Protein-protein interactions are fundamental to virtually every aspect of cellular functions. Blue, green and yellow bimolecular fluorescence complementation (BiFC) systems based on GFP and its variants allow the investigation of protein-protein interactions in vivo. We have developed the first red BiFC system based on an improved monomeric red fluorescent protein (mRFP1-Q66T), expanding the range of possible applications for BiFC.With interactome data available for several model organisms, a challenging next step in post-genomic research is to analyze protein complex formation in vivo. The recently developed BiFC assay is a comparably fast and simple noninvasive technology to study protein interactions inside living cells. BiFC is based on the reconstitution of the yellow fluorescent protein (YFP) from two nonfluorescent fragments when they are brought into close proximity by a physical interaction between proteins fused to each fragment 1 . This approach has proven to be robust and versatile, and multicolor versions using spectral variants of GFP further increase the potential of this technology 2 . The color spectrum available for BiFC, however, has been limited to blue, green and yellow.The red fluorescent protein from Discosoma sp. (DsRED) and its variants are established intracellular reporter proteins, but the characteristics of most DsRED variants, particularly their obligate tetramerization, impede their application in BiFC 3,4 . An extensively mutated monomeric DsRED variant (mRFP1) has been generated 5 , but substantially altered spectra, poor brightness and low photostability limit its usefulness as a reporter protein.We have identified improved mRFP1 mutants, which allowed us to establish a red fluorescent reporter system for detection of protein interactions in vivo. We used site-directed mutagenesis to generate mRFP1 variants with a randomized first position of the fluorophore. Screening of 5,000 colonies resulted in the identification of 50 clones exhibiting strong red fluorescence representing three different mutations (Q66C, Q66S and Q66T) with frequencies of 5%, 43% and 52%, respectively.

SummaryThe suitability of the recently described red¯uorescent protein dsRED from reef corals for use as a reporter in plant molecular biology was investigated. Based on the clone pDSRED (Clontech), plant expression vectors were constructed for constitutive dsRED expression in the cytosol, the endoplasmic reticulum and the vacuole. Fluorescence microscopy of tobacco BY2 suspension culture cells transiently expressing the plant vectors generated proved that cytosolic expression of the dsRED gives rise to readily detectable levels of red¯uorescence, whereas expression in the ER was poor. Vacuolar dsRED expression did not result in any signi®cant¯uorescence. dsRED transgenic tobacco SR1 plants were generated to test the sensitivity of dsRED as a reporter in an auto¯uorescent background, and to identify the possible impact of the introduced¯uorescent protein on morphogenesis, plant development and fertility. During the transformation and regeneration phase plants did not show any abnormalities, indicating that dsRED is not interfering with plant development and morphogenesis. Regenerated plants were analysed by PCR, Western blot and¯uorescence microscopy for the presence and expression of the transferred genes. The ®lter sets chosen for¯uorescence microscopy proved to be able to block the red chlorophyll¯uorescence completely, allowing speci®c dsRED detection. Best expression levels were obtained with dsRED targeted to the cytosol or chloroplasts. ER-targeted expression of dsRED also gave rise to readily detectable¯uorescence levels, whereas vacuolar expression yielded no¯uorescence. dsRED transgenic plant lines expressing the protein in the cytosol, ER or chloroplast proved to be fertile. Seed set and germination were normal, except that the seeds and seedlings maintained the red uorescence phenotype.