Bimolecular fluorescence complementation: lighting up seven transmembrane domain receptor signalling networks

Abstract

There is increasing complexity in the organization of seven transmembrane domain (7TM) receptor signalling pathways, and in the ability of their ligands to modulate and direct this signalling. Underlying these events is a network of protein interactions between the 7TM receptors themselves and associated effectors, such as G proteins and beta-arrestins. Bimolecular fluorescence complementation, or BiFC, is a technique capable of detecting these protein-protein events essential for 7TM receptor function. Fluorescent proteins, such as those from Aequorea victoria, are split into two non-fluorescent halves, which then tag the proteins under study. On association, these fragments refold and regenerate a mature fluorescent protein, producing a BiFC signal indicative of complex formation. Here, we review the experimental criteria for successful application of BiFC, considered in the context of 7TM receptor signalling events such as receptor dimerization, G protein and beta-arrestin signalling. The advantages and limitations of BiFC imaging are compared with alternative resonance energy transfer techniques. We show that the essential simplicity of the fluorescent BiFC measurement allows high-content and advanced imaging applications, and that it can probe more complex multi-protein interactions alone or in combination with resonance energy transfer. These capabilities suggest that BiFC techniques will become ever more useful in the analysis of ligand and 7TM receptor pharmacology at the molecular level of protein-protein interactions.

Figure 1

Principle of bimolecular fluorescence complementation (BiFC) and design of Aequorea fluorescent protein fragments. (A) Protein partners can associate and dissociate reversibly until the split fluorescent protein tags refold to form a stable β-barrel structure. The subsequent maturation of the chromophore generates a fluorescent BiFC complex. Folding and maturation times are for yellow fluorescent protein (YFP) (1–154) and YFP (155–238) tags at 25°C in vitro (Hu et al., 2002). (B) The structure of Aequorea fluorescent proteins consists of 11 β-strands, with the amino acids that develop into the chromophore shown in a helix between strands 3 and 4. Venus mutations on an enhanced YFP background are highlighted in red (Nagai et al., 2002). Fluorescent fragments known to display complementation (YFP unless otherwise stated) are shown at positions 1: 154–155 (e.g. Hu et al., 2002), 2: 158–159 (Mervine et al., 2006), 3: 172–173 (Hu and Kerppola, 2003), 4: superfolder green fluorescent protein (GFP) 222–223 (Cabantous and Waldo, 2006) and 5: 144–145 (Nagai et al., 2001). Sequence references refer to original native GFP.

Figure 2

Bimolecular fluorescence complementation (BiFC) imaging assay for seven transmembrane domain receptor–β-arrestin2 interaction. Dual stably transfected HEK293 cells were established, which expressed β-arrestin2-Yn and C terminal Yc-tagged β2-adrenoceptors. In (A), the same field of living cells was imaged before addition of agonist, and then again following 60 min isoprenaline treatment (10 µM). Yellow fluorescent protein (YFP) BiFC acquisition was performed on a Zeiss LSM 510 (Jena, Germany) confocal microscope (63× plan apochromat/1.4 NA oil objective) using 514 nM excitation and a 530 nM LP emission filter, and the same laser power and gain settings. Scale bar: 40 µm. In (B), automated images were acquired from cells seeded into a 96-well plate (IX ultra platereader, Molecular Devices, Sunnyvale, CA, USA) of both the nuclear stain (H33342) and YFP BiFC fluorescence (scale bar: 80 µm). As for the high-resolution confocal images in (A), 10 µM isoprenaline (37°C, 60 min) increased BiFC in perinuclear vesicular compartments. Granularity analysis classified puncta by size (white dots 1–3 µm diameter, red ‘vesicles’ 3 µm + diameter), and normalized to nuclear-based cell count (green). Quantification as vesicle average intensity per cell, expressed as a percentage of the 10 µM isoprenaline response, enabled concentration response curves to different adrenoceptor ligands to be constructed (C, pooled responses, n= 4 or more). Full agonists (isoprenaline and formoterol), partial agonists (salbutamol and salmeterol) and antagonist/inverse agonist ligands (e.g. ICI118551) can be separated in this manner.

Figure 3

Mutational analysis of Y receptor–β-arrestin2 interaction detected by bimolecular fluorescence complementation (BiFC). Dual stably transfected HEK293 cells established Yc-tagged neuropeptide Y (NPY) receptors at equivalent expression levels on a common β-arrestin2-Yn cell line. NPY-stimulated Y receptor/β-arrestin2 BiFC was measured by automated image acquisition and granularity analysis as described in Figure 2. The left hand panel (A) shows the effect of mutating a C tail-phosphorylated sequence in the Y1 receptor (Ser352–Thr361), previously identified as important for arrestin interaction (Holliday et al., 2005; Ouedraogo et al., 2008). Successive pairs of alanine mutations (2A, 4A and 6A illustrated in inset) reduce the maximal receptor association with β-arrestin2 induced by NPY (pEC₅₀ range 8.55–8.73; n= 4+). On the right (B), the greater affinity of the Y2 receptor for β-arrestin2 after H155P mutation (intracellular loop 2; Marion et al., 2006) is revealed by an increase in NPY potency (Y2 pEC₅₀: 7.18 ± 0.06, Y2H155P pEC₅₀: 8.28 ± 0.18; n= 6) as well as elevated basal and maximum responses. Experiment details are provided in Kilpatrick et al. (2008).

Figure 4

Three ways of examining seven transmembrane domain (7TM) receptor oligomers by bimolecular fluorescence complementation (BiFC). Multicolour BiFC allows competitive formation of alternative 7TM receptor dimers to be assessed by the development of cyan fluorescent protein (CFP) or yellow fluorescent protein (YFP)-like BiFC fluorescence (A). Alternatively, BiFC YFP formed by a 7TM receptor dimer can act as an energy transfer acceptor. Typically, Renilla luciferase (LUC) provides the donor for bioluminescence resonance energy transfer (BRET), carried as a full-length protein by a third 7TM receptor monomer (B) or reconstituted by complementation in a second 7TM receptor dimer (tetrameric interaction, C).

Figure 5

Analysis of histamine H₁ receptor membrane diffusion by fluorescence correlation spectroscopy (FCS). Both H₁-yellow fluorescent protein (YFP) (A) and H₁-Yn + H₁-Yc complemented fluorescence (BiFC) (B) are mainly localized on the plasma membrane of transiently transfected CHO-K1 cells. Fusion to YFP allows detection of all receptors, whether monomers or oligomers (C). In contrast, a minimum of a dimer is required for detection of BiFC, and such dimers will only be visible when complementary fragments are present in the complex (D). For FCS, fluctuations in fluorescence intensity are recorded with time (E, F) from the cell membrane above the nucleus (marked ‘+’ in A and B). Subsequent analysis produces autocorrelation curves (G, H) from which the diffusion coefficients of H₁Yn + H₁Yc BiFC dimers (7.0 ± 0.4 × 10⁻⁹ cm⁻²·s⁻¹; n= 77 cells) and of H₁-YFP (5.3 ± 0.2 × 10⁻⁹ cm⁻²·s⁻¹; n= 59 cells) can be obtained (see Rose et al., 2008 for more experimental details). The significantly more rapid diffusion of the homo-oligomeric H₁ BiFC receptor population suggests a subset of protein interactions or microdomain localization that differ from other receptor pools labelled only by H₁-YFP.