Full text
| |
| Author(s): |
Antonio Marinho da Silva Neto
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Carlos. |
| Institution: | Universidade de São Paulo (USP). Instituto de Física de São Carlos (IFSC/BT) |
| Defense date: | 2017-12-19 |
| Examining board members: |
Rinaldo Wander Montalvão;
Jorge Chahine;
Kaline Rabelo Coutinho;
Luiz Carlos Gomide Freitas;
João Renato Carvalho Muniz
|
| Advisor: | Glaucius Oliva; Rinaldo Wander Montalvão |
| Abstract | |
One of the major challenges of modern structural biology is how to deal with protein flexibility. Besides the experimental difficulties, a relatively overlooked theoretical challenge is the lack of a proper mathematical language to represent proteín conformational space. The most popular representations have severe limitations, which reflects on the difficulties associated with conformational ensemble analyses. However, differential geometry (GD) and information theory (TI) can help to overcome such difficulties and were not well explored in this context. Here we investigate the usage of DG and TI as a mathematical representation of protein conformational space applied to the analyses of conformational ensembles. The DG descriptors calculation consists of representing protein backbone as a spatial curve and describes it by its curvature, κ, and torsion, τ . Based on those values, the distance between conformation and flexibility measurements were defined and a clustering algorithm was applied to identify conformational states. For the application of TI, a coding system for DG descriptors was developed to express each conformation as a sequence of finite symbols. Based on those sequences, information measurements associated to a residue, Rres, and to a conformation, Rconf , were defined. To investigate its efficacy, the proposed method was applied to conformation ensembles of three test systems: 1) Ubiquitin, 2) E1-DBD of HPV18 and 3) the steps of c-Myb-KIX binding. The DG analyses show equally good or superior performance when compared with popular methods on all tested system. In addition, the methods are especially useful to monitoring helix stability and analyses of very flexible proteins (or regions), since avoids the necessity of superposing structures. The values of Rconf are useful to compare different steps of a folding process and residues near regions involved in binding events tend to present higher values of Rres. However, those residues importance is uncertain and further studies are necessary to determinate if and how those can contribute to protein function. Nevertheless, the information measurements were informative on the comparison of compare conformational states and allow to formulate a testable hypothesis. On the other hand, the GD representation is computationally convenient, intuitive and avoid most of the limitations of the popular method applied to conformational ensemble analyses. (AU) | |
| FAPESP's process: | 13/18398-0 - Determination of Protein-Protein and Protein-Ligand Complexes from Sparse Experimental Data |
| Grantee: | Antonio Marinho da Silva Neto |
| Support Opportunities: | Scholarships in Brazil - Doctorate |