, as the aim of this work was identifying the best drugs to inhibit COVID-19, a library of approved drugs, found in Supplementary [Table S1], were downloaded from DrugBank website in xml format. The version of database is 5.1.5 released on January 3, 2020, and contained more than 3000 approved drugs. After that, xml format was changed to the sdf and then all molecules were extracted as a pdb formatted file. After that, because of the lack of hydrogen in the converted structures, all molecular hydrogens were added with Open Babel software, automatically. Besides, one of the main steps in computational drug design is optimizing 3D structures of ligands. There are several tools available to generate 2D/3D structure/conformers, but, because of existing huge number of molecules as well as time-consuming optimization procedure, Open Babel was utilized to optimize 3D structure, automatically, with the aim of finding low-energy conformations via conformer searching. In this software, accessing several algorithms for conformer searching is performed by Gen3D library with steepest descent geometry optimization and the MMFF94 force field. Finally, to confirm the optimization procedure, some of the structures were optimized with HyperChem software and the results were compared with Open Babel outputs. Finally, all optimized structures were converted to pdbqt format via Open Babel software for docking studies.
High-throughput docking-based virtual screening
HTVS methods have been one of the main computational drug design approaches to rapidly investigate hundreds number of chemical compounds which could be appropriate for finding the best drug-like molecules. One of the common computational methods in HTVS is molecular docking. The main goal of molecular docking is given a ligand-target interaction in two main steps: (i) minimizing ligand conformation in receptor binding site and (ii) scoring these conformations. Vina AutoDock is one of the docking methods useful for HTVS by using multithreading on multicore machines and widely utilized for drug repurposing. The main drawback of Vina docking is that did not contain flexible binding sites residues.
Structure-based virtual screening
Structure-based virtual screening (SBVS) has been one of the main computational drug design approaches based on the simulation ligand-target interaction for identification of hit molecules. Molecular docking is one of the SBVS methods with the aim of optimized matching prediction of ligand orientation according to the 3D structures of receptors. The main goal of molecular docking is investigating affinity and binding energy of DTI. Finally, all suggested conformers are clustered according to the computed free energies and grouped together by scoring function.
| Results and Discussion|| |
As mentioned, this study was founded on two different parts, i.e., HTVS and molecular docking, which were discussed in the following by details.
High-throughput virtual screening
To perform the docking study, two main input datasets must be prepared, containing approved drugs and protein crystallography structure. For the first one, all FDA-approved drugs (more than 3500 molecules) were downloaded from DrugBank website and their structures were optimized (discussed in material section). 3D structure of ACE2 was the second essential information which was downloaded from PDB, i.e., ID 1r4l, with resolution After that, protein PDB structure was rectified with AutoDock4 software; then, molecular docking simulation was performed between drugs and receptor via AutoDock Vina software. According to the root mean square error (RMSE), the best conformers of each ligand were extracted, and finally, the ligands were sorted based on their affinities. More than 200 drugs had an affinity with score better than 11 Kcal/Mol. Hence, ligands with pose scores of more than 15 Kcal/Mol and weight <1000 (g/mol) were selected (corresponding to 30 molecules, i.e., about the top 10% compounds) [Table 1]. Among these drugs, some of them are used widely, especially in patients with underlying disease such as antiviral. As shown, it was figured out seven different potent drugs – with ΔG <−15 (μM) and weight <700 (g/mol) – that could be effective to inhibit ACE2 enzyme activity.
Before to consider molecular interaction precisely, we were interested to investigate proposed drug performance in the body, i.e., their side effects and mechanism of actions which were extracted from DrugBank website [Table 2].
|Table 1: Fifteen percent of top-ranked extracted drug after docking simulation based on virtual screening|
Click here to view
At the second step, seven suggested molecules were docked to the receptor via AutoDock4. There are two main steps for docking procedure, i.e., setting the search space and optimization and docking procedure. In the first step, choosing appropriate grid box to search in 3D space of protein is a critical point which conducted researcher to have reliable decision. Three different parameters are vital which must be adjusted that are selecting suitable box center, number of points in each dimension, and spacing between the points. To achieve the best selection for box center, existed ligand in protein crystallography was extracted via Schrodinger software and docked again via blind docking method. The result was matched with experimental approach. Its binging energy was 10.7 (μM). Hence, the center of grid box was defined on the center of ligands (x = 40.12, y = 1.32, z = 23.68) in crystallography of proteins with size 60 × 60 × 60 and 0.375°A. After that, according to the defined atoms, probes were serially located at each grid point and internal energy between the probs and protein were computed for each atom type, individually. Computed energies for each point were utilized as a lookup table during the docking simulation.
The second step was docking procedure with the aim of finding out binding energy of each ligand as well as its interaction with the target. Thus, Lamarckian genetic algorithm (LGA) was utilized to extract the population of ligand conformations, randomly. In LGA method, two optimization methods, genetic algorithm and local search, are combined to enhance docking performance. Van der Waals potentials and a dihedral angle term are two main critical parameters to calculate internal energy. The results of interaction-binding energies and interactions are summarized in [Table 3]. As shown, indinavir, retapamulin, and saquinavir have the lowest binding energies (<−12.5 [μM]).
|Table 3: The binding energy of molecular docking computed with AutoDock4|
Click here to view
Besides, the interactions between suggested drugs, indinavir, retapamulin, and saquinavir, and protein are illustrated in [Figure 2]a, [Figure 2]b, [Figure 2]c. As demonstrated, various amino acids were contributed on the interaction, which are helpful to increase binding affinities.
|Figure 2: ACE2 in complex with (a) Indinavir, (b) Retapamulin, (c) Saquinavir|
Click here to view
| Conclusion|| |
High-blood pressure (hypertension) has been one of the main regular situations reported in most patients with severe illness in COVID-19. The main worry in the medical treatment of these conditions, such as using RAAS inhibitors, is occurring adverse outcomes which emerge as the robust estimator of COVID-19-related death. Hence, hypertension condition has been key determined prognostic. In the recent studies, it was released that spike protein of coronaviruses binds to the human receptor in the cell surface, i.e., ACE2 which its expression is increased in the patients with type 1 or type 2 diabetes. Hence, it can be concluded that ACE2 expression and activity have a significant rule in SARS-CoV-2 patients. Therefore, the drugs with high affinity for ACE2 enzymatic site should be prescribed with caution in these patients
In this study, it was attempted to find the main ACE2 inhibitors among 3981 approved drugs downloaded from DrugBank website. The receptor structure was downloaded from PDB website (PDB ID: 1r4l, resolution 3 Å) and its three-dimensional conformation was refined with AutoDock4 software, such as add hydrogen and remove nonpolar hydrogen and add total Kollman charge. To evaluate docking results, firstly, the ligands in the 1r4l crystallography were docked again. The results revealed that selected centers of grid box were appropriate. In HTVS step, approved drugs were docked to the receptor via AutoDock Vina. The best conformers of each ligand were selected based on the RMSE, and finally, the ligands were sorted based on their affinities. Ligands with pose scores <20 (Kcal/Mol) were extracted and are shown in [Table 1]. After that, to find the best inhibitors, drugs with weights <700 (g/mol) were investigated via AutoDock4. Outputs of the proposed computational model revealed that indinavir, retapamulin, and saquinavir can significantly bind to ACE2 and may interface with its normal activity.
This project was supported by the vice chancellery of research at Isfahan University of Medical Sciences (Contract grant number: 198220).
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Xu YH, Dong JH, An WM, Lv XY, Yin XP, Zhang JZ, et al.
Clinical and computed tomographic imaging features of novel coronavirus pneumonia caused by SARS-CoV-2. J Infect 2020;80:394-400.
Hamming I, Timens W, Bulthuis ML, Lely AT, Navis G, van Goor H. Tissue distribution of ACE2 protein, the functional receptor for SARS coronavirus. A first step in understanding SARS pathogenesis. J Pathol 2004;203:631-7.
Keidar S, Kaplan M, Gamliel-Lazarovich A. ACE2 of the heart: From angiotensin I to angiotensin (1-7). Cardiovasc Res 2007;73:463-9.
Meng Y, Yu CH, Li W, Li T, Luo W, Huang S, et al.
Angiotensin-converting enzyme 2/angiotensin-(1-7)/Mas axis protects against lung fibrosis by inhibiting the MAPK/NF-κB pathway. Am J Respir Cell Mol Biol 2014;50:723-36.
Towler P, Staker B, Prasad SG, Menon S, Tang J, Parsons T, et al.
ACE2 X-ray structures reveal a large hinge-bending motion important for inhibitor binding and catalysis. J Biol Chem 2004;279:17996-8007.
Zhang Y, Zheng N, Hao P, Cao Y, Zhong Y. A molecular docking model of SARS-CoV S1 protein in complex with its receptor, human ACE2. Comput Biol Chem 2005;29:254-7.
Rella M, Rushworth CA, Guy JL, Turner AJ, Langer T, Jackson RM. Structure-based pharmacophore design and virtual screening for novel angiotensin converting enzyme 2 inhibitors. J Chem Inf Model 2006;46:708-16.
Zhang Y, Zheng N, Nan P, Cao Y, Hasegawa M, Zhong Y. Computational simulation of interactions between SARS coronavirus spike mutants and host species-specific receptors. Comput Biol Chem 2007;31:134-7.
Lokeshwari D, Krishna Kumar NK, Manjunatha H. Multiple mutations on the second acetylcholinesterase gene associated with dimethoate resistance in the melon aphid, Aphis gossypii
). J Econ Entomol 2016;109:887-97.
Valera-Vera EA, Sayé M, Reigada C, Miranda MR, Pereira CA. In silico
repositioning of etidronate as a potential inhibitor of the Trypanosoma cruzi
enolase. J Mol Graph Model 2020;95:107506.
Zhu Z, Wang X, Yang Y, Zhang X, Mu K, Shi Y, et al
. D3Similarity: A Ligand-Based Approach for Predicting Drug Targets and for Virtual Screening of Active Compounds Against COVID-19. ChemRxiv. Cambridge: Cambridge Open Engage; 2020; This content is a preprint and has not been peer-reviewed.
Talluri S. Molecular Docking and Virtual Screening based prediction of drugs for COVID-19, Combinatorial Chemistry & High Throughput Screening 2020;23:1. DOI: 10.2174/1386207323666200814132149.
Zhou Y, Hou Y, Shen J, Huang Y, Martin W, Cheng F. Network-based drug repurposing for novel coronavirus 2019-nCoV/SARS-CoV-2. Cell Discov 2020;6:14.
Basu A, Sarkar A, Maulik U. Molecular docking study of potential phytochemicals and their effects on the complex of SARS-CoV2 spike protein and human ACE2. Sci Rep 2020;10:17699.
Joshi T, Joshi T, Sharma P, Mathpal S, Pundir H, Bhatt V, et al. In silico
screening of natural compounds against COVID-19 by targeting Mpro and ACE2 using molecular docking. Eur Rev Med Pharmacol Sci 2020;24:4529-36.
Jia H, Neptune E, Cui H. Targeting ACE2 for COVID-19 therapy: Opportunities and challenges. Am J Respir Cell Mol Biol 2021;64:416-25.
Ghasemi F, Mehridehnavi A, Pérez-Garrido A, Pérez-Sánchez H. Neural network and deep-learning algorithms used in QSAR studies: Merits and drawbacks. Drug Discov Today 2018;23:1784-90.
Zhang W, Xu H, Li X, Gao Q, Wang L. DRIMC: An improved drug repositioning approach using Bayesian inductive matrix completion. Bioinformatics 2020;36:2839-47.
Gao K, Nguyen DD, Wang R, Wei GW. Machine intelligence design of 2019-nCoV drugs. bioRxiv 2020;1-16, doi: 10.1101/2020.01.30.927889.
Smith M, Smith JC. Repurposing Therapeutics for COVID-19: Supercomputer-Based Docking to the SARS-CoV-2 Viral Spike Protein and Viral Spike Protein-Human ACE2 Interface. ChemRxiv. Cambridge: Cambridge Open Engage; 2020; This content is a preprint and has not been peer-reviewed.
Guy JL, Jackson RM, Acharya KR, Sturrock ED, Hooper NM, Turner AJ. Angiotensin-converting enzyme-2 (ACE2): Comparative modeling of the active site, specificity requirements, and chloride dependence. Biochemistry 2003;42:13185-92.
Dhasmana A, Raza S, Jahan R, Lohani M, Arif JM. High-throughput virtual screening (HTVS) of natural compounds and exploration of their biomolecular mechanisms: An in silico approach. In: Ahmad Khan MS, Ahmad I, Chattopadhyay D, editors. New Look to Phytomedicine. Ch. 19. Advancements in Herbal Products as Novel Drug Leads: Academic Press; 2019. p. 523-48.
Zhang DH, Wu KL, Zhang X, Deng SQ, Peng B. In silico
screening of Chinese herbal medicines with the potential to directly inhibit 2019 novel coronavirus. J Integr Med 2020;18:152-8.
Meng XY, Zhang HX, Mezei M, Cui M. Molecular docking: A powerful approach for structure-based drug discovery. Curr Comput Aided Drug Des 2011;7:146-57.
Trott O, Olson AJ. AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem 2010;31:455-61.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3]