2011 OSU Molecular Life Sciences
Interdisciplinary Graduate Programs Symposium
Talk abstracts
Abstract:
The marginal stability of natural proteins presents challenges in disease elucidation as well as production and administration of protein drugs. Currently, there are no reliable methods to predict the thermodynamic consequence of even a single mutation; therefore, the state of the art for stabilizing proteins is immature. In the post-genomic era the application of sequence statistics is an attractive candidate for designing proteins and further understanding the sequence-structure relationship. Our lab has developed and empirically validated the roles of conservation and statistical correlation in the triosephosphate isomerase family.
Consensus mutation-that is, mutation to the most common amino acid at a position in a multiple sequence alignment-increases the stability of native proteins half the time. To date, there is no a priori method to predict which consensus mutations will be stabilizing. First, we have developed a double-sieve statistical filter that selects stabilizing consensus mutations with >90 % accuracy. We have used this algorithm to engineer a mesophilic triosephosphate isomerase into a thermophile with 15 consensus mutations.
Secondly, we have designed and engineered a fully-consensus TIM from a raw and curated dataset. These two TIMs differ at only a handful of nonconserved positions, but differ dramatically in their activities and biophysical characterization. The first triosephosphate isomerase is molten globular, monomeric, yet weakly active. The second enzyme is well folded, dimeric and wild-type active. These two proteins differ only in the completeness of their statistical correlations and networks.
Keywords: Protein Engineering, Sequence Statistics, Protein Stability