Combination of novel green fluorescent protein mutant TSapphire and DsRed variant mOrange to set up a versatile in planta FRET-FLIM assay

Förster resonance energy transfer (FRET) measurements based on fluorescence lifetime imaging microscopy (FLIM) are increasingly being used to assess molecular conformations and associations in living systems. Reduction in the excited-state lifetime of the donor fluorophore in the presence of an appropriately positioned acceptor is taken as strong evidence of FRET. Traditionally, cyan fluorescent protein has been widely used as a donor fluorophore in FRET experiments. However, given its photolabile nature, low quantum yield, and multiexponential lifetime, cyan fluorescent protein is far from an ideal donor in FRET imaging. Here, we report the application and use of the TSapphire mutant of green fluorescent protein as an efficient donor to mOrange in FLIM-based FRET imaging in intact plant cells. Using time-correlated single photon counting-FLIM, we show that TSapphire expressed in living plant cells decays with lifetime of 2.93 ± 0.09 ns. Chimerically linked TSapphire and mOrange (with 16-amino acid linker in between) exhibit substantial energy transfer based on the reduction in the lifetime of TSapphire in the presence of the acceptor mOrange. Experiments performed with various genetically and/or biochemically known interacting plant proteins demonstrate the versatility of the FRET-FLIM system presented here in different subcellular compartments tested (cytosol, nucleus, and at plasma membrane). The better spectral overlap with red monomers, higher photostability, and monoexponential lifetime of TSapphire makes it an ideal FRET-FLIM donor to study protein-protein interactions in diverse eukaryotic systems overcoming, in particular, many technical challenges encountered (like autofluorescence of cell walls and fluorescence of pigments associated with photosynthetic apparatus) while studying plant protein dynamics and interactions.Single- and dual-color fluorescence imaging with intrinsically fluorescent proteins is increasingly being used to study the expression, targeting, colocalization, turnover, and associations of diverse proteins involved in different plant signal transduction pathways (for review, see Fricker et al., 2006). Concurrent with the use of fluorescence-based cell biology, Förster resonance energy transfer (FRET) has emerged as a convenient tool to study the dynamics of protein associations in vivo. The technique exploits the biophysical phenomenon of nonradiative energy transfer from a donor fluorophore to an appropriately positioned acceptor at a nanometer scale (1–10 nm; Jares-Erijman and Jovin, 2003). In living cells, FRET occurs when two proteins (or different domains within a single protein) fused to suitable donor and acceptor fluorophores physically interact, thus bringing the donor and the acceptor within the favorable proximity for energy transfer (Immink et al., 2002; Bhat et al., 2006). This results in a decrease in the donor''s fluorescence intensity (or quantum yield QY]) and excited-state lifetime (Gadella et al., 1999). Furthermore, if the acceptor molecule is a fluorophore, then FRET additionally results in an increase in the acceptor''s emission intensity (sensitized emission; Shah et al., 2001; Bhat et al., 2006).However, the exploitation and use of fluorescent marker proteins to study protein trafficking and associations in plants can be problematic because plant cells contain a number of autofluorescent compounds (e.g. lignin, chlorophyll, phenols, etc.) whose emission spectra interfere with that of the most commonly used green or red fluorescent protein fluorophores and/or their spectral variants. For example, lignin fluorescence in roots, vascular tissues, and cell walls of aerial plant parts interferes with imaging at wavelengths between 490 and 620 nm, whereas the chlorophyll autofluorescence in green aerial plant parts is prevalent between 630 and 770 nm (Chapman et al., 2005). Consequently, conventional imaging of GFP and its closest spectral variants (like cyan fluorescent protein CFP] and yellow fluorescent protein YFP]) is most likely to be problematic in roots, whereas red-shifted intrinsic fluorescent proteins (including monomeric red fluorescent protein and recently identified spectral variants like mStrawberry and mCherry) may be hard to discriminate in chloroplast-containing aerial tissues (Chapman et al., 2005). The problems get further compounded in FRET assays because the autofluorescence arising from phenols, lignin, and chlorophyll can limit the choice of fluorophores suitable for in planta FRET assays.CFP and YFP have been widely used as a donor-acceptor pair in in planta FRET measurements (Bhat et al., 2006; Dixit et al., 2006). However, in photophysical terms, this pair is less than ideal for FRET imaging. Both have broad excitation and emission spectra with a small Stokes shift (Chapman et al., 2005). Second, QY of CFP (QY = 0.4) is relatively lower than that of YFP (QY = 0.61), and thus a significantly higher (and rather cell damaging) amount of excitation energy is needed to induce FRET (Dixit et al., 2006). Additionally, CFP displays multiexponential lifetimes with a shorter (1.3 ns) and a longer (2.6 ns) component (Becker et al., 2006). Although the deviation from the single-component decay is reasonably small (Tramier et al., 2002; Becker et al., 2006), the shorter CFP lifetime component can erroneously be interpreted as being the result of lifetime reduction due to energy transfer. At the same time, weak or transient protein associations may get masked and thus remain undetected. Whereas the parental wild-type GFP is extremely photostable and shows a monoexponential decay pattern (excited-state lifetime 3.16 ± 0.03 ns; Striker et al., 1999; Volkmer et al., 2000; Shaner et al., 2005), its close spectral overlap with YFP makes it unsuitable as a donor in GFP-YFP FRET experiments. Likewise, wild-type or enhanced GFP (or YFP) as a donor to red-shifted monomers as acceptors is suboptimal because the 488-nm (or 514-nm) laser line commonly used to excite GFP (or YFP) cross excites most of the red monomers (e.g. mOrange, mStrawberry) because of their broad excitation spectra (Zapata-Hommer and Griesbeck, 2003; Shaner et al., 2004).Recently, TSapphire (Q69M/C70P/V163A/S175G; excitation/emission 399/511 nm), a variant of the Sapphire (T203I) mutant of wild-type GFP with improved folding properties and better pH sensitivity, was described (Zapata-Hommer and Griesbeck, 2003). The T203I mutation in TSapphire (and original Sapphire as well) abolishes the 475-nm excitation peak found in the wild-type GFP (Tsien, 1998). TSapphire is efficiently excited below 410 nm, which makes it ideal for studying plant protein dynamics and interactions because, at this wavelength, there is negligible excitation of the autofluorescing chlorophyll pigments. Furthermore, TSapphire also represents a good donor to red monomer acceptors that are negligibly excited at this wavelength (Shaner et al., 2004). Using a purified Zn²⁺ sensor with TSapphire and mOrange as a donor-acceptor pair, Shaner and colleagues demonstrated the ratiometric intramolecular FRET between the two fluorophores in vitro (Shaner et al., 2004). The sensor yielded a 6-fold ratiometric increase (562/514-nm mOrange/TSapphire emission ratio) upon Zn²⁺ binding.However, currently there are no reports demonstrating the application and use of TSapphire and monomeric red-shifted fluorophores as donor-acceptor FRET pairs to probe intermolecular protein-protein interactions in vivo. In this article, we demonstrate in vivo FRET-fluorescence lifetime imaging microscopy (FLIM) between the donor TSapphire and the acceptor mOrange. We show that TSapphire expressed in living plant cells decays with a monoexponential lifetime of 2.93 ± 0.09 ns, which is in agreement with the published lifetime for its parent wild-type GFP (3.2 ns; Striker et al., 1999; Volkmer et al., 2000). Furthermore, we demonstrate intramolecular FRET-FLIM between chimerically linked TSapphire and mOrange (with a 16-amino acid linker in between). When fused to genetically known interacting proteins and expressed in intact living cells, the donor and the acceptor fluorophores show energy transfer in different subcellular compartments indicative of intermolecular protein-protein interactions. These results validate the versatility of the proposed in vivo FRET-FLIM assay based on the donor TSapphire and the acceptor mOrange, which turns out to work with both soluble and membrane proteins.