Poster abstracts

Poster number 14 submitted by Brandon Sullivan

Sequence statistics and biophysic to probe the Protein Folding Problem

Brandon J. Sullivan (Ohio State Biochemistry Program, The Ohio State University), Venuka Durani (Department of Chemistry, The Ohio State University), Thomas J. Magliery (Departments of Chemistry and Biochemistry, The Ohio State University)

Abstract:
Proteins are responsible for many of life’s critical processes. As such, their folding, stability and fidelity must be exquisitely controlled to maintain homeostasis. Mutations that disrupt this balance give rise to many diseases. These illnesses are increasingly difficult to prevent and treat because the mechanism of protein folding and the details of protein stability are poorly understood. Computational and mutagenic approaches towards these topics are powerful, but currently lack accuracy and throughput, respectively. Nature, on the hand, has sampled sufficient sequence space and lends itself as the perfect database for studying the inverse folding problem. Here, we describe bioinformatic approaches to understand two levels of information encoded within sequences; conservation and correlation. First, we have designed and characterized a consensus enzyme from the multiple sequence alignment of triosephosphate isomerase. This enzyme is molten globular and weakly active despite its low sequence identity and rare monomeric structure. A second variant of this protein was constructed to repair the dimer interface, but global dynamics, not the interface sequence, prohibit oligomers in both proteins. A third consensus TIM was designed using a curated database and differs at only 36 nonconserved positions. This protein is well folded, wild-type active and dimeric which further stresses the importance of curation in bioinformatic endeavors. Second, we have applied relative entropy calculations to improve our ability to select for stabilizing consensus mutations in wild-type proteins. Finally, we have developed two mathematical tests to determine how residues statistically interact pairwise and within networks. These pairs and networks will be mutated into consensus and wild-type hosts to elucidate the kinetic and thermodynamic roles of statistical correlation within proteins.

Keywords: Protein Engineeering, Sequence Stastics, TIM-barrel