A protein 3D reduced representation wherein each amino acid is represented by two centroids: that of the backbone atoms (N, Ca, C’, O) and that of the sidechain atoms (Cß and beyond)
Published in Chapter:
Visualization of Protein 3D Structures in ‘Double-Centroid’ Reduced Representation: Application to Ligand Binding Site Modeling and Screening
Vicente M. Reyes (School of Life Sciences, College of Science, Rochester Institute of Technology, Rochester, USA) and Vrunda Sheth (School of Life Sciences, College of Science, Rochester Institute of Technology, Rochester, USA)
Copyright: © 2011
|Pages: 16
DOI: 10.4018/978-1-60960-491-2.ch026
Abstract
This article is of two parts: (a) the development of a protein reduced representation and its implementation in a Web server; and (b) the use of the reduced protein representation in the modeling of the binding site of a given ligand and the screening for the model in other protein 3D structures. Current methods of reduced protein 3D structure representation such as the Ca trace method not only lack essential molecular detail, but also ignore the chemical properties of the component amino acid side chains. This chapter describes a reduced protein 3D structure representation called “double-centroid reduced representation” and presents a visualization tool called the “DCRR Web Server” that graphically displays a protein 3D structure in DCRR along with non-covalent intra- and intermolecular hydrogen bonding and van der Waals interactions. In the DCRR model, each amino acid residue is represented as two points: the centroid of the backbone atoms and that of the side chain atoms; in the visualization Web server, they and the non-bonded interactions are color-coded for easy identification. The visualization tool in this chapter is implemented in MATLAB and is the first for a reduced protein representation as well as one that simultaneously displays non-covalent interactions in the molecule. The DCRR model reduces the atomicity of the protein structure by ~75% while capturing the essential chemical properties of the component amino acids. The second half of this chapter describes the application of this reduced representation to the modeling and screening of ligand binding sites using a data model termed the “tetrahedral motif.” This type of ligand binding site modeling and screening presents a novel type of pharmacophore modeling and screening, one that depends on a reduced protein representation.