Potential inhibitors targeting RNA-dependent RNA polymerase activity (NSP12) of SARS-CoV-2

A novel coronavirus (SARS-CoV-2) that is initially found to trigger human severe respiratory illness in Wuhan City of China in 2019, has killed 2,718 people in China by February 26, 2020, and which has been recognized as a public health emergency of international concern as well. And the virus has spread to more than 38 countries around the world. However, the drug has not yet been officially licensed or approved to treat SARS-Cov-2 infection. NSP12-NSP7-NSP8 complex of SARS-CoV-2, essential for viral replication and transcription, is generally regarded as a potential target to fight against the virus. According to the NSP12-NSP7-NSP8 complex (PDB ID: 6NUR) structure of SARS, two homologous models were established for virtual screening in present study, namely NSP12-NSP7 interface model and NSP12-NSP8 interface model. Seven compounds (Saquinavir, Tipranavir, Lonafarnib, Tegobuvir, Olysio, Filibuvir and Cepharanthine) were selected for binding free energy calculations based on virtual screening and docking scores. All the seven compounds can combine well with NSP12NSP7-NSP8 in the homologous model, providing drug candidates for the treatment and prevention of SARS-CoV-2.


Introduction
Since the outbreak of pneumonia caused by a novel coronavirus  in Wuhan, China in December 2019, the world have faced unprecedented challenges in treating the disease caused by this virus [1][2][3]. China has implemented strict epidemic prevention and control measures across the country to avoid larger scale epidemic [4].
However, as of February 26, 2020, there were 78,195 confirmed cases and more than 2,718 deaths. The virus now presents in more than 30 countries. The infectious power and harmful effects of SARS-CoV-2 should not be underestimated. To date, there are no officially licensed or approved drugs against this novel coronavirus. There is an urgent need to find new targets for development of anti-SARS-CoV-2 agents.
ORF1a and ORF1b at the 5'-terminus of the Coronavirus (CoV) genome encode polyprotein 1a and polyprotein 1b, the two proteins could be cleaved into 16 nonstructural proteins (NSPs), which are essential for viral replication and transcription, thus being regarded as a potential virulence factor and a target for CoV [5,6]. Among these NSPs , the NSP12 subunit is the essential RdRp (rna-dependent RNA polymerase) of the coronavirus replicative machinery, which was even able to extend a homopolymeric primer-template substrate by a few dozen nucleotides in vitro [7].
The 3.1Å cryo-EM structure of the SARS-CoV RNA polymerase NSP12 shows it can bind with its essential co-factors NSP7 and NSP8 [8]. The replication of the SARScoronavirus genome involves two RNA-dependent RdRps. The first is primerdependent and associated with the NSP12, whereas the second is catalyzed by NSP8.
NSP8 is capable of de novo initiating replication process and has been proposed to operate as a primase [9]. In addition, NSP7, a component of the CoV replicase polyprotein, also participates in viral replication processed by binding to NSP12 as another primase [9]. The NSP12 needs to associate with NSP7 and NSP8 to activate its capability to replicate long RNA [7]. This elicit us to identify the particularly interesting compound disrupt the binding of NSP7 or NSP8 to NSP12, thus which could be used to inhibit the RdRp activity of NSP12, acting as novel antiviral agents and therapies of SARS-Cov-2.
The amino acid sequence alignment revealed that the NSP12 of SARS-CoV-2 shared 96.35% similarity with the NSP12 (PDB ID: 6NUR) of SARS ( Figure 1). In addition, comparative analyses of their deduced amino acid sequences revealed that NSP7 and NSP8 of SARS-CoV-2 shared 98.8% and 97.5% similarity with that of SARS-CoV respectively. Therefore, the SARS-CoV-2 NSP12 structures could be constructed by the performance of homology modeling using SARS NSP12 (PDB ID: 6NUR) as template. The NSP7 and NSP8 binding pocket of NSP12 were designated as active sites for screening compounds. Here, through high-throughput screening methods using a pool of 30,000 small molecules, several potential drug candidates were identified for preventing the binding NSP7 or NSP8 to NSP12, suggesting further assessment of the anti-SARS-CoV-2 activity of these compounds in cell culture.
Then the model was converted to pdbqt format by prepare_receptor4.py script with assigning atomic types and atomic charges. All rotatable bonds in the molecule are set to be flexible for flexible docking.

Preparation of target proteins
Homology model of the target protein (NSP12, NSP7 and NSP8 of SARS-CoV-2) was built by modeller 9.18 using crystal structure of SARS NSP12 (PDB ID: 6NUR) as template. 100 independent structures were constructed and the one with best DOPE score was chosen for further energy minimization by Amber. The relaxed model was saved as pdb file, and which was converted to pdbqt format as docking receptor using AutoDockTools-1.5.6, with assigned atom type and partial charge.

Molecular docking
Vina1.1.2 was used to perform molecular docking. The docking boxes were set at the NSP12-NSP7 interface and NSP12-NSP8 interface, respectively. The search exhaustiveness was set to be 32, and the number of binding modes was set to be 9.
Other parameters were set as default. During docking, NSP7 (or NSP8) was removed from the complex and only NSP12 was left as receptor.

Binding free energy calculation
Each simulation system was immersed in a cubic box of TIP3P water with 10 Å distance from the solute. The Na + or Clwas applied to neutralize the system. General Amber force field (GAFF) 15 and Amber ff14SB force field were used to parameterize the ligand and protein respectively. 10,000 steps of minimization with constraints (10 kcal/mol/Å2) on heavy atoms of complex, including 5,000 steps of steepest descent minimization and 5,000 steps of conjugate gradient minimization, was used to optimize each system. Then each system was heated to 300 K within 0.2 ns followed by 0. Lonafarnib were all closely associated with antiviral activity.  Table 2 Eleven compounds selected from the NSP12-NSP8 interface of homology model

NSP7 model
Saquinavir, the first HIV protease inhibitor was introduced into the market in 1995.
Tipranavir, a novel non-peptide protease inhibitor approved for use in patients with resistant strains of HIV. Both of which are safe and generally well-tolerated in HIV-1infected adults [10,11]. Our docking results showed that five of the hydrogen bonds involving GLY-297, PHE-299, PHE-325, PHE-326 and ALA-327 maintained upon the binding of saquinavir with interface between SARS-CoV-2 NSP12 and NSP7 ( Figure   1A). As for Tipranavir, hydrogen bonds involving PHE-325 maintained upon the binding of Tipranavir with interface between SARS-CoV-2 NSP12 and NSP7 ( Figure   1C). Saquinavir and Tipranavir could bind to the interface active pockets of the SARS-CoV-2 NSP12 and NSP7 ( Figure 1B, D). The previous study showed that saquinavir could bind to the SARS-CoV-2 RNA-dependent RNA polymerase and inhibit the enzyme activity [12]. Our observations further confirm that saquinavir and tipranavir can bind to the NSP12-NSP7 interface as interfacial blockers, thus making them as candidates for further in vitro evaluation of anti-SARS-CoV-2 activity.

Docking results of Lonafarnib against SARS-CoV-2 NSP12-NSP7 model
Lonafarnib as a non-peptidomimetic inhibitor of farnesyltransferase has been used for progeria [13,14]. Our docking results showed that lonafarnib was mainly combined with the interface between SARS-CoV-2 NSP12 and NSP7 through van der Waals potential energy and hydrophobic accumulation, involving PHE-299, PHE-727, PHE-324, PHE-325 and PHE-326 ( Figure 2A). Lonafarnib could bind to the interface active pockets between the SARS-CoV-2 NSP12 and NSP7 ( Figure 2B). Therefore, we speculate that Lonafarnib has potential activity for the treatment of SARS-Cov-2 infection.

Docking results of Olysio against SARS-CoV-2 NSP12-NSP8 model
Olysio, a hepatitis c virus (HCV) NS3/4A protease inhibitor approved for the treatment of genotype 1 chronic hepatitis C in combination with pegylated interferon and ribavirin [17]. Our docking results showed that the hydrogen bonds involving VAL-214 and van der Waals forces maintained upon the binding of Olysio and SARS-CoV-2 NSP12-NSP8 interface ( Figure 4A). Olysio could also bind to the interface active pockets of the SARS-CoV-2 NSP12-NSP8 ( Figure 4B). Thus, based on the present results, Olysio may be considered as a candidate for further in vitro evaluation of anti-SARS-CoV-2 activity.

Docking results of Cepharanthine against SARS-CoV-2 NSP12-NSP8 model
Cepharanthine, an alkaloid tetrandrine isolated from Stephania tetrandra was found to exert strong anti-cancer, anti-inflammatory and antioxidant activities [18]. In addition, it shows in vitro inhibitory effect on Herpes simplex virus type 1 (HSV-1) infected cells [19]. Our docking results showed that the hydrogen bonds involving ARG-215 maintained upon the binding of Cepharanthine and SARS-CoV-2 NSP12-NSP8 interface, with additionally van der Waals forces ( Figure 5A). Cepharanthine could bind to the interface active pockets of the SARS-CoV-2 NSP12-NSP8 ( Figure  5B). The previous study showed Cepharanthine could significantly inhibit the replication of human coronavirus strains OC43 [18]. Taken together, Cepharanthine coule be a potential natural antiviral compound for the prevention and treatment of SARS-CoV-2 infection.

Docking results of Filibuvir against SARS-CoV-2 NSP12-NSP8 model
Filibuvir is an effective oral non-nucleoside HCV NS5B polymerase inhibitor for the potential treatment of chronic HCV infection [20]. Studies have shown that Filibuvir was well tolerated and could be considered in combination with other antiviral drugs to achieve better safety and efficacy for chronic HCV [21]. Our docking results showed that Filibuvir was mainly combined with the SARS-CoV-2 NSP12-NSP8 interface through van der Waals potential energy ( Figure 6A). Filibuvir could bind to the interface active pockets of the SARS-CoV-2 NSP12-NSP8 ( Figure 6B). Thus, Filibuvir can be considered as a candidate drug for treating SARS-CoV-2 infection, providing evidence for further research.