Histidine-engineered Cu-BTC nanozyme with enhanced laccase-like activity combining the machine learning for precise recognition of Beta-lactam antibiotics.

Although nanozyme sensor arrays can simultaneously recognize multiple target substances, they are currently rarely used for identifying Beta-lactam antibiotics (BLs). This may be due to the lower catalytic performance of some nanozymes in practical applications, which further limits the detection performance of nanozyme sensor arrays. Therefore, developing highly active nanozymes is particularly important. Here, we introduced histidine during the preparation of Cu-1,3,5-benzenetricarboxylic acid (Cu-BTC) to obtain Cu-BTC@His nanozymes with high laccase-like (LAC) catalytic activity. Due to the unique physicochemical properties of BLs, they can inhibit the LAC activity of Cu-BTC@His, and the degree of inhibition increases with the increase of reaction time. A three-channel nanozyme sensor array was constructed based on reaction kinetics and applied to the discrimination of nine BLs. In addition, by optimizing multiple machine learning (ML) algorithms, the accuracy of the neglected concentration detection model constructed based on this array has been improved from 31.27 % to 95.92 %, which is beneficial for identifying unknown samples in real samples. This work is not only of great significance for improving the identification of BLs in complex samples, but also provides some reference and guidance for the design of highly active laccase-like nanozymes in the future.