From principles to practice: a comprehensive guide to FRET-FLIM in plants

1.1 Overview

Proteins are the most abundant biomolecules in cells, playing critical roles in nearly every biological process. Over the years, their structure, localization, and indispensable roles in structural support, enzymatic catalysis, transport and storage, signaling, and defense were vastly explored. Traditionally, protein biochemistry is studied by their extraction, purification, and in vitro characterization via diverse techniques, such as electrophoresis and chromatography. Additionally, other advanced techniques, such as X-ray crystallography and nuclear magnetic resonance (NMR) spectroscopy, allow the precise and accurate characterization of protein structures, providing insights into their functions. Later on, protein studies which were performed in simplified in vitro environments expanded to real-time observations within live systems. This breakthrough was made possible by the discovery and use of fluorescent proteins (FPs), such as the green fluorescent protein (GFP) from Aequorea victoria [1], [2]. FP-fusion and FP-based confocal microscopy allow real-time observations of protein localization and dynamics in diverse systems including plants [3]. In fact, plants present unique challenges due to their relatively high autofluorescence, arising from pigments such as chlorophyll and carotenoids in plastids, as well as lignin and other phenolic compounds in cell walls [4]. Nevertheless, advanced microscopes offering enhanced sensitivity, superior signal-to noise ratio, and broad spectral variability enable the detection of weak and often mixed signals, offering more accurate analyses of protein behavior in live plant systems.

Within cells, proteins do not exist as isolated entities, they rather interact and form dynamic interaction networks that regulate intricate biological processes. The discovery of diverse FPs with different excitation and emission spectra enabled tagging and visualizing multiple proteins within the same system, allowing the in vivo study of protein complexes. Yet, the co-localization of proteins does not necessarily indicate their physical interaction, it merely serves as a prerequisite, constrained by the diffraction limit (∼250 nm) inherent to light microscopy. Therefore, further techniques were developed to investigate protein–protein interactions (PPIs) in vivo, such as techniques based on Förster resonance energy transfer (FRET), which enables high precision measurements of PPIs beyond the diffraction limit [5]. In plants, FRET is now commonly used to investigate PPIs, including those among transcription factors, signaling proteins, and membrane transporters [6]. For instance, several studies unravelled the formation of homo- and heterodimers [7], [8], [9], [10], [11], as well as higher order complexes among proteins [12]. Additionally, functional analyses were performed to investigate which protein domains serve as interaction sites [13]. This has been achieved in various plant organelles, such as the plasma membrane [7], [8], Golgi stacks [10], [11], [14], [15], endoplasmic reticulum [9], plasmodesmata [16] and nuclei [13]. In this context, it is worth noting that despite the profound capabilities of super resolution microscopy, it faces several challenges in plants, including complex tissue structure, dynamic growth and movement, sensitivity to photobleaching, cell wall thickness, and high levels of autofluorescence.

1.2 FRET principles

FRET was described by Theodor Förster in 1948 and refers to the non-radiative transfer of energy from an excited donor fluorophore to an acceptor fluorophore [17]. The process begins when a donor fluorophore absorbs a photon, promoting its electrons to a higher energy state. In this excited state, the donor can transfer energy to an acceptor fluorophore through a quantum mechanical process known as dipole-dipole coupling. As the energy is transferred, the donor returns to its ground state, while the acceptor becomes excited and subsequently emits light.

However, FRET occurs only when there is an overlap between the emission spectrum of the donor and the excitation spectrum of the acceptor. Additionally, the two fluorophores should be in a very close vicinity (<10 nm), and their transition dipole moments should ideally be oriented in parallel. Upon FRET, the fluorescence intensity and lifetime of the donor decrease, while the fluorescence intensity of the acceptor increases. These changes in the fluorescence parameters are measurable, and they provide the foundation for using FRET to asses PPIs [18].

Accordingly, the FRET transfer rate (k _FRET ) is dependent on the distance between the donor and acceptor fluorophores (R) as well as their relative orientation. Therefore, any conformational changes in protein complexes impact this rate. Additionally, k _FRET is inversely related to the sixth power of the donor-to-acceptor separation distance (R₀), reflecting the dipole-dipole coupling mechanism. R₀ represents the Förster radius, which is the distance between the donor and acceptor at which the FRET efficiency is 50 %. τ _D represents the excited state lifetime of the donor fluorophore in the absence of the acceptor [19].

1.3 FRET-based techniques

FRET can be measured using mainly three microscopic methods. Here, we discriminate between the two intensity-based approaches sensitized emission FRET (SE-FRET) and FRET-Acceptor photobleaching (FRET-APB) and the non-intensity-based FRET-Fluorescence Lifetime Imaging Microscopy (FRET-FLIM). SE-FRET monitors the changes in donor and acceptor emission by acquiring either a complete spectrum or by using appropriate filters for donor and acceptor. In case the two fluorophores are not in close proximity, excitation of the donor does not lead to excitation of the acceptor and therefore no detectable emission of the acceptor. In the case of FRET, due to the radiation-free transfer from the donor to the acceptor, the emission of the acceptor will increase while the emission of the donor decreases. SE-FRET is often used for FRET-based biosensors to monitor real-time metabolite concentrations and cellular processes such as phytohormone and calcium signaling [20], [21], [22]. When using FRET-APB, the acceptor fluorophore is bleached, resulting in a measurable increase of the donor fluorescence in case of FRET. While FRET-APB is simple and practical, it has two main limitations. First, it involves photobleaching the acceptor which prevents repeated measurements on the same cell. Second, it is intensity-dependent as it relies on the expression levels of the FP-tagged proteins, which can be highly heterogeneous in biological systems. Interestingly, FRET-FLIM can surpass these challenges as it needs only low excitation power, reducing the risks of photobleaching and phototoxic processes. Also, it measures the fluorescence lifetimes of fluorophores, an intrinsic characteristic unique to each fluorophore, rather than their intensities. Therefore, FLIM enables studying the dynamics of proteins complexes with high spatial and temporal resolutions, offering more robust and reproducible in vivo measurements.

1.4 FRET pairs

Optimal FRET measurements require FRET pairs with spectral overlap, photostability, and compatibility with the plant autofluorescence [23]. For instance, ECFP and EYFP is a commonly used FRET pair with high spectral overlap. However, the donor ECFP is prone to photobleaching and is relatively dim due to its low quantum yield [24]. Also, it shows a multi-exponential fluorescence lifetime decay, i.e., its intensity decreases over time at varying rates, with each rate corresponding to a different fluorescence lifetime, complicating FRET-FLIM analyses. Therefore, mCerulean, which is brighter and in the same spectral range, presents an alternative FRET pair with mVenus [25]. Moreover, the FRET pair should not have crosstalk with the autofluorescence of the studied tissue or cellular compartment while preserving the correct localization of the fluorescent protein. A previous study compared 12 different FRET pairs in transient Nicotiana benthamiana experiments and showed that eGFP-mCherry is still a good choice [26]. However, several alternative FRET pairs have been tested in plants i.e. mVenus that could serve as a donor for mRuby3 or mKate2 as well as an acceptor for mTurquoise2. Furthermore, a combination of mTurquoise2 and mNeonGreen proved to be an efficient FRET pair [26]. Since improved versions of well-studied fluorophores are developed regularly, the FPbase (www.fpbase.org/) offers an invaluable tool for selecting potential FRET pairs as it provides detailed descriptions of fluorescent protein properties, including spectral attributes, FRET relationships, and compatibility with various filters and light sources (Figure 1) [27].

Figure 1:

Spectral characteristics and FRET parameters for mV and mCh. (A) Excitation and emission spectra of the fluorescent proteins mV (donor) and mCh (acceptor). The integral overlap area, J(λ), highlights the spectral overlap between mV emission and mCh excitation. Figure data obtained from the fluorescent protein database (FPbase) [27]. (B) Table summarizing key photophysical properties of mV and mCh including quantum yield (QY), extinction coefficient of the acceptor (EC_Acc), Förster radius (R₀, in Å), and overlap integral (J(λ)).

The following sections entail a detailed overview of the practical application of FRET-APB and FRET-FLIM in plants.

2.1 Materials

FRET-APB requires a microscope with a strong laser or appropriate illumination source, such as a confocal or wide-field microscope with suitable filter sets and mirrors to bleach the acceptor. The tools needed for preparing a plant sample include microscope slides, coverslips (#1.5, 170 μm thickness), razor blades or biopsy puncher, forceps, and fluorescence-free water. The plant material should express an appropriate FRET pair, such as mVenus and mCherry, for accurate FRET measurement. In the absence of stable transgenic lines, transient expression systems, such as N. benthamiana (N. benthamiana), offer a useful alternative for FRET experiments. To illustrate, N. benthamiana can be transformed with FRET constructs via Agrobacterium tumefaciens-mediated infiltration, allowing the visualization and analyses of proteins within a few days [29]. However, the expression of FRET constructs under strong constitutive or native promoters can lead to a rapid protein overexpression, creating artifacts that can influence FRET measurements. Therefore, the use of inducible expression systems, such as estradiol-inducible promoters, allows a precise control of gene expression preventing protein overaccumulation [30], [31].

2.2 Methods

3.1 Materials

FRET-FLIM requires a confocal microscope equipped with pulsed lasers, single-photon sensitive detectors such as avalanche photodiodes (APDs), hybrid detectors (HyDs), or fast photo multiplier tubes (PMTs), in addition to the FLIM module, i.e., TCSPC (time-correlated single photon counting) electronics. Additionally, a power meter to measure the power of the excitation light at the objective is needed. The tools and plant material are the same as in FRET-APB. Additionally, rhodamine 110 or Atto425, and erythrosine B solutions are needed for calibration. Rhodamine 110 has a stable fluorescence lifetime (4 ns in water), a high quantum yield, photostability, and it shows a mono-exponential decay behavior. These measurements are not necessary for the fitting process but are needed to verify the accuracy of the FRET-FLIM setup. When the donor fluorophore is in the blue spectrum, Atto425 can be used for a similar calibration. Atto425 has a shorter fluorescence lifetime (3.6 ns in water), making it convenient for comparisons with donor fluorophores of shorter lifetimes. Erythrosine B has an extremely short lifetime and can therefore be used to record the instrument response function (IRF). The IRF measurement reflects the influence of all components of the instrument on the incoming photons and is therefore individual for each setup. The IRF is later on used to deconvolve the whole decay histogram during the fitting process. Alternatively, a mirror can be used to record the IRF [38].

3.2 Methods

4.1 PLT3-WOX5 interaction using FRET-APB

PLT3 tagged with mVenus (PLT3-mV) and WOX5-tagged to mCherry (WOX5-mCh) were co-expressed in N. benthamiana leaves using A. tumefaciens-mediated transformation [13], [29]. PLT3-mV localized in the nucleus, where it formed biomolecular condensates. WOX5-mCh colocalized with PLT3-mV and was recruited into its condensates. FRET-APB was performed as described, measuring the fluorescence intensity of the two FP-tagged proteins before and after bleaching the WOX5-mCh acceptor (Figure 2A–C).

Figure 2:

FRET-APB experiment in nuclei of transiently expressing N. benthamiana leaf epidermis cells. PLT3-mV and WOX5-mCh intensity before (A–C) and after bleaching (A′–C′) of the acceptor. The white square indicates the bleached area. (D) Graphic representation of the fluorescence intensity of mV and mCh, the black arrow indicates the time point of bleaching. Scale bars represent 5 μm.

Before bleaching the acceptor, the intensity of the FP-tagged proteins slightly decreased due to acquisition bleaching, which should not exceed 10 %. Upon bleaching the acceptor, the intensity of WOX5-mCh dropped to nearly zero and then gradually recovered. However, PLT3-mV intensity increased after bleaching, indicating FRET occurrence (Figure 2D). Measurements were conducted on a minimum of ten nuclei, with the experiment repeated at least three times. Consequently, the FRET efficiency E _app calculated as per equation (2) was 21 %.

4.2 PLT3-WOX5 interaction using FRET-FLIM

PLT3-WOX5 interaction was further investigated via FRET-FLIM [12], [13], determining the intensity-weighted fluorescence lifetime measurements and the relative number of molecules undergoing FRET, known as binding, by using OPA (Figure 3, Table 1). First, WOX5-mV was expressed alone to measure the donor-only fluorescence lifetime. Second, WOX5-mV was co-expressed with free mCherry to control for background fluorescence, spectral crosstalk, or non-specific FRET between free mCherry and WOX5-mV. Finally, two FRET samples were analyzed, one with WOX5-mV and full-sequence PLT3-mCh which contains three PrDs, and the other expressing WOX5-mV and PLT3 with mutated or deleted PrDs (PLT3ΔPrD-mCh). In the PLT3ΔPrD-mCh construct, PrD1 at the N-terminus was substituted by a 27 amino acid linker, while PrD2 and PrD3 at the C-terminus were deleted, to test whether the PrDs serve as interaction sites between PLT3 and WOX5.

Figure 3:

Comparison of traditional FRET-FLIM analysis and OPA. Summary of quantification of (A) intensity weighted fluorescence lifetime [in ns] of WOX5-mV and (B) one pattern analysis. Statistical groups were assigned after non-parametric Kruskal Wallis with post-hoc Dunn’s test (α = 0.05, p-values adjusted after Benjamini and Hochberg). (C–F) Representative images of fluorescence lifetime imaging microscopy (FLIM) measurements of nuclei in N. benthamiana epidermal leaf cells after pixel-wise multiexponential fitting. The fluorescence lifetime of the donor WOX5-mV in absence or presence of the indicated acceptor (mCherry-NLS, PLT3-mCh or PLT3ΔPrD-mCh) is color-coded: blue (2.5) refers to low fluorescence lifetime [in ns], red (3.1) indicates high fluorescence lifetime. Scale bars represent 6 µm.

The fluorescence lifetime of WOX5-mV decreased significantly when co-expressed with PLT3-mCh, indicating FRET and thereby confirming an interaction between PLT3 and WOX5 (Table 2). This lifetime reduction was notably diminished when WOX5-mV was co-expressed with PLT3ΔPrD-mCh, suggesting that the PrDs in PLT3 are critical for the interaction between these two proteins (Figure 3A, C–F, Table 1).

Furthermore, using equation (3) allows for the calculation of the FRET efficiency based on the lifetime reduction and results in a FRET efficiency of around 10 %. This value is lower than that obtained from the FRET-APB approach. This discrepancy possibly results from the choice of the donor molecule, WOX5 or PLT3, respectively. PLT3 causes the formation of nuclear condensates that exhibit a strong fluorescence signal which can cause an artificial lifetime reduction when using detectors, that have a long dead time, triggering the so called pile-up effect. OPA provided further insights, showing that while FRET efficiency remained unchanged, WOX5 binding to full-length PLT3 was significantly stronger than to PLT3ΔPrD. This result reflects a higher affinity between the full-length proteins, highlighting the added detail and value OPA offers (Figure 3B, Tables 1, 3 and 4).