Label-Free, AI-Driven Evanescent Microscopy Decodes Single-Cell Membrane Protein Binding Kinetics and Adhesion
Caixin Huang1,2, Jingbo Zhang1,3, Zhaoyang Liu1,3
1Beijing National Laboratory for Molecular Sciences, Key Laboratory of Analytical Chemistry for Living Biosystems, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China.
None:
Quantifying ligand interactions with membrane proteins in their native cellular environment is crucial for understanding cellular processes and drug screening, yet it remains a challenging task. Here, we introduce an artificial intelligence (AI)-driven evanescent microscopy that enables label-free quantification of kinetics for the binding of both large and small molecules to membrane proteins on multiple individual cells. The optical platform continuously monitors ligand interactions with membrane proteins at each cell adhesion site in real time across a millimeter-scale field of view. An AI-based multifeature recognition framework addresses the high noise floor issue resulting from the intrinsic nonspecificity of label-free imaging, enabling automated, subcellular quantification of ligand binding kinetics for individual cells. Using an explainable machine learning approach, we demonstrate how cellular adhesion biomechanics contribute to heterogeneous membrane protein binding kinetics, with the dominant biomechanical features varying depending on the specific ligands and membrane proteins involved. This work presents a scalable solution for label-free single-cell analysis, paving the way for exploring cellular features that regulate ligand interactions with membrane proteins, which is essential for understanding the heterogeneity that drives drug resistance.
