Construction of a highly selective and sensitive carbohydrate-detecting biosensor utilizing Computational Identification of Non-disruptive Conjugation sites (CINC) for flexible and streamlined biosensor design.

Fluorescently-labeled solute-binding proteins that alter their fluorescence output in response to ligand binding have been utilized as biosensors for a variety of applications. Coupling protein ligand binding to altered fluorescence output often requires trial and error-based testing of both multiple labeling positions and fluorophores to produce a functional biosensor with the desired properties. This approach is laborious and can lead to reduced ligand binding affinity or altered ligand specificity. Here we report the Computational Identification of Non-disruptive Conjugation sites (CINC) for streamlined identification of fluorophore conjugation sites. By exploiting the structural dynamics properties of proteins, CINC identifies positions where conjugation of a fluorophore results in a fluorescence change upon ligand binding without disrupting protein function. We show that a CINC-developed maltooligosaccharide (MOS)-detecting biosensor is capable of rapid (kon = 20 μM-1s-1), sensitive (sub-μM KD) and selective MOS detection. The MOS-detecting biosensor is modular with respect to the spectroscopic properties and demonstrates portability to detecting MOS released via α-amylase-catalyzed depolymerization of starch using both a stopped-flow and a microplate reader assay. Our MOS-detecting biosensor represents a first-in-class probe whose design was guided by changes in localized dynamics of individual amino acid positions, supporting expansion of the CINC pipeline as an indispensable tool for a wide range of protein engineering applications.