In silico analysis of the inhibitory activities of GABA derivatives on 4-aminobutyrate transaminase

2.1 Structure prediction

3D structure of human protein 4-aminobutyrate transaminase (from here-on referred to as target) is not known experimentally yet. For elucidation of the 3D structure of this protein, structure of pig 4-aminobutyrate transaminase (PDB ID: 1OHV; Storici et al., 2004; from here-on referred to as template) was used as a template with high sequence similarity. Protein structure prediction, carried out using Swiss-Model server (Arnold et al., 2006), was followed by model optimization and validation using Ramachandran plot (Kleywegt and Jones, 1996), Procheck, Errat (Colovos and Yeates, 1993), Verify3D (Luthy et al., 1992; Bowie et al., 1991) and WhatIF (Vriend, 1990) tools.

2.2 Ligand dataset

Yadav et al. reported a series of new γ-aminobutyric acid (GABA) derivatives obtained from 4-(1,3-dioxoisoindolin-2-yl) butanoic acid by coupling it with various substituted amines by using DCC as a coupling reagent, which was used in docking against the target and its template. Fig. 1 shows 2D structures of this series designated as compounds 4a–4m (Yadav et al., 2012).

Close

Figure 1

2D structures of GABA derivatives (4a–4m). (4a) 4-(1,3-Dioxoisoindolin-2-yl)-N-phenylbutanamide; (4b) 4-(1,3-Dioxoisoindolin-2-yl)-N-(4-fluorophenyl)butanamide; (4c) N-(4-Chlorophenyl)-4-(1,3-dioxoisoindolin-2-yl)butanamide; (4d) N-(4-Bromophenyl)-4-(1,3-dioxoisoindolin-2-yl)butanamide; (4e) 4-(1,3-Dioxoisoindolin-2-yl)-N-(4-iodophenyl)butanamide; (4f) 4-(1,3-Dioxoisoindolin-2-yl)-N-(4-nitrophenyl)butanamide; (4g) 4-(1,3-Dioxoisoindolin-2-yl)-N-p-tolylbutanamide; (4h) 4-(1,3-Dioxoisoindolin-2-yl)-N-(4-ethylphenyl)butanamide; (4i) 4-(1,3-Dioxoisoindolin-2-yl)-N-(4-methoxyphenyl)butanamide; (4j) N-(2,4-Dichlorophenyl)-4-(1,3-dioxoisoindolin-2-yl)butanamide; (4k) N-(2,4-Difluorophenyl)-4-(1,3-dioxoisoindolin-2-yl)butanamide; (4l) N-(4-Chloro-2-iodophenyl)-4-(1,3-dioxoisoindolin-2-yl)butanamide; (4m) N-(4-Chloro-2-methylphenyl)-4-(1,3-dioxoisoindolin-2-yl)butanamide.

Export to PPT

2.3 Molecular docking studies

Using the tool AutoDock 4.2 (Sanner, 1999), automated dockings were performed for finding the conformations of selected ligands (compounds) binding to the active site of 4-aminobutyrate transaminase. Polar hydrogen atoms were added to the receptor proteins and Kollman charges were alloted. Non-polar hydrogen atoms, in ligands, were merged and Gasteiger partial charges were assigned. Also, rotations were allowed for ligands’ bonds, and random orientations and torsions were used during the procedure. AutoGrid was then used to generate grid maps by locating an 80 × 80 × 80 Å3 grid around each receptor model, separated by 0.375 Å. The Lamarckian genetic search algorithm was used by setting the program to 100 runs, 150 population size, 2.5 × 10⁶ energy evaluations, 27,000 maximum iterations, 0.02 mutation rate, 0.80 crossover rate and an RMS tolerance of 1.0 Å for cluster analysis. The most occurring docked conformation with the lowest binding energy was selected for that particular ligand.

2.4 Binding residues

From PDB (ID: 1OHV), the active site residues of the template protein were extracted followed by inference of target protein residues by alignment of the two proteins.

Template: Gly136A, Asp298A, Thr353B, Gln301A, Ser137A, Arg192A, Val300A.

Target: Gly164, Asp326, Thr381, Gln329, Ser165, Arg220, Val328.

3.1 3D structure prediction of human 4-aminobutyrate transaminase

Crystal structure of pig 4-aminobutyrate-aminotransferase (PDB ID: 1OHV; resolution 2.30 Å) was used as template for the prediction of 3D model for human 4-aminobutyrate transaminase. MSA analysis indicated 80% sequence identity between target and template sequences and an RMSD of 0.2 Å between the two models. Fig. 2 shows the protein structures of template (Fig. 2a) and target protein (Fig. 2b) along with superimposed pose of both (Fig. 2c).

Close

Figure 2

3D structures of template and target proteins, and their superimposition. (a) Structure of pig 4-aminobutyrate-aminotransferase used as template; (b) Structure of human 4-aminobutyrate transaminase obtained through homology modeling; (c) template-target superimposition with an RMSD of 0.2 Å.

Export to PPT

3.2 Structure validation

Ramachandran plot indicated 99% residues in the allowed regions and none of the active site residues in the disallowed region. Moreover, parameters like peptide bond planarity, non-bonded interactions, Cα tetrahedral distortion, main chain H-bond energy and overall G factor for the predicted structure lie within the favorable range.

The homology models were further verified using Errat, Verify3D and WhatIF tools. Errat scored the overall quality factor for non-bonded atomic interactions very much above the accepted range giving a score of 91.3, which represent the high quality of our predicted model. Fig. 3 shows the structure validation by Ramachandran plot, Procheck and Errat programs.

Close

Figure 3

Structure validation of modeled protein, human ABAT. (a) Ramachandran plot showing 99% residues in the allowed region; (b) Procheck validation plot for main chain parameters; (c) Errat quality check with an overall quality factor of 91.372.

Export to PPT

Main chain parameters in terms of conformation of the predicted model were checked for deviations from the template using Procheck (Fig. 3b). Verify3D produced averaged data points with 96.97% residues exhibiting a score above 0.2, assuring a considerable high model quality. In the case of human 4-aminobutyrate transaminase, none of the scores for any residue given by WhatIF was found less than −5.0 indicating structure reliability. These data indicated that the predicted model is of good quality to allow further study.

3.3 Molecular dockings

Molecular docking studies were carried out with both target and template 4-aminobutyrate transaminase. All the compounds were found to strongly inhibit by completely occupying the active sites in the case of both the target protein, and its template model. All inhibitors showed low energy values as indicated in Table 1 for target and for template in Table 2. The inhibition constant values indicated the efficiency of the compounds to inhibit the protein, especially compounds 4f, 4h, and 4m in the case of target and 4c, 4d, 4f, 4g, 4h, and 4l in the case of the template. These exceptionally low Ki values and low energy values prove drug efficiency for these compounds.

For target protein, binding energy values range from −6.89 to −8.34 kcal/mol while for template protein, binding energy values range from −7.11 to −9.29 kcal/mol. Along with these, intermolecular energy values, vdW + Hbond + desolv energy values, electrostatic energy values, internal energy values and system's unbound energy also lie in a favorable range for both target and template as clearly indicated in Tables 1 and 2.

3.4 Binding mode of GABA derivatives

All GABA derivatives bind efficiently with binding sites of pig and human 4-aminobutyrate transaminase. Fig. 4a and b shows the binding mode of GABA derivatives with pig 4-aminobutyrate transaminase and human 4-aminobutyrate transaminase respectively. It is clearly shown that all compounds bind in the same cavity for both template and target.

Close

Figure 4

Binding of all GABA derivatives in (a) template binding site pig ABAT and (b) target binding site human ABAT. Ribbon represents the receptors while ligands are shown in sticks encapsulated in a mesh.

Export to PPT

Fig. 5 further shows the residues involved in binding with both template and target receptors. Fig. 5a shows that Gly136, Asp298, Thr353, Gln301, Ser137A and Arg192 residues of template are involved in interaction while Fig. 5b shows that Gly164, Asp326, Thr381, Gln329, Ser165, and Arg220 residues of target are involved in interactions.

Close

Figure 5

Residues involved in binding for (a) template and (b) target. All the ligands are shown in sticks enveloped by the mesh surface. Receptor residues are shown in sticks.

Export to PPT

Human ABAT interaction with the compounds 4a–4m was analyzed further in detail. Table 3 shows the binding residues’ information for compounds 4a–4m, we observed that all active site residues for Human ABAT were involved in docking with nearly all the compounds.

Strong inhibitor binding is also reflected by the frequency of hydrogen bonds as shown in Table 4 and Fig. 6. Compounds 4a, 4c, 4e and 4k each made three hydrogen bonds with two residues. Ser165, Arg220 and Thr381 of human ABAT protein were mostly seen in making hydrogen bonds with these compounds, however, Arg220 was the most frequently occurring residue in hydrogen bonding. The details of hydrogen bonds formed by the compounds with binding residues, atoms involved in the bonds and distance are given in Table 4.

Close

Figure 6

Interacting residues involved in docking of inhibitors at active site of human ABAT protein (H-atoms are colored gray, O-atoms red, N-atoms blue, I-atoms purple, C-atoms of inhibitors are colored green while C-atoms of receptor residues are colored yellow, and hydrogen bonds are shown with green dashed line labeled with distances in Å). (a–m) Compounds 4a–4m mode of interaction with the human ABAT protein binding site.

Export to PPT

Fig. 6 shows the interacting residues involved in docking of inhibitors at the active site of human ABAT protein. Compound 4a forms two hydrogen bonds with Arg220 (2.35 Å and 2.35 Å) and a single hydrogen bond with Thr381 (1.82 Å), while it interacts hydrophobically with Gly164, Ser165 and Val328 as shown in Fig. 6a. Compound 4b makes hydrogen bonds with Arg220 (1.78 Å and 1.95 Å) and forms hydrophobic interactions with Gly164, Ser165, Val328, Gln329 and Thr381 (Fig. 6b). Compound 4c forms hydrogen bonds with Arg220 (2.09 Å and 2.45 Å) and Thr381 (1.84 Å), while hydrophobic interactions are observed with Gly164, Ser165, Asp326 and Val328 (Fig. 6c). Compound 4d forms a single hydrogen bond with Arg220 (2.05 Å), while Gly164, Ser165, Asp326, Val328 and Thr381 are shown to form hydrophobic interactions as shown in Fig. 6d. Compound 4e shows hydrogen bonding with Ser165 (2.11 Å and 2.49 Å) and Thr381 (2.13 Å), and hydrophobic interactions with Gly164, Arg220, Val328 and Gln329 (Fig. 6e). Compound 4f is shown to be involved in hydrogen bonding with Arg220 (1.90 Å) and Thr381 (1.83 Å), and in hydrophobic interactions with Gly164, Ser165, and Gln329 (Fig. 6f). Compound 4g hydrogen bonds with Arg220 (2.07 Å) and forms hydrophobic interactions with Ser165, Asp326, Val328, Gln329 and Thr381 (Fig. 6g). Compound 4h is shown to form hydrogen bonds with Ser165 (2.03 Å) and Thr381 (1.91 Å) and hydrophobic interactions with Gly164, Arg220, Asp326, Val328 and Gln329 (Fig. 6h). Compound 4i forms a hydrogen bond with Ser165 (2.01 Å), and hydrophobic interactions with Gly164, Arg220, Asp326, Val328 and Thr381 (Fig. 6i). Compound 4j interacts by forming a hydrogen bond with Arg220 (1.79 Å), and hydrophobically with Ser165, Asp326, Gln329 and Thr381 as shown in Fig. 6j. Compound 4k makes hydrogen bonds with Ser165 (2.12 Å and 3.10 Å) and Thr381 (2.24 Å), and hydrophobic interactions with Gly164, Arg220, Val328 and Gln329 as shown in Fig. 6k. Compound 4l is shown to have two hydrogen bonds with Arg220 (2.01 Å and 2.26 Å), and hydrophobic interactions with Ser165, Asp326, Val328, Gln329 and Thr381 (Fig. 6l). Compound 4m forms hydrogen bonds with Arg220 (1.78 Å and 2.13 Å), while the residues Ser165, Asp326, Val328, Gln329 and Thr381 interact in a hydrophobic manner with the compound (Fig. 6m).