Identification Of Protein Structures:From Native Folding To Induced Folding

Generally a protein can spontaneouly fold into its well-defined three-dimensional structure from random coil when translated from mRNA to a linear chain of amini acids. The resulting folded shape is known as protein native state(or native structures) and the physical process of tertiary structure formation is known as native folding. A correctly folded form is typically required for a protein to perform its specific functions. However, there are increasing evidences suggesting that many proteins lack the ordered or stable 3D structures as a whole or in part and yet possess significant functions among all kingdoms of life. These proteins are called intrinsically disordered proteins or intrinsically disordered regions(IDPs or IDRs) and those having stable structures under native state are called ordered proteins correspondingly. Although IDPs or IDRs do not fold in isolation, they often undergo disorder-to-order transition upon binding interactions. This process is termed the coupled folding and bindng or induced folding. The main part of the current study includes four sections:(1) we first presented an integrated score function SVR_CAF to identify protein native structures among decoys(which are native-like modes generated by computers). The discriminative ability of SVR_CAF was evaluated by several decoys sets and it showed comparative performance in comparison with the state-of-art score functions.(2) Next, we collected a set of Pfam domain sequences that are predicted to be completely disordered and yet form structures in Protein Data Bank(PDB). We predicted the disorder for all the seeds and structural sequences of those domains and found significant diversity of disordered prediction among the sequences for the same domains. We also divided the structures into groups by their likely machinsms of structural formations, including disulfide bonds, ion corrdination and partner binding. For the proteins have both monomer and dimer structures, and proteins bound to several different parnters, we both observed structural changes upon complex formation and partners binding.(3) After that, we paid more detailed attentions to the partner-binding group and named them disordered binding domains. We identified three types of interactions for those disordered binding domains: DNA binding, disordered partner binding and ordered partner binding. Each type was discussed by gaving examples.(4) Finally, instead of disordered prediction on sequences, we predicted structural disorder based on surface area and interface area on a non-redundant set of PDB files and identified more potential IDPs or IDRs in the context of complexes. Furthermore, we presented several examples to demonstrate the role of intrinsic disorder in the regulations of protein-protein interactions and also showed the diversity of binding interactions for the complexes including three different chains. Overall, we identified native folding structures from decoy sets for ordered proteins and identified induced folding structures for candidate disordered proteins or regions. These two idenitfications undergo “order-to-disorder transition” and repsent an important complementation to the traditional sequence-structure-function paradigm.