New Strategies To Reveal Protein Candidates In Protein-Protein Interactome Study

Meng-Chieh Chen. M.S., Purdue University, December 2014. New Strategies to Reveal Protein Candidates in Protein-Protein Interactome Study. Major Professor: Weiguo A. Tao. Comprehensive protein-protein interaction network analysis can help reveal protein functions in a system-wide manner. A reliable knowledgebase of interaction networks is not only important for selecting the candidates for drug therapies, but also for evaluating the disease risk. In current interaction databases, 322579 interactions comprised of 56460 proteins have been reported (statistical analysis from APID: Agile Protein Interaction DataAnalyzer; http://bioinfow.dep.usal.es/apid/index.htm). The huge datasets are contributed mainly by yeast -two -hybrid (Y2H) screening and affinity-purification followed by mass spectrometry (AP-MS). High false positive rates and failing to cover certain interaction categories are the limitations of these two methods. Here, we developed two proteomic strategies to address the limitations. First, we attempt to expand the dynamic of protein abundance in peptides mixtures by identifying phosphorylated peptides. The phosphorylation is a ubiquitous post-translation modification which is involved in regulating the signaling pathway and protein-protein interaction. The step of phosphopeptide enrichment after immunoprecipitation was employed as the second purification to reduce the complexity of the samples. With this approach, the sample complexity was significantly reduced because most of the unmodified peptides were removed after the phosphopeptide enrichment. In this simplified sample, the number of many high abundant proteins and peptides identified, such as ribosomal proteins, was significantly reduced, resulting in improved identification of Plk1 interacting proteins. We also applied this approach to the PRMT5 network, in which few phosphorylation