Interpretation of biphasic protein modification and modification-induced enzyme inactivation reaction plots

Protein engineering of a bacterial N-acyl-d-glucosamine 2-epimerase for improved stability under process conditions

Enzymatic cascade reactions, i.e. the combination of several enzyme reactions in one pot without isolation of intermediates, have great potential for the establishment of sustainable chemical processes. However, many cascade reactions suffer from cross-inhibitions and enzyme inactivation by components of the reaction system. This study focuses on the two-step enzymatic synthesis of N-acetylneuraminic acid (Neu5Ac) using an N-acyl-d-glucosamine 2-epimerase from Anabaena variabilis ATCC 29413 (AvaAGE) in combination with an N-acetylneuraminate lyase (NAL) from Escherichia coli. AvaAGE epimerizes N-acetyl-d-glucosamine (GlcNAc) to N-acetyl-d-mannosamine (ManNAc), which then reacts with pyruvate in a NAL-catalyzed aldol condensation to form Neu5Ac. However, AvaAGE is inactivated by high pyruvate concentrations, which are used to push the NAL reaction toward the product side. A biphasic inactivation was observed in the presence of 50-800mM pyruvate resulting in activity losses of the AvaAGE of up to 60% within the first hour. Site-directed mutagenesis revealed that pyruvate modifies one of the four lysine residues in the ATP-binding site of AvaAGE. Because ATP is an allosteric activator of the epimerase and the binding of the nucleotide is crucial for its catalytic properties, saturation mutagenesis at position K160 was performed to identify the most compatible amino acid exchanges. The best variants, K160I, K160N and K160L, showed no inactivation by pyruvate, but significantly impaired kinetic parameters. For example, depending on the mutant, the turnover number kcat was reduced by 51-68% compared with the wild-type enzyme. A mechanistic model of the Neu5Ac synthesis was established, which can be used to select the AvaAGE variant that is most favorable for a given process condition. The results show that mechanistic models can greatly facilitate the choice of the right enzyme for an enzymatic cascade reaction with multiple cross-inhibitions and inactivation phenomena.