Artificial allosteric protein switches with machine-learning-designed receptors.

Area of Science:

• Synthetic biology
• Biotechnology
• Protein engineering

Background:

• Protein allostery is crucial for biological information and energy processing.
• Developing artificial allosteric proteins is a major goal in synthetic biology and biotechnology.

Purpose of the Study:

• To engineer minimal ligand-binding domains as efficient receptors for single-component allosteric switches.
• To create synthetic allosteric switches for biosensing and logic gate applications.
• To demonstrate the practical utility of these synthetic switches in engineered cells and bioelectronic devices.

Main Methods:

• Machine learning for engineering minimal ligand-binding domains.
• Construction of colorimetric, luminescent, and electrochemical biosensors.
• Compilation of biosensors into intramolecular YES and AND logic gates.
• Hydrogen/deuterium exchange mass spectrometry and 19F nuclear magnetic resonance analyses.
• Engineering Escherichia coli for steroid-dependent antibiotic resistance.
• Development of bioelectronic devices for steroid hormone quantification.

Main Results:

• Machine-learning-engineered minimal domains function as efficient receptors in allosteric switches without global conformational change.
• Synthetic allosteric switches were successfully compiled into intramolecular YES and AND logic gates.
• Ligand binding was shown to reduce conformational entropy, enhancing reporter domain catalytic activity.
• Demonstrated utility through engineered E. coli with steroid-dependent antibiotic resistance.
• Developed bioelectronic devices capable of quantifying steroid hormones.

Conclusions:

• Minimal engineered domains can serve as effective receptors in synthetic allosteric switches.
• These synthetic switches offer versatile platforms for biosensing, logic operations, and bioelectronic applications.
• The approach holds significant potential for advancing synthetic biology and biotechnology tools.