Biocatalysts are increasingly used in industry to create enantiomerically
pure compounds; routes employing dehydrogenases have been especially
successful. Glucose dehydrogenase belongs to the extensive superfamily
of SDRs (short chain dehydrogenase/reductase) and is currently a favored
enzyme for cofactor regeneration. Glucose is its only substrate at a
reasonable rate and the reaction is favored towards oxidation of glucose.
A thermostable glucose dehydrogenase, previously developed at GeorgiaTech, was submitted to several rounds of site-directed mutagenesis, supported by database analysis and structure-guided design, to create an alcohol dehydrogenase based on a glucose dehydrogenase scaffold with a markedly changed substrate specificity. Key to the identification of the necessary amino acid hot spots within the sequence was a bioinformatics approach that involved an extensive data base analysis of existing SDRs
using specific algorithms. The resulting 3-4 amino acid exchanges led to
a complete change of substrate specificity away from glucose but
respectable activity towards substrates such as cyclopentanol and
1,3-cyclohexandiol. By achieving the goal of de novo design of a glucose
dehydrogenase, we demonstrate the successful implementation of
bioinformatic tools and modeling that can accelerate development of novel
substrate specificity in an otherwise limited enzyme system.