key: cord-0714106-xtaeb8xd
authors: Batool, Abida; Bibi, Nousheen; Amin, Farhat; Kamal, Mohammad Amjad
title: Drug designing against NSP15 of SARS-COV2 via high throughput computational screening and structural dynamics approach
date: 2020-12-01
journal: Eur J Pharmacol
DOI: 10.1016/j.ejphar.2020.173779
sha: ff496730992391ad65f0cead782d14c7e8857238
doc_id: 714106
cord_uid: xtaeb8xd

The rapid outbreak of the COVID-19 also known as SARS-CoV2 has been declared pandemic with serious global concern. As there is no effective therapeutic against COVID-19, there is an urgent need for explicit treatment against it. The focused objective of the current study is to propose promising drug candidates against the newly identified potential therapeutic target (endonuclease, NSP15) of SARS-CoV2. NSP15 is an attractive druggable target due to its critical role in SARS-CoV2 replication and virulence in addition to interference with the host immune system. Here in the present study, we integrated the high throughput computational screening and dynamic simulation approach to identify the most promising candidate lead compound against NSP15. 5-fluoro-2-oxo-1H-pyrazine-3-carboxamide (favipiravir), (3R,4R,5R)-3,4-Bis(benzyloxy)-5-((benzyloxy) methyl) dihydrofuran-2(3H)-one) remedesivir, 1,3-thiazol-5-ylmethyl N-[(2S,3S,5S)-3-hydroxy-5-[[(2S)-3-methyl-2-[[methyl-[(2-propan-2-yl-1,3-thiazol-4-yl)methyl]carbamoyl]amino]butanoyl]amino]-1,6-diphenylhexan-2-yl]carbamate (ritonavir), ethyl (3R,4R,5S)-4-acetamido-5-amino-3-pentan-3-yloxycyclohexene-1-carboxylate (oseltamivir), and (2S)-N-[(2S,4S,5S)-5-[[2-(2,6-dimethylphenoxy)acetyl]amino]-4-hydroxy-1,6-diphenylhexan-2-yl]-3-methyl-2-(2-oxo-1,3-diazinan-1-yl)butanamide (lopinavir) were chosen as a training set to generate the pharmacophore model. A dataset of ∼140000 compounds library was screened against the designed pharmacophore model and 10 unique compounds were selected that passed successfully through geometry constraints, Lipinski Rule of 5, and ADME/Tox filters along with a strong binding affinity for NSP15 binding cavity. The best fit compound was selected for dynamic simulation to have detailed structural features critical for binding with the NSP15 protein. Given our detailed integrative computational analysis, a Small molecule (3,3-Dimethyl-N-[4-(1-piperidinylcarbonyl) phenyl] butanamide) with drug-like properties and high binding affinity with the NSP15 is proposed as a most promising potential drug against COVID-19. The current computational integrative approach may complement high-throughput screening and the shortlisted small molecule may contribute to selective targeting of NSP15 to stop the replication of SARS-CoV2.

target due to its critical role in SARS-CoV2 replication and virulence in addition to interference with the host immune system. Here in the present study, we integrated the high throughput computational screening and dynamic simulation approach to identify the most promising candidate lead compound against NSP15. 5-fluoro-2-oxo-1H-pyrazine-3- (lopinavir) were chosen as a training set to generate the pharmacophore model. A dataset of ~140000 compounds library was screened against the designed pharmacophore model and 10 unique compounds were selected that passed successfully through geometry constraints, Lipinski Rule of 5, and ADME/Tox filters along with a strong binding affinity for NSP15 binding cavity. The best fit compound was selected for dynamic simulation to have detailed structural features critical for binding with the NSP15 protein.

In late 2019, the rapid outbreak of a novel coronavirus, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) was found in Wuhan, China as a root cause for many cases of respirational ailment. Presumably, the virus instigated as a result of zoonotic transmission among animals like bats and humans but heap on by human to human through common droplet contagion (Wang et al., 2020 , Li et al., 2020 and Sohrabi et al., 2019 .

The virus hastily blowout starting from China to above 212 countries worldwide and, till this date, infested almost more than 3 million people, and over 0. well as it will exert massive social impact and economic loss in the billions of dollars (Stoermer et al., 2020; Keogh-Brown et al., 2008) .

used to remove redundancy among these libraries. Open Babel (O'Boyle et al., 2011) was used for format conversion of library compounds like from .sdf to .mol or .pdb and vice versa.

The common featured ligand-based pharmacophore model for the training set of the molecules (Favipiravir, Remedesivir, Ritonavir, Oseltamivir, and Lopinavir) was created using Ligand Scout 4.3 (www.inteligand.com) (Amanlou and Mostafavi, 2017).

Pharmacophore modelling is based on the assembly of chemical functionalities and then the alignment of shared features of the training compounds set. Then by resulted pharmacophore model sensitivity and specificity check was applied to optimize and refine the libraries of small molecules to find only active hits and eliminate the inactive hits from initial screening.

The refined pharmacophore model was used as a query for virtual screening of the 0.14 million compounds of Antiviral, Antiviral HBV, Coronavirus, and FDA approved drug libraries using Ligand Scout 4.3. The basic key of success for the drug designing process is the enrichment of chemical database by which all the compounds with poor drug-like properties are removed and remaining filtered out hit molecules passes through a series of filters.

Initially, PAINS server (Baell and Holloway, 2010) was used to remove false positive hits, and then Osiris Property Explorer Applet (Dillard and Goldberg, 1978) (http://www.openmolecules.org/propertyexplorer/applet.html) and Osiris Data Warrior (Sander et al., 2015) (http://www.openmolecules.org/datawarrior/) were used to check Lipinski-filter (Lipinski, 2004) and geometry constraint from the screened compound.

Through this enrichment process, we filtered out ~3500 compounds with optimized geometrical constraints, Mwt 180-450kDa, LogP value of 1-5, five or fewer rotatable bonds, less than five hydrogen bond donors (HBD's), and less than ten hydrogen bond acceptors (HBAs). Moreover, SCIfinder (Ridley, 2009) was used to explore whether our active hits are previously reported in the literature, and also similarity search was done to assemble and inspect all those hits with> 80 % similarity.

The optimized screened compounds (~3500), with adequate pharmacological features, 

The present study was designed to perform high throughput virtual screening extended with a structural dynamics approach to identify the potential candidate drug against NSP15.

Ritonavir, remedesivir, oseltamivir, lopinavir, and favipiravir ( Fig. S1 ) were used as a training set to generate the pharmacophore model. In an optimized model, their pharmacophore-fit score was 45.70, 47.48, 46.46, 48.14, and 46 .09 respectively and they shared 4 common features i.e., each has two hydrogen bonds donors and two hydrogen bond acceptors. 2D and 3D view of the pharmacophore model are shown in Fig. 1 A and B. This model was used as a template for a screening of 140000 compounds from Antiviral, Antiviral HBV, Coronavirus, and FDA approved drug libraries. By the end of this screening process, we got ~3500 compounds with shared featured same as the pharmacophore model. We further screened out these 3500 compounds by removing false positive and by taking Lipinski rule of five and geometry constraints as standard. ~500 compounds were shortlisted. These ~500 compounds were in full agreement with pharmacophore features, geometry constraints, and Lipinski Rule of 5.

To investigate the critical interactions with key residues of active binding sites of NSP15 protein of SARS-CoV2, an improved dataset of 500 compounds was subjected to molecular docking studies against the active binding pocket of NSP15 protein of SARS-CoV2. Docking analysis resulted in 50 docked inhibitors with strong binding affinities towards NSP15 protein ( phenyl] butanamide pretty well within the binding pocket of the target protein ( Fig. 3 A and B ).

To gauge stability, convergence, energetic and structural properties, 20 ns MD simulation was performed with selected putative lead compound and NSP15 of SARS-CoV2. Resulting J o u r n a l P r e -p r o o f trajectories were cautiously analysed to determine the stability, energetics, and structural properties and interactions during MD simulations.

Plots of root mean square deviation (RMSD) and root mean square fluctuation (RMSF) were generated to assess the stability and fluctuations. RMSD profile for bound NSP15 (NSP15screening hit) was compared with the apo-state (NSP15) as a reference. An overall convergence of energies indicated well-behaved behaved-systems with equilibration. The RMSD plot shows values between the C-alpha atom of complex and apo-proteins were below 2Å, signifying the system stability ( Fig. 4A ). RMSD plot verified more binding stability in NSP15 when it is bound to lead compound. These results strengthened our docking results as they were in good agreement with each other. By thorough analysis of dynamic trajectories at different ns conformational changes in NSP15 binding cavity were captured most likely to critical for binding with the potential NSP15 inhibitor. Through the comparative analysis of NSP15 apo and NSP15 bound state, significant conformational changes were observed in the loop region and secondary structure surrounding the binding cavity while the residues involved in the binding showed stability except small conformational switches to facilitate binding with the inhibitor. Lys252 and Lys277 tilted towards the cavity to make tight π-π stacking interaction and Pi-alkyl interaction with the bound inhibitor while Lys 71 and ser 198 showed inward push to make H-bonding with the inhibitor. Apart from these significant fluctuations in the loop region J o u r n a l P r e -p r o o f surrounding the binding residues were observed to help key residues to attain favourable binding pose (Fig. 5 ).

Currently, COVID19 has been proved the most threatening and most frightening epidemic of this century so far. And most alarming is the fact that still there is no therapeutic available against this disease. In vitro process of drug designing that includes screening and testing of millions of compounds is time-consuming and very costly. The lack of explicit therapeutics against the novel COVID-19 urge for the active drug discovery for which computational methods offer a fast and cost-efficient approach (Murumkar et al., 2010) .

The present study was designed to use high throughput computational screening and structural dynamics approach to find out putative inhibitor against NSP15 protein of SARS-CoV-2. NSP15 that is considered vital for the SARS CoV-2 life cycle and virulence is for Ritonavir (Zhang & Yap, 2004) . Binding analysis of these training compounds against NSP15 revealed the preference for those specific amino acids (Fig. S2) .

Initially, by using 5 compounds (ritonavir, remedesivir, oseltamivir, lopinavir, and favipiravir) that showed positive therapeutic potential for coronavirus patients in different countries, a pharmacophore model was generated. The generated pharmacophore model was screened against a small molecule dataset of 140000 compounds with subsequent filtration to increase specificity, resulting in the isolation of 10 hits as putative NSP15 specific inhibitors.

Finally, 3,3-Dimethyl-N-[4-(1-piperidinylcarbonyl) phenyl] butanamide was selected for detailed structural demonstration as representative screening hit for specific targeting of NSP15 protein of SARS-CoV2 (Table S3 and S4). Docking analysis reveals that along with Ser198, Lys71, Lys90, Arg199, Leu252, and Tyr279 are also more critical residues in the binding cavity of NSP15 (Fig. S2 ).

Detailed scrutinization of the dynamic behaviour of the final screening hit revealed important structural details of NSP15 binding pocket upon binding to lead compound. Comparative analysis of apo-state and inhibitor bound NSP15 complex uncovered significant structural compactness that may prove crucial to disrupting the SARS-CoV-2 virulence in the human body. During MD simulations, key substrate interacting residues (Lys71, Lys90, Ser198, and Leu252) were mainly detected in inhibitor binding which showed that isolated screening hot ma has a strong affinity towards the binding cavity of NSP15 ( Fig. 4a and 4b) . Therefore, detailed structural behavior monitoring strengthens our screening approach as NSP15 protein showed more stability in behavior upon binding to a putative inhibitor molecule. Our binding J o u r n a l P r e -p r o o f analysis with our pharmacophore compounds strengthens our selection of lead compounds by exactly fitting within the same cavity.

Given our results analysis, we propose that our lead compound may prove more effective and specific for NSP15 targeted therapy. Our lead compound i.e. 3,3-Dimethyl-N-[4-(1piperidinylcarbonyl) phenyl] butanamide is a small molecule with highly positive features of drug-likeness and Medicinal Chemistry can be the best potential drug against SARS-CoV2 inhibition, and it can be used with confidence for further in vitro analysis to investigate it with prospective of medicinal use.

To this date, no specific drugs or vaccines are available to treat SARS-CoV2 despite its close -(3-methyl-5-sulfamoyl-1,3,4thiadiazol-2-ylidene)acetamide 4 hydron;8,12,14-trihydroxy-5methyl-11,13bis(methylamino)-2,4,9trioxatricyclo[8.4.0.0 3,8 [(8S,9R,10S,11S,13S,14S,17R)-9-fluoro-11,17-dihydroxy-10,13-dimethyl-3-oxo-1,2,6,7,8,11,12,14,15,16- 

Rufener Data Warrior: An open-source program for chemistry aware data visualization and analysis

PatchDock and SymmDock: Servers for rigid and symmetric docking

Overcoming nonstructural protein 15-nidoviral uridylate-specific endoribonuclease (nsp15/NendoU) activity of SARS-CoV-2

Agha World Health Organization declares global emergency: A review of the 2019 novel coronavirus (COVID-19)

From SARS to MERS, thrusting coronaviruses into the spotlight

Homology Models of the Papain-Like Protease PLpro from Coronavirus 2019-nCoV ChemRxiv

Thailand sees apparent success treating Coronavirus with drug cocktail

Review of the 2019 novel coronavirus (SARS-CoV-2) based on current evidence