To ensure proper folding, cells have evolved a sophisticated and essential machinery of proteins called molecular chaperones that assist the folding of newly made polypeptides. The importance of proper protein folding is underscored by the fact that a number of diseases, including Alzheimer's and those involving infectious proteins (prions), result from protein-misfolding events. My research focuses on identifying and understanding the machinery necessary for efficient folding, as well as studying the mechanism and consequences of protein misfolding. We are also developing experimental and analytical approaches for exploring the organizational principles of complex biological systems.
Mechanism of Amyloid Formation and Propagation
In contrast to the more frequently observed disordered aggregates, some proteins form ordered aggregates, termed amyloid fibrils, that accumulate in a number of human diseases. Our goal is to provide a mechanistic understanding of how amyloid fibers propagate, to elucidate how molecular chaperones control amyloid growth in vivo, and to determine the role of amyloid formation in both disease and normal physiology. Efforts to understand amyloid propagation have been hampered by the lack of a facile genetic or biochemical system for studying this process. This situation has improved greatly with the finding that the [URE3] and [PSI+] states of yeast result from the prion-like aggregation of endogenous proteins. My laboratory has taken advantage of the PSI phenomenon to identify and characterize properties of the protein Sup35p that allow it to propagate a β-sheet–rich prion form. These efforts have been greatly aided by our development of strategies that let us create distinct synthetic prion conformations of Sup35p (the protein determinant of [PSI+]) and introduce them into yeast. This has provided the first demonstration of the "protein-only" hypothesis, identified amyloid as the infectious conformation, and enabled mechanistic investigations into several of the most perplexing features of prion biology in terms of the structural and biophysical properties of amyloids and cellular factors that act on such forms. These features include the ubiquitous presence of "species barriers," which inhibit transmission between even closely related prion proteins, and prion strains wherein infectious particles composed of the same protein give rise to a range of different heritable prion states.
Protein Folding in the Endoplasmic Reticulum
The ER is a highly specialized folding compartment responsible for the structural maturation of all secreted proteins. The long-term goal of our ER studies is to understand how the ER maintains an environment that allows efficient folding. Secreted proteins typically depend on disulfide bonds to fold. To support disulfide formation, the ER must closely regulate its redox potential. Although the chemical requirement for an oxidizing environment has been long appreciated, prior to our work how this was accomplished was unknown. We established that a specific cellular redox machinery is required for disulfide formation and identified Ero1, a novel and ubiquitous protein, as an essential component of this machinery. More recently, my lab has reconstituted Ero1-mediated oxidative folding from pure components. This work established that Ero1 is a novel, conserved FAD-dependent enzyme and defined the in vivo pathway of disulfide formation: Ero1 is oxidized by molecular oxygen and in turn acts as a specific oxidant of protein disulfide isomerase (PDI), which then directly oxidizes disulfide bonds in folding proteins. While providing a robust driving force for disulfide formation, the use of molecular oxygen as the terminal electron acceptor can lead to oxidative stress through the production of reactive oxygen species and oxidized glutathione. How Ero1 distinguishes between the many different PDI-related proteins and how the cell minimizes the effects of oxidative damage from Ero1 remain important open questions.
To obtain a more global view of how efficient folding is achieved in the ER, we have exploited DNA microarray "chips" to examine the transcriptional response to accumulation of unfolded proteins in the ER. Our studies dramatically changed the view of the unfolded protein response (UPR). Rather than primarily up-regulating molecular chaperones, the UPR results in a concerted reorganization of the ER, altering polypeptide flux from the cytosol to ER, retrograde export from the ER, transport of proteins from the ER to the Golgi, and retrieval of ER proteins from the Golgi. We are extending our studies of the regulation of protein folding in the ER to metazoan systems by combining traditional cell biological and biochemical techniques with genomic approaches, including microarrays and comprehensive RNA interference libraries.
Understanding the Organizational Principles of Cellular Systems
Comprehensive description of the yeast proteome. A central challenge of the postgenomic era is to identify and quantitate an organism's complete set of expressed proteins. To facilitate such efforts, my laboratory, in collaboration with Erin O'Shea's group (HHMI, University of California, San Francisco), has produced and characterized two comprehensive yeast libraries in which we attach a carboxyl-terminal tag to each gene in its native chromosomal context. The collections consist of ~6,000 yeast strains, each expressing a single protein fused to either the tandem affinity purification (TAP) tag, which provides an extremely high affinity epitope tag, or to the green fluorescent protein (GFP), which allows us to monitor protein localization. These libraries let us follow the expression level and localization of the yeast proteome with unprecedented completeness. In addition, these libraries and the data generated from them enable our efforts, described below, to provide comprehensive strategies for dissecting complex biological systems.
Single-cell proteomics. Sixty years ago Erwin Schrödinger pointed out that many biological processes depend on molecules that exist at low copy number per cell and thus are subject to stochastic noise. More recent studies establish that biological responses (e.g., oocyte maturation or yeast nutrient responses) often exhibit switch-like behavior that is only evident in single-cell analyses. Yet, historically, mRNA and protein measurements are done on bulk samples because of the technical obstacles in obtaining single-cell data. Consequently, many questions concerning the scope and control of stochastic noise, or the nature of signaling events, remain largely unaddressed.
We have developed a strategy using the yeast GFP library and flow cytometry to measure protein abundance at the single-cell level. With this system, we can measure more than 4,000 strains per day, collecting about 50,000 single-cell measurements per sample. We have provided three key validations of this approach. First, GFP fluorescence is proportional to protein abundance measured by Western blotting. Second, measurement errors are far lower than encountered with other strategies. Third, variation (noise) in abundance detected using cytometry is proportional to variation measured using the more traditional technique of microscopy.
The precision and nature of our high-throughput cytometry measurements enable the exploration of a range of biological questions that were previously inaccessible, including the global analysis of changes in protein abundances in response to changes in cellular states, as well as the origin and consequences of differences in noise levels in different classes of proteins.
Construction and analysis of high-density genetic interaction maps. In addition to a description of the molecular activity of individual proteins, a comprehensive description of a cell requires an understanding of how groups of proteins act together to carry out specific biological processes, and how those different processes communicate among themselves. Theoretical considerations suggest that information regarding each of these organizational levels can be obtained from the analysis of genetic interaction maps (pairwise descriptions of the extent to which the loss of one gene will aggravate or buffer the effect of the loss of a second one). The spectrum of genetic interactions is also crucial for analyzing the manifestation of quantitative traits, for guiding efforts to tailor drug treatments to an individual's genetic makeup, and for developing rational strategies for drug-combination therapies. Finally, central features of evolutionary biology, including maintenance of genetic variability, sexual reproduction, and speciation, depend on the structure of genetic interactions.
Prior to our work, however, there was no effective strategy to construct such maps, and few analytical tools available to extract information from them even if they existed. We have developed an integrated set of experimental and computational strategies that now make it possible to measure and analyze high-density genetic interaction maps. We also developed a robust approach, termed DAmP (decreased abundance by mRNA perturbation), for creating hypomorphic alleles in yeast, which allowed us to explore the interactions of essential genes and now makes the essential part of the yeast genome accessible to the comprehensive chemical sensitivity, genetic, and functional studies that have revolutionized the analysis of nonessential genes. The application of this strategy to several fundamental processes, including the yeast early secretory pathway, demonstrated the immediate value of such maps. These studies revealed or clarified the role of many proteins.
At a broader level, the genetic interaction maps delineate pathway organization and components of physical complexes and illustrate the interconnection between the various secretory processes. The wealth of biological information yielded by these studies, alongside the potential of tailoring genetic interaction maps to a wide variety of problems, argues that this strategy will provide a general framework for obtaining a holistic understanding of complex biological systems.