In Silico Screening of Sugar Alcohol Compounds to Inhibit Viral Matrix Protein VP40 of Ebola Virus

Molecular Biology Reports (2019) 46:3315–3324 https://doi.org/10.1007/s11033-019-04792-w

ORIGINAL ARTICLE

In silico screening of sugar alcohol compounds to inhibit viral matrix protein VP40 of Ebola virus

Nagasundaram Nagarajan1 · Edward K. Y. Yapp2 · Nguyen Quoc Khanh Le1 · Hui‑Yuan Yeh1

Received: 28 December 2018 / Accepted: 28 March 2019 / Published online: 13 April 2019 © Springer Nature B.V. 2019

Abstract Ebola virus is a virulent pathogen that causes highly lethal hemorrhagic fever in human and non-human species. The rapid growth of this virus infection has made the scenario increasingly complicated to control the disease. Receptor viral matrix protein (VP40) is highly responsible for the replication and budding of progeny virus. The binding of RNA to VP40 could be the crucial factor for the successful lifecycle of the Ebola virus. In this study, we aimed to identify the potential drug that could inhibit VP40. Sugar alcohols were enrich with antiviral properties used to inhibit VP40. Virtual screening analysis was perform for the 48 sugar alcohol compounds, of which the following three compounds show the best binding afnity: Sorbitol, Mannitol and Galactitol. To understand the perfect binding orientation and the strength of non-bonded interactions, individual molecular docking studies were perform for the best hits. Further molecular dynamics studies were conduct to analyze the efcacy between the protein–ligand complexes and it was identify that Sorbitol obtains the highest efcacy. The best-screened compounds obtained drug-like property and were less toxic, which could be use as a potential lead compound to develop anti-Ebola drugs.

Keywords Ebola · VP40 · Sugar alcohols · Molecular docking · Molecular dynamics simulations

Introduction each named after the location where it first identified: Zaire (EBOV-Z), Sudan (EBOV-S), Thai forest (EBOV-T), Ebola is a flamentous, negative-stranded and highly patho- Bundibugyo (EBOV-B) and Reston (EBOV-R) with vary- genic RNA virus, which most often results in fatal illness ing fatality rates [3]. The RNA genome sequence length is in humans and other primates [1]. It typically causes hem- about ~ 19 Kb that encodes seven structural proteins such orrhagic fever that leads to severe bleeding from diferent as nucleoprotein (NP), viral matrix proteins VP24, VP30, parts of the body and causes death [2]. As of 30 March VP35, VP40, and RNA polymerase [3, 4]. A vaccine against 2016, 28,610 suspected cases and 11,307 deaths have been EBOV-Z has shown the potential of immune responses report in the most afected countries of Guinea, Liberia, and against surface glycoproteins and nucleoproteins [5]. Several Sierra Leone. The Ebola virus genus comprises fve species investigations on anti-Ebola drugs have carried out, but no efective drug has yet been approve by the FDA. * Nagasundaram Nagarajan Among the viral matrix protein, VP40 plays a crucial role [email protected] in the process of viral transcription at early stages of infec- * Hui‑Yuan Yeh tion [6]. These studies strongly suggested that the binding [email protected] of RNA to VP40 could be the critical factor for the suc- Edward K. Y. Yapp cessful lifecycle of the Ebola virus. Drugs targeting VP40 [email protected]‑star.edu.sg may alter the conformation of the protein and thus it could Nguyen Quoc Khanh Le afect the binding with RNA. The viral matrix protein VP40 [email protected] appears to be highly expressed in Ebola virus and plays a vital role in the budding of Ebola virus from the plasma 1 School of Humanities, Nanyang Technological University, membrane [7]. VP40 is made up of 326 amino acids and 14 Nanyang Dr, Singapore 637332, Singapore has two domains connected by a fexible linker, in which 2 Singapore Institute of Manufacturing Technology, 2 the N-terminal domain is responsible for oligomerization, Fusionopolis Way, Singapore 138634, Singapore

Vol.:(0123456789)1 3 3316 Molecular Biology Reports (2019) 46:3315–3324 while the C-terminal domain is responsible for membrane Sugar alcohols are low digestible carbohydrates com- binding. Viral matrix proteinVP40 shows oligomerize monly found in plant products. These are hydrogenated in both hexamers and octamers, both of which consist of forms of carbohydrate in which the carbonyl group (alde- antiparallel viral matrix protein [VP40] dimers. It have to hyde or ketone) has reduced to a primary or secondary be noted that the formation of these diferent oligomeric hydroxyl group and can obtain by natural or synthetic form. forms was determined by diferences in the interdimeric The clinical efcacy and bioavailability of anti-viral drugs interface, while the monomer–monomer interface within are the important factors that have to taken in concern to intradimeric interface is similar in both octamers and hex- treat viral infections. Sugar alcohols have antiviral proper- amers. In most cases, it was noted that oligomeric VP40 ties [11] and it could be consider in combating Ebola. Sugar expressed in UV-inactivated virion’s and in mammalian alcohol are devoid of toxicity with high compatibility and cells expressing VP40, whereVP40 observed with high lipid biodegradability hence it can be used to enhance the efec- layers. Further VP40 octamer have shown to interact with tiveness against viral which improve the patient compliance RNA molecule at defned sequence pattern [8]. Although and decrease the adverse efect particularly by targeting the we know that octamerization of VP40 is vital for the Ebola specifc active sites for viral target. virus replication, the functions of both the octameric and In this study, we adopted a computational approach to hexameric forms of VP40 still under investigation. The Co- identify the potential lead molecules from the sugar alco- crystal 3D structure (PDB id: 1H2C) of VP40-RNA (Fig. 1) hols. Identifcation of the compounds based on the unique is an octamer shows that the two key amino acids PHE125 conformation, which may interact with specifc binding site and ARG134 of VP40 directly interacted with RNA [6]. The of the target to create a perfect molecular interaction. We RNA-protein structure is stabilised by 140 amino acid resi- used this platform to identify drug candidates that bind and dues of VP40 (including residues THR123, PHE125, and inhibit viral matrix protein VP40 of Ebola virus strains. This ARG134 of a fragment of the N-terminal domain) and UGA approach integrated with the retrieval of sugar alcohol com- (stop codon) of RNA. The crystal structure elucidate that pounds and virtual screening of compounds to identify the VP40 has to undergo two diferent conformational changes top ranking drug like candidates; molecular docking analysis in order to obtained oligomerization. This involves in the to identify the bind afnity of the screened lead compounds displacement of the N-terminal region (residues 31 to 70) and molecular dynamics simulation analysis to understand and the movement of the C-terminal domain, which then the efcacy of the best binding sugar alcohols. Our results forms the binding pocket for the specifc recognition of the explain the efciency of integrated molecular level inves- ssRNA motif U-G-A, Which could be consider as a potential tigation in prioritizing the compounds for further in vitro, target for antiviral drug design [9, 10]. in vivo and clinical testing. This approach will signifcantly

Fig. 1 Surface view of Ebola VP40 bound with RNA showing the amino acid in RNA vicinity

1 3 Molecular Biology Reports (2019) 46:3315–3324 3317 reduce the time, risk, cost, and resources required to deter- position restrained dynamics simulations (NVT and NPT) mine efcacious therapies against future Ebola outbreaks. at 300 K for 300 ps. The equilibrated structures were then subject to molecular dynamics simulations for 50,000 ps at a constant temperature of 300 K and pressure of 1 atm, and the Materials and methods integration time step was set to 2 fs. The non-bonded list was created using an Atom-based threshold of 8 Å. Long-range Dataset electrostatic interactions were managed using the particle- mesh Ewald algorithm and Lennard-Jones mathematical The crystal structure of the matrix protein VP40 of Ebola model were applied with the threshold value of 0.9 nm for virus was retrieved from RCSB PDB database with the cor- interatomic potential calculation. During the simulations, the responding PDB ID 1H2C [8]. The three-dimensional struc- lengths of all bonds containing hydrogen atoms were con- tures of the sugar alcohols compounds were retrieve from strained by utilising the Lincs algorithm [17]; the trajectory PubChem database. The three-dimensional structures of the snapshots were stored for structural analysis every picosec- target protein was energy-minimized using the GROMACS ond. The RMSD, the hydrogen bonds, minimum distance 5.1.2 [12] adopting the GROMOS53a6 [13, 14] force feld were calculate using the Gromacs utilities g_rms, g_hbond parameters before performing the docking analysis. and g_mindist.

Virtual screening and molecular docking analysis ADME and drug likeliness analysis

Docking is often approximate to a lock and key process Lipinski’s rule of fve used to test the bioavailability char- where the confrmation of a ligand and receptor do not acteristics, such as the absorption, distribution, metabolism change during binding. Ligands are often fexible and occupy and elimination (ADME) of the ligands. In this present multiple conformations in solution, and although the confor- study, these molecular properties and the drug-likeness mation of receptors is better defned, they too can change, of the ligands were calculate using the Mol soft program particularly on ligand binding in the so-called “Induced ft (http://molsoft.com/mprop/). model”. VcPpt is the tool used for high throughput virtual screening; it was independently develop as an extension of Prediction of toxicity risk and oral toxicity (LD50) AutoDock vina. The top-ranked ligands further docked with molecular docking suit AutoDock Vina [15]. The AutoDock The ProTox web server used to predict the preclinical oral tools used for the addition of charges and polar hydrogen and toxicity (LD50) of the drug candidates [18]. The ProTox the adjustment of other parameters. Additionally, Auto grid server predicts the toxicity level of candidate drug molecule used to generate grid maps and spacing [16]. The molecu- by comparing the two dimensional chemical alignment of lar docking analysis performed by applying the Lamarck- given molecule with toxicity level known approved com- ian genetic algorithm (LGA). Each of the docking analysis pounds and on the identifcation of over-represented frag- repeated for 10 times with diferent conformation and which ments in the toxic compound. was robust to complete after 250,000 energy evaluations. The docking calculations included a population size of 150 and translational step of 0.2 Å and the docking results were Results and discussion ranked according to the binding free energy and the fre- quency of the most probable binding site. Virtual screening and docking analysis

Molecular dynamics simulations Virtual screening performed for the 48-sugar alcohol compounds (Table 1) with the target octameric structure of RNA The structures of the docked complexes of VP40 with the free viral matrix protein VP40 of Ebola virus using VcPpt screened ligands were used as the starting point for MD tool. Sugar alcohol compounds Sorbitol, Mannitol and Galac- simulations using the GROMACS 5.1.2 [12] adopting the titol (Fig. 2) showed good binding energy and obtained strong GROMOS53a6 force feld parameters. The structures were hydrogen bond interactions with the target VP40. Rest of solvated in a cubic box with a size of 0.9 nm using periodic the compounds obtained less binding energy with less for- boundary conditions and the SPC water model. The topol- mation of non-bonded interactions. Hence, the best binding ogy of the ligands was generate using the PRODRG server. compounds have considered for molecular docking analysis Subsequently, energy minimization carried out for both com- in order to identify the lead compounds atomic pattern in the plex structures using the steepest descent energy protocol. interactions and the amino acids involved in the bindings. Indi- Furthermore, the systems were equilibrated by function a vidual molecular docking analyses were perform for the three

1 3 3318 Molecular Biology Reports (2019) 46:3315–3324

Table 1 List of virtually screened top ten best sugar alcohol compounds S. no PUBCHEM ID Chemical name Molecular Molecular InChI key Structure formula weight (g/ mol)

1. 5780 d-Sorbitol C6H14O6 182.172 FBPFZTCFMRRESA- JGWLITMVSA-N

2. 6251 d-Mannitol C6H14O6 182.172 FBPFZTCFMRRESA- KVTDHHQDSA-N

3. 11850 Galactitol C6H14O6 182.172 FBPFZTCFMRRESA- GUCUJZIJSA-N

4. 61892 (3R,4R)-Hexane- C6H14O6 182.172 FBPFZTCFMRRESA- 1,2,3,4,5,6-hexol SXEFJBAOSA-N

5. 82170 l-Sorbitol C6H14O6 182.172 FBPFZTCFMRRESA- FSIIMWSLSA-N

6. 90540 d-Idito C6H14O6 182.172 FBPFZTCFMRRE- SAZXXMMSQZSA-N

7 120700 Allitol C6H14O6 182.172 FBPFZTCFMRRESAF- BXFSONDSA-N

8 134038 (2S,3S,4R,5R)- C6H14O6 184.164 FBPFZTCFMRRESADY- Hexane- JOCEOGSA-N 1,2,3,4,5,6-hexol

9. 136460 l-Mannitol C6H14O6 182.172 FBPFZTCFMRRESA- BXKVDMCESA-N

10. 151263 d-Altritol 151263 C6H14O6 182.172 FBPFZTCFMRRESAKA- ZBKCHUSA-N

compounds with target VP40 using AutoDock Vina. The drug energy of − 5.4 kcal/mol, − 5.1 kcal/mol and − 4.3 kcal/mol binding sites of VP40 defned as HIS124, PHE125, ARG134, respectively (Table 2) (Fig. 3). Compound Sorbitol and Manni- ASN136, THR123, PHE172, THR173, and TRY171. Com- tol formed three hydrogen bonds with VP40 RNA-binding res- pounds Sorbitol, Mannitol and Galactitol obtained binding idues THR123, HIS124 and THR173 (Fig. 4a, b). Compound

1 3 Molecular Biology Reports (2019) 46:3315–3324 3319

Fig. 2 3D structure of virtually screened three best hits of sugar alcohols a Sorbitol b Mannitol c Galactitol

Table 2 Binding energy and S. no Sorbitol compounds Binding energy Number of Hydrogen bond interacting residues hydrogen bonding residues (Kcal/mol) hydrogen bonds of VP40 with sugar alcohol compounds 1. Sorbitol − 5.4 3 THR123, HIS124,THR173 2. Mannitol − 5.1 3 THR123, HIS124,THR 173 3. Galactitol − 4.3 2 PHE125, GLN170

Fig. 3 Molecular docking of viral matrix protein VP40 with sugar alcohols, green dotted lines represents conventional hydrogen bonds. a Molec- ular interaction of sorbitol with VP40. b Interaction of Mannitol with VP40. c Interaction Galactitol with VP40. (Color fgure online)

Galactitol formed two hydrogen bonds with residues PHE125 in complex with Ebola viral target VP40. From the 50 ns and GLN170 of Ebola target VP40 (Fig. 4c). Furthermore, it MD simulations run, the stability and the interactions has been report in previous studies that two conserved residues between the complexes were analyze for VP40-Sorbitol, PHE125 and ARG134 are the most crucial residues for RNA VP40-Mannitol and VP40-Galactitol. Root mean square binding in the Ebola virus and ARG134 plays a critical role deviation (RMSD), hydrogen bonds and minimum distance in the replication of this virus. Compounds galactitol forming formation between the protein–ligand complexes analyzed. hydrogen bond with amino acid PHE125 of target receptor The RMSD is the basic criteria to measure the convergence VP40 thus inhibition of this compounds can abolish the inter- and the stability of the systems and this can be determined action of VP40 with RNA. by the equilibrium phase obtained by the systems. The backbone RMSD of the protein have calculated, it was observe Molecular dynamics simulations of Sorbitol, that the complex VP40-Sorbitol obtained stable equilibrium Mannitol and Galactitol in complex with VP40 after ~ 20 ns of simulations period (Fig. 5a). However, the VP40-Mannitol and VP40-Galactitol complexes main- Molecular dynamics simulations studies performed for the tained unstable equilibrium throughout the simulations sugar alcohol compounds; Sorbitol, Mannitol and Galactitol period (Fig. 5b, c). Further, the RMSD of the ligands have

1 3 3320 Molecular Biology Reports (2019) 46:3315–3324 calculated from the three complex molecular dynamics sim- From RMSD analysis, it is infer from the protein and ligand ulation, sorbitol in complex VP40 obtained stable RMSD RMSD value the VP40-Sorbitol complex has been well con- value of ~ 0.01 nm to ~ 0.04 nm (Fig. 6a), mannitol obtained verge in compared with VP40-Mannitol and VP40-Galactitol RMSD value of ~ 0.01 nm to ~ 0.06 nm (Fig. 6b) and galacti- complexes. This shows that the compound Sorbitol has more tol obtained RMSD value of ~ 0.05 nm to ~ 0.3 nm (Fig. 6c). stable interactions with VP40 in the dynamics condition. A

Fig. 4 Ligplot analysis shows the 2D images of VP40 interacting ing and hydrogen bond forming amino acid residues of VP40 with residues with sugar alcohol compounds. a Surrounding and hydrogen Mannitol. c Surrounding and hydrogen bond forming amino acid resi- bond forming amino acid residues of VP40 with sorbitol. b Surround- dues of VP40 with Galactitol

Fig. 5 Protein backbone root mean square deviation of viral matrix protein VP40 in complex with sugar alcohols

Fig. 6 Root mean square deviation of ligand heavy atoms in complex with viral matrix protein VP40

1 3 Molecular Biology Reports (2019) 46:3315–3324 3321 delicate balance among all types of weak bonded interac- intermolecular minimum distances between protein–ligand tions determines the stable protein–ligand complex, among complexes have calculated. It was observe that minimum which hydrogen bonds are consider as the most important distance between VP40-Sorbitol, VP40-Mannitol and VP40- bond. Hydrogen bonds were determine by various energies Galactitol as ~ 0.15 to ~ 0.25 nm, ~ 0.2 to ~ 0.25 nm and ~ 0.15 among that electrostatics energy and polarization plays a to ~ 0.3 nm (Fig. 8). A proximate distance indicates a more predominant role. We calculated the number of hydrogen stable complex VP40-Sorbitol maintained a shorter and bonds formed between the VP40-Sorbitol, VP40-Mannitol more stable complex as compared to VP40-Mannitol and and VP40-Galactitol complexes. The hydrogen bond angle VP40-Galactitol. Although VP40-Galactitol complex cutofs and length were 120° and 3.5 Å respectively. The obtained less minimum distance equal to VP40-Sorbitol, it total numbers of hydrogen bonds found between VP40- is not stable during the well-equilibrated simulation period. Sorbitol, VP40-Mannitol and VP40-Galactitol complexes were count from the last 10 ns of simulations period (Fig. 7). ADME analysis It have noted that all the three complexes maintained 1 to 3 hydrogen bonds in the most equilibrated simulations About 60% of potential drug-like compounds were reject period. Hydrogen bond, analysis elucidated that all the best- in the clinical phase because of its molecular property. In screened sugar alcohol compounds obtained stable bindings the earlier days of drug discovery, the drug likeliness and with Ebola target VP40 in the dynamics condition. Further toxicity of the drug-like molecules have been analyze at the

Fig. 7 Number of hydrogen bonds formed between sugar alcohols compounds and viral matrix protein VP40

Fig. 8 Minimum distance between sugar alcohols compounds and viral matrix protein VP40

1 3 3322 Molecular Biology Reports (2019) 46:3315–3324 end of the drug discovery process. However, in recent days Toxicity risks and oral toxicity (LD50) analysis the toxicity test performed as an initial step of drug discovery in order to neglect compounds with poor molecular Drug development is an arduous process that requires small property. Hence, pre-assessment of molecular properties molecules to be highly bioavailable and void of toxic to the of drug-like compounds may rapid the drug discovery pro- entire clinical phase. During drug developmental stage, tox- cess and with the support of experimental results, currently icity and side efects are the major reason that lead to the computational algorithms may predict the drug likeliness, unsuccessful discovery. The toxicity of the drug candidates which limited the cost and time in drug discovery process. are being analyzed through animal models which are time Molecular properties such as the molecular weight (MW), consuming and costly and for each dosage animal scarifca- the number of hydrogen bond acceptors and donors, and tion requires. In silico, prediction serves as an alternative partition coefcient (logP) in a molecule are always crucial approach for simplifying and rationalising drug development when examining the bioavailability of drug-like molecules at the preclinical stage [23]. In this study, we used the Pro- [19]. By considering, the above listed molecular properties toxweb server to calculate the LD50 value of the screened “rule of fve” have been formulate to determine the drug lead compounds. The higher the LD50 dose, the lower the likeness of compounds. The cut-of range of each property toxicity of the compound. The calculated number of atoms, is explained as a drug like compound should have ≤ 500 g/ number of bonds, number of rings, number of rotable bonds, mol molecular weight thus have high membrane permeabil- total charge, molecular polar surface area (A 2), oral toxicity, ity. The calculated octanol–water partition coefcient logP toxicity class, average similarity and prediction accuracy is value should be ≤ 5. The hydrogen bond acceptors should 16, 15, 0, 5, 0, 121.38, 13,500 (mg/kg), 6, 100% and 100% be ≤ 10 and hydrogen bond donors should be ≤ 5. Therefore, respectively for the best screened three sugar alcohol com- the bioavailability and molecular properties of the Sorbi- pounds (Table 4). In compare with the toxicity profle of tol, Mannitol, and Galactitol were calculate based on the drug like molecules screened from the traditional Chinese Lipinski rule of fve using the Molsoft program (http://molso medicine database against the Ebola target VP40 [22], the ft.com/mprop/). All the three screened sugar alcohol com- sugar alcohol compounds obtained better toxicity profle pounds obtained same molecular properties; the number of in term of Oral toxicity and Toxicity Class. These results hydrogen bond acceptor and the number of hydrogen bond indicate that all the screened sugar alcohol displays a better donor have been calculated as six, MolLogP value is − 3.60, safety profle. MolLogS value is 0.30 (in mg/L), Mol PSA value is 121.38 A2, MolVol value is 151.83 A 3 and the number of stereo centers observed as four. The results shows that all the three- Conclusion screened compounds accepted the Lipinski rule of fve [20, 21] (Table 3). Even though all the three compounds obtained Organic compounds represent the majority of new chemi- six hydrogen bond donors, it can be accept as a drug mol- cal entities in drug discovery studies worldwide. Bio- ecule in accordance with the Lipinski rule of fve, which logically potential organic isoforms continue to provide accepts one violation in the rule. The molecular properties of a strong impetus for the development of novel drugs that the sugar alcohol compounds are best in terms of molecular enhance the implicit value in providing molecules for weight, MolLogP and hydrogen acceptors in compare with experiments. More recently, organic chemists are involved previous studies that identify compounds to inhibit Ebola in designing new synthetic short routes for the efective VP40 from the traditional chinese medicine data set [22]. synthesis of drug molecules. More emphasis have given Hence, these sugar alcohol compounds could considered as to multiple compounds base synthesis, which is expect to valid lead compounds to inhibit VP40 of Ebola virus. yield more potential and bioavailability. Novel methods for

Table 3 ADME and drug Drug likeliness score Sugar alcohol compounds Target likeliness analysis for sugar alcohol compounds Sorbitol Mannitol Galactitol

No. of hydrogen bond acceptor 6 6 6 VP40 No. of hydrogen bond donor 6 6 6 MolLogP − 3.60 − 3.60 − 3.60 MolLogS 0.30 (in mg/L) 0.30 (in mg/L) 0.30 (in mg/L) Mol PSA 121.38 A2 121.38A2 121.38A2 MolVol 151.83 A3 151.83A3 151.83A3 No. of stereo centers 4 4 4

1 3 Molecular Biology Reports (2019) 46:3315–3324 3323

Table 4 Drug toxicity analysis Drug toxicity score Sugar alcohol compounds Target for the best-screened hits Sorbitol Mannitol Galactitol

No. of atoms 16 16 16 VP40 No. of bonds 15 15 15 No. of rings 0 0 0 No. of rotable bonds 5 5 5 Total charge 0 0 0 Molecular polar surface area (A2) 121.38 121.38 121.38 Predicted LD50 13,500 (mg/kg) 13,500 (mg/kg) 13,500 (mg/kg) Predicted toxicity class 6 6 6 Average similarity 100% 100% 100% Prediction accuracy 100% 100% 100% combinatorial synthesis permit us to selectively obtained Acknowledgements The authors take this opportunity to thank the enantiomers that display interesting biological activity. It Nanyang Technological University for providing the facilities and for encouragement to carry out this work. is a testament to the value of these discoveries that they are now consider mainstream. Advanced computational tech- Author contributions NN and HYY were involved in designing the nologies applied to identify the drugs from the enriched experiments. The acquisition, analysis and the interpretation of the source of organic molecules in the drug discovery process. data were carry out by NN, HYY, EKYY, and NQKL. All the authors Computational methods can quickly predict the perfect approved the manuscript. binding orientation of all potential drug candidates that Compliance with ethical standards preferred the best potential drug. Finding new drugs for infectious disease is complicated because of potential tar- Conflict of interest The authors declared no confict of interest. get identifcation and validation. Knowledge of the target structure leads to the structure-based drug design which is considered as one of the most efcient and efective methods for tailoring suitable therapeutics. In this aspect, References this study adopted a computational pipeline to identify potential inhibitors to inhibit Ebola viral target VP40 from 1. Sullivan N, Yang ZY, Nabel GJ (2003) Ebola virus pathogenesis: sugar alcohol compounds. 48 sugar alcohol compounds implications for vaccines and therapies. J Virol 77:9733–9737 were subjected for virtual screening, of which three-sugar 2. Dessen A, Volchkov V, Dolnik O, Klenk HD, Weissenhorn W alcohol compounds obtained maximum binding afnity (2000) Crystal structure of the matrix protein VP40 from Ebola virus. EMBO J 19:4228–4236 with VP40: Sorbitol, Mannitol, and Galactitol. Individual 3. Johnson RF, McCarthy SE, Godlewski PJ, Harty RN (2006) Ebola molecular docking analysis performed for the screened hits virus VP35-VP40 interaction is sufcient for packaging 3E-5E with viral target VP40 and Sorbitol identifed as the best minigenome RNA into virus-like particles. J Virol 11:5135–5144 binding compound based on the binding energy and the 4. Hoenen T, Biedenkopf N, Zielecki F, Jung S, Groseth A, Feld- mann H, Becker S (2010) Oligomerization of Ebola virus VP40 number of hydrogen bond interactions. In order to under- is essential for particle morphogenesis and regulation of viral stand the docking results are robotics or fortune binding transcription. J Virol 84:7053–7063 efcacy between the VP40 and inhibitor was calculate for 5. Sullivan N, Yang ZY, Nabel GJ (2003) Ebola virus pathogenesis: all the three complexes. Molecular dynamics simulation implications for vaccines and therapies. J Virol 77:9733–9737 6. Oany AR, Sharmin T, Chowdhury AS, Jyoti TP, Hasan MA (2015) analysis clearly elucidated that VP40-Sorbitol complex Highly conserved regions in Ebola virus RNA dependent RNA obtained stable RMSD in the well-equilibrated simulation polymerase may be act as a universal novel peptide vaccine target: period, while hydrogen bond and minimum distance analy- a computational approach. Silico Pharmacol 3(1):7 sis explain the high efcacy of VP40-Sorbitol complex. 7. Timmins J, Schoehn G, Ricard-Blum S, Scianimanico S, Vernet T, Ruigrok RW, Weissenhorn W (2003) Ebola virus matrix protein Further oral toxicity and drug likeliness analysis has been VP40 interaction with human cellular factors Tsg101 and Nedd4. conducted to explore the ability of these sugar alcohols J Mol Biol 326:493–502 and predicted that could use as a potential lead to treat 8. Gomis-Ruth FX, Dessen A, Timmins J, Bracher A, Kolesnikowa Ebola. Overall, this computational investigation comes out L, Becker S, Klenk HD, Weissenhorn W (2003) The matrix protein VP40 from Ebola virus octamerizes into pore-like structures with possible inhibitors that could use as potential drugs with specifc RNA binding properties. Structure 11(4):423–433 to treat Ebola. Further in vitro and in vivo testing are 9. Hartlieb B, Weissenhorn W (2006) Filovirus assembly and bud- required to evaluate the anti-Ebola activity of these drugs. ding. Virology 344(1):64–70

1 3 3324 Molecular Biology Reports (2019) 46:3315–3324

10. Lee JE, Fusco ML, Hessell AJ, Oswald WB, Burton DR, Saphire 18. Drwal MN, Banerjee P, Dunkel M, Wettig MR, Preissner R (2014) EO (2008) Structure of the Ebola virus glycoprotein bound to an ProTox: a web server for the in silico prediction of rodent oral antibody from a human survivor. Nature 454(7201):177–182 toxicity. Nucleic Acids Res 42:W53–W58 11. Mathew S, Al Thani AA, Yassine HM (2018) Computational 19. Ertl P, Rohde B, Selzer P (2000) Fast calculation of molecular screening of known broad-spectrum antiviral small organic mol- polar surface area as a sum of fragment based contributions and ecules for potential infuenza HA stem inhibitors. PLoS ONE 5:4. its application to the prediction of drug transport properties. J Med https://doi.org/10.1371/journal.pone.0203148 Chem 43:3714–3717 12. Lindahl E, Hess B, van der Spoel D (2001) GROMACS 3.0: a 20. Muegge I (2003) Selection criteria for drug like compounds. Med package for molecular simulation and trajectory analysis. J Mol Res Rev 23:302–321 Model 7:306–317 21. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ (2001) Experi- 13. Hess B, Kutzner C, van der Spoel D, Lindahl E (2008) GROMACS mental and computational approaches to estimate solubility and 4: algorithms for highly efcient, load-balanced, and scalable permeability in drug discovery and development settings. Adv molecular simulation. J Chem Theory Comput 4(3):435–447 Drug Deliv Rev 46:3–26 14. Van Gunsteren FW, Billeter SR, Eising AA, Hunenberger PH, 22. Karthick V, Nagasundaram N, Doss CGP, Chakraborty C, Siva R, Kruger P, Mark AE, Scott WRP, Tironi IG (1996) Biomolecu- Lu A, Zhang G, Zhu H (2016) Virtual screening of the inhibitors lar simulation: the GROMOS96 manual and user guide. Zurich, targeting at the viral protein 40 of Ebola virus. Infect Dis Poverty Biomos 5:12 15. Trott O, Olson AJ (2010) AutoDock Vina: improving the speed 23. Rispin A, Farrar D, Margosches E, Gupta K, Stitzel K, Carr G, and accuracy of docking with a new scoring function, efcient Greene M, Meyer W, McCall D (2002) Alternative methods for optimization, and multithreading. J Comput Chem 31:455–461 the median lethal dose (LD(50)) test: the up-and-down procedure 16. Morris GM, Huey R, Lindstrom W, Sanner MF, Belew RK, Good- for acute oral toxicity. ILAR J 43:233–243 sell DS, Olson AJ (2009) AutoDock4 and AutoDockTools4: auto- mated docking with selective receptor fexibility. J Comput Chem Publisher’s Note Springer Nature remains neutral with regard to 30:2785–2791 jurisdictional claims in published maps and institutional afliations. 17. Lindahl Erik, Hess Berk, van der Spoel David (2001) GROMACS 3.0: a package for molecular simulation and trajectory analysis. J Mol Model 7:306–317

1 3