Key words

1 Introduction

1.1 Chemokines and Their Receptors : A Dynamic System

Chemokines are a family of structurally related, low-molecular-weight pro-inflammatory cytokines involved in recruitment of specific cell populations to target tissues by interacting with members of the G protein -coupled receptor (GPCR ) family [1–4]. Although originally described as specific mediators of leukocyte directional movement , they are now implicated in a much wider variety of physiological and pathological processes including tumor cell growth and metastasis , atherosclerosis, angiogenesis , chronic inflammatory disease, and HIV -1 infection [2, 5–10].

In man, more than 50 chemokines and 20 receptors have been described and classified according to functional criteria as pro-inflammatory or homeostatic chemokines, depending on their role in inflammation or in immune system homeostasis, respectively [1, 11]. Although chemokines and their receptors were once considered independent, isolated entities, the situation is much more complex than initially anticipated. Chemokines can form dimers, tetramers, and even oligomers [12, 13]. They interact not only with chemokine receptors but can also bind cell surface glycosaminoglycans as well as non-signaling scavenger receptors [1, 14, 15]. Chemokine receptors interact with chemokines and also other molecules, such as defensins and virally encoded chemokines [2, 16–19].

Whereas some data indicate that GPCR family members can function as monomers [20], increasing experimental evidence indicates that these receptors form homodimeric and heterodimeric complexes at the cell membrane [21–25] not only with other chemokine receptors, but also with other GPCR (e.g., EBI2, delta opioid receptor) [26, 27] and other membrane proteins (e.g., CD4, CD26, T cell receptor, tetraspanins) [28–32]. These receptor complexes are functional entities that mediate biological responses and are associated with modulation of ligand binding and of G protein associations, and with activation of signaling events distinct from those triggered by individual receptors [33, 34]. Heterodimerization is the mechanism that underlies delayed AIDS progression in HIV -1 -infected patients bearing the CCR2V64I polymorphism [35]; CCR5 expression alters CD4/CXCR4 heterodimer conformation, thus blocking M-tropic HIV-1 binding and infection [28]. Heterodimers activate specific signaling cascades that differ from those activated by homodimers [33, 36]. In contrast to homodimer-triggered Gi signaling, CCR2/CCR5 heterodimers activate G11, which alters PI3K induction kinetics [33]. CCR5 and CXCR4 heterodimerization with opioid receptors modulates chemokine responses [27, 37, 38]. Oligomeric complexes can also regulate ligand affinity; for example, CXCR5 and EBI2 heterodimers reduce CXCL13 affinity for CXCR5, as well as subsequent Gi protein activation [26]. In this scenario, chemokine-mediated signaling properties depend on the receptor complex stabilized. This defines a very dynamic universe that offers ample possibilities for regulating chemokine function, and allows the design of innovative drugs to target specific chemokine-mediated functions without altering others.

Although biochemical approaches were initially used to determine chemokine receptor expression at the cell membrane as well as the signaling molecules involved in chemokine function, they render a static view of the system that can lead to misinterpretation of results. For example, western blot and coimmunoprecipitation showed CCR2 dimers only after ligand activation, which led us to describe an active role for ligands in receptor dimerization [23, 39]. Later studies using imaging-based techniques showed that dimers form in the absence of chemokines [21, 40, 41]. Our initial conclusion was thus incorrect; the difficulty in detecting these complexes by immunoprecipitation might be due to dimer conformational instability in the absence of ligand.

1.2 Resonance Energy Transfer Technology

Newer methods now in wide use for evaluating chemokine receptor oligomerization in living cells are based on resonance energy transfer (RET). These techniques are also useful for determining conformation dynamics, identifying the role of ligands and/or receptors in this process, and defining the dimerization site in the cell [42].

RET measurement allows identification of molecular interactions using techniques based on the quantitative theory developed by Förster in the 1940s. RET is a non-radioactive quantum mechanical process that neither requires electron collision nor involves heat production. There are two main types of RET, fluorescence resonance energy transfer (FRET ) and bioluminescence resonance energy transfer (BRET); in the former, the donor fluorochrome transfers energy to an acceptor fluorochrome and in the latter, the donor molecule is luminescent [43, 44].

Both methods require generation of fusion proteins between the receptor and the fluorescent/luminescent donor and acceptor proteins, as well as the use of transfected cells [45]. Although BRET has been used for single-cell analysis [46], it is a more appropriate approach for cell suspensions [47]. It allows measurement of energy transfer between receptors , independently of their expression pattern. Dimers at the plasma membrane cannot be distinguished from those being synthesized or trafficking through the endoplasmic reticulum. BRET saturation curves can nonetheless be quantitated, as reported for CCR5 /CCR2 heterodimers [48] and CXCR4 homodimers [40]. In contrast, FRET imaging using confocal microscopy allows measurements in single cells and identification of specific cell locations. FRET requires robust controls to discard direct acceptor activation by the light used to excite the donor, to eliminate nonspecific random collisions, and to monitor receptor overexpression [49].

2 Materials

2.1 Materials Common to all Techniques Described

  1. 1.

    HEK-293T cells (ATCC® CRL-11268™, Manassas, VA) (see Note 1 ).

  2. 2.

    DMEM culture medium containing 10 % fetal calf serum, 1 mM sodium pyruvate, and 2 mM l-glutamine (complete DMEM).

  3. 3.

    Tissue culture 6-well plates.

  4. 4.

    Cell incubator with 5%CO2.

  5. 5.

    Polyethylenimine (PEI, Sigma-Aldrich, St Louis, MO).

  6. 6.

    150 mM NaCl.

  7. 7.

    Hank’s balanced salt solution (HBSS) supplemented with 0.1 % glucose.

  8. 8.

    Bradford protein assay (Bio-Rad, Hercules, CA).

  9. 9.

    Black 96-well microplates (OptiPlate, PerkinElmer, Waltham, MA).

  10. 10.

    Multilabel fluorescent plate reader (EnVision, Perkin Elmer).

  11. 11.

    Statistical analysis software (GraphPad PRISM, GraphPad Software Inc., San Diego, CA and MATLAB, The Mathworks Inc., Natick, MA).

2.2 Additional Material for BRET , BRET-BiFC , and SRET Analysis

  1. 1.

    Flat-bottom white 96-well microplates (Corning 3912, Corning, NY).

  2. 2.

    Coelenterazine H (p.j.k GmbH, Germany).

  3. 3.

    DeepBlueC (Biotium Inc., Hayward, CA).

2.3 Expression Vectors for Fluorescent and Luminescent Proteins

  1. 1.

    pECFP-N1, pEYFP-N1, pEdsRed-N1 (Clontech, Mountain View, CA) (see Notes 2 and 3 ).

  2. 2.

    For BRET assays use C-terminal part of the chemokine receptor fused to the Renilla luciferase gene using the pRLuc-N1 plasmid (PerkinElmer).

  3. 3.

    For BiFC assays, use N- and C-terminal nonfluorescent fragments of a specific fluorescent protein (see Table 1) cloned in pcDNA3.1 (Addgene, Cambridge, MA) (see Note 4 ).

    Table 1 Combinations of fluorescent protein fragments for BiFC
    Full size table

3 Methods

3.1 Fluorescence Resonance Energy Transfer

The FRET mechanism involves a donor fluorophore in an electron-excited state that can transfer its excitation energy to a nearby acceptor chromophore in a non-radioactive fashion through long-range dipole–dipole interactions. This phenomenon is not mediated by photon emission and in many applications, energy transfer results in quenching of donor fluorescence and subsequent reduction of fluorescence lifetime; this process is obviously also accompanied by an increase in acceptor fluorescence emission. The range over which energy transfer can take place is limited to approximately 10 nm (100 Å), a sufficient distance to consider that molecular interactions take place. FRET efficiency is thus extremely sensitive to the distance between donor and acceptor.

Although there are many fluorescent proteins pairs used as FRET donor/acceptors (Table 2), they respond to several common requirements.

Table 2 Properties of fluorescent protein pairs for FRET
Full size table
  1. 1.

    The donor emission spectrum must overlap the acceptor excitation spectrum. If there is donor–acceptor interaction, donor excitation with the maximum absorbance wavelength must increase the intensity of the maximum emission fluorescence of the acceptor.

  2. 2.

    Energy transfer efficiency depends on the relative orientations of the donor emission dipole and the acceptor absorption dipole.

  3. 3.

    Donor and acceptor molecules must be located within 1–10 nm of one another. Energy transfer efficiency between donor and acceptor molecules decreases as the sixth power of the distance between the two fluorescent molecules as described by the Förster equation:

    $$ {E}_{\mathrm{t}}={R_0}^6/{R}^6+{R_0}^6 $$

    where E t represents energy transfer efficiency, R is the distance between the fluorescent proteins , and R 0 is the donor–acceptor distance at which 50 % of the excitation energy is transferred whereas the remaining 50 % decays as non-radioactive or radioactive energy. This efficiency depends on the fluorescent partners used (Table 1).

  4. 4.

    The fluorescence lifetime of the donor molecule (quantum yield of the donor) must be long enough to allow energy transfer to the acceptor.

Several methods are used to determine and quantify FRET efficiency (see Note 5 ). To study receptor interactions in dynamic processes the most adequate FRET technique is based in the sensitized emission of the acceptor.

3.1.1 Sensitized Emission of the Acceptor

FRET efficiency can be determined in individual (fixed) cells using confocal microscopy or in living cell populations using a multilabel fluorescent plate reader with specific detectors. The method requires specific excitation of the donor fluorescent protein and use of specific light filters to determine donor and acceptor fluorescent emissions. Specific detectors must be used to collect the maximum peak of fluorescent emission for both donor and acceptor molecules.

As in other FRET methods, the donor emission spectrum must overlap the acceptor excitation spectrum. For many FRET partners, the requirement for spectral proximity can lead to strong fluorophore cross talk and cross-excitation processes that might alter quantitative analysis. To correct for detection of ratiometric signals and sensitized emission, a linear unmixing method is used that is based on the assumption that total detected signal (S) for each channel (λ) can be expressed as a linear combination of the contributing fluorophores (FluoX), according to the formula:

$$ S\left(\lambda \right)={A}_1\times \mathrm{Fluo}1\left(\lambda \right)+{A}_2\times \mathrm{Fluo}2\left(\lambda \right) $$

where A n is the contribution of a specific fluorophore, and fluorophore concentration in the signal observed therefore determines their respective contribution to the total signal [50, 51]. This signal is redistributed into the specific fluorescence channels and can be analyzed quantitatively. To calculate fluorophore contributions, linear unmixing requires knowledge of reference values for samples that contain the fluorophores of interest separately. For successful separation of overlapping signals, the number of detection channels used and the number of fluorophores in the sample must be the same.

For highly efficient FRET partners such as CFP and YFP, which have very close overlapping signals, the FRET pair must be excited by a single excitation wavelength and FRET interactions resolved using linear unmixing (Fig. 1, example spectrum CFP-YFP, see FRET measurement). This is the best way to analyze the contribution of individual signals to the mixture.

Fig. 1
figure 1

Excitation and emission spectra of a CFP–YFP FRET pair. The scheme shows absorbance (exc) and emission (em) spectra of CFP (cyan fluorescent protein; donor; D) and YFP (yellow fluorescent protein; acceptor, A). Spectral overlap between CFP emission and YFP excitation (shaded region) is a prerequisite for FRET

Full size image

3.1.2 FRET Saturation Curves by Sensitized Emission

  1. 1.

    Plate HEK293T cells (3.5 × 105 cells/well in 2 ml complete DMEM, using 6-well plates) and culture (24 h, 37 °C, 5 % CO2).

  2. 2.

    To obtain FRET saturation curves, prepare two transfection mixtures: (a) mix 25 μl of 150 mM NaCl/μg of DNA and vortex (10 s); (b) at a 4:1 ratio, mix 150 mM NaCl and PEI (5.47 mM in nitrogen residues). Add mixture b to mixture a at a 1:1 ratio, vortex (10 s) and incubate (15–30 min, room temperature). Add the mixture (a + b) to the cells in 1 ml serum-free DMEM, incubate (4 h, 37 °C), and replace medium with complete DMEM (see Note 6 ).

  3. 3.

    At 48 h post-transfection, wash cells twice in HBSS supplemented with 0.1 % glucose and resuspend in this same buffer. Determine total protein concentration for each cell sample using a Bradford assay kit. Pipette Cell suspensions (20 μg protein/100 μl) into black 96-well microplates and read in a multilabel fluorescent plate reader equipped with a high-energy xenon flash lamp, a short-wavelength filter (8 nm bandwidth, 405 nm) and a long-wavelength emission filter (10 nm bandwidth, 486 nm for CFP channel and 10 nm bandwidth, 530 for YFP channel) (see Note 7 ). To determine the spectral signature, HEK293T cells are transiently transfected with the chemokine receptor coupled to CFP or to YFP separately. The contributions of CFP and YFP alone are measured in each detection channel and normalized to the sum of the signal obtained in both channels. Analyze the spectral signatures of CFP or YFP fused to chemokine receptors and check variability (it must not vary significantly (p > 0.05)) from the signatures determined for each fluorescent protein alone. For FRET quantitation and receptor-YFP expression quantitation in FRET saturation curves, the spectral signature is taken into account for linear unmixing in order to separate the two emission spectra. To determine the fluorescence emitted by each fluorophore in FRET experiments, apply the following formulas:

$$ \mathrm{C}\mathrm{F}\mathrm{P}=S/\left(1+1/R\right)\kern0.28em \mathrm{and}\kern0.28em \mathrm{Y}\mathrm{F}\mathrm{P}=S/1+R $$

where

$$ S=\mathrm{ChCFP}+\mathrm{ChYFP}, $$
$$ R=\left({\mathrm{YFP}}_{530}Q\hbox{--} {\mathrm{YFP}}_{486}\right)/\left({\mathrm{CFP}}_{486}-{\mathrm{CFP}}_{530}Q\right),\kern0.28em \mathrm{and} $$
$$ Q=\mathrm{ChCFP}/\mathrm{ChYFP} $$

ChCFP and ChYFP represent the signal detected for CFP in the 486 nm and 530 nm detection channels (Ch), respectively; CFP486, CFP530, YFP530, and YFP486 are the normalized contributions of CFP and YFP to 486 and 530 nm channels, as determined from their spectral signatures.

  1. 4.

    Sensitized emission FREt allows measurement of the energy transferred relative to the acceptor/donor ratio to generate FRET saturation curves. The curves show FRET efficiency as a function of the acceptor–donor ratio and are characterized by two important parameters; FRETmax, is the (asymptotic) maximum of the curve and FRET50 corresponds to the acceptor–donor ratio that yields half FRETmax efficiency (only when energy transfer reaches saturation and the curve is hyperbolic). FRETmax is associated with the number of receptor complexes formed and/or changes in complex conformation, and FRET50 allows estimation of the apparent affinity between the two partners that form the complex [52, 53]. These two parameters are deduced from data analysis using a nonlinear regression equation applied to a single binding site model (based on Michaelis–Menten model, Fig. 2) (see Note 8 ).

    Fig. 2
    figure 2

    FRET saturation curve generated by the sensitized emission method. A representative FRET saturation curve for a chemokine receptor pair. The curve reaches FRETmax (maximum FRET efficiency detected) and is hyperbolic. FRET50 indicates the FRET efficiency value that corresponds to the acceptor–donor ratio (YFP–CFP ratio) that yields half FRETmax

    Full size image

3.2 BRET

The BRET technique is based on non-radiative energy transfer between a bioluminescent donor (usually Renilla luciferase (Luc)) and a fluorophore acceptor. Like FRET , this method requires the generation of fusion proteins for donors between the receptor and luciferase (Luc) and for acceptors fusion between receptor and YFP, CFP, or GFP2. When Luc is present, it oxidizes its substrate (coelenterazine or DeepBlueC) and triggers photons release. Close proximity (10 nm) of an appropriate fluorophore ensures its excitation; that is, electron movement to a higher energetic state and thus, photon emission of longer wavelengths (see Note 9 ).

3.2.1 BRET Titration Assays

  1. 1.

    Plate HEK293T cells (3.5 × 105 cells/well in 2 ml complete DMEM, using 6-well plates) and culture (24 h, 37 °C, 5 % CO2).

  2. 2.

    Cotransfect the cells using the PEI protocol (see Subheading 3.1.1 above) with a constant amount of donor (Luc-fused chemokine receptor) and increasing amounts of acceptor protein (YFP-, CFP-, or GFP2-fused chemokine receptor, depending on the luciferase substrate; see above) (see Note 10 ).

  3. 3.

    At 48 h post-transfection, wash cells twice in HBSS supplemented with 0.1 % glucose and resuspend cells in this buffer. Using a Bradford assay, determine total protein concentration in each well.

  4. 4.

    Quantify the fluorescent protein (20 μg) using a multilabel plate reader equipped with a high-energy xenon flash lamp (for CFP or GFP2 acceptor, 8 nm bandwidth excitation filter at 405 nm; for YFP, 10 nm bandwidth excitation filter at 510 nm). Receptor fluorescence expression is determined as fluorescence of the sample minus the fluorescence of cells that express donor alone. For BRET2 and BRET1 measurements, the equivalent of 20 μg cell suspension is distributed in 96-well microplates, followed by 5 μM DeepBlueC (for BRET2) or coelenterazine H (for BRET1). For BRET2 experiments, signals are obtained immediately (30 s) after DeepBlueC addition using the multilabel plate reader, which allows integration of signals detected in the short- (8 nm bandwidth, 405 nm) and the long-wavelength filters (10 nm bandwidth, 486 nm). For BRET1, readings are collected 1 min after coelenterazine H addition, as the plate reader allows integration of signals detected in the short- (10 nm bandwidth, 510 nm) and long-wavelength filters (10 nm bandwidth, 530 nm). Receptor-Luc luminescence signals should be acquired 10 min after coelenterazine H (5 μM) addition. BRET efficiency (BRETeff) is defined as:

$$ {\mathrm{BRET}}_{\mathrm{eff}}=\left[\left(\mathrm{long}\kern0.28em \mathrm{wavelength}\kern0.28em \mathrm{emission}\right)/\left(\mathrm{short}\kern0.28em \mathrm{wavelength}\kern0.28em \mathrm{emission}\right)\right]-{C}_{\mathrm{f}} $$

where C f is [(long wavelength emission)/(short wavelength emission)] for the Luc construct expressed alone in the same experiment (control).

  1. 5.

    Statistical analysis (see Note 5 and Subheading 3.1.2 above).

3.3 Bimolecular Fluorescence Complementation (BiFC)

Bimolecular fluorescence complementation (BiFC ) technology enables simple, direct visualization of protein–protein interactions in living cells [54–56]. The BiFC assay uses receptors fused to two fluorescent protein fragments that are nonfluorescent individually; fluorescence is recovered only when both fragments interact, that is, when the accompanying receptor form complexes. The approach can be used for analysis of interactions between many types of proteins and does not require information about the structures of the interaction partners (see Note 11 ).

Like FRET and BRET , BiFC requires fusion of the fluorescent protein fragments to chemokine receptors , as well as testing that neither protein expression nor receptor function is modified by the fluorescent fragments (see Notes 12 and 13 ).

3.3.1 BiFC Measurement

  1. 1.

    Plate HEK293T cells 24 h before transfection (3.5 × 105 cells/2 ml, in 6-well plates) and cultured (37 °C, 5 % CO2).

  2. 2.

    Cotransfect using PEI or JetPei methods (according to manufacturer’s protocol), with chemokine receptors fused to the nonfluorescent fragments at a 1:1 ratio (i.e., 0.7 μg CXCR4 -nYFP + 0.7 μg CXCR4-cYFP) and culture (48 h, 37 °C, 5 % CO2).

  3. 3.

    Wash cells with HBSS supplemented with 0.1 % glucose and resuspend in the same buffer. Determine the protein concentration in the cell pool using a Bradford assay kit (Bio-Rad).

  4. 4.

    Pipette cells into black 96-well plates, at 20 μg of protein in 100 μl (0.2 μg/μl) per well.

  5. 5.

    Quantify the fluorescent protein using a multilabel plate reader equipped with a high-energy xenon flash lamp (for YFP, 10 nm bandwidth excitation filter at 510 nm and 10 nm bandwidth emission filter at 530 nm).

3.4 BRET -BiFC

BRET -BiFC allows identification of heterotrimeric complexes in living cells. This method combines BRET and BiFC techniques sequentially. It is thus necessary take into account all controls and considerations described in Subheadings 2 and 3. In BRET-BiFC, the donor is a Luc-fused chemokine receptor or protein and the acceptor is formed by interaction of the two nonfluorescent fragments fused to the chemokine receptors of interest.

  1. 1.

    Plate HEK293T cells (3.5 × 105 cells/well in 2 ml complete DMEM, use 6-well plates) and culture (24 h, 37 °C, 5 % CO2).

  2. 2.

    Using the PEI method (see Subheading 3.1.2, step 2 above), cotransfect cells with a constant amount of donor (Luc-fused chemokine receptor) and increasing amounts of a mixture of acceptor proteins at a 1:1 ratio (i.e., nYFP- and cYFP-fused chemokine receptors ) (see Note 14 ).

  3. 3.

    At 48 h post-transfection, wash cells twice in HBSS supplemented with 0.1 % glucose and resuspend the pellet in the same buffer. Determine protein concentration for the cell pool using a Bradford assay kit.

  4. 4.

    Quantify the fluorescent protein in samples containing 20 μg protein, using a multilabel plate reader equipped with a high-energy xenon flash lamp (for YFP, 10 nm bandwidth excitation filter at 510 nm). Expression of the fluorescent protein-fused receptor is determined as fluorescence of the sample minus fluorescence of cells expressing donor-Luc alone. Distribute the cell suspension (20 μg/well) in 96-well microplates (Corning 3912, flat-bottom white plates). Add coelenterazine H (for BRET1, 5 μM/well; PJK GmbH) and read 1 min later in the plate reader, which allows integration of signals detected in the short- (10 nm bandwidth, 510 nm) and long-wavelength filters (10 nm bandwidth, 530 nm). Receptor-Luc luminescence signals are acquired 10 min after coelenterazine H addition. BRET -BiFC efficiency is calculated using the same formula given for BRET efficiency in Subheading 3.2.1 above.

    In each BRET-BiFC titration curve, the relative amount of acceptor is given by the ratio between acceptor (YFP) formed by the two nonfluorescent proteins and the luciferase activity of the donor (Luc).

  5. 5.

    Statistical analysis as in Subheading 3.1.2 above (see Note 5 ).

3.5 Sequential BRET-FRET (SRET)

The SRET method allows detection of complexes of three proteins in living cells. There are several possible combinations depending on the donor and acceptor fusion proteins (Fig. 3).

Fig. 3
figure 3

Scheme of BRET -BiFC , SRET1, SRET2, and SRET3 techniques. (a) BRET-BiFC for the chemokine receptors CKR1, CKR2, and CKR3 fused to Rluc, cYFP, and nYFP, respectively. Following interaction between chemokine receptors CKR2 and CKR3 (colored halos ), YFP is reconstituted and is susceptible to excitation by Rluc activated by its substrate, coelenterazine H (yellow dots ). EYFP emission is detected at 530 nm. (b) SRET1 for the chemokine receptors CKR1, CKR2, and CKR3 fused to Rluc, CFP, and YFP, respectively. Due to activation by its substrate DeepBlueC (blue dots ), Rluc excites CFP (BRET), which in turn excites YFP (FRET ) that is detected at 530 nm. (c) In SRET2, initial BRET between Luc and YFP triggered by coelenterazine H (yellow dots ) excites dsRed, the last acceptor fluorescence protein, which then emits light at 590 nm. (d) SRET3 is a FRET1 and FRET2 sequence. Due to the interaction between CFP, YFP, and dsRed fused to a chemokine receptor (CKR1, CKR2, CKR3, respectively), CFP excitation (at 405 nm) triggers dsRed light emission at 590 nm

Full size image

SRET1 combines BRET1 and FRET . The first donor protein is a Luc-fused chemokine receptor. The first acceptor protein is YFP, which is then used as a second donor to excite the next acceptor, a chemokine receptor fused to dsRed, which is the signal detected. This method uses coelenterazine H as a Luc substrate.

SRET2 combines BRET2 and FRET . The Luc-fused chemokine receptor is the first donor (using DeepBlueC as a specific Luc substrate). T CFP- or GFP2-fused receptor is then excited and its emission energy (BRET2) excites the YFP-fused receptor, which is then the last SRET acceptor.

SRET3 combines two sequential FRET methods. The donor protein is the CFP- or GFP2-fused chemokine, the first acceptor/second donor is the YFP-fused receptor which excites the last acceptor, which is the dsRed-fused receptor.

Here we describe in detail the SRET2 method, which can easily be adapted for SRET1 or SRET3 using the detection filters for each specific fluorophore and the appropriate Luc substrate in the case of SRET1.

  1. 1.

    Plate HEK293T cells (3.5 × 105 cells/well in 2 ml complete DMEM, use 6-well plates) and culture (24 h, 37 °C, 5 % CO2).

  2. 2.

    Using the PEI method (Subheading 3.1.2, step 2 above), cotransfect the cells with a constant amount of Luc-fused receptor and increasing amounts of receptor fused to acceptors (CFP or GFP2 and YFP at 1:1 ratio); culture (48 h, 37 °C, 5 % CO2).

  3. 3.

    Wash cells with HBSS supplemented with 0.1 % glucose and resuspend the pellet in the same buffer. Total protein concentration is determined for the cell pool using a Bradford assay kit (Bio-Rad).

  4. 4.

    Use aliquots of transfected cells (20 μg protein/100 μl in a 96-well microplate) to perform three experiments in parallel.

  5. 5.

    In the first experiment, quantify FRET efficiency for the FRET pairs used (CFP or GFP2 and YFP) and the amount of acceptor protein–YFP at each ratio. Distribute cells (20 μg) in black 96-well microplates and read in a multilabel plate reader equipped with a high-energy xenon flash lamp, using an excitation filter at 405 ± 8 nm and 10 nm bandwidth emission filters corresponding to 530 nm (channel 1) and 486 nm (channel 2). As in FRET, separate the relative contribution of fluorophores to each detection channel for linear unmixing. Then measure the contribution of CFP or GFP2 and YFP proteins alone to the two detection channels and normalize to the sum of the total signal obtained in the two detection channels. Quantify the total amount of receptor-YFP at each ratio in the same equipment using a 510 ± 10 nm excitation filter at and a 530 ± 10 nm emission filter.

  6. 6.

    In the second experiment, quantify the receptor-Luc expression by determining its luminescence. Pipette the cells (20 μg) into 96-well microplates (white-bottomed white plates), add the substrate (5 μM coelenterazine H); after 10 min, detect luminescence in a multilabel plate reader.

  7. 7.

    For SRET1 evaluation in the third experiment, distribute cells (20 μg) in 96-well microplates (white-bottomed white plates) and add 5 μM DeepBlueC as Luc substrate; after 30 s, collect the SRET signal using a multilabel plate reader with detection filters for short (486 nm) and long wavelengths (530 nm).

  8. 8.

    By analogy with BRET , net SRET is defined as:

$$ \begin{array}{l}\mathrm{netSRET}=(\left(\mathrm{long}\kern0.28em \mathrm{wavelength}\kern0.28em \mathrm{emission}\right)/\\ {}\kern7em \left(\mathrm{short}\kern0.28em \mathrm{wavelength}\kern0.28em \mathrm{emission}\right))-{C}_{\mathrm{f}}\end{array} $$

where

$$ {C}_{\mathrm{f}}=(\left(\mathrm{long}\kern0.28em \mathrm{wavelength}\kern0.28em \mathrm{emission}\right)/\left(\mathrm{short}\kern0.28em \mathrm{wavelength}\kern0.28em \mathrm{emission}\right) $$

for cells expressing the receptor-Luc, receptor-CFP or -GFP2 separately. SRET is only detected if all receptors interact and the corresponding pairs (Luc/CFP or Luc/GFP2 and CFP/YFP or GFP2/YFP or Luc/YFP) are located at <10 nm distance.

  1. 9.

    For statistical analysis, see Subheading 3.1.2 above and Note 5 .

4 Notes

  1. 1.

    Any other cell line with high transfection efficiency for the receptors of interest can be employed. In this case the transfection method should be optimized to these cell lines. Take into account that most cell lines can express endogenously chemokine receptors (most frequently CXCR4 ) that may alter chemokine oligomers.

  2. 2.

    Several plasmids are commercially available to fuse fluorescent proteins to a receptor (Clontech).

  3. 3.

    For chemokine receptors and to avoid interference with ligand binding it is best to fuse fluorescent proteins to the C-terminal end of the receptor.

  4. 4.

    All the techniques described in this chapter require fusion of chemokine receptors to fluorescent (FRET ) or luminescent proteins (BRET ), or N- or C-terminal fragments of fluorescent proteins (BiFC ). Insertion of a fluorescent probe in the C-terminal region of the receptor involves eliminating the receptor stop codon, whereas N-terminal insertion requires elimination of the fluorescent protein stop codon. Transfected cells should therefore be analyzed for receptor expression and function.

    The constructs are obtained using standard molecular biology techniques in commercially available vectors that bear the luminescent or fluorescent proteins (Clontech). In the case of BiFC, the fluorescent protein used must be cut into N- and C-terminal fragments, which must then be included in separate expression vectors (i.e., pcDNA3.1; see Table 2). Insertion of these fragments in-frame with the chemokine receptor C-terminal region requires elimination of the receptor stop codon. Transfected cells with these chimeric proteins should be tested for receptor expression and function.

    Several experiments are needed to test whether the chimeric proteins maintain expression and function. Using flow cytometry , stain the chemokine receptors expressed at the cell surface with specific antibodies. Chemokine receptor function is usually tested using calcium mobilization assays or migration in transwells. The chemokine receptors fused to fluorescent/luminescent proteins must behave similarly to the wild type receptor.

  5. 5.

    Methods used to determine and quantify FRET efficiency include:

    1. (a)

      Donor quenching or acceptor photobleaching is a method based on quenching donor fluorescence. As some donor photons excite the acceptor, a decrease is detected in donor emission energy. After acceptor photobleaching, increased donor light emission is detected. Due to cell damage caused by the extended laser exposure needed to photobleach the acceptor, this method cannot be used in live cells.

    2. (b)

      For sensitized emission of the acceptor, the acceptor signal is quantitated after donor excitation. This method is useful to determine FRET in dynamic processes in the cell, such as the consequences of ligand stimulation, the effects triggered by other proteins coexpressed on cell surface, and so on.

    3. (c)

      Fluorescence lifetime imaging microscopy (FLIM) is a method based on measurement of a constant parameter of each fluorescent protein in each experimental condition, termed lifetime. Lifetime is the duration of the excited state of a fluorophore before returning to its ground state. This technique allows spatial resolution of biochemical processes. The fluorescence lifetime of a donor molecule decreases in FRET conditions independently of fluorescent protein concentrations and of excitation intensities.

  6. 6.

    The cells must be transiently cotransfected with a constant amount of donor (chemokine receptor fused to the donor fluorophore, CFP) and increasing amounts of acceptor protein (chemokine receptor fused to the acceptor fluorophore, YFP) using the polyethylenimine method (PEI).

  7. 7.

    Gain settings must be identical for all experiments to maintain a constant relative contribution of each fluorophore to the detection channels for spectral imaging and linear unmixing.

  8. 8.

    Statistical analysis is needed to reduce experimental variability (at least five replications of the saturation curves are usually generated), and will finally determine the homogeneous curves. We currently use three distinct statistical methods (bootstrap, F test, and Akaike information criterion (AIC)), which usually lead to similar conclusions [57]. For example, we can determine the conformational changes in dimers promoted by a given ligand (“the treatment”). Curves for untreated and ligand-treated groups are naturally paired when dimerization is evaluated in the same group of cells before and after “treatment.” To determine which model best fits the data for pairs of saturation curves, we use the AIC method corrected for small size samples (n ≤ 10), in which the null hypothesis is one curve for all data sets (before and after “treatment”) and the alternative hypothesis is the existence of different curves for each data set. If the majority of the AICc difference (Δ) is positive, the preferred model is a distinct curve for all data sets, whereas if Δ is negative, the preferred model is a single curve for all data sets [28]. When the number of individual determinations (n) in each curve is >10 observations, a t-test can be used to compare the components of each pair [58–60]. When the p value is <0.05, we can conclude that the changes observed after a specific treatment are statistically significant.

  9. 9.

    Distinct luciferase substrates can be chosen depending on the acceptor fluorophore. Requirements for energy transfer in BRET are the same that those for FRET (see Subheading 3.1 above). The most used luciferase substrates are coelenterazine, used in BRET1 assays, whose maximal emission is at 515 nm and which is used in combination with YFP as acceptor. DeepBlueC, used in BRET2 assays, is an analogue of the natural luciferase substrate with maximal emission wavelength at 410 nm; it is used in combination with CFP or GFP2, which emit at 485 nm and 515 nm respectively, yielding a spectral separation of >100 nm.

  10. 10.

    The cells must also be transfected separately with donor and acceptor. Use untransfected HEK293T cells as a background control for each experiment.

  11. 11.

    One advantage of this technique is that the complex formed has strong intrinsic fluorescence, which allows direct visualization of the protein interaction without exogenous agents. This avoids disturbance of the cells, but also is a clear disadvantage, as the user must know that there is a delay between the time needed by the proteins to interact and the time required for the reconstituted complex to become fluorescent [54].

  12. 12.

    Flexible linker sequences are recommended, to allow maximal mobility of the nonfluorescent fragments after complex formation.

  13. 13.

    To determine BiFC specificity, many controls are included in each experiment, such as nonfluorescent fragment-fused receptors with point mutations in the interaction interface that impede reconstitution of the complete protein [55, 61]. Control proteins with the same cell expression pattern as the chemokine receptors should be included.

  14. 14.

    Controls include cells transfected with the receptor-Luc construct alone and with receptors fused to the acceptor fragments. In addition, untransfected HEK293T cells are needed to determine the background.