Deep learning-guided engineering of pectinase for enhanced catalytic performance in tobacco processing

Area of Science:

• Biotechnology
• Enzyme Engineering
• Computational Biology

Background:

• Enzyme engineering is crucial for industrial applications.
• Deep learning offers novel approaches for protein redesign.
• Pectinase is vital in food processing and biomass degradation.

Purpose of the Study:

• To engineer a highly active and stable pectinase using deep learning.
• To investigate the structural basis for enhanced enzyme performance.
• To evaluate the enzyme's efficacy in industrial applications and its impact on flavor.

Main Methods:

• Utilized deep learning model ProteinMPNN guided by multiple sequence alignment for enzyme redesign.
• Introduced 72 mutations into the pectinase sequence.
• Performed structural and molecular dynamics analyses to understand performance enhancements.

Main Results:

• The engineered pectinase variant DS-5 exhibited an 8.9-fold increase in catalytic activity.
• DS-5 showed enhanced thermostability (10°C higher optimal temperature) and broad pH activity (7.0-11.0).
• Structural analysis revealed improved substrate binding potential due to altered surface electrostatics.

Conclusions:

• Deep learning provides an effective framework for engineering highly efficient enzymes.
• The redesigned pectinase demonstrates superior performance in apple juice clarification and tobacco processing.
• Enzyme engineering improvements directly translate to enhanced end-product quality and sensory profiles.