Label-Free Imaging of Single Proteins and Binding Dynamics via Deep Learning-Enhanced Plasmonic Scattering Microscopy

Area of Science:

• Biophysics
• Biochemistry
• Nanotechnology

Background:

• Label-free imaging of single-protein binding dynamics is crucial for understanding biochemical processes and developing biosensors.
• Current methods face challenges in accurately tracking unlabeled proteins in complex environments.

Purpose of the Study:

• To develop a high-throughput, automatic, label-free method for tracking single unlabeled proteins.
• To enable real-time analysis of protein binding dynamics, including residence time and thermodynamics.
• To establish a quantitative framework for distinguishing specific from nonspecific protein interactions.

Main Methods:

• Integration of plasmonic scattering microscopy with a spatiotemporal deep-learning framework.
• Utilizing a tailored recurrent neural network for continuous isolation and tracking of single proteins.
• Leveraging label-free imaging and real-time trajectory analysis.

Main Results:

• Achieved high-throughput, automatic, label-free tracking of single proteins from complex backgrounds.
• Successfully discriminated between transient and stable binding events through residence time measurements.
• Resolved nanoscale protein motions and quantified binding thermodynamics.
• Established a quantitative framework to differentiate specific from nonspecific interactions.

Conclusions:

• The developed method provides a powerful tool for label-free single-protein analysis.
• It sets new benchmarks for detecting specific protein binding in low-abundance immunosensing.
• The approach is extendable to analyze protein interactions in complex biological fluids.