Identification of potential inhibitors of SARS-CoV-2 S protein–ACE2 interaction by in silico drug repurposing

Background: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a new coronavirus discovered that appeared in Wuhan, China, in December 2019, causes COVID-19 disease which have resulted in cases similar to SARS-atypical pneumonia. Worldwide, around 116 million cases and 2.57 million deaths are reported with new cases and increasing mortality every day. To date, there is no specific commercial treatment to control the infection. Repurpose drugs targeting the angiotensin-converting enzyme 2 (ACE2) receptor represents an alternative strategy to block the binding of SARS-CoV-2 protein S and forestall virus adhesion, internalization, and replication in the host cell. Methods: We performed a rigid molecular docking using the receptor binding domain of the S1 subunit of S protein (RBD S1)-ACE2 (PDB ID: 6VW1) interaction site and 1,283 drugs FDA approved. The docking score, frequency of the drug in receptor site, and interactions at the binding site residues were used as analyzing criteria. Results: This research yielded 40 drugs identified as a potential inhibitor of RBD S1-ACE2 interaction. Among the inhibitors, compounds such as ipratropium, formoterol, and fexofenadine can be found. Specialists employ these drugs as therapies to treat chronic obstructive pulmonary disease, asthma and virtually any respiratory infection. Conclusions: Our results will serve as the basis for in vitro and in vivo studies to evaluate the potential use of those drugs to generate affordable and convenient therapies to treat COVID-19.


Introduction
Emerging viruses can be defined as those whose incidence has increased in the last twenty years or whose presence has a high probability of increasing in the near future. Diseases caused by emerging viruses are one of the biggest public health threats globally 1 . Some of the viruses that fall within this catalog are the avian influenza virus subtype H5N1, severe acute respiratory syndrome (SARS), Ebola, Zika, and MERS-CoV, to name a few 2 . Coronaviruses (CoVs) are classified into four genera, α-CoV, β-CoV, γ-CoV, and δ-CoV2. α and β infect mammals, γ birds and δ birds and mammals, respectively 3 . These viruses are of public health importance because they cause enteric, renal, and neurological respiratory diseases that range from asymptomatic to fatal 4,5 .
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) appeared in Wuhan, China, in December 2019, causing cases of SARS-like atypical pneumonia 6,7 , with a clinical picture of fever, general malaise, dry cough, shortness of breath and was called the coronavirus disease 2019 (COVID-19) 8 . It can be asymptomatic, develop mild-to-severe symptoms, or may cause death in patients with chronic diseases, such as hypertension, diabetes, and obesity 9 . On January 31 st 2020, the World Health Organization (WHO) declared COVID-19 a public health emergency of international concern and, on March 12 th , it was declared a global pandemic 10 . In Mexico, local transmission (phase 2 of transmission) was declared on March 24 th 2020, which resulted in the suspension of non-essential activities in the country, generating economic losses in addition to public health problems and deaths associated with the disease. As of March 1 st 2021, Mexico had reached 2.1 million cases of COVID-19 and 186 thousand deaths; around 116 million cases and 2.57 million deaths have been reported worldwide.
To date, there is no specific commercial treatment to control the infection [11][12][13] . Measures such as early detection, blocking the route of transmission through social isolation, isolation of suspected cases, disinfection of objects, as well as frequent hand washing with soap, in addition to the use of biosafety equipment such as surgical masks for health personnel, may reduce the transmission of COVID-19 among population.
Coronaviruses, such as SARS-CoV-2, are positive-stranded RNA viruses enveloped on a membrane. The coronaviral genome is composed of approximately 30,000 nucleotides containing the envelope (E), membrane (M), spike (S), nucleocapsid (N) and ORFs, that encode non-structural proteins, including enzymes that appear during their in-host reproductive cycle-genes 14 .
This virus measures 70 to 100 nm and belongs to the genus β-CoV 15 and it has been proposed that any of the aforementioned proteins that make up CoVs may be targets for the development of vaccines or drugs 4 . Protein S plays an essential role on COVID-19 infection as it mediates the internalization on host cell and for the spread of the virus in the infected host 16,17 . This starts when the receptor binding domain of the S1 subunit (RBD S1 ) of S protein binds to the peptidase domain of angiotensin-converting enzyme 2 receptor (ACE2) 18 and it is know that disrupting the binding of S protein to ACE2 prevents the attaching an the later internalization of the virus to the host cell 19 .
This protein interaction has recently been crystallized and deposited in the Protein Data Bank database 20 , allowing us to use it as a model of study to test different strategies to counter SARS-CoV-2 infection, like blocking S glycoprotein-ACE2 interaction through the discovery of sites of potential pharmaceutical interest.
In 2019, Research and Development (R&D) spending in the pharmaceutical industry totaled 186 billion U.S. dollars globally and its projected to reach 233 billion U.S. dollars to 2026. Unfortunately, drug development takes large time and financial resources that not all countries possess, especially developing countries, like Mexico.
In this sense, drug repurposing or repositioning allow us to integrate all evidence, pharmacodynamics/kinetics, bioavailability, among other important parameters, from an existing and approved drug in order to manage emerging diseases, like COVID-19. All this translates into a considerable decrease in research time and investment of resources in R&D 21 .
Different approaches have been taken in order to disrupt SARS-CoV-2 protein S-ACE2 interaction, as an example, many works has focus on finding potential biding sites on protein S structure, however, new variant strains has been detected worldwide, like B117 in UK, P1351 in South Africa, P1 and P2 in Brazil. All variant strains display the N501Y mutation, which is located on the RBD of the S protein, making the interaction more effective. In this sense, targeting RBD may be a transitory approach, therefore, an alternative strategy would be aiming at the ACE2 receptor. Some authors have pointed out some concerns about using drugs that targeting the renin-angiotensin signaling (RAS) pathway 22-24 , but Jia and collaborators 25 highlight current efforts of exploiting ACE2 as therapeutic target, like the use of pseudo-ligands to dominate the binding site for SARS-CoV-2 as an example. Therefore, inhibition of the SARS-CoV-2 protein S-ACE2 interaction trough aiming ACE2 receptor it is a plausible strategy. In this study, we screened a library consisting of 1,283 FDA-approved drugs and acquired by Ministry of Health of Mexico in order to identify potential inhibitors of SARS-CoV-2-ACE2 interaction.

Methods
Molecular modeling, electric partial-charge assignation, ligand conformer, searching of potential binding sites, energy minimizations, visualization and docking were performed with Molecular Operating Environment package 26 .

Amendments from Version 1
Edits to the manuscript focus on an improved and better organization on the abstract, methods and discussion sections, accordingly to reviewers suggestions. New references were cited on manuscript to support Introduction and Discussion sections in order to addresses the reviewers comments.
Any further responses from the reviewers can be found at the end of the article

Ligand preparation
The chemical structure of 1,283 drugs that comprises the updated list of reference drugs, as well as the National Compendium of Health Supplies of Mexico (June 2020 update) were obtained from the DrugBank, ZINC15 and PubChem database in September 2020. In order to simulate ligand flexibility for our rigid docking simulations, we generated a set of low-energy conformer for each drug with Conformer Import tool, with an imposed conformational cut-off energy of 3 kcal/mol from minimum energy structure of each compound, calculated with the AMBER10-EHT force field. The resulted in-house molecular data base (mdb) contain multiple conformers for each molecule and were used for rigid docking simulation.
Protein selection for ligand docking The X-ray crystal structure of SARS-CoV-2 RBD S1 in a complex with the ACE2 (PDB ID: 6VW1, resolution of 2.68 Å) was selected as a protein target for docking simulations. Importantly, this engineered structure is the first to presents all the functionally important epitopes in the SARS-CoV-2 receptor binding motif 20 . Potential binding sites in ACE2 near the interface region between the SARS-CoV-2 RBD S1 and ACE2 proteins, were identified with Site Finder tool. All crystallographic water and ligands molecules were removed from the system (chains B, E, F). Hydrogen atoms (Protonate 3D tool) and partial charges (Potential Setup tool) were added to ACE2 assuming pH equal to 7.0 and using the AMBER10-EHT force field, respectively. Before docking, the ACE2 protein structure was subjected to energy minimization using the same forcefield, in order to optimize atomic contacts. Docking simulations between the optimized ACE2 structure and each of the conformers contained in the in-house database, was carried out under the rigid-docking protocol. The docking parameters were set to take each ligand conformation as unique molecule, using the Alpha Triangle algorithm as placement method (at least 100 different orientations or poses on potential binding site) and further evaluation keeping the thirty best poses accordingly the London scoring function for binding affinity with a second refinement as a Rigid Receptor using Affinity dG algorithm keeping the ten best poses. The results were analyzed by docking score, frequency of the chemical compound as a stable conformation and the types of interactions at the binding site residues.

Structural analysis of SARS-CoV-2 -ACE2 interaction
The structural analysis for the SARS-CoV-2 RBD S1 of the spike protein in a complex with the ACE2 (PDB ID: 6VW1; Figure 1A) revealing a potential site for ligand binding inside ACE2 structure ( Table 1). The identified receptor site ( Figure 1B) is proximal to the binding site of RBD S1 with a size of 86, therefore it can be used for simulating rigid molecular docking since receptor atoms are in an exposed region of the structure, which could be in favor of drug binding.
Virtual screening and molecular docking An average of 78 conformations were generated for each ligand by Conformation Import MOE, generating 100,450 ligand conformations of the FDA approved and prescript drugs by the Mexican Public Health System.
The docking results were sorted and analyzed based on their S score, binding frequency which the drug binds to the receptor Figure 1. SARS-CoV-2 RBD S1 interaction with human ACE2 receptor. A) Crystallographic structure (PDB ID: 6VW1) of RBD S1 (red) and ACE2 receptor (blue); B) Molecular surface of the selected binding site in ACE2.
site and type of interactions, preferably, hydrogen bond, of the ligand with the selected site. We selected 38 drugs ( Table 2) that presents the best docking score between −10.04 and −4.04.
Subsequently, we shortlisted nine drugs based on their risk of teratogenicity, route of administration, interaction with other drugs, side effects and by their background as pharmacological therapy for the treatment of respiratory diseases 27,28.   Figure 2I) and a docking score of −9.02. Some pharmacokinetics characteristics 27,28 of the shortlisted potential inhibitors of the RBD S1 -ACE2 interaction are summarized in Table 3.

Discussion
It has been established that S protein of SARS-CoV-2 virus plays a major role during viral infection. The S protein mediates receptor recognition, cell attachment and fusion of viral membrane with host cell membrane 29-34 . The S protein binds to ACE2 receptor through the RBD S1 , mediating viral attachment to host cell 35 . Expression of ACE2 is ubiquitous in lung, intestine, heart and kidney, also alveolar epithelial type II cells had higher expression levels 23 . SARS-CoV-2, as one RNA viruses, has shown a high mutation rate as a result of lack of proofreading mechanisms, which leads to gain the ability to rapidly adapt to changes in their environment, which in turn leads to a great challenge for treating and preventing infections 36 . In this sense, the RBD region is a critical therapeutic target (vaccines and drugs) due to its indispensable function; however, it is suggested that mutations in this region may render pharmacological or immunological therapies ineffective 37,38 , therefore, it is needed to search and design alternative treatments.
In order to block this event, we propose an in silico approach to identify potential inhibitors of the SARS-CoV-2 -ACE2 interaction aiming at the ACE2 receptor, blocking the virus accessibility to the membrane-bound ACE2. In this sense, Jia and collaborators 25 present an extensive review for this underexplored approach to treat COVID-19, pointing that it is imperative to determine, by clinicians, the stage of the disease and comorbidities that could prove consequential for an ACE2-targeting regimen. Here, we screened a drug library consisting of 1,283 drugs, FDA approved and prescribed by the Mexican Public Health System, for potential SARS-CoV-2−ACE2 inhibitors, using a rigid receptor docking approach. Utilization of an FDA-approved drug library is an effective and ideal tool for drug repurposing in antiviral research 39  potentially inhibitor drugs of SARS-CoV-2−ACE2 interaction and these are listed on Table 2. Several of those drugs were previously reported to be used for the treatment of respiratory diseases.
Within this list of potential inhibitors of the SARS-CoV-2−ACE2 interaction, is fexofenadine, a third generation antihistamine whose therapeutic indication is the treatment of symptoms of stationary allergies through the selective blockade of H1 receptors 44 . It possesses direct effect on combating the cytokine storm caused by SARS-CoV-2 through inhibition of histamine and interleukin-6 (IL-6) release. In silico evidence 45,46 suggest that it may interact with the SARS-CoV-2 main protease enzyme M Pro , a key enzyme in viral replication 47 , acting as a potential inhibitor.
Cefixime is a third-generation antibiotic derived from cephalosporin whose use is indicated for the treatment of infections in the upper and lower respiratory tract, otorhinolaryngological 48 and urinary tract 49 , inhibiting the synthesis of the bacterial wall by binding to specific binding proteins for penicillin and is currently used as a secondary therapy to prevent opportunistic infections during the development of COVID-19 50 .
Pitavastatin is a statin indicated for lowering blood cholesterol levels by inhibiting HMG-CoA reductase, preventing cholesterol synthesis 51 , and it has also been observed that statin treatments can interfere with viral infectivity through inhibition of glycoprotein processing 52 also, they modulates the inflammatory process at cellular level 53 , which is a remarkable characteristic of the SARS-CoV-2 infection. Additionally, in silico findings suggest that could be an efficient inhibitor of SARS-CoV-2 M Pro54 and SARS-CoV-2 RNA-dependent RNA polymerase (RdRp) 55 thru active site binding.
Lopinavir is a protease inhibitor indicated as first barrier therapy, in conjunction with Ritonavir, to treat infection caused by the HIV virus by inhibiting the HIV-1 protease 56 in addition, studies in cell cultures have shown its effectiveness as an inhibitor of the replication of the MERS-CoV virus 57 and SARS-CoV-1 58 , while in severe cases of SARS-CoV-2 infection, the results of clinical trials indicate that it is not useful 59 .
Formoterol and aformoterol, anenantiomer of formoterol, are long-lasting selective β agonists indicated for the treatment of chronic obstructive pulmonary disease (COPD) and bronchospasms 60 , in the same way there is evidence of the Vitamin use of these drugs as a partial inhibitor of viral replication in primary epithelial cells cultures 61 and in silico data suggest their binding to the papain-like protease PL pro , a coronavirus enzyme essential for viral spread 62 .
Ipratropium is a bronchodilator anticholinergic indicated for the treatment of asthma, shortness of breath, cough and tightness in the chest in patients with COPD 63,64 . Inhalation therapy with ipratropium is currently in use to dilate bronchioles in COVID-19 patients to increase oxygen saturation levels (from <80% to 94%) 65 .
Pargeverine is an antispasmodic opioid alkaloid whose therapeutic indication is aimed at the treatment of painful spasms 66 , also, acts as anticholinergic and has a moderate and non-selective blockade of muscarinic cholinergic fibers 67 . Since cholinergic activity contribute to airway narrowing, this might be a potential agent to open airway obstruction.
Cholecalciferol, is a form of vitamin D (vitamin D3) that can be synthesized naturally in the skin and acts as a hormonal precursor, being converted into calcitriol, and therapeutically is used as a vitamin supplement to treat deficiencies of this vitamin 68 . In addition, it has been observed that vitamin D supplementation is favorable to reduce viral infections such as influenza 69,70 or more aggressive cases such as HIV 71 and it has recently been suggested that it also presents favorable effects before and during the infection caused by SARS-CoV-2 72 .
Likewise, it is important to take into account that ACE2 plays an important biological role since regulates cardiovascular functions and innate immune system 73 and, therefore cautions must be taken. Another point to consider is the delivery method of the drug, since the primary target must be smooth muscle, like the one surrounding the bronchioles, and lung epithelial cells in the airway and airspace compartments, hence, inhalable delivery would be the acceptable choice to deliver the drug in a selectively and localized manner.
Given these characteristics, the results obtained through our in silico approach, we consider that the aforementioned drugs are outlined as possible inhibitors of the RBD S1 -ACE2 interaction. These drugs are well tolerated, commonly used and affordable, hence, most of the drugs on this list can be tested in vitro, and even in vivo and, consequently, in clinical trials for the development of adjuvant therapies to treat COVID-19.

Conclusion
In the absence of approved therapies for treatment or prevention, drug repurposing has provided fast and valuable insight into the treatment of COVID-19. Targeting ACE2 receptor as a COVID-19 therapy it is a conceivable approach since it is essential for the viral internalization. However, this approach requires an integrative evaluation of the pros and cons by a clinical context since ACE2 is a multifunctional protein. Several drugs are currently investigated by clinical trials or are already in use to treat COVID-19 patients, like lopinavir or ipratropium. In this in silico study using structure-bases virtual screening, we identified potential inhibitors of SARS-CoV-2-ACE2 by their interaction with ACE2 receptor. Based on desired characteristics like pharmacokinetics, route of administration or by their background as pharmacological therapy, we propose a shortlist of drugs suitable for testing their potential RBD S1 -ACE2 inhibitory activity: fexofenadine, cefixime, pitavastatine, lopinavir, arformoterol, formoterol, ipratropium, pargeverine and cholecalciferol. Our identification of potential inhibitors of the SARS-CoV-2-ACE2 interaction among commonly use drugs highlights their potential use for treating COVID-19. Further in vitro, in vivo or clinical trial are needed to validate their potential use as inhibitors of SARS-CoV-2-ACE2 interaction.

Data availability
Source data Protein Data Bank: Crystal structure of SARS-CoV-2 receptor binding domain (RBD S1 ) of the spike protein in a complex with the ACE2 receptor. https://identifiers.org/pdb:6vw1.

Open Peer Review
Results: Table 3, the details regarding generation of data appears missing. I feel the detailed legends can be given.

○
The current gold standard for molecular activity prediction is MD simulation, this can be tried by authors.

Discussion:
In the methodology the number of drugs selected were 1283, but here it changed to 1300. This looks inconsistent.

○
In the methodology, authors have mentioned that they have generated some conformers. Which conformation of drugs were active? ○ Overall, I feel the discussion part appears very weak and lacks substantial evidence for how these drugs could be used in COVID-19.

Conclusions:
The conclusion can be made simple by reducing the number of sentences. Most of the journals will not accept citations (references) in conclusions. Hence, the authors can give a thought to modify the conclusion by keeping their views only on objectives/ aims.

Guillermo A Silva-Martinez, Tecnologico Nacional de Mexico en Celaya, Celaya, Mexico
We thank the reviewer for the detailed revision of our paper and for suggesting improvements.

Abstract:
As the pandemic is dynamically changing the numbers, giving a country specific might not hold good. I suggest to add the data related to the world-wide data can be put. Response: We thank the reviewer for the comments and for suggesting improvements. Changes on the subsections mentioned above were made. Resolution of 6VW1, the rationale of using that structure and accession date to databases were added to the manuscript. In the "Protein selection for ligand docking", it is now clearly specified that rigid molecular docking is carrying out. In this type of docking, protein orientation is fixed and only ligands are allowed to vary their orientation, therefore each conformer allow to simulate this variation. We assumed that no further explanation was needed.

Fernando Yepes-Calderon
GYM Group SA, Cali, Colombia Authors use molecular docking tools to test existing medicine in their efficacy to impede the S protein-ACE2 binding and thus, disabling the entry point of Cov2 to human cells. The authors use a solid justification to support this project. Since medicine creation is an expensive and long-lasting task, verifying if some active molecules in already FDA-approved medicine might block the binding between the S-protein and the ACE2 is an excellent alternate path.
The document is generally well organized and pleasant to read. I suggest adding some corrections to the abstract as follows.

Background
A new coronavirus outbreak, firstly reported in Wuhan, China, in December 2019, causes COVID-19, which symptoms are similar to SARS-atypical pneumonia. Worldwide, around 116 million cases and 2.57 million deaths are reported, with new cases increasing mortality every day. To date, there is no specific commercial treatment to control the infection.
Repurpose drugs targeting the angiotensin-converting enzyme 2 (ACE2) receptor represent an alternative strategy to block the binding of CoV2's protein S and forestall virus adhesion, internalization, and replication in the host cell.

Methods
We performed rigid molecular docking using the receptor-binding domain at the S1 subunit of the S protein (RBDS1)-ACE2 (PDB ID:6VW1) interaction site and 1,283 FDA-approved drugs. The docking score, frequency of the drug in the receptor site, and interactions at the binding site residues were used as analyzing criteria.

Results
This research yielded 40 drugs labeled as potential inhibitors of RBD S1-ACE2 interaction. Among the inhibitors, compounds such as ipratropium, formoterol, and fexofenadine can be found. Specialists employ these drugs to treat chronic obstructive pulmonary diseases, asthma, and almost any respiratory infection. Conclusions: Our results will serve as the basis for in vitro and in vivo studies to evaluate the potential use of those drugs to generate affordable and convenient therapies to treat COVID-19.
The proposed modifications focus on keeping the communications global. The CoV2 numbers in Mexico are absent in this suggested abstract since the authors provide worldwide numbers. I also shortened some sentences and fixed some wrongly used singulars. Correct the hyphenation in the whole document. If the authors are using a latex template, use the command \hyphenation{co-rrect, se-pa-ra-tion, sche-me}.

1.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Response: We do not have any issue during the structure preparation; therefore, reproducibility can be achieved without any problem.
The list of potential inhibitors is highly appreciated. The authors presented the interactions of the selected drugs nicely.
Response: We appreciate that the reviewer finds our data well-presented.
I suggest the authors reinforcing their discussion with arguments derived from the mutation capacity of the retrovirus that might render current vaccination not entirely compelling. Therefore, the presented alternatives could set as plausible options for treatment.
Response: We thank the reviewer for suggesting improvements to our discussion section. Reinforced of the discussion section were made in the manuscript taking reviewers suggestion.