Embracing enzyme promiscuity with activity-based compressed biosensing.

Area of Science:

• Biochemistry
• Enzymology
• Proteomics
• Biotechnology

Background:

• Developing protease-activatable drugs and diagnostics necessitates identifying protease-specific substrates.
• Most substrates are cleaved by multiple proteases, complicating the identification of individual enzyme activity in mixtures.
• Existing methods struggle with scalability as the number of target proteases increases.

Purpose of the Study:

• To introduce a novel method, Substrate Libraries for Compressed Sensing of Enzymes (SLICE), for selecting substrate libraries to classify protease mixtures.
• To enable protease mixture classification without signal deconvolution or highly specific substrates.
• To develop a metric for ranking substrate libraries based on orthogonality and protease coverage.

Main Methods:

• Developed the Substrate Libraries for Compressed Sensing of Enzymes (SLICE) method.
• Introduced a compression score (C) to quantify substrate orthogonality and protease coverage.
• Validated the compression score's predictive accuracy for classification using 140 in silico and 55 in vitro libraries.

Main Results:

• The compression score (C) showed strong predictive accuracy for classification (Pearson r = 0.71 for in silico, r = 0.55 for in vitro).
• A two-substrate library selected using SLICE successfully classified 11 enzymes in 28 plasma samples (AUROC = 0.93).
• SLICE enables classification of complex protease mixtures with high accuracy using minimal substrates.

Conclusions:

• SLICE provides an efficient strategy for selecting substrate libraries to classify protease mixtures.
• The method overcomes limitations of substrate promiscuity and the need for specific substrates.
• SLICE is poised to advance the development of activity-based sensors for diagnostics and imaging.