Eine kritische Evaluierung FRET-basierter Biosensoren als Werkzeuge für die quantitative Metabolitanalytik

In the recent years, a huge set of genetically-encoded fluorescence biosensors has been developed, allowing the detection of signaling intermediates and metabolites in real time with subcellular spatial resolution. Many of these biosensors are based on Foerster Resonance Energy Transfer (FRET). They represent the basis for numerous non-invasive cell-based assays to monitor signal transduction in living cells. The two FRET-based biosensors used in this work are composed of periplasmatic binding proteins from E. coli (PBP) fused to two derivatives of the green fluorescent protein (GFP) at the Cand N-terminus, respectively. Structural changes of the PBPs in such “venus-flytrap” sensors upon metabolite binding alter the relative orientation and/or distance between both terminal fluorescent proteins thereby changing the FRET signal. Specifically two well established sugar sensors for maltose- (FLIPmal25μ) and glucose detection (FLII12Pglu-) were investigated (Fehr, Frommer et al. 2002; Deuschle, Okumoto et al. 2005), which have already been used in several mammalian-, lower eukaryotic-, bacterial- and plant cells (Bermejo, Haerizadeh et al. 2011; Hou, Takanaga et al. 2011). In order to evaluate both biosensors as in vivo tools for quantitative metabolite analyses the effects of varied environmental conditions of biochemically essential parameters on the signal intensity and the Kd-value of both biosensors and on the single fluorescence proteins were thoroughly studied in this PhD-project. Specifically the influences of pH, buffer salts, ionic strength, temperature and several intracellular metabolites were studied. Further, investigations with different molecular crowding agents such as polyethyleneglycol and Ficoll demonstrated that the sensor specific parameters are also influenced by the viscosity of the medium. The results demonstrate that both biosensors are significantly affected by numerous of the tested parameters. Almost all micro-environmental variations led to considerably different FRET-signals or Kd-values, because either the fluorescent proteins or the binding domain are affected by the tested parameters. As a consequence, the common practice to apply such sensors for quantitative in vivo measurements based on an in vitro calibration in an arbitrarily chosen buffer system can easily result in false concentration determination in vivo, because of the multiple influence factors in living cells, which cannot be controlled.