Boosting Array-based Diagnosis of Urinary Tract Infections via a Heterobifunctional Bridge-Driven Strategy

Array-based sensing platforms have witnessed significant progress in recent years, demonstrating considerable potential for high-throughput and multiplexed analytes detection; however, their clinical translation remains hindered by inadequate sensitivity in complex biofluids. Herein, we introduce a heterobifunctional bridge-driven (HB-Bridge) strategy for constructing the sensor array, which incorporates asymmetric termini to enable dual-mode sensing through concurrent covalent binding and electrostatic interactions. Upon addition of bacteria or lipopolysaccharide analogues, the array constructed by covalently complexing HB-Bridge HB1-HB3 with perylene diimide fluorophores (PDIs) P1-P6 allow both fluorescence turn-on through covalent competition and quenching through noncovalent aggregation. The optimized array enabled the simultaneous identification of 21 uropathogens in urine, achieving 84.52% blind-test accuracy. Ultrasensitive detection was achieved for Escherichiacoli, the predominant cause of over 80% of clinical urinary tract infections (UTIs), with a limit of OD₆₀₀ = 0.0000417. Notably, in a double-blind study of 150 clinical urine samples, the top-performing multilayer perceptron (MLP), screened via multiple machine-learning classifiers, achieved 98.50% accuracy in simultaneously distinguishing multiple bacterial types within 30 min using only 12.5 µL of urine, and remarkably maintained 97.92% accuracy even in clinically challenging polymicrobial cases. Overall, this study establishes a dual-mode sensing platform enabling rapid and precise UTI diagnosis with strong potential for clinical translation.