Physical interactions among plant MADS-box transcription factors and their biological relevance

The biological interpretation of the genome starts from transcription, and many different signaling pathways are integrated at this level. Transcription factors play a central role in the transcription process, because they select the down-stream genes and determine their spatial and temporal expression. In higher eudicot species around 2000 specific transcription factors are present, which can be classified into families based on conserved common domains. The MADS-box transcription factor family is an important family of transcription regulators in plants and genetic studies revealed that members of this family are involved in various developmental processes, like floral induction, floral organ formation and fruit development. In contrast to this wealth of information concerning MADS-box gene functions, the molecular mode of action of the encoded proteins is far from completely understood. Biochemical and yeast η-hybrid experiments performed in the past showed that MADS-box proteins are able to interact mutually, and based on these findings a hypothetical quaternary model has been proposed as molecular working mechanism. According to this model two MADS-box protein dimers assemble into a higher order complex, which binds DNA and regulates target gene expression. Although, this molecular mechanism sounds plausible, it still lacks evidence from in vivo studies. In this study we investigated physical interactions among members of the Petunia hybrida and Arabidopsis thaliana MADS-box transcription factor families in living plant cells. For this purpose, sophisticated micro-spectroscopy techniques have been implemented and in addition, some novel fluorescent-protein-based tools were developed. The first chapter gives an introduction about the dynamic transcriptional process and describes our current knowledge about transcriptional regulation in eukaryotes. The central question of this chapter is how transcription factors are able to find their specific binding sites (ci's-elements) within the huge genome. The various mechanisms, such as "looping" and "sliding", that have been proposed are discussed, as well as the relevance of direct interactions between transcription factors for the control of gene expression. In a first attempt to detect protein interactions in living cells, we transiently expressed combinations of petunia MADS-box transcription factors labeled with different color variants of the Green Fluorescent Protein (GFP) in leaf protoplasts (Chapter 2). Subsequently, the transfected protoplasts were analyzed by means of FRET-FLIM (Fluorescence Resonance Energy Transfer - Fluorescence Lifetime Imaging) to identify specific dimerization. In addition, we have obtained indirect evidence for higher-order complex formation of the petunia MADS-box proteins FLORAL BINDING PROTEIN2 (FBP2), FBP11, and FBP24 in living cells. Similar kind of analyses for Arabidopsis MADS-box proteins involved in petal and stamen development revealed clear differences in interaction affinities in vivo and furthermore, many homodimers were identified that could not be detected by yeast-based systems in the past (Chapter 3). This result demonstrated the robustness of the FRET-FLIM approach. Based on our observations, we hypothesize that 'partner selectivity' plays an important role in complex formation at particular developmental stages. To study differences in interaction affinity and selectivity and the consequences for complex formation in more detail, a novel method was developed (Chapter 4). The technique, designated "Competition-FRET", allows the verification of competition effects between proteins, and furthermore, it may provide information about the formation of higher-order complexes between different proteins under study. The developed method was implemented to investigate in depth the preference for homo- or heterodimer interactions of the Arabidopsis MADS-box proteins AGAMOUS (AG) and SEPALLATA3 (SEP3). The detection of interactions in living cells by FRET as it has been done in the studies described above demands a sophisticated microscopy set-up, and therefore, we decided to test and implement an alternative and theoretically simple technique (Chapter 5). This method for the in vivo detection of protein-protein interaction is called BiFC (Bimolecular Fluorescence Complementation), or "Split-YFP". In this system, a fluorescent molecule is split into two inactive domains and these two non-fluorescent parts are fused to the proteins under study. Only upon interaction of the two protein partners the two non-fluorescent parts of the fluorescent molecule are brought into close proximity, which enables the recovery of fluorescence. We used the EYFP (Enhanced Yellow Fluorescence Protein) molecule as fluorescent molecule and were able to detect the interaction between AG and SEP3 in nuclei of Arabidopsis leaf protoplasts. Techniques like this and FRET-FLIM allow the analyses of interactions between proteins in living cells, but give no information about the size of the formed complexes. To get a first indication about the stoichiometry of protein complexes, we monitored the diffusion time of in vitro synthesized AG-EYFP and SEP3-EYFP fusion proteins by means of FCS (Fluorescence Correlation Spectroscopy). From these experiments described in Chapter 6, we could speculate that SEP3 is present as a dimer and also as a higher order complex, whilst AG on its own is able to assemble into larger complexes. The diffusion time of the product formed upon co-translation of both AG and SEP3, suggests that a multimenc protein complex with a high molecular weight is formed upon interaction between AG and SEP3. Even though FCS is a powerful technique, these interpretations should be taken cautiously, mainly because these experiments were done in vitro instead of in living cells. Finally, in the last chapter we discuss the various methods that have been implemented and developed to monitor protein-protein interactions and complex formation of MADS-box transcription factors in living plant cells. Furthermore, we made a first step to monitor interactions in intact tissues under endogenous expression levels, and the preliminary results obtained from these in planta FRET-FLIM measurements are discussed.