Proteomics in 3D: Development of new technology and computational tools for structural analysis

Abstract

Proteins form complexes with specific structures to carry out biological function. Better understanding of protein complex structures will allow scientists to gain deeper insights on biological function, pathway regulation and disease mechanisms. But characterization of protein interactions is challenging. High-throughput methods, such as yeast two-hybrid and co-immunoprecipitation, provide little structural information for the formed complexes. X-ray crystallography and NMR enable high resolution structure measurement, but those methods have very low through-put, and could only be used on purified complexes. Chemical cross-linking coupled with mass spectrometry is a high through-put structure characterization method. It has not been widely applied to complex biological systems due to the technical difficulties of sample preparation and data analysis. Protein Interaction Reporter (PIR) recently developed in Bruce lab in UW, has overcome some of the major technical difficulties in cross-linking technique. PIR has been proved to be applicable in multiple biological systems, and many novel structure measurements were made. Some of the sample preparation method improvements were covered in this dissertation. In addition to the advancements of sample preparation methods, a public database for cross-linking data, XLink-DB, was developed to accommodate the fast growing amount of data. XLink-DB is a data storage, analysis, and visualization platform. It has a web interface for user to upload and analyze their own results. XLink-DB automatically fetches related information from other public databases such as UniProt and PDB to help user analyze and visualize their results. It also contains a protein modeling and docking pipeline to generate model complex structures with cross-linking distance constraints. As the first data analysis and storage platform for cross-linking results, XLink-DB fills the need of a high through-put pipeline of data processing and visualization, greatly improves the efficiency of data analysis, and provides deeper insight into the data.