The computational analysis of ligands binding to various proteins is a powerful method to quickly identify potential molecules for further analysis. These methods have been successfully used in various studies (32). In the first stage, the molecules were docked into the SARS-CoV2 protein snapshots using Glide module of Schrodinger in standard precision (SP) mode. The molecules were then ranked using average Glide score. The MM-GBSA was then performed to ensure the appropriate selection of top hits. All hits were visually inspected for interactions with receptor residues.
3.1 Molecules docking to SARS-CoV2 Structural proteins:
The hallmark feature of coronaviruses is their transmembrane spike (S) glycoprotein as this protein is reason for its name “Corona” in Latin meaning, "Crown". SARS-CoV-2 uses the spike (S) protein to attach to host cells. The spike protein exists as homo-trimers. Each monomer is about 180 kDa and has two distinct subunits S1 and S2. While the receptor binding is mediated by S1 subunit with the help of receptor binding domain (RBD), the fusion between the viral envelope and the host cellular membranes is facilitated by the S2 subunits upon the cleavage of S1-S2 junction by host proteases (33). The S1 subunit of spike protein in SARS-CoV2 has four distinct domains: NTD, CTD1, CTD2 and CTD3, of these the “up” conformation of CTD1 is responsible for binding with ACE2 receptor (34). The S protein, due to its important role in the very first stage of infection, is an important target for development of therapeutics and vaccines. The co-crystal structure of the S-protein with small molecule ligand is not available, therefore we used the sitemap algorithm in Schrodinger to identify the active site on S-protein. The sitemap revealed a site that is very close to the receptor binding domain and trimerization interface lined by the residues Ser 46, Leu 48, Leu 303, Lys 304, Ser 305, Glu 309, Thr 732, Asn 758, Thr 827, Phe 833, Tyr 837, Arg 847, Lys 854, Asn 856, Val 860, Gln 949, Val 952, Asn 955, Gln 957, Asn 960, Val 963, and His 1058. Many of these residues are highly conserved among coronaviruses. The site is overlapping to the site suggested by Kalathiya et al (35). Recently, some of the SARS-CoV2 strains containing mutants of the S-protein (D614G) with high infectivity have been reported. This mutant does not change the structure of S-protein but increases its binding with human TMPRSS2 protein (36). This mutation is away from the identified binding site.
Our molecular docking analysis suggest that capreomycin, posaconazole, mefloquine, nebivolol, angiotensin II, celecoxib and trimethoprim bind to spike protein with appreciable affinity (Supplementary Table 1) (Fig. 2). Other groups have also predicted the binding of posaconazole to spike protein which further substantiates our analysis (37). Posaconazole is an antifungal agent used in the prevention of invasive fungal infections and is also shown to inhibit the entry of Chikungunya virus (38) and replication of Zika and Dengue viruses by binding to oxysterol-binding protein (sterol transporter) (39). Mefloquine is an antimalarial drug used in chloroquine resistant malaria. Nebivolol is an antihypertensive molecule with a very good safety profile in subjects with obstructive respiratory comorbidities (40) and can be an important drug to consider in SARS like diseases. Capreomycin is a polypeptide (isolated from Streptomyces capreolus) used in the treatment of multidrug resistant tuberculosis. Its mechanism is similar to aminoglycosides and used in the inhalation therapy of pulmonary tuberculosis by spray-drying technology (41, 42). It can be a promising prophylactic agent against SARS-CoV2 using similar application strategy (Fig. 2).
The nucleocapsid (N) protein is crucial for the viral RNA packaging. It is made up of two distinct RNA-binding domains (the N-terminal and the C-terminal domain) linked by serine/arginine-rich (SR-rich) domain (SRD)(43). Previous studies with SARS-CoV1 suggest that N protein inhibits TGF-beta, AP-1, NF-kB signaling and type 1 interferon production but induces apoptosis. The sera of COVID-19 patients shows the presence of IgG, IgA, and IgM antibodies against N protein suggesting its role in eliciting humoral immune response (44, 45). In the current study the crystal structure of N-terminal dimerization domain of nucleocapsid phosphoprotein with a ligand (PDB ID: 6WKP) is used. The ligand binds at the dimer interface and has interactions with residues of both of the chains. The active site was defined as residues lying within 5 Å of the cocrystallized ligand. Our study predicts that nelarabine, paclitaxel, regadenoson, quinaprilat and bromfenac are among top molecules binding to N protein (Supplementary Table 1) (Fig. 3).
3.2. Molecules docking to SARS-CoV2 enzymes:
2′-O-Methyl Transferase (Nsp16) of SARS-CoV2 belongs to the S-adenosylmethionine-dependent methyl transferase family and is activated upon binding to Nsp10. Capping of viral mRNA at 5′-end is one of the viral strategy for protecting viral transcripts from host 5′ exoribonucleases and escaping the host innate immune response by mimicking as host mRNAs, thus Nsp16 is the potential target for antiviral therapeutics (46).
The crystal structure of SARS-CoV2 Nsp16 (PDB ID: 6W4H, co-crystallized with S-adenosyl methionine) was used in the current studies. The binding site was found to be lined by the residues Phe 70, Gly 71, Ala 72, Gly 73, Asp 99, Leu 100, Leu 111, Gly 113, Met 131, Tyr 132, Asp 133, Phe 149, Asp 114, Ala 116, Cys 115, and Val 118. Our study shows that methotrexate, viomycin, saralin, saquinavir, venetoclax, vidarabine, histrelin, triptorelin and ribavirin binds to Nsp16 with high affinity (Supplementary Table 1). Methotrexate forms hydrogen bonds with Asn6841, Asp6928, Lys6968, Asp6897, Asn6899 and Asp6876 of NSP16. (Fig. 4).
Methotrexate acts as an antimetabolite and thus used as an antineoplastic drug. It is also used in treatment of inflammatory diseases like rheumatoid arthritis. It decreases the de novo synthesis of purines and pyrimidines and forms dimers with thymidylate synthase (TS), hence also has anti-parasitic effect (47). Methotrexate is also shown to effectively reduce replication of Zika and Dengue viruses (48). Zidovudine is used in HIV1 treatment (21), histrelin and triptorelin are gonadotropin-releasing hormone analogs used in the treatment of central precocious puberty and endometriosis (49). Lanreotide is a long-acting analog of somatostatin and is used for the management of acromegaly, a condition caused by excess secretion of growth hormone. Octreotide is also a somatostatin analog currently used for the treatment of watery diarrhoea and flushes caused by certain carcinoid tumors. Vidarabine (ara-A) is a purine analog and an antiviral drug used for infections caused by herpes simplex and varicella zoster viruses.
Among all the proteins encoded by SARS-CoV2 genome, PLpro (papain-like protease) and 3CLpro (3C chymotrypsin-like protease) are two important viral proteases that cleave the two polyproteins (pp1a and pp1ab) into individual functional viral proteins (Nsp2-Nsp16). The two proteases are important for replication and controlling the host cell response and hence they are among the key targets for the development of therapeutics against SARS-CoV2. These proteases have cysteine in the active site that has also been targeted for the development of covalent inhibitors. There are many small molecules, peptides and peptidomimetics that have been developed against these proteases (50–52).
The 3CLpro is a cysteine protease having three domains: β-barrel Domain I (residues 8–101) and II (residues 102–184) and α-helix domain III (residues 201–306) similar in structure to chymotrypsin (53). The functional protease is a dimer that cleaves polyprotein 1ab in 11 regions at its specific cleavage site (P1) of Leu-Gln↓(Ser, Ala, Gly). The sequences of SARS-CoV2 and SARS-CoV main protease are highly similar (96% identity) and so their 3D structures, barring some surface residues. However, enigmatically the inhibitors of SARS-CoV 3CLpro lopinavir and ritonavir that were also recommended for use against SARS-CoV2 have not shown expected results in the clinical trials for COVID-19 (54). The binding site for 3CLpro was defined as residues falling within 5 Å of the co-crystallized ligand (PDB: 6W63). The residues Thr 25, His 41, Cys 44, Thr 45, Ser 46, Met 49, Asn 142, Gly 143, Ser 144, Cys 145, His 164, Met 165, Glu 166, Leu 167, Pro 168, Asp 187, Arg 188, Gln 189, and Gln 192 were used for defining the active site. Rupintrivir, alatrofloxacin, cangrelor, capreomycin, naldemedine, lopinavir and indinavir are among the drugs predicted to bind to 3CLpro (Supplementary Table 1) (Fig. 5). It is important to note that most of the molecules are making HB interactions with the oxyanion hole residues (Asn 142, Gly 143, Ser 144) of the 3CLpro.
Previous studies report α-ketoamides, lopinavir and ritonavir as inhibitor of 3CLpro (55, 56). Ruprintrivir inhibits human rhinovirus (HRV) 3C protease and has shown broad-spectrum anti-HRV activity (57). Others have also indicated it to be useful against SARS-CoV2 main protease (58). Indinavir is shown to inhibit HIV protease by blocking its active site and leads to immature virus particle formation, however high doses have been linked to lipodystrophy syndrome (59). Naldemedine, is a µ-opioid receptor antagonist used for the treatment of opioid-induced constipation (60).
PLpro is a domain within nsp3 of pp1a/pp1ab with proteolytic activity. It cleaves three sites at 181–182, 818–819, and 2763–2764 at the N-terminus of PP1ab (61). It is the least explored among coronavirus proteins and only a few inhibitors are known for this protein (62). Our study predicts that galidesivir, pralatrexte, methotrexate, daunorubicin, ganciclovir, folic acid, montelukast and itraconazole are among interesting molecules binding to the protease PLpro (Supplementary Table 1). Galidesivir has broad-spectrum antiviral activity (in vitro) against many RNA viruses in nine different families, including the coronaviruses (63). The binding of galidesivir with PLpro is shown in Fig. 6. This drug has been under clinical trials for COVID-19 (NCT03891420). Daunorubicin (DNR) is the anthracycline compound used in the Kaposi's sarcoma and lymphomas treatment of HIV-1 infected patients (64). Moreover its derivative N,N-dimethyl daunomycin (NDMD) is used as the inhibitor of Herpes simplex virus (HSV) (65). Montelukast is shown to reduce proinflammatory cytokines e.g. TNF-α, IL-6 and IL-1β levels (66, 67). A previous study suggests that it inhibits Zika virus by disrupting the integrity of the virions (68). It has been predicted by other groups as well to bind to main protease of SARS-CoV2 (69). An interesting observation is the identification of folic acid as a high affinity ligand of PLpro.
Helicase enzyme (Nsp13) of SARS-CoV2 is motor protein essential for unwinding of both dsDNA and dsRNA and has metal binding (Zn2+) N-terminal and helicase domain (Hel). It is involved in formation of RTC of SARS-CoV2 along with RdRp, which is known to enhance its activity (70). The SARS-CoV2 helicase has 99.8% sequence similarity with that of SARS-CoV. Since it is one of the most conserved proteins in Nidoviruses and is essential for viral RNA synthesis, it is an attractive target for antiviral drug development. A recent review summarizes its importance as a drug target in COVID-19 (71). In the current studies the cryo electron microscope structure of helicase-RdRp (PDB: 6XEZ) was used. The residues within 5 Å of the ADP bound to helicase enzyme were defined as the active site. Our analysis shows that eratapenem, methotrexate, clofarabine, trimethoprim, ascorbic acid, cefixime, and pibrentasvir bind to the helicase with high affinity (Fig. 7). Clofarabine is a potent HIV-1 inhibitor (72). Pibrentasvir, is a HCV NS5A inhibitor effective against all HCV genotypes (73).
The most vital enzyme responsible for the replication/transcription of the viral genome is the RNA-dependent RNA polymerase (RdRp) also known as Nsp12. The primer for RdRp RNA synthesis is synthesized by Nsp8 (74). It has two main functional domains namely nidovirus RdRp associated nucleotidyl transferase (NIRAN) domain and RNA dependent RNA polymerase (RdRp) domain. The NIRAN domain helps in nucleotide transfer while RdRp domain is involved in the polymerisation. The RdRp is conserved in structure and function among RNA viruses (75). This enzyme, due to its importance in viral replication and also to the fact that humans are devoid of it, is a very attractive target (76, 77). Moreoever due to the availability of its structure with cofactors Nsp7 and Nsp8 (PDB: 6M71) and remdesivir (PDB: 7BV2) the structure based design is feasible. A number of studies have been done on development of RdRp inhibitors and some molecules e.g. remdesivir, favipiravir etc. have been approved for emergency use in COVID-19.
In the current studies, we have used the structure of RdRp complexed with remdesivir (PDB: 7BV2). The residues falling withing 5 Å of the remdesivir were defined as active site. Our analysis shows that fludarabine, cobicistat, capreomycin, regadenoson, doxazocin, pibrentasvir, elbasvir, indinavir and remdesivir among others that can bind with RdRp (Fig. 8).
Fludarabine is used for the treatment of hematological malignancies. It inhibits various critical enzymes and results in the inhibition of DNA synthesis. It has been predicted to be ative against SARS-CoV2 RdRp by other groups as well (69, 78). Ribavirin is broad spectrum antiviral used for treatment of RSV infection, hepatitis C and viral hemorrhagic fevers (79). It is a well known RdRp inhibitor. Cobicistat is known to inhibit the cytochrome-mediated metabolism of HIV protease and was approved in 2012 by FDA as pharmacoenhancer for HIV treatment (80). Other groups have also predicted that cobicistat and capreomycin can inhibit SARS-CoV2 protease (81) (82). Pibrentasvir and elbasvir are HCV NS5A inhibitors and indinavir is potent HIV protease inhibitor (83). Another molecule monteleukast, a leukotrine inhibitor used as antihistaminic was also showing good affinity towards RdRp (docking score − 9.42). The molecules we identified to bind to RdRp can serve as potential alternatives to remdesivir.
The Nsp15 is EndoRNase with endoribonuclease activity. It cleaves the 5′ and 3′ of uridylate residues in RNA by forming 2′-3′cyclic phosphodiester. Its mechanism is similar to that of RNase A, RNAse T1 and XendoU (84). Its NendoU activity can interfere with the host’s innate immune response and masks the exposure of viral dsRNA to host dsRNA sensors (85). The crystal structure of SARS-CoV2 Nsp15 cocrystallized with U5P (PDB: 6WLC) was used in the current studies. The active site was defined by the residues falling with 5 Å of the co-crystallized ligand. The active site is situated near the N-terminal and is surrounded by beta sheets and a helix. In our analysis, drugs such as quinapril, octreotide, folic acid, and macimorelin were found to bind to Nsp15 with appreciable affinity (Fig. 9). Quinapril is an angiotensin converting enzyme (ACE) inhibitor and the ACE inhibitors have been suggested to be beneficial for COVID-19 patients (86). Folic acid is essential for DNA and protein synthesis and in the adaptive immune response (87). The dose dependent effect of folic acid on rotavirus infected mice has been reported indicting its antiviral activity (88). Additionally, the role of folic acid in the prevention of cellular entry of SARS-CoV2 has been reported (89). Macimorelin is used for the diagnosis of adult growth hormone deficiency (90). Interestingly other groups have also predicted it to be active against SARS-CoV2 (91, 92).
Nsp14 is the 3’-5’exonuclease that plays a role in proofreading mechanism (93). Nsp14 contains four conserved DE-D-D acidic and a zinc-finger (ZnF) domain (94). The homology model of Nsp14 based on the crystal structure of closely related Nsp14 of SARS CoV (PDB: 5C8T_chainB, 95.07% identity) was used for the current studies. The binding site was defined by comparison with the cocrystallized ligand (PDB: 5C8T, chainB). The SARS-CoV2 Nsp14 active site was found to be lined by the residues Arg289, Trp292, Asn306, Arg310, Asp324, Lys336, Asp362, Ala363, Leu366, Asn386, Asn388, Phe401, Tyr419, Asn422, Phe426, His455, Arg476, Tyr491, and Phe506.
Our molecular docking predicted that cangrelor, venetoclax, pimozide, nilotinib, droperidol, nebivolol, indacaterol, ezetimibe, simeprevir, siponimod, lapatinib, elagolix bind to Nsp14 (Fig. 10).
Pimozide, a calmodulin inhibitor is shown to inhibit Chikungunya virus secretion (95). Moreover, it binds to the envelope protein of HCV and inhibits infection with many HCV genotypes (96). Droperidol is also predicted by other groups to be effective against SARS-CoV2 infection (97). Ezetimibe is shown to inhibit formation of capsid-associated relaxed circular DNA of Hepatitis B Virus (HBV) (98) and is also shown to inhibit Dengue infection by interfering in formation of replication complex (99). Indacaterol is the β2-adrenoceptor agonist and used in the treatment of chronic obstructive pulmonary disease (COPD) since it induces bronchodilation (100). It is a promising candidate for therapeutics against SARS-CoV2 due to its ability to regulate genes involved in suppressing proinflammatory cytokine production and attenuation of airway hyper-responsiveness (101). However, dose and treatment schedule needs to be evaluated due to its counter effect on the expression of RNase L which is vital for antiviral response.
Since one of our major objectives was modeling of the intrinsic flexibility of the SARS-CoV2 proteins by molecular dynamics simulation and finding drugs that can adjust with the site flexibility. We provide a summary of the top drugs for individual proteins and their docking scores in the frames generated by molecular dynamics along with the average MMGBSA score Table 2. The drugs with consistently good docking scores will have a better average. This approach is novel and is not reported anywhere before for screening of drugs against SARS-CoV2 as per the best of our knowledge.
Table 2
Docking and MMGBSA scores of top drugs targeting different SARS-CoV2 proteins.
RdRp |
Drugs/Frames | F1$ | F2$ | F3$ | F4$ | F5$ | Avg_score* | MMGBSA_Avg# |
Fludarabine | -11.32 | -10.85 | -9.83 | -11.34 | -11.86 | -11.04 | -126.47 |
Ribavirin | -10.83 | -9.63 | -10.24 | -10.06 | -11.38 | -10.43 | -110.14 |
Acrobase | -12.94 | -10.58 | -8.08 | -9.28 | -10.63 | -10.30 | -99.25 |
Remdesivir | -10.53 | -9.69 | -8.49 | -10.02 | -10.21 | -9.79 | -107.23 |
Cangrelor | -9.39 | -8.36 | -9.03 | -9.42 | -10.36 | -9.31 | -118.86 |
Nebivolol | -9.56 | -9.89 | -8.69 | -7.98 | -10.13 | -9.25 | -88.87 |
Spike |
Capreomycin | -9.61 | -8.82 | -9.16 | -8.24 | -9.68 | -9.10 | -168.25 |
Trimethoprim | -9.39 | -9.02 | -8.97 | -8.14 | -9.26 | -8.96 | -85.29 |
Mefloquine | -8.56 | -8.69 | -7.89 | -9.02 | -9.10 | -8.65 | -105.84 |
Nebivolol | -8.56 | -8.59 | -8.47 | -7.89 | -9.26 | -8.55 | -135.63 |
Angiotensin II | -10.77 | -8.68 | -7.28 | -8.02 | -7.89 | -8.53 | -96.82 |
Celecoxib | -9.23 | -7.58 | -8.08 | -7.89 | -8.95 | -8.35 | -83.21 |
Main Protease (3CLpro) |
Octreotide | -11.08 | -11.58 | -11.03 | -11.03 | -10.25 | -10.99 | -78.23 |
Rupintrivir | -10.68 | -10.56 | -9.87 | -10.78 | -11.03 | -10.58 | -110.61 |
Lopinavir | -11.04 | -9.58 | -9.97 | -10.89 | -11.31 | -10.56 | -92.06 |
Lapatinib | -10.03 | -10.56 | -10.00 | -10.97 | -11.03 | -10.52 | -87.03 |
Ritonavir | -10.77 | -10.26 | -10.36 | -10.02 | -10.98 | -10.48 | -85.06 |
Fosaprepitant | -9.89 | -9.85 | -10.24 | -10.36 | -10.58 | -10.18 | -93.05 |
Exonuclease |
Cangrelor | -11.79 | -10.58 | -11.06 | -11.03 | -11.28 | -11.15 | -147.54 |
Venetoclax | -9.90 | -10.03 | -10.07 | -9.58 | -10.01 | -9.92 | -165.58 |
Pimozide | -9.68 | -9.50 | -9.89 | -10.05 | -10.21 | -9.87 | -90.28 |
Nebivolol | -9.23 | -9.45 | -10.25 | -10.84 | -9.58 | -9.87 | -115.42 |
Nilotinib | -9.46 | -9.65 | -9.89 | -10.58 | -9.25 | -9.77 | -105.87 |
Droperidol | -9.34 | -9.26 | -10.21 | -9.87 | -8.51 | -9.44 | -100.52 |
EndoNuclease |
Octreotide | -9.40 | -8.12 | -8.58 | -10.87 | -11.03 | -9.60 | -86.25 |
Quinapril | -9.63 | -8.69 | -9.02 | -9.36 | -10.02 | -9.34 | -108.02 |
Metaraminol | -9.47 | -9.57 | -8.24 | -10.24 | -9.02 | -9.31 | -98.57 |
Ribavirin | -8.06 | -9.28 | -8.26 | -10.03 | -10.58 | -9.24 | 111.58 |
Folic Acid | -8.89 | -8.98 | -9.36 | -8.14 | -10.28 | -9.13 | -99.58 |
Macimorelin | -8.52 | -7.85 | -9.28 | -8.97 | -10.03 | -8.93 | -108.20 |
Helicase |
Eratapenem | -10.34 | -9.51 | -9.67 | -10.21 | -9.58 | -9.86 | -120.58 |
Methotrexate | -9.99 | -8.28 | -10.69 | -9.14 | -10.78 | -9.78 | -111.21 |
Trimethoprim | -8.25 | -8.59 | -9.41 | -9.26 | -9.57 | -9.02 | -105.13 |
Cefixime | -9.03 | -8.52 | -8.66 | -8.87 | -9.25 | -8.87 | -60.25 |
Clofarabine | -8.33 | -8.55 | -8.69 | -8.78 | -9.87 | -8.84 | -99.58 |
Ascorbic Acid | -8.56 | -8.77 | -8.69 | -8.96 | -9.09 | -8.81 | -88.26 |
PLpro |
Methotrexate | -11.99 | -11.99 | -11.99 | -11.99 | -12.00 | -11.99 | -80.26 |
Galidesivir | -12.36 | -11.39 | -12.06 | -11.03 | -12.68 | -11.90 | -126.95 |
Pralatrexate | -11.10 | -11.25 | -11.54 | -12.02 | -12.24 | -11.76 | -92.36 |
Ganciclovir | -10.84 | -10.76 | -10.76 | -10.76 | -10.76 | -10.78 | -83.26 |
Daunonubicin | -9.42 | -9.04 | -8.98 | -10.06 | -11.06 | -9.71 | -82.06 |
Itraconazole | -8.95 | -8.78 | -9.06 | -10.25 | -10.17 | -9.44 | -69.20 |
Methyl transferase |
Methotrexate | -8.57 | -8.98 | -9.26 | -8.98 | -9.58 | -9.07 | -90.58 |
Vinadarabine | -8.56 | -9.12 | -8.56 | -8.69 | -8.77 | -8.74 | -65.89 |
Saquinavir | -8.36 | -8.55 | -8.45 | -8.60 | -9.02 | -8.60 | -92.05 |
Venetoclax | -8.10 | -8.68 | -8.96 | -8.45 | -8.78 | -8.59 | -74.28 |
Viomycin | -7.85 | -7.89 | -8.26 | -8.65 | -8.79 | -8.29 | -100.60 |
Saralasin | -8.29 | -8.88 | -8.76 | -8.23 | -7.26 | -8.28 | -95.28 |
Nucleocapsid |
Nelarabine | -8.51 | -8.56 | -8.40 | -9.02 | -9.14 | -8.73 | -90.58 |
Thioguanine | -8.48 | -8.55 | -8.50 | -8.56 | -8.66 | -8.55 | -80.25 |
Paclitaxel | -8.20 | -8.54 | -8.64 | -8.42 | -8.74 | -8.51 | -77.01 |
Quinaprilat | -8.75 | -8.79 | -7.95 | -8.40 | -8.51 | -8.48 | -94.02 |
Regadenoson | -8.95 | -8.05 | -8.74 | -8.09 | -7.99 | -8.36 | -65.84 |
Bromfenac | -8.41 | -8.96 | -8.10 | -7.58 | -8.74 | -8.36 | -84.85 |
$docking score in individual snapshots generated from molecular dynamics (F1-F5), *average docking score. #average MMGBSA score. |