Protein-ligand binding site recognition using complementary binding-specific substructure comparison and sequence profile alignment

Abstract

Motivation: Identification of protein-ligand binding sites is critical to protein function annotation and drug discovery. However, there is no method that could generate optimal binding site prediction for different protein types. Combination of complementary predictions is probably the most reliable solution to the problem.

Results: We develop two new methods, one based on binding-specific substructure comparison (TM-SITE) and another on sequence profile alignment (S-SITE), for complementary binding site predictions. The methods are tested on a set of 500 non-redundant proteins harboring 814 natural, drug-like and metal ion molecules. Starting from low-resolution protein structure predictions, the methods successfully recognize >51% of binding residues with average Matthews correlation coefficient (MCC) significantly higher (with P-value <10(-9) in student t-test) than other state-of-the-art methods, including COFACTOR, FINDSITE and ConCavity. When combining TM-SITE and S-SITE with other structure-based programs, a consensus approach (COACH) can increase MCC by 15% over the best individual predictions. COACH was examined in the recent community-wide COMEO experiment and consistently ranked as the best method in last 22 individual datasets with the Area Under the Curve score 22.5% higher than the second best method. These data demonstrate a new robust approach to protein-ligand binding site recognition, which is ready for genome-wide structure-based function annotations.

Availability: http://zhanglab.ccmb.med.umich.edu/COACH/

Fig. 1.

Flowchart of (A) TM-SITE and (B) S-SITE for protein–ligand binding site prediction

Fig. 2.

An illustrative example of TM-SITE binding site prediction on the phosphatidylinositol 4,5-bisphosphate 3-kinase protein. (A) I-TASSER model with TM-score = 0.74 and global/local-binding RMSD = 19.8/1.5 Å. (B) Binding pocket identification by ConCavity (green mesh). (C) Query SSFL definition based on predicted binding pockets for the query protein (green cartoon) which has a RMSD to the native 3.8 Å. (D) Template SSFLs recognized by TM-align. (E) Superposition of all template SSFLs on the query structure. (F) Final model of the predicted binding sites: native/predicted ligands are shown in magenta/blue sticks; true/false positive binding sites are highlighted in green/red ball-sticks

Fig. 3.

Summary of blind LBS predictions in the COMEO experiment during last 22 weeks. The mean AUC was computed over all targets tested in the corresponding weeks

Fig. 4.

Illustrative examples of successful predictions by COACH in the CAMEO. The receptor structures are shown in gray cartoon buried in transparent surface. The native and predicted ligands are in magenta and orange colors, respectively. The true positive, false positive and false negative predictions of the ligand-binding residues by COACH are highlighted in green, red and blue sticks, respectively. (A) DNA binding ETS domain of FEV in complex with DNA. (B) beta'-COP/Insig-2 complex. (C) Glutathione transferase homolog from Lodderomyces elongisporus bound with CIT. (D) R39-imipenem Acyl-enzyme bound with Mg²⁺ ion