Implementation of The Double-Centroid Reduced Representation of Proteins and its Application to the Prediction of Ligand Binding Sites and Protein-Protein Interaction Partners Using FORTRAN 77/90 Language

TL;DR

This paper details the implementation of a reduced protein structure representation and its application to predicting ligand binding sites and protein-protein interaction partners using FORTRAN codes, validated on specific protein complexes.

Contribution

It introduces a FORTRAN-based method for transforming protein structures into a reduced form and applies it to predict ligand binding sites and interaction partners, demonstrating its efficacy.

Findings

Successful construction of 3D search motifs for GTP and sialic acid.

Effective screening for potential receptor proteins and protomers.

Validation on known protein complexes shows promising results.

Abstract

Transformation of protein 3D structures from their all-atom representation (AAR) to the double-centroid reduced representation (DCRR) is a prerequisite to the implementation of both the tetrahedral three-dimensional search motif (3D SM) method for predicting specific ligand binding sites (LBS) in proteins, and the 3D interface search motif tetrahedral pair (3D ISMTP) method for predicting binary protein-protein interaction (PPI) partners (Reyes, V.M., 2015a & c, 2015b, 2009a, b & c). Here we describe results demonstrating the efficacy of the set of FORTRAN 77 and 90 source codes used in the transformation from AAR to DCRR and the implementation of the 3D SM and 3D ISMTP methods. Precisely, we show here the construction of the 3D SM for the biologically important ligands, GTP and sialic acid, from a training set composed of experimentally solved structures of proteins complexed with the pertinent ligand, and their subsequent use in the screening for potential receptor proteins of the two ligands. We also show here the construction of the 3D ISMTP for the binary complexes, RAC:P67PHOX and KAP:phospho-CDK2, from a training set composed of the experimentally solved complexes, and their subsequent use in the screening for potential protomers of the two complexes. The 15 FORTRAN programs used in the AAR to DCRR transformation and the implementation of the two said methods are: get_bbn.f, get_sdc.f, res2cm_bbn.f, res2cm_sdc.f, nrst_ngbr.f, find_Hbonds.f, find_VDWints.f, find_clusters.f90, find_trees.f90, find_edgenodes.f90, match_nodes.f, fpBS.f90, Gen_Chain_Separ.f, remove_H_atoms.f and resd_num_reduct.f. Two flowcharts - one showing how to implement the tetrahedral 3D SM method to find LBSs in proteins, and another how to implement the 3D ISMTP method to find binary PPI partners - are presented in our two companion papers (Fig. 2, Reyes, V.M., 2015a, Fig. 1 & 2, Reyes, V.M., 2015c).