Direct comparison of bioluminescence-based resonance energy transfer methods for monitoring of proteolytic cleavage

Bioluminescence resonance energy transfer (BRET) is a powerful tool for the study of protein–protein interactions and conformational changes within proteins. Two common implementations of BRET are BRET1 with Renilla luciferase (RLuc) and coelenterazine h (CLZ, λem ∼ 475 nm) and BRET2 with the substrate coelenterazine 400a (CLZ400A substrate, λem = 395 nm) as the respective donors. For BRET1 the acceptor is yellow fluorescent protein (YFP) (λem ∼ 535 nm), a mutant of green fluorescent protein (GFP), and for BRET2 it is GFP2 (λem ∼ 515 nm). It is not clear from previous studies which of these systems has superior signal-to-background characteristics. Here we directly compared BRET1 and BRET2 by placing two different protease-specific cleavage sequences between the donor and acceptor domains. The intact proteins simulate protein–protein association. Proteolytic cleavage of the peptide linker simulates protein dissociation and can be detected as a change in the BRET ratios. Complete cleavage of its target sequence by thrombin changed the BRET2 ratio by a factor of 28.9 ± 0.2 (relative standard deviation [RSD], n = 3) and changed the BRET1 ratio by a factor of 3.05 ± 0.07. Complete cleavage of a caspase-3 target sequence resulted in the BRET ratio changes by factors of 15.45 ± 0.08 for BRET2 and 2.00 ± 0.04 for BRET1. The BRET2 assay for thrombin was 2.9 times more sensitive compared with the BRET1 version. Calculated detection limits (blank signal + 3σb, where σb = standard deviation [SD] of blank signal) were 53 pM (0.002 U) thrombin with BRET1 and 15 pM (0.0005 U) thrombin with BRET2. The results presented here suggest that BRET2 is a more suitable system than BRET1 for studying protein–protein interactions and as a potential sensor for monitoring protease activity.