Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Jul 1.
Published in final edited form as: Methods. 2008 Jun 27;45(3):207–213. doi: 10.1016/j.ymeth.2008.06.008

Multicolor BiFC analysis of competition among G protein β and γ subunit interactions

Thomas R Hynes 1, Evan Yost 1, Stacy Mervine 1, Catherine H Berlot 1,*
PMCID: PMC2688734  NIHMSID: NIHMS66236  PMID: 18586104

Abstract

We have applied multicolor BiFC to study the association preferences of G protein β and γ subunits in living cells. Cells co-express multiple isoforms of β and γ subunits, most of which can form complexes. Although many βγ complexes exhibit similar properties when assayed in reconstituted systems, knockout experiments in vivo suggest that individual isoforms have unique functions. BiFC makes it possible to correlate βγ complex formation with functionality in intact cells by comparing the amounts of fluorescent βγ complexes with their abilities to modulate effector proteins. The relative predominance of specific βγ complexes in vivo is not known. To address this issue, multicolor BiFC can determine the association preferences of β and γ subunits by simultaneously visualizing the two fluorescent complexes formed when β or γ subunits fused to amino terminal fragments of yellow fluorescent protein (YFP-N) and cyan fluorescent protein (CFP-N) compete to interact with limiting amounts of a common γ or β subunit, respectively, fused to a carboxyl terminal fragment of CFP (CFP-C). Multicolor BiFC also makes it possible to determine the roles of interacting proteins in the subcellular targeting of complexes, study the formation of protein complexes that are unstable under isolation conditions, determine the roles of co-expressed proteins in regulating the association preferences of interacting proteins, and visualize dynamic events affecting multiple protein complexes. These approaches can be applied to studying the assembly and functions of a wide variety of protein complexes in the context of a living cell.

Keywords: multicolor bimolecular fluorescence complementation, heterotrimeric G protein, G protein βγ complex, fluorescence microscopy, spectrofluorometer, CFP, YFP, subcellular targeting, live cell imaging

1. INTRODUCTION

Most combinations of the 5 G protein β subunits and 12 γ subunits that have been identified in mammals form dimers that exhibit similar abilities in vitro to modulate the activities of effectors such as adenylyl cyclase [1], phospholipase C [2], and GIRK channels [3]. However, there is evidence suggesting that specific αβγ combinations are required for particular receptor-G protein signaling pathways in vivo [4]. In most cases it is not known which αβγ heterotrimers mediate particular GPCR signaling pathways, nor has it been possible to determine which βγ combinations predominate in a particular cell type. The relative amounts of particular βγ complexes in a cell will be determined both by the expression levels of each of the β and γ subunits as well as by their accessibilities to and relative affinities for each other. Multicolor BiFC makes it possible to quantify the association preferences of β and γ subunits for each other in intact cells.

Multicolor BiFC consists of the simultaneous visualization of the two fluorescent complexes formed when proteins fused to amino terminal fragments of YFP and CFP (YFP-N and CFP-N, respectively) interact with a common binding partner fused to a carboxyl terminal fragment of CFP (CFP-C). Complexes containing YFP-N and CFP-C fusion proteins are yellow, while those containing CFP-N and CFP-C fusion proteins are cyan, because the amino terminal fragment of the fluorescent protein, which contains the chromophore, determines the spectral properties of the complex [5]. In the experiments described here the fluorescent proteins are split at residue 158 such that the amino terminal fragment consists of residues 1–158 of the fluorescent protein and the carboxyl terminal fragment consists of residues 159–238 of the fluorescent protein. For multicolor BiFC, it is important to use CFP rather than YFP to produce the carboxyl terminal fragment, because BiFC is not obtained when the closely related fluorescent protein fragments CFP(1–154) and YFP(155–238) are used [5]. For competition analysis, we use Cerulean, a modified version of ECFP that is 2.5-fold brighter than ECFP [6], to produce Cer-N fusion proteins, because Cer-N fusions compete more effectively than do CFP-N fusions with YFP-N fusions. The reason for this difference is not clear, but could indicate greater stability of Cer-N-fusion/CFP-C-fusion complexes compared to CFP-N-fusion/CFP-C-fusion complexes.

The interaction preferences of β and γ subunits identified using BiFC probably reflect association preferences, because most available evidence indicates that BiFC is irreversible [7, 8]. For βγ complexes, this is not a concern, because they generally associate irreversibly. The only reported exceptions are β5γ2 [9] and β4γ11 [10], which are unstable in vitro. Efforts are underway in several laboratories to develop a reversible form of BiFC by means of splitting the fluorescent protein in different places and mutagenesis of the fluorescent protein fragments, which would enable dynamic analysis of reversible protein-protein interactions. Nevertheless, the relative association preferences of reversibly associating proteins can also be estimated using this approach.

2. DESCRIPTION OF METHOD

2.1 Producing and assessing functionality of BiFC βγ complexes

The amino termini of the G protein β and γ subunits are associated as a coiled coil in βγ complexes [11] (Fig. 1). Fusion of fluorescent protein fragments to the amino termini of each subunit (Fig. 1) results in BiFC [12–14]. As will be discussed at greater length in section 2.4, the abilities of different γ subunits to compete for a common β subunit are compared using fusions of γ subunits (red in Fig. 1A, green in Fig. 1B) to the carboxyl termini of YFP-N (yellow in Fig. 1A) and Cer-N (cyan in Fig. 1B). The β subunit to be competed for (magenta in Fig 1, A and B) is fused to the carboxyl terminus of CFP-C (dark blue in Fig. 1, A and B). Conversely, the abilities of different β subunits to compete for a common γ subunit are compared using an alternative BiFC configuration; the β subunits (red in Fig. 1C, green in Fig. 1D) are fused to the carboxyl termini of YFP-N (yellow in Fig. 1C) and Cer-N (cyan in Fig. 1D). The γ subunit to be competed for (magenta in Fig 1, C and D) is fused to the carboxyl terminus of CFP-C (dark blue in Fig. 1, C and D).

FIG. 1.

FIG. 1

Open in a new tab

Models of fluorescent βγ complexes produced with multicolor BiFC. The split fluorescent protein at the top of each model is joined by linkers (orange) to the βγ dimer at the bottom. The CFP-C fragment (dark blue) is combined with either the YFP-N fragment (yellow) to produce yellow fluorescence or the Cer-N fragment (cyan) to produce cyan fluorescence. (A and B) YFP-N-γ and Cer-N-γ compete for CFP-C-β. β is magenta and γ is red (A) or green (B). (C and D) YFP-N-β and Cer-N-β compete for CFP-C-γ. γ is magenta and β is red (C) or green (D). The structures of the fluorescent protein fragments are adapted from the structure of GFP [33]. The structures of β and γ are from the structure of an αt/αi1 chimera complexed with βtγt [34].

It is important to demonstrate that both of these configurations of BiFC βγ complexes are functional in assay(s) relevant to the planned studies. We have demonstrated that YFP-N-β1YFP-C-γ complexes potentiate αs-mediated activation of adenylyl cyclase in COS-7 cells [12] and that CFP-C-β5Cer-N-γ complexes activate phospholipase C-β2 expressed in HEK-293 cells [14]. Interestingly, co-expressed GFP- β1 and untagged γ2, and co-expressed CFP-γ2 and untagged β1, but not co-expressed GFP-β1 and CFP-γ2, can potentiate activation of adenylyl cyclase [12]. This result suggests that although the βγ structure can accommodate the equivalent of one fluorescent protein (produced by BiFC) at its amino terminus, two fluorescent proteins distort the structure enough to interfere with potentiation of αs-mediated adenylyl cyclase activation. However, FRET has been reported between YFP-β1 and CFP-γ2 and YFP-β1/CFP-γ2, like untagged β1γ2, has been reported to increase basal N-type Ca2+ channel facilitation and attenuate noradrenaline-mediated inhibition of Ca2+ channels, as well as activate G-protein-gated inwardly rectifying K+ (GIRK) channels [15]. Therefore, the functional consequences of fusing fluorescent proteins and their fragments to β and γ subunits may vary and should be tested for the effector of interest.

To apply multicolor BiFC to other proteins, it may be necessary to fuse the fluorescent protein fragments to regions other than the amino terminus. For instance, we have obtained BiFC with the β2-adrenergic receptor by fusing fluorescent protein fragments to the carboxyl terminus (data not shown). Moreover, sometimes, as with G protein α subunits, both the amino and carboxyl termini are important for function. We have identified internal regions of αq [16] and αs [17] in which insertion of full-length fluorescent proteins bordered by linkers consisting of Gly-Ser-Ser-Ser-Ser-Gly maintains full retention of function. Inserting fluorescent protein fragments at these sites also permits BiFC (data not shown).

2.2. Correlating complex formation with functionality

BiFC makes it possible to correlate βγ function directly with the amount of βγ complex formation in intact cells. For instance, we have compared the abilities of different CFP-C-β5Cer-N-γ complexes to activate phospholipase C-β2 [14]. Phospholipase C-β2 activity in cells expressing CFP-C-β5 and different Cer-N-γ subunits was compared with the relative amounts of fluorescence produced by the CFP-C-β5Cer-N-γ complexes. Four of seven CFP-C-β5Cer-N-γ complexes (those containing Cer-N-γ2, Cer-N-γ5, Cer-N-γ7, or Cer-N-γ12) activated co-expressed phospholipase C-β2, while the other 3 complexes (those containing Cer-N-γ1, Cer-Nγ10, or Cer-N-γ11) produced no activation above that seen in cells transfected with empty vector (Fig. 2A). No activity was obtained when CFP-C-β5 was expressed without a Cer-N-γ subunit or when any of the Cer-N-γ subunits was expressed without CFP-C-β5 (Fig. 2A). The three CFP-C- β5Cer-N-γ complexes that did not stimulate phospholipase C-β2 exhibited only minimal fluorescence, indicating that lack of activity was due to the absence of complex formation (Fig. 2B). For the four CFP-C-β5Cer-N-γ complexes that activated phospholipase C-β2, the ratios of CFP-C-β5Cer-N-γ-stimulated activity (Fig. 2A) to amount of complex formation (Fig. 2B) were similar (Fig. 2C), indicating that the different efficacies of the CFP-C-β5Cer-N-γ combinations were due primarily to different amounts of complex formation. These results are in agreement with in vitro studies that found similar abilities of different βγ complexes to modulate effectors [1–3]. In contrast, different βγ complexes have been shown to exhibit a wider range of abilities to interact with G protein-coupled receptors [18–21].

FIG. 2.

FIG. 2

Open in a new tab

Comparisons of the abilities of different β5 and γ combinations to activate phospholipase C-β2 and to form complexes. (A) Activation of phospholipase C-β2 in cells expressing CFP-C-β5 and Cer-N-γ subunits. HEK-293 cells were transfected with 3 µg of phospholipase C-β2 plasmid and where indicated, 2.4 µg of CFP-C-β5 plasmid, and 0.3 µg of Cer-N-γ plasmids. The total amount of plasmid in each transfection was maintained at 5.7 µg by making up the difference with pcDNAI/Amp. Inositol phosphate (IP) formation in cells labeled with [3H]inositol was measured as described [35]. Values represent the means ± S.E. from 6 experiments performed in triplicate. (B) Fluorescence intensities of CFP-C-β5Cer-N-γ complexes expressed under the same transfection conditions as in (A). Values represent the means ± S.E. from 3 experiments performed in duplicate. (C) Normalization of phospholipase C-β2 activation by coexpressed CFP-C-β5 and Cer-N-γ subunits to the fluorescence intensities of the CFP-C-β5Cer-N-γ complexes. IP levels obtained in cells transfected with pcDNAI/Amp (background) were subtracted from the levels in cells expressing CFP-C-β5 and either Cer-N-γ2, Cer-N-γ5, Cer-N-γ7, or Cer-N-γ12. These background-subtracted activities were divided by the corresponding CFP-C-β5Cer-N-γ intensities. Reproduced with permission from [14].

2.3. Quantifying the amount of complex formation under non-competing conditions

Single color BiFC is used to quantify the formation of βγ complexes in the absence of competition. For instance, to compare the abilities of different γ subunits to form complexes with the same β subunit, each Cer-N-γ subunit is co-expressed in limiting amounts with an excess of the CFP-C-β subunit. The amounts of plasmid that will produce a linear dose-response of fluorescence versus amount of transfected plasmid need to be determined empirically. For HEK-293 cells (1.6 × 106 per 60-mm dish) transiently transfected with 6 µl of Lipofectamine 2000 (Invitrogen) and analyzed 48 hours after transfection, 2.4 µg of CFP-C-β1 plasmid co-transfected with 0.0033 to 0.03 µg of Cer-N-γ plasmid yielded a linear relationship [13]. In the case of β5γ complexes, the fluorescence intensities of the complexes spanned a larger range and for the less intense ones, 0.1 to 0.3 µg of Cer-N-γ plasmid was necessary to produce detectable signals and yielded linear relationships between amounts of fluorescence and amounts of transfected plasmid (Fig. 3A) [14]. The relative amount of complex formation is calculated as the slope of fluorescence versus amount of transfected plasmid (Fig. 3A). It is important to keep the total amount of plasmid constant using empty vector to control for promoter competition. We have less experience with Cer-N-β subunits and CFP-C-γ subunits, but in a study using Cer-N-β1, Cer-N-β5, and CFP-C-γ2, we found that 2.4 µg of CFP-C-γ2 co-expressed with 0.00033 to 0.03 µg of either Cer-N-β1 or Cer-N-β5 plasmid yielded a linear relationship of fluorescence to Cer-N-β dose [14].

FIG. 3.

FIG. 3

Open in a new tab

β5 interacts preferentially with γ2 rather than γ1, γ5, γ7, γ10, γ11, or γ12. (A) Fluorescence intensities of CFP-C-β5Cer-N-γ complexes when CFP-Cβ5 is not limiting. HEK-293 cells were transfected with 2.4 µg of plasmid expressing CFP-C-β5 and the indicated µg of Cer-N-γ plasmids. The total amount of plasmid in each transfection was maintained at 2.7 µg by making up the difference with pcDNAI/Amp. Values represent the means ± S.E. from 3 experiments performed in duplicate. CFP-C-β5Cer-N-γ intensities/µg Cer-N-γ plasmid, determined from linear fits to the data were as follows (means ± S.E, × 10-4): β5γ1, 3.93 ± 0.27; β5γ2, 408 ± 44.8; β5γ5, 21.8 ± 2.36; β5γ7, 106 ± 4.34; β5γ10, 6.72 ± 0.94; β5γ11, 11.3 ± 1.40; β5γ12, 22.0 ± 0.93. (B) Expression levels of Cer-N-γ subunits when co-expressed with an excess of CFP-C-β5. HEK-293 cells were transfected as in (A). Values represent the means ± S.E from 3 experiments. Cer-N-γ expression/µg plasmid, determined from linear fits to the data were as follows (means ± S.E, × 10−6): γ1, 1.82 ± 0.29; γ2, 16.87 ± 1.50; γ5, 4.77 ± 1.00; γ7, 10.51 ± 2.13; γ10, 2.78 ± 0.05; γ11, 8.48 ± 0.28; γ12, 5.54 ± 0.59. (C) Normalization of CFP-C-β5Cer-N-γ intensities to Cer-N-γ expression levels. The intensity/µg Cer-N-γ plasmid of each CFP-C-β5Cer-N-γ complex was divided by the corresponding Cer-N-γ expression/µg plasmid. (D) Competition between Cer-N-γ subunits and YFP-N-γ2 for limiting amounts of CFP-C-β5. The intensity of CFP-C-β5YFP-N-γ2 was measured in the presence of each Cer-N-γ subunit or empty vector. HEK-293 cells were transfected with 0.6 µg each of plasmids expressing CFP-C-β5 and YFP-N-γ2, and the indicated µg of each Cer-N-γ plasmid. The total amount of plasmid in each transfection was maintained at 3.63 µg using pcDNAI/Amp. Values represent the means ± S.E. from 3 experiments performed in duplicate. (E) CFP-C-β5YFP-N-γ2 intensity is expressed as a function of the relative amounts of co-expressed Cer-N-γ. Expression levels were determined in HEK-293 cells transfected with 0.6 µg each of plasmids expressing CFP-C-β5 and pcDNAI/Amp, and 0.03, 0.09, 0.27, or 2.43 µg of each Cer-N-γ plasmid. The total amount of plasmid in each transfection was maintained at 3.63 µg using pcDNAI/Amp. The expression levels of the Cer-N-γ subunits varied linearly and the data were fit by linear regressions. The plasmid amounts used in (D) were multiplied by Cer-N-γ expression/µg plasmid to yield the normalized amount of each Cer-N-γ subunit. CC indicates CFP-C and YN indicates YFP-N. Reproduced with permission from [14].

2.3.1. Measurement of BiFC in cell populations

Fluorescence can be quantified in individual cells using fluorescence microscopy [5, 13, 22], but we have found that it is easier and less time-consuming to obtain quantitative results by measuring the fluorescence in cell populations using a spectrofluorometer. Two days after transient transfection of HEK-293 cells (1.6 × 106 per 60-mm dish) the cells are washed once in 4 mL of HBSS+ CaCl2 media (20 mM Hepes, pH 7.2, 118 mM NaCl, 4.6 mM KCl, 10 mM D-glucose, 1 mM CaCl2). 2 mL of HBSS+EDTA media (20 mM Hepes, pH 7.2, 118 mM NaCl, 4.6 mM KCl, 10 mM D-glucose, 0.5 mM EDTA) are then added to the dish, cells are scraped off the dishes with a rubber policeman, passed through a pipet several times to break up clumps, and suspended in a 1 cm square glass cuvette with a magnetic stir bar.

When measuring cells in suspension it is important to minimize the signal due to light scattering, and to subtract out the signal due to autofluorescence. The monochrometers used in most spectrofluorometers transmit enough light at wavelengths outside the band-pass region to generate a large background signal in highly scattering samples. This scattering signal can be eliminated through the use of band-pass and long-pass filters, which are generally much better than the monochrometer at blocking light outside the band-pass region. Autofluorescence cannot be eliminated, but varies as a function of cell density, which can be determined with a light scattering measurement that is made at a wavelength outside that of the fluorescent signal emission.

We collect our data on a PC1 photon-counting spectrofluorometer (ISS, Champaign, IL) configured with motorized filter wheels on both the excitation path between the excitation monochrometer and the sample, and on the emission path between the sample and the emission monochrometer as described [13]. The slits on the excitation and emission monochrometers are set to 16 nm. To eliminate the light scattering signal from the cell suspensions during measurements of CFP and YFP intensity, band-pass filters are used in the excitation filterwheel in combination with long-pass filters in the emission filterwheel. For the CFP channel measurement the excitation monochrometer is set to 430 nm with a 430/25 band-pass filter, and the emission monochrometer is set to 480 nm with a 455 long-pass filter. For YFP channel measurements the excitation is set to 492 nm with a 492/18 band-pass filter, and emission is set to 530 nm with a 515 long-pass filter. A dilute solution of glass milk, which scatters significantly, but is not fluorescent, should produce little if any signal in the CFP and YFP channels. For the light scattering channel measurement, the excitation monochrometer is set to 650 nm with a 1.3 OD neutral density filter in combination with a 590 long-pass filter, and the emission monochrometer is also set to 650 nm with no filter. All filters are from Chroma. Control of the monochrometers, motorized filterwheels and data acquisition is done using the Vinci software program (ISS) and allows the measurements of CFP, YFP and light scattering to be made in quick succession without repositioning the sample.

To subtract the signal due to autofluorescence, measurements using the CFP, YFP, and light scattering channels are made each day on a dilution series of cells transfected with vector alone. The linear relationships between the autofluorescence signals in the CFP and YFP channels versus the light scattering signal are calculated. CFP, YFP and light scattering measurements are then made for samples containing fluorescent proteins, and the scatter value is used to calculate the amount of autofluorescence to subtract from the CFP and YFP measurements.

2.3.2. Controlling for the expression levels of Cer-N-fusion proteins

Immunoblots are used to quantify expression under the same transfection conditions as the fluorescence measurements are made. We have found that polyclonal antibodies to residues 3 to 17 of GFP (Anti-GFP, N-terminal; Sigma-Aldrich, St. Louis, MO) and to full-length GFP (Rockland Immunochemicals, Gilbertsville, PA) work well for detecting and comparing expression levels of Cer-N fusion proteins. Fig. 4A shows an immunoblot of Cer-N-γ subunits using the polyclonal antibody to residues 3 to 17 of GFP and Fig. 4B shows an immunoblot of Cer-N-β subunits using the polyclonal antibody to full-length GFP. The antigen-antibody complexes were detected using the ECL Western blotting protocol (GE Healthcare, Chalfont St. Giles, Buckinghamshire, UK) (Fig. 4A) or with SuperSignal West Pico Chemiluminescent Substrate (Pierce Biotechnology, Rockford, IL) (Fig. 4B) and were imaged using a Lumi-Imager (Roche Applied Science, Indianapolis, IN) (Fig. 4A) or with a FluorChem SP Imaging System (Alpha Innotech, San Leandro, CA) (Fig. 4B). Bands are quantified using IPLab software (BD Biosciences) (Fig. 3B and Table 1). We have found that under the conditions in which BiFC signals vary linearly as a function of the amount of transfected plasmid (Fig. 3A), the expression levels as determined by immunoblots also vary linearly (Fig. 3B) [14]. The relative amount of expression is calculated as the slope of band intensity versus amount of transfected plasmid (Fig. 3B and Table 1). The relative amounts of complex formation are calculated as the ratio of BiFC intensity to expression level (Fig. 3C) [14].

FIG. 4.

FIG. 4

Open in a new tab

Expression levels of Cer-N-subunits are controlled for using immunoblots. (A) Expression levels of Cer-N-γ subunits co-expressed with CFP-C-β5. HEK-293 cells were transfected as in Figs. 3A and 3B. µg indicate the amount of transfected Cer-N-γ plasmid or vector. The vector control lane was taken from a gel run, transferred, and immunoblotted in parallel with the one containing CFP-C-β5 and Cer-N-γ subunits. (B) Expression levels of Cer-N-β1 and Cer-N-β5 co-expressed with CFP-C-γ2. HEK-293 cells were transfected with 2.4 µg of CFP-C-γ2 plasmid and the indicated amounts (in µg) of plasmid expressing Cer-N-β1 or Cer-N-β5. The total amount of plasmid in each transfection was maintained at 2.43 µg by making up the difference with pcDNAI/Amp. For each lane, 15 µg of total cell lysates were resolved on NuPage Bis-Tris 4–12% gels (Invitrogen), transferred to nitrocellulose, and processed as described in the text.

TABLE 1.

Competition of Cer-N-γ subunits with YN-γ2 for dimerization with CC-β5 in live HEK-293 cells Values represent the mean ± S.E. from 3 experiments.

γsubunit IC50 (µg CerN-γ)a Cer-N-γ expression/µgb(X 10−5) Normalized IC50c(X 10−5)
γ1 3.10 ± 1.08 1.38 ± 0.23 4.27 ± 1.65
γ2 0.15 ± 0.02 4.44 ± 0.77 0.66 ± 0.14
γ5 2.15 ± 0.59 4.39 ± 0.17 9.43 ± 2.64
γ7 0.40 ± 0.06 6.78 ± 0.58 2.68 ± 0.49
γ10 2.66 ± 0.48 3.27 ± 0.07 8.71 ± 1.57
γ11 1.94 ± 0.54 7.88 ± 0.58 15.27 ± 4.41
γ12 1.60 ± 0.31 5.62 ± 0.46 9.01 ± 1.88
Open in a new tab
a

Defined as µg of Cer-N-γ plasmid that produced a 50% decrease in the intensity of CC-β5YN-γ2, calculated from the data in Fig. 3D.

b

Calculated from linear fits of Cer-N-γ expression/µg plasmid. Expression levels were determined as described in the legend to Fig. 3E.

c

Defined as IC50 X Cer-N-γ expression/µg plasmid.

Reproduced with permission from [14].

2.4. Using multicolor BiFC to compare the abilities of subunits to compete for a common binding partner that is expressed in limiting amounts

To compare the abilities of different γ subunits to compete for a common β subunit, one of the γ subunits (red in Fig. 1A) is fused to the carboxyl terminus of YFP-N (yellow in Fig. 1A) and each of the γ subunits (green in Fig. 1B) to be compared is fused to the carboxyl terminus of Cer-N (cyan in Fig. 1B). The β subunit to be competed for (magenta in Fig 1, A and B) is fused to the carboxyl terminus of CFP-C (dark blue in Fig. 1, A and B). We quantify competition as the loss of yellow fluorescence of the CFPC-β/YFP-N-γ complex when Cer-N-γ subunits are co-expressed (Fig. 3, D and E). Conversely, to compare the abilities of different β subunits to compete for a common γ subunit, one of the β subunits (red in Fig. 1C) is fused to the carboxyl terminus of YFP-N (yellow in Fig. 1C) and each of the β subunits to be compared (green in Fig. 1D) is fused to the carboxyl terminus of Cer-N (cyan in Fig. 1D). The γ subunit that is competed for (magenta in Fig. 1, C and D) is fused to the carboxyl terminus of CFP-C (dark blue in Fig. 1, C and D). Competition is quantified as the loss of yellow fluorescence of the CFP-C-γ/YFP-N-β complex when Cer-N-β subunits are co-expressed.

The relative preferences of the limiting subunit for the competing subunits are expressed as the IC50 for inhibition of yellow fluorescence produced by the CFP-C-subunit/ YFP-N-subunit complex. For example, for inhibition of association of YFP-N-γ2 with CFP-C-β5 by Cer-N-γ subunits, the IC50 is defined as µg of Cer-N-γ subunit plasmid that produces a 50% decrease in the intensity of CFP-C-β5YFP-N-γ2. To determine IC50 values, the data are fit to:

Y=(100)/(1+X/a)b) Eq. (1)

where X is µg of transfected Cer-N-γ plasmid, Y is the % of maximal fluorescence produced by CFP-C-β5YFP-N-γ2, a is the half-maximal inhibitory concentration (IC50) of the Cer-N-γ subunit, and b is the slope factor. These IC50 values are normalized by multiplying by the relative expression levels of the Cer-N-γ subunits (Fig. 3E and Table I).

2.4.1. Optimization of amounts of transfected Cer-N-subunit, YFP-N-subunit, and CFP-C-subunit plasmids

In competition experiments, the subunit that is competed for is expressed in a limiting amount compared to the competing subunits. The optimum amounts of plasmids to be transfected are determined empirically by measuring competition between fusions of the same subunit to Cer-N and YFP-N. Loss of YFP fluorescence from CFP-C-β/YFP-N-γ or YFP-N-β/CFP-C-γ complexes is measured in the presence of a range of amounts of Cer-N-γ or Cer-N-β subunits, respectively. The initial fluorescence of the CFP-C-β/YFP-N-γ or YFP-N-β/CFP-C-γ complex needs to be high enough to provide a workable dynamic range of fluorescent intensities in the presence of the competing Cer-N-γ or Cer-N-β subunits. The β subunit fusions generally express at lower levels than do the γ subunit fusions, so the optimal conditions for competition of γ subunits for β subunits may be different from those for competition of β subunits for γ subunits. For example, to compare γ subunits competing for β1 or β5, we transfected HEK-293 cells (1.6 × 106 per 60-mm dish) with 0.6 µg each of plasmids expressing CFP-C-β1 or CFPC-β5 and YFP-N-γ2 and 0, 0.01, 0.03, 0.09, 0.27, 0.81, or 2.43 µg of each Cer-N-γ subunit using 6 µl of Lipofectamine 2000 [13, 14](Fig. 3D). The total amount of plasmid was maintained at 3.63 µg by making up the difference with empty vector (pcDNAI/Amp). To compare β1 and β5 competing for γ2, we transfected HEK-293 cells with 0.3 µg of CFP-C-γ2 plasmid, 0.6 µg of YFP-N-β1 plasmid, and 0.033, 0.1, 0.3, 0.9, 2.7, or 8.1 µg of plasmids encoding either Cer-N-β1 or Cer-N-β5. The total amount of plasmid was maintained at 9 µg using pcDNAI/Amp (see Fig. 8 of [14]). The expression levels of the competitors should be determined under the conditions of competition so that IC50 values can be normalized to the amounts of the expressed competitors (Fig. 3E and Table 1).

2.5. Determining the roles of interacting proteins in the subcellular targeting of complexes

Multicolor BiFC makes it possible to simultaneously image the subcellular localization patterns of two complexes that share a common subunit. In this way, the roles of alternative binding partners in regulating the localization patterns of complexes can be determined. These studies have demonstrated that the binding partners of both β and γ subunits play roles in subcellular targeting. For example, dual imaging of β5γ2 and β1γ2 in HEK-293 cells demonstrated a role for β subunits in targeting γ subunits (Fig. 5, A–C). CFP-C-γ2Cer-N-β5 localizes to discrete cytoplasmic structures (Fig. 5, A and C) that exhibit co-localization with markers for the ER and Golgi apparatus [14], while CFP-C-γ2YFP-N-β1 associates with the plasma membrane (Fig. 5, B and C), confirming previous studies using single color BiFC [12]. Conversely, dual imaging of β5γ2 and β5 complexed with RGS7, a member of the R7 family of regulators of G protein signaling (RGS) proteins, revealed that β5 localization can be determined by its binding partner. In contrast to the localization of CFP-C-β5YFP-N-γ2 on discrete cytoplasmic membranes (Fig. 5, D, F, G, and I), CFP-C-β5Cer-N-RGS7 produces a diffuse cytoplasmic signal (Fig. 5, E and F), and CFP-C-β5Cer-N-RGS7-t, containing a truncated form of RGS7 lacking the DEP domain, localizes preferentially in the nucleus (Fig. 5, H and I). The roles of co-expressed proteins in regulating the targeting of these complexes can also be determined. For instance, we observed that Cer-C-β5Cer-N-γ2 and Cer-C-β5Cer-N-RGS7 were targeted preferentially to the plasma membrane by wild-type αo and a constitutively activated αo mutant, respectively [14].

FIG. 5.

FIG. 5

Open in a new tab

Simultaneous imaging of two complexes sharing a common subunit demonstrates the roles of alternative binding partners in subcellular localization. CFP (A) and YFP (B) images from the same cell expressing Cer-N-β5CFP-Cγ2 and YFP-N-β1CFP-C-γ2. In the merge image (C), Cer-N-β5CFP-C-γ2 is green and YFP-N-β1CFP-C-γ2 is red. HEK-293 cells were transfected with the following amounts (in µg) of plasmids: CFP-C-γ2, 0.4; Cer-N-β5, 0.1; YFP-N-β1, 0.2; and mCherry-Mem, 0.0025 (a plasma membrane marker [14]). YFP (D) and CFP (E) images from the same cell expressing CFP-C-β5YFP-N-γ2 and CFP-C-β5Cer-N-RGS7. In the merge image (F), CFP-C-β5YFP-N-γ2 is green and CFP-C-β5Cer-N-RGS7 is red. HEK-293 cells were transfected with the following amounts (in µg) of plasmids: CFP-C-β5, 0.3; YFP-N-γ2 and Cer-N-RGS7, 0.15; and mCherry-Mem, 0.0025. (G) YFP and (H) CFP images from the same cell expressing CFP-C-β5YFP-N-γ2 and CFP-C-β5CFP-N-RGS7-t, which contains a truncated form of RGS7 lacking the DEP domain. HEK-293 cells were transfected with the following amounts (in µg) of plasmids: CFP-C-β5, 0.3; YFP-N-γ2 and CFP-N-RGS7-t, 0.15; and mCherry-Mem, 0.0025. In the merge image (I), CFP-C-β5YFP-N-γ2 is green and CFP-C-β5Cer-N-RGS7-t is red. Scale bar = 10 µm. Figures (D), (E), (G), and (H) are reproduced with permission from [14].

2.6. Studying the formation of protein complexes that are unstable in vitro

The G protein β5 subunit differs from other β subunits in having divergent sequence and subcellular localization patterns. Although β5γ2 can modulate effector proteins [3, 23–27] (Fig. 2), β5 associates with R7 family regulators of G protein signaling (RGS) proteins when purified from tissues [28]. The fact that β5γ2 dimers are unstable under non-denaturing buffer conditions [9, 29] may explain why they have yet to be isolated. Using multicolor BiFC, we found that Cer-N-γ2 competed three times more effectively than did Cer-N-RGS7 for CFP-C-β5 [14] (Fig. 6A), illustrating the advantage of using BiFC in intact cells to identify protein interaction partners.

FIG. 6.

FIG. 6

Open in a new tab

Multicolor BiFC demonstrates that multiple factors influence formation of β5 complexes. (A) In HEK-293 cells co-expressing CFP-C-β5, YFP-N-γ2, and either Cer-N-γ2 or Cer-N-RGS7, Cer-N-γ2 competes 3-fold more effectively for CFP-C-β5 than does Cer-N-RGS7. (B) Co-expression of R7BP influences competition between γ2 and RGS7 for β5. In the presence of R7BP, Cer-N-RGS7 and Cer-N-γ2 compete with similar efficacies for CFP-C-β5. In addition, other β subunits may affect the relative amounts of β5 complexes. Cer-N-β1 competes 4-fold more effectively for CFP-C-γ2 than does Cer-N-β5. Data for these results are shown in [14].

2.7. Determining the roles of co-expressed proteins in regulating the association preferences of interacting proteins

Using multicolor BiFC, the potential roles of co-expressed proteins in determining the association preferences of alternative binding partners can be investigated. R7 family binding protein (R7BP), which, like β5 and RGS7, is highly expressed in the nervous system, can target β5RGS7 complexes to the plasma membrane [30]. We found that in the presence of co-expressed R7BP, the preference of CFP-C-β5 for Cer-N-γ2 over Cer-N-RGS7 was lost. Cer-N-γ2 and Cer-N-RGS7 now competed similarly with YFP-N-γ2 for CFP-C-β5[14] (Fig. 6B). The formation of β5 complexes will also be influenced by the association preferences of its alternative binding partners. For instance, γ2 can also form complexes with other β subunits. Using multicolor BiFC, we found that Cer-N-β1 competed four times more effectively than did Cer-N-β5 for CFP-C-γ2 [14] (Fig. 6B). This preferential interaction of γ2 with β1 rather than β5 could decrease formation of β5γ2 complexes by sequestering some of the available γ2 away from β5.

2.8. Visualizing dynamic events affecting multiple protein complexes

Previously, we visualized the β2-adrenergic receptor (β2AR)-stimulated trafficking patterns of αs and β1γ7 in HEK-293 cells and found that both Gs subunits internalized and co-localized on vesicles upon stimulation of the β2AR [17]. αs and β1γ7 internalized more rapidly than did the β2AR and, in contrast to the β2AR, their internalization was insensitive to a dominant negative mutant of dynamin 1 [17]. Multicolor BiFC makes it possible to monitor the β2AR-stimulated internalization of pairs of βγ complexes simultaneously. In this way, differences in the kinetics and localization patterns of βγ complexes can be visualized and quantified. For example, we have found that β1γ11 internalizes more rapidly than does β1γ7 and appears to localize on different subcellular structures (Fig. 7). Further experiments will reveal if these differences result from the different prenyl groups on the γ subunits (geranylgeranyl on γ7, farnesyl on γ11) [31].

FIG. 7.

FIG. 7

Open in a new tab

The stimulus-induced localization pattern of β1γ11 differs from that of β1γ7. YFP (top), CFP (middle), and merge (bottom) images from the same cell expressing CFP-C-β1YFP-N-γ7 and CFP-C-β1Cer-N-γ11 acquired at the indicated times before and after stimulation with 10 µM isoproterenol. In the merge image, colocalization of CFP-C-β1YFP-N-γ7 (red) and CFP-C-β1Cer-N-γ11 (green) is indicated in yellow. HEK-293 cells were transfected with the following amounts (in µg) of plasmids: αs, 0.3; CFP-C-β1, 0.3; YFP-N-γ7, 0.1125; Cer-N-γ11, 0.0375; β2AR, 0.1; and mCherry-Mem, 0.0025. More YFP-N-γ7 than Cer-N-γ11 plasmid was used to normalize the fluorescence intensities of the two βγ complexes. The same result was obtained in cells expressing CFP-C-β1YFPN-γ11 and CFP-C-β1Cer-N-γ7.

3. CONCLUDING REMARKS

The formation of βγ complexes in vivo is likely to reflect the affinities of β and γ subunits for each other as well as their accessibilities to each other and their associations with other co-expressed proteins. By estimating the relative association preferences of potential binding partners in the context of a living cell, multicolor BiFC takes all of these factors into account. This approach quantifies the interaction preferences of β and γ subunits expressed heterologously in intact cells. In the particular cell of interest, it is important to determine which β and γ subunits are expressed endogenously at the protein level and to quantify these expression levels. Combining this information with the relative association preferences determined with multicolor BiFC should provide a reasonable estimate of the predominance of specific endogenous βγ complexes. In addition, a recently described proximity ligation in situ assay that enables visualization of endogenous protein complexes in fixed samples [32] may provide a complementary means of corroborating the results of multicolor BiFC.

ACKNOWLEDGMENTS

This work was supported by National Institutes of Health grant GM50369.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  • 1.Iniguez-Lluhi JA, Simon MI, Robishaw JD, Gilman AG. J Biol Chem. 1992;267:23409–23417. [PubMed] [Google Scholar]
  • 2.Ueda N, Iniguez-Lluhi JA, Lee E, Smrcka AV, Robishaw JD, Gilman AG. J Biol Chem. 1994;269:4388–4395. [PubMed] [Google Scholar]
  • 3.Mirshahi T, Robillard L, Hebert TE, Logothetis DE. J Biol Chem. 2002;277:7348–7355. doi: 10.1074/jbc.M109999200. [DOI] [PubMed] [Google Scholar]
  • 4.Robishaw JD, Berlot CH. Curr Opin Cell Biol. 2004;16:206–209. doi: 10.1016/j.ceb.2004.02.007. [DOI] [PubMed] [Google Scholar]
  • 5.Hu CD, Kerppola TK. Nat Biotechnol. 2003;21:539–545. doi: 10.1038/nbt816. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Rizzo MA, Springer GH, Granada B, Piston DW. Nat Biotechnol. 2004;22:445–449. doi: 10.1038/nbt945. [DOI] [PubMed] [Google Scholar]
  • 7.Kerppola TK. Nat Rev Mol Cell Biol. 2006;7:449–456. doi: 10.1038/nrm1929. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Magliery TJ, Wilson CG, Pan W, Mishler D, Ghosh I, Hamilton AD, Regan L. J Am Chem Soc. 2005;127:146–157. doi: 10.1021/ja046699g. [DOI] [PubMed] [Google Scholar]
  • 9.Jones MB, Garrison JC. Anal Biochem. 1999;268:126–133. doi: 10.1006/abio.1998.3064. [DOI] [PubMed] [Google Scholar]
  • 10.McIntire WE, MacCleery G, Murphree LJ, Kerchner KR, Linden J, Garrison JC. Biochemistry. 2006;45:11616–11631. doi: 10.1021/bi0604882. [DOI] [PubMed] [Google Scholar]
  • 11.Sondek J, Bohm A, Lambright DG, Hamm HE, Sigler PB. Nature. 1996;379:369–374. doi: 10.1038/379369a0. [DOI] [PubMed] [Google Scholar]
  • 12.Hynes TR, Tang V, Mervine SM, Sabo JL, Yost EA, Devreotes PN, Berlot CH. J Biol Chem. 2004;279:30279–30286. doi: 10.1074/jbc.M401432200. [DOI] [PubMed] [Google Scholar]
  • 13.Mervine SM, Yost EA, Sabo JL, Hynes TR, Berlot CH. Mol Pharmacol. 2006;70:194–205. doi: 10.1124/mol.106.022616. [DOI] [PubMed] [Google Scholar]
  • 14.Yost EA, Mervine SM, Sabo JL, Hynes TR, Berlot CH. Mol Pharmacol. 2007;72:812–825. doi: 10.1124/mol.107.038075. [DOI] [PubMed] [Google Scholar]
  • 15.Ruiz-Velasco V, Ikeda SR. J Physiol. 2001;537:679–692. doi: 10.1111/j.1469-7793.2001.00679.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hughes TE, Zhang H, Logothetis DE, Berlot CH. J Biol Chem. 2001;276:4227–4235. doi: 10.1074/jbc.M007608200. [DOI] [PubMed] [Google Scholar]
  • 17.Hynes TR, Mervine SM, Yost EA, Sabo JL, Berlot CH. J Biol Chem. 2004;279:44101–44112. doi: 10.1074/jbc.M405151200. [DOI] [PubMed] [Google Scholar]
  • 18.Richardson M, Robishaw JD. J Biol Chem. 1999;274:13525–13533. doi: 10.1074/jbc.274.19.13525. [DOI] [PubMed] [Google Scholar]
  • 19.McIntire WE, MacCleery G, Garrison JC. J Biol Chem. 2001;276:15801–15809. doi: 10.1074/jbc.M011233200. [DOI] [PubMed] [Google Scholar]
  • 20.Lim WK, Myung CS, Garrison JC, Neubig RR. Biochemistry. 2001;40:10532–10541. doi: 10.1021/bi010950c. [DOI] [PubMed] [Google Scholar]
  • 21.Gibson SK, Gilman AG. Proc Natl Acad Sci U S A. 2006;103:212–217. doi: 10.1073/pnas.0509763102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Grinberg AV, Hu CD, Kerppola TK. Mol Cell Biol. 2004;24:4294–4308. doi: 10.1128/MCB.24.10.4294-4308.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Watson AJ, Katz A, Simon MI. J Biol Chem. 1994;269:22150–22156. [PubMed] [Google Scholar]
  • 24.Zhang S, Coso OA, Lee C, Gutkind JS, Simonds WF. J Biol Chem. 1996;271:33575–33579. doi: 10.1074/jbc.271.52.33575. [DOI] [PubMed] [Google Scholar]
  • 25.Lindorfer MA, Myung CS, Savino Y, Yasuda H, Khazan R, Garrison JC. J Biol Chem. 1998;273:34429–34436. doi: 10.1074/jbc.273.51.34429. [DOI] [PubMed] [Google Scholar]
  • 26.Lei Q, Jones MB, Talley EM, Garrison JC, Bayliss DA. Mol Cells. 2003;15:1–9. [PubMed] [Google Scholar]
  • 27.Zhou JY, Siderovski DP, Miller RJ. J Neurosci. 2000;20:7143–7148. doi: 10.1523/JNEUROSCI.20-19-07143.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Witherow DS, Wang Q, Levay k, Cabrera L, Chen j, Willars GB, Slepak VZ. J Biol Chem. 2000;275:24872–24880. doi: 10.1074/jbc.M001535200. [DOI] [PubMed] [Google Scholar]
  • 29.Jones MB, Siderovski DP, Hooks SB. Mol Interv. 2004;4:200–214. doi: 10.1124/mi.4.4.4. [DOI] [PubMed] [Google Scholar]
  • 30.Drenan RM, Doupnik CA, Jayaraman M, Buchwalter AL, Kaltenbronn KM, Huettner JE, Linder ME, Blumer KJ. J Biol Chem. 2006;281:28222–28231. doi: 10.1074/jbc.M604428200. [DOI] [PubMed] [Google Scholar]
  • 31.Wedegaertner PB, Wilson PT, Bourne HR. J Biol Chem. 1995;270:503–506. doi: 10.1074/jbc.270.2.503. [DOI] [PubMed] [Google Scholar]
  • 32.Soderberg O, Gullberg M, Jarvius M, Ridderstrale K, Leuchowius KJ, Jarvius J, Wester K, Hydbring P, Bahram F, Larsson LG, Landegren U. Nat Methods. 2006;3:995–1000. doi: 10.1038/nmeth947. [DOI] [PubMed] [Google Scholar]
  • 33.Ormo M, Cubbitt AB, Kallio K, Gross LA, Tsien RY, Remington SJ. Science. 1996;273:1392–1395. doi: 10.1126/science.273.5280.1392. [DOI] [PubMed] [Google Scholar]
  • 34.Lambright DG, Sondek J, Bohm A, Skiba NP, Hamm HE, Sigler PB. Nature. 1996;379:311–319. doi: 10.1038/379311a0. [DOI] [PubMed] [Google Scholar]
  • 35.Medina R, Grishina G, Meloni EG, Muth TR, Berlot CH. J. Biol. Chem. 1996;271:24720–24727. doi: 10.1074/jbc.271.40.24720. [DOI] [PubMed] [Google Scholar]

RESOURCES