Bacterial proteases are sporadic contributors to milk spoilage, reducing the quality of ultra-heat treated (UHT) milk and other dairy products. Current methods for measuring bacterial protease activity in milk are insensitive and too slow to be used in routine testing in dairy processing plants. We have designed a novel bioluminescence resonance energy transfer (BRET)-based biosensor to measure the activity of proteases secreted by bacteria in milk. The BRET-based biosensor is highly selective for bacterial protease activity compared with other proteases tested, notably including plasmin, which is abundant in milk. It incorporates a novel peptide linker that is selectively cleaved by P. fluorescens AprX proteas... More
Bacterial proteases are sporadic contributors to milk spoilage, reducing the quality of ultra-heat treated (UHT) milk and other dairy products. Current methods for measuring bacterial protease activity in milk are insensitive and too slow to be used in routine testing in dairy processing plants. We have designed a novel bioluminescence resonance energy transfer (BRET)-based biosensor to measure the activity of proteases secreted by bacteria in milk. The BRET-based biosensor is highly selective for bacterial protease activity compared with other proteases tested, notably including plasmin, which is abundant in milk. It incorporates a novel peptide linker that is selectively cleaved by P. fluorescens AprX proteases. The peptide linker is flanked by green fluorescent protein (GFP) at the N-terminus and a variant Renilla luciferase (RLuc2) at the C-terminus. Complete cleavage of the linker by bacterial proteases from Pseudomonas fluorescens strain 65, leads to a 95% decrease in the BRET ratio. We applied an azocasein-based calibration method to the AprX biosensor using standard international enzyme activity units. In a 10-min assay, the detection limit for AprX protease activity in buffer was equivalent to 40 pg/mL (≈0.8 pM, 22 μU/mL) and 100 pg/mL (≈2pM, 54 μU/mL) in 50% (v/v) full fat milk. The EC values were 1.1 ± 0.3 ng/mL (87 μU/mL) and 6.8 ± 0.2 ng/mL (540 μU/mL), respectively. The biosensor was approximately 800x more sensitive than the established FITC-Casein method in a 2-h assay, the shortest feasible time for the latter method. The protease biosensor is sensitive and fast enough to be used in production settings. It is suitable for measuring bacterial protease activity in raw and processed milk, to inform efforts to mitigate the effects of heat-stable bacterial proteases and maximise the shelf-life of dairy products.