Harnessing bacterial periplasmic binding proteins as biosensors for on-line process analytical technologies

Abstract

Ligand-binding proteins have good to moderate binding affinity for various small molecules, usually in the millimolar to micromolar range. Many industrially and clinically relevant metabolites in cell culture also fall in these concentration ranges. Hence, naturally occurring ligand-binding proteins provide a good basis for further engineering efforts to construct biosensors. In this work, we deployed protein modelling tools to identify and design a library of binding proteins from various species to evaluate their suitability for use as biosensors. Candidates were selected based on extent of their known sequence and functional characterization in literature. We also streamlined our developmental pipeline for these biosensors by optimizing our in vitro screening and characterization assays. Stages in the protein purification, fluorophore-conjugation, dialysis were fine-tuned to improve the overall efficiency and screening throughput. As a result, promising biosensors for the following metabolites, branched-chain amino acids (BCAA), calcium, glucose, glutamine, lactate, and phosphate, were obtained. Out of this suite of biosensors, the BCAA, glucose, and phosphate binding proteins exhibited desirable characteristics for out-of-the-box use as biosensors, while the other candidates would require further engineering to achieve the parameters needed, such as high signal-to-noise ratio, good detection range and sensitivity. Next, the glucose biosensor candidate was immobilized in a microcolumn packed with Ni-NTA resin and interfaced it with a miniature fluorometer, which in turn transmitted output to a LabVIEW GUI that collected and displayed the data in a time-resolved manner. A proof-of-concept was performed with bacteria cultures to measure and record glucose concentration in the media at regular intervals until the end of fermentation. We also carried out preliminary tests with the BCAA biosensor by measuring analytes in samples obtained from a time-course hybridoma cell culture experiment and validated the measurements with a commercial BCAA kit. From this study, we developed multiple protein-based biosensors of metabolic importance and demonstrated the integration of these sensors with a microfluidic device that records and displays data in real time. This establishes the viability of implementing protein biosensors in real-world applications such as bioprocess monitoring.