An evaluation of genetically encoded FRET-based biosensors for quantitative metabolite analyses in vivo.

A broad range of genetically-encoded fluorescence biosensors has been developed, allowing the detection of signaling intermediates and metabolites in real time. Many of these biosensors are based on Foerster Resonance Energy Transfer (FRET). The two biosensors of the well-known "Venus-flytrap" type exemplarily studied in this work are composed of a central sugar binding protein flanked by two green fluorescent protein derivatives, namely ECFP as well as Citrine and EYFP, respectively. In order to evaluate FRET-based biosensors as an in vivo tool for quantitative metabolite analyses, we have thoroughly studied the effects of pH, buffer salts, ionic strength, temperature and several intracellular metabolites on the signal intensity of both biosensors and both fluorescence proteins. Almost all micro-environmental variations led to considerably different FRET signals, because either the fluorescent proteins or the metabolite binding domains were affected by the tested parameters. This resulted not only in altered FRET ratios between the apo state and the saturated state but also in significant shifts of the apparent binding constant. This underlines the necessity of careful controls in order to allow reliable quantitative measurements in vivo.