Analysis and Prediction of Chymotrypsin Substrate Preferences through Large Data Acquisition with Target-Free mRNA...

Area of Science:

• Biochemistry
• Computational Biology
• Drug Discovery

Background:

• Oral peptide drug delivery is hindered by degradation from gastrointestinal proteases, such as chymotrypsin.
• Current peptidase databases are insufficient for systematic analysis of protease substrate preferences, particularly for non-natural amino acids, making peptide stability optimization difficult.

Purpose of the Study:

• To accelerate the stability optimization of peptide drug candidates.
• To develop a predictive model for identifying chymotrypsin-resistant peptide sequences.
• To explore the utility of mRNA display for generating large datasets for protease resistance prediction.

Main Methods:

• Generated large datasets of chymotrypsin-resistant peptides using mRNA display.
• Analyzed enriched sequence motifs to understand chymotrypsin cleavage patterns.
• Developed and validated a machine-learning model to predict peptide resistance to chymotrypsin.
• Simulated stability predictions for non-natural amino acids using a leucine hold-out model.

Main Results:

• Recapitulated known chymotrypsin cleavage sites and identified novel protective and destabilizing residues.
• Demonstrated position-dependent effects of amino acids on peptide cleavage.
• Validated the predictive model's performance using experimental chymotrypsin half-life measurements.
• Showcased robust performance of the model in predicting stability for non-natural amino acids.

Conclusions:

• mRNA display is a powerful tool for generating large-scale data for protease resistance studies.
• Combining mRNA display with machine learning provides valuable predictive models for chymotrypsin cleavage.
• This workflow can be expanded to other proteases, aiding future peptide drug discovery efforts.