Virtual screening workflow
In this in silico drug repurposing study, we identified potential inhibitors of TMPK and D9 decapping enzyme of monkeypox virus using the virtual screening workflow shown in Fig. 2. First, we prepared a library of 202 US FDA approved drugs which are either antivirals or antibiotics (Methods; Supplementary Table S1). Second, the three-dimensional (3D) structures of the TMPK and D9 decapping enzyme in monkeypox virus were constructed using homology modelling based on published crystal structures (PDB 2V54, 7SEZ) of the corresponding proteins in vaccinia virus (Methods; Fig. 1).
Third, to assess the stability of the modelled protein structures, the free protein structures for TMPK and D9 decapping enzyme were subjected to MD simulations of 100 ns in triplicate (Methods). For TMPK and D9 decapping enzyme, the root mean square deviation (RMSD) of the Cα atoms show fewer fluctuations after 20 ns (Fig. 3a,d). The root mean square fluctuation (RMSF) shows the dynamics of the amino acid residues in a protein. For TMPK and D9 decapping enzyme, the RMSF of the amino acid residues show fluctuations primarily in the loop region indicating that the secondary structures of the two proteins are stable (Fig. 3b,e). The radius of gyration (Rg) of a protein in the MD trajectory shows the compactness of the protein. For TMPK and D9 decapping enzyme, the Rg show little variation indicating that the two protein structures are compact during the MD simulation (Fig. 3c,f). Subsequently, we used the stable structures of TMPK and D9 decapping enzyme at the end of their MD simulation trajectories at 100 ns for molecular docking.
Fourth, we performed molecular docking of the 202 US FDA approved drugs in the compiled ligand library of antivirals or antibiotics against the stable protein structures of TMPK and D9 decapping enzyme (Methods). The binding site residues important for the activity of TMPK in monkeypox were determined by comparing the modelled protein structure with the crystallized protein structure of TMPK from the vaccinia virus (PDB 2V54). TMPK belongs to the Nucleotide monophosphate kinase (NMP) family and contains 9 important binding site residues (Fig. 1a). TMPK binds to Thymidine diphosphate (TDP) at its NMP binding site. In particular, the base binds to Arg72 (R72) and Phe68 (F68), and the sugar group is bound by Tyr101 (Y101), Leu53 (L53) and Asp13 (D13), and the phosphate group is bound by Arg93 (R93), Arg41 (R41) and Lys17 (K17). Asn65 (N65) is present in the binding pocket and is also important for the interaction [38]. The binding site residues important for the decapping activity of the D9 decapping enzyme in monkeypox were determined by comparing the modelled protein structure with the crystallized protein structure of D9 decapping enzyme from the vaccinia virus (PDB 7SEZ). The D9 decapping enzyme contains 5 important residues in the m7GDP binding pocket (Fig. 1b). The m7GDP is sandwiched between the two aromatic residues Phe54 (F54) and Tyr158 (Y158) which are important for its recognition. The Asp151 (D151) and Glu16 (E16) interact with the guanine base of m7GDP. The phosphate chain of the m7GDP is stabilized by interactions with Arg50 (R50) [44].
Fifth, considering these important binding site residues, an interaction cutoff of 4 or more non-covalent interactions between ligand and important binding sites in the target protein in the best docked pose was used to filter top hits. After imposing the interaction cutoff, the top 2 drugs were selected based on the lowest binding (docking) energy. For TMPK, the drugs Tipranavir (T1) and Cefiderocol (T2) were the top hits, and for D9 decapping enzyme, the drugs Tipranavir (D1), Doxorubicin (D2) and Dolutegravir (D3) were the top hits (Fig. 4; Table 1). Sixth, to examine the stability of the protein-ligand complexes for the top hits, we performed MD simulations of the docked complexes for both target proteins. The free energy of the protein-ligand complexes was computed using g_mmpbsa [61, 62].
Table 1
Binding energy of the top potential inhibitors of the two target proteins, TMPK and D9 decapping enzyme, of monkeypox virus. For each ligand or drug, the table provides the target protein, drug identifier, drug name, docking based binding energy and MM-PBSA based binding energy in kcal mol− 1.
Target protein
|
Drug identifier
|
Drug name
|
Docking based binding energy (kcal mol− 1)
|
MM-PBSA based binding energy (kcal mol− 1)
|
TMPK
|
T1
|
Tipranavir
|
-7.6
|
-9.04 ± 6.37
|
T2
|
Cefiderocol
|
-7.6
|
-18.74 ± 4.53
|
D9
|
D1
|
Tipranavir
|
-11
|
-26.65 ± 4.29
|
D2
|
Doxorubicin
|
-9.9
|
-32.57 ± 4.17
|
D3
|
Doultegravir
|
-9.9
|
-9.87 ± 3.98
|
Top potential inhibitors of TMPK and D9 decapping enzyme of monkeypox virus among approved drugs
Supplementary Table S1 gives the drug name, drug identifier and therapeutic class, of the 202 US FDA approved drugs, which are either antivirals or antibiotics, and are part of the ligand library in this virtual screening study. The binding site residues and residues involved in non-covalent interactions in the best docked poses for the top 2 inhibitors for TMPK and D9 decapping enzyme are given in Table 2. In case of TMPK, the top 2 predicted inhibitors are Tipranavir and Cefiderocol, and in case of D9 decapping enzyme, the top 2 predicted inhibitors are Tipranavir, Doxorubicin and Dolutegravir. Since Doxorubicin and Dolutegravir have the same docking based binding energy for the target protein D9 decapping enzyme, both drugs are considered in the list of top 2 predicted inhibitors for the protein (Table 1). A three-dimensional (3D) and two-dimensional (2D) visualization of the non-covalent interactions in the best docked pose for these top inhibitors with residues of the drug target proteins, TMPK and D9 decapping enzyme, are shown in Figs. 5 and 6, respectively.
Table 2
Non-covalent interactions between ligand and protein residues in the best docked poses of top potential inhibitors for TMPK and D9 decapping enzyme of monkeypox virus. For each protein-ligand complex, the table gives the residues in the ligand binding site, residues forming hydrogen bond interactions, hydrophobic interactions, halogen interactions and aromatic interactions with the ligand.
Drug identifier
|
Drug name
|
Binding site residues
|
Hydrogen bond interaction
residues
|
Hydrophobic
interaction residues
|
Halogen interaction residues
|
Aromatic interaction residues
|
T1
|
Tipranavir
|
D50, D92, E141, E142, I49, K17, L53, N37, P39, Q40, R41, R93, S15, T18, Y35
|
D50, N37, P39, R41
|
D50, I49, L53, N37, Q40, R41, Y35
|
D92, K17, S15, T18
|
-
|
T2
|
Cefiderocol
|
D92, E141, E142, K17, L53, N37, P39, Q40, R41, R93, S15, T18, T54, Y35
|
K17, L53, N37, Q40, R41, S15, T54, Y35
|
E141, L53, Q40, T18, Y35
|
N37
|
-
|
D1
|
Tipranavir
|
A58, C162, D151, E105, E16, F154, F35, F54, G160, I159, K198, L108, L147, L202, Q62, Q63, R15, T149, Y158, Y201
|
C162, I159, K198, R15, T149, Y158, Y201
|
A58, F154, F35, F54, G160, I159, K198, L108, L147, L202, T149, Y158
|
Q62, Q63
|
F35, F54, F154
|
D2
|
Doxorubicin
|
D151, E16, E183, F154, F35, F54, G160, H33, I159, L147, Q62, T149, Y158
|
D151, T149
|
F35, F54, F154, H33, I107, L147, Y158
|
-
|
F54, Y158
|
D3
|
Dolutegravir
|
C162, D151, E16, F154, F35, F54, G160, K198, L147, T149, Y158
|
C162, K198, Q62, T149, Y158
|
F54, F35, Y158
|
L147
|
F35, Y158
|
Drug T1 Tipranavir [63], has a docking binding energy with TMPK of -7.6 kcal mol− 1. It is a nonpeptidic HIV protease inhibitor [64]. The nonpeptidic nature offers molecular flexibility to Tipranavir, and makes it easier to fit into the active site [65] of TMPK. In the best docked pose, Tipranavir binds with the TMPK residues E142, N37, P39, R41, D50, I49, L53, R93, T18, Y35, E141, K17, D92, S15 and Q40. Further, Tipranavir forms hydrogen bonds with the TMPK residues R41, P39, D50 and N37, and trifluoromethyl group of Tipranavir forms halogen bond interactions with the TMPK residues S15, K17, T18 and D92.
Drug T2 Cefiderocol [66], has a docking binding energy with TMPK of -7.6 kcal mol− 1. It is a siderophore cephalosporin antibiotic which has been approved by the US FDA for the treatment of complicated urinary tract infections. Moreover, the drug is also included in the WHO list of essential medicines [67, 68]. Cefiderocol contains a pyrrolidinium group on the C-3 side chain and an aminothiazole group in the C-7 side chain which improve its antibacterial activity [66]. In the best docked pose, Cefiderocol binds with the TMPK residues E142, Q40, Y35, P39, R41, L53, T54, E141, T18, R93, S15, N37, D92 and K17. In the best docked pose, the pyrrolidinium group forms hydrogen bonds with the TMPK residues T54 and L53, the chlorocathecol group forms hydrogen bonds with the TMPK residues S15, K17 and N37, the aminothiazole group forms hydrogen bonds with the TMPK residues Q40 and Y35, and the carboxylic acid group forms hydrogen bond with the TMPK residue N37. Further, a hydrogen bond is also seen between the drug and TMPK residue R41. Moreover, the chlorine in the chlorocatechol group of Cefiderocol forms halogen bond interactions with the TMPK residue N37.
Drug D1 Tipranavir also has a docking binding energy with D9 decapping enzyme of -11 kcal mol− 1. In the best docked pose, Tipranavir binds with the D9 residues R15, F54, Y158, F35, T149, E16, K198, E105, A58, F154, I159, G160, T149 and Y201. In the best docked pose, Tipranavir forms hydrogen bonds with the D9 residues R15, T149, Y158, I159, Y201, K198 and C162. Further, halogen bond interactions are observed between the trifluoromethyl group of Tipranavir and D9 residues Q62 and Q63. In particular, we highlight that Tipranavir is a top hit for both drug target proteins considered in this study.
Drug D2 Doxorubicin [69], also called Adriamycin, was originally isolated from Streptomyces peucetius var. caesius, and has a glycosidic structure [70]. Doxorubicin is an anthracycline antibiotic with antitumour activity, and is used for the treatment of breast cancer, AIDS-related Kaposi’s sarcoma and other cancers [69, 71]. The drug is also included in the WHO list of essential medicines [67]. Doxorubicin has a docking binding energy with D9 decapping enzyme of -9.9 kcal mol− 1. In the best docked pose, Doxorubicin binds with the D9 residues E183, F35, Y158, T149, E16, F54, D151, F154, H33, Q62, L147, I159 and G160. In the best docked pose, the chromophore moiety in Doxorubicin forms hydrogen bonds with the D9 residues T149 and D151.
Drug D3 Dolutegravir [72], has a docking binding energy with D9 decapping enzyme of -9.9 kcal mol− 1. Dolutegravir is a HIV integrase inhibitor [72, 73], which is also included in the WHO list of essential medicines [67]. As per WHO recommendation, Dolutegravir can be used as first and second line treatment for HIV infection across all populations [74]. In the best docked pose, Dolutegravir binds with the D9 residues C162, D151, E16, F154, F35, F54, G160, K198, L147, T149 and Y158. In the best docked pose, the electron withdrawing aromatic ring of Dolutegravir forms hydrogen bonds with D9 residues K198 and C162, and further, the ligand forms hydrogen bonds with D9 residues T149, Y158 and Q62. Fluorine atom of Dolutegravir forms halogen bond interactions with D9 residue L147, and Dolutegravir forms aromatic interactions with D9 residues F35 and Y158.
The above-mentioned 4 US FDA approved drugs which are identified as top potential inhibitors of TMPK and D9 decapping enzyme in monkeypox, have also previously been reported in the literature as promising candidates for repurposing against other viral diseases. Tipranavir, which shows binding with both TMPK and D9 decapping enzyme, has been reported previously as a promising candidate for repurposing against SARS-CoV-2 [75, 76] and flaviviruses like West Nile virus and Zika virus [77]. Cefiderocol has also been shown to be affective against melioidosis [78], ventilator associated bacterial pneumonia, and other Gram-negative bacterial infections [79]. Doxorubicin and Dolutegravir have also been reported previously as a promising candidate for repurposing against SARS-CoV-2 [76, 80]. These studies suggest that the approved drugs predicted as top inhibitors in this study are promising candidates for repurposing against monkeypox virus infections. Lastly, the docking based binding energies for an expanded list of top inhibitors predicted in this study for TMPK and D9 decapping enzyme are given in Supplementary Tables S2 and S3, respectively.
MD based stability analysis of protein-ligand complexes of top inhibitors
We performed 150 ns MD simulations for the protein-ligand docked complexes of the top 2 inhibitors for TMPK and D9 decapping enzyme (Methods). We observed that all the protein-ligand complexes for the top 2 inhibitors for TMPK were stable during these MD simulations. The average RMSD of Cα atoms in TMPK for TMPK-T1 = 1.95 ± 0.39 Å and for TMPK-T2 = 1.62 ± 0.21 Å (Fig. 7a). The RMSF of the amino acid residues in TMPK-T1 and TMPK-T2 complexes closely followed the RMSF values of apo TMPK residues (Figs. 3c and 7b). The Rg of the TMPK shows considerably small deviation in both protein-ligand complexes with 16.68 ± 0.09 Å for TMPK-T1 and 16.57 ± 0.08 Å for TMPK-T2 (Fig. 7c). This suggests that TMPK was stable and compact during the MD simulation. However, we observed some deviation in the stability of the ligand in the binding site of the protein during the MD simulations of the protein-ligand complexes. Both the ligands were stable after 100 ns with RMSD values of ligand heavy atoms for TMPK-T1 = 3.64 ± 0.98 Å, and TMPK-T2 = 7.47 ± 0.54 Å (Fig. 7d).
Similar to the protein-ligand complexes of TMPK, we also analyzed the MD trajectories of the protein-ligand complexes for top inhibitors of D9 decapping enzyme. We observed that all the protein-ligand complexes for the top 2 inhibitors for D9 decapping enzyme were stable during the MD simulation. The average RMSD of Cα atoms in D9 decapping enzyme for D9-D1 = 2.28 ± 0.34 Å, D9-D2 = 2.82 ± 0.44 Å, and D9-D3 = 2.02 ± 0.3 Å (Fig. 8a). The RMSF of the amino acid residues in D9-D1, D9-D2 and D9-D3 complexes closely followed the RMSF values of apo D9 decapping enzyme residues (Figs. 3f and 8b). The Rg of the D9 decapping enzyme shows considerably small deviation in all three protein-ligand complexes with 18.56 ± 0.13 Å for D9-D1, 18.22 ± 0.14 Å for D9-D2 and 18.53 ± 0.14 Å for D9-D3 (Fig. 8c). This suggests that D9 decapping enzyme was stable and compact during the MD simulation. We also observed that the three ligands are stable in the binding site of the D9 decapping enzyme during the MD simulations of the protein-ligand complexes. In particular, the RMSD values of the ligand heavy atoms for D9-D1 = 5.06 ± 0.4 Å, D9-D2 = 2.7 ± 0.62 Å, and D9-D3 = 3.32 ± 0.44 Å (Fig. 8d).
The Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) method is routinely used to better estimate the binding energy of a ligand in a protein-ligand complex. We also computed the MM-PBSA based binding energy of the top 2 inhibitors of TMPK and D9 decapping enzyme in this study (Methods; Table 1). For TMPK, the MM-PBSA based binding energy for Tipranavir (T1) in complex TMPK-T1 = -9.04 ± 6.37 kcal mol− 1, and cefiderocol (T2) in complex TMPK-T2 = -18.74 ± 4.53 kcal mol− 1. Similarly, for D9 decapping enzyme, the MM-PBSA based binding energy for Tipranavir (D1) in complex D9-D1 = -26.65 ± 4.29 kcal mol− 1, Doxorubicin (D2) in complex D9-D2 = -32.57 ± 4.17 kcal mol− 1, and Dolutegravir (D3) in complex D9-D3 = -9.87 ± 3.98 kcal mol− 1.