Assessing substrate scope of the cyclodehydratase LynD by mRNA display-enabled machine learning models

Area of Science:

• Biochemistry
• Molecular Biology
• Natural Products Chemistry

Background:

• Ribosomal synthesized and post-translationally modified peptides (RiPPs) are natural products synthesized via complex enzymatic pathways.
• Multi-domain enzymes in RiPP biosynthesis exhibit a "division of labor," with distinct domains for substrate binding and modification.
• The promiscuous nature of RiPP enzyme active sites allows for modification of multiple residues within a peptide substrate, irrespective of sequence context.

Purpose of the Study:

• To investigate the substrate promiscuity of the RiPP cyclodehydratase, LynD, using a high-throughput assay.
• To develop predictive models for RiPP enzyme substrate processing.
• To explore the application of these models in enzyme elucidation and the engineering of peptide-based therapeutics.

Main Methods:

• Utilized mRNA display, a high-throughput peptide display technology, for extensive substrate profiling of LynD.
• Constructed deep learning models based on the generated substrate profiling data.
• Analyzed epistatic interactions within enzymatic processing pathways.

Main Results:

• Demonstrated the capability of mRNA display to perform vast substrate profiling for RiPP enzymes.
• Developed accurate deep learning models for predicting LynD's substrate processing.
• Gained insights into epistatic interactions governing enzymatic activity.

Conclusions:

• High-throughput mRNA display is effective for characterizing RiPP enzyme substrate promiscuity.
• Deep learning models can accurately predict RiPP enzyme activity and inform on mechanistic details.
• This approach facilitates the discovery and engineering of RiPP enzymes for therapeutic applications, including peptide-based inhibitors.