Engineering the substrate binding site of benzoylformate decarboxylase.

Benzoylformate decarboxylase (BFDC) and pyruvate decarboxylase (PDC) are both thiamin diphosphate-dependent enzymes. The two share a common three-dimensional structure and catalyze a similar chemical reaction, i.e., decarboxylation of 2-keto acids. However, they vary significantly in their substrate utilization pattern. In particular, BFDC has extremely limited activity with pyruvate, while PDC has no activity with benzoylformate. Here we report our progress, using a semirandom approach, toward converting BFDC into an efficient pyruvate decarboxylase. From the structure of BFDC in complex with R-mandelate, 12 residues within a 5 A radius from the inhibitor molecule were selected and subjected individually to site-saturation mutagenesis. Each variant was screened for its ability to decarboxylate five different substrates, i.e., benzoylformate, 2-ketohexanoate, 2-ketopentanoate, 2-ketobutanoate, and pyruvate. The first round of mutagenesis showed that changes in Thr377 and Ala460 resulted in an altered substrate spectrum which included higher activity toward pyruvate. Two variants, T377L and A460Y, were selected as the starting point of a second round of site-saturation mutagenesis. In both cases, the T377L-A460Y double mutant proved to be the only new variant with significantly improved catalytic activity toward pyruvate. When compared to the wild-type enzyme, based on k(cat)/K(m) values, the T377L-A460Y variant showed an 11000-fold improvement in the ratio between pyruvate and benzoylformate utilization. This double mutant displays a K(m) value for pyruvate of 2 mM as well as a k(cat)/K(m) value for pyruvate which is only 70-fold lower than that of Zymomonas mobilis PDC.