Engineering the Substrate Specificity of Toluene Degrading Enzyme XylM Using Biosensor XylS and Machine Learning

Area of Science:

• Biotechnology
• Synthetic Biology
• Enzyme Engineering

Background:

• High-throughput enzyme activity evaluation is crucial for machine learning-driven enzyme engineering.
• Current methods can be limiting for generating large training datasets.
• Biosensor-based approaches offer a potential solution for quantitative and rapid activity assessment.

Purpose of the Study:

• To investigate the applicability of a biosensor-based method for enzyme engineering using machine learning.
• To engineer the substrate specificity of XylM (a rate-determining enzyme in XylMABC) from *Pseudomonas putida*.
• To improve the conversion efficiency of XylM for the non-native substrate 2,6-xylenol.

Main Methods:

• Developed a biosensor in *Escherichia coli* utilizing a fluorescent protein reporter regulated by the XylS transcriptional regulator, which responds to 3-methylsalicylic acid.
• Used fluorescence intensity as a proxy for 3-methylsalicylic acid productivity to evaluate XylM variants.
• Applied machine learning-assisted directed evolution, using biosensor data as training input, to optimize XylM.

Main Results:

• The biosensor system demonstrated a concentration-dependent fluorescence response to 3-methylsalicylic acid, enabling indirect enzyme activity evaluation.
• Two cycles of machine learning-assisted directed evolution yielded an XylM variant (XylM-D140E-V144K-F243L-N244S) with 15-fold higher productivity compared to the wild-type.
• The generated data proved sufficiently quantitative and high-throughput for machine learning training.

Conclusions:

• Biosensor-based indirect enzyme activity evaluation is a viable, high-throughput method for generating training data in machine learning-driven enzyme engineering.
• This approach significantly enhances the directed evolution of enzymes, as demonstrated by the improved XylM variant.
• The study expands the utility of machine learning in enzyme engineering and protein design.