Characterizing the co-evolution of protein-protein interactions.

摘要

Protein-protein interactions play crucial roles in many biological systems by controlling various processes involved in metabolic, signaling, and regulatory pathways. Due to the wealth of experimental information collected by recent genome sequencing efforts, computational techniques have become increasingly useful in analyzing large datasets to provide inferences about protein function. In this thesis, several computational approaches are presented that can be utilized to characterize and identify protein-protein interactions. In Chapter 1, a novel approach is developed for quantifying the co-evolution between two protein families that interact. This technique is applied to chemokines and their receptors to show that they co-evolve. Using this analysis, inferences about the binding partners for proteins with uncharacterized binding specificities can be made. Chapter 2 demonstrates how the co-evolutionary analysis can be used to study the putative co-evolution of chemokines and chemokine receptors of both human and viral origin. Identifying possible interactions between the human cellular and the human cytomegalovirus (CMV)-encoded chemokines and chemokine receptors and defining the complex interplay between these proteins can further our understanding of the mechanisms controlling virus trafficking and evasion of the human immune system. In Chapter 3, the co-evolutionary analysis is applied to study possible domain-domain interactions of the mildew resistance gene o (MLO) protein family. A high degree of co-evolution between domains could identify potential domain interactions and/or domains that share a common binding partner. In Chapter 4, the co-evolutionary analysis is extended to quantitate the degree of co-evolution between proteins that interact. This approach allows for fast and objective identification of possible protein-protein interactions. Six systems of interacting protein families were studied---the syntaxin/unc-18 protein families, the adrenergic receptors and their G-alpha subunits, the TGF-beta proteins and their receptors, the colicin/immunity protein families, the chemokines and their receptors, and the VEGF proteins and their receptors. From this analysis, inferences about the interaction partners for proteins of uncharacterized binding specificities in the TGF-beta and syntaxin families were made.