AI-guided Cas9 engineering provides an effective strategy to enhance base editing

Area of Science:

• Molecular Biology
• Bioengineering
• Genetics

Background:

• Precise genome editing is essential for biological research and therapeutic applications.
• Current base editors require optimization to improve their editing efficiency.
• Advancements in protein design and AI have enabled sophisticated protein engineering.

Purpose of the Study:

• To optimize base editor efficiency using AI-driven protein engineering.
• To develop a high-performance Cas9 variant with enhanced editing capabilities.
• To validate the universal applicability of AI-guided strategies for improving gene editing tools.

Main Methods:

• Utilized the Protein Mutational Effect Predictor (ProMEP) to predict single-site mutations in Cas9.
• Engineered and tested 18 point mutations in the AncBE4max base editor prototype.
• Predicted and combined multiple mutations to create an optimized variant, AncBE4max-AI-8.3.

Main Results:

• Developed the AncBE4max-AI-8.3 variant, demonstrating a 2-3 fold increase in average editing efficiency.
• Successfully improved the performance of multiple base editors including CGBE, YEE-BE4max, ABE-max, ABE-8e, and HF-BEs.
• Observed stable enhancements in editing efficiency across seven cancer cell lines and human embryonic stem cells.

Conclusions:

• AI models are effective tools for protein engineering in the context of gene editing.
• The developed AI-guided strategy provides a universal approach to enhance the efficiency of various gene editing tools.
• Engineered base editors show broad applicability and stable performance in diverse cellular contexts.