A new protein binding pocket similarity measure based on comparison of clouds of atoms in 3D: application to ligand prediction

Abstract

Background: Predicting which molecules can bind to a given binding site of a protein with known 3D structure is important to decipher the protein function, and useful in drug design. A classical assumption in structural biology is that proteins with similar 3D structures have related molecular functions, and therefore may bind similar ligands. However, proteins that do not display any overall sequence or structure similarity may also bind similar ligands if they contain similar binding sites. Quantitatively assessing the similarity between binding sites may therefore be useful to propose new ligands for a given pocket, based on those known for similar pockets.

Results: We propose a new method to quantify the similarity between binding pockets, and explore its relevance for ligand prediction. We represent each pocket by a cloud of atoms, and assess the similarity between two pockets by aligning their atoms in the 3D space and comparing the resulting configurations with a convolution kernel. Pocket alignment and comparison is possible even when the corresponding proteins share no sequence or overall structure similarities. In order to predict ligands for a given target pocket, we compare it to an ensemble of pockets with known ligands to identify the most similar pockets. We discuss two criteria to evaluate the performance of a binding pocket similarity measure in the context of ligand prediction, namely, area under ROC curve (AUC scores) and classification based scores. We show that the latter is better suited to evaluate the methods with respect to ligand prediction, and demonstrate the relevance of our new binding site similarity compared to existing similarity measures.

Conclusions: This study demonstrates the relevance of the proposed method to identify ligands binding to known binding pockets. We also provide a new benchmark for future work in this field. The new method and the benchmark are available at http://cbio.ensmp.fr/paris/.

Figure 1

AUC score versus classification error as an evaluation of binding pocket similarity measure. Red circles represents pockets fixing ligand L₁, blue squares represents pockets fixing ligand L₂. The AUC score does not reflect the fact of good pocket clusterization, while the classification error does.

Figure 2

Relationship between AUC performances of the methods tested. (a) on the Kahraman dataset (b) on the Homogenous dataset. Each node corresponds to a particular method, parent nodes perform significantly better than child nodes according to the Wilcoxon signed-rank test.

Figure 3

Superposition of the binding pockets of two structurally different proteins binding ATP. A) overall structures of pdb: PDB: 1E8X in grey and PDB: 1DV2 in red superposed according to their binding sites using Sup-CK. ATP molecules are represented in blue. B) Superposition of the ATP molecules from PDB: 1DV2 and PDB: 1E8X when their binding sites are superposed. C) Positively charged protein regions around ATP molecules of PDB: 1E8X in grey and PDB: 1DV2 in red. D) Protein hydrophobic patches around ATP molecules of PDB: 1E8X in grey and PDB: 1DV2 in red.

Figure 4

Alignment two ATP binding pockets. Alignment of two ATP pockets made by sup-CK, atoms of each pockets are represented by blue and red points, two ATP ligands are traced in licorice.

Figure 5

Projection of the ext-KD dataset on the first two kernel principal components defined by the similarity measure sup-CK. Clustering of binding pockets according to their ligands, which illustrates the performance of this method for ligand prediction.

Figure 6

Performance on the HD dataset. (a) Mean AUC score and prediction error as functions of σ in the sup-CK method (pure geometrical version, λ = ∞), (b) mean AUC score and (c) classification error as functions of σ and λ when information on atoms partial charges is used.

Figure 7

Classification error of the sup-CK_L algorithm. as a function of R and σ (λ = 0.25): (a) Kahraman dataset, (b) HD dataset.

Figure 8

Projection of ATP binding pockets on the two first kernel principal components of sup-CK. Repartition of the ATP binding pockets generated by the sup-CK similarity measure on the extended Kahraman dataset. Red squares represent ligases, blue stars represent transferases.