Modular Design and Selection of Phosphospecific Tetratricopeptide Repeat Affinity Proteins

Protein phosphorylation is one of the most abundant protein post-translational modifications (PTMs), allowing cells to rapidly adapt to an ever-changing extracellular environment. Since its discovery over a century ago, research efforts have revealed a complex but elegant picture of protein phosphorylation events organized into finely-tuned networks. These efforts also elucidated how perturbations to phosphorylation networks lead to aberrant signaling associated with diseases such as Alzheimer's disease, neurodegeneration, and cancer.;At the crux of many such studies is the ability to observe and quantify specific phosphorylation events in complex cellular samples. Of the various technologies developed to quantify phosphorylation, phosphospecific affinity reagents have the advantage of sequence-specific phosphorylation detection in complex samples without the need for sophisticated instrumentation or purification. Phosphospecific antibodies have become increasingly popular for detecting specific phosphorylation in vitro with many modifications to standard antibody immunization and selection strategies. However, antibody selection is inherently random and restricted in live-cell applications by the instability and sensitivity of antibodies to the reducing cell environment. To provide antibody alternatives, a few research groups have recently pursued the design of phosphospecific affinity reagents using other protein scaffolds.;In this dissertation, we describe the development and application of a new method to design and select phosphospecific affinity reagents that are functional in living cells. Three core elements are featured in our strategy: 1) the tetratricopeptide repeat affinity protein (TRAP) as a design scaffold, 2) a new fluorescence-based selection strategy using intracellular detection of phosphoprotein-protein interactions, and 3) a modular approach toward re-design TRAP phosphopeptide binding specificity.;The TRAP scaffold is a native protein-protein interaction (PPI) domain found in all three domains of life. Unlike antibodies, TRAPs are small, stable, and functional in bacteria, yeast, and mammalian cells. Furthermore, TRAP peptide binding specificity can be finely-tuned or thoroughly re-designed using rational design, statistical analysis of homologous TRAP sequences, and selection from large combinatorial TRAP libraries. Here we first describe a simple TRAP redesign to bind a model phosphopeptide using a straightforward charge complementation approach. This TRAP-phosphopeptide interaction served as an experimental benchmark for further studies.;Next we developed a more general strategy to select for TRAPs that specifically interact with any phosphopeptide of interest. Paramount to this goal is a high-throughput strategy to select for phosphopeptide-binding TRAPs from combinatorial TRAP libraries. We developed a fluorescence-based strategy in E. coli using split mCherry enhanced reassembly and fluorescence (SMERF) that we coupled to fluorescence-activated cell sorting (FACS). Using this strategy, we identified phosphopeptide-binding TRAPs from combinatorial TRAP libraries using several phosphopeptide targets to show the generality of the approach.;Finally, we leveraged our SMERF technique with modular TRAP design to create TRAPs that bind a human protein kinase D2 (PKD2) phosphopeptide target. Our modular design strategy was motivated by the fact that selection of protein-specific PPI domains is currently limited by the need to perform an independent selection for each new target protein. Instead, we hypothesized that the TRAP-peptide interface could be subdivided into independent TRAP binding pockets that each interact with specific peptide residues. Thus, we designed libraries. where a single pocket is randomized and used SMERF to select for binding pockets complementing peptide mutations. We then combinatorially mixed-andmatched the selected binding pockets on individual TRAPs, coining these . proteins MoUSeTRAPs (modular union of selected TRAPs). Using SMERF, we showed that our best MoUSeTRAP interacts with the PKD2 peptide specifically.;Thus, we have shown that TRAP peptide binding specificity can be extensively re-designed for binding to a variety of phosphopeptide targets, including a PKD2 peptide required for kinase auto-activation. In the process, we developed a general strategy for observing phosphoprotein-protein interactions in E. coli that can also be extended to other PPIs involving non-canonical amino acids. We also show that TRAPs can be designed in a modular fashion to bind a peptide target. This advantage can be capitalized upon for the rapid design of TRAPs that interact with a large population of peptide targets. Thus, the TRAP design scaffold is versatile for the design of reagents for specific detection and quantification of various proteins of interest.