One pattern analysis (OPA) for the quantitative determination of protein interactions in plant cells

BACKGROUND: A commonly used approach to study the interaction of two proteins of interest (POIs) in vivo is measuring Förster Resonance Energy Transfer (FRET). This requires the expression of the two POIs fused to two fluorescent proteins that function as a FRET pair. A precise way to record FRET is Fluorescence Lifetime IMaging (FLIM) which generates quantitative data that, in principle, can be used to resolve both complex structure and protein affinities. However, this potential resolution is often lost in many experimental approaches. Here we introduce a novel tool for FLIM data analysis of multiexponential decaying donor fluorophores, one pattern analysis (OPA), which allows to obtain information about protein affinity and complex arrangement by extracting the relative amplitude of the FRET component and the FRET transfer efficiency from other FRET parameters. RESULTS: As a proof of concept for OPA, we used FLIM-FRET, or FLIM-FRET in combination with BiFC to reassess the dimerization and tetramerization properties of known interacting MADS-domain transcription factors in Nicotiana benthamiana leaf cells and Arabidopsis thaliana flowers. Using the OPA tool and by extracting protein BINDING efficiencies from FRET parameters to dissect MADS-domain protein interactions in vivo in transient N. benthamiana experiments, we could show that MADS-domain proteins display similar proximities within dimeric or tetrameric complexes but bind with variable affinities. By combining FLIM with BiFC, we were able to identify SEPALLATA3 as a mediator for tetramerization between the other MADS-domain factors. OPA also revealed that in vivo expression from native promoters at low levels in Arabidopsis flower meristems, makes in situ complex formation of MADS-domain proteins barely detectable. CONCLUSIONS: We conclude that MADS-domain protein interactions are transient in situ and may involve additional, so far unknown interaction mediators. We conclude that OPA can be used to separate protein binding from information about proximity and orientation of the interacting proteins in their complexes. Visualization of individual protein interactions within the underlying interaction networks in the native environment is still restrained if expression levels are low and will require continuous improvements in fluorophore labelling, instrumentation set-ups and analysis tools.