Please use this identifier to cite or link to this item: http://hdl.handle.net/2445/99692
Title: Structural Modeling and Characterization of Protein Interactions of Biomedical Interest: The Challenge of Molecular Flexibility
Author: Pallara, Chiara
Director: Fernández-Recio, Juan
Keywords: Pèptids
Proteïnes
Macromolècules
Metabolisme
Peptides
Proteins
Macromolecules
Metabolism
Issue Date: 4-Feb-2016
Publisher: Universitat de Barcelona
Abstract: Las proteínas son grandes biomoléculas que desarrollan funciones esenciales en las células, muy a menudo mediante la formación de complejos altamente específicos con otras proteínas y biomoléculas. Por tanto, uno de los mayores retos científicos en la actualidad es el estudio completo a nivel estructural y energético de todas las interacciones entre proteínas de interés biológico y terapéutico. Sin embargo, la consideración precisa de la plasticidad de las proteínas en los métodos computacionales de modelado molecular no es trivial, debido a limitaciones tanto técnicas como metodológicas. En este contexto, el objetivo principal de esta tesis doctoral ha sido el desarrollo, aplicación y evaluación de herramientas computacionales para la caracterización estructural, energética y dinámica de las proteínas y sus interacciones. Para cumplir con estos objetivos, durante la primera parte de la tesis se ha llevado a cabo la revisión de varios protocolos computacionales para la caracterización de las superficies de interación entre proteínas. Los métodos analizados proporcionan unas predicciones razonablemente consistentes y fiables. También se ha llevado a cabo la evaluación de la eficacia predictiva de nuestro método pyDock en CAPRI, un experimento comunitario de evaluación de métodos de modelado estructural de complejos entre proteínas. En general, a pesar de los avances metodológicos en los protocolos de docking, el modelado eficaz de la plasticidad de las proteínas sigue siendo un reto importante en el campo. En base a los análisis anteriores, se ha llevado a cabo un estudio sistemático sobre la importancia de la heterogeneidad conformacional en el reconocimiento entre proteínas. Los resultados indican que los ensamblados conformacionales generados a partir de proteínas en solución contienen confórmeros con mejor complementariedad energética que la estructura cristalográfica de dichas proteínas y que favorecen su reconocimiento intermolecular. A partir de estos resultados, se ha propuesto un nuevo métodode docking que usa ensamblados conformacionales generados a partir de las proteínas en solución. Esta estrategia resulta particularmente efectiva en casos poco o medianamente flexibles. Finalmente, en la última parte de la tesis se ha llevado a cabo la aplicación de métodos computacionales al modelado de varios casos de interés biomédico. En conclusión, los avances metodológicos en cuanto al modelado de proteínas y sus interacciones, junto a la inclusión eficaz de la flexibilidad conformacional, permiten tener herramientas computacionales cada vez más útiles para complementar los datos experimentales y mejorar la comprensión de procesos biológicos relevantes
Proteins are large biomolecules that play essential functional and structural roles within cells and that typically act through their interaction with other proteins and biomolecules, forming highly specific functional complexes. Thus, a major biological challenge is to provide structural and energetics details for such interactions. In this context, computational methods can successfully contribute to predict and characterize the mechanistic aspects of protein function, in which conformational flexibility plays a major role. However, an accurate consideration of protein plasticity within computational modeling of protein function at molecular level is still far from trivial, mostly because of both technical and methodological limitations. Thus, the main purpose of this PhD thesis is the assessment, development and application of computational tools for the structural, energetic and dynamic characterization of protein molecules and their interactions. To fulfill these objectives, the first part of the thesis consists in the review and comparison of several existing computational protocols for the characterization of protein-protein interfaces, as well as in the evaluation and discussion of the performance of pyDock, the docking protocol previously developed in our group, as resulted from the last CAPRI (Critical Assessment of PRediction of Interactions) editions. These analyses confirm that current computational protocols aimed to model the phylogenetic, structural and energetic properties of residues within protein-protein interfaces show reasonably good predicting performance and consistency. However, as shown by CAPRI experiment, despite recent methodological advances in docking, dealing with protein plasticity is still a crucial bottle-neck. Based on these premises, the second part of the thesis is focused on understanding the role of protein conformational heterogeneity in protein-protein recognition. Subsequently, a novel protocol to integrate unbound conformational ensembles within a docking framework has been devised and systematically tested. The analysis of conformational heterogeneity in precomputed unbound ensembles reveals that docking encounters are favoured by improving the energetic complementarity of the docking partners rather than the geometrical similarity to the bound state. Moreover, the unbiased use of such ensembles is a successful strategy to incorporate flexibility into a docking approach for low-medium flexible cases, especially those that presumably follow a conformational selection mechanism. Finally the last part of the thesis consists in the application of computational methods to the modeling of protein interactions and the exhaustive exploration of their conformational space within different realistic contexts, thanks to the expertise previously acquired on the theoretical basis of protein interactions. Thus, the main lines of research include the energetic characterization of host-pathogen protein interactions (i.e., host GTPase Rab5 with pathogen phospholipase VipD), the ab-initio modeling of the encounter complex ensembles of redox proteins (i.e. PSI with alternative electron donors, cytochrome c6 and plastocyanin), and finally the description of the structural and dynamic basis of a protein kinase dysfunction under pathological conditions (i.e., MEK1 oncogenic and CFC-related mutants). The results confirmed that computational modeling can complement experimental data to improve the understanding of biological processes involving protein interactions and can help to rationalize and quantify the structural, energetic and dynamic effects of pathological mutations at molecular level.
URI: http://hdl.handle.net/2445/99692
Appears in Collections:Tesis Doctorals - Facultat - Farmàcia

Files in This Item:
File Description SizeFormat 
Chiara_Pallara_THESIS.pdf38.3 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.