ProBiS-ligands: a web server for prediction of ligands by examination of protein binding sites

Abstract

The ProBiS-ligands web server predicts binding of ligands to a protein structure. Starting with a protein structure or binding site, ProBiS-ligands first identifies template proteins in the Protein Data Bank that share similar binding sites. Based on the superimpositions of the query protein and the similar binding sites found, the server then transposes the ligand structures from those sites to the query protein. Such ligand prediction supports many activities, e.g. drug repurposing. The ProBiS-ligands web server, an extension of the ProBiS web server, is open and free to all users at http://probis.cmm.ki.si/ligands.

Figure 1.

Ligand prediction by the ProBiS-ligands server, starting from a query protein structure (light green).

Figure 2.

ProBiS-ligands output page. Left: query protein (green cartoon model) and two predicted ligands (CPK colored stick models). Invariant binding sites residues are thinner CPK-colored sticks. Right: table with predicted small-molecule ligands clustered according to their predicted location on the query protein and transposed from different binding sites; the two selected ligands are highlighted.

Figure 3.

Predicted protein–ligand complexes. Query proteins are green cartoon models and invariant binding site residues are CPK-colored stick models (Ligand View 1). (A) Three predicted ion ligand clusters (ions are spheres) on Glyoxalase family protein (PDB ID: 2qqz). (B) Predicted protein ligand (yellow–pink cartoon) on SH2 domain protein (3tkz). (C) Two predicted small molecule ligands, i.e. ATP and an inhibitor of biotin carboxylase (thick CPK sticks) on D-alanine:D-alanine ligase (1iov). (D) Predicted DNA ligand on endonuclease IV protein (4hno).