An Automatable Analytical Algorithm for Structure-Based Protein Functional Annotation via Detection of Specific Ligand 3D Binding Sites: Application to ATP (ser/thr Protein Kinases) and GTP (Small Ras-type G-Proteins) Binding Sites

TL;DR

This paper introduces an automated, scalable 3D motif-based algorithm for identifying ligand-specific binding sites in proteins, enabling functional annotation with high specificity and sensitivity, applicable to various protein families.

Contribution

The study presents a novel 3D motif detection method that does not rely on structural alignment, improving accuracy in ligand-binding site prediction and functional annotation.

Findings

Nearly 100% specificity in identifying binding sites.

Sensitivity over 60% for kinase family and 93% for G-proteins.

Effective discrimination between GTP/GTP-like and ATP/ATP-like ligands.

Abstract

We have developed an analytical, ligand-specific and scalable algorithm that detects a "signature" of the 3D binding site of a given ligand in a protein 3D structure. The said signature is a 3D motif in the form of an irregular tetrahedron whose vertices represent the backbone or side-chain centroids of the amino acid residues at the binding site that physically interact with the bound ligand atoms. The motif is determined from a set of solved training structures, all of which bind the ligand. Just as alignment of linear amino acid sequences enables one to determine consensus sequences in proteins, the present method allows the determination of three-dimensional consensus structures or "motifs" in folded proteins. Although such is accomplished by the present method not by alignment of 3D protein structures or parts thereof (e.g., alignment of ligand atoms from different structures) but by nearest-neighbor analysis of ligand atoms in protein-bound forms, the same effect, and thus the same goal, is achieved. We have applied our method to the prediction of GTP- and ATP-binding protein families, namely, the small Ras-type G-protein and ser/thr protein kinase families. Validation tests reveal that the specificity of our method is nearly 100% for both protein families, and a sensitivity of greater than 60% for the ser/thr protein kinase family and approx. 93% for the small, Ras-type G-protein family. Further tests reveal that our algorithm can distinguish effectively between GTP and GTP-like ligands, and between ATP- and ATP-like ligands. The method was applied to a set of predicted (by 123D threading) protein structures from the slime mold (D. dictyostelium) proteome, with promising results.