key: cord-0769663-46ot0fry authors: Fayed, Marwa A.A.; Farrag El-Behairy, Mohammed; Abdallah, Inas A.; Mohamed Abdel-Bar, Hend; Elimam, Hanan; Mostafa, Ahmed; Moatasim, Yassmin; A. M Abouzid, Khaled; A. M. M. Elshaier, Yaseen title: Structure- and Ligand-Based In silico Studies toward the Repurposing of Marine Bioactive Compounds to Target SARS-CoV-2 date: 2021-02-25 journal: Arab J Chem DOI: 10.1016/j.arabjc.2021.103092 sha: c604cb8a3836bccb91ebb2c2836ed37a0adb9150 doc_id: 769663 cord_uid: 46ot0fry This work was a structured virtual screening for marine bioactive compounds with reported antiviral activities which were subjected to structure-based studies against SARS-CoV-2 co-crystallized proteins. The molecular docking of marine bioactive compounds against the main protease (Mpro, PDB ID: 6lu7 and 6y2f), the spike glycoprotein (PDB ID: 6vsb), and the RNA polymerase (PDB ID: 6m71) of SARS-CoV-2 was performed. Ligand-based approach with the inclusion of rapid overlay chemical structures (ROCS) was also addressed in order to examine the probability of these marine compounds sharing relevance and druggability with the reported drugs. Among the examined marine library, the highest scores in different virtual screening aspects were displayed by compounds with flavonoids core, acyl indole, and pyrrole carboxamide alkaloids. Moreover, complete overlay with the co-crystallized ligands of Mpro was revealed by sceptrin and debromo-sceptrin. Thalassoilin (A-B) which found in the Red Sea exhibited the highest binding and similarity outcomes among all target proteins. These data highlight the importance of marine natural metabolites in regard of further studies for discovering new drugs to combat the COVID-19 pandemic. The coronavirus disease 2019 (COVID-19) is known to affect the respiratory system, causing severe acute respiratory syndrome; it was declared by the WHO in March 2020 as a global pandemic. The severity and spreading of the disease are attributed to different types of human coronaviruses [1] . A wide range of symptoms can be caused by these types, similar to those of the infection by influenza such as in case of the Middle East Respiratory Syndrome (MERS) and Severe Acute Respiratory Syndrome (SARS) [2] . The SARS-CoV-2 was identified to be a single-stranded positive-sense RNA virus with a size of ~30 kb that belongs to the genus Beta-coronaviruses with a crown due to the presence of spike glycoproteins on the viral envelope [3] . It can infect the throat, the lower respiratory tract and cause pneumonia in humans, although it seems that the symptoms are milder than those of the other SARS viruses [3] . To date, more than 59.81 million cases in total have been confirmed worldwide along with 1.41 million death, according to the World Meter [4] . Each SARS-CoV-2 protein has its relevant functional role, active site and subsequently, its own druggable target. These proteins as non-structural proteins (nsp) e.g. nsp 3, the papain-like protease (PLpro, nsp3 domain), main protease (M pro , nsp5), the RNA polymerase complex (nsp12-nsp7-nsp8), and other targets [5] . The SARS-CoV-2 main protease (M pro ) is a cysteine protease engaged in many cleavage steps in the precursor polyproteins, and so it has a critical role in the viral life cycle. The active form of M pro is a homodimer containing two protomers composed of three domains each: domain I, domain II, and domain III, domains II and III are connected by a long loop region [6] [7] [8] . The binding of SARS-CoV-2 to angiotensin-converting enzyme 2 (ACE-2) receptors in type II pneumocytes in the lungs triggers the cascade of an inflammatory response in the lower respiratory tract. SARS-CoV-2 is very similar to SARS-CoV-1 regarding the biochemical interactions and pathogenesis [9] . The spike glycoprotein is composed of a transmembrane trimetric glycoprotein protruding from the viral surface; it determines the diversity of coronaviruses and host tropism [10] . Moreover, SARS-CoV-2 spike glycoproteins also contain a variable receptor-binding domain (RBD) in the S1 functional subunit, which binds to the ACE-2 receptor and subsequently to the type 2 transmembrane protease serine 2 (TMPRSS2). The interaction complex is cleaved by TMPRSS2, leading to ACE-2 and the activation of the spike glycoprotein (Figure 1) , facilitating viral entry into the target cell [11] [12] . Inside the host cell, viral RNA hijacks the host cell machinery, inducing it to produce RNA and proteins that contribute to the assembly of new viral particles. Natural bioactive products are considered to be one of the main and most diverse sources of a variety of therapeutic agents, providing more than 50 % of the available drugs which are in clinical use around the world [13] . Natural compounds with marine origin are of great importance as a source of novel and potentially life-saving bioactive secondary metabolites, since the marine environment represents 95% of the biosphere [14] . There are more than 40,000 different phytoplankton species, including 680 species of marine algae belonging to Rhodophyta, Phaeophyta, and Chlorophyta, commonly known as red, brown, and green seaweed, respectively, in addition to 71 mangrove plant species that were previously documented in the global marine biotope. They provide essential metabolites, including fatty acids, ionic trace minerals, vitamins, enzymes, bioflavonoids, amino acids, as well as other nutrients [13] , which possess several beneficial biological effects, such as anti-cancer, antibacterial, and antiviral activity even against resistant strains [13] . Various aquatic organisms were previously reported to provide a variety of unique lead compounds with antiviral activity [15] . The drug repurposing approach provides a quick tool to overcome the coronavirus pandemic. Most drugs subjected to repurposing with the aim of overcoming the COVID-19 outbreak are commercially available and their dosage and toxicity in humans are well known with a history of decades of clinical use in some cases [16] . Many valuable efforts in drug repurposing approaches have been recognized against SARS-CoV-2 utilizing cheminformatics and bioinformatics tool [17] [18] [19] [20] [21] [22] . Recently, Claudio N. C. et. al. addressed three drug discovery strategies in the frame structural and druggability studies of the SARS-CoV-2 proteome. [5] These includes interfering proteins responsible for genome stability as non-structural proteins (nsp) 13 , blocking proofreading proteins as nsp 14, and targeting nsp15 responsible for viral replication and also the accumulation of cytokine-producing macrophages [23] . The repositioning of antiviral drugs that are already used against SARS-CoV-2 proteins, including remdesivir (an RNA polymerase inhibitor), umifenovir (a membrane fusion inhibitor targeting viral entry), and lopinavir/ritonavir (a drug combination targeting viral proteases) was reported and currently being considered in different combinations in a Phase IV clinical trial, as shown in Figure 1 [24] . Therefore, the drug repurposing approach could be a promising tool for the discovery of an effective therapy against this novel coronavirus. This oriented us to search for a COVID-19 remedy among the known marine metabolites through in-silico screening for their activity against the COVID-19 protease, polymerase, as well as spike glycoproteins and comparing their activity to already used anti-coronavirus drugs [25] [26] [27] [28] . This study included a screening for marine bioactive compounds with reported antiviral activities ( discussed for the highest-scoring compounds. Our attention was afterwards directed to the ligand-based drug repurposing tactic, which included shape alignment, rapid overlay chemical structures (ROCS), and structure-property relationship to examine the druggability of the prioritized marine compounds. These bioactive compounds were made publicly accessible for facilitating the conduction of further studies and the optimization by the scientific community [29] . The study was performed using the Openeye scientific software [30] (academic license 2020). The compounds were ranked and sorted according to their consensus score values. Docking results of marine compounds against M pro (PDB ID: 6lu7 [31] and 6y2f [32] ), spike glycoprotein (PDB ID: 6vsb [33]), and RNA polymerase (PDB ID: 6m71 [34] ) were subsequently filtered based on their obtained consensus score values. The lowest consensus score value indicated the strongest binding to the receptor. The compounds shown in Figure 3 displayed an interaction to the targeted receptors that was stronger than that of standard ligands (N3, α-ketoamide for M pro , S ligand 1 for spike protein, and remdesivir for RNA polymerase, respectively, shown in Figure 1 ). These compounds were subjected to the acquisition of further details, as shown in Tables 2-5 . The top hits compounds, as shown in Figure 3 , were classified based on their chemical structures into 1) hydrocarbon-based (enyne, conjugated ene, or alkanes) compounds, Sceptrin was found to form HBs with key amino acids, including Glu166 and Gln189 as well as two HBs with Asn142. Moreover, sceptrin was observed to overlay with cocrystallized ligands through a strong binding interaction, as shown in Figure 4c . Thalassoilin A and thalassoilin B exhibited the same binding mode and pose through the formation of HB with Gly143, although both displayed different modes of interaction in case of thalassoilin C, as shown in Figure 4d . Snapshot of these three flavonoids indicated high similarity in case of ligand pose. Interestingly, moderate in vitro antiviral HCV protease activity with an IC50 value of 16 μM was displayed by these flavonoids [35] . This result suggests that the total extract of this metabolite has the potential to be further investigated. Agelifenin exhibited weak bonding and no overlay with standard ligands in comparison to agelifenin 1, which was assumed due to extra bromination Some compounds still exhibited a stronger binding mode with PDB ID 6y2f than in case of 6lu7. Debromo sceptrin showed unique and strong interaction with the receptor through the formation of a binding mode and poses like the co-crystallized ligand, as shown in Figure 5a . Debromo sceptrin (pink) and sceptrin (blue) showed no co-overlay, as shown in Figure 5b . Moreover, a complete overlay was identified for thalassoilin B and thalassoilin A, while thalassoilin C exhibited different binding modes and pose, as shown in Figure 5c . In contrast, an overlay was identified for all acyl indole alkaloids (topsentin derivatives), as shown in Figure 5d . Although more than 50 compounds showed stronger interaction than the standard ligand, most of them exhibited different binding modes or docked outside the created receptor. Compounds containing long-chain hydrocarbon parts behaved stickier in association with standard ligands. Figure 6a Remdesivir is an anti-polymerase drug prescribed for the alleviation of the COVID-19 disease. Its docking mode showed its ability to form two HBs with Gln166 through its oxygen of the furanose ring (strong) and via the oxygen of the carbonyl functionality. Additionally, it participated in another HB with Asn79 through the phosphor atom, as shown in Figure 7a . Rapid Overlay Chemical Structure (ROCS) is a chemoinformatic screening technique utilized to perceive a similarity between chemical entities based on their three-dimensional shape [36] . The 3D shape structure displays good neighborhood behavior in which high a similarity in the shape reflects the high similarity in their biology in cases where a high similarity in biology is not reflected in a similarity in the structure. [37, 38] Shape similarity using the ROCS tool has different applications, including virtual screening, lead-hopping, molecular alignment, pose generation, as well as structural predictions [36] . The approach of this study was to identify molecules that can adopt shapes which are extraordinarily similar to reported drug candidates (the query), as shown in Figures 8-10 . Based on this concept, matches are created according to only the volume of overlapping of optimally aligned molecules, which are virtually independent from the atom types and the bonding patterns of the query. The ROCS study requires two files which must be in 3D format in case of most stable conformers generated by the Omega application in the Openeye software: a) a database file, which includes the collection of all compounds for this study; b) a query file, which includes the standard (reference) molecules or the lead compound used for the inhibition of SARS-CoV-2 replications. The outputs from the ROCS analysis were 1) a shape counter, shape atoms, and color atom labels for the database set as well as the query compounds, as shown in Table 6 and Figures 8-10 . The shape was represented with a series of dotted lines around the molecule, while the color feature was shown as a filled colored circle representing the different kinds of chemical features; 2) the overlay, which was considered to be the alignment between the database as well as the molecules and the query, as visualized by the vROCS and VIDA applications; and 3) the set of scores were expressed in Tanimoto scores. The most important score is the Tanimoto Combo (TC) that includes both shape fit and color. This score had a value between 0 and 2 and was the score used for ranking the hit list. Based on previous docking results, the compounds were found to express multi target actions as most of them exhibited high scores with the different targets (main protease, RNA polymerase, or spike glycoprotein). In this regard, these marine candidates were selected to be the subjects of further cheminformatics studies. Compounds formed flavonoids nucleolus (thalassoilin B, thalassoilin C and subsequently thalassoilin A) showed higher TC scores (0.6800-0.4980) than PF-00835231 and α-ketomide (0.4960, 0.4950 respectively) in comparison with remdesivir, as shown in Table 6 . Table 6 . Concerning the comparative analysis of these compounds in association with drugs inhibiting the viral entry, such as umifenovir, the following marine compounds were analyzed: acylindole marine alkaloids (topsentin, bromotopsentin, 4,5-dihydro-6deoxybromotopsentin respectively), lornemides A, dragmacidin F, atomeric acid, thalassoilin C, macrolactin A, PF-00835231, and subsequently thalassoilin A, exhibiting a high TC score (0.946-0.685, respectively), as shown in Table 6 . All marine compounds exhibited TC score more than remdesivir in case of umifenovir as a query. As a result, flavonoid compounds and acyl indoles exhibited high 3D shape similarity in association with most of the known reported drug candidates. Shape color and volume shape of all compounds were illustrated in Figure 6 including supplementary data for their figures. Physicochemical properties determine the drug-likeness score in case of orally administered drugs. Drug candidates with high drug-likeness scores were previously reported to exhibit higher absorption and bioavailability in lower doses and show fewer drug-drug interaction warnings [39] . Absorption, distribution, metabolism, excretion, and toxicity (ADMET) calculations contribute to the determination of the failure of approximately 60% of all drugs in the different clinical trial phases. In this regard, ADMET is determined at the beginning of the drug discovery phases to eliminate molecules coming with poor ADMET properties from an earlier drug discovery pipeline to save research costs. As demonstrated in Table 7 and among the selected compounds, atomeric acid displayed the highest drug-likeness score (0.56), followed by flavonoids marine compounds (thalassoilin B, thalassoilin C, thalassoilin A) with a drug-likeness score in the range of 0.39-0.36. Atomeric acid had the highest lipophilicity value with a logP value of 5.86 and, therefore, low hydrophilicity, causing poor absorption or permeation. This result suggested us to semi-synthesize this compound in an ester form to optimize the lipophilicity to be a potential orally active compound. 4. Experimental Marine bioactive compounds were retrieved from literatures, and reported drugs were identified from known data base collection. The molecular docking studies were operated using the OpenEye Basic method to represent shape and color features in ROCS is using ROCS application Open Eye scientific software. Gaussian overlap to the query and the best scoring parameters is Tanimoto Combo scores (shape + color), the highest score is the best matched with query compound. Lipinski´s rule (Rule of five) and molecular property prediction was calculated at the following free access website https://www.molsoft.com/servers.html. Regarding the PreADMET estimation, it was determined through utilizing the free access of the website https://preadmet.bmdrc.kr/. The paper was based on work supported by the University of Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and coronavirus disease-2019 (COVID-19): The epidemic and the challenges Overview of lethal human coronaviruses Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China Lancet Coronavirus Update: Cases And Deaths From COVID-19 Virus Pandemic-Worldometer Functional and druggability analysis of the SARS-CoV-2 proteome Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors Structural basis for the inhibition of SARS-CoV-2 main protease by antineoplastic drug carmofur SARS coronavirus unique domain: three-domain molecular architecture in solution and RNA binding Molecular Immune Pathogenesis and Diagnosis of COVID-19 Angiotensin-converting enzyme 2 is a functional receptor for the SARS coronavirus Evidence that TMPRSS2 Activates the Severe Acute Respiratory Syndrome Coronavirus Spike Protein for Membrane Fusion and Reduces Viral Control by the Humoral Immune Response TMPRSS2 and ADAM17 Cleave ACE2 Differentially and Only Proteolysis by TMPRSS2 Augments Entry Driven by the Severe Acute Respiratory Syndrome Coronavirus Spike Protein Anticancer Drugs from Marine Flora: An Overview New marine derived anticancer therapeutics-a journey from the sea to clinical trials Marine Products as a Source of Antiviral Drug Leads Drug repurposing against COVID-19: focus on anticancer agents Repurposing approved drugs as potential inhibitors of 3CL-protease of SARS-CoV-2: Virtual screening 60 and structure based drug design In vitro screening of a FDA approved chemical library reveals potential inhibitors of SARS-CoV-2 replication Drug discovery strategies for SARS-CoV In silico Drug Repurposing for COVID-19: Targeting SARS-CoV-2 Proteins through Docking and Consensus Ranking Ligand and structurebased virtual screening applied to the SARS-CoV-2 main protease: an in silico repurposing study A Computational Approach to Identify Potential Novel Inhibitors against the Coronavirus SARS-CoV-2 Functional and druggability analysis of the SARS-CoV-2 proteome Drug repurposing strategies for COVID-19 Natural-like products as potential SARS-CoV-2 Mpro inhibitors: in-silico drug discovery In silico drug discovery of major metabolites from spices as SARS-CoV-2 main protease inhibitors In-silico drug repurposing and molecular dynamics puzzled out potential SARS-CoV-2 main protease inhibitors In Silico Evaluation of Prospective Anti-COVID-19 Drug Candidates as Potential SARS-CoV-2 Main Protease Inhibitors Anna Maria Almerico and Marco Tutone Exploring the SARS-CoV-2 Proteome in the Search of Potential Inhibitors via Structure-Based Pharmacophore Modeling/Docking Approach. Computation An in silico perception for newly isolated flavonoids from peach fruit as privileged avenue for a countermeasure outbreak of COVID-19 Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors Crystal structure (monoclinic form) of the complex resulting from the reaction between SARS-CoV-2 (2019-nCoV) main protease and tert-butyl Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation SARS-Cov-2 RNA-dependent RNA polymerase in complex with cofactors Thalassiolin D: a new flavone Oglucoside Sulphate from the seagrass Thalassia hemprichii Comparison of Shape-Matching and Docking as Virtual Screening Tools Identification and Molecular Modeling of New Quinolin-2-One Thiosemicarbazide Scaffold with Molecular Shape Comparison of Angiotensin II Receptor Antagonists How drug-like are 'ugly' drugs: do druglikeness metrics predict ADME behaviour in humans? Drug Discovery Today