ABSTRACT in the pathology and replication of the


Ebola is a deadly virus that has recently emerged as an enormous public health concern which causes dangerous illness with high fatality rates of 90%. Potential treatments for the disease are currently being studied although no specific treatment or vaccine exists and hence the need to identify novel inhibitors to aid combat the disease. The VP24 protein of the Ebola virus (EBOV) is believed to play a very key role in the pathology and replication of the Ebola virus disease. VP24 is also known to interfere with the human system’s response to viral infections, thus making it a viable target for combating Ebola Virus disease. This study therefore seeks to identify potential lead compounds from the African flora that can inhibit the activity of the Ebola virus VP24 protein by computer-aided virtual screening. Molecules from AfroDB were used for docking studies with the protein target. Drugs that have shown to inhibit the activity VP24 protein were also docked and their binding energy were examined and compared to those obtained from AfroDB. 21 compounds passed as potent VP24-inhibiting lead compounds. The docking results showed that AfroDB ligand ZINC95486070 demonstrated high binding properties with VP24 than all the VP24-inhibitor drugs (binding energy -9.7 kcal/mol) based on the molecular docking simulation. The leads from AfroDB generally demonstrated better binding potential than the VP24-inhibitor compounds in the computational studies. ****












Ebola virus disease (EVD) is a viral hemorrhagic fever which affects humans and other primates, and is caused by ebolaviruses. Similar to other filoviruses, Ebola virus (EBOV) replicates very efficiently in many cells, giving rise an appreciable number of virus in cells of the mononuclear phagocyte system (MPS) and other cells including liver cells, fibroblasts, and adrenal gland cells. EBOV is believed to be transmitted to humans through direct contact with blood, mucous membranes or through skin contact. The disease is associated with a high risk of death, killing an average of about 50 percent of those infected with the highest death rate being around 90 percent.

The World Health Organization reported a total of 24 Ebola outbreaks involving 1,716 cases from 1976 to 2013. The largest outbreak ever recorded was the epidemic in West Africa mostly in Liberia, Sierra Leone and Guinea, claiming about 11,000 lives out of about 28,000 cases which occurred from December 2013 to January 2016 and the most recent outbreak occurring in May 2017 in the Democratic Republic of the Congo.

EBOV is a single-stranded RNA virus which encodes seven proteins: nucleoproteins (NP), glycoprotein (GP), polymerase (L), VP24, VP30, VP35, and VP40. VP24 is a secondary matrix protein that plays a very key role in the pathology of the Ebola virus disease. The VP24 and VP35 structural proteins of EBOV are believed to play a very important role in interfering with the human immune system’s response to viral infections.

VP24 disrupts the signalling pathway of STAT1. VP24 inhibits the function of karyopherins (KPNA) by binding to it in a region which overlaps with the region where STAT1 binds to the KPNA due to the higher binding affinity between the VP24-KPNA complex. As a result, STAT1 is not able to elicit an immune response and is able to transport viral components into the nucleus or the target cell. VP24 is also responsible for forming fully functional and infectious virus-like particles (VLPs), the promotion of viral nucleocapsid formation and the regulation of the replication. The broad roles the VP24 protein in the replication of the virus and immune-suppression makes it a viable target for combating Ebola Virus disease.

Potential treatments for the disease are currently being studied although no specific treatment or vaccine exists. Few currently available drugs including Miglustat, Ouabain, Nilotinib, Clomiphene, Toremifene among others have shown to be possible VP24-inhibitors. Ouabain, Nilotinib and Miglustat are responsible for blocking viral replication. BCX4430 (Immucillin-A), an antiviral drug, has also been developed as a potential treatment for deadly filovirus infections such as Ebola virus disease and Marburg virus disease

Computer-aided drug design (CADD) has evolved to become one of the most progressive and effective research fields, which also uses cost-efficient and reliable techniques to identify and discover lead compounds for many diseases. Many drugs for combating different diseases on the market are either produced from natural compounds or derived from natural sources. The African flora remains an untapped reservoir of new drug candidates for combating various kinds of diseases.  The African continent can boast of its wealth in biodiversity, which can be exploited to produce novel drug candidates from its natural sources.

This study therefore seeks to discover potential drug candidates from African natural compounds for combating Ebola virus disease by inhibiting the VP24 protein through in silico drug design method, and comparing their interactions with the protein to other VP24-inhibitor drugs.















This study has deployed some software such as GROMACS, UCSF Chimera 1.12, Pymol, Pyrx and Marvin from ChemAxon.

Obtaining VP24 Protein, African Natural Compounds and VP24-Inhibitor Drugs

In this study, the 3D structure of the VP24 protein was retrieved from RCSB PDB website (http://www.rcsb.org/pdb/home/home.do) with corresponding PDB ID: 4M0Q. The African natural compounds were also obtained from AfroDB, a database containing naturally occurring African compounds 1. A total of 833 compounds were obtained from AfroDB. Drugs from previous works that have shown to inhibit the activities of the VP24 protein (Miglustat, Ouabain, Nilotinib, Clomiphene, Toremifene and BCX4430) were retrieved from PubChem (https://pubchem.ncbi.nlm.nih.gov/).

ADMET Screening of AfroDB Compounds

Admet Predictor 8.0 by Simulations Plus, Inc. was used to screen the 833 AfroDB compounds in order to eliminate any unwanted ligands that have high toxicity and very high (>350gmol-1) or very low (<250gmol-1) molecular weight. Also the "ADMET_Risk" model in Admet Predictor 8.0 summarizes the range of potential liabilities that can derail a drug candidate's development, from low solubility to unacceptably fast CYP metabolism to causing phospholipidosis. A reference set of 2270 commercial drugs from the World Drug Index (WDI) were used in Admet Predictor 8.0 to build the models. 90% of the reference set have an ADMET_Risk score less than 7.5. Preparation of VP24 Protein and AfroDB Ligands The AfroDB ligands obtained in ".sdf" were converted to ".pdbqt" using "Open Babel" in PyRx. GROMACS was used to perform the energy minimization of the protein using the OPLS-AA force field. The energy minimized VP24 protein which was in ".gro" was then converted to ".pdb" using Open Babel. The water molecules were then removed from the protein using Pymol and the molecule was saved and then converted to ".pdbqt". Molecular Docking CASTp server was used to determine the binding sites of the protein. Also, from literature, the binding sites of the VP24 protein have been determined using Site Finder. Using Chimera 1.12, the sites predicted by CASTp were analysed. The sites with no openings were eliminated. Also the sites with relatively small volume and area were also eliminated. The virtual screening was performed focusing on the sites predicted. Autodock Vina in PyRx was used to run the screening process and the results were visually inspected in PyMol 2, 3.  ADMET Testing The compounds with high binding affinity and better pose with the energy-minimized VP24 protein were selected and ADMET test was carried out on them again using the Swiss Adme server 4. Also the pharmacokinetic properties of the potential lead compounds were determined. ADMET test is a means of in vitro screening for Absorption, Distribution, Metabolism, Excretion and Toxicity of pharmaceutical compounds. It determines the likelihood of a compound to be a drug candidate. Ligplot+ v1.4.5 was then used to determine the interactions between the protein and the selected ligands. The default parameters were used in running the interaction profile. Marvin from ChemAxon was used for drawing, displaying and characterizing chemical structures 5.                       RESULTS AND DISCUSSION This study used the 3D structure of EBOV VP24 with corresponding PDB ID: 4M0Q, which has been observed with a resolution of 1.92? through X-ray diffraction method. The high-resolution value of the protein indicates that the 3D structure provided by the PDB is good enough to be used for the study. The surface view of the 3D structure of EBOV VP24 protein is shown in Fig. 1. Figure 1: Surface view of the Ebola virus VP24 protein. 247 ligands out of the 833 from AfroDB passed after the screening process with Admet Predictor 8.0. Most of the compounds failed due to their heavy molecular weight and relatively high ADMET_Risk score (above 7.5). Site Finder predicted the binding sites of the EBOV 24 protein as Ala99, Gln103, Leu106, Gly117, Gly120, Leu121, Ser123, Asp124, Leu127, Thr128, Thr183, Gln184, Asn185, and His186. CASTp however predicted about 72 sites on the protein which were inspected using Chimera 1.12 and sites with no openings were eliminated. Also the predicted sites with very small volume and area such that no ligand can fit, were also ignored. Only four sites from the CASTp prediction were considered. The four binding sites of the EBOV VP24 protein from CASTp that were considered are: Pocket 1:  Lys39, Gly44, Ile45, Glu46, Phel47, Asp48, Ser151, Arg154, Ser155, Ile157, Leu158, Ile161, Ser225, Thr226, Ala229, Phe230, Thr231. Pocket 2: Val96, Ala99, Ala100, Gln103, Leu106, Glu113, Ala116, Gly117, Gly120, Ser123, Asp124, Leu127, Thr128, Ile181, Thr183, Gln184, Asn185, His186, Ile188. Pocket 3: Gly101, Ile102, Gln105, Gln109, Ser110, Ile112, Leu115, Leu119, Gln144, Ser146, Lys148, Met149, Leu152, Ile153, Asn156, Ile157. Pocket 4: Ile102, Gln103, Ile107, Gly120, Ser123, Asp124, Leu127, Thr128, Thr183, Gln184, Asn185, His186. The binding sites predicted by Site Finder are the same as that predicted by CASTp in Pocket 2. This makes the other sites predicted by CASTp potential targets for inhibiting the activity of the VP24 protein The binding sites predicted by both CASTp and Site Finder of the EBOV VP24 protein are shown in Fig. 2 (in red).   Figure 2: Predicted binding sites of EBOV VP24 protein by Site Finder (Left) and CASTp (Right). The 247 ligands that passed the ADMET screening were then docked against the VP24 protein. The VP24-inhibitor drugs from PubChem were also added. Ligands that possessed binding energy greater than 8.0kcal/mol with the protein were selected. 45 ligands had a binding energy greater than 8.0kcal/mol. They were further passed through the Swiss Adme server in order to obtain very safe and potent VP24-inhibiting lead compounds. 21 natural products from AfroDB which obey the Lipinski rule of 5, finally passed as lead compounds. The free binding energy of the VP24-ligand complex is a very important factor to consider when selecting possible drug candidates that will inhibit the activity of the protein. Based on the molecular docking simulation performed, ZINC95486070 demonstrated the highest binding properties with VP24 as compared to all the VP24-inhibitor drugs with a binding energy of -9.7 kcal/mol. Clomiphene, Torimefene, BCX4430 and Miglustat possessed the least binding affinity to the VP24 protein with binding energies of -7.5,-7.4, -6.5 and -5.2 kcal/mol respectively. Miglustat is believed to block viral replication when bounded to the VP24 protein. Table 1: Comparing the Binding Energies of AfroDB Compounds and Drugs that Bind to VP24 Protein ZINC ID/ DRUG NAME. 2D MOLECULAR STRUCTURE. BINDING ENERGY (kcal/mol). ZINC000095486070 -9.7 Ouabain -9.3 ZINC000003594643 -9.1 ZINC000095486008 -9.1 Nilotinib -9.1 ZINC000095485910 -8.8 ZINC000040393112 -8.7 ZINC000005839739 -8.6 ZINC000014612849 -8.6 ZINC000100003285 -8.6 ZINC000014611375 -8.5 ZINC000031168265 -8.5 ZINC000100003068 -8.5 ZINC000095486221 -8.4 ZINC000006119190 -8.3 ZINC000014780240 -8.3 ZINC000014887523 -8.3 ZINC000095486121 -8.3 ZINC000014645872 -8.2 ZINC000095486224 -8.2 ZINC000013411309 -8.1 ZINC000028109109 -8.1 ZINC000095486030 -8.1 Clomiphene -7.5 Torimefene -7.4 BCX4430 -6.5 Miglustat -5.2   The molecular interaction between the ligands and the binding sites of the VP24 protein is an important factor when considering a ligand to be a plausible drug candidate. The VP24 protein- ligand interaction of the top five (5) ligands with high binding energy are shown From the Fig. 3 below, ZINC000095486070 is shown to interact with Gly44, Ile45, Arg154, Leu158, Phe230 via hydrophobic bonds. In Fig. 4, ZINC000003594643 also forms hydrophobic bonds with Ala229, Asp48, Glu46, Arg154, and Phe230. ZINC000095486008 also forms hydrophobic bonds with Ala229, Glu46, Arg154, Phe230 and hydrogen bonds with Arg154 and Glu46 of lengths 3.08 and 2.94 respectively as shown in Fig. 5. ZINC000095485910 also forms hydrophobic bonds with Lys39, Asp48 and Glu46 from Fig. 6. Also, it forms very strong hydrogen bonds with Lys39 and Asp48 with bond lengths, 2.91 and 2.89 respectively. Fig. 7, it can be seen that ZINC000040393112 also forms hydrophobic bonds with Glu46 and Arg154.         Figure 3: ZINC000095486070 docked into VP24 (left) and its interaction with VP24 (right).     Figure 4: ZINC000003594643 docked into VP24 (left) and its interaction with VP24 (right).               Figure 5: ZINC000095486008 docked into VP24 (left) and its interaction with VP24 (right).                 Figure 6: ZINC000095485910 docked into VP24 (left) and its interaction with VP24 (right).            Figure 7: ZINC000040393112 docked into VP24 (left) and its interactions with VP24 (right).     Table 2: Molecular properties of the 21 lead compounds and VP24-inhibor drugs COMPOUND logP(o/w) TPSA (Ų) WEIGHT (g/mol) H-Acc H-Don ZINC000095486070 2.53 23.55 292.37 2 0 Ouabain -0.52 206.60 584.65 12 8 ZINC000003594643 2.18 42.01 336.38 4 0 ZINC000095486008 3.23 58.92 338.40 4 2 Nilotinib 4.68 97.62 529.52 8 2 ZINC000095485910 2.66 91.78 321.28 5 3 ZINC000040393112 3.01 63.32 331.49 3 2 ZINC000005839739 2.09 80.67 338.35 5 1 ZINC000014612849 3.05 59.29 338.31 6 0 ZINC000100003285 3.26 61.72 337.41 4 3 ZINC000014611375 2.09 85.27 328.32 6 2 ZINC000031168265 3.05 54.37 292.33 3 1 ZINC000100003068 2.01 89.90 348.39 6 1 ZINC000095486221 3.74 32.59 313.48 2 1 ZINC000006119190 3.11 50.06 308.28 5 0 ZINC000014780240 2.67 100.13 326.30 6 3 ZINC000014887523 2.99 54.37 292.33 3 1 ZINC000095486121 3.20 60.69 322.48 3 3 ZINC000014645872 2.31 98.36 328.27 7 2 ZINC000095486224 1.66 47.53 267.33 3 3 ZINC000013411309 2.13 35.83 262.31 2 1 ZINC000028109109 2.16 41.93 299.36 4 1 ZINC000095486030 2.60 40.16 349.38 4 0 Clomiphene 6.00 12.47 405.96 9 2 Torimefene 5.98 12.47 405.96 2 0 BCX4430 1.51 140.31 265.27 6 6 Miglustat -0.56 84.16 219.28 5 4   ****   RELEVANT FINDING NAME (AfroDB ID) CONSTITUENT COMMENT ZINC000005839739   Aldehyde Aldehyde groups can be used as chemical agents used in hospitals for disinfecting properties. Aldehyde has the ability kill the Ebola virus 6.  ZINC000014780240 and ZINC000095485910 Catechol   ZINC000014887523 and ZINC000031168265 Quinone A (cyclohexane?1,4?dione) Quinone derivatives have been developed to combat infections by the viruses of the family Poxviridae. Also, quinone derivatives have shown antiviral activity against viral infections, especially HIV infections 7. ZINC000100003285 Mannich A       Centers for Disease Control and Prevention (CDC) announced in a publication on their website that chemical agents such as aldehyde, bleach, ammonia phenolics, peracetic acid among others can kill the Ebola virus 6. ZINC000005839739 possesses an aldehyde group in its structure, thus making it a potential drug candidate for tackling the Ebola virus disease. It also has a very high binding affinity to the VP24 protein (-8.6 kcal/mol). Quinone A derivatives have also been developed as antiviral drugs to combat viral infections such as HIV and  those caused by viruses of the family Poxviridae 7. ****   CONCLUSION     BIBLIOGRAPHY 1      F. Ntie-Kang, D. Zofou, S. B. Babiaka, R. Meudom, M. Scharfe, L. L. Lifongo, J. A. Mbah, L. M. Mbaze, W. Sippl, and S. M. N. Efange, "AfroDb: A Select Highly Potent and Diverse Natural Product Library from African Medicinal Plants," PLoS One, vol. 8, no. 10, p. e78085, 2013. 2      O. Trott and A. J. Olson, "AutoDock Vina: Improving the Speed and Accuracy of Docking with a New Scoring Function, EfficientOptimization, and Multithreading," J. Comput. Chem., vol. 31, no. 2, pp. 455–61, 2010. 3      K. Rother, "Introduction to PyMOL," Methods Mol. Biol. Clift. Nj, vol. 635, no. 8, pp. 0–32, 2005. 4      A. Daina, O. Michielin, V. Zoete, C. L. Brooks, and R. Huang, "SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules," Sci. Rep., vol. 7, p. 42717, Mar. 2017. 5      ChemAxon, "Marvin Sketch," https://www.chemaxon.com/products/marvin/, 2013. . 6      "Information on the Survivability of the Ebola Virus in Medical Waste | Ebola Hemorrhagic Fever | CDC." Online. Available: https://www.cdc.gov/vhf/ebola/healthcare-us/cleaning/ebola-virus-survivability.html. Accessed: 11-Jan-2018. 7      J. Koyama, "Anti-infective quinone derivatives of recent patents.," Recent Pat. Antiinfect. Drug Discov., vol. 1, pp. 113–125, 2006.