Figure 1.
Tetrapeptide backbone modifications.
Figure showing various modifications done to the N-terminal of the tetrapeptides for obtaining the peptidomimetics. In molecules 28 and 31, amide bonds are replaced with thioamide bonds between Glu-Thr and Thr-Ala of N-methylated ETAV, respectively.
Figure 2.
(A) Heatmap depicting the clustering of known inhibitors of PDZ domains on the basis of their similarities in chemical structures. (B) Dendrogram depicting the clustering of known inhibitors of PDZ domains. Chemical structures of representative inhibitors from each cluster are also shown.
Figure 3.
Differences between PDZ2 and PDZ3 of PSD-95.
(A) Local alignment of the sequences of 2nd (PDB ID: 1QLC) and 3rd (PDB ID: 1BE9) PDZ domains of PSD-95 protein by BLAST and the conservation of the binding pocket residues is highlighted. (B) Ribbon diagrams depicting the structures 1BE9 and 1QLC. Regions in both 1BE9 and 1QLC, which show alignment in Figure 3A, are colored in cyan, while the N- and C-terminus residues, which did not align, are depicted in magenta. The longer loop region in 1QLC corresponding to the insertion seen in the local alignment is colored in purple. The binding pocket residues made flexible during docking are shown in red color. (C) Schematic depiction of the grid box used for docking of ligands on both PDZ domains. The dimension are x = 38Å, y = 74Å & z = 34Å.
Figure 4.
Docking output of 3,4-dichlorophenylethylATAV onto the PDZ3 domain.
(A) Typical example of the output of docking done by VINA showing 9 top scoring docking solutions. The lower and upper bound RMSDs for the 9 binding poses have been calculated with respect to the conformation highlighted in orange. (B) Image showing all 9 docked conformations of one of the ligands (3,4-dichlorophenylethylATAV) under study on 1BE9. (C) 1BE9 with its native peptide (red colored) in cartoon representation. The docked conformation of ligand chosen for MD simulation is shown in stick representation.
Figure 5.
(A) RMSD vs time plot for the MD simulation on the docking complex involving the inhibitor 3,4-dichlorophenylethylATAV and 1BE9. (B) RMSD vs time plot for the MD simulation on the docking complex involving the inhibitor 3,4-dichlorophenylethylATAV and 1QLC.
Figure 6.
RMSD vs Time plot after excluding certain residues.
(A) RMSD vs time plot for one of the MD simulations (simulation 2 from Figure 5B) on the docking complex involving the inhibitor 3,4-dichlorophenylethylATAV and 1QLC. RMSD has been calculated after excluding residues 12–15 from N-terminal and loop regions comprising of residues 42–49 & 65–70. The inset picture shows the superposition of the docked complex before and after MD. (B) RMSD vs time plot for one of the MD simulations (simulation 1 from Figure 5A) on the docking complex involving the inhibitor 3,4-dichlorophenylethylATAV and 1BE9. RMSD has been calculated after excluding residues 1–10 from N-terminal and 100–115 from C-terminal of the protein. The inset picture shows the superposition of the docked complex before and after MD.
Figure 7.
The binding modes of the ligand 3,4-dichlorophenylethylATAV to the 2nd and 3rd PDZ domains of PSD95 as predicted by docking and MD simulations. The residues of the PDZ domains, which are in contact with the inhibitor for more than 50% of the simulation time during last 1ns of the 5ns MD simulations, are shown in RED color.
Table 1.
Table summarizing the list of inhibitors for which MD simulations performed in complex with PDZ2 (1QLC) and PZD3 (1BE9).
Table 2.
Table summarizing the ranking of the inhibitors according to their Ki values.
Figure 8.
Comparative analysis for determining binding selectivity of PDZ2 and PDZ3.
Bar graph showing the comparison of binding energy values determined experimentally as well as computed by using MM/PBSA after 5 ns of MD simulation for the set of 38 peptidomimetic ligands for both PDZ2 and PDZ3 domains. The ligand numbers given on x-axis is according to the numbering given in Table 1. The difference in binding energy values between PDZ2 and PDZ3 for all 38 ligands is plotted on y-axis.
Figure 9.
Correlation with experiments (PDZ2).
(A) Correlation between experimentally determined binding energy values and binding energy values calculated from MM/PBSA for the energy minimized docked complexes for 2nd PDZ domain of PSD-95 protein. (B) Correlation between experimentally determined binding energy values and binding energy values calculated from MM/PBSA after performing MD simulation on the docked complexes for 2nd PDZ domain of PSD-95 protein.
Figure 10.
Comparison before and after MD (PDZ2).
(A) Structural superimposition of docked complex of one of the ligands (PyrenbutylETAV) before and after MD simulation on 2nd PDZ domain of PSD-95 protein (PDBID: 1QLC) where the ligand is moving out of the binding pocket of 1QLC during the course of simulation. (B) Structural superimposition of docked complex of one of the ligands (3,4-dichlorophenylethylETDV) before and after MD simulation on 2nd PDZ domain of PSD-95 protein (PDBID: 1QLC) where the ligand is repositioning itself in the binding pocket of 1QLC during the course of simulation.
Figure 11.
Correlation with experiments (PDZ3).
(A) Correlation between experimentally determined binding energy values and binding energy values calculated from MM/PBSA for the minimized docked complexes for 3rd PDZ domain of PSD-95 protein. (B) Correlation between experimentally determined binding energy values and binding energy values calculated from MM/PBSA after performing MD simulation on the docked complexes for 3rd PDZ domain of PSD-95 protein.
Figure 12.
Comparison before and after MD (PDZ3).
(A) Structural superimposition of docked complex of one of the ligands (3,4-dichlorophenylethylATAV) before and after MD simulation on 3rd PDZ domain of PSD-95 protein (PDBID: 1BE9) where the ligand is moving out of the binding pocket of the 1BE9 during the course of simulation. (B) Structural superimposition of docked complex of one of the ligands (AnthracenethylETAV) before and after MD simulation on 3rd PDZ domain of PSD-95 protein (PDBID: 1BE9) where the ligand is repositioning itself in the binding pocket of the 1BE9 during the course of simulation.