Docking-based identification of small-molecule binding sites at protein-protein interfaces

Abstract

Protein-protein interactions play an essential role in many biological processes, and their perturbation is a major cause of disease. The use of small molecules to modulate them is attracting increased attention, but protein interfaces generally do not have clear cavities for binding small compounds. A proposed strategy is to target interface hot-spot residues, but their identification through computational approaches usually require the complex structure, which is not often available. In this context, pyDock energy-based docking and scoring can predict hot-spots on the unbound proteins, thus not requiring the complex structure. Here, we have devised a new strategy to detect protein-protein inhibitor binding sites, based on the integration of molecular dynamics for the generation of transient cavities, and docking-based interface hot-spot prediction for the selection of the suitable cavities. This integrative approach has been validated on a test set formed by protein-protein complexes with known inhibitors for which complete structural data of unbound molecules and complexes is available. The results show that local conformational sampling with short molecular dynamics can generate transient cavities similar to the known inhibitor binding sites, and that docking simulations can identify the best cavities with similar predictive accuracy as when knowing the real interface. In a few cases, these predicted pockets are shown to be suitable for protein-ligand docking. The proposed strategy will be useful for many protein-protein complexes for which there is no available structure, as long as the the unbound proteins do not deviate dramatically from the bound conformations.

Keywords: Cavity detection; Drug discovery; Interface hot-spot residues; Modulation of protein–protein interactions; Molecular dynamics; Protein docking simulations.

Graphical abstract

Fig. 1

Prediction of protein–ligand and protein-PPI inhibitor pockets by Fpocket. (A) Performance of fpocket on predicting known protein–ligand (DUD-e database) and protein-PPI inhibitor (TIMBAL/2P2I databases) pockets, considering the best-rank (orange) or 3 best-ranked (blue) predictions. (B) Positive predicted value (PPV) and coverage (COV) of best-scoring predicted pockets on DUD-e (left) and TIMBAL/2P2I (right) databases. The plot shows the best-scoring predicted pocket for each case, with those considered a hit by Pocket Picker Criterion (PPC) represented as * symbol. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Predicted pockets in IL-2 protein using different conformational states. IL-2 protein is shown in grey surface, with IL-2R partner protein in blue ribbon, and the inhibitor with the best IC50 in green. (A) Predicted pockets (best-scoring in orange, the other one in yellow) on the unbound IL-2. (B) 10 best-scoring pockets predicted on MD side-chain conformers generated from unbound IL-2 (best-scoring pocket in orange, the others in yellow). (C) 10 best-scoring pockets from MD, restricted to the known protein–protein interface (best-scoring pocket in orange, the others in yellow). (D) 10 best-scoring pockets from MD, containing ≥ 3 predicted hot-spots (best-scoring pocket in orange, the others in yellow). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Docking-based hot-spot predictions. Protein residues are colored by NIP value, resulting from the docking calculations on the unbound proteins known to bind PPI inhibitors. For comparison, the partner protein is shown in green ribbon. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Assessment of the identification of PPI inhibitor pockets by integrating MD simulations and docking-based hot-spot predictions. PPV and coverage of the best-scoring predicted pocket on the MD-based side-chain conformers, which contain at least three predicted hot-spots.

Fig. 5

Docking of inhibitor on the predicted cavities of IL-2. IL-2 is shown in grey surface, and the different binding modes of the inhibitor are shown as ball and sticks. (A) IL-2 bound to FRH inhibitor (PDB ID: 1PY2). (B) Best-scoring docking model by Glide, using the 1st ranked pocket from MD and hot-spot prediction (from a MD snapshot at 2.241 ns). (C) The closest docking model to the reference in terms of ligand RMSD, obtained with the 9th ranked pocket from MD and hot-spot prediction (from a MD snapshot at 2.261 ns).