CORD-19:06a1002f9fbea7179ac3572843f66b14568af6e4 JSONTXT 8 Projects

Nelfinavir was predicted to be a potential inhibitor of 2019-nCov main protease by an integrative approach combining homology modelling, molecular docking and binding free energy calculation Abstract 2019-nCov has caused more than 80 deaths as of 27 January 2020 in China, and infection cases have been reported in more than 10 countries. However, there is no approved drug to treat the disease. 2019-nCov M pro is a potential drug target to combat the virus. We built homology models based on SARS M pro structures, and docked 1903 small molecule drugs to the models. Based on the docking score and the 3D similarity of the binding mode to the known M pro ligands, 4 drugs were selected for binding free energy calculations. Both MM/GBSA and SIE methods voted for nelfinavir, with the binding free energy of -24.69±0.52 kcal/mol and -9.42±0.04 kcal/mol, respectively. Therefore, we suggested that nelfinavir might be a potential inhibitor against 2019-nCov M pro . author/funder. All rights reserved. No reuse allowed without permission. In December 2019, cluster of patients with pneumonia were reported in Wuhan, Hubei Province, China. 1, 2 Shortly, a new coronavirus, temporally named 2019-nCov, was identified to be the cause of the disease, which is the seventh member of the family betacoronavirus. 1 More than 2,700 infection cases and 80 deaths were reported as of 27 January 2020 from China. In addition, infection cases have been reported from Thailand, Australia, Malaysia, Singapore, France, Japan, South Korea, the United States, Vietnam, Canada and Nepal, indicating that the disease is a potential threat to the global health. 3 Sadly, the number is still growing rapidly and no drug has been approved to be effective. Therefore, it is urgent to discover and develop drugs to cure the disease. Based on its function, the main protease (M pro ) or chymotrypsin-like protease (3CL pro ) 4 is suggested to be a potential drug target to combat 2019-nCov, which is highly conservable among coronaviruses. Sequence alignment revealed that the M pro of 2019-nCov shares 96% similarity with that of SARS (severe acute respiratory syndrome) (Figure 1 ). Studies for identifying the inhibitors of 2019-nCov M pro were quickly performed for discovering and developing drugs against the disease. For example, Hualiang Jiang and collaborators identified 30 drugs and compounds as the M pro inhibitors via protein modelling and virtual screening, which is a rapid progress in the way to cope with the crisis. In addition, one of the 30 drugs/compounds, remdesivir, was also suggested to be potential inhibitor against 2019-nCov by Liu et al, 5 For finding more potential drugs as inhibitors of the protein, we modelled the 2019-nCov M pro structures using SARS M pro (PDB ID: 2GTB) as template and docked 1903 approved drugs against the model in this study. Fifteen drugs were selected based on the docking score and three dimensional (3D) similarity to available M pro inhibitors against other coronavirus. For validation, we modelled 10 additional new models of 2019-nCov M pro and docked the 15 drugs against the new models, which revealed that 6 drugs (nelfinavir, praziquantel, pitavastatin, perampanel, eszopiclone, and zopiclone) have good binding modes with the new models. Binding free energy calculation were then performed for 4 of the 6 drugs using MM/GBSA and SIE methods. Nelfinavir was identified as the best one with predicted binding free energies of -24.69±0.52 kcal/mol by MM/GBSA and -9.42±0.04 kcal/mol by SIE, respectively. Taking into account its high 3D similarity of binding mode to the known M pro inhibitor, we proposed that nelfinavir should be a potential inhibitor against 2019-nCov M pro . 2.1 Homology modelling. 43 M pro complexes with ligands were downloaded from protein data bank 6 (PDB IDs: 1WOF, 2A5I, 2A5K, 2ALV, 2AMD, 2GTB, 2GX4, 2GZ7, 2GZ8, 2OP9, 2QIQ, 2V6N, 2ZU4, 2ZU5, 3SN8, 3SND, 3SZN, 3TIT, 3TIU, 3TNS, 3TNT, 3V3M, 4F49, 4MDS, 4TWW, 4TWY, 4WY3, 4YLU, 4YOG, 4YOI, 4YOJ, 4ZRO, 5C5N, 5C5O, 5EU8, 5N5O, 5N19, 5NH0, 5WKJ, 5WKK, 5WKL, 5WKM, author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.01.27.921627 doi: bioRxiv preprint 6FV1) and aligned to 2GTB in PyMOL. 7 with improving scoring and minimization. The hydrogens were added to 2019-nCov M pro model by pdb2pqr (--ff=amber --ffout=amber --chain --with-ph=7). 13 Then the model was converted to pdbqt format by prepare_receptor4.py script in MGLToos version 1.5.6. 10 The ligand in 2GTB was used to define the grid and the buffer space was set to 6.0 Å (autobox_add). The random seed was explicitly set to 0 (seed). The exhaustiveness of the global search was set to 32 (exhaustiveness) and at most 1 binding mode was generated for each drug (num_modes). MolShaCS, which utilized Gaussianbased description of molecular shape and charge distribution, was used to calculate the 3D similarities between approved drugs and available M pro inhibitors. 14 2.4 Molecular dynamics simulation. Each simulation system was immersed in a cubic box of TIP3P water that was extended by 9 Å from the solute, with a rational number of counter ions of Na + or Clto neutralize the system. General Amber force field (GAFF) 15 and Amber ff03 force field 16 were used to parameterize the ligand and protein, respectively. 10,000 steps of minimization with constraints (10 kcal/mol/Å 2 ) 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 0.1 ns equilibration in NPT ensemble. Finally, 5 ns MD simulation on each system at 300 K was performed. The minimization, heating and equilibrium are performed with sander program in Amber16. The 5 ns production run was performed with pmemd.cuda. Based on the 5 ns MD simulation trajectory, bind author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.01.27.921627 doi: bioRxiv preprint free energy (ΔG) was calculated with MM/GBSA 17, 18 and SIE 19 approaches. In the MM/GBSA, the ΔG was calculated according to equation (1), where ΔEele and ΔEVDW refer to electrostatic and van der Waals energy terms, respectively. ΔGgb and ΔGnp refer to polar and non-polar solvation free energies, respectively. Conformational entropy (TΔS) was calculated by nmode module in Amber16. The dielectric constants for solvent and solute were set to 80.0 and 1.0, respectively, and OBC solvation model (igb = 5 and PBradii = 5) 20 was used in this study. Other parameters are set to default values. In the SIE, the ΔG was calculated based on equation (2), where EC and EVDW refer to the sum of intermolecular Coulomb and van der Waals author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not peer-reviewed) is the . After visualizing the docked complexes carefully, we selected 15 drugs (Table 2) for further analysis. There are some differences between the conformations of the protein in the 3D similarity reference and 2GTB model. Therefore, we modelled 10 new homology models using the proteins in 3D similarity reference as templates and we redocked the 15 drugs to the 10 new homology models. 6 drugs (nelfinavir, pitavastatin, perampanel, praziquantel, zopiclone, and eszopiclone) show good docking scores and binding modes (Table 3) . Because eszopiclone and zopiclone was used to treat insomnia in a low dosage which may not be suitable to treat pneumonia, we carried out further binding free energy calculation for the rest 4 drugs. author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.01.27.921627 doi: bioRxiv preprint praziquantel-3V3M, were subjected to 5 ns molecular dynamics simulations using author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.01.27.921627 doi: bioRxiv preprint Amber 16. To provide insight into their binding mechanisms, the binding free energies were calculated by MM/GBSA and SIE approaches. In results of the MM/GBSA approach, the calculated binding free energies of nelfinavir-2GX4, pitavastatin-2GTB, perampanel-5NH0, and praziquantel-3V3M, are -24.69±0.52, -12.70±0.38, -14.98±0.34, and -6.51±0.21 kcal/mol, respectively (Table 4) , which highlight nelfinavir as the most active one. In pitavastatin-2GTB, perampanel-5NH0, and praziquantel-3V3M, the van der Waals interaction (E vdw ) makes a more significant contribution than the electrostatic interaction (E ele ) ( As shown in Figure 2 , the binding model of nelfinavir in its docking complex turned out to be very similar with that of the original ligand (TG-0205221) 25 author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not peer-reviewed) is the . 2019-nCov caused more than 80 deaths in China as of 27 January 2020 and is a potential threat to the global health. However, there is no approved drug to treat the disease. 2019-nCov M pro is a potential drug target to combat the virus, which shares 96% sequence similarity with the corresponding one in SARS. We built 11 homology models of 2019-nCov M pro and docked 1903 approved small molecule drugs to the 2GTB model. Based on the docking score and the 3D similarity of binding mode to 39 known M pro binders, 15 drugs were selected for further evaluation. The 15 drugs were then docked to all the 11 homology models, leading to 4 drugs for binding energy calculations. Both MM/GBSA and SIE calculations voted for nelfinavir, a HIV-1 protease inhibitor to treat HIV. Therefore, we suggested that nelfinavir might be active against 2019-nCov M pro . In addition, pitavastatin, perampanel, and praziquantel might author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.01.27.921627 doi: bioRxiv preprint also have moderate activities against 2019-nCov.