key: cord-0928170-o0jysed8
authors: Bahadur Gurung, Arun; Ajmal Ali, Mohammad; Elshikh, Mohamed S.; Aref, Ibrahim; Amina, Musarat; Lee, Joongku
title: An in silico approach unveils the potential of antiviral compounds in preclinical and clinical trials as SARS-CoV-2 Omicron inhibitors
date: 2022-04-22
journal: Saudi J Biol Sci
DOI: 10.1016/j.sjbs.2022.103297
sha: ed3d26ca9d50d189e55dd9d9439dd793dc140625
doc_id: 928170
cord_uid: o0jysed8

The increased transmissibility and highly infectious nature of the new variant of concern (VOC) that is severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Omicron and lack of effective therapy need the rapid discovery of therapeutic antivirals against it. The present investigation aimed to identify antiviral compounds that would be effective against SARS-CoV-2 Omicron. In this study, molecular docking experiments were carried out using the recently reported experimental structure of omicron spike protein in complex with human angiotensin-converting enzyme 2 (ACE2) and various antivirals in preclinical and clinical trial studies. Out of 36 tested compounds, Abemaciclib, Dasatinib and Spiperone are the three top-ranked molecules which scored binding energies of -10.08 kcal/mol, -10.06 kcal/mol and -9.54 kcal/mol respectively. Phe338, Asp339, and Asp364 are crucial omicron receptor residues involved in hydrogen bond interactions, while other residues were mostly involved in hydrophobic interactions with the lead molecules. The identified lead compounds also scored well in terms of drug-likeness. Molecular dynamics (MD) simulation, essential dynamics (ED) and entropic analysis indicate the ability of these molecules to modulate the activity of omicron spike protein. Therefore, Abemaciclib, Dasatinib and Spiperone are likely to be viable drug-candidate molecules that can block the interaction between the omicron spike protein and the host cellular receptor ACE2. Though our findings are compelling, more research into these molecules is needed before they can be employed as drugs to treat SARS-CoV-2 omicron infections.

The increased daily reported cases of coronavirus disease 2019 due to the Omicron variant (B.1.1.529) of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has added further to the global health crisis. This new variant of concern (VOC) was first detected on November, 1 in Botswana and reported to the World Health Organization (WHO) on November 26, 2021. Since then, it has rapidly spread to at least 38 countries (WHO Classification of Omicron (B.1.1.529), 2021). Omicron features a few deletions and over 30 mutations, some of which (e.g., 69-70del, T95I, G142D/143-145del, K417N, T478K, N501Y, N655Y, N679K, and P681H) overlap with those seen in the alpha, beta, gamma, or delta VOC (GISAID, 2021) . These deletions and mutations have been linked to its increased transmissibility, stronger viral binding affinity, and antibody escape. Other known omicron mutations include those that improve transmissibility and alter binding affinity (Greaney et al., 2021; Harvey et al., 2021) . Unfortunately, the implications of the majority of the remaining omicron mutations are unknown, leaving a lot of questions about how the whole set of deletions and mutations may affect viral behaviour and vulnerability to natural and vaccine-mediated immunity (Karim and Karim, 2021) . This fifth VOC is considered to be three times more contagious than the original SARS-CoV-2 and maybe even more so than the delta variant (Gao et al., 2021) .

Computational drug discovery is a successful technique for speeding up and reducing the cost of drug development. The applicability of computational drug discovery has been extended and broadly applied to nearly every stage in the drug discovery and development workflow, including target identification and validation, lead discovery and optimization, and preclinical tests, due to the dramatic increase in the availability of biological macromolecule and small molecule information (Ou-Yang et al., 2012) . One of the most extensively utilised structurebased drug design (SBDD) methods is virtual screening based on molecular docking. The molecular docking approach may be used to represent the atomic level interaction between a small molecule and a protein, allowing us to define small molecule behaviour in target protein binding sites and elucidate essential biochemical processes (McConkey et al., 2002) .

The technique is becoming increasingly popular in computer-aided drug design as more biomolecular structures are being submitted to the protein data bank and made freely available. However, standard molecular docking approaches do not take into account the target's flexibility, which could lead to the omission of some active molecules. Molecular dynamics (MD) simulations, on the other hand, may obtain numerous target conformations and have been frequently employed in molecular docking to account for target flexibility (Liu et al., 2018) . MD simulations paired with binding free energy estimates have also been found to be an effective technique for increasing the enrichment factor of virtual screening by improving the accuracy of target-ligand binding ability prediction (Rastelli et al., 2010) . The recently solved experimental structure of SARS-CoV-2 Omicron spike protein in complex with human ACE2 (PDB ID: 7WBL) (Han et al., 2022) offers an opportunity to identify small molecule inhibitors. The transmembrane spike (S) glycoprotein, which forms homotrimers projecting from the viral surface, is responsible for coronavirus (CoV) entry into host cells (Tortorici and Veesler, 2019) . S is made up of two functional subunits that are responsible for viral and cellular membrane fusion (S1 subunit) as well as binding to the host cell receptor (S2 subunit). S is cleaved at the interface between the S1 and S2 subunits in many CoVs, although they remain non-covalently coupled in the prefusion conformation (Kirchdoerfer et al., 2016; Walls et al., 2016) . S is cleaved further by host proteases at the socalled S2′ site, which is positioned immediately upstream of the fusion peptide in all CoVs (Madu et al., 2009; Millet and Whittaker, 2015) . It's been suggested that this cleavage prepares the protein for membrane fusion by causing substantial irreversible conformational changes (Park et al., 2016) . The coronavirus S glycoprotein is the principal target of neutralising antibodies (Abs) and the focus of therapeutic and vaccine design because it is surface-exposed and mediates viral entry into host cells (Walls et al., 2020) . In this study, we have used the molecular docking technique and molecular dynamics simulations to screen small antiviral molecules which can effectively bind to the omicron spike protein and hinder its interaction with the host cellular receptor. We have proposed a few molecules which may be developed into antiviral drugs against omicron.

2.1. Retrieval of ligands and protein structure A total of 36 compounds with preclinical or clinical trial evidence against previous variants of SARS-CoV-2 were selected for the study (Xiang et al., 2021) . The three-dimensional structures of the compounds were obtained from the PubChem database (Kim et al., 2016) .

The compounds with the 2D structure were converted into the 3D structure using OpenBabel version 2.4.1 (O'Boyle et al., 2011) . The structural optimization of the compounds was performed using the Merck molecular force field 94 (MMFF94) (Halgren, 1996) . The threedimensional structure of SARS-CoV-2 Omicron spike protein in complex with human ACE2 was retrieved from protein data bank (PDB) using accession PDB ID: 7WBL (Han et al., 2022) . The cryogenic electron microscopy (cryo-EM) structure has been resolved at a resolution of 3.40 Å.

The protein-protein interface statistics such as interface residues, interface area, hydrogen bonds, disulphide bonds and non-bonded contacts were evaluated using the PDBsum program (Laskowski, 2009 ).

Each compound was prepared for molecular docking experiments by the addition of Gasteiger charges, hydrogen atoms and optimally defined torsions using AutoDock Tools-1.5.6 (Morris et al., 2009) . The heteroatoms, including ions, co-crystallized ligand and water molecules, were removed from the Omicron spike protein. AutoDock Tools-1.5.6 tool (Morris et al., 2009) was used for adding a suitable number of polar hydrogen atoms and Kollman charges to the target protein.

The Lamarckian genetic algorithm was employed for molecular docking experiments where the docking parameters were selected from our previously reported methodology (Gurung, 2020) . AutoDock4.2 software was used to perform blind molecular docking ( 

The DataWarrior version 4.6.1 program was used to determine the physicochemical characteristics of the compounds (Sander et al., 2015) .

GROningen MAchine for Chemical Simulations (GROMACS) 2019.2 software (Hess et al., 2008) with the GROMOS96 43a1 force field was used to run 100-ns MD simulations on the free omicron spike protein and its complexes with lead molecules. Using a three-point model for water termed simple point charge (SPC216), the systems were solvated in a water-filled 3-D cube with 1 Å spacing. A leap-frog temporal integration technique was used to integrate Newton's equations of motion. The systems were neutralised, and the quantity of energy used was optimised. The temperature was set to 300 K, and the systems were equilibrated for 300 ps in the NVT (Number of particles, Volume, and Temperature) ensemble, and another 300 ps in the NPT (Number of particles, Pressure, and Temperature) ensemble.

Essential dynamics (ED), also known as principal component analysis (PCA), was investigated by diagonalizing a covariance matrix of C-α atoms of the system. Eigenvectors were used to define the directions of massive coordinated motions, and the eigenvalues corresponded to mean-square positional fluctuation (Amadei et al., 1993) .

A total of 36 small molecule inhibitors with preclinical and clinical trial data against COVID-19 were selected for the present study ( Table 1 ). The rationale behind the selection of these molecules is to see if they work against the omicron variant of SARS-CoV-2 as well.

We have utilised the very recently reported experimental structure of omicron SARS-CoV-2 spike with human ACE2 receptor for our studies. PDBsum analysis reveals that the proteinprotein complex is stabilized through one salt bridge, ten hydrogen bonds and 87 non-bonded contacts ( Table 2 ). The interaction of omicron spike protein to ACE2 involves Ser19, Gln24, Thr27, Phe28, Lys31, His34, Glu35, Asp38, Tyr41, Gln42 Leu79, Met82, Tyr83, Lys353, Gly354, Asp355 and Arg357 of ACE2 and Tyr449, Arg493, Tyr453, Leu455, Phe456, Ala475, Gly476, Asn477, Phe486, Asn487, Tyr489, Ser496, Arg498, Thr500, Tyr501, Gly502 and His505 of omicron spike protein (Figure 1) . The top 3 lead molecules which show considerable binding affinity to the omicron spike protein receptor include Abemaciclib (ΔG=-10.08 kcal/mol), Dasatinib (ΔG=-10.06 kcal/mol) and Spiperone (ΔG=-9.54 kcal/mol) ( Table 1) . These molecules bind to the spike interface region hydrophobic interactions and hydrogen bonds. For example, Abemaciclib binds to the spike protein through hydrophobic interactions with Tyr365, Leu368, Tyr369, Asn370, Phe374, Phe377, Cys379, Val382, Ser383, Pro384, Leu387, Asn388, Cys432, Ile434 and Phe515 (Figure 2A) Figure 2B) . The third lead molecule Spiperone binds to the spike protein receptor through hydrophobic interactions with Phe318, Asp339, Asp364, Tyr365, Val367, Leu368, Tyr369, Phe377, Pro384, Cys432, Ile434 and Leu513 ( Figure 2C) . The compounds were also evaluated for their oral bioavailability which includes parameters such as molecular weight (MW), LogP (partition coefficient between n-octanol and water), number of hydrogen bond donors (HBD), number of hydrogen bond acceptors (HBA), rotatable bond (RB) count, topological polar surface area (TPSA) and drug-likeness scores ( Table 3) . All the three identified lead molecules-Abemaciclib, Dasatinib and Spiperone show good drug-likeness scores of 6.2983, 8.0376 and 9.9352 respectively. Although Abemaciclib's molecular weight exceeds 500, its other physicochemical properties are within the permissible range of Lipinski's rule of five (Lipinski, 2004 ) and Veber's rule (Veber et al., 2002) , and the other two lead compounds, Dasatinib and Spiperone, totally obey these two drug-likeness rules.

The conformational changes and stability of the free omicron spike protein and its complexes with the top three lead compounds (Abemaciclib, Dasatinib, and Spiperone) were investigated using a 100-ns molecular dynamics simulation in an aqueous environment. The geometric parameters of the systems were computed from their trajectories, including root mean square deviation (RMSD), root mean square fluctuation (RMSF), the radius of gyration (Rg), solvent accessible surface area (SASA), and the number of hydrogen bonds (NHBs) ( Table 4 ). The RMSD is a typical way of determining the structural distance between coordinates which allows us to calculate the average distance between atoms in a group (e.g. The highly coordinated modes of fluctuations for Omicron spike and Omicron spike complexed with the top 3 ranked compounds were studied using essential dynamics (ED). A covariance matrix was generated for this by selecting the protein's Cα atoms (N = 195). After diagonalizing the covariance matrix for free omicron spike protein, Omicron spike_Abemaciclib, Omicron spike_Dasatinib and Omicron spike_Spiperone, the trace of the covariance matrix was found to be 5.98045 nm 2 , 5.0047 nm 2 , 6.31747 nm 2 and 7.47092 nm 2 respectively. The first few eigenvectors, out of a total of 585 (3N), play a significant role in the system's fluctuations. The eigenvalue curve of Omicron spike_Dasatinib and Omicron spike_Spiperone complex were steeper than that of free omicron spike protein except for Omicron spike_Abemaciclib indicating that fewer eigenvectors are required to comprehend the same degree of total mean square fluctuations (Figure 8) (18.1754%), with corresponding cosine values of 0.00296267, 0.0241793, 0.0791846 and 0.00316657. We chose only eigenvectors 1 and 2 for our investigations since they contribute more to total mean square fluctuations and their cosine values are less than 1 (sufficient conformational sampling). The free omicron spike protein is more compact and clustered than omicron spike-ligand complexes, indicating that its fluctuations are less coordinated (Figure 9 ). The fluctuation represented by PC1 and PC2 is also projected as a free energy landscape. The Gibbs free energies of conformation sub-spaces of in free omicron spike protein, Omicron spike_Abemaciclib, Omicron spike_Dasatinib and Omicron spike_Spiperone were determined to be 0 to 7.87 kJ/mol, 0 to 8.88 kJ/mol, 0 to 8.17 kJ/mol and 0 to 8.3 kJ/mol, respectively (Figure 10) , showing that free omicron spike protein has more energetically favourable conformational transitions than its complexes. In addition, across the free energy landscape, a few meta-stable states or energy minima (in blue) were found in a free omicron spike protein and omicron spike-ligand complexes. Furthermore, at T = 300 K, the absolute entropy computed using the quasi-harmonic approximation method (Andricioaei and Karplus, 2011 ) was found to be 41977.5 J/molK, 42356.0 J/molK, 42269.8 J/molK and 41989.3 J/molK for free omicron spike protein, Omicron spike_Abemaciclib, Omicron spike_Dasatinib and Omicron spike_Spiperone respectively. The results indicate that the disorder or randomness in the omicron spike protein structure increases upon interaction with the lead molecules.

The increased number of new cases due to the new variant of concerns, SARS-CoV-2 omicron, offers a challenge to global health systems as the world fights COVID-19. With a significant number of mutations and deletions, the fifth variant of concern is thought to be more infectious than previously found variants (Gao et al., 2021) . More clinical evidence is needed to determine the virus's actual pathogenicity, transmissibility, and ability to neutralize serum antibodies. Because there are no viable treatments for the previous variants as well, discovering therapeutic compounds for the omicron is a difficult challenge that will necessitate much scientific research. SARS-CoV-2 uses the spike glycoprotein to attach to the cellular receptor ACE2 to enter into the host cells and mediate its infectious cycle. Our current study is aimed at the identification of small molecules which can inhibit the interaction between omicron spike protein and ACE2. intolerance to previous therapy (Kantarjian et al., 2006) . Spiperone is an antagonist for the dopamine D2 receptor. It is not used even though it has antipsychotic properties. It's more commonly utilised as a pharmacological tool for studying neurotransmitter receptors (Davis, 2007) . The molecular dynamics simulation results revealed stability of the trajectories through 100 ns and displayed differences in the geometric properties between the free omicron protein and its complexes with the lead molecules. Essential dynamics and entropic analysis also indicate the modulation of the structure of the spike protein by these lead molecules indicating that these compounds can impair the function of the omicron spike protein.

To determine the binding strength and thermodynamic parameters of interaction between omicron spike protein and the lead compounds, experimental approaches such as surface plasmon resonance (SPR) and isothermal titration calorimetry (ITC) as well as spectroscopic techniques are recommended. Despite the fact that the current research is totally computational, the outcomes of our studies will aid in further experimental investigations and the development of therapies for SARS-CoV-2 omicron infections.

The extremely contagious nature of the novel omicron variant necessitates the development of effective drug molecules to combat its widespread spread. We found three candidate drugs Abemaciclib, Dasatinib and Spiperone with high binding affinity for omicron spike glycoprotein and good drug-likeness scores using molecular docking and physicochemical property analysis. The binding of the lead compounds causes aberrations in the structural properties of the omicron spike protein as revealed from molecular dynamics simulation investigations. These compounds could be developed into SARS-CoV-2 omicron entry inhibitors, which would help to reduce infection and spread.

The authors report no conflicts of interest in this work. 

Protein-protein complex between human ACE2 (pink) and omicron spike protein (aquamarine) and with the interface residues shown in yellow and orange spheres for ACE2 and spike protein respectively. (B) PDBSum analysis shows the interface residues and molecular interactions between ACE2 (Chain A) and omicron spike (Chain B). 

The authors report no conflicts of interest in this work. 

Essential dynamics of proteins

On the calculation of entropy from covariance matrices of the atomic fluctuations

XPharm: The Comprehensive Pharmacology Reference

Abemaciclib, a third CDK 4/6 inhibitor for the treatment of hormone receptor-positive, human epidermal growth factor receptor 2-negative advanced or metastatic breast cancer

Omicron variant (B. 1.1. 529) of SARS-CoV-2, a global urgent public health alert!

Tracking of Variants

Complete mapping of mutations to the SARS-CoV-2 spike receptor-binding domain that escape antibody recognition

In silico structure modelling of SARS-CoV-2 Nsp13 helicase and Nsp14 and repurposing of FDA approved antiviral drugs as dual inhibitors

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

others, 2022. Receptor binding and complex structures of human ACE2 to spike RBD from Omicron and Delta SARS-CoV-2. Cell

SARS-CoV-2 variants, spike mutations and immune escape

GRGMACS 4: Algorithms for highly efficient, load-balanced, and scalable molecular simulation

Omicron SARS-CoV-2 variant: a new chapter in the COVID-19 pandemic

PubChem Substance and Compound databases

Pre-fusion structure of a human coronavirus spike protein

PDBsum new things

LigPlot+: multiple ligand-protein interaction diagrams for drug discovery

Lead-and drug-like compounds: the rule-of-five revolution

Molecular dynamics simulations and novel drug discovery

Characterization of a highly conserved domain within the severe acute respiratory syndrome coronavirus spike protein S2 domain with characteristics of a viral fusion peptide

The performance of current methods in ligand--protein docking

Host cell proteases: critical determinants of coronavirus tropism and pathogenesis

AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility

Open Babel: An open chemical toolbox

Computational drug discovery

Proteolytic processing of Middle East respiratory syndrome coronavirus spikes expands virus tropism

Fast and accurate predictions of binding free energies using MM-PBSA and MM-GBSA

DataWarrior: an open-source program for chemistry aware data visualization and analysis

Structural insights into coronavirus entry

Molecular properties that influence the oral bioavailability of drug candidates

Structure, function, and antigenicity of the SARS-CoV-2 spike glycoprotein

Cryo-electron microscopy structure of a coronavirus spike glycoprotein trimer

Classification of Omicron (B.1.1.529): SARS-CoV-2 variant of concern

Recent advances in developing small-molecule inhibitors against SARS-CoV-2

The authors would like to extend their sincere appreciation to the Researchers Supporting Project number (RSP-2021/306), King Saud University, Riyadh, Saudi Arabia.