Biosensor-based monitoring of the central metabolic pathway metabolites

In vivo biosensors have a wide range of applications, from the detection of metabolites to the regulation of metabolic networks, providing a versatile tool for cell biology and metabolic engineering. However, compared with the vast array of small molecules present in nature, the existing range of biosensors is far from sufficient. Here we describe the use of human hypoxanthine guanine phosphoribosyltransferase (HGPRT) as a ligand binding domain (LBD) protein, that acts by ligand-dependent stabilization, to build a biosensor for detection of the pentose phosphate pathway metabolite 5-phospho-α-D-ribose 1-diphosphate (PRPP). Using this protein as a template, we computationally redesigned a new pocket de novo according to the pose of the ligand, creating a binding mode exclusive to recognize another pentose phosphate metabolite, D-erythrose 4-phosphate (E4P), and glycerate-3-phosphate (3PG), from the glycolysis pathway. Furthermore, E4P biosensor was developed by fluorescence-activated cell sorting (FACS) and application of it enabled successful screening for the highest phenylalanine-producing strain reported to date. This work provides a strategy for computational design and development of biosensors for a broad range of molecules.