Orchestrated Nanozyme/Natural Enzyme Combination Strategy for a Machine Learning-Powered Biosensor Array toward High-Through-put Screening of Antidiabetic Herbs

Abstract Screening for α-glucosidase (α-GLU) inhibitors is vital for the discovery of antidiabetic drugs, and traditional Chinese herbal medicines (TCMs) are promising antidiabetic agents. However, traditional high-throughput screening strategies are labor intensive. The integration of nanozymes with machine learning provides prospects for convenient and highly efficient screening strategies. To advance this status, a dual-active MnO2-decorated Ce metal−organic framework nanozyme (MnPCe) with outstanding oxidase-like (OXD-like) and laccase-like (LAC-like) activity was synthesized for the first time. After the conjugation of MnO2 on CeMOF, the OXD-like and LAC-like activities of MnPCe increased. Inspired by the cascade reaction and signal complementarity between the dual-active MnPCe nanozyme and natural α-GLU, a machine learning-assisted triple-channel biosensor array for α-GLU inhibitors was constructed based on the combination of the nanozyme and natural enzyme catalytic channels. The three colorimetric signals were integrated into a unique fingerprint pattern, enabling the sensor array to accurately discriminate nine common α-GLU inhibitors. It also successfully distinguished TCMs containing different types of α-GLU inhibitors, highlighting its potential for high-throughput screening of antidiabetic agents from complex herbal matrices. It inspires the innovation of multifunctional nanozymes, the development of high-throughput screening of promising drugs, and the extraction of valuable compounds from herbal medicines.