Small-molecule binding and sensing with a designed protein family.

Area of Science:

• Protein Engineering
• Computational Biology
• Biosensor Development

Background:

• De novo protein design is crucial for creating custom molecular tools.
• Developing sensors for arbitrary small molecules requires tailored binding proteins.

Purpose of the Study:

• To create a versatile platform for designing small-molecule-binding proteins using computational methods.
• To demonstrate the utility of these designed proteins in creating functional biosensors.

Main Methods:

• Integration of deep learning algorithms with physics-based simulations for protein pocket engineering.
• Computational design of protein binders for six distinct small-molecule targets.
• Biophysical characterization to validate binding affinities and design accuracy.
• Construction of a chemically induced dimerization (CID) system for biosensing.

Main Results:

• Generated a diverse family of proteins with tunable pocket geometries.
• Achieved nanomolar to low micromolar binding affinities for designed protein binders.
• Demonstrated atomic-level accuracy in protein design.
• Successfully developed a cortisol biosensor using a designed cortisol binder and CID system.

Conclusions:

• The NTF2 protein fold is amenable to deep learning-based design for creating specific binders.
• This integrated approach accelerates the development of custom protein binders and biosensors.
• The methodology holds potential for applications in analytical, environmental, and biomedical fields.