Our proteomics studies, currently focused on E. coli, are based on using mass spectrometry to identify proteins that copurify through multidimensional native-state chromatographic separations. Such proteins are candidates for the subunits in heterotypic complexes. This approach generates hundreds of candidate interactions, and simultaneously detects expression for about 10-15% of the E. coli proteins that likely to be expressed under a given physiological condition. We are working to improve our coverage of the proteome, and to apply our proteomics analysis to cells under different conditions, to examine how changes in physiology are correlated with changes in what proteins are expressed, and, moreover, changes in the subunit composition of complexes.

We are currently focusing on two models for studying the basis for the specificity of protein-protein interactions. The first is dimer formation by leucine zippers. Leucine zippers are short alpha-helical coiled-coils and are used as a dimerization motif in many eukaryotic transcription factors. In addition, on the order of 10% of all proteins encoded in most genomes are predicted to have zipper-like coiled coils. Our second model for interaction specificity is the LysR-type Transcriptional Regulators (LTTRs). LTTRs are the most abundant family of bacterial transcription factors, and face a specificity problem similar to the one faced by leucine zippers. Our approach to this important question is based on a combination of computational methods to predict protein-protein contacts, and experimental methods to test the importance of different contacts in the interaction surfaces. Understanding how leucine zippers and LTTRs interact to form homodimers and heterodimers and why these interactions are specific will help us understand not only fundamental principles of protein architecture, but also how these proteins function as genetic and developmental switch components. Ultimately, we hope to understand the molecular basis for recognition specificity and design proteins with novel properties.

We are also beginning to work on how to apply our experimental approaches to study the localization and topology of membrane proteins. A question of particular interest is whether membrane proteins are uniformly distributed over the surfaces of the inner and outer membranes of E. coli. A few examples are known where specific proteins are localized to the poles or septa of E. coli. We are examining whether membrane vesicles prepared from E. coli contain subpopulations containing different proteins. We are also exploring high throughput methods to determine which parts of proteins are exposed to the inner and outer surfaces of the membranes.