Binding and sensing diverse small molecules using shape-complementary pseudocycles

Area of Science:

• Protein engineering
• Biotechnology
• Computational biology

Background:

• Designing proteins with specific small molecule-binding capabilities is challenging.
• Existing methods often lack the versatility for diverse molecular targets.

Purpose of the Study:

• To develop a novel computational approach for designing high-affinity small molecule-binding proteins.
• To create proteins suitable for downstream applications like sensing and chemically induced dimerization.

Main Methods:

• Utilized deep learning to generate protein pseudocycles with tunable binding pocket shapes.
• Employed computational docking to identify complementary protein designs for target small molecules.
• Optimized interaction surfaces for high binding affinity and performed experimental screening.

Main Results:

• Successfully designed and validated high-affinity binders for four distinct small molecules, including methotrexate and thyroxine.
• Demonstrated the modularity of the designed proteins by creating chemically induced dimerization systems.
• Engineered low-noise nanopore sensors utilizing the designed protein domains.

Conclusions:

• The deep learning-based pseudocycle design approach is effective for creating high-affinity small molecule binders.
• The modular protein designs facilitate the development of advanced biosensing and molecular control systems.
• This method offers a versatile platform for protein engineering with broad biotechnological potential.