ABSTRACT
Förster resonance energy transfer (FRET) microscopy is a valuable methodology to infer distances of 1-10 nm in living specimens, greatly surpassing the resolution of light at ~200 nm and doing so with virtually any light microscopy system. The availability of different types of fluorophores and imaging methodologies allows for a tailored approach to specific research goals. This chapter describes the use of the three main fluorophore categories: organic dyes, conjugated to a protein of interest; GFP-type fusion proteins expressed in living specimens; and semiconductor nanocrystals, widely known as quantum dots (QDs), conjugated to target proteins. Each category presents its own challenges to the researcher, and we
stress therefore the importance of the judicious choice of a FRET pair by considering its spectral and other properties. FRET microscopy methodologies cover intensity-based emission, filter-based or spectral-based confocal imaging, and fluorescence lifetime imaging (FLIM-FRET)-based approaches. We present actual research data with extensive background information for each of the fluorophore choices and imaging methodologies, with emphasis on QDs. 10.1 IntroductionFörster resonance energy transfer (FRET) is a valuable methodology to conduct measurements at the nanometer scale in the life sciences. FRET imaging assays the distance between fluorophorelabeled molecules within 1-10 nm. Live and fixed cells-grown in suspension or attached to culture plates-microorganisms, and tissue specimens, as well as purified molecules and other types of biological samples have been used in FRET imaging. A wide variety of FRET-based biological experiments have been developed to determine proximities within macromolecular assemblies, such as signaling complexes and receptor-ligand clusters in cellular membrane compartments.1-5 Depending on the research objectives, a wide range of different microscopy systems can be used for FRET: widefield, confocal, spectral, fluorescence lifetime imaging (FLIM), and anisotropy. The increasing availability of pH-and photo-stable organic dyes with high quantum yields, the ever-growing family of green fluorescent protein (GFP)-type fluorescent fusion proteins, and latterly the development of semiconductor nanocrystals, widely known as quantum dots (QDs), have accelerated the use and widened the utility of FRET microscopy. The recent introduction of spectral imaging to measure FRET events has increased even more the choice of fluorophores and FRET pairs, opening the door to the development of multiplex FRET imaging. Nevertheless, the selection of a fluorophore FRET pair remains a careful consideration that will be influenced by the research goals and availability of suitable instruments; specimen and imaging parameter optimization steps are also important to get the full benefit of FRET microscopy. QDs in particular are still at an early development stage for live-specimen microscopy, and their use should be considered in light of both their advantages and pitfalls, as described in this chapter. Therefore, FRET
microscopy is a particularly useful technique to take advantage of the ever-increasing opportunities to conduct nanometer-scale measurements in live biological specimens. Since there are many excellent publications that cover FRET microscopy in detail2,6-9, we will focus here on the use of different types of fluorophore FRET pairs, such as, organic dyes, GFP-like fluorescent proteins (FPs), and QDs in biological applications. 10.2 FRET MicroscopyIn the late 1940s, Förster proposed the theory of FRET, which described how energy could be nonradiatively transferred directly from a fluorophore in the excited state (the donor, D) to a nonidentical neighbor fluorophore (the acceptor, A) (10-12). The energy from a donor molecule can be transferred directly to an acceptor molecule under the following four conditions: (i) when the emission spectrum of the donor overlaps significantly with the absorption spectrum of the acceptor, (ii) when the two fluorophores are within ~1 to ~10 nm of each other, (iii) when the donor emission and acceptor absorption dipole moments are in favorable mutual orientations, and (iv) when the donor has a reasonably high quantum yield. For a FRET system composed of a single donor and a single acceptor, the efficiency of the energy transfer from the donor to the acceptor (E) is described as the ratio of energy transfer rate (kT) to the sum of all deactivation rates of the excited state of the donor (kT + kD), where kD is the sum of its deactivation rates other than FRET (Eq. 10.1): E = kT/(kT + kD) (10.1) E varies inversely with the sixth power of the distance separating the donor and acceptor molecules (r), as described by Eq. 10.2 (12;13): E = R06/(R06 + r6) ´ r = R0[(1/E) – 1]1/6 (10.2) R0 is the Förster distance for the FRET pair, where E is equal to 50%, and it is described by Eq. 10.3:
R0 = 0.211 ◊ {k2 ◊ n-4 ◊ QYD ◊ J}1/6, J = eAf f d f d
( ) ( )( )l l l ll l 40
(10.3)
where κ2 ranging from 0 to 4 is the dipole orientation factor and usually assumed to be 2/3 (the average value integrated over all possible angels), n is the refractive index of the medium, QYD is the donor quantum yield, and J expresses the degree of spectral overlap between the donor emission and the acceptor absorption (14). Moreover, εA is the extinction coefficient of the acceptor at its peak absorption wavelength, λ is the wavelength, fD(λ) is the donor emission spectrum, and fA(λ) is the normalized acceptor absorption spectrum (peak absorbance = 1). Although the distance over which FRET can occur is limited to the 1-10 nm range, in practice robust energy transfer can only be detected when the distance between donor-and acceptor-labeled molecules lies within the 2-8 nm range, depending on the microscopy system’s capabilities and on the donor-acceptor R0 values. This FRET-based inferred proximity between two labeled cellular components considerably surpasses the resolution of general light microscopy, which at best can resolve distances of ~200 nm. Since in the majority of biological applications, donor and acceptor molecules are not separated by a fixed distance at a 1:1 ratio, nanoscale FRET measurements can only provide relative distances between donor-and acceptor-labeled molecules generated from averages of spatial distribution maps. Nevertheless, relative distances are extremely valuable information when following proximity changes within molecular assemblies, since the distance between interacting proteins falls within the 2-8 nm FRET distance range. Given a suitable donor-acceptor fluorophore pair, FRET can be executed on different microscopy systems using distinct methodological approaches with the goal of determining the relative measurement of nanometer-scale distances between donor-and acceptor-labeled molecules. Many FRET microscopy techniques have been developed and are generally categorized into intensity-based and lifetime-based methods. Steady-state intensity-based measurements performed under identical imaging conditions are typically used to measure the “apparent” energy transfer efficiency (E%) as shown in Eq. 10.4-called “apparent,” because most of intensity-based FRET microscopy techniques cannot differentiate between the donor molecules that participate in FRET events (FRET donors) and those that do not (non-FRET donors). In Eq. 10.4, qD represents the average intensities of the quenched donor in the presence of the acceptor (i.e., samples
containing donor-and acceptor-labeled molecules) and D is the average intensities of the unquenched donor in the absence of the acceptor (i.e., samples containing only donor-labeled molecules). E% = 100 ¥ (1 – qD/D) (10.4) Two main intensity-based approaches are routinely applied in FRET studies using most of the standard microscope systems available, such as wide-field, confocal, spectral, or total internal reflection fluorescence (TIRF). One is acceptor photobleaching (AP) or donor de-quenching FRET microscopy, where the E% (Eq. 10.4) is directly estimated by measuring the donor intensities before (quenched, qD) and after (unquenched, D) photobleaching the acceptor in samples containing donor-and acceptor-labeled molecules (15;16). This method is straightforward and can be used to collect measurements of FRET events in live cells under certain conditions, including reduced diffusion of the donor and acceptor molecules, quick and efficient bleaching of the acceptor molecules, and reduced acceptor-donor photoconversion, which has been shown to occur even in the well-known CFP-EYFP FRET pair (17). In the second intensity-based FRET methodology, including filter-based and spectral FRET, the unquenched donor signal (D) cannot be directly measured from samples containing donor-and acceptor-labeled molecules; instead it is usually determined by adding the energy transfer levels estimated from the acceptor intensities upon donor excitation (acceptor sensitized emission) to the measured quenched donor signal (qD) to calculate E% (Eq. 10.4). However, accurate quantification of the energy transfer levels requires the removal of the spectral bleedthrough (SBT) contaminations, which are a direct result of the spectral overlap required for FRET events. Algorithms have been developed for various microscopy techniques to identify and remove the SBT contaminations, allowing the accurate measurement of the E%. In filter-based FRET microscopy, the donor, acceptor, and FRET signals are measured in channels separated using band-pass filters (18-25). In contrast, spectral FRET microscopy uses a spectral detector to measure signals acquired over a continuous emission spectrum, followed by the use of spectral linear unmixing to separate the signals emitted from the donor and acceptor (2629). Both filter-based and spectral FRET microscopy can be used for time-lapse live-cell FRET imaging involving relatively complex postcollection image processing. The spectral FRET approach has the
advantage of allowing more flexibility in the choice of fluorophores, such as QDs (30). Moreover, spectral imaging can also be used to develop multiplexed FRET approaches to follow multiple signaling events in live cells by using more than one donor-acceptor pair (31). In lifetime-based FRET microscopy, FRET events can be identified by measuring the reduction in the donor lifetime that results from quenching, and the energy transfer efficiency (E) is estimated from the donor lifetimes determined in the absence (tD – unquenched lifetime) and the presence (tDA – quenched lifetime) of the acceptor based on Eq. 10.5 (14): E = 1 – (tDA/tD) (10.5) The fluorescence lifetime refers to the average time a molecule stays in its excited state before emitting a photon and is an intrinsic property of a fluorophore; it can be measured in different ways, and the experimental techniques are generally divided into time and frequency domain methods (32). In the time domain, a fluorophore is usually excited with a pulsed light source and its decay profile is directly measured at different time windows using high-speed detectors plus fast synchronization electronics; the fluorescence lifetime of the fluorophore is estimated from analyzing the recorded decay profile. In the frequency domain, the excitation light is modulated at different frequencies, which are chosen upon the lifetime scale of a fluorophore to be measured (megahertz for nanosecond decays). At each frequency, the phase shift and amplitude attenuation (modulation) of the fluorescence emission relative to the phase and am-plitude of the excitation light are measured. Analyzing the recorded phase shifts or amplitude attenuations or both can estimate the fluorescence lifetime of the fluorophore. In FLIM-FRET microscopy, SBT is not an issue since only the donor signals are monitored, and the fluorescence lifetime is not influenced by probe concentration, exci-tation light intensity, or light scattering; these facts make FLIM-FRET microscopy an accurate method to measure nanoscale distance distributions. However, fluorescent lifetimes are sensitive to environmental conditions. Thus, in live cells, for example, fluorophores may be exposed to different pH values and therefore may show multiple lifetime values, increasing the complexity of FRET measurements using FLIM. Changes in the fluorescence lifetime of fluorescence pro-teins have also been detected (33) and should be considered when
using FLIM to image FRET between fluorescence proteins in live cells. 10.3 Choosing FRET Pairs
Considering only the fluorescence properties to choose fluorophores for FRET is not a trivial decision. Quantum yields, extinction coefficients, degree of donor-acceptor spectral overlap, and SBT levels, all play a role to be considered and to be optimized. Nevertheless, the final selection of the right donor-acceptor pair should also include the actual biological question to be addressed, the type of biological specimen to be imaged, and the instrument available to measure FRET levels. After selection, optimization steps should be performed to attain sufficient intensity/lifetime levels of donor and acceptor molecules that should result in E levels above background/negative controls. This would entail experimenting with different cell lines, fixed or live specimens, fluorescence intensity levels of donor-and acceptor-labeled proteins due to different expression or labeling levels, different imaging conditions such as laser power, gain, and image collection speed, and other variables. Here we will review FRET methodological approaches from the point of view of different donor-acceptor pairs instead of the more standard approach that considers different microscopy-based approaches. Fluorophores broadly fall into three categories: organic dyes, FPs, and inorganic semiconductor nanocrystal QDs. Organic dyes are small and well established, and the more recent Alexa Fluor (AF) and Cy families of fluorophores are quite photostable at a wide range of pH and temperature. FPs are normally overexpressed by the cells, and new, brighter varieties with higher quantum yields are added constantly. The newer inorganic semiconductor nanocrystal QDs have been used mostly for in vitro FRET biological assays. QDs are still in the early application phases of biomedical FRET imaging but hold considerable promise and are described here in greater detail. We will focus on in vivo FRET applications using organic dye pairs (e.g., AF488-AF555), new FPs variants (e.g., mTFP-KO2), and QD-organic dye pairs (e.g., QD566-AF568 and QD580-AF594). An important FRET property that should considered before the donor-acceptor pair selection is their Förster distances (R0), which can be estimated from the photophysical properties of the donor and
acceptor dyes based on Eq. 10.3. Different FRET pairs have different R0 with the majority being ~5-6 nm (30;34;35). However, it is difficult to obtain the accurate R0 value in some cases. For example, different values of quantum yield of the particular dye-protein conjugate may alter the R0 value (36), which is normally estimated based on the quantum yield of the dye itself. Moreover, the dipole orientation factor assumed to be 2/3 may not reflect the truth (14), although it has been recently demonstrated that the dipole-dipole interactions mechanism used in R0 calculationsaccurately describes the energy transfer process in QD-dye FRET pairs and that the point dipole approximation is correct for QD donors in FRET reactions (35). As mentioned above, the practical experimental range of FRET measurements is ~2-8 nm, depending on the microscopy system’s capabilities and on the R0 values of the selected FRET pairs. Therefore, the choice of a FRET pair and its R0 can play a role in extending the distance range that will be assayed in a particular FRET-based assay. When choosing a FRET pair, one needs to keep in mind that a larger R0 will increase the likelihood of a FRET event. Increasing R0 may be achieved by using a donor with a higher quantum yield, an acceptor with a larger extinction coefficient, and a pair with a larger spectral overlap. Here, we have used three different types of donor-acceptor pairs with increasing R0 values: mTFP-mKO2 with R0 = 5.29 nm (37); QD566-AF568 and QD580-AF594 with R0 = 6.0 nm (30); andAF488-AF555 with R0 = 7.0 nm (https://www.invitrogen.com/site/us/en/home/References/Molecular-Probes-The-Handbook/tables/R0-values-for-some-Alexa-Fluor-dyes.html) (Fig. 10.1). These donor-acceptor FRET pairs can assay different ranges of donor-acceptor distances.
