key: cord-0856815-nn98mtdr
authors: Özdemir, Mücahit; Köksoy, Baybars; Ceyhan, Deniz; Sayın, Koray; Erçağ, Erol; Bulut, Mustafa; Yalçın, Bahattin
title: Design and in silico study of the novel coumarin derivatives against SARS-CoV-2 main enzymes
date: 2020-12-27
journal: Journal of biomolecular structure & dynamics
DOI: 10.1080/07391102.2020.1863263
sha: 51bbfd87394f5650200ff332c0b811c579052d6f
doc_id: 856815
cord_uid: nn98mtdr

The novel coronavirus (SARS-CoV-2) causes severe acute respiratory syndrome and can be fatal. In particular, antiviral drugs that are currently available to treat infection in the respiratory tract have been experienced, but there is a need for new antiviral drugs that are targeted and inhibit coronavirus. The antiviral properties of organic compounds found in nature, especially coumarins, are known and widely studied. Coumarins, which are also metabolites in many medicinal drugs, should be investigated as inhibitors against coronavirus due to their pharmacophore properties (low toxicity and high pharmacokinetic properties). The easy addition of substituents to the chemical structures of coumarins makes these structures unique for the drug design. This study focuses on factors that increase the molecular binding and antiviral properties of coumarins. Molecular docking studies have been carried out to five different proteins (Spike S1-subunit, NSP5, NSP12, NSP15, and NSP16) of the SARS-CoV-2 and two proteins (ACE2 and VKORC1) of human. The best binding scores for 17 coumarins were determined for NSP12 (NonStructural Protein-12). The highest score (–10.01 kcal/mol) in the coumarin group is 2-morpholinoethan-1-amine substituted coumarin. Molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) analyses of selected ligand-protein complexes were performed. The binding energies in each 5 ns were calculated and it was found that the interaction between ligand and target protein were stable. Communicated by Ramaswamy H. Sarma

When the coronavirus first appeared in Wuhan, Hubei province, China in December 2019, its source was unknown and traces of pneumonia were visible in patients carrying the virus Lu et al., 2020; Zu et al., 2020) . This virus, which was found to belong to the coronavirus family, was first named 2019-nCoV (2019-novel CoronaVirus), and then the World Health Organization changed this name to SARS-CoV-2 (Severe Acute Respiratory Syndrome-Corona Virus-2) due to its close similarity to SARS-CoV (Gorbalenya et al., 2020; Guo et al., 2020) . The disease caused by the coronavirus (COVID-19) is highly contagious and spreads rapidly from person to person. Since the virus spread rapidly all over the world, it has been declared a global pandemic by the WHO (Ghebreyesus, 2020) .

COVID-19 affects people in various ways and no generalization can be made. Many people infected with the virus have mild to moderate symptoms and recover without hospitalization (Al-Tawfiq, 2020) . In severe cases, common symptoms include respiratory distress, fever, weakness, dry cough, shortness of breath, headache, and pain of the bones (Singhal, 2020) . The group in which the symptoms of the disease are most effective are those with diabetes, heart, cancer, hypertension, and chronic respiratory diseases (Casey et al., 2020; Fang et al., 2020; Ganatra et al., 2020; Kreutz et al., 2020; Zheng et al., 2020) . For patients included in the risk group, coronavirus is very dangerous and in advanced cases, the infection can cause pneumonia, acute respiratory failure, kidney failure, pulmonary embolism, and even death (Jordan et al., 2020; . Coronavirus effects can be prevented by inhibiting the binding of the virus to human cell receptors, the spread of their genetic material, and the replication. Several antiviral drug combinations have been tried for the treatment of COVID-19, but it has been reported that these drugs are not excellent effective. Many scientists around the world are researching to discover new drugs for inhibition of coronavirus. Phytochemicals, as well as normal chemical drugs against SARS-CoV-2, are tested and satisfactory results are obtained (Chikhale et al., 2020; da Silva Antonio et al., 2020; Kumar et al., 2020; Pandey et al., 2020; Ul Qamar et al., 2020) .

Coumarins are natural organic phytochemicals found in some plants as well as easily synthesized (Murray et al., 1982) . As a result of extensive research, antiviral, anticoagulant, antibacterial, and anticancer properties have been proven (de Souza et al., 2005; Hassan et al., 2016; Koch-Weser & Sellers, 1971; Sashidhara et al., 2010; Thakur et al., 2015) . For example, warfarin (coumadin) is used as a vitamin K-antagonist for cardiac patients (warfarin achieves its anticoagulant effect by competitive inhibition with vitamin K epoxide reductase) and it has been proven by experiments that it does not have a high toxic effect in clinical studies. The excellent pharmacophore properties of coumarins make this group prominent in drug design.

Coronaviruses are enveloped, single-stranded, positive RNA viruses that have been almost completely analyzed, all their receptors published in the protein data bank (PDB). SARS-CoV-2 contains two characteristic protein groups and accessory factors. The first is structural proteins and consist of the Spike (S), Membrane (M), Nucleocapsid (N), and Envelope (E) protein group. The second is non-structural proteins, and consist of the NSP1, NSP2, NSP3 (Papain-like), NSP4, NSP5 (Main protease, 3CLpro), NSP6, NSP7, NSP8, NSP9 (RNA replicase), NSP10, NSP11, NSP12 (RNA-dependent RNA polymerase-RdRp), NSP13 (Helicase), NSP14 (Exonuclease), NSP15 (NendoU), NSP16 (2'O-Methyltransferase). The main protease cleaves at 11 sites within the polyprotein to release NSP4-NSP16. It is also responsible for NSP maturation . The spike protein is split into glycosylated subunits (S1 and S2). S1 binds to host receptor ACE2 while S2 mediates viral and host membrane fusion (Hoffmann et al., 2020) . NSP12 is an RNA-dependent RNA polymerase (RdRp) that performs both replication and transcription of the viral genome (Q. . NSP15 is an endoribonuclease that promotes cleavage of RNA at the 3 0 -ends of uridylates. Loss of NSP15 affects both viral replication and pathogenesis. It is also necessary to avoid host cell dsRNA sensors (Kim et al., 2020) . NSP16 interacts with the NSP10 and is activated. 2 0 -O-methyltransferase activity is required for viral mRNA capping. It can also work against host cell antiviral sensors (Viswanathan et al., 2020) . Human Angiotensin converting enzyme 2 (hACE2) is an enzyme bound to the cell membranes of cells in the lungs, arteries, heart, kidney and intestines (Donoghue et al., 2000; Hamming et al., 2004) . ACE2 catalyzes the conversion of angiotensin II to angiotensin by hydrolysis and lowers blood pressure (Keidar et al., 2007) . ACE2 interacts with spike proteins around the coronavirus, is entry point and allows the virus to enter the cell . Vitamin K epOxide Reductase protein complex is responsible for its reduction to the active form of vitamin K 2,3-epoxide, which is important for coagulation. As a result of the inhibition of VKORC1, vitamin K-hydroxylation cannot become active and coagulation does not occur (Czogalla et al., 2015; Oldenburg et al., 2006) .

The reason for the selection of secondary amine-containing coumarins in this study is that coumarins are natural, easy to synthesize and derivatize, have high biological activity and low toxicity, and such an important group has been scarcely investigated for SARS-CoV-2. Molecular docking studies, which are the first step of drug design, have been carried out on target coumarins, and the high score results we have obtained will be a stepping stone for scientists who will conduct in vitro andin vivo studies.

MM-PBSA analysis is significant to determine the relationship between drug candidate and target protein. In this analysis, the interaction energy can be calculated in picosecond level to evaluate the stability of drugprotein complex (Aktaş et al., 2020) . MM-PBSA analyses between coumarins 3, 5, 16 and remdesivir are performed by using nanoscale molecular dynamics (NAMD) and visual molecular dynamics (VMD) software.

Ethyl 7-Hydroxy-4,8-dimethylcoumarin-3-propanoate (1) can be synthesized according to the synthesis method in the literature (Abuknesha & Darwish, 2005) . Amide structure is obtained by nucleophilic reaction of ethyl carboxylate moiety of 7-hydroxy-4,8-dimethylcoumarin-3-propanoate (1) compound with free amines. N,N-dimethylethylenediamine (for 2), 4-(2-aminoethyl)morpholine (for 3), 1-(2-aminoethyl)piperidine (for 4), 1-(2-aminoethyl)piperazine (for 5), propan-1amine (for 6), aniline (for 7), pyridin-4-amine (for 8), 1H-imidazol-2-amine (for 9), 2-aminobenzimidazole (for 10), 2-aminooxazole (for 11), 2-aminobenzoxazole (for 12), 5aminoindole (for 13), 2-amino-1,9-dimethyl-1,9-dihydro-6Hpurin-6-one (for 14), cytosine (for 15), 5,6-dihydro-3H-purin-6-amine (for 16) and 2-aminobenzothiazole (for 17) were used for amidation reaction of 7-hydroxy-4,8-dimethylcoumarin-3-propanoate (Scheme 1).

Molecular docking studies were conducted to provide a theoretical perspective for possible molecular interactions of compounds with the target proteins. The theoretical binding affinities of these compounds for targeted proteins were also determined by docking calculations. Molecular docking calculations, energy minimization, and molecular visualization of docking results were carried out by using the Molecular Operating Environment software package (MOE, v2019.0102, Chemical Computing Group ULC) (ULC, 2019). Preparation of coumarin derivatives (1-17) compounds and model inhibitor molecules for molecular docking was performed with MarvinSketch software (ChemAxon, 2016) . Before the docking process, the drawing and editing of the newly designed coumarin compounds in the SD File format were done with the MarvinSketch suit program (ChemAxon, 2016). These molecular structures have been protonated, were charges added, and conformation minimization was performed with the root mean square gradient (RMS 0.001 kcal/mol/A2) by using the MMFF94 Forcefield parameters, which can be accessed in Energy Minimization protocols of this software (Halgren, 1996) . In this study, docking analyzes of coumarin derivatives (1-17) for 7 target proteins were performed. Of these target proteins Spike S1 subunit, NSP5, NSP12, NSP15, and NSP16 are COVID-19 proteins, ACE2 and VKORC1 are human proteins. The VKORC1 target protein is not directly related to COVID-19, but is involved in vitamin K-metabolism. Since it is the natural potential target protein for the coumarin derivatives we have developed, it has been subject to study. In docking calculations, the homology model with UniProtKB-code-Q9BQB6-in SWISS-MODEL-repository was used for the VKORC1 3 D protein structure (Bienert et al., 2017; Consortium, 2019) . Favipiravir is effective against coronavirus proteins due to its covalent interaction, coumarins containing morpholine, piperidine and piperazine exhibit NH þ or NH 2 þ ionic interaction, (Since HOMO orbitals are on amide derivatives and LUMO orbitals are on the coumarin structure, electrons migrate from the substituent to the main skeleton). As shown in their calculations, it is important that NSP12 has the potential to covalently bond with RNA components, similar to favipiravir. This indicates that compounds with NH þ properties may have an irreversibly strong inhibitory effect on NSP12-RNA. It will be important to demonstrate experimentally at the next stage (for HOMO-LUMO, Figure 3 and for NH þ ionic interaction of compound 3, Figure 4 ). The X-ray crystal structures as three-dimensional coordinates of these target proteins were obtained from the Scheme 1. Representation of ethyl 7-hydroxy-4,8-dimethylcoumarin-3-propanoate (1) and its amido derivatives (2-17).

Protein Data Bank (http://www.rcsb.org/) (Berman et al., 2000) . For use in docking calculations, structures with PDB IDs of 7BQY for 3CLpro, 7BV2 for NSP12, 6WXC (Kim et al., 2020) for NSP15, 6WKQ (Minasov et al., 2020) for NSP16, and 6M0J (Lan et al., 2020) for Spike S1 subunit and ACE2 protein were chosen as crystal structure models corresponding to these target proteins ( Figure 1 ). Structural defects in these target proteins were eliminated automatically with the "Structure Preparation" module of MOE suit software. Default parameters of MOE Protonate 3D Module were used to determine and optimize the overall low potential energy configuration of the terminal amides, hydroxyls, thiols, histidines, and hydrogenation positions of the titratable groups in a certain pH throughout the system (Temperature 300 K, pH 7, solvent 0.1 M, electrostatic energy cutoff 15 A with Generalized Born Model_GB/VI; van der Waals 800R3 cutoff 10 A, MOE Protonate 3D) (Lan et al., 2020) . The energy minimization of the system was performed with the Amber12: EHT Forcefield parameters, RMS with the gradient of 0.001 kcal/mol/A2, which can be reached in MOE Energy Minimization protocols (Case et al., 2012; Labute, 2007) . Possible ligand binding sites in the minimized protein were determined by MOE' SiteFinder module. The Site Finder methodology is a geometric method based on Alpha Shapes, a generalization of convex surfaces developed by Edelsbrunner (Gerber & M€ uller, 1995) .

Local docking of coumarin compounds (1-17) and model inhibitors to the active site of these targets proteins was performed via MOE using the default docking calculation parameters. The top-ranked pose defined by minimum energy (kcal/mol) for every compound, was used as the final molecular docking results. London DG scoring function was used for docking calculations. The London DG scoring function estimates the free energy of binding the ligand at a particular pose in a target structure. This scoring function is explained in detail in the user manual of the MOE software. After the initial scoring function for the obtained docking poses, the GBVI/WSA DG scoring function was used as the final docking scoring methodology. The GBVI/WSA DG is a forcefield-based scoring function that estimates the free energy of binding of the ligand from a given pose (Edelsbrunner & M€ ucke, 1994) . This scoring function is explained in detail in the user manual of the MOE software.

The Gaussian 09 (Frisch et al., 2009) program was used to perform theoretical calculations, and the computational results were visualized through GaussView 5.0 (Dennington et al., 2009) . The B3LYP functional [B3: Becke's three-parameter nonlocal exchange functional, (Becke, 1988 (Becke, , 1993 LYP: Lee-Yang-Parr's correlation function (Lee et al., 1988) ] was used for density functional theory (Kohn & Sham, 1965) with the 6-311 G(d,p) basis set. Geometry optimization was carried out in the gas phase. For geometry optimization of the compounds was chosen as a stable form with C1 symmetry. Also, frequency analysis of coumarins was performed. Total SCF Density (isoval ¼ 0.0004 e/au 3 ) mapped with electrostatic potential (ESP) was utilized to visualize the total electron density.

In drug design, the estimation of the pharmacophore properties of the target molecules saves time. ADMET is an acronym for absorption, distribution, metabolism, excretion, and toxicity, and a pharmacophore is an approach that offers parameters about the placement of a compound into a living organism. All of these five criteria determine the pharmacological activity of the compounds. pkCSM, a free online web server (http://structure.bioc.cam.ac.uk/pkcsm) (Pires et al., 2015) was used to predict the ADMET properties of the designed coumarin derivatives (1-17).

Molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) calculations were performed for coumarins 3, 5, 16 and Remdesivir. Binding free energies were determined for these ligand-protein structures in 7BV2 from 0 to 100 ns (Kollman et al., 2000) . For these calculations, Nanoscale Molecular Dynamics (NAMD) (Nelson et al., 1996) and Visual Molecular Dynamics (VMD) software were used. Binding energy was calculated by using Equation (1).

where DG Binding is the binding free energy, G Complex , G Protein , and G Ligand demonstrates the total free energy of the protein-ligand complex and total free energies of the isolated protein and ligand, respectively.

Ethyl 7-hydroxy-4,8-dimethylcoumarincoumarin-3-propanamido coumarin derivatives designed for the inhibition of SARS-CoV-2 proteins have been optimized by B3LYP/6-311G (d,p) level of theory. Possible molecular geometries of coumarins are shown in Figure 2 .

Electronic energy values for all designed coumarin compounds increased negatively as secondary amines were added to the structure compared to the main coumarin structure 1 (Figure 3 ). This indicates that the secondary amine addition to the coumarin structure makes the structure more active. The largest electronic energy value À27128.87 kcal/mol for ethyl 7-hydroxy-4,8-dimethycoumarinl-3-propanoate (1). The lowest electronic energy value is À44083.76 kcal/mol for 7-hydroxy-4,8dimethyl-3-N-(benzo[d]thiazol-2-yl)-propanamido-coumarin (17). HOMO (Highest Occupied Molecular Orbital) and LUMO (Lowest Unoccupied Molecular Orbital) energy levels, and the gap value between these two orbitals provide information about electronic transitions for the target molecules and are very important parameters. Also, the HOMO and LUMO values show nucleophilic and electrophilic attraction in the molecule as well as electronic transitions. The band gap (HOMO-LUMO gap) is a parameter that gives information about chemical reactivity and a narrow band gap means high reactivity and a wide band gap means lower reactivity according to the relevant molecule groups. Very useful for comparing chemical reactivity of compounds in a series. Band gaps were lowered by substituting secondary amines into ethyl 7-hydroxy-4,8-dimethylcoumarin-3-propanoate coumarin (1).

Docking analyzes were performed to understand the molecular interaction mechanisms between coumarin derivatives (1-17) and the target proteins. Also of these, docking analyzes of model inhibitor warfarin and some drugs (favipiravir, hydroxychloroquine, and remdesivir) reported to be effective against COVID-19 were also performed (Barnes et al., 2020; Beigel et al., 2020; Cai et al., 2020; Gautret et al., 2020; Geleris et al., 2020; Grein et al., 2020; . Docking results of all compounds are given in Tables 1, 2 , and Figures 4, 5, S1-S4 (Supporting material). According to docking calculations between the coumarin compound series (1-17) and NSP12, the compounds presenting the highest docking scores are 3, 5, and 16 with average binding affinities of À10.01, À9.36, and À9.09 kcal mol À1 , respectively. The binding affinities of these compounds to NSP12-RNA complex are À8.30, À9.10, and À8.39 kcal/mol, respectively.

The catalytic site major residues of NSP12 that interact with coumarin compounds (1-17) are Asp452, Tyr455, Met542, Arg553, Arg555, Thr556, Asp618, Tyr619, Pro620, Lys621, Cys622, Asp623, Arg624, Thr680, Ser 681, Ser682, Asn691, Phe694, Ser759, Asp760, Ser795. NSP12 (PDB ID: 7BV2) has magnesium ions in its catalytic domain. Since magnesium ions take part in natural catalytic processes, they are preserved in docking calculation and considered as a catalytic site component (as residues MG1004 and MG1005). It was seen in docking calculation that these magnesium ions, especially MG1004, form metal ligation with all coumarin compounds (1-17).

The residues of the NSP12 protein that show the highest interaction with coumarins (1-17) are Asp623, Thr556, Ly621, and Pro620, respectively. Asp623 and Thr556 residues usually show strong sidechain hydrogen bond (donor) interaction in different docking poses of coumarin compounds (1-17). When the different docking poses of coumarin derivatives with NSP12 were examined together, it was observed that the Lys621 residue had strong sidechain hydrogen bond Table 1 . Docking scores of coumarin derivatives (1-17) on SARS-CoV-2 receptors (7BQY3 for 3CLpro, 7BV23 for NSP12, 6WXC for NSP15, 6WKQ for NSP16, and 6M0J for Spike S1 subunit and ACE2 protein).

Docking scores 3CLPro

Spike-S1 ACE2 Spike-ACE2 NSP12 NSP12-RNA NSP15 NSP16 VKORC1 -6.46 -5.85 -6.50 -6.14 -6.82 -7.12 -6.59 -6. (acceptor) interactions with compounds 2, 9, 12, and 15; backbone hydrogen bond (acceptor) interactions with compounds 2,491,015 and 17; strong "arene attraction" with compounds 31,012 and 13; weak "arene attraction" with compounds 5, 7, 9 and 14. Pro620 residue shows strong arene attraction with compounds 121,415 and 17, while weak arene attraction with compounds 25,910 and 13. The residues of the NSP12 protein that show the highest interaction with model compounds (warfarin, favipiravir, hydroxychloroquine, and remdesivir) are Arg555 and Arg553, respectively. Arg555 and Arg553 residues usually show strong sidechain hydrogen bond (acceptor) interaction in different docking poses of model (reference) compounds. When the different docking poses of coumarin compounds with NSP12 were examined together, it was observed that the Arg555 residue had strong sidechain hydrogen bond (acceptor) interactions with model compounds warfarin and remdesivir; strong arene attraction interaction with model compound hydroxychloroquine in only one docking pose. Arg553 residue shows sidechain hydrogen bond (acceptor) interactions with all reference compounds in the different docking poses. At the same time, it was observed that warfarin and favipiravir created a strong arene attraction interaction with Ser682 residue of NSP12 protein.

While complex with the template and primary RNA components of NSP12 protein, the residues most frequently interacting with the coumarin compounds (1-17) are Arg555, Thr556, and Asp623, in docking calculations. The Arg555 residue generally shows side-chain hydrogen bond (acceptor) interactions with coumarin compounds series (except compounds 191,011 and 14) , weak arene attraction with compounds 23,414 and strong arene attraction interactions with compound 13. The Thr556 and Asp623 residues generally show side-chain hydrogen bond (donor) interactions with the coumarin compounds (1-17) other than compounds 5, 15, and 17. Strong metal ligation interactions are generally observed between the magnesium ions (especially the magnesium ion called MG1004 residue) and the coumarins (1-17) , which are the natural components of NSP12 or NSP12-RNA complex. However, while the magnesium ions in 

13 -Ã Values obtained by energy minimization. the NSP12 and/or NSP12-RNA complex had no interaction with the reference molecules warfarin and favipiravir, it was observed that they formed strong interactions with the reference molecules hydroxychloroquine and remdesivir. In docking analysis, the primary and template RNA components in the NSP12-RNA complex, especially template uracil nucleotide (t-U10) and the primary uracil nucleotide (p-U20), have strong molecular interactions with the coumarin derivatives (Figures 4 and 5) .

The reference compound warfarin has established sidechain hydrogen bond (acceptor) interactions with Arg555, Ser759, t-U10, and p-U20 residues, and arene attraction interactions with Ser682 and p-U20 residues of the NSP-RNA complex. Molecular interactions of other model compounds with the NSP12-RNA complex are given in Figures 4 and 5 .

The protein-ligand docked complexes of the compounds 3, 5, and 16 with the highest docking scores were subjected to energy minimization. As a result of this process, the final docking scores of compounds 3, 5, and 16 were found to be À10.20, À9.86, and À10.10 kcal/mol, respectively (Figures 4  and 5) . However, in docking analysis, the binding affinities of the model compounds warfarin, favipiravir, hydroxychloroquine, and remdesivir with the NSP12 protein were found to be À6.66, À4.81, À7.59, and À9.06 kcal/mol, respectively. When energy minimization was applied to the protein-ligand complexes obtained from the docking process of the model compounds with the NSP12 protein, the final docking scores of these compounds were found to be À7.83, À5.33, À11.03, and À11.20 kcal/mol, respectively (Figures S1-S4, Supporting material and Table 2 ).

In the coumarin compound series (1-17), the compounds with the highest docking score for the NSP12-RNA complex were again 3, 5, 16, and their binding affinities were found as À9.10, À8.39, and À8.30 kcal/mol, respectively. The docking scores of these compounds as a result of energy minimization were found as À9.33, À9.49, and À9.62 kcal/mol, respectively (Table 1 and Figures 4, 5) . In docking analysis, the binding affinities of the model compounds warfarin, favipiravir, hydroxychloroquine, and remdesivir for the NSP12-RNA complex were found to be À7.06, À4.85, À7.92, and À9.51 kcal/mol, respectively. The docking scores of these compounds as a result of energy minimization were found as À7.88, À5.43, À11.12, and À12.61 kcal/mol, respectively (Table 2 and Figures S1-S4, Supporting material) . According to the docking results, with the exception of 3, 5, and 16, both coumarin compounds (1-17) and selected model compounds generally bind to the NSP12-RNA complex with higher affinity than the NSP12 single protein. The specific Coumarin series compounds (2-17), except coumarin 1, almost all gave a higher docking score against VKORC1 than the warfarin model compound. While the warfarin docking score was À7.41 kcal/mol, the coumarin derivatives (2-17) were observed to have binding affinities between À7.58 and À8.66 kcal/mol and higher energetic values than warfarin, except for coumarin 1, which has a value of À7.28 kcal/mol. Another model compound, hydroxychloroquine, was observed to have a binding affinity of À8.07 kcal/mol with VKORC1. Nine coumarin derivatives gave a higher docking score against VKORC1 than the hydroxychloroquine model compound (Tables 1 and 2) .

Possible interactions of other model inhibitors used in the study with VKORC1 were investigated by docking calculations ( Figure 6 ). According to non-covalent docking calculations, "Remdesivir", an antiviral drug approved for use in European countries for the treatment of COVID-19, whose primary target is viral proteins, interestingly, showed a very high binding affinity for VKORC1 with À10.03 kcal/mol. The clinical success of remdesivir against COVID-19 has suggested that both the compound's direct targeting of virus proteins and its anticoagulant effect may be important in secondary clinical situations (tables) developing after COVID-19 infection. We can suggest that it may be beneficial to address this situation in the clinical examinations of remdesivir related to COVID-19. Favipiravir is a compound that shows a covalent and irreversible inhibition mechanism with its primary target molecule. Therefore, a comparison of non-covalent docking scores would not be appropriate.

Molecular docking has some disadvantages in the analysis of binding stability. In docking calculations, protein accepts as rigid while ligand accepts flexible. However, both protein and ligand are flexible in the living. The flexibility of both protein and ligand should be taken into consideration. For this aim, MM-PBSA calculations are performed for drug candidate, Coumarins 3, 5, 16 and Remdesivir. The binding energy is calculated in each 5 ns and given in Table 3 . Additionally, standard deviation (SD) is calculated and given in same table.

The binding energies from 0 to 100 ns are calculated for selected complex structures. The energy values are represented in Figure 7 .

According to Table 3 and Figure 7 , it can be said that selected drug candidates are effective against RNA dependent RNA polymerase (RdRp). However, the most effective one is remdesevir in the inhibition of NSP12 which is RdRp. Also, coumarin 3 and coumarin 5 have similar effectivity and their activities are better than coumarin 16. The complex structure between coumarin 3 and 7BV2 at 0,255,075 and 100 ns are represented in Figure 8 . 

In this study, coumarin derivatives containing secondary amine substituents were designed to develop effective antiviral agents to inhibit coronavirus. DFT method was used to estimate the ideal geometries and physicochemical properties of molecules. 17 coumarin derivatives were studied using molecular docking studies against MPro, NSP12 with RNA, NSP15, NSP16, and Spike with ACE2, which are important receptors of SARS-CoV-2, and their binding scores were compared with those of warfarin, favipiravir, remdesivir, and hydroxychloroquine model inhibitors. The coumarin series was found to be specific for NSP12 among all receptors, and the highest scores were obtained for NSP12 or NSP12/RNA. In the coumarin series, the compounds with the highest docking score for the NSP12-RNA complex were again 5, 16, and 3, and their binding affinities were found as À9.10, À8.39, and À8.30 kcal/mol, respectively. These scores are higher than the binding affinities of warfarin, favipiravir, and hydroxychloroquine, but lower than the binding affinity of remdesivir. These differences in docking calculation results between the coumarin compounds (1-17) and the main reference molecule warfarin, and it is thought that the coumarin compounds (1-17) may be more active than warfarin. Since coronavirus is known to cause pulmonary embolism, binding affinities against vitamin K epoxide reductase complex inhibitor have also been investigated, taking into account the anticoagulant property, which is an important property of coumarins. Other coumarins, except the aminefree starting coumarin, scored better than the blood-thinning drug warfarin, but again remdesivir scored best against the vitamin K epoxide reductase complex. In conclusion, the coumarin series in this study is specific against NSP12 and viral replication can be prevented by inhibition of this enzyme. Coumarins with high binding scores are suitable compounds for new designs and organic reactions, and new designs may be made using the functional groups of remdesivir compound and will be promising for the development of multifunctional and powerful antiviral drugs. MM-PBSA calculations are performed and it is found that coumarin 3, 5, and 16 are found as active against 7BV2. These ligands are in a stable interaction.

Coupling of enzymatic and immunoassay steps to detect E. coli: A new, highly sensitive tandem technique for the analysis of low levels of bacteria

New antiviral drugs for the treatment of COVID-19 instead of favipiravir

Asymptomatic coronavirus infection: MERS-CoV and SARS-CoV-2 (COVID-19)

Thromboembolism and anticoagulant therapy during the COVID-19 pandemic: Interim clinical guidance from the anticoagulation forum

Density-functional exchange-energy approximation with correct asymptotic behavior

Becke's three parameter hybrid method using the LYP correlation functional

Remdesivir for the treatment of Covid-19-Final report

The protein data bank

The SWISS-MODEL repository-new features and functionality

Experimental treatment with favipiravir for COVID-19: An open-label control study

Comparative genetic analysis of the novel coronavirus (2019-nCoV/SARS-CoV-2) receptor ACE2 in different populations

COVID-19 pneumonia with hemoptysis: Acute segmental pulmonary emboli associated with novel coronavirus infection

MarvinSketch: Marvin was used for drawing, displaying and characterizing chemical structures, substructures and reactions

In-silico investigation of phytochemicals from Asparagus racemosus as plausible antiviral agent in COVID-19

UniProt: A worldwide hub of protein knowledge

Structural modeling insights into human VKORC1 phenotypes

Natural products' role against COVID-19

Antibacterial activity of coumarins

GaussView (Version 5.0)

A novel angiotensin-converting enzyme-related carboxypeptidase (ACE2) converts angiotensin I to angiotensin

Three-dimensional alpha shapes

Are patients with hypertension and diabetes mellitus at increased risk for COVID-19 infection?

The novel coronavirus disease (COVID-19) threat for patients with cardiovascular disease and cancer

Hydroxychloroquine and azithromycin as a treatment of COVID-19: Results of an open-label non-randomized clinical trial

Observational study of hydroxychloroquine in hospitalized patients with Covid-19

MAB, a generally applicable molecular force field for structure modelling in medicinal chemistry

WHO Director-General's opening remarks at the media briefing on COVID-19-11

The species severe acute respiratory syndrome-related coronavirus: Classifying 2019-nCoV and naming it SARS-CoV-2

Compassionate use of remdesivir for patients with severe Covid-19

The origin, transmission and clinical therapies on coronavirus disease 2019 (COVID-19) outbreak-an update on the status

Merck molecular force field. I. Basis, form, scope, parameterization, and performance of MMFF94

Tissue distribution of ACE2 protein, the functional receptor for SARS coronavirus. A first step in understanding SARS pathogenesis

Therapeutic potential of coumarins as antiviral agents

SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by a clinically proven protease inhibitor

Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. The Lancet

VMD: Visual molecular dynamics

Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors

Covid-19: Risk factors for severe disease and death

ACE2 of the heart: From angiotensin I to angiotensin (1-7)

Crystal structure of Nsp15 endoribonuclease NendoU from SARS-CoV-2

Drug interactions with coumarin anticoagulants

Self-consistent equations including exchange and correlation effects

Calculating structures and free energies of complex molecules: Combining molecular mechanics and continuum models

Hypertension, the renin-angiotensin system, and the risk of lower respiratory tract infections and lung injury: Implications for COVID-19 European Society of Hypertension COVID-19 Task Force Review of Evidence

Identification of phytochemical inhibitors against main protease of COVID-19 using molecular modeling approaches

Protonate 3D: Assignment of macromolecular protonation state and geometry

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor

Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density

Outbreak of pneumonia of unknown etiology in Wuhan, China: The mystery and the miracle

Angstrom resolution crystal structure of NSP16-NSP10 heterodimer from SARS-CoV-2 in complex with sinefungin

The natural coumarins: Occurrence, chemistry, and biochemistry

NAMD: A parallel, object-oriented molecular dynamics program

Vitamin K epoxide reductase complex subunit 1 (VKORC1): The key protein of the vitamin K cycle

Targeting SARS-CoV-2 spike protein of COVID-19 with naturally occurring phytochemicals: An in silico study for drug development

How to discover antiviral drugs quickly

pkCSM: Predicting small-molecule pharmacokinetic and toxicity properties using graphbased signatures

Synthesis and in vitro evaluation of novel coumarin-chalcone hybrids as potential anticancer agents

A review of coronavirus disease-2019 (COVID-19)

Coumarins as anticancer agents: A review on synthetic strategies, mechanism of action and SAR studies

Structural basis of SARS-CoV-2 3CLpro and anti-COVID-19 drug discovery from medicinal plants

Molecular operating environment (MOE) (Version

Structural basis of RNA cap modification by SARS-CoV-2

Structural basis for RNA replication by the SARS-CoV-2 polymerase

Comorbid chronic diseases and acute organ injuries are strongly correlated with disease severity and mortality among COVID-19 patients: A systemic review and meta-analysis

Remdesivir in adults with severe COVID-19: A randomised, double-blind, placebo-controlled, multicentre trial

Analysis of therapeutic targets for SARS-CoV-2 and discovery of potential drugs by computational methods

Structural basis for inhibition of the RNA-dependent RNA polymerase from SARS-CoV-2 by remdesivir

COVID-19 and the cardiovascular system

Coronavirus disease 2019 (COVID-19): A perspective from China

The numerical calculations reported in this paper were fully performed at TUBITAK ULAKBIM, High Performance, and Grid Computing Center (TRUBA resources). We thank MNG Holding for their support. In this study, machinery and equipment obtained from the RGD-020 project of Sivas Cumhuriyet University Scientific Research Project Directorate were used in MM-PBSA Analyses. For this reason, we thank Sivas Cumhuriyet University Scientific Research Project Directorate. 

No potential conflict of interest is reported by the authors.