Resonance energy transfer microscopy: observations of membrane-bound fluorescent probes in model membranes and in living cells

A conventional fluorescence microscope was modified to observe the sites of resonance energy transfer (RET) between fluorescent probes in model membranes and in living cells. These modifications, and the parameters necessary to observe RET between membrane-bound fluorochromes, are detailed for a system that uses N-4-nitrobenzo-2-oxa-1,3-diazole (NBD) or fluorescein as the energy donor and sulforhodamine as the energy acceptor. The necessary parameters for RET in this system were first optimized using liposomes. Both quenching of the energy donor and sensitized fluorescence of the energy acceptor could be directly observed in the microscope. RET microscopy was then used in cultured fibroblasts to identify those intracellular organelles labeled by the lipid probe, N-SRh-decylamine (N-SRh-C10). This was done by observing the sites of RET in cells doubly labeled with N-SRh-C10 and an NBD-labeled lipid previously shown to label the endoplasmic reticulum, mitochondria, and nuclear envelope. RET microscopy was also used in cells treated with fluorescein-labeled Lens culinaris agglutinin and a sulforhodamine derivative of phosphatidylcholine to examine the internalization of plasma membrane lipid and protein probes. After internalization, the fluorescent lectin resided in most, but not all of the intracellular compartments labeled by the fluorescent lipid, suggesting sorting of the membrane-bound lectin into a subset of internal compartments. We conclude that RET microscopy can co-localize different membrane-bound components at high resolution, and may be particularly useful in examining temporal and spatial changes in the distribution of fluorescent molecules in membranes of the living cell.     

Covalent labeling of cell-surface proteins for in-vivo FRET studies

Fluorescence resonance energy transfer (FRET) is a powerful technique to reveal interactions between membrane proteins in live cells. Fluorescence labeling for FRET is typically performed by fusion with fluorescent proteins (FP) with the drawbacks of a limited choice of fluorophores, an arduous control of donor-acceptor ratio and high background fluorescence arising from intracellular FPs. Here we show that these shortcomings can be overcome by using the acyl carrier protein labeling technique. FRET revealed interactions between cell-surface neurokinin-1 receptors simultaneously labeled with a controlled ratio of donors and acceptors. Moreover, using FRET the specific binding of fluorescent agonists could be monitored.     

Effect of phospholipid:protein ratio on the state of aggregation of the (Ca2+-Mg2+)-ATPase

The organization of the (Ca2+-Mg2+)-ATPase has been studied in reconstituted systems by fluorescence polarization of the ATPase labeled with fluorescein isothiocyanate (FITC) and resonance energy transfer between ATPase labeled with FITC and with eosin isothiocyanate (EITC). The fluorescence polarization of FITC-ATPase was found to decrease with increasing labeling ratio FITC:ATPase, indicating depolarization as a result of resonance energy transfer between ATPase molecules. Fluorescence polarization was, however, independent of the molar ratio of phospholipid to protein above a molar ratio of 50:1. Resonance energy transfer between FITC-ATPase and EITC-ATPase was also found to be independent of phospholipid:protein ratio. It is suggested therefore that the ATPase is not randomly distributed in the plane of the membrane but rather forms ordered clusters (probably rows of monomers or dimers) on the fluorescence time scale (nanoseconds) even in the presence of a large excess of phospholipid. This organization within the membrane is dependent both on the chemical structure of the phospholipid and on its physical phase.     

Biochemical and biophysical characterization of serotonin 5-HT2C receptor homodimers on the plasma membrane of living cells

While many studies have provided evidence of homodimerization and heterodimerization of G-protein-coupled receptors (GPCRs), few studies have used fluorescence resonance energy transfer (FRET) combined with confocal microscopy to visualize receptor dimerization on the plasma membrane, and there have been no reports demonstrating the expression of serotonin receptor dimers/oligomers on the plasma membrane of living cells. In the study presented here, biochemical and biophysical techniques were used to determine if 5-HT(2C) receptors exist as homodimers on the plasma membrane of living cells. Immunoprecipitation followed by Western blotting revealed the presence of immunoreactive bands the predicted size of 5-HT(2C) receptor monomers and homodimers that were detergent and cross-linker sensitive. Bioluminescence resonance energy transfer (BRET) was assessed in HEK293 cells expressing 5-HT(2C) receptors labeled with Renilla luciferase and yellow fluorescent protein. BRET levels were not altered by pretreatment with serotonin. Confocal microscopy provided direct visualization of FRET on the plasma membrane of live cells expressing 5-HT(2C) receptors labeled with cyan (donor) and yellow (acceptor) fluorescent proteins. FRET, assessed by acceptor photobleaching, was dependent on the donor/acceptor ratio and independent of acceptor expression levels, indicating that FRET resulted from receptor clustering and not from overexpression of randomly distributed receptors, providing evidence for GPCR dimers/oligomers in a clustered distribution on the plasma membrane. The results of this study suggest that 5-HT(2C) receptors exist as constitutive homodimers on the plasma membrane of living cells. In addition, a confocal-based FRET method for monitoring receptor dimerization directly on the plasma membrane of living cells is described.     

Physical association between the adipocyte fatty acid-binding protein and hormone-sensitive lipase: a fluorescence resonance energy transfer analysis

Previous in vitro studies have established that hormone sensitive lipase (HSL) and adipocyte fatty acid-binding protein (AFABP) form a physical complex that presumably positions the FABP to accept a product fatty acid generated during catalysis. To assess AFABP-HSL interaction within a cellular context, we have used lipocytes derived from 293 cells (C8PA cells) and examined physical association using fluorescence resonance energy transfer. Transfection of C8PA cells with cyan fluorescent protein (CFP)-HSL, yellow fluorescent protein (YFP)-adipocyte FABP, or YFP-liver FABP revealed that under basal conditions each protein was cytoplasmic. In the presence of 20 microm forskolin, CFP-HSL translocated to the triacylglycerol droplet, coincident with BODIPY-FA labeled depots. Fluorescence resonance energy transfer analysis demonstrated that CFP-HSL associated with YFP-adipocyte FABP in both basal and forskolin-treated cells. In contrast, little if any fluorescence resonance energy transfer could be detected between CFP-HSL and YFP-liver FABP. These results suggest that a pre-lipolysis complex containing at least AFABP and HSL exists and that the complex translocates to the surface of the lipid droplet.     

Flow cytometric detection of transbilayer movement of fluorescent phospholipid analogues across the boar sperm plasma membrane: elimination of labeling artifacts

Reliable protocols were established for investigating asymmetric distributions of 6-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino-caproyl (C6NBD) phospholipids in the plasma membrane of boar sperm cells under physiological conditions. A method based on fluorescence resonance energy transfer was used to ensure that incorporation of the fluorescent phospholipids into the sperm proceeded via monomeric transfer. The total amount of incorporated phospholipid fluorescence and the proportion of translocated phospholipid fluorescence were determined by flow cytometric analysis before, and after, dithionite destruction of outer leaflet fluorescence. Catabolism of incorporated fluorescent phospholipids was blocked with phenylmethylsulfonyl fluoride. Membrane-damaged cells were detected with impermeant DNA stains, thereby enabling their exclusion from subsequent analyses of the flow cytometric data, whence it could be demonstrated that the labeled phospholipids were incorporated only via the outer plasma membrane leaflet in living sperm cells. Phospholipid uptake and internalization was followed at 38 degrees C. After 1 hr of labeling, about 96% of the incorporated C6NBD-phosphatidylserine, 80% of C6NBD-phosphatidylethanolamine, 18% of C6NBD-phosphatidylcholine, and 4% of C6NBD-sphingomyelin were found to have moved across the plasma membrane bilayer to the interior of the spermatozoa. These inward movements of fluorescent phospholipids were ATP-dependent and could be blocked with sulfhydryl reagents. Movements from the inner to the outer leaflet of the sperm plasma membrane were minimal for intact fluorescent phospholipids, but were rapid and ATP-independent for fluorescent lipid metabolites. The described method enables, for the first time, assessment of changes in lipid asymmetry under fertilizing conditions.     

Visualization of epidermal growth factor (EGF) receptor aggregation in plasma membranes by fluorescence resonance energy transfer. Correlation of receptor activation with aggregation

Fluorescence resonance energy transfer between epidermal growth factor (EGF) molecules, labeled with fluorescent reporter groups, was used as a monitor for EGF receptor-receptor interactions in plasma membranes isolated from human epidermoid A431 cells. Epidermal growth factor molecules labeled at the amino terminus with fluorescein isothiocyanate served as donor molecules in these energy transfer measurements, while EGF molecules labeled with eosin isothiocyanate at the amino terminus served as the energy acceptors. Both of these derivatives were shown to be active in binding to membrane receptors and in the activation of the endogenous receptor/tyrosine kinase activity. We found that membranes in the absence of added metal ion activators showed relatively little energy transfer (approximately 10% donor quenching) between the labeled growth factors. However, divalent metal ion activators of the EGF receptor/tyrosine kinase caused a significant increase in the extent of energy transfer between the labeled EGF molecules. Specifically, in the presence of 20 mM MgCl2, the extent of quenching of the donor fluorescence increased to 25% (from 10% in the absence of metal), while in the presence of 4 mM MnCl2, the extent of energy transfer was increased still further to 40-50%. The addition of an excess of EDTA resulted in the reversal of the observed energy transfer to basal levels. The increased energy transfer in the presence of these divalent cations correlated well with the ability of these metals to stimulate the EGF receptor/tyrosine kinase activity. However, the extent of receptor-receptor interactions measured by energy transfer was independent of receptor autophosphorylation. Overall, these results suggest that conditions under which the EGF receptor is primed to be active as a tyrosine kinase, within a lipid milieu, result in an increased aggregation of the receptor.     

Proximity of lectin receptors on the cell surface measured by fluorescence energy transfer in a flow system.

Molecules of the lectin concanavalin A have been labeled separately with the fluorescein and rhodamine chromophores and jointly bound to the surface of transformed Friend erythroleukemia cells. The two dyes constitute an ideal donor-acceptor pair for fluorescence resonance energy transfer thereby permitting the determination of the proximity relationships between bound ligand molecules and the corresponding surface receptors. The transfer efficiency at saturation (about 57%) was measured in a multiparameter flow system using laser excitation at 488 nm and detection of fluorescein and rhodamine emission intensities as well as the emission anisotropy of the rhodamine fluorescence for each cell. The degree of energy transfer was estimated from the quenching of donor emission, the sensitization of acceptor emission, and the depolarization of acceptor fluorescence. The system has been modeled according to a formalism developed by Gennis and Cantor (Biochemistry 11: 2509, 1972). We estimate the separation between the surfaces of bound lectin molecules at saturation to be 0-40 A, a range possibly characteristic for micropatches induced by ligand binding.     

Voltage-sensitive cyanine dye fluorescence signals in lymphocytes: plasma membrane and mitochondrial components

The origin of the cyanine dye fluorescence signal in murine and human peripheral blood leukocytes was investigated using the oxa- and indo-carbocyanines di-O-C5(3) and di-I-C5(3). Fluorescence signals from individual cells suspended with nanomolar concentrations of the dyes were measured in a flow cytometer modified to permit simultaneous four-parameter analysis (including two-color fluorescence or fluorescence polarization measurements). The contributions of mitochondrial membrane potential (psi m) and plasma membrane potential (psi pm) to the total voltage-sensitive fluorescence signal were found to depend on the equilibrium extracellular dye concentration, manipulated in these experiments by varying the ratio of dye to cell density. Hence, conditions could be chosen that amplified either the psi m or the psi pm component. Selective depolarization of lymphocytes or polymorphonuclear leukocytes (PMN) in mixed cell suspensions demonstrated that defining the partition of dye between cells and medium is requisite to assessing the heterogeneity of cell responses by cyanine dye fluorescence. At extracellular dye concentrations exceeding 5 nM in equilibrated cell suspensions, both mitochondrial and plasma membrane dye toxicity were observed. In murine splenic lymphocytes, plasma membrane toxicity (dye-induced depolarization) was selective for the B lymphocytes. Certain problems in calibration of psi pm with valinomycin at low dye concentrations and perturbations of psi pm by mitochondrial inhibitors are presented. These findings address the current controversy concerning psi m and psi pm measurement in intact cells by cyanine dye fluorescence. The finding of selective toxicity at low cyanine dye concentrations suggest that purported differences in resting psi m among cells or changes in psi pm with cell activation may reflect variable susceptibility to dye toxicity rather than intrinsic cell properties.     

Imaging protein-protein interactions using fluorescence resonance energy transfer microscopy

Fluorescence resonance energy transfer (FRET) detects the proximity of fluorescently labeled molecules over distances >100 A. When performed in a fluorescence microscope, FRET can be used to map protein-protein interactions in vivo. We here describe a FRET microscopy method that can be used to determine whether proteins that are colocalized at the level of light microscopy interact with one another. This method can be implemented using digital microscopy systems such as a confocal microscope or a wide-field fluorescence microscope coupled to a charge-coupled device (CCD) camera. It is readily applied to samples prepared with standard immunofluorescence techniques using antibodies labeled with fluorescent dyes that act as a donor and acceptor pair for FRET. Energy transfer efficiencies are quantified based on the release of quenching of donor fluorescence due to FRET, measured by comparing the intensity of donor fluorescence before and after complete photobleaching of the acceptor. As described, this method uses Cy3 and Cy5 as the donor and acceptor fluorophores, but can be adapted for other FRET pairs including cyan fluorescent protein and yellow fluorescent protein.     

Applications of fluorescence resonance energy transfer for mapping biological membranes

The interaction of the cell surface proteins plays a key role in the process of transmembrane signaling. Receptor clustering and changes in their conformation are often essential factors in the final outcome of ligand receptor interactions. Fluorescence resonance energy transfer (FRET) is an excellent tool for determining distance relationships and supramolecular organization of cell surface molecules. This paper reviews the theoretical background of fluorescence resonance energy transfer, its flow cytometric and microscopic applications (including the intensity based and photobleaching versions), and provides a critical evaluation of the methods as well. In order to illustrate the applicability of the method, we summarize a few biological results: clustering of lectin receptors, cell surface distribution of hematopoietic cluster of differentiation (CD) molecules, and that of the receptor tyrosine kinases, conformational changes of Major Histocompatibility Complex (MHC) I molecules upon membrane potential change and ligand binding.     

Flow cytometric detection of membrane potential changes in murine lymphocytes induced by concanavalin A.

The effect of the mitogenic lectin concanavalin A on the membrane potential of murine lymphocytes was investigated by observing the fluorescence of cells stained with carbocyanine and oxonol dyes. We describe a rapid and reliable method for detecting lectin-induced membrane potential changes in individual cells by flow cytometric analysis of oxonol fluorescence. By 10 min after addition of lectin to suspensions of isolated cells from lymph node, 7-15% of the cells have responded by releasing oxonol dye, indicating a membrane hyperpolarization. The dose onset of this response is similar to that for mitogenesis, which was assessed by measuring [3H]thymidine incorporation. The effect is abolished by alpha-methyl mannoside (100mM), which prevents concanavalin A from binding to the cells, but not by fucose (100mM). When cells are treated with lectin in medium from which Ca2+ has been omitted or to which quinine (0.5mM) has been added, a membrane depolarization is observed. Since these are conditions under which activation of plasma membrane Ca2+-dependent K+ channels is prevented, these findings support the view that the early hyperpolarization of these cells is brought about by an increase in intracellular free [Ca2+].     

The use of the potential-sensitive fluorescent probe bisoxonol in mast cells.

The regulation of the plasma membrane potential of rat peritoneal mast cells at the resting state and during activation was investigated using bisoxonol as a potential-sensitive fluorescent dye. Fluorescence microphotography showed that this negatively charged probe was not only present in the plasma membrane, but was also distributed in the cytoplasm. The intracellular localization of bisoxonol was confirmed by conducting experiments which showed that bisoxonol fluorescence was not enhanced in ATP-permeabilized mast cells. Rotenone (10(-7) M) and oligomycin (10(-6) M) did not change the fluorescence of bisoxonol showing, therefore, mitochondrial depolarization was not recorded with bisoxonol and suggesting that bisoxonol may represent a useful probe to study plasma membrane potential changes in the absence of exocytosis. We showed that, in non-stimulated mast cells, the blockade of the sodium pump enhanced the fluorescence of bisoxonol as did gramicidin a non selective ionophore used to fully depolarize the cells. High concentration of potassium (30 mM) as well as different ionic channel blockers did not significantly change the fluorescence intensity of bisoxonol, suggesting that ionic channel permeabilities were not involved in maintaining the resting plasma membrane potential of mast cells. Mast cells stimulated by compound 48/80 completely lost the fluorescence, shown by fluorescence microphotography, suggesting that exocytotic phenomena might induce a dye redistribution which is not only due to changes in the plasma membrane potential. In mast cells pretreated with pertussis toxin, which blocks mast cell-exocytosis, compound 48/80 induced a delayed (2 min) decrease of bisoxonol fluorescence which was shown to be dependent on the activity of the sodium pump. Considering that bisoxonol is a useful potential-sensitive probe in exocytosis-deprived mast cells, our results suggest that the sodium pump is mainly involved in the changes of plasma membrane potential of mast cells.     

Nod factors integrate spontaneously in biomembranes and transfer rapidly between membranes and to root hairs, but transbilayer flip-flop does not occur.

Three novel nodulation (Nod) factors were synthesized from chitotetraose and three structurally different fluorescent BODIPY-tagged fatty acids. With fluorescence spectroscopic and microscopic techniques, the following aspects were studied: whether these amphiphilic molecules insert in membranes, whether they transfer between different membranes, and whether they are able to transfer from a membrane to a legume root hair. Fluorescence correlation spectroscopy showed that fluorescent Nod factors are present as monomers in PBS buffer at a concentration of 10 nM, but that when either Triton X-100 micelles or dioleoylphosphatidylcholine (DOPC) vesicles are present, the Nod factors are associated with these particles. With time-correlated single-photon counting fluorescence spectroscopy, it was shown that upon Nod factor insertion in the membrane, the rotation of the fluorescent acyl chain was markedly reduced. A fluorescence resonance energy transfer assay was used to study the transfer of Nod factors from one membrane to the other, or from vesicles to root hairs. Nod factors transfer rapidly between membranes or from vesicles to root hair cell walls. However, they do not flip-flop between membrane leaflets. The results provide novel insights for the mode of secretion and transfer of Nod factors during the early steps of the Rhizobium-legume interaction.     

Viral DNA packaging studied by fluorescence correlation spectroscopy

The DNA packaging machinery of bacteriophage T4 was studied in vitro using fluorescence correlation spectroscopy. The ATP-dependent translocation kinetics of labeled DNA from the bulk solution, to the phage interior, was measured by monitoring the accompanied decrease in DNA diffusibility. It was found that multiple short DNA fragments (100 basepairs) can be sequentially packaged by an individual phage prohead. Fluorescence resonance energy transfer between green fluorescent protein donors within the phage interior and acceptor-labeled DNA was used to confirm DNA packaging. Without ATP, no packaging was observed, and there was no evidence of substrate association with the prohead.     

Membrane fluidity in normal and cystic fibrosis fibroblasts

We have observed distinct differences in the polarization of fluorescence and temperature dependent emission intensity of the highly fluorescent phospholipid derivative (1-acyl-2-(N-4-nitrobenzo-2-oxa-1,3-diazole)--aminocaproyl phosphatidylcholine (NBD-PC), when incorporated in the plasma membranes of normal and cystic fibrosis fibroblasts. Fluorescence polarization measurements indicate that the fluorochrome has a much higher degree of rotational mobility in cystic fibrosis fibroblasts as compared with normal cells. Temperature dependent transitions in the emission intensity of NBD-PC incorporated in normal fibroblasts are indicated at 17.7 and 21.2° C while the abnormal cell membranes apparently undergo transitions at 8.7 and 13.5° C. These differences might be due to changes in plasma membrane composition and/or organization, in the case of the cystic fibrosis cells.     

Intensity-based energy transfer measurements in digital imaging microscopy

Investigation of protein-protein associations is important in understanding structure and function relationships in living cells. Using Förster-type resonance energy transfer between donor and acceptor labeled monoclonal antibodies we can assess the cell surface topology of membrane proteins against which the antibodies were raised. In our current work we elaborated a quantitative image microscopic technique based on the measurement of fluorescence intensities to calculate the energy transfer efficiency on a pixel-by-pixel basis. We made use of the broad excitation and emission spectrum of cellular autofluorescence for background correction of images. In addition to the reference autofluorescence images (UV background) we recorded three fluorescent images (donor, acceptor and energy transfer signal) of donor-acceptor double labeled samples, and corrected for spectral spillage of the directly excited donor and acceptor fluorescence into the energy transfer image. After careful image registration we were able to calculate the energy transfer efficiency on a pixel-by-pixel basis. In this paper, we also present a critical comparison between results obtained with this method and other approaches (photobleaching and flow cytometric energy transfer measurements).     

Monitoring agonist-induced phospholipase C activation in live cells by fluorescence resonance energy transfer

Agonist-induced intracellular Ca(2+) signals following phospholipase C (PLC) activation display a variety of patterns, including transient, sustained, and oscillatory behavior. These Ca(2+) changes have been well characterized, but detailed kinetic analyses of PLC activation in single living cells is lacking, due to the absence of suitable indicators for use in vivo. Recently, green fluorescent protein-tagged pleckstrin homology domains have been employed to monitor PLC activation in single cells, based on (confocal) imaging of their fluorescence translocation from the membrane to the cytosol that occurs upon hydrolysis of phosphatidylinositol bisphosphate. Here we describe fluorescence resonance energy transfer between pleckstrin homology domains of PLCdelta1 tagged with cyan and yellow fluorescent proteins as a sensitive readout of phosphatidylinositol bisphosphate metabolism for use both in cell populations and in single cells. Fluorescence resonance energy transfer requires significantly less excitation intensity, enabling prolonged and fast data acquisition without the cell damage that limits confocal experiments. It also allows experiments on motile or extremely flat cells, and can be scaled to record from cell populations as well as single neurites. Characterization of responses to various agonists by this method reveals that stimuli that elicit very similar Ca(2+) mobilization responses can exhibit widely different kinetics of PLC activation, and that the latter appears to follow receptor activation more faithfully than the cytosolic Ca(2+) transient.     

Interaction of negatively charged liposomes with nuclear membranes: adsorption, lipid mixing and lysis of the vesicles

Fluorescence energy transfer studies reveal that negatively charged lipid vesicles interact with nuclei from mouse liver cells. This interaction was observed with charged lipid vesicles composed of PA or PS but not with the uncharged PC or PE:PC vesicles. The vesicles were prepared by bath sonication and contained either a fluorescent marker in the lipid bilayer or in the vesicular interior. The negatively charged vesicles showed an adsorption to the nuclear membrane visible by fluorescence microscopy. The results obtained by resonance energy transfer experiments are interpreted in terms of a mixing of the lipids from the vesicles with the nuclear membrane. Encapsulation studies documented a staining of the nuclei only if the dye molecules of high or low molecular weight were encapsulated inside negatively charged vesicles. As consequence of the vesicle-nuclei interaction morphological changes on the nuclear surface became visible.     

Xenopus oocyte plasma membrane sheets for FRET analysis

Plasma membrane sheets from Xenopus oocytes have been isolated for use in fluorescence resonance energy transfer (FRET) measurements. This system has the following advantages: 1) fluorescent recordings from a large surface area to maximize the signal-to-noise ratio, 2) reduction in background fluorescence from proteins retained in intracellular compartments, and 3) access to the cytoplasmic surface of the plasma membrane for rapid solution changes. To demonstrate the utility of this approach, we have examined a previously published FRET-based Ca²⁺ sensor, namely, the Cameleon-PM. This construct targets to the plasma membrane and, upon various Ca²⁺ additions to the cytoplasmic face of the membrane, shows ratiometric FRET changes. From the ratiometric changes recorded, an apparent Ca²⁺ affinity of 1.65 µM was determined. Thus preparation of Xenopus oocyte plasma membrane sheets and FRET measurements demonstrates all three of the advantages outlined above. fluorescence resonance energy transfer