Partition-Based Engineering of Glutamate Decarboxylase Unlocks Substrate Promiscuity for Synthesis of Nonprotein Amino Acids

Glutamate decarboxylase (GAD) is a highly specific pyridoxal 5′-phosphate (PLP)-dependent enzyme widely used in the biosynthesis of gamma-aminobutyric acid (GABA), renowned for its high activity, stability, and conversion rate. However, its strict substrate specificity for glutamate confines its application solely to GABA production, and the extremely narrow substrate scope severely limits its potential in synthesizing other valuable amines from diamino acids (e.g., NH2–x–COOH). Here, we report the first successful engineering of selective substrate promiscuity toward structurally similar acidic amino acids in GAD from Escherichia coli (EcGadB). Using a partition-based engineering (PBE) strategy, we simultaneously engineered two key regions: the γ-carboxyl binding site (T62) for substrate affinity and the PLP-binding site (T212) for catalytic efficiency. This yielded double mutants M4D and M4X, which efficiently decarboxylate the non-native substrates l-aspartate and L-2-aminoadipic acid into β-alanine and 5-aminovaleric acid, respectively, while retaining excellent pH adaptability and thermostability. An in vitro crude enzyme system achieved high titers (up to 38 g/L) and near-quantitative conversion (>97%) for 5-aminovaleric acid, outperforming whole-cell catalysis. Molecular dynamics and free energy calculations revealed the mechanistic basis for altered substrate specificity. Our work provides efficient biocatalysts for amine synthesis and a generalizable framework for engineering PLP-dependent enzymes.