USING HIGH­‐THROUGHPUT METHODS TO CHART PROTEIN‐PROTEIN AND PROTEIN­‐LIPID INTERACTIONS IN EUKARYOTIC CELLS

Abstract

To truly understand the self-replicating eukaryotic cell we need to make significant progress unraveling the interactome, the sum of all the interactions between the proteins and the metabolites of the cell. Here, I present two projects to that end: The first study maps protein complexes in a unicellular thermophilic eukaryote, Chaetomium thermophilum, using an innovative approach integrating several biological techniques. Thermophilic proteins are, by their nature, more stable than their mesophilic counterparts and C. thermophilum has been described as a potential model organism for structural studies. We use a size exclusion chromatography (SEC) to separate high-molecular weight protein complexes from cell lysate. Coeleuting proteins are identified by mass spectrometry and inferred to be in a protein complex together. Chemical crosslinking combined with mass spectrometry (XL-MS) is applied to the SEC fractions to provide direct biochemical confirmation of the predicted protein-protein interactions. Together these methods have allowed us to identify protein complexes with novel subunits and also functionally related coeluting proteins not hitherto known to form protein complexes. Additionally, negative-stain electron microscope (EM) images of protein mixtures from the SEC fractions are correlated with the elution patterns of the identified complexes to distinguish the structural signatures (Shapes) of specific complexes. This enabled manual picking of particles from cryo-EM micrographs to solve the molecular structure of C. thermophilum fatty acid synthase (FAS) to 4.7A resolution directly from the SEC fractions without further purification. A novel binder of FAS was identified by EM and confirmed with XL-MS as a branching biotin dependent carboxylase, which together constitutes a potential metabolon for the production of branch chain fatty acids. This project facilitates the use of C. thermophilum as a model organism for structural biology and the methods may open the way for high-throughput structural biology. The second project used a high-content screen to discern the roles of lipid classes on the localization of proteins, protein complexes and biological processes in the cell. This approach attempts to shed light on this protein-metabolite network by genetically depleting selected lipid classes by perturbing their biosynthesis and using in vivo imaging to map localization changes of proteins in the cells and therefore identification of lipid dependent localizations. The database created in this project will facilitate the development of follow-up studies that will further discern the structural roles of lipids in the organization of the proteome.