Repurposing the Combination Drug of Favipiravir, Hydroxychloroquine and Oseltamivir as a Potential Inhibitor against SARS-CoV-2: A Computational Study

The virus SARS-CoV-2 has created a situation of global emergency all over the world from the last few months. We are witnessing a helpless situation due to COVID-19 as no vaccine or drug is effective against the disease. In the present study, we have tested the applicability of some combination drugs against COVID-19. We have tried to understand the mechanism of action of some repurposed drugs: Favipiravir (F), Hydroxychloroquine (H) and Oseltamivir (O). The ADME analysis have suggested strong inhibitory possibility of F, H, O combination towards receptor protein of $3CL^{pro}$ of SARS-CoV-2 virus. The strong binding affinity, number of hydrogen bond interaction between inhibitor, receptor and lower inhibition constant computed from molecular docking validated the better complexation possibility of F+H+O: $3CL^{pro}$ combination. Various thermodynamical output from Molecular dynamics (MD) simulations like potential energy ($E_g$), temperature (T), density, pressure, SASA energy, interaction energies, Gibbs free energy ($\Delta G_{bind}$) etc., also favored the complexation between F+H+O and CoV-2 protease. Our In-Silico results have recommended the strong candidature of combination drugs Favipiravir, Hydroxychloroquine and Oseltamivir as a potential lead inhibitor for targeting SARS-CoV-2 infections.


Introduction
In 21 st century, from the last few months the whole world is witnessing the pandemic due to the recent outbreak by the disease COVID-19 caused by novel coronavirus. It is believed that the virus was originated from the wet meat market of Wuhan city of China sometime in December in 2019 Guan et al., 2020]. Corona viruses are not new to the mankind. From the last few centuries, mankind had perceived the presence of these viruses in the form of avian flu virus around 2003 [Keil, et al., 2006], Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV) around 2003 [ Rota, et al., 2003], Middle East Respiratory Syndrome Coronavirus (MERS-CoV) around 2012 [Su, et al., 2016]. The latest pandemic disease COVID-19 that has created a public health emergency in this series aroused due to novel coronavirus: SARS-CoV-2 . The whole scientific fraternity in the world is working hard day and night to get a medicine/vaccine to fight against this deadly virus, but the reality is that till date we have no definite medication for COVID-19. As of today, the 5 th November, 2020 according to WHO data there are 47,930,397 confirmed patients of COVID-19 disease and 1,221,781 deaths worldwide have been reported [https://www.who.int]. The transmission of the disease as well as the death toll due to COVID-19 expanding exponentially day by day. CoV-2 can easily spread by contact transmission. The infection can easily be transmitted through respiratory droplets and also through surface contamination. Aerosol transmission may be considered as another way of transmission of the disease [Rothe, et al., 2020]. Till now research outputs says that the virus can easily be transmitted by symptomatic asymptomatic and presymptomatic patients [Furukawa, et al., 2020;Chunyang, et al., 2020;Ghinai, et al., 2020]. There is an urgent need of effective treatment method to limit the transmission of this disease as early as possible.
Coronavirus is spherical in shape having diameter between 80 -160 mm. The spherical envelope surface is covered with spike glycoproteins (S), membrane proteins (M) and envelope proteins (E) [Woo, et al., 2005;Gorden, et al., 2020]. The main envelope of virus contains a spiral nucleocapsid which is formed by genomic RNA and phosphorylated nucleocapsid (N) protein [Priyadarsini, et al., 2020]. The main genome of CoV is comprised of a longest known genome among RNA viruses which is a single-stranded positive-strand RNA ranging from 26 Kb to 32 Kb in length. Coronaviruses can be divided into four categories: α, β, γ, and δ. α and β coronaviruses only infect mammals [Heidary, et al., 2020], while γ and δ mainly infect birds [Muradrasoli, et al., 2010]. SARS-CoV-2 is a novel βcoronavirus. The S proteins of the virus initiate the attachment and entry to the host cells through the current medical industry we have many examples of such type of antiviral drugs which have multiple use. Similarly, overlapped molecular pathways are also observed for many diseases. For example, Sofosbuvir, Ribavirin and Remdesivir are well known approved drugs for hepatitis C virus [Fried, et al., 2002]. Remdesivir has also established its strong repurposing potential against Ebola, Zika viruses and now CoV-2 virus [Elfiky et al., 2020]. Remdesivir has shown very effective impact on CoV-2 infection by blocking the replication of the virus inside human body . Favipiravir and Oseltamivir have been investigated for the treatment of Ebola virus, Lassa virus and influenza viruses A and B. Similarly, Lopinavir and Ritonavir were originally developed for the treatment of HIV patients. Now in current CoV-2 emergency situation all the above-mentioned viral medicines are being used to treat COVID-19 patients [Costanzo, et al., 2020]. Like antiviral drugs, many available antibiotic drugs like Doxycycline, Chloroquine, Hydroxychloroquine etc., also are being used as repurposed drugs for the treatment of various bacterial and viral diseases. Normally antibiotic drugs attacked the walls of a bacteria. Also, these drugs prevent the bacteria from synthesizing a molecule in the cell wall called peptidoglycan. Peptidoglycan provides the wall with the strength it needs to survive in the human body [Kapoor, 2017]. Doxycycline is frequently used for the treatment of various infections by gram-positive and gram-negative bacteria, aerobes and anaerobes and also for other types of bacteria [Grant et al., 2020]. For the treatment of malaria, rheumatoid arthritis, chronic discoid lupus erythematosus [Rainsford, et al., 2015] Hydroxychloroquine is being actively used. For COVID-19 treatment some individual antiviral and antibiotic drugs have already shown lower to moderate effectiveness when tested against infections in patients. Some of them are: Chloroquine, Hydroxychloroquine, Nafamostat, Nitazoxanide, Remdesivir, Ribavirin, Penciclovir, Ritonavir, AAK1, Baricitinib, Choline and Arbidol [Muralidharana, et al., 2020;Chowdhury et al., 2020].
Another effective way of treatment which clinicians are using to fight against COVID-19 is combination drug therapy where two or more different known antiparasitic drugs, immunomodulators or natural remedies are used to treat COVID-19 patients. Combination drugs have already proved their effectiveness to combat against HIV and other coronaviruses earlier Ter Meulen, et al;Khan, et al., 2020]. Combination drugs have shown their efficiency for the treatment of MERS-CoV infection [Falzarano, et al., 2013]. Lopinavir and Ritonavir combination is one of the mostly tested repurposed combination medicine for the treatment of MERS-CoV infection [Chan, et al., 2015]. Another combination of drug of Mycophenolic acid (MPA) and IFN- has proved its strong effectiveness against MERS-CoV affected diseases . In most of the above-mentioned cases, application of combination drugs has been proved to be success as most of the patients were declared clinically recovered [Shalhoub, et al., 2015]. For COVID-19 treatment antibiotic and antiviral drug combination of Hydroxychloroquine and Favipiravir is already being tested [Costanzo et al., 2020].
Similarly, the combination of Hydroxychloroquine + Azithromycin [Gautret, et al., 2020], Favipiravir+ Nafamostat mesylate [Doi et, al. 2020] or Lopinavir+ Oseltamivir + Ritonavir [Muralidharana, et al., 2020] are already in use against SARS-CoV-2 infection. In the present study, we have tried to understand the mechanism of action of some mostly applicable proposed antiviral and antibiotic drug combination for COVID-19. Our proposed combination is Favipiravir +Hydroxychloroquine +Oseltamivir. Favipiravir, a synthetic prodrug has already tested for its antiviral activity against the influenza virus [Furuta, et al., 2002]. It is already been used for the treatment of Ebola virus, Lassa virus, and now COVID-19 . Hydroxychloroquine is an FDA approved drug which is usually being used for the treatment of malaria, rheumatoid arthritis, chronic discoid lupus erythematosus, and systemic lupus erythematosus [Ben-Zvi, et al., 2012].
Oseltamivir is an antiviral drug used for the treatment and prophylaxis of infection with influenza viruses A and B [Hurt, et al., 2009]. In the present work we have represented the challenges and applicability of repurposing of drugs Favipiravir, Hydroxychloroquine and Oseltamivir in terms of their applicability individually and in combination modes against SARS-CoV-2 infections. To understand the interaction between receptor 3CL pro protease and inhibitor ligand drugs we have applied some simulation techniques like energy minimization, molecular docking and molecular dynamics (MD) simulations.

Procedure of potential target protein structures for SARS-CoV-2
SARS-CoV-2 is a virus having positively sensed single stranded RNA. The protein structure of CoV-2 contains spike (S) , membrane (M), envelope (E) and nucleocapsid (N) [Woo, et al., 2005;Gorden, et al., 2020]. The structure of CoV-2 virus has been characterized very rapidly after its appearance to the world since the genomic sequence of the virus is already known to the research world [Wu, et al., 2020]. The structures of newly invented CoV-2 virus and already known CoV virus is very similar [Yin, et al., 2020]. The similarity index is about 95%. So the identification process of 3chymotrypsinlike viral protease (3CL pro ) of CoV-2 was appeared to be much faster .
For quick drug discovery the similarity index plays a major role. To generate the non-structural proteins the 3CL pro split the poly-protein at 11 distinct sites. The process play an significant role in the way of viral replication. 3CL pro is located at the 3 ends, which exhibits excessive variability. Therefore, 3CL pro is a potential target or anti-coronaviruses inhibitors screening [Deng, et al., 2014].
In the present study, we have used 3CL pro proteases(6LU7) as main target protein of drug molecules. 3D structure of the 6LU7 of CoV-2 was retrieved from the Protein Data Bank website (https://www.rcsb.org) [Burley et al., 2019]

Procedure of Ligand drug molecules preparations
There are many drugs available in medical industry, which have several protein targets. Similarly, several diseases are there which share overlapping molecular pathways. So, the concept of drug repurposing has become a very familiar treatment protocol for such type of diseases. In this study, we have used well known and FDA approved drugs Favipiravir (C5H4FN3O2), Hydroxychloroquine (C18H26ClN3O), Oseltamivir (C16H28N2O4). Favipiravir (T-705) is a synthetic prodrug, which is used in the antiviral activity of chemical agents against the influenza virus [Furuta, et al., 2002]. It has a good bioavailability (∼94%), 54% protein binding affinity. Favipiravir-RTP binds to inhibits RNA dependent RNA polymerase (RdRp), which ultimately prevents viral transcription and replication [Furuta, et al., 2017]. Hydroxychloroquine is used in the treatment of malaria, systemic lupus erythematosus [Ben-Zvi, et al., 2012]. The bioavailability of hydroxychloroquine is 67-74%. According to available data the Hydroxychloroquine rises the pH value in human organelles [D'Acquarica, et al., 2020]. The raised pH in human organelles can prevent virus particles (such as SARS-CoV and SARS-CoV-2) from utilizing their activity for fusion and entry into the cell. Oseltamivir is an antiviral drug, which is used for the treatment and prophylaxis of infection with influenza viruses A and B [Hurt, et al., 2009]. Oseltamivir exerts its antiviral activity by inhibiting the activity of the viral neuraminidase enzyme found on the surface of the virus, which prevents budding from the host cell, viral replication, and infectivity [Muralidharana, et al., 2020]. Virtual screening of these drugs have done before the checking of inhibition capability of the F, H and O drugs. SWISS ADME software (https://www.swissadme.ch) and ADMET (https://vnnadmet.bhsai.org/) software were used for the virtual screening of above mentioned drugs [Daina, et al., 2017].
For virtual screening of different available drugs there are some Drug-likeness rules like Lipinski's rule of five (Ro5), Veber's rule, MDDR-like rule, Egan rule, Ghose filter, Muegge rule etc., which are used for preliminary drug screening [Lipinski, 2004, Apparsundaram, et. al., 2000, Veber, et. al., 2002. Among all of them Ro5 is the most important rule and also known as ''a rule of thumbˮ. The filters of Ro5 are followed by: Molecular weight less than equals to 500, H-bond donors less than equals to 5, H-bond acceptor less than equals to 10, MLOGP less than equals to 4.15 and molar refractivity between 30 and 140. Those drugs which follows the Ro5 with some required pharmacological properties can be used as a potential candidate for vocally active drug in humans [Lipinski, 2004]. For Drug preparation, the ligands in 'SDF' format were obtained directly from the PubChem (National Library of Medicine) (https://pubchem.ncbi.nlm.nih.gov/) and converted to 'PDB' format with the help of Auto Dock tools [Morris et al., 2009]. After that, all molecular structures of the ligands were optimized by using density functional theory (DFT) with the basis set 6.311G (d,p) [Becke, 1997] using the Gaussian 09 program [Frisch, 2004]. The optimized structure was visualizing with the help of Gauss View 5 molecular visualization program [Dennington, et al., 2007].

Molecular docking and visualization
To examine the protein-ligand interaction we have used molecular docking using AutoDock 4.2 and AutoDock vina [Morris, et al., 2009]. MGL tools were used for the preparation of protein and ligands for molecular docking. Molecular docking is the computation method which is used to perform the binding energy calculation and energy minimization of the protein-ligand complex. With the help of molecular docking we can identify that the ligand is at their energy minimized state or not with some scoring function. From docking score we can also predict that the ligand is properly docked with host protein (6LU7) or not [Yuriev, et al., 2011].  , 2017]. After that the MD simulation is used for the already docked structure [Berendsen, et al., 1995].

Molecular Dynamic (MD) Simulations
From MD simulation results, we can find out the structural dynamics of protein: ligand interaction at an atomistic level. In MD simulation for calculating thermodynamics parameters of the ligand: protein complex structure the LINUX based platform ''GROMACS 5.1 Package ˮ [Berendsen et al., 1995] with GROMOS43A2 force fields [Gunsteren, et al., 1996] was used. The thermodynamic parameters computed were : Epot, T, D, Rg, RMSD, RMSF, SASA, H-bonds and interaction energies. Firstly, all three ligand structures (F, H, O) were optimized by using DFT with the basis set 6.311G (d,p) [Becke, 1997] by using Gaussian 09 package [Frisch, 2004]. All ligands should be at their optimized minimum energy states for molecular dynamics simulation so energy minimization was the first step for MD simulation. For energy minimization of complex, time varying (1ps-100ps) steepest descent algorithm with for 500,00 steps was used. The algorithm has a cut off up to 1000 KJ/mol for minimizing the steric clashes [Anuj, et al., 2020]. There are two phases to obtain the energy minimization of complex with each having 500,000 steps. In the first phase, minimization was obtained through a binary condition

Computational details
MD simulations and corresponding energy calculations have been computed using HP Intel Core i5 -1035G1 CPU and 8 GB of RAM with Intel UHD Graphics and a 512 GB SSD. proposed as chemical compound which have certain compulsory pharmacological properties which can make the drug as a potential candidate for orally active drug in humans [Lipinski, 2004]. Also RO5 helps to avoid preclinical and clinical failures by primary preclinical progress of drugs by differentiating between nondrug and drug like molecules. From the ADMET analysis we can conclude that none of our tested drug has any cyto-toxicity effect. According to data calculated from ADMET software   (Table 3 and SD1).  receptor protein and ligand form the complex then they shows the less value of dreiding energy than the individual value. Lowest dreiding energy means the most favorable structure of protein: ligand complex structure [Stephen, et al., 1990].

MD Simulation analysis
According to the procedure followed for MD simulation, TIP3P water model has been used and 4Na + ions were added to maintain the neutrality of ligand: protein complex structure. To calculate the interaction free energies for the protein: ligand complex structure (ΔGbind), the MMPBSA (Molecular Mechanics Poisson-Boltzmann Surface Area) method [Rashmi, et al., 2014] sourced from the APBS and GROMACS packages have been used. This model has both repulsive and attractive components [Kollman, et al., 2000]. To predict calculate the binding energy the data was collected at every 100ps between 0ps and 10000ps. In aqueous environment, the binding free energy of the ligand: receptor complex can be shown by following equations. To obtain the data at atomistic level MD simulation is used. It can simulate the data in picosecond (ps)/nanosecond(ns) and so it is the best verified silico method among the researchers [McDowell, et al., 2007;Benson & Daggett, 2012].  After the implementation of NVT and NPT conditions the each complex system was completely stable under equilibrium conditions. So system was ready for equilibration process. After running the equilibration process in water medium the MD simulations were operated for complete time trajectory 0-10000ps. Throughout the MD simulation process many thermodynamic parameters for the complex system were computed. These thermodynamic parameters of the complex system are used to know the stability and to observe possible configuration changes of receptor protein in its bare state and complex structure at time resolved simulation trajectory. Some important thermodynamic parameters are RMSD, RMSF, , potential energy, Rg, inter-molecular H-bonds, SASA various non-bonded interaction energies for protein: ligand complex structures. Output for all thermodynamic parameters for protein in its bare state and with all complex formations with possible ligands are shown in Table 4, SD 3.   Figure 6b, SD 4). From the graph, it is observed for the complex (F+H+O: 6LU7) structure the value of RMSD showed a variation between 0.14-0.31 (0.01) nm compared to receptor protein 6LU7 variation 0.13-0.32, which indicates that possibility of less fluctuation when the ligands are binding with protein (Table 4). Also the result observed for the complex structures (F: 6LU7, H: 6LU7, O: 6LU7, F+H: 6LU7) showed a variation between 0.10-0.40, 0.23-0.24, 0.09-0.48, 0.12-0.29 nm compared to receptor protein 6LU7 variation 0.13-0.32 again indicates that the less fluctuation in every complex structure during binding with receptor protein (Table 4, SD 3).     6LU7 was obtained as -214.372±47.627 kJ/mol (Table 4)  In the present study the MD simulation results validated that repurposing of combination drug F+H+O can make an remarkably stable complex with SARS-CoV-2 protein (6LU7) after binding to the active sites of this protein.

Conclusion
Repurposing of combination drugs have already proved their effectiveness to combat against many viral diseases like HIV, Ebola and against other coronaviruses earlier. For the present study we have tried to recognize the mechanism of action of some combination drugs by repurposing some common antiviral and antibiotic drugs as Favipiravir, Hydroxychloroquine and Oseltamivir against COVID-19. The properties like physiochemical , medical chemistry and pharmacokinetics from ADME 6LU7 (0.22nm) and host protein (0.22nm) confirmed the total inheritance of proposed combination drug inside host protein. Above result has been further verified by the RMSF variation data which clearly showed that F+H+O: 6LU7 complex structure does not affected the protein backbone. Lowest SASA energy for F+H+O: 6LU7 complex (7.5 nm 2 ) also validated the best stability of F+H+O combination.
The stability of F+H+O combination as complex with 6LU7 was further verified by the lowest value of ΔGbind (-214.372±47.627 kJ/mol) compared to all other combination and individual forms. By analyzing all the in-silico results obtained from molecular docking and MD simulations we can definitely conclude that repurposing of our proposed antiviral antibiotic drug combination: Favipiravir + Hydroxychloroquine + Oseltamivir has established its strong candidature to be used as a promising potential inhibitor for targeting SARS-CoV-2 virus. We have recommended that our In-Silico results have the strong candidature of combination drugs Favipiravir, Hydroxychloroquine and Oseltamivir as a potential lead inhibitor for targeting SARS-CoV-2 infections.