proteins STRUCTURE O FUNCTION O BIOINFORMATICS Defining and characterizing protein surface using alpha shapes Laurent-Philippe Albou, 1 y Benjamin Schwarz, 1,2 y Olivier Poch, 1 * Jean Marie Wurtz, 1 and Dino Moras 1 1 Department of Biology and Structural Genomics, IGBMC, CNRS, INSERM, ULP, Ilkirch, France 2 LSIIT UMR 7005 CNRS, Universite ´ de Strasbourg, Strasbourg, France INTRODUCTION ABSTRACT The biological function of a protein essentially relies on The alpha shape of a molecule is a geometrical representation its interactions with solvent and other biomolecules. that provides a unique surface decomposition and a means to Chemical and structural diversity observed at molecular filter atomic contacts. We used it to revisit and unify the defi- surfaces allow for the wide variety of interactions necessary nition and computation of surface residues, contiguous for cellular life. To decipher biological processes, it is thus patches, and curvature. These descriptors are evaluated and crucial to accurately define the nature and shape of these compared with former approaches on 85 proteins for which surfaces. The determination of the surface in terms of both bound and unbound forms are available. Based on the atoms, residues, and surface patches has already allowed to local density of interactions, the detection of surface residues shows a sensibility of 98%, whereas preserving a well-formed conduct numerous studies in the protein–protein interac- tion fields1–3 as well as to develop several prediction algo- protein core. A novel conception of surface patch is defined by traveling along the surface from a central residue or atom. rithms for the detection of binding sites and the modeling of complexes.4–7 More detailed characterization of the By construction, all surface patches are contiguous and, there- fore, allows to cope with common problems of wrong and surface in terms of clefts and knobs (respectively concav- nonselection of neighbors. In the case of protein-binding site ities and convexities) was also used for the study of inter- prediction, this new definition has improved the signal-to- face complementarity and docking of molecules.8–11 noise ratio by 2.6 times compared with a widely used Amongst the methodologies used for the description of approach. With most common approaches, the computation the surface of a molecule, the alpha shape theory12 is of surface curvature can be locally biased by the presence of probably one of the most promising. The alpha shape subsurface cavities and local variations of atomic densities. A model of a molecule is a polyhedral representation that novel notion of surface curvature is specifically developed to uniquely decomposes the space occupied by its atoms and avoid such bias and is parametrizable to emphasize either local or global features. It defines a molecular landscape com- retains interesting characteristics such as the shape of the posed on average of 38% knobs and 62% clefts where interact- molecule and a notion of interatom neighborhood. De- ing residues (IR) are 30% more frequent in knobs. A statisti- spite the relative complexity of the theory, alpha shapes cal analysis shows that residues in knobs are more charged, have been used to address a wide variety of problems in less hydrophobic and less aromatic than residues in clefts. IR structural biology, such as the computation of protein sur- in knobs are, however, much more hydrophobic and aromatic face and volume13 as well as their derivatives,14 the detec- and less charged than noninteracting residues (non-IR) in tion of pockets in known structures,15–17 the construc- knobs. Furthermore, IR are shown to be more accessible than tion of molecular surface meshes,18,19 the validation of non-IR both in clefts and knobs. The use of the alpha shape structures,20,21 or the study of interfaces.22,23 as a unifying framework allows for formal definitions, and fast and robust computations desirable in large-scale projects. Additional Supporting Information may be found in the online version of this This swiftness is not achieved to the detriment of quality, as article. proven by valid improvements compared with former Grant sponsors: The Centre National de la Recherche Scientifique (CNRS), the approaches. In addition, our approach is general enough to be ´ et de la Recherche Me ´dicale (INSERM), Structural Pro- Institut National de la Sante teomics in Europe (SPINE2-Complexes, CEE FP7 LSHG-CT-2006-031220), the applied on nucleic acids and any other biomolecules. ´crypthon program initiated by the Association Franc De ¸aise contre les Myopathies (AFM, CAMI 12727), IBM, the Ligue Nationale contre le Cancer, comite ´ du Haut- Proteins 2009; 76:1–12. Rhin and the Universite ´ Louis Pasteur de Strasbourg (ULP). C 2008 Wiley-Liss, Inc. V V y Laurent-Philippe Albou and Benjamin Schwarz contributed equally to this work. *Correspondence to: Olivier Poch, 1 rue Laurent Fries, BP 10142, 67404 Illkirch Key words: surface; alpha shape; patch; curvature; binding CEDEX, France. E-mail: poch@igbmc.fr Received 26 June 2008; Revised 30 September 2008; Accepted 1 October 2008 site; interaction; knob; cleft; structural bioinformatics; com- Published online 21 October 2008 in Wiley InterScience (www.interscience.wiley. putational biology. com). DOI: 10.1002/prot.22301 PROTEINS 1 C 2008 W ILEY- L ISS, I NC. V V
Recommend
More recommend
