Inhibitor designing, virtual screening, and docking studies for methyltransferase: A potential target against dengue virus.

AIM: Aim of this work was to design and identify some S-adenosyl-L-homocysteine (SAH) analogs as inhibitors of S-adenosyl-L-methionine-dependent methyltransferase (MTase) protein using computational approaches.
INTRODUCTION: According to the current scenario the dengue has been a global burden. The people are being killed by dengue virus in an abundant number. Despite of lot of research being going on dengue worldwide, there is no single drug which can kill its virus. This creates an urge for new drug target identification and designing. MTase has been reported as an effective target against dengue virus as it catalyzes an essential step in methylation and capping of viral RNA for viral replication.
MATERIALS AND METHODS: The crystal structure of MTase in complex with SAH was used for designing new analogs of SAH. SAH analogs designed were analyzed on the basis of docking, ADMET, and toxicity analysis done using Discovery Studio 3.5.
RESULTS: Seventeen analogs found noncarcinogenic, nonmutagenic, as well as good ADMET properties and good drug-like profile.
CONCLUSION: These SAH analogs, inhibitors of MTase may act as drugs against dengue virus. Further synthesis and biological testing against dengue virus is under observation.

Dengue fever is a mosquito-borne infection found to infect people mainly in tropical and subtropical regions around the world.[1] The primary carrier of dengue virus is Aedes aegypti mosquito. An infected mosquito becomes capable of transmitting the dengue virus to humans after an incubation period of 4–10 days. After infection, humans are main doormatries for multiplication, serving as a source of the virus for uninfected mosquitoes. An infected human can transmit the infection for 4–5 days and maximum 12 via Aedes mosquitoes after their first symptoms appear. The dengue fever has been a serious concern around the world in recent decades. More than 40% of the world's population is at risk of infection from dengue.[2]

The current estimates by WHO there may be 50–100 million dengue infections worldwide every year. This is now endemic in more than 100 countries where the people from the America, South-East of Asia, and the Western Pacific regions are the most seriously affected. The rate of dengue fever infection has exceeded 1.2 million cases in 2008 and over 2.3 million in 2010 as per official data submitted by the Member States in Americas, South-East Asia, and Western Pacific. In 2013, out of 2.35 million cases of dengue, 37,687 cases were of severe dengue were reported in the Americas alone. There is not any permanent treatment of dengue is available. Hence, there is need of any drug against dengue virus. Many proteins and nuclear enzymes of have been reported as important for development and survival of dengue virus.[3]

Nowadays, dengue hemorrhagic fever is a serious cause of hospitalization and death of peoples around the world, mostly in Asian and Latin American countries. Recovery from infection by one serotype provides lifelong immunity against that particular serotype out of all four dengue virus serotypes DEN-1, DEN-2, DEN-3, and DEN-4. However, cross-immunity to the other serotypes is only partial and temporary after recovery. Subsequent infections by other serotypes increase the risk of developing severe dengue.[4]

The nonstructural proteins five contains methyltransferase (MTase) and RNA-dependent RNA polymerase (RdRp) activity. The function of MTase is methylation and catalyzing the capping of viral RNA, which is an essential step for viral replication.[5] MTase catalyzes the transfer of methyl group from substrate methyl donor S-adenosyl-L-methionine (SAM). So MTase inhibition can prevent the methylation step of viral RNA. Different crystal structures of the MTase have been developed which provide information about the structure function.[678] MTase crystal structure in complex with S-adenosyl-L-homocysteine (SAH) was used as a drug target for designing and virtual screening of inhibitors. The nonstructural proteins five C-terminal domain contains the RdRp activity. Biochemical high-throughput assays have been developed to test the efficacy of small molecule RdRp inhibitors and found inhibition of this protein can lead to the death of virus.[9]

Egloff et al. developed the crystal structure of MTase (PDB ID - 2p41) in complex with guanine-N7 and adenosine-2'O methylation (7-MeGpppG2'OMe) and SAH. The position of SAH an amino acid derivative which is formed by demethylation of SAM indicates the binding site of the methyl donor SAM.[9] Derivatives of SAH were designed and screened for affinity and ADMET profile using Discovery Studio 3.5. Some of the designed compounds have an affinity toward MTase better than SAH. The few of analogs were found to have good ADME profiles and found to be noncacinogenic and nonmutagenic. Selected designed can be further be investigated as against dengue virus.

Materials and Methods

Compound library designing

The endogenous ligand SAH is a demethylated product of SAM. MTase catalyzes the transfer of methyl groups using SAM as a methyl donor. Bioisosteric analogs of SAH were designed using Discovery Studio 3.5. Replacement of the fragments of the compound SAH was done using the lead optimization module of Discovery Studio 3.5. Structure-based drug design and often ligand scaffold hopping is performed in the absence of a protein structure.[1011] Scaffold SAH was divided into three substructure parts as Group-1 (2-amino-4-mercaptobutanoic acid), Group-2 (9H-purine), and Group-3 (tetrahydrofuran-3,4-diol) shown in Figure 1. Bioisosterics replacement these three substructures gave total 314 analogs of SAH.

Figure 1

Crystal structure of methyltransferase with its inhibitors S-adenosyl-L-homocysteine in dark yellow color

Docking studies

Docking is drug designing program on the basis of receptor and ligand. It is applied to predict the structure and binding free energy of receptor and ligand complex. Based on the extent of flexibility docking can be classified into four classes: Both the receptor and ligand are flexible, receptor is flexible and ligand is rigid, ligand is flexible and receptor is rigid, and both the receptor and ligand are rigid. LibDock module of Discovery Studio 3.5 was used for docking of the analogs of SAH and target protein. It works on an algorithm for docking small molecules into an active receptor site.[12]

The binding site was created by selecting SAH using Discovery Studio 3.5 having binding site dimensions X (−15.12), Y (13.23), and Z directions (−17.28) for docking. The compounds were prepared and minimized. The conformations of compounds were generated with maximum 225 numbers of conformations to be generated, and conformations of separate isomers were also created within the threshold of 20.0 kcal/mol relative energy. The CHARMm force field was applied to minimize the compounds with 1000 steps of steepest descent (SD) algorithm. The having a root mean square deviation (RMSD) <1.0 Å were considered as duplicates as the higher specified RMSD value will reduce the number of ligand poses returned and an RMS gradient 0.001.[13]

LibDock is a high-throughput algorithm developed by Diller and Merz, which consist of the polar and apolar features as “Hotspots.” The molecules which passed the applied filters were docked with the crystal structure of MTase (PDB ID - 2p41) in complex with guanine-N7 and adenosine-2’O methylation (7-MeGpppG2’OMe) and SAH. Selected compounds were docked to analyze and compare interacting amino acids. CHARMm force field was applied for docking and scoring, and other docking parameters were kept default. Ligand-receptor minimization in situ during dockingwas performed on the complexes to remove any ligand van der Waals clashes prior scoring and calculating binding energy. The 5000 steps of SD with free movement of atoms within the binding site sphere were used during the minimization.[12] The binding energy of protein-ligand complexes was calculated from the free energies of the complex, and free energies of individual protein and the ligand using CHARMm force field and implicit salvation method.[14] Ligands were prepared for use in docking and other applications by using Discovery Studio 3.5, particularly for those which require a three-dimensional (3D) coordinates and biological ionization and tautomerization states. When studying receptor-ligand interactions and other areas, it is important to correctly prepare the ligands. Different protonation states, isomers and tautomers typically have different 3D geometries and binding characteristics.

ADME and toxicity prediction

ADMET properties of the drug should be optimum at least to get drug absorbed and distributed to the site of action. The ratio of lipophilicity and water solubility should be according to the Lipinski rule of five. As a highly soluble drug cannot cross the membranes and a highly lipophilic drug gets the problem to get dissolved. Hence, ADME and toxicity prediction of compounds was done using ADMET and toxicity prediction modules of Discovery Studio 3.5. ADME descriptors allow eliminating the compounds with unfavorable ADMET characteristics early on to avoid expensive reformulation cost later and, for structural refinements that are designed to improve ADME properties, prior to resource expenditure on synthesis.[1516] Analogs were also subjected for analysis of Topkat_FDA_male/female mouse_noncarcinogenicity versus carcinogenicity probability where carcinogen versus noncarcinogen, probability below 0.3 indicates a noncarcinogen, and probability above 0.7 signifies a carcinogen. Cytochrome P450 2D6 (Cyp2d6) plays an essential role in the metabolism of a wide range of xenobiotics in the liver and its inhibition by a drug constitutes a majority cases of drug-drug interaction. Inhibitory effect of analogs was determined computationally. TOPKAT module of Discovery Studio 3.5 validates toxicity estimation through quantitative structural toxicity relationship and similarity searching of the model's database. On the basis of query compound structure, predictive models perform an accurate and valid estimation of the query structure because the model is accurately predictive in that region of space.[17]

Results and Discussion

PDB structure of MTase protein (PDBID -2p41.A) was retrieved from Protein Drug Bank (http://www.rcsb.org/pdb/home/home.do). MTase contains 7-MeGpppG2’OMe and SAH where SAH is shown in the binding pocket in dark yellow color in Figure 1. SAH was selected as lead and divides into three groups for bioisosteric replacement.

Compound library

SAH was selected as lead. The lead modification was done by replacing the groups with their bioisosteres. The lead compound was divided into three groups [Figure 2]. Each group was replaced with bioisosteres. The bioisosteric replacement of Group-1 (2-amino-4-mercaptobutanoic acid), Group-2 (9H-purine), and Group-3 (tetrahydrofuran-3,4-diol) [Figure 2] generated 93, 89, and 132 new bioisosteric analogs of SAH, respectively. Therefore, total 314 analogs were generated.

Figure 2

Structure of S-adenosyl-L-homocysteine showing groups modified

ADME and toxicity analysis

All the designed analogs of SAH were subjected for in-silico ADME analysis using Discover Studio 3.5. Only 41 analogs could pass the ADMET filter (Lipinski’ rule of five) and toxicity filters (carcinogenicity, mutagenicity, Cyp2d6 prediction, and toxicity). A drug should have to get dissolved, i.e., the drug should be soluble because solubility plays a significant role in drug action. Hence, solubility range should between the −8.0 (extremely low solubility) and 0.0 (optimal solubility). Levels defined for solubility signified that the compounds with solubility levels 3 and 4 have good and optimal solubility, respectively, therefore these analogs have solubility as per RO5 as tabulated in Table 1.

Table 1

Inhibitor analogs with good ADMET properties

Absorption of the drug should also be good so that a large amount of drug can reach at the site of action. Absorption of a compound depends mainly on ALogp and polar surface area (PSA) of that compound. ALogp98 denoted lipophilicity (ALogp) value of the compound where the value of it should be between the −2.0 and 5.0 for good intestinal absorption of a drug. PSA denoted by PSA_2d defines the PSA of the compound. Drug to permeate cells and to be absorbed orally it should have PSA not more than 140 A.[2] Likewise, absorption level 0 and 1 defined that the compounds had good and moderate absorption Level, respectively, according to RO5 as given in Table 1. The analogs with nontoxic properties and better ADME properties were screened out. Only 41 analogs found to have well to optimum druglikeness.

All the 41 analogs were subjected for toxicity (mutagenicity and carcinogenicity) analysis using Discovery Studio 3.5. The mutagenic or carcinogenic analogs of SAH were eliminated where only 26 compounds were found nonmutagenic and noncarcinogenic [Table 2]. Carcinogen versus noncarcinogen, probability below 0.3 indicates that only 26 analogs were found a noncarcinogen, in male, as well as in female, according to the data made available by the FDA. The negative effect of selected compounds on Cyp2d6 was analyzed and found that none of selected 26 compounds may inhibit the Cyp2d6. Hence, there may be no chances of drug-drug interaction as shown in Table 2. The bioisosteric analogs of Group-2 (9H-purine) substructure only showed better druglikeness profiles.

Table 2

Inhibitor analogs of SAH which could pass toxicity screening

Docking studies of compounds

Before docking all the selected druglike 26 analogs of SAH and target protein, MTase were prepared using Discover studio 3.2. The binding site was defined by selecting the lead compound SAH in light red color circle around the SAH compound as shown in Figure 1. Prepared compounds were docked with target protein using LibDock module of Discover studio 3.2 and hundreds of pose conformations were generated for each compound. A total of 1040 conformations were generated within the binding pocket. Only 17 analogs were showing better dock score than SAH and SAM. Lead compound SAH found to have LibDock score 124.5 and binding energy (142.4 Kcl/mol) which formed hydrogen bonds with two residues LYS.181 and SER.150 as shown in Figure 3. The methyl donor substrate SAM had LibDock score 125.3 and binding energy (145.3 Kcl/mol) and formed a hydrogen bond with two residues GLY.148 and ARG.84 different than the SAH within the same binding pocket as shown in Figure 3.

Figure 3

Interaction of (a) S-adenosyl-L-homocysteine (LibDock score - 124.5) and (b) S-adenosyl-L-methionine (LibDock score - 125.3) with the methyltransferase target protein. S-adenosyl-L-methionine and S-adenosyl-L-homocysteine in yellow color, interacting amino acids in element color, and hydrogen bonds in green color are shown

Hence, analogs with more LibDock score than SAH and SAM toward target protein (more than 126) and good binding energy was selected as better compound analogs. Only 17 analogs [Table 3] were found to have better interaction affinity (LibDock score) than endogenous ligands SAH and SAM and lower binding energy and were also interacted with the amino acid residues participated in interacting with SAH and SAM given in Table 4.

Table 3

Structures of final 17 designed bioisosteric analogs of SAH ligand

Table 4

Interaction score and interacting residues of better scoring analogs of SAH

The compound Analog 1 with (137.8) highest Libdock score formed H-bond with amino acids residues LYS.181 and ARG.57 shown in Figure 4. The compound Analog 15 scored (135.9) formed the hydrogen bond with SER.150 residue as shown in Figure 5 and Analog 17 scored (136.1) also found to interact with the amino acid residues LYS.181 and ARG.84 as shown in Figure 6. These most scoring three compound analogs (Analog 1, Analog 15, and Analog 17) and all other selected analogs scored better dock score at lower binding energy and formed the hydrogen bonds with the residues interacting with SAH within the same binding pocket. Computational analysis suggested that the designed compounds interacted in a similar manner of SAH within the binding pocket of SAH. The compounds attained similar binding orientation occupying the common amino acids within the binding site Figure 7.

Figure 4

Interaction of Analog 1 most scoring analogs of S-adenosyl-L-homocysteine interacting residues information within the binding pocket of target protein (element color - ligand; dotted blue color line - hydrogen bond)

Figure 5

Interaction of Analog 15 most scoring analogs of S-adenosyl-L-homocysteine within the binding pocket of target protein and interacting amino acid residues (element color - ligand; dotted blue color line - hydrogen bond)

Figure 6

Analog 17 of most scoring analogs of S-adenosyl-L-homocysteine with interacting residues within the binding pocket of target protein (element color - ligand; dotted blue color line - hydrogen bond)

Superimposition of all 17 S-adenosyl-L-homocysteine (yellow color) designed analogs (green color) within the binding site (pale red color) of target protein

Out of analogs designed by bioisosteric replacement of three substructures of SAH, Group-2 (9H-purine) substructure bioisosteric analogs of SAH only showed better interaction affinity than lead SAH toward dengue MTase, as well as druglikeness. These selected analogs will be synthesized and investigated for their activity toward dengue virus.

Conclusion

The drug discovery for dengue fever is the serious concern around the world in recent decades because there is no permanent good drug against dengue. MTase plays a crucial role in methylation and catalyzing the capping of viral RNA, which is an important step for viral replication. Out of the modified analogs of SAH, 17 analogs were found to have a better affinity for MTase protein and ADMET property. Seventeen compounds found nonmutagenic, as well as noncarcinogenic and had a good druglike profile. The dock score of all selected 17 inhibitor analogs of MTase found better than SAH and even SAM also. Three analogs (Analog 1, Analog 15, and Analog 17) were found best scoring among all. Compounds interacted in the pattern of SAH and SAM-forming hydrogen bond with LYS.181, SER.150, and ARG.84 amino acid residues found to interact with SAM and SAH. So the computational study suggested that the designed analogs may act as inhibitors of MTase protein and may lead to the better drug against dengue virus. These analogs need further wet laboratory investigation. The synthesis and biological testing of these selected compound analogs are under observation.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.