(185f) Disulfide Stabilization of Noncovalent Quaternary Structures On the Yeast Surface

Proteins in nature often form permanent or transient complexes to achieve biological effects, including signal transduction, regulation, allostery, stability and genomic economy. Although the structural information is available at atomic resolution for many protein oligomers, the creation, evolution and current biological roles of protein quaternary structures are difficult to model. Whereas the contributions of individual residues in globular proteins are studied by evaluating the consequences of site directed mutations, similar studies are more difficult to perform for protein complexes. For example, assembling permanent protein complexes requires co-expressing and/or co-purifying the constituent subunits whereas analyzing transient complexes requires performing in vitro binding assays of individually purified subunits. The technical difficulty of characterizing protein complexes in high throughput is thus a formidable challenge when studying protein association. In this study, our goal is to develop a strategy to enable the reconstitution of a broad spectrum of quaternary structures for characterization regardless of the stability of the complex. To this end, we developed structure-based disulfide trapping to help express noncovalent protein complexes on the yeast surface. Whereas the instability of a native complex results in rapid dissociation of the subunits, the engineered disulfide can shift the equilibrium toward the bound complex for efficient display. The surface displayed complex is analyzed using flow cytometry, allowing high throughput assays to be developed based on disulfide trapping. To demonstrate that disulfide trapping is general and can be broadly applied, we used the technique to express three unstable quaternary structures: antibody variable domains, interleukin-8 homodimer and p53 peptide bound to MDM2. The expression level of the protein complex varies with the interaction strength as well as the position of the engineered disulfide. Using disulfide trapping, therefore, it is feasible to evaluate the effects of interfacial mutations, reconstitute protein complexes for functional characterization without lengthy biochemical purification, and target difficult protein surfaces to engineer peptide ligands with high specificity.