Substrate Recognition by the Phenylalanine- Adenylating Domain of Gramicidin Synthetase, and Redesign of Nonribosomal Peptide Synthetases by Modulation of Substrate Specificity

Non-ribosomal peptide synthetases (NRPS) are a family of enzymes that assemble a variety of pharmacologically interesting polypeptides from canonical and non-canonical amino acids. The identity and connectivity of the monomers in the final product are directly determined by the order of domains in the enzyme that are specific for the recognition and incorporation of a particular amino acid. Here we use K*, an ensemble-based, statistical mechanics–derived approximation to the binding constant, to predict mutations to the phenylalanine adenylation domain (PheA) of gramcidin synthetase (GrsA) that will improve binding of a miscognate amino acid, either leucine or tyrosine. PheA mutants predicted by K* to have improved binding of leucince or tyrosine have been made and demonstrate preferred binding of the targeted amino acid over phenylalanine. The catalytic specificity (kcat/KM) of mutants was also evaluated by steady-state kinetics, and improvement for targeted substrate was observed, though not enough to become the catalytically preferred substrate. To better understand the lack of congruency between experimentally determined dissociation constants and steady-state kinetics data, stop-flow tryptophan quenching experiments were done to determine pre-steady-state association and dissociation rates. Stop-flow experiments revealed a previously unknown PheA binding event that is specific for phenylalanine and nucleotide dependent. Additional stop-flow experiments show that PheA binds phenylalanine differently in the presence of ATP and AMP.