Brent Carroll, BS
New York University
High-Throughput Dissection of Ultra-Specific Molecular Recognition by Botulinum Proteases
Scanning Site Saturation Mutagenesis (SSSM) substitutes all possible single amino acid substitutions in a protein. However, these pooled oligo-based libraries generally only include sidechains that reveal positive or negative effects from the screen. Our system, termed Digital Directed Evolution (DDE), sorts the mutations into individual wells first and quantifies complete enzyme kinetics for whole-protein SSSM. We use custom-built robotic systems to quantitatively assay thousands of mutants per day and reduce the cost by 100-fold. Importantly, thorough enzyme kinetics in every position for every sidechain can uncover the basis for enzyme recognition and catalysis. This data will provide an unprecedented training set for AI/ML-guided enzyme design.
Here, this approach focuses on one of nature’s most impressive feats of molecular recognition – an enzyme capable of recognizing and cleaving a single peptide bond in the human proteome. This protease, called light chain E (LC/E) derives from Botulinum Neurotoxin serotype E (BONT/E). Control over such proteolytic specificity could form the basis for high-potency, low-dose protein-ablative therapeutics to treat a wide variety of diseases.
Scanning Site Saturation Mutagenesis (SSSM) substitutes all possible single amino acid substitutions in a protein. However, these pooled oligo-based libraries generally only include sidechains that reveal positive or negative effects from the screen. Our system, termed Digital Directed Evolution (DDE), sorts the mutations into individual wells first and quantifies complete enzyme kinetics for whole-protein SSSM. We use custom-built robotic systems to quantitatively assay thousands of mutants per day and reduce the cost by 100-fold. Importantly, thorough enzyme kinetics in every position for every sidechain can uncover the basis for enzyme recognition and catalysis. This data will provide an unprecedented training set for AI/ML-guided enzyme design.
Here, this approach focuses on one of nature’s most impressive feats of molecular recognition – an enzyme capable of recognizing and cleaving a single peptide bond in the human proteome. This protease, called light chain E (LC/E) derives from Botulinum Neurotoxin serotype E (BONT/E). Control over such proteolytic specificity could form the basis for high-potency, low-dose protein-ablative therapeutics to treat a wide variety of diseases.
