Protein Engineering, Vol. 15, No. 7, 603-610,
July 2002
© 2002 Oxford University Press
Exploring routes to stabilize a cationic pyridoxamine in an artificial transaminase: site-directed mutagenesis versus synthetic cofactors
1 Departments of Chemistry and 2 Biochemistry, University of Minnesota, Minneapolis, MN 55455, USA
Two artificial transaminases were assembled by linking a pyridoxamine derivative within an engineered fatty acid binding protein. The goal of mimicking a native transamination site by stabilizing a cationic pyridoxamine ring system was approached using two different strategies. First, the scaffold of intestinal fatty acid binding protein (IFABP) was tailored by molecular modeling and site-directed mutagenesis to position a carboxylate group close to the pyridine nitrogen of the cofactor. When these IFABP mutants (IFABP-V60C/L38K/E93E and -V60C/E51K/E93E) proved to be unstable, a second approach was explored. By N-methylation of the pyridoxamine, a cationic cofactor was created and tethered to Cys60 of IFABP-V60C/L38K and -V60C/E51K; this latter strategy had the effect of permanently installing a positive charge on the cofactor. These chemogenetic assemblies catalyze the transamination between 
-ketoglutarate and various amino acids with enantioselectivities of up to 96% ee. The pH profile of the initial rates is bell shaped and similar to native aminotransferases. The kcat values and the turnover numbers for these new constructs are the highest achieved to date in our system. This success was only made possible by the unique flexibility of the underlying enzyme design concept employed, which permits full control of both the protein scaffold and the catalytically active group.