Computational analysis on protein binding and interactions-a water exclusion perspective

• Main Author: Liu, Qian
• Other Authors: Li Jinyan
• Format: Theses and Dissertations
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer applications::Computer-aided engineering
• Online Access: https://hdl.handle.net/10356/50692
• Institution: Nanyang Technological University

Protein binding plays a fundamental role in biological processes and molecular functions. In this thesis, we introduce a `water exclusion' perspective and a new contact definition to conduct structural analysis on protein binding at both residue and atomic levels. In particular, our analysis is concentrated on binding hot spots in interfaces. We first propose the hypothesis of `double water exclusion' (DWE) to model binding hot spots. This hypothesis is intended to refine the seminal O-ring theory, an influential proposition stating that binding hot spots at protein interfaces are often surrounded by a ring of energetically less important residues. Beside this ring topology, DWE also describes the organizational topology of the ring-inside, energetically more important hot spot residues; this topology is not specified by the O-ring theory. Furthermore, DWE strengthens the `hot region' principle and the `coupling' proposition of binding hot spots. At a residue level, the requirements for a cluster of residues to form a hot spot under the DWE hypothesis can be mathematically satisfied by a biclique subgraph of residues in a pair of interacting proteins. In our evaluation according to ASEdb (a database storing hot spots determined by wet-lab experiments), we found that these DWE bicliques are rich in true hot spot residues. Further, we suggest a Z-score method to assess the biological significance of DWE bicliques and binding hot spots. This Z-score notion is designed to capture the probability of a contact residue occurring in or contributing to protein binding interfaces. Z-score is integrated by the water exclusion hypothesis in DWE and uses crystal packing as the reference state; crystal packing is first used in pairwise potential calculation. Our evaluation shows that Z-score outperforms earlier methods for classifying binding hot spots from non-hot spots. To study physicochemical properties of DWE hot spots and bicliques, we dissect DWE bicliques on a non-redundant dataset with 867 interactions with three different types of protein interactions. Our investigation on these DWE bicliques suggests that biological interactions have unique DWE bicliques compared to crystal packing, indicating the specificity of biological binding behaviors. Further, the unique DWE bicliques to the obligate and to non-obligate interactions also illustrate the different binding behaviors of these two types of interactions. Specifically, residues and their pairs in these DWE bicliques are found to have diverse preference for these three types of interactions.