Quantification of Dark Protein Populations in Fluorescent Proteins by Two-Color Coincidence Detection and Nanophotonic Manipulation

[Image: see text] Genetically encoded visible fluorescent proteins (VFPs) are a key tool used to visualize cellular processes. However, compared to synthetic fluorophores, VFPs are photophysically complex. This photophysical complexity includes the presence of non-emitting, dark proteins within the ensemble of VFPs. Quantitative fluorescence microcopy approaches that rely on VFPs to obtain molecular insights are hampered by the presence of these dark proteins. To account for the presence of dark proteins, it is necessary to know the fraction of dark proteins (f(dark)) in the ensemble. To date, f(dark) has rarely been quantified, and different methods to determine f(dark) have not been compared. Here, we use and compare two different methods to determine the f(dark) of four commonly used VFPs: EGFP, SYFP2, mStrawberry, and mRFP1. In the first, direct method, we make use of VFP tandems and single-molecule two-color coincidence detection (TCCD). The second method relies on comparing the bright state fluorescence quantum yield obtained by photonic manipulation to the ensemble-averaged fluorescence quantum yield of the VFP. Our results show that, although very different in nature, both methods are suitable to obtain f(dark). Both methods show that all four VFPs contain a considerable fraction of dark proteins. We determine f(dark) values between 30 and 60% for the different VFPs. The high values for f(dark) of these commonly used VFPs highlight that f(dark) has to be accounted for in quantitative microscopy and spectroscopy.