Machine Learning Optimizing Enzyme/ZIF Biocomposites for Enhanced Encapsulation Efficiency and Bioactivity

Area of Science:

• Biochemistry and Materials Science
• Enzyme immobilization
• Nanotechnology

Background:

• Enzyme immobilization in zeolitic imidazolate frameworks (ZIFs) is crucial for biocatalysis.
• Optimizing synthesis for enhanced enzyme performance (encapsulation efficiency, retained activity, thermal stability) remains challenging.
• Machine learning (ML) offers a powerful tool for complex optimization problems.

Purpose of the Study:

• To present the first ML-assisted design strategy for optimizing enzyme/ZIF synthesis.
• To enhance the performance of glucose oxidase (GOx) and horseradish peroxidase (HRP) within ZIFs.
• To establish a workflow for accelerating the development of enzyme/ZIF systems.

Main Methods:

• Utilized an iterative ML-driven training-design-synthesis-measurement workflow.
• Employed glucose oxidase (GOx) and horseradish peroxidase (HRP) as model enzymes.
• Selected Zn(eIM)2 (eIM = 2-ethylimidazolate) as the model ZIF for optimization.

Main Results:

• Discovered optimized GOx/ZIF (G151) and HRP/ZIF (H150) formulations.
• Achieved overall performance index (OPI) values at least 1.3 times higher than baseline studies.
• Identified key relationships between synthesis parameters, structure, and enzyme performance using random forest models.

Conclusions:

• The ML-assisted design strategy effectively accelerates the optimization of enzyme/ZIF synthesis.
• This approach provides valuable insights for future enzyme immobilization studies.
• The strategy holds significant promise for biocatalysis, drug delivery, and sensing applications.