3.1 Papain-like Proteinase (PLpro, NSP-2-3) (PDB code 6w9c)
PLpro is responsible for cleavaging the N-terminus of the replicase polyprotein and thus production of NSP- 1-3. This way it plays an important role in viral replication. It is also shown to be suppressing type 1 interferon signalling thus coMpromising the host immunity (31). For all these reasons, it is a valuable target for inhibiting viral replication.
We first performed amino acid sequence alignment between the selected new crystal structure (PDB code 6w9c) and its closest match in humans SARS-CoV PLpro (PDB code 3e9s). There was a high sequence identity of about 82% with several conserved regions [Fig 3].
In order to find the docking site, we used the above information of highly conserved sites and an existing PDB crystal structure containing a bound ligand called 5-Amino-2-Methyl-N-[(1r)-1-Naphthalen-1-Ylethyl]benzamide (an engineered hydrolase of class: papain-like protease deubiquitinase inhibitor) (32) locked in with the crystal structure of SARS CoV-1 PLpro (PDB code 3e9s) (33).
We then performed structural residue alignment. The cartesian RMSD value between the new SARS-CoV-2 PLpro and earlier reported SARS-CoV PLpro came out to be 1.48 which shows a close structural similarity between the two structures [Fig 4]. It also implies that the ligand binding site identified in the later can be used for the new crystal structure of SARS-CoV-2 as well. The ligand receptor site was thus selected by analysing this existing crystal structure (A-chain) and its attached ligand. This site was prepared by adding hydrogen ions, electrostatic charges and removing water molecules.
The screening results were plotted with ICM score on x-axis while the mfscore was on y-axis. This provided the best selection of ligands which performed equally well on both scoring systems [Fig 5].
The results are displayed in Table 2 which shows a selection of lowest scoring ligands on ICM and mfscores.
It was found that Theasinensin A which is a polyphenol flavonoid found in black tea (Camellia sinensis) happened to score the lowest. It was found to be surrounded by 9 amino acid residues in the pocket with G266 being the closest with a distance between it and the ligand’s oxygen-8 calculated to be around 2.9 Å. The ligand was also forming at least 1 more hydrogen bond with E161 and several hydrophobic interactions (D164, L162, P248, Y264, N267, G163, K157) [Fig 6].
Curucumin (principal curcuminoid of turmeric), Quercetin (found in red onions and kale), Mitoxantrone (an anthracenedione antineoplastic agent), Amentoflavone (a biflavonoid constituent of a number of plants including Ginkgo biloba) and Colistin (an antibiotic for multi-resistant gram-negative bacteria) were some of the most notable ligands with high affinity towards the receptor pocket.
Table 2. Potential PLpro inhibitors from the curated database of herbal and antiviral drugs.
3.2 ADP Ribose phosphatase macro domain of NSP-3 SARS-CoV-2 (PDB code 6vww)
Coronaviridae (genera Coronavirus and Torovirus) possess one of the largest single-stranded RNA genomes (27 to 31.5 kb) (34). One of the non structural proteins made by polypeptide 1a of SARS-CoV-2 is NSP-3 which contains certain well preserved macro domains (35). One such domain has been identified as ADP ribose phosphatase which is believed to be involved in diverse pathways, including ADP-ribose metabolism and posttranslational protein modification. The SARS NSP-3 domain readily removes the 1'' phosphate group from Appr-1''-p in in vitro assays, confirming its phosphatase activity (36).
We screened multiple potential ligands against the AMP binding site of this protein [Fig 7].
The results show [Table 3] that Epigallocatechin gallate, also known as epigallocatechin-3-gallate (most abundant catechin in tea) came on top with the highest ICM score and shows very high promise in binding this active receptor site. This drug has been shown to be effective in treatment of colon cancer cells as well (37) and is known for its antioxidant properties. The ligand was shown to be forming 6 hydrogen bonds (G46, A50, G47, S128, W360, A129) with the closest one with G46 (distance 2.6 Å) [Fig 8]. There were also several hydrophobic interactions as well (I131, A38, G130, F132, G48).
The second notable mention is of Cimicifugic acid (extracted from cimicifuga racemosa). Its role has been well documented in preventing collagen degradation by collagenases or collagenolytic enzymes under pathological conditions, wound healing, or inflammation (38). Its role has also been documented for its antiviral activity against enterovirus A71 infection (39).
Table 3. Potential NSP-3 inhibitors from the curated database of herbal and antiviral drugs.
3.3 Main protease (Mpro also called 3CL pro) (PDB code 6w6y)
The protein crystal structure used in this analysis represents the PanDDA analysis of the crystallographic fragment screening of SARS-CoV-2 main protease A-chain. The structure came with a ligand molecule docked around the active catalytic site. This site was subsequently selected for a further docking experiment. Mpro along with PLpro is considered an essential protein for SARS-CoV-2 replication and forms a part of the replicase polyprotein complex. It is found to operate at at-least 11 active cleavage sites on the large polyprotein 1ab (~790 kDa). If the action of this protein is blocked, it would most certainly hinder the replication of this virus. After literature review and using the ligand position that came with the crystal structure, the protein receptor site was identified. It was made sure that the receptor site also includes Cys145 and His41 (catalytic dyad) which are the main amino acid residues of the active catalytic site.
The ligands with the lowest scores on ICM and mfscore scales included Theaflavin, 5,7,3′,4′-Tetrahydroxy-2'-(3,3-dimethylally and Silymarin. This is shown in Fig 9 and Table 4].
Table 4. Potential Mpro inhibitors from the curated database of herbal and antiviral drugs.
Theaflavin, an antioxidant polyphenol which is often formed from the condensation of flavan-3-ols in tea leaves, had the lowest ICM score of -25. Theaflavin and its derivatives have also been shown to be equally effective as an antioxidant in comparison with catechins (40). Furthermore, it was shown to have lower docking scores against RNA dependent RNA polymerase of SARS-CoV-2 in a recent study (41). Though the most important finding was a strong hydrogen bond between its hydrogen (H8) and oxygen of His41 residue which also serves a critical role in forming the active catalytic site. Cys145 was also in close proximity, involved in hydrophobic interactions, thus the ligand was able to fully block this active catalytic site. Apart from these, Theaflavin was forming a total of four more strong hydrogen bonds (T26, T24, Q189, S46) with the closest being 2.7Å in distance. [Fig 9]. There were also several hydrophobic interactions (M49, T25, T45).
The second most interesting ligand was Sillymarin as it had equally low ICM and mfscore. It is.known mostly for its hepatoprotective functions, but recently has been shown possessing antiviral properties against hepatitis C virus (42). Sillymarin is a standardized extract of the milk thistle seeds, containing a mixture of flavonolignans (43) and has also been shown to exert membrane-stabilizing and antioxidant activity (44). It also possesses antifibrotic, immunomodulating and anti-inflammatory effects as mentioned in another study (45). Our findings suggest that these two ligands can act as potential viral Mpro inhibitors of SARS-CoV-2 and can be selected for subsequent testing.
3.4 Spike Protein (PDB code 6m0j)
The selected viral protein crystal structure represents the main interaction site of SARS-CoV-2 spike protein bound with human ACE-2 receptor (hACE2). For our target receptor site for docking, we chose the hACE2 interaction domain on the spike protein (receptor binding domain). Spike protein is considered to be the key for entry into the body cells. It has been shown to be interacting with hACE2 receptors on several human cells including type-2 pneumocytes and enterocytes where it helps in the protein mediated membrane fusion (46). It has also been shown to infect T lymphocytes despite a low expression of hACE2 receptors (47). Thus by blocking this receptor binding domain by a ligand, we can potentially inhibit its entry into human cells.
The docking results showed C20H19F2N3O5 (Dolutegravir) with the lowest ICM score of -19.29 while Chebulagic had the lowest mfscore of -133.7 [Table 5]. The ligands with equally good scores on both scales included Theasinensin with an ICM score of -19.22 and an mfscore of -98 and Colistin (ICM score -16.69, mfscore -126) [Fig 11].
Table 5. Potential spike protein inhibitors from the curated database of herbal and anti-viral drugs.
The chosen ligand (Theasinensin A) showed about 3 different hydrophobic interactions (Y505, Y495, G496) while it formed close hydrogen bonds with 6 residues (T500, Q498, Y449, G502, N501, R403) [Fig 12] with the closest having a distance of 2.47Å from T500. Theasinensin A is among those herbal agents which have already been shown to possess antiviral activity (48).
3.5 RNA Replicase NSP-9 (PDB code 6w9q)
Non structural protein 9 has been found to be playing a key role in binding of RNA-Polymerase to the RNA strand and interacts with NSP-8, thus forming an important component of the RNA-polymerase complex (49). Inhibiting its action may play a role in keeping the RNA-polymerase from binding to the RNA strand thus inhibiting the viral replication. The receptor pocket for action was chosen to be around the bound peptide site of the crystal structure. This was then verified by the pocket finder screening results of the ICM software which proved the selected pocket to be showing maximum hydrophobicity and receptivity for a drug like ligand.
Our docking results showed Epigallocatechin with the lowest ICM score of around -26 while Chebulagic, a benzopyran tannin, had the lowest mfscore of -117 [Fig 13 and Table 6]. Epigallocatechin (found in black tea) has been shown to be an effective antioxidant in several studies (50) while Chebulagic acid has been shown to be inhibiting enterovirus 71 replication in some studies (51). Cimicifugic acid, Scutellarin and Rosemarinic acid were other compounds with lower ICM and mfscores. Epigallocatechin is also found to have a similar zinc ionophore activity as shown by chloroquine in one of the studies (50) thus being a herbal agent, it can be safely used in higher concentrations to achieve better inhibition compared to chloroquine.
Epigallocatechin had the lowest overall score in both metrics. It was shown to be forming 2 hydrophobic interactions (T77, V76) and 4 hydrogen bonds (R111, D78, A107, V110) with the smallest distance of 2.8Å with R111 [FIG 14].
Table 6. Potential NSP-9 inhibitors from the curated database of herbal and antiviral drugs.
These results show the potential role of these ligands in inhibiting NSP-9 thus they may hinder RNA replication.
3.6 Non structural proteins 10-16 (PDB code 6w75)
NSP- 10-16 is a complex of several individual proteins including RNA polymerase (NSP-12), helicase (NSP-13), exonuclease (NSP-14), endoribonuclease (NSP-15) and methyl-transferase (NSP-16) domains. Each one plays its role in bringing about the replication of the virus. The ICM pocket finder was run on this complex to locate the most suitable target pockets for ligand docking [Fig 15]. The selected pocket was found on the A-chain of NSP-16 molecules that had the highest hydrophobicity and DLID scores (Drug-Likeness) (27) and a volume of 348A3.
After selecting the target pocket, virtual screening was performed [Fig 16] which showed the antiviral Favipiravir (Avigan) a novel RNA polymerase inhibitor having the highest affinity (ICM score -36, mfscore -42) with receptor site [Table 7]. Favipiravir is a novel broad-spectrum antiviral drug and has been found to be effective against all the subtypes of influenza virus including those resistant to neuraminidase and M2 inhibitors (52). In a recently published study it was found to be reducing viral infection (53).
Table 7. Potential NSP-16 inhibitors from the curated database of herbal and antiviral drugs.
Favipiravir was developed by Fujifilm Toyama Chemical and approved in Japan in 2014 for the treatment of influenza and is currently under trial for its potential role in inhibiting viral replication of SARS-CoV-2. In our analysis, it was found to be making 5 hydrophobic interactions (D6897, D6898, D6912, M6929, F6947) while it was forming 4 hydrogen bonds (G6869, G6911, C6913, S6896) with the closest one of 2.8Å with G6869 [Fig 17].
Among other ligands, Amentoflavone (a biflavonoid constituent of a number of plants including Chinese plant Selaginella tamariscina and Ginkgo biloba) also showed a high affinity with an ICM score of -33. It has been shown to possess antiviral and anti cancer effects in some studies (54,55). Theaflavin was also among potential inhibitors of this site.
The following table summarizes the key ligands found for all 6 proteins [Table 8].
Table 8. Summary The most potent compounds with their relative binding force scores, immune effects and sources
Proteins
|
Drugs
|
ICM score
|
mfScore
|
Immune effects
|
Viral Effects
|
Source
|
PLpro
|
Theasinensin A
|
-32.25
|
-119.14
|
Reduces the levels of pro-inflammatory mediators (56)
|
May have antiviral effects (57)
|
Flavonoid found in black tea
|
Curucumin
|
-24.23
|
-24.41
|
Reduces the levels of Proinflammatory Cytokines (58)
|
May have antiviral effects (59)
|
Principal curcuminoid of turmeric
|
Quercetin
|
-21.06
|
-58.94
|
Can Reduce the LPS-induced macrophage inflammation (60)
|
May have antiviral effects (61)
|
Red onions and kale
|
Mitoxantrone
|
-22.83
|
-91.00
|
No significant effects on proinflammatory cytokines (62)
|
May have antiviral effects (63)
|
Anthracenedione antineoplastic agent
|
Amentoflavone
|
-23.21
|
-91.43
|
Reduces proinflammatory cytokines (64)
|
May have antiviral effects (65)
|
Plant Constituent including Ginkgo biloba
|
Colistin
|
-11.42
|
-168.50
|
Increases the secretion of proinflammatory cytokines (66)
|
Not Known
|
Antibiotic gram-negative bacteria
|
ADP Ribose
|
Epigallocatechin
|
-43.68
|
-18.75
|
Reduces proinflammatory cytokines (67)
|
May have antiviral effects (68)
|
Most abundant catechin in tea
|
Cimicifugic acid
|
-36.47
|
-96.23
|
Reduces proinflammatory cytokines (69)
|
May have antiviral effects (39)
|
Cimicifuga racemosa
|
Quercetin
|
-36.43
|
-52.34
|
Can Reduce the LPS-induced macrophage inflammation (60)
|
May have antiviral effects (61)
|
Red onions and kale
|
Mpro
|
Theaflavin
|
-25.07
|
-78.31
|
Reduces proinflammatory cytokines (70)
|
May have antiviral effects (71)
|
Tea leaves
|
Silymarin
|
-23.86
|
-110.89
|
Reduces proinflammatory cytokines (72)
|
May have antiviral effects (42)
|
Milk thistle seeds
|
Spike protein
|
Theasinensin A
|
-19.21
|
-98.12
|
Reduces the levels of pro-inflammatory mediators (56)
|
May have antiviral effects (57)
|
Flavonoid found in black tea
|
Chebulagic
|
-12.95
|
-133.70
|
Reduces proinflammatory cytokines (73)
|
May have antiviral effects (74)
|
benzopyran tannin and an antioxidant
|
Dolutegravir
|
-19.28
|
-37.62
|
Not Known
|
Known anti-retroviral used in HIV (75)
|
Antiretroviral medication used, together with other medication, to treat HIV/AIDS (75)
|
Colistin
|
-16.69
|
-126.28
|
Increases the secretion of proinflammatory cytokines (66)
|
Not Known
|
Antibiotic gram-negative bacteria
|
NSP-9
|
Epigallocatechin
|
-26.68
|
-11.13
|
Reduces proinflammatory cytokines (67)
|
May have some antiviral properties (68)
|
Most abundant catechin in tea
|
Chebulagic
|
-16.45
|
-29.97
|
Reduces proinflammatory cytokines (73)
|
May have antiviral effects (74)
|
Cimicifuga racemosa
|
NSP-16
|
Favipiravir
|
-36.04
|
-42.26
|
Reduces proinflammatory cytokines (76)
|
Known antiviral with a variable spectrum against viruses (77)
|
Treatment of influenza
|
Amentoflavone
|
-33.1
|
-117.56
|
Reduces proinflammatory cytokines (64)
|
May have antiviral effects (65)
|
Plants including Chinese plant Selaginella tamariscina and Ginkgo biloba
|
Cimicifugic acid
|
-29.28
|
-109.08
|
Reduces proinflammatory cytokines (69)
|
May have antiviral effects (39)
|
Cimicifuga racemosa
|
3.7 Comparison with other classes as controls
We also tested some of the most mentioned (12,78–81) ligands from different drug classes (Antivirals, Antibacterials, Antifungals, Antimalarials, Antihypertensive, Cholesterol lowering, Anticancer, Antiallergics, AntiInflammatory and membrane stabilizers) as controls for each of the protein targets. Favipiravir and Ribavirin in general, Atoquone and Rosuvastatin against NSP-3, Mitoxantrone against PLpro, Montelukast against Spike protein and Silymarin against NSP-3 and NSP-16 showed high binding affinities (Table 9). However, as mentioned above, most of these control ligands did not score the best affinity scores when compared to other phytochemicals in each protein receptor docking experiment except for Favipiravir.
Table 9. Shows the comparative ICM scores of some of the most representative ligands from different classes tested as controls against the phytochemicals
Ligands
|
PLPro
|
NSP-3
|
Mpro
|
Spike-Pro
|
NSP-9
|
NSP-16
|
Favipiravir
|
-14.80
|
-18.52
|
-9.51
|
-11.17
|
-11.18
|
-36.04
|
Ribavirin
|
-14.01
|
-30.33
|
-8.77
|
-10.16
|
-9.31
|
-22.36
|
Darunavir
|
-12.29
|
-8.82
|
-10.41
|
-7.06
|
0.69
|
-4.43
|
Ritonavir
|
-5.04
|
2.82
|
4.63
|
-4.74
|
6.86
|
10.69
|
Saquinavir
|
-5.94
|
-8.78
|
-14.31
|
-13.19
|
0.18
|
2.86
|
Remdesivir
|
-2.33
|
-8.23
|
-6.49
|
-2.23
|
2.15
|
-1.26
|
Azithromycin
|
7.41
|
31.1
|
5.34
|
1.27
|
7.72
|
21.22
|
Colistin
|
-11.42
|
11.44
|
0.33
|
-16.69
|
27.99
|
29.53
|
Prulifloxacin
|
-14.44
|
-14.69
|
-13.32
|
-8.80
|
-6.84
|
-18.67
|
Atovaquone
|
-12.04
|
-29.46
|
-10.44
|
-7.25
|
-9.83
|
-15.52
|
Artemisinin
|
-8.26
|
2.77
|
-5.12
|
-4.80
|
-7.28
|
-6.94
|
Chloroquine
|
-6.75
|
-11.61
|
-8.37
|
-2.30
|
-14.03
|
-10.64
|
Hydroxychloroquine
|
-4.66
|
-9.54
|
-10.33
|
-6.19
|
-4.60
|
-17.46
|
Ramipril
|
1.43
|
-20.20
|
-8.30
|
-7.02
|
-6.91
|
-3.34
|
Rosuvastatin
|
4.10
|
-33.45
|
-7.47
|
-13.15
|
-7.45
|
-9.39
|
Mitoxantrone
|
-22.83
|
-15.99
|
-6.50
|
-12.63
|
-14.62
|
-9.62
|
Daunorubicin
|
-13.00
|
-12.41
|
-2.78
|
-11.20
|
-3.66
|
-1.83
|
Montelukast
|
-17.02
|
-11.87
|
-17.9
|
-20.24
|
-18.04
|
-15.13
|
Silymarin
|
-11.88
|
-29.79
|
-23.86
|
-5.53
|
-4.50
|
-26.19
|