CREATION OF AN NADP-DEPENDENT PYRUVATE-DEHYDROGENASE MULTIENZYME COMPLEX BY PROTEIN ENGINEERING

Systematic replacement of a set of amino acids in the betaalphabeta-fold of the NAD-binding domain of Escherichia coli dihydrolipoamide dehydrogenase has been used to convert its coenzyme specificity from NAD to NADP. After comparison with the homologous enzyme glutathione reductase, Glu 203 was replaced with a valine residue, thereby eliminating the potential to form hydrogen bonds with the 2'- and 3'-OH groups of the adenine ribose in NAD. Similarly, Met 204, Pro 210, Phe 205, and Asp 206 were replaced by an arginine, an arginine, a lysine, and a histidine residue, respectively, to provide a nest of positive charge to accommodate the 2'-phosphate group of the incoming NADP. In addition, Gly 185 and Gly 189 in the betaalphabeta motif were replaced with alanine residues to facilitate the positioning of the newly introduced Val 203 by allowing a flip of the peptide bond between residues Gly 180 and Gly 181. Wild-type dihydrolipoamide dehydrogenase is inactive with NADP, but the mutant enzyme displayed high levels of activity with this coenzyme, the values of K(m), k(cat), and k(cat)/K(m) comparing favorably with those found for the wild-type enzyme operating with NAD. The mutant enzyme was also capable of assembly in vitro to form an active pyruvate dehydrogenase multienzyme complex, the coenzyme specificity of which reflected that of its dihydrolipoamide dehydrogenase component. These experiments should make it possible now to study the effects in vivo of requiring a crucial catabolic enzyme to function with the wrong coenzyme, an important extension of protein engineering into the living cell.