Abstract
Dengue virus (DENV) is a mosquito-borne flavivirus which causes Dengue fever and severe Dengue. It exists as four antigenically different serotypes that are further classified into genotypes with varying degrees of pathogenicity. The non-structural protein 1 (NS1) of DENV has an important role in viral replication and its pathogenesis. NS1 is also considered as an important diagnostic marker for Dengue pathogenesis. To the best of our knowledge, there are no attempts to explore small molecule drugs to target the NS1 of all the serotypes. Here, we have taken the DENV 2 NS1 crystal structure as a reference to model the NS1 structure of the other three serotypes. Once the active site of the NS1 is identified, virtual screening of plant flavonoids is carried out against the NS1 of all the four serotypes. The top 200 molecules in the library with high binding affinities are further analysed to find the common ones having comparable affinities to all the four serotypes. The predicted common flavonoids are subjected to ADMET profiling to further select the most potential molecules that can be used to target NS1 of all the four serotypes.
Electronic supplementary material
The online version of this article (10.1007/s13337-019-00561-2) contains supplementary material, which is available to authorized users.
Keywords: Dengue virus, Non-structural protein 1, Flavonoids, Virtual screening, ADMET
Introduction
Dengue virus (DENV) is a mosquito-borne flavivirus transmitted by Aedes mosquito. It exists as four different serotypes: DENV 1, 2, 3, and 4, all are known to cause Dengue fever (DF) and severe Dengue (SD) [11]. A recent report estimates that annually around 390 million people in the world are affected by Dengue [5]. But still, there are no vaccines or antivirals available to treat Dengue. The positive sense genomic RNA of DENV encodes three structural (Capsid, premembrane and Envelope) and seven non-structural (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) proteins [38]. Several studies have explored these proteins as potential drug candidates to combat Dengue. Several plant derived flavonoids are screened either in vitro or in silico to identify potential leads that can be developed against Dengue. The target proteins included NS1 [26], NS2B/3 protease [13, 27, 30, 31], and NS5 RNA dependent RNA polymerase [8, 9]. Apart from the flavonoids, several plant extracts are also explored to target Dengue NS1 [24, 28, 33, 36, 41]. Though these studies have led to potential lead identification, they are very specific to a particular serotype of Dengue. As the pathogenicity and disease outbreak potential differs among the serotypes and the genotypes within the serotypes, it is desired to identify potential drug candidates with pan-serotypic activity.
The non-structural protein 1 (NS1) of Dengue is a 48 kDa glycoprotein expressed by the virus as a monomer in the infected cells. Post-translational modifications lead to the formation of a homodimer that later gets associated with the membranes of the organelles and the cell itself. NS1 also gets secreted by the infected cells in a soluble hexameric form (trimer of dimers) and hence is used as a diagnostic marker for Dengue infections [2, 7, 16, 42]. NS1 has three distinct structural domains (1) β-roll dimerization domain (1-29); (2) Wing domain (30-180); and (3) Core β ladder domain (181-352). The wing domain has two glycosylation sites at Asn 130 and 175, whereas the core β ladder domain houses the third glycosylation site at Asn 207 [1]. Though NS1 is shown to be involved in viral RNA replication [21–23, 25, 29], immune system evasion and pathogenesis [3, 4, 14, 18, 19], the underlying mechanisms are yet to be elucidated in detail. Recently, it has been shown that the concentrations of secreted NS1 are associated with the onset of SD [17]. Recent reports suggest a role for NS1 in vascular leakage, coagulopathy, and thrombocytopenia [20]. Of the three suggested implications of NS1, role in vascular leakage is the most studied and the molecular mechanism is currently being investigated. Owing to the multifunctional role and its implication in disease severity, NS1 has been studied as a drug target to combat Dengue. Both small molecule and antibody based therapeutics have been explored to target NS1. Furthermore, NS1 based vaccines have been tested in different mouse models. However, the small molecule based therapies have the limitation of insufficient efficacy against one or more serotypes. The NS1 antibodies have shown cross reactivity with host proteins and the vaccines also suffer the same limitation as that of the small molecules.
In the current study, we explored a new strategy to identify unique small molecules that show pan-serotype activity with comparable efficacies. Recently, Sood et al. [33] have shown an alcoholic extract prepared from Cissampelos pareira having pan-serotype inhibition of DENV as assessed by decrease in viral NS1 secretion and viral replication. It is well known that the plant-derived flavonoids are considered to be good antiviral agents due to their minimal side effects and high abundance in nature. Due to the above reasons, we explored flavonoids to target the NS1 protein of Dengue and more specifically identify those flavonoids that will have comparable binding affinities to all the four serotypes of Dengue. Using the recently solved three-dimensional structure of the DENV-2 NS1 [1] as a template, we have modelled the NS1 of the three other serotypes. Since the active sites and the highly conserved N-glycosylation sites (Asn 130 and 175) of NS1 are present in the wing domain, we have decided to target the flavonoids to this region. We have screened more than 2500 flavonoids against NS1 of all the Dengue serotypes at these active sites. We selected the common flavonoids that bind to all the four serotypes with higher binding affinity. ADMET profiling of these pan-serotype ligands further funnelled our search to two unique molecules.
Materials and methods
Amino acid variation among Dengue NS1 proteins
The reference sequences of DENV 1, 2, 3, and 4 were extracted from the NCBI reference genomes (https://www.ncbi.nlm.nih.gov/). The NS1 sequences of the four serotypes were provided in Supplementary Table 1. The alignment of the sequences was built using BioEdit version 7.2.5 [12]. It uses the Clustal omega program to generate the alignment file [32].
Template based modelling of NS1
Only DENV 2 NS1 X-ray crystal structure (4O6B) is available. So the NS1 sequences of DENV 1, 3, and 4 were modeled by using DENV 2 NS1 structure as a template. The homology modelling was performed using the SWISS-MODEL web server (https://swissmodel.expasy.org/interactive) [6]. The best models were selected based on different parameters like GMQE, QMEAN, sequence identity, and coverage of the sequence. The higher the value of GMQE the higher the reliabilty of the predicted structure. GMQE values usually lie between 0 and 1. QMEAN value around 0 is indicated to be in good agreement between modelled and experimental structures. QMEAN Scores of − 4.0 or below are considered to be a low quality model. Structure refinement and energy minimization of the modelled protein structures was done using the MODREFINER web server (https://zhanglab.ccmb.med.umich.edu/ModRefiner/) [40]. MODREFINER uses C-alpha trace, main-chain model, or full-atomic model for refinement. It keeps both side chains and backbone atoms flexible during the time of structural refinement. After refinement, the structural assessment of all the models was performed using ProSA-web (https://prosa.services.came.sbg.ac.at/prosa.php) [39] and Verify3D web server (http://servicesn.mbi.ucla.edu/Verify3D/) [10]. ProSA-web generates two plots, Z-score and residue score plot. Z-score indicates the overall model quality. In Verify3D, the criteria for qualifying a model is that at least 80% of the amino acid residues of a modelled structure should score > = 0.2 in the 3D/1D profile. The stereo chemical assessment of the modelled structure of NS1 proteins was checked by analysing residue by residue geometry and overall structure geometry. Finally, the Ramachandran plots were generated using the PROCHECK (http://servicesn.mbi.ucla.edu/PROCHECK/) [15]. Once the predicted models were cross checked for the quality of prediction, active site residues were identified using 3Dligandsite (http://www.sbg.bio.ic.ac.uk/3dligandsite/) web server [35]. 3DLigandSite finds the ligand bound homologous structures for the query protein structure and then superimposes the modelled structure on the ligand bound structure to determine the ligand binding sites.
Virtual screening studies of flavonoids against Dengue NS1
The flavonoid library used in this study was obtained from Timtec (https://www.timtec.net/). Before the screening, the docking input files were prepared using the Autodock Tools (http://autodock.scripps.edu/resources/adt). For the protein structures, polar hydrogens and Kollman charges were added. The three dimensional coordinates for ligands were obtained from PubChem (http://pubchem.ncbi.nlm.nih.gov/compound). Partial charges were added to the ligands according to the Gasteiger calculation by ADT. These ligands and proteins were saved as .pdbqt files as input files for virtual screening. The virtual screening of flavonoids was performed using the Autodock Vina software (http://vina.scripps.edu/) [34]. Autodock vina is an open source software used for virtual screening. It uses Lamarkian genetic algorithm (LGA) for calculating the possible ligand binding conformations. About 2566 flavonoids were screened against NS1 of all the four serotypes of Dengue. The search grid size kept for the docking was X = 38, Y = 34 and Z = 38 with X centre at − 31.948, Y centre at − 22.696, and Z centre at − 2.487. The exhaustiveness of the search was kept at 56. The energy range was kept at 4 and the number of modes was restricted to 20. The details of the parameters taken for analysis for each serotype were provided in Supplementary Table 2. The top 200 ligands from each of the serotypes were selected and rescored. The ranking of the ligands was done according to the binding affinity for all the four serotypes. After rescoring, only those ligands having activity against all the four serotypes, were selected for further analysis. The interaction maps of the ligands with the protein were generated using the visualization module of the Discovery studio (http://www.3dsbiovia.com).
Druglikeness and toxicity prediction of the ligands
Druglikeness and toxicity analysis predicts whether a particular ligand has the properties to become an orally active drug. For this prediction, Lipinski’s rule of five is a key component taken into account. According to Lipinski’s rule, a ligand molecule should have the following properties to become a good drug candidate. The molecular weight of that ligand should be below 500 Daltons, the cLogP value should be less than 5. It shouldn’t have more than 5 H-bond donors and 10 H-bond acceptors. Here we have used Osiris property explorer (https://www.organic-chemistry.org/prog/peo/) to explore the drug-likeness of the select ligands. Besides taking the Lipinski’s rule of five into account, Osiris predicts the drug like ness and toxicity in a fragment based approach. The input for Osiris is the SMILES notation of the ligands (Supplementary Table 3) and the output is a tabulation of parameters that indicate the likelihood of a molecule to be drug like or not.
Results
The NS1 sequence variation among the serotypes
In the current study, we performed virtual screening of plant derived flavonoids to target NS1 protein of all four serotypes of Dengue in anticipation of identifying unique flavonoid based molecules with pan-serotype activity.
The NS1 sequences for DENV serotypes were obtained from NCBI. Sequence alignment was performed among the reference NS1 sequences of DENV 1, 2, 3, and 4 using the BioEdit program (Fig. 1). Though there is a high degree of conservation among the four serotypes, they do differ in about 15% (out of 352 amino acids) of the positions. Most of the variations tend to fall in the conservative category (a hydrophobic amino acid replaced by another hydrophobic amino acid). Since we are targeting the wing domain of NS1, a closer look at the sequence variation in this region revealed several non-conservative amino acid changes. However, most of this variation is from a charged residue to a hydrophilic residue. One interesting variation we have observed is in DENV-4, wherein a compensatory co-variation of charged amino acids at positions 51 and 85 (E51 K and K85D). Surprisingly, in the crystal structure of DENV2 NS1, these two amino acids are not making any kind of interactions. Please refer to Supplementary Table 4 for the observed amino acid variation for the entire NS1 protein across the four serotypes.
Fig. 1.
The alignment of NS1 0f DENV 1, 2, 3, and 4. Conserved amino acids are represented by “.” This conserved position is according to the conservation of first sequence. The beginning and end of an arrow (shown above the sequence) indicates the beginning and end of a specific domain in the protein
Homology modelling and active site prediction
The X-ray crystal structure of DENV 2 NS1 is available making it a perfect template to model the NS1 of DENV-1, 3, and 4. SWISS-MODEL was used for the homology modelling. The quality of the models was assessed by using sequence similarity and Z-scores. The backbone RMSD between the structures is within 1 Å to each other. The modelled structures were subjected to structure refinement by the MODREFINER. The stereo-chemical assessments of the refined modelled structures were performed by the Verify3D and the ProSA-web. In the case of Verify3D, at least 80% of the amino acids should have scored > = 0.2 in the 3D/1D profile. The modelled DENV 1 NS1 has 83.38%, DENV 3 has 83.67%, and DENV 4 has 84.53% of the amino acids that scored > = 0.2 in the 3D/1D profile, so all the Dengue NS1 models passed the criteria of Verify3D (Supplementary Figure 1). To further confirm the quality of the modelled structures, the Z-scores of the predictions were obtained from ProSA-web. All the models have shown Z-scores within the range for “good models” (Supplementary Figure 2). ProCheck further confirmed the quality of the predicted models with the Ramachandran plots derived from the structures and their agreement with the expected values of Phi and Psi angles for a good model (Supplementary Figure 3).
The active site was determined using the 3Dligandsite. For all the serotypes, the positions 87T and 130N fall in the active site region. Asn at 130 is the glycosylation site present in the wing domain (Table 1). Mapping of the active site residues onto the structure of NS1 revealed the presence of these residues in the wing domain of the NS1 (Supplementary Figure 4).
Table 1.
Active site residues predicted by 3DLigandsite
| Protein name | Serotypes | Active site residues |
|---|---|---|
| NS1 | DENV 1 | 87T; 128V; 130N; 132T |
| DENV 2 | 87T; 130N; 131Q | |
| DENV 3 | 87T; 130N | |
| DENV 4 | 87T; 130N; 131S; 132T |
Molecular docking and ADMET analysis
Virtual screening of ~ 2500 flavonoid derived molecules was carried out against the NS1 of four serotypes of Dengue. The top 200 ligands from the VS run of each of the serotypes were chosen based on the binding affinities of the ligands to the target protein as calculated by the AutoDockVina program. After ranking and rescoring (see “Materials and methods”), 29 ligands were selected for further analysis (Table 2). Most of these molecules have a common coumarin ring structure with very few of them having a chalcone ring (Supplementary Figure 5). These 29 ligands show comparable affinities to NS1 of all four serotypes. Interestingly, all the ligands are ranked in the top 30 of the respective serotypes (Table 3). To elucidate the drug-likeness of these ligands ADMET analysis was carried out using Osiris property explorer.
Table 2.
The PubChem CID numbers with their IUPAC names and their molecular formula of the 29 ligands selected after the virtual screening
| PubChem CID | IUPAC name | Formula |
|---|---|---|
| 5044499 | 2-Oxo-~{N}-[[1,3,3-trimethyl-5-[(2-oxochromene-3-carbonyl)amino]cyclohexyl]methyl]chromene-3-carboxamide | C30H30N2O6 |
| 3981274 | Spiro[1H-indole-3,13′-2,10,16-trioxapentacyclo[12.8.0.03,12.04,9.017,22]docosa-1(14),4,6,8,17,19,21-heptaene]-2,11′,15′-trione | C26H15NO6 |
| 2171774 | 2-Oxo-~{N}-[[1-(2-oxochromene-3-carbonyl)piperidin-4-yl]methyl]chromene-3-carboxamide | C26H22N2O6 |
| 2958078 | 1-[(1-Benzyl-2,3-dimethylindole-5-carbonyl)amino]-3-(4-methyl-2-oxochromen-7-yl)thiourea | C29H26N4O3S |
| 3619355 | [3-(2,3-Dihydro-1,4-benzodioxin-6-yl)-4-oxo-2-(trifluoromethyl)chromen-7-yl] 3-phenyl-2-(phenylmethoxycarbonylamino)propanoate | C35H26F3NO8 |
| 1247446 | ~{N}-[3-(Benzylcarbamoyl)phenyl]-6-chloro-2-oxochromene-3-carboxamide | C24H17ClN2O4 |
| 2971079 | ~{N}-(2-Methyl-4-oxo-5,6,7,8-tetrahydro-[1]benzothiolo[2,3-d]pyrimidin-3-yl)-2-oxochromene-3-carboxamide | C21H17N3O4S |
| 2977427 | 9-(4-Chlorophenyl)-6-(4-oxochromen-3-yl)-5,6,8,9,10,11-hexahydrobenzo[b][1, 4]benzodiazepin-7-one | C28H21ClN2O3 |
| 1263810 | ~{N}-[2-[2-Methyl-5-(trifluoromethoxy)-1~{H}-indol-3-yl]ethyl]-2-oxochromene-3-carboxamide | C22H17F3N2O4 |
| 1278058 | ~{N}’-[2-(1-Bromonaphthalen-2-yl)oxyacetyl]-2-oxochromene-3-carbohydrazide | C22H15BrN2O5 |
| 2210663 | 4-(3-Chloro-6-nitro-1-benzothiophene-2-carbonyl)-~{N}-(4-methyl-2-oxochromen-7-yl)piperazine-1-carbothioamide | C24H19ClN4O5S2 |
| 3143550 | 6-(6-Methyl-4-oxochromen-3-yl)-9-phenyl-5,6,8,9,10,11-hexahydrobenzo[b][1, 4]benzodiazepin-7-one | C29H24N2O3 |
| 3289191 | 3-[2-(4-Methylphenyl)-2,3-dihydro-1~{H}-1,5-benzodiazepin-4-yl]chromen-2-one | C25H20N2O2 |
| 633683 | 7-Benzo[e]benzotriazol-2-yl-3-phenylchromen-2-one | C25H15N3O2 |
| 3798264 | 2-(1-Methylbenzimidazol-2-yl)benzo[f]chromen-3-one | C21H14N2O2 |
| 2965736 | 2-(2-Chlorophenyl)-~{N}-(4-methyl-2-oxochromen-7-yl)quinoline-4-carboxamide | C26H17ClN2O3 |
| 5896476 | (10~{Z})-8,8-dimethyl-10-[(naphthalen-2-ylamino)methylidene]pyrano[2,3-f]chromene-2,9-dione | C25H19NO4 |
| 5940184 | ~{N},~{N}’-bis[4-[(~{E})-2-phenylethenyl]phenyl]propanediamide | C31H26N2O2 |
| 9682442 | 3-Benzyl-4-methyl-7-[(2~{Z})-2-(4-methylphenyl)-2-[(4-nitrophenyl)hydrazinylidene]ethoxy]chromen-2-one | C32H27N3O5 |
| 17250958 | 4-[[4-[[2-(4-Chlorophenyl)-5-methyl-1,3-oxazol-4-yl]methyl]piperazin-1-yl]methyl]-7-methylchromen-2-one | C26H26ClN3O3 |
| 17250975 | 7-Methyl-4-[[4-[[5-methyl-2-(4-methylphenyl)-1,3-oxazol-4-yl]methyl]piperazin-1-yl]methyl]chromen-2-one | C27H29N3O3 |
| 3822224 | 3-(2-Anthracen-9-yl-2,3-dihydro-1,5-benzothiazepin-4-yl)chromen-2-one | C32H21NO2S |
| 9678018 | 3-Benzyl-7-[(2~{Z})-2-[(2,4-dinitrophenyl)hydrazinylidene]-2-(4-methylphenyl)ethoxy]-4-methylchromen-2-one | C32H26N4O7 |
| 17251245 | ~{N}-(2,6-dimethyl-4-oxothieno[2,3-d]pyrimidin-3-yl)-2-[4-[(7-methyl-2-oxochromen-4-yl)methyl]piperazin-1-yl]acetamide | C25H27N5O4S |
| 1822279 | 6-Bromo-3-[2-(naphthalen-1-ylamino)-1,3-thiazol-4-yl]chromen-2-one | C22H13BrN2O2S |
| 1177264 | 4-Methyl-7-[2-oxo-2-(4-phenylphenyl)ethoxy]chromen-2-one | C24H18O4 |
| 17251029 | 4-[[4-[[2-(4-Methoxyphenyl)-5-methyl-1,3-oxazol-4-yl]methyl]piperazin-1-yl]methyl]-7-methylchromen-2-one | C27H29N3O4 |
| 1905602 | 6-(1,1,2,2,3,3,4,4,4-Nonafluorobutyl)-3-oxa-13-azatetracyclo[7.7.1.02,7.013,17]heptadeca-1(17),2(7),5,8-tetraen-4-one | C19H14F9NO2 |
| 1261801 | 6-Methyl-~{N}-(4-methyl-2-oxochromen-7-yl)-2-pyridin-2-ylquinoline-4-carboxamide | C26H19N3O3 |
Table 3.
The rank and binding affinities of the 29 ligands in each of the serotypes
| Ligand number | PubChem CID | DENV 1 | DENV 2 | DENV 3 | DENV 4 | ||||
|---|---|---|---|---|---|---|---|---|---|
| Rank | Binding affinity | Rank | Binding affinity | Rank | Binding affinity | Rank | Binding affinity | ||
| 1 | 3981274 | 02 | − 9.4 | 10 | − 7.9 | 07 | − 9.0 | 06 | − 9.4 |
| 2 | 2171774 | 03 | − 9.1 | 06 | − 8.3 | 10 | − 8.7 | 13 | − 8.7 |
| 3 | 1247446 | 08 | − 8.5 | 09 | − 8.0 | 14 | − 8.3 | 12 | − 8.8 |
| 4 | 2971079 | 08 | − 8.5 | 09 | − 8.0 | 15 | − 8.2 | 08 | − 9.2 |
| 5 | 1278058 | 09 | − 8.4 | 13 | − 7.6 | 13 | − 8.4 | 11 | − 8.9 |
| 6 | 17251245 | 13 | − 8.0 | 15 | − 7.4 | 11 | − 8.6 | 16 | − 8.4 |
| 7 | 1261801 | 16 | − 7.7 | 10 | − 7.9 | 12 | − 8.5 | 05 | − 9.5 |
| 8 | 5044499 | 01 | − 9.5 | 01 | − 9.2 | 02 | − 9.9 | 09 | − 9.1 |
| 9 | 2958078 | 09 | − 8.4 | 09 | − 8.0 | 06 | − 9.1 | 01 | − 10.0 |
| 10 | 3619355 | 06 | − 8.7 | 13 | − 7.6 | 04 | − 9.4 | 06 | − 9.4 |
| 11 | 2977427 | 09 | − 8.4 | 13 | − 7.6 | 15 | − 8.2 | 06 | − 9.4 |
| 12 | 1263810 | 09 | − 8.4 | 14 | − 7.5 | 10 | − 8.7 | 14 | − 8.6 |
| 13 | 2210663 | 09 | − 8.4 | 10 | − 7.9 | 07 | − 9.0 | 16 | − 8.4 |
| 14 | 3143550 | 09 | − 8.4 | 13 | − 7.6 | 13 | − 8.4 | 09 | − 9.1 |
| 15 | 3289191 | 09 | − 8.4 | 05 | − 8.4 | 14 | − 8.3 | 15 | − 8.5 |
| 16 | 633683 | 09 | − 8.4 | 06 | − 8.3 | 12 | − 8.5 | 08 | − 9.2 |
| 17 | 3798264 | 10 | − 8.3 | 10 | − 7.9 | 15 | − 8.2 | 12 | − 8.8 |
| 18 | 2965736 | 10 | − 8.3 | 09 | − 8.0 | 07 | − 9.0 | 08 | − 9.2 |
| 19 | 5896476 | 11 | − 8.2 | 09 | − 8.0 | 07 | − 9.0 | 13 | − 8.7 |
| 20 | 5940184 | 11 | − 8.2 | 14 | − 7.5 | 15 | − 8.2 | 09 | − 9.1 |
| 21 | 9682442 | 12 | − 8.1 | 08 | − 8.1 | 13 | − 8.4 | 04 | − 9.6 |
| 22 | 17250958 | 13 | − 8.0 | 14 | − 7.5 | 06 | − 9.1 | 13 | − 8.7 |
| 23 | 17250975 | 11 | − 8.2 | 13 | − 7.6 | 05 | − 9.2 | 12 | − 8.8 |
| 24 | 3822224 | 12 | − 8.1 | 02 | − 9.0 | 01 | − 10.9 | 02 | − 9.9 |
| 25 | 9678018 | 17 | − 7.6 | 12 | − 7.7 | 15 | − 8.2 | 06 | − 9.4 |
| 26 | 1822279 | 14 | − 7.9 | 15 | − 7.4 | 07 | − 9.0 | 13 | − 8.7 |
| 27 | 1177264 | 12 | − 8.1 | 15 | − 7.4 | 13 | − 8.4 | 14 | − 8.6 |
| 28 | 17251029 | 15 | − 7.8 | 15 | − 7.4 | 07 | − 9.0 | 16 | − 8.4 |
| 29 | 1905602 | 14 | − 7.9 | 13 | − 7.6 | 11 | − 8.6 | 07 | − 9.3 |
Rank The ranks are given according to the binding affinity
ADMET analysis
Of the 29 ligands, eight of them have a predicted cLogP value higher than five, indicating their lack of penetrating the lipid bilayer. Interestingly only two ligands have shown solubility less than − 4, this is expected as all these ligands are plant derived flavonoids and are expected to be highly soluble in aqueous solution. The ‘drug likeness’ of a molecule is calculated by a fragment based approach in Osiris. A positive value indicates a molecule to be more drug like, however since this approach is based on fragments, one can have a molecule with only lipophilic fragments that contribute favourably to the ‘drug likeness’ score when in reality this molecule may not be a good drug candidate because of very low solubility. The ‘drug likeness’ scores for the 29 ligands ranged anywhere from − 102.4 to 5.47. By taking the molecular weight, cLogP, LogS, drug likeness, and toxicity risks into account Osiris calculates an overall ‘Drug Score’. A value of more than 0.5 is considered to be a good score for a molecule to be a good drug candidate (Table 4). The toxicity risks of the 29 ligands are provided in the Table 5. The structures and the interaction profiles are given in Supplementary Figure 5 and Supplementary Table 5.
Table 4.
ADMET property analysis of the screened ligands generated by Osiris property explorer
| Ligand number | PubChem CID | cLogP | Solubility | Molecular weight | TPSA | Druglikeness | Drug score |
|---|---|---|---|---|---|---|---|
| 1 | 3981274 | 2.58 | − 5.05 | 437 | 90.9 | 4.13 | 0.48 |
| 2 | 2171774 | 2.33 | − 3.99 | 458 | 102 | 4.13 | 0.54 |
| 3 | 1247446 | 3.79 | − 5.53 | 432 | 84.5 | 0.56 | 0.42 |
| 4 | 2971079 | 1.85 | − 4.81 | 407 | 116.3 | − 1.09 | 0.33 |
| 5 | 1278058 | 2.63 | − 5.87 | 466 | 93.7 | − 2 | 0.14 |
| 6 | 17251245 | 0.79 | − 3 | 493 | 123 | 2.33 | 0.68 |
| 7 | 1261801 | 4.22 | − 5.34 | 421 | 81.12 | − 16.68 | 0.15 |
| 8 | 5044499 | 3.42 | − 5.49 | 514 | 111 | 0.06 | 0.28 |
| 9 | 2958078 | 4.2 | − 6.08 | 510 | 116 | − 13.56 | 0.12 |
| 10 | 3619355 | 6.29 | − 8.01 | 645 | 109 | − 44.4 | 0.05 |
| 11 | 2977427 | 4.95 | − 7.18 | 468 | 67.4 | 5.47 | 0.33 |
| 12 | 1263810 | 4 | − 5.31 | 430 | 80.4 | − 5.19 | 0.21 |
| 13 | 2210663 | 3.86 | − 5.91 | 542 | 168 | − 16 | 0.12 |
| 14 | 3143550 | 4.69 | − 6.79 | 448 | 67.4 | 2.76 | 0.36 |
| 15 | 3289191 | 4.19 | − 5.25 | 380 | 50.7 | 2.92 | 0.43 |
| 16 | 633683 | 4.62 | − 5.85 | 389 | 57 | − 1.04 | 0.05 |
| 17 | 3798264 | 3.57 | − 4.09 | 326 | 44.1 | 1.39 | 0.24 |
| 18 | 2965736 | 5.4 | − 6.83 | 440 | 68.29 | − 15.15 | 0.1 |
| 19 | 5896476 | 3.84 | − 6.06 | 397 | 64.6 | − 6.19 | 0.07 |
| 20 | 5940184 | 6.62 | − 7.47 | 458 | 58.2 | − 3.5 | 0.03 |
| 21 | 9682442 | 7.02 | − 7 | 533 | 106 | − 11.4 | 0.07 |
| 22 | 17250958 | 3.98 | − 5.18 | 463 | 58.8 | − 2.5 | 0.27 |
| 23 | 17250975 | 3.72 | − 4.79 | 443 | 58.8 | − 5.3 | 0.29 |
| 24 | 3822224 | 7.28 | − 8.87 | 483 | 63.96 | 2.68 | 0.04 |
| 25 | 9678018 | 6.1 | − 7.49 | 578 | 152 | − 10.89 | 0.06 |
| 26 | 1822279 | 5.88 | − 7.1 | 448 | 79.5 | 1.35 | 0.1 |
| 27 | 1177264 | 4.12 | − 5.82 | 370 | 52.6 | − 17.34 | 0.15 |
| 28 | 17251029 | 3.31 | − 4.46 | 459 | 68 | − 6.51 | 0.31 |
| 29 | 1905602 | 5.15 | − 6.27 | 459 | 29.5 | − 102.4 | 0.09 |
cLogP It indicates the ability of a molecule to pass through the lipid bilayer. Higher cLogP value indicates poor absorption
Solubility Solubility of a ligand in an aqueous medium
TPSA Total polar surface area
Druglikeness Whether the ligand is a drug like molecule. A positive value of ‘druglikeness’ indicates a molecule to be more drug like
Drug score Drug score combines cLogP, Solubility, TPSA, Druglikeness, and toxicity risks to provide a score. The value is between 0 and 1. A value higher than 0.5 is considered favourable
Table 5.
Toxicity analysis of the screened ligands generated by Osiris property explorer
| Ligand number | PubChem CID | Mutagenic | Tumorigenic | Irritant | Reproductive effect |
|---|---|---|---|---|---|
| 1 | 3981274 | No | No | No | Medium risk fragment (MRF) |
| 2 | 2171774 | No | No | No | MRF |
| 3 | 1247446 | No | No | No | No |
| 4 | 2971079 | No | No | No | MRF |
| 5 | 1278058 | MRF | MRF | No | MRF |
| 6 | 17251245 | No | No | No | No |
| 7 | 1261801 | No | High-risk fragment (HRF) | No | No |
| 8 | 5044499 | No | No | No | MRF |
| 9 | 2958078 | No | No | No | MRF and HRF |
| 10 | 3619355 | No | No | No | MRF and HRF |
| 11 | 2977427 | No | No | No | No |
| 12 | 1263810 | No | No | No | MRF |
| 13 | 2210663 | HRF | No | No | No |
| 14 | 3143550 | No | No | No | No |
| 15 | 3289191 | No | No | No | MRF |
| 16 | 633683 | HRF | MRF | HRF | HRF |
| 17 | 3798264 | HRF | HRF | No | No |
| 18 | 2965736 | No | HRF | No | No |
| 19 | 5896476 | HRF | HRF | HRF | No |
| 20 | 5940184 | HRF | HRF | HRF | No |
| 21 | 9682442 | No | No | No | HRF |
| 22 | 17250958 | No | No | No | No |
| 23 | 17250975 | No | No | No | No |
| 24 | 3822224 | HRF | HRF | HRF | MRF |
| 25 | 9678018 | No | No | No | HRF |
| 26 | 1822279 | HRF | HRF | No | No |
| 27 | 1177264 | No | No | No | HRF |
| 28 | 17251029 | No | No | No | No |
| 29 | 1905602 | HRF | No | HRF | No |
No These ligands possess no high risk or medium risk fragments
Mutagenic Whether the ligand has any fragment that can cause mutation
Tumorigenic Whether the ligand has any fragment that can cause tumours
Irritant Whether the ligand has any fragment that can cause Irritation
Reproductive effect Whether the ligand has any fragment that can cause any reproductive effect
Only ligand 2 (pubchem ID 2171774) and ligand 6 (Pubchem ID 17251245) have shown a drug score of more than 0.5 (Table 4). The chemical structure of these two ligands has revealed that these molecules have a core coumarin ring structure (Fig. 2).
Fig. 2.
Chemical structure of the final two ligands selected after the ADMET analysis
Interaction profile analysis
The docking poses of the ligands 2 and 6 have revealed two different binding poses. Both the molecules bind in a vertical orientation to DENV 1, 3, and 4 while they are binding in a horizontal geometry to DENV 2 (Fig. 3).
Fig. 3.
Docking poses of ligands 2171774 and 17251245 with the wing domain of NS1 protein. Ligand 2171774 is represented in green and 17251245 is represented in Blue. The figure a and b represents DENV1, c and d is DENV 2, e and f is DENV3 and g and h DENV 4
Interaction profile of ligand 2
The interactions between the ligand and the protein are majorly through hydrophobic interactions and H-bonding is also observed with one or two amino acids at the active site. All but DENV-4 NS1, have 85 K interacting with the ligand either through hydrophobic (DENV 1 and 2) or H-bond (DENV 3) interactions. This particular residue is D in DENV 4. Tryptophan at position 50 makes hydrophobic interactions with the ligand in DENV 1 and 3, whereas this residue is Histidine in DENV 2 and 4. Leucine at position 86 makes H-bonding with ligand in DENV 2, while F at the same position in DENV 1 makes hydrophobic contacts with the ligand. Surprisingly, though DENV 3 and 4 have Leucine at this position, they do not interact with the ligand. Only DENV 2 shows a binding interaction with the glycosylation site N130. Nevertheless, in case of DENV 1 and 3, it is observed that the nearest neighbours of N130, are indeed making either hydrophobic or H-bonding interactions with the ligand. In case of DENV 4, the amino acids interacting with the ligand are completely different than observed for the other three serotypes (Table 6).
Table 6.
Interacting residues and close neighbouring residues of the final two ligands in each serotype
| Serotype | Ligand CID | Residues interacting with ligands | Close neighbouring residues |
|---|---|---|---|
| DENV 1 | 2171774 | Trp 50, Lys 85, Phe 86, Lys 122, Ile 124, Ala 126, Thr 131 | Gly 125, Asp 127, Ile 123, Gln 129, Ala 121, Lys 120, Gly 119 |
| 17251245 | Trp 50, Met 84, Lys 85, His 111, Lys 120, Ala 121, Ala 126, Gln 129 | Thr 131, Asn 130, Phe 86, Ser 117, Ser 114, Tyr 113, Asp 83, Gly 119 | |
| DENV 2 | 2171774 | Thr 72, Pro 73, Asn 76, Lys 85, Leu 86, Asn 130 | Leu 79, Ser 80, Glu 83, Val 84, Thr 87, Ile 88, Met 89, Thr 90 |
| 17251245 | Lys 69, Pro 73, Asn 76 | Thr 72, His 77, Ser 80, Met 89, Thr 90, Gly 91, Asp 92 | |
| DENV 3 | 2171774 | Trp 50, Ile 84, Lys 85, Ala 121, Lys 122, Thr 125, Ser 131 | Glu 51, Asn 83, Tyr 113, Thr 117, Val 124, |
| 17251245 | Ala 47, Trp 50, Ile 84, Lys 85, Leu 86, Ala 121, Lys 122 | Glu 51, Thr 117, Val 124, Thr 125, Ala 126, Ser 131 | |
| DENV 4 | 2171774 | Ala 40, Ala 43, Ser 44, Pro 107, Tyr 113, Trp 115, | Arg 41, Leu 47, Val 78, Leu 79,, Gly 82, Ser 109 |
| 17251245 | Ala 40, Arg 41, Ala 43, Leu 47, Leu 79, Pro 107, Lys 112, Tyr 113, Trp 115 | Ser 44, Val 78, His 84, Gly 82, Ser 109 |
The residues highlighted in bold are interacting through H-bonding
Interaction profile with ligand 6
Compared to ligand 2, ligand 6 shows a slightly different binding interaction profile in all the serotypes. In case of DENV 1, A121 and Q129 are making H-bonding interactions with the ligand while W50 and K85 are making hydrophobic interactions as seen in case of ligand 2. In the case of DENV 2, N76 makes hydrogen bonding interactions with the ligand while K69 and P73 makeup the other interactions. Note that ligand 6 has the lowest binding affinity for DENV 2 NS1 among the four serotypes (Table 3). DENV 3 NS1 ligand profile seems to be similar to that of DENV 1. As is seen before, DENV 4 has a completely different set of amino acids interacting with the ligand (Table 6 and Fig. 4d).
Fig. 4.
Interaction profile of ligands 2171774 and 17251245 with the amino acid residues of DENV 1 (a), 2 (b), 3 (c), and 4 (d). The legend for the color coding of the type of interactions is given at the bottom panel of the figure
Discussion
In the current work, we have proposed a method to find novel small molecules to target Dengue proteins with pan-serotypic activity. The DENV 2 NS1 was used as a template to model the NS1 of the three other serotypes. Virtual screening was done with a flavonoid derived library of ~ 2500 molecules to find molecules that will target all four serotypes.
Evolutionary analysis of Dengue genomes revealed DENV 4 to be in a distant clad than the other three serotypes [37]. Here, the NS1 sequences show a similar trend with the DENV 4 NS1 being the most distant compared to the other three. Nevertheless, the active site residues in all the four serotypes are quite similar (Table 1). Our virtual screening studies have shown that coumarin based flavonoids are the most efficient binders of the wing domain of NS1, albeit with differences in the interacting amino acids across the four serotypes.
Out of the ~ 2500 ligands, based on the binding affinities as calculated by Autodock and the criteria used for selection (see “Materials and methods”), 29 ligands have shown the potential to bind all the four serotypes with comparable affinities and would be good lead molecules for further functionalization. If one were to go by the “drug score” of Osiris, only two molecules (ligand 2 and 6) have all the features required to be good drug candidates. However, upon closer inspection we have found few other molecules (ligands with Pubchem IDs 3981274, 1247446, and 5044499) that upon further functionalization can lead to potential drug candidates.
Looking at the interaction profiles (i.e. the amino acids in the protein that the ligands are making interactions with), it is evident that DENV 1 and 3 have quite similar interaction profiles irrespective of the ligand (Fig. 3a, e and b, f). DENV 2 has unique interaction profile, quite different from the other three serotypes. Interestingly, the orientation of the ligand at the binding pocket is also different in DENV 2 compared to the other three serotypes (Fig. 3). On the other hand, DENV 4, the distant of the four serotypes, evolutionarily speaking, has a ligand binding pocket quite different than the other three. This result is not very surprising when considering the evolutionary analysis of the Dengue (Fig. 2 in [37]). The phylogenetic analysis of the available open reading frames of Dengue clearly shows the close relation between DENV 1 and 3. DENV 2 diverged from the 1 and 3 very early and DENV 4 forms a distinct clade, diverging as a separate clade. As one would assume from the evolutionary history of Dengue, the current study could not find any common amino acid (s) that are targeted by the flavonoids. Nevertheless, we found two molecules, ligand 2 and 6 with compound CID 2171774 (2-oxo-~{N}-[[1-(2-oxochromene-3-carbonyl)piperidin-4-yl]methyl]chromene-3-carboxamide) and 17251245 (~{N}-(2,6-dimethyl-4-oxothieno[2,3-d]pyrimidin-3-yl)-2-[4-[(7-methyl-2-oxochromen-4-yl)methyl] piperazin-1-yl]acetamide) fulfil the criteria to be good drug candidates. Further optimization of these molecules, besides the three other molecules having very good drug scores as predicted by Osiris, with the knowledge of the interaction profiles derived from this study, should lead to drug candidates with pan serotypic activity.
Here, we presented an in silico approach to explore small molecule libraries to identify unique molecules with pan-serotypic activity against Dengue. Using the proposed methodology, we have identified two flavonoid derived molecules with comparable affinities and ‘drug like’ properties to target NS1 protein of all the four serotypes of Dengue. Though the similarities (or lack thereof) in the binding interactions of these molecules are consistent with the evolutionary divergence of the Dengue serotypes, both the molecules are found to be potential drug candidates to target all the four serotypes in our in silico approach. Further functionalization of these molecules will enhance their efficacy and may lead to the discovery of the eluding single molecule to target all the four serotypes. We anticipate that the work flow presented in the current study help in designing new drug leads against viruses having strains of varying serological properties.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Acknowledgements
The authors declare no funding source. We would like to thank Dr. Madduri Srinivasarao for insightful discussions. We would like to thank Simran Agarwal for her assistance in developing the python scripts. We also thank Akshay Shinde for his help in the generation of interaction profiles. BM is funded by the Ph.D. fellowship from BITS Pilani.
Compliance with ethical standards
Conflict of interest
The authors declare no conflict of interest.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Akey DL, Brown WC, Dutta S, Konwerski J, Jose J, Jurkiw TJ, DelProposto J, Ogata CM, Skiniotis G, Kuhn RJ, Smith JL. Flavivirus NS1 structures reveal surfaces for associations with membranes and the immune system. Science. 2014;343:881–885. doi: 10.1126/science.1247749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Alcon S, Talarmin A, Debruyne M, Falconar A, Deubel V, Flamand M. Enzyme-linked immunosorbent assay specific to Dengue virus type 1 nonstructural protein NS1 reveals circulation of the antigen in the blood during the acute phase of disease in patients experiencing primary or secondary infections. J Clin Microbiol. 2002;40:376–381. doi: 10.1128/JCM.40.2.376-381.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Avirutnan P, Fuchs A, Hauhart RE, Somnuke P, Youn S, Diamond MS, Atkinson JP. Antagonism of the complement component C4 by flavivirus nonstructural protein NS1. J Exp Med. 2010;207:793–806. doi: 10.1084/jem.20092545. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Avirutnan P, Hauhart RE, Somnuke P, Blom AM, Diamond MS, Atkinson JP. Binding of flavivirus nonstructural protein NS1 to C4b binding protein modulates complement activation. J Immunol. 2011;187:424–433. doi: 10.4049/jimmunol.1100750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bhatt S, Gething PW, Brady OJ, Messina JP, Farlow AW, Moyes CL, Drake JM, Brownstein JS, Hoen AG, Sankoh O, Myers MF, George DB, Jaenisch T, Wint GR, Simmons CP, Scott TW, Farrar JJ, Hay SI. The global distribution and burden of dengue. Nature. 2013;496:504–507. doi: 10.1038/nature12060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Biasini M, Bienert S, Waterhouse A, Arnold K, Studer G, Schmidt T, Kiefer F, Gallo Cassarino T, Bertoni M, Bordoli L, Schwede T. SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nucl Acids Res. 2014;42:W252–W258. doi: 10.1093/nar/gku340. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chaiyaratana W, Chuansumrit A, Pongthanapisith V, Tangnararatchakit K, Lertwongrath S, Yoksan S. Evaluation of dengue nonstructural protein 1 antigen strip for the rapid diagnosis of patients with dengue infection. Diagn Microbiol Infect Dis. 2009;64:83–84. doi: 10.1016/j.diagmicrobio.2009.01.004. [DOI] [PubMed] [Google Scholar]
- 8.Coulerie P, Eydoux C, Hnawia E, Stuhl L, Maciuk A, Lebouvier N, Canard B, Figadere B, Guillemot JC, Nour M. Biflavonoids of Dacrydium balansae with potent inhibitory activity on dengue 2 NS5 polymerase. Planta Med. 2012;78:672–677. doi: 10.1055/s-0031-1298355. [DOI] [PubMed] [Google Scholar]
- 9.Coulerie P, Nour M, Maciuk A, Eydoux C, Guillemot JC, Lebouvier N, Hnawia E, Leblanc K, Lewin G, Canard B, Figadere B. Structure-activity relationship study of biflavonoids on the Dengue virus polymerase DENV-NS5 RdRp. Planta Med. 2013;79:1313–1318. doi: 10.1055/s-0033-1350672. [DOI] [PubMed] [Google Scholar]
- 10.Eisenberg D, Luthy R, Bowie JU. VERIFY3D: assessment of protein models with three-dimensional profiles. Methods Enzymol. 1997;277:396–404. doi: 10.1016/s0076-6879(97)77022-8. [DOI] [PubMed] [Google Scholar]
- 11.Gubler DJ. Dengue and dengue hemorrhagic fever. Clin Microb Rev. 1998;11:480. doi: 10.1128/cmr.11.3.480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hall T, Biosciences I, Carlsbad C. BioEdit: an important software for molecular biology. GERF Bull Biosci. 2011;2(1):60–61. [Google Scholar]
- 13.Heh CH, Othman R, Buckle MJ, Sharifuddin Y, Yusof R, Rahman NA. Rational discovery of dengue type 2 non-competitive inhibitors. Chem Biol Drug Des. 2013;82:1–11. doi: 10.1111/cbdd.12122. [DOI] [PubMed] [Google Scholar]
- 14.Kurosu T, Chaichana P, Yamate M, Anantapreecha S, Ikuta K. Secreted complement regulatory protein clusterin interacts with dengue virus nonstructural protein 1. Biochem Biophys Res Commun. 2007;362:1051–1056. doi: 10.1016/j.bbrc.2007.08.137. [DOI] [PubMed] [Google Scholar]
- 15.Laskowski RA, MacArthur MW, Moss DS, Thornton JM. PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Crystallogr. 1993;26:283–291. [Google Scholar]
- 16.Lemes EM, Miagostovicsh MP, Alves AM, Costa SM, Fillipis AM, Armoa GR, Araujo MA. Circulating human antibodies against dengue NS1 protein: potential of recombinant D2V-NS1 proteins in diagnostic tests. J Clin Virol. 2005;32:305–312. doi: 10.1016/j.jcv.2004.08.015. [DOI] [PubMed] [Google Scholar]
- 17.Libraty DH, Young PR, Pickering D, Endy TP, Kalayanarooj S, Green S, Vaughn DW, Nisalak A, Ennis FA, Rothman AL. High circulating levels of the dengue virus nonstructural protein NS1 early in dengue illness correlate with the development of dengue hemorrhagic fever. J Infect Dis. 2002;186:1165–1168. doi: 10.1086/343813. [DOI] [PubMed] [Google Scholar]
- 18.Lin CF, Lei HY, Shiau AL, Liu HS, Yeh TM, Chen SH, Liu CC, Chiu SC, Lin YS. Endothelial cell apoptosis induced by antibodies against dengue virus nonstructural protein 1 via production of nitric oxide. J Immunol. 2002;169:657–664. doi: 10.4049/jimmunol.169.2.657. [DOI] [PubMed] [Google Scholar]
- 19.Lin CF, Wan SW, Cheng HJ, Lei HY, Lin YS. Autoimmune pathogenesis in dengue virus infection. Viral Immunol. 2006;19:127–132. doi: 10.1089/vim.2006.19.127. [DOI] [PubMed] [Google Scholar]
- 20.Lin SW, Chuang YC, Lin YS, Lei HY, Liu HS, Yeh TS. Dengue virus nonstructural protein NS1 binds to prothrombin/thrombin and inhibits prothrombin activation. J Infect. 2012;64(3):325–334. doi: 10.1016/j.jinf.2011.11.023. [DOI] [PubMed] [Google Scholar]
- 21.Lindenbach BD, Rice CM. Flaviviridae: the viruses and their replication. In: Knipe DM, Howley PM, Griffin DG, Lamb RA, Martin MA, Roizman B (eds) Fields virology; 2001. p. 991–1041. Walters Kluwer/Lippincott Williams & Wilkins, New York.
- 22.Lindenbach BD, Rice CM. trans-Complementation of yellow fever virus NS1 reveals a role in early RNA replication. J Virol. 1997;71:9608–9617. doi: 10.1128/jvi.71.12.9608-9617.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mackenzie JM, Jones MK, Young PR. Immunolocalization of the dengue virus nonstructural glycoprotein NS1 suggests a role in viral RNA replication. Virology. 1996;220:232–240. doi: 10.1006/viro.1996.0307. [DOI] [PubMed] [Google Scholar]
- 24.Mohan S, McAtamney S, Jayakanthan K, Eskandari R, von Itzstein M, Pinto BM. Antiviral activities of sulfonium-ion glucosidase inhibitors and 5-thiomannosylamine disaccharide derivatives against dengue virus. Int J Antimicrob Agents. 2012;40:273–276. doi: 10.1016/j.ijantimicag.2012.05.002. [DOI] [PubMed] [Google Scholar]
- 25.Muylaert IR, Chambers TJ, Galler R, Rice CM. Mutagenesis of the N-linked glycosylation sites of the yellow fever virus NS1 protein: effects on virus replication and mouse neurovirulence. Virology. 1996;222:159–168. doi: 10.1006/viro.1996.0406. [DOI] [PubMed] [Google Scholar]
- 26.Qamar MT, Mumtaz A, Naseem R, Ali A, Fatima T, Jabbar T, Ahmad Z, Ashfaq UA. Molecular docking based screening of plant flavonoids as dengue NS1 inhibitors. Bioinformation. 2014;10:460–465. doi: 10.6026/97320630010460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Qamar MT, Ashfaq UA, Tusleem K, Mumtaz A, Tariq Q, Goheer A, Ahmed B. In-silico identification and evaluation of plant flavonoids as dengue NS2B/NS3 protease inhibitors using molecular docking and simulation approach. Pak J Pharm Sci. 2017;30:2119–2137. [PubMed] [Google Scholar]
- 28.Rathore AP, Paradkar PN, Watanabe S, Tan KH, Sung C, Connolly JE, Low J, Ooi EE, Vasudevan SG. Celgosivir treatment misfolds dengue virus NS1 protein, induces cellular pro-survival genes and protects against lethal challenge mouse model. Antivir Res. 2011;92:453–460. doi: 10.1016/j.antiviral.2011.10.002. [DOI] [PubMed] [Google Scholar]
- 29.Sampath A, Padmanabhan R. Molecular targets for flavivirus drug discovery. Antivir Res. 2009;81:6–15. doi: 10.1016/j.antiviral.2008.08.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sarwar MW, Riaz A, Dilshad SMR, Al-Qahtani A, Nawaz-Ul-Rehman MS, Mubin M. Structure activity relationship (SAR) and quantitative structure activity relationship (QSAR) studies showed plant flavonoids as potential inhibitors of dengue NS2B-NS3 protease. BMC Struct Biol. 2018;18:6. doi: 10.1186/s12900-018-0084-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Senthilvel P, Lavanya P, Kumar KM, Swetha R, Anitha P, Bag S, Sarveswari S, Vijayakumar V, Ramaiah S, Anbarasu A. Flavonoid from Carica papaya inhibits NS2B-NS3 protease and prevents Dengue 2 viral assembly. Bioinformation. 2013;9:889–895. doi: 10.6026/97320630009889. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sievers F, Higgins DG. Clustal omega. Curr Protoc Bioinform. 2014;48(3):11–16. doi: 10.1002/0471250953.bi0313s48. [DOI] [PubMed] [Google Scholar]
- 33.Sood R, Raut R, Tyagi P, Pareek PK, Barman TK, Singhal S, Shirumalla RK, Kanoje V, Subbarayan R, Rajerethinam R, Sharma N, Kanaujia A, Shukla G, Gupta YK, Katiyar CK, Bhatnagar PK, Upadhyay DJ, Swaminathan S, Khanna N. Cissampelos pareira Linn: natural source of potent antiviral activity against all four dengue virus serotypes. J Immunol. 2015;9:e0004255. doi: 10.1371/journal.pntd.0004255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2010;31:455–461. doi: 10.1002/jcc.21334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wass MN, Kelley LA, Sternberg MJ. 3DLigandSite: predicting ligand-binding sites using similar structures. Nucl Acids Res. 2010;38:W469–W473. doi: 10.1093/nar/gkq406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Watanabe S, Rathore AP, Sung C, Lu F, Khoo YM, Connolly J, Low J, Ooi EE, Lee HS, Vasudevan SG. Dose- and schedule-dependent protective efficacy of celgosivir in a lethal mouse model for dengue virus infection informs dosing regimen for a proof of concept clinical trial. Antivir Res. 2012;96:32–35. doi: 10.1016/j.antiviral.2012.07.008. [DOI] [PubMed] [Google Scholar]
- 37.Weaver SC, Vasilakis N. Molecular evolution of dengue viruses: contributions of phylogenetics to understanding the history and epidemiology of the preeminent arboviral disease. Infect Genet Evolut. 2009;9:523–540. doi: 10.1016/j.meegid.2009.02.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Whitehead SS, Blaney JE, Durbin AP, Murphy BR. Prospects for a dengue virus vaccine. Nat Rev Microbiol. 2007;5:518–528. doi: 10.1038/nrmicro1690. [DOI] [PubMed] [Google Scholar]
- 39.Wiederstein M, Sippl MJ. ProSA-web: interactive web service for the recognition of errors in three-dimensional structures of proteins. Nucl Acids Res. 2007;35:W407–W410. doi: 10.1093/nar/gkm290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Xu D, Zhang Y. Improving the physical realism and structural accuracy of protein models by a two-step atomic-level energy minimization. Biophys J. 2011;101:2525–2534. doi: 10.1016/j.bpj.2011.10.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Yao X, Ling Y, Guo S, Wu W, He S, Zhang Q, Zou M, Nandakumar KS, Chen X, Liu S. Tatanan A from the Acorus calamus L. root inhibited dengue virus proliferation and infections. PLoS Comput Biol. 2018;42:258–267. doi: 10.1016/j.phymed.2018.03.018. [DOI] [PubMed] [Google Scholar]
- 42.Young PR, Hilditch PA, Bletchly C, Halloran W. An antigen capture enzyme-linked immunosorbent assay reveals high levels of the dengue virus protein NS1 in the sera of infected patients. J Clin Microb. 2000;38:1053–1057. doi: 10.1128/jcm.38.3.1053-1057.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.