Thermodynamic Microenvironment Engineering in Mesoporous Nanoreactors to Enhance Biocatalysis for AI-Empowered Ultrasensitive Pathogen Detection
Yuechun Li, Chenjie Nie, Chunyan Ji, Zhaowen Cui, Yanwei Ji, Min Ma, Wentao Zhang, Leina Dou, Qianjin Liu, Jianlong Wang
Abstract
Harmonizing enzyme-support microenvironments to govern thermodynamic interaction landscapes presents a critical yet underexplored frontier in nanobiocatalysis for pathogen detection. Herein, we architecturally engineer mesoporous resorcinol formaldehyde nanospheres (mRFNSs, 9.95 nm pores) with tailored surface chemistry to elucidate how microenvironment modulation dictates enzyme immobilization energetics. Thermodynamic dissection demonstrates that betaine-tailored mRFNSs with optimal immobilization efficiency and activity dramatically reshape binding energetics, achieving record affinity through optimized electrostatic complementarity, hydrogen-bond networks, and hydrophobic effect. This microenvironment engineering strategy delivers an unprecedented 4.01-fold enhancement in binding constant ( K a = 1.12 × 10 8 vs 2.79 × 10 7 M –1 ) and superior thermodynamic spontaneity (Δ G = −46.0 vs −42.5 kJ mol –1 ). Leveraging this, we develop a paradigm-shifting ratiometric fluorescence immunoassay where ALP triggers in situ silicon quantum dot (SiQDs) synthesis (530 nm) against tetraphenylbenzidine reference (620 nm), achieving ultrasensitive Salmonella typhimurium ( S. typhimurium ) detection (100 CFU mL –1 ), which is 50-fold lower than that of conventional ELISA. A convolutional neural network (CNN) decodes smartphone-captured fluorescence hues, enabling portable classification (93.75% accuracy) of pathogen levels. Validated in food matrices (81.44–116.93% recovery), this work establishes thermodynamic-microenvironment correlations as a blueprint for next-generation nanobiocatalysts, bridging biointerface science with artificial intelligence (AI)-enhanced diagnostics.