Nanopore sensing with machine learning-assisted analysis for real-time single-molecule detection of MDM2-p53TAD interactions.

Protein-protein interactions (PPIs) play pivotal roles in physiological processes but remain challenging drug targets. Conventional PPI detection methods often suffer from limited sensitivity and require time- and labor-intensive procedures, such as fluorophore labeling, restricting their utility in drug discovery. Here, we present a YaxAB nanopore sensing platform coupled with machine learning-assisted signal analysis for label-free, real-time, single-molecule detection of PPIs between the oncogenic murine double minute 2 (MDM2) protein and p53TAD1. Proteins were efficiently captured into the YaxAB nanopore, and free MDM2 generated ionic current signatures readily distinguishable from those of the MDM2-p53TAD1 complex. To enable rapid and quantitative event analysis, we developed a machine learning-based classification framework that accurately distinguished free MDM2, MDM2-p53TAD1 complexes, and MDM2-drug complexes directly from nanopore current traces. This integrated platform enabled automated interpretation of single-molecule events, real-time monitoring of PPI inhibition by small-molecule MDM2 antagonists, and estimation of drug binding affinity without labeling or ensemble averaging. Taken together, the YaxAB nanopore sensing and machine learning analysis platform provides a versatile approach for single-molecule interaction analysis, with potential applications in drug discovery, diagnostics, including protein probe-biomarker interaction assays.