The FASEB Journal • Research Communication

Stability of GABAB receptor oligomers revealed by dualTR-FRET and drug-induced cell surface targeting

Laëtitia Comps-Agrar,*,†,‡,1 Julie Kniazeff,*,†,2 Carsten Brock,*,† Eric Trinquet,†,‡

and Jean-Philippe Pin*,†,2

*Institut de Génomique Fonctionnelle, Centre National de la Recherche Scientifique (CNRS) UnitéMixte de Recherche (UMR)-5203, Institut National de la Santé et de la Recherche Médicale(INSERM) U661, Montpellier, France; †Universités Montpellier 1 and 2, Montpellier, France;and ‡Cisbio Bioassays, Codolet, France

ABSTRACT The function of cell surface proteinslikely involves the formation and dissociation of oligo-meric complexes. However, the dynamics of this pro-cess are unknown. Here we examined this process forthe GABAB receptors that assemble into oligomers ofheterodimers through the association of their GABAB1

subunit. We report a method to study oligomer dynam-ics based on a drug-controlled cell surface targeting ofintracellularly retained receptors and a parallel mea-surement of two FRET signals in HEK293 cells.GABAB1 subunits at the cell surface (4.0�0.6 a.u.) arelabeled with a pair of fluorophores (donor and redacceptor). New receptors are then targeted to the cellsurface during 3h treatment with AP21967 such that thenumber of receptors is doubled (9.1�0.7 a.u.). Afterlabeling these new receptors with a second acceptor(green), the red FRET remained unchanged (5189�36vs. 4783�32 cps), supporting the stability of the pre-formed oligomers. However, new oligomers are de-tected by the green FRET signal indicating both recep-tor populations are in the same microdomains. As acontrol, we confirmed the strict stability of theGABAB heterodimer itself. Herein, using a novelmethod to monitor the dynamics of cell surfacecomplexes, we provide evidence for the stability ofGABAB oligomers.—Comps-Agrar, L., Kniazeff, J.,Brock, C., Trinquet, E., Pin, J.-P. Stability of GABAB

receptor oligomers revealed by dual TR-FRET anddrug-induced cell surface targeting. FASEB J. 26,3430–3439 (2012). www.fasebj.org

Key Words: GPCRs � protein complex dynamics � protein-pro-tein interaction

G-protein-coupled receptors (GPCRs) form themost important family of cell surface receptors, repre-senting �3% of our genes. They are involved in manyphysiological processes and respond to a large variety ofstimuli (1). For over 10 years, even though monomericGPCRs represent the minimal entity able to activate Gproteins (2–4), it has been established that GPCRs canassemble into dimers that influence receptor proper-ties, such as cell surface targeting, pharmacology, sig-naling, or desensitization (5–8). More recently, severalstudies reported the existence of higher-ordered GPCRoligomers that could add even more complexity toreceptor functioning (9–11).

The �-aminobutyric acid B (GABAB) receptor repre-sents a good illustration of this phenomenon. It is anobligatory heterodimer composed of 2 subunits, GABAB1(GB1), which bears the agonist binding site (12, 13), andGABAB2 (GB2), which is responsible for G-protein activa-tion (14, 15). Expressed alone, GB1 is retained in theendoplasmic reticulum (ER) due to a retention signal inits C terminus, and only its coiled-coil interaction with theC-terminal tail of GB2 prevents the retention and pro-motes the cell surface targeting of the functional het-erodimer (16, 17). Recently, we reported that, both intransfected cells and in the brain, GABAB heterodimerscan form higher-ordered oligomers (likely tetramers)through the interaction of their GB1 subunits (10, 18).Such GABAB oligomers were detected through cell sur-face Förster resonance energy transfer (FRET) experi-ments using either Snap-tagged GB1 and GB2 (10) orfluorescent antibodies to label native GABAB receptors inbrain membranes (18). Western blot analysis of brainGABAB receptor complexes on native gels is also compat-ible with the existence of GABAB tetramers (19). No such

1 Current address: Genentech Inc., Department of ProteinChemistry, 1 DNA Way, South San Francisco, CA 94080, USA.

2 Correspondence: Institut de Génomique Fonctionnelle,Molecular Pharmacology Department, 141 rue de la Cardo-nille, F-34094 Montpellier Cedex 5, France. E-mail: J.-P.P.,jppin@igf.cnrs.fr; J.K., julie.kniazeff@igf.cnrs.fr

doi: 10.1096/fj.12-203646This article includes supplemental data. Please visit http://

www.fasebj.org to obtain this information.

Abbreviations: CT, Clip tag; DMEM, Dulbecco’s modifiedEagle’s medium; ER, endoplasmic reticulum; FKBP, FK506-binding protein; FRET, Förster resonance energy transfer;GABA, �-aminobutyric acid; GB1, �-aminobutyric acid Bsubunit 1 (GABAB1); GB2, �-aminobutyric acid B subunit 2(GABAB2); GPCR, G-protein-coupled receptor; HA, hemag-glutinin; ST, Snap tag; popA, population A; popB, populationB; TR-FRET, time-resolved Förster resonance energy transfer;WT, wild type

3430 0892-6638/12/0026-3430 © FASEB

higher-ordered oligomers was observed with other class CGPCRs such as metabotropic glutamate receptors (10,18). Surprisingly, GABAB receptor oligomeric assemblylimits receptor coupling to G protein (18). This led to theidea that the formation of these oligomers could be thesubject of specific regulation, promoting a tight control ofGABAB receptor coupling efficacy.

Here, we develop a direct method to address thedynamics of the oligomeric GABAB complexes at thesurface of transfected living cells. It is based onthe analysis of the association and exchange betweentwo receptor populations: one already present at thecell surface and a second, originally retained in the ER,but targeted to the plasma membrane after drug appli-cation (see Fig. 1). The interactions are detected usingtime-resolved FRET (TR-FRET), a technique giving ahigh signal to noise ratio (20, 21). It is based on the useof a single donor fluorophore, Lumi4Tb, characterizedby a very long fluorescence half-life and by a spikyemission spectrum, conferring it the ability to transferto both green- and far-red-emitting acceptor fluoro-phores (ref. 22 and Supplemental Fig. S1). Here, wecombine TR-FRET with Snap-tag (ST) technology thatallows the specific and covalent labeling of cell surfaceproteins without altering the functional properties(10). When applied to study the stability of the GABABoligomers, this approach brought evidence for a lim-ited association of subunits between those already atthe cell surface and the new one reaching the surface.Not only does this study bring new information regard-ing the intrinsic stability of the GABAB receptor oligom-ers as determined in heterologous cells, but it alsoreports a new approach to examine the cell surfacedynamics of many other membrane protein complexes.

MATERIALS AND METHODS

Fluorophores

The SNAP-Lumi4Tb (SSNPTBE), SNAP-Red (SSNPREDE),SNAP-Green (SSNPGRNE), CLIP-Lumi4Tb (SCLPTBE),CLIP-Red (SCLPREDE) and CLIP-Green (SCLPGRNE) sub-strates are commercially available from Cisbio Bioassays(Codolet, France).

Plasmids and site-directed mutagenesis

Plasmids encoding the wild-type (WT) GABAB1a and GABAB2subunits bearing a hemagglutinin (HA), Flag, or c-Mycepitope or an ST at their N terminus were described previ-ously (10, 23). The sequence encoding FRB*-KKXX-stopcodon was inserted between a NotI site, engineered at posi-tions 843, 878, 905, and 940 of GB2 (Quick-Change; Strat-agene, Santa Clara, CA, USA), and HindIII.

Cell culture and transfection

HEK293 cells were cultured in Dulbecco’s modified Eagle’smedium (DMEM) supplemented with 10% FBS and tran-siently transfected with the Lipofectamine 2000 following themanufacturer’s instructions. Cells were then seeded out in a

96-well CellStar black plate (Greiner Bio-One, Frickenhausen,Germany) at 100,000 cells/well. All media used for cellculture were purchased from Life Technologies (Grand Is-land, NY, USA).

Cell surface expression quantification by ELISA

ELISA on intact cells was performed as described previously(10). Briefly, after fixation (paraformaldehyde 4%) andblocking (phosphate-buffered saline with 1% fetal calf se-rum), cells were incubated with anti-HA monoclonal antibody(clone 3F10; Roche Applied Science, Basel, Switzerland) oranti-Flag-M2 monoclonal antibody (Sigma-Aldrich, Saint-Louis, MO, USA), both conjugated with horseradish peroxi-dase (30 min at 0.5 �g/ml). After washes, bound antibody wasdetected by chemoluminescence using SuperSignal substrate(Pierce, Rockford, IL, USA) and a Wallac Victor2 counter(Molecular Devices, Sunnyvale, CA, USA).

TR-FRET between two STs or between ST and Clip tag(CT) using sequential labeling

At 24 h after transfection, cells were washed with DMEM andincubated with BG or BC substrates diluted in DMEM for 1 h(for SNAP substrates) or 2 h (for CLIP substrates) at 37°C, 5%CO2 (labeling 1, as indicated in Supplemental Table S1) (10,24). After 4 washes with Tris-KREBS buffer (20 mM Tris, pH7.4; 118 mM NaCl; 5.6 mM glucose; 1.2 mM KH2PO4; 1.2 mMMgSO4; 4.7 mM KCl; and 1.8 mM CaCl2), FRET signal wasrecorded at 665 nm on a RubyStar plate reader (BMGLabtech, Orternberg, Germany). Cells were then incubatedwith 100 �l of 1 �M AP21967 diluted in DMEM (or DMEMfor negative control) for 2 or 1 h when SNAP or CLIPderivatives, respectively, were used for labeling 2. Then, 50�l/well of AP21967 (or DMEM) was removed, and 50 �l of2� fluorescent SNAP or CLIP substrates were added andincubated for 1 or 2 h, respectively (labeling 2, SupplementalTable S1). After 4 washes with Tris-KREBS buffer, the FRETsignals at 665 and 520 nm were recorded on a RubyStar platereader and an Infinite 500 reader (Tecan, Männedorf, Swit-zerland), respectively. FRET signal was calculated as the FRETintensity, where the nonspecific FRET due to random colli-sions and the contamination of the donor at the FRETwavelength (665 or 520 nm) were subtracted; i.e., FRETintensity � (signal at 665 or 520 nm measured on cellscolabeled with the donor and the acceptor) � (signal re-corded on the same cells labeled with the donor only). ForFRET calculations, we verified that donor emission was con-stant for all conditions. Population A (popA) � population B(popB) FRET (at 665 and 520 nm) was calculated by subtract-ing the FRET signal measured on cells incubated 3 h withDMEM (instead of AP21967) to that measured on the samecells treated with 1 �M AP21967 in order to take into accountall the cellular processes (internalization, recycling, andneosynthesis) during the 3-h drug incubation.

Cell imaging

Transiently transfected HEK293 cells were seeded out oncoverslips pretreated with poly-d-lysine. At 24 h aftertransfection, cells were labeled with SNAP-Red (0.3 �M)diluted in DMEM for 1 h at 37°C, 5% CO2. After 4 washeswith Tris-KREBS buffer, the cells were either fixed (4%PFA) or treated with AP21967. After 2 h treatment withAP21967, half of the volume was removed and replaced bySNAP-Green (0.3 �M final concentration) for 1 h. After 4washes, the coverslips were mounted on a microscope slideusing Prolong Gold. Fixed cells were permeabilized with

3431GABAB RECEPTOR TETRAMERS ARE STABLE COMPLEXES

Triton (0.2%) and then labeled with SNAP-Green (0.3 �Mfinal concentration) for 1 h. Cell nuclei were counter-stained with Hoechst 33342. The images were taken on aLSM 510 Zeiss microscope using a �40 objective (CarlZeiss, Oberkochen, Germany) at the Montpellier RIOImaging facility (Institute of Human Genetics) and pro-cessed using ImageJ software (U.S. National Institutes ofHealth, Bethesda, MD, USA).

Intracellular calcium measurements

At 24 h after transfection, Ca2� mobilization in response toincreasing concentrations of GABA was measured fromHEK293 cells transiently transfected with plasmids encodingthe indicated GABAB receptor subunits and the chimeric Gprotein Gqi9 using Fluo4 a.m. (Molecular Probes, Eugene,OR, USA) and a Flexstation microplate reader (MolecularDevices), as previously described (10).

Binding assay

At 24 h after transfection, Bmax was determined from HEK293cells transiently transfected with the GABAB receptor subunitsusing the nonpermeant [3H]-CGP54626A, as previously de-scribed (10).

RESULTS

Drug-induced targeting of GABAB receptor to the cellsurface

To monitor the stability of GABAB receptor tetramers,we aimed at measuring the association of cell surfaceoligomers (popA), when new receptors are targeted tothe cell surface (popB) (Fig. 1). Our strategy was basedon a drug-inducible targeting of popB receptorsadapted from the Ariad “regulated heterodimerization”method (Ariad Pharmeceuticals, Inc., Cambridge, MA,USA; refs 25–27). This system uses a “dimerizer” drug(AP21967) that promotes the high-affinity interactionbetween the cytosolic FK506-binding protein (FKBP)family (endogenously expressed in HEK293 cells) and aprotein domain FRB*. Here, FRB* was inserted in theC-terminal tail of GB2 together with the C-terminal ERretention COP1 binding motif KKXX (the KKDL motif

was used) (28, 29). When AP21967 is applied to thecells, the ternary complex FKBP/AP21967/FRB* isexpected to form and to prevent the KKXX interactionwith COP1, hence releasing GB2 from the ER andallowing it to reach the cell surface (Fig. 2A).

The sequence encoding FRB*-KKXX was fused at theC terminus of the GB2 subunit at 4 different positions(Fig. 2B). All 4 constructs were efficiently retainedinside the cells, as assessed by ELISA against the Flagepitope inserted at the N terminus of GB2 performedon both intact and permeabilized cells (Fig. 2C). Afteraddition of AP21967, only 3 Flag-GB2-FRB*-KKXX con-structs were correctly targeted to the cell surface.Because a better ratio between the basal and theAP21967-induced cell surface expressions was obtainedwith construct 2 (insertion at position 878; Fig. 2D), itwas used for the rest of the study. As expected, Flag-GB2-FRB*-KKXX did not alter the HA-GB1 retention inthe ER, and the heterodimer was targeted to the cellsurface in a drug-inducible manner (Fig. 2E). We alsooptimized both the AP21967 concentration (1 �M) andthe incubation time (3 h) (Fig. 2F, G). Such modifica-tion of the GB2 subunit did not affect GABAB receptorfunction (GABA pEC50 of 7.17�0.07 and 7.19�0.09 forreceptors containing WT-GB2 and GB2-FRB*-KKXX,respectively, in a Ca2� mobilization assay; Supplemen-tal Fig. S2).

To monitor the association/dissociation dynamics ofthe GABAB oligomers, a combination of plasmids en-coding GB1, GB2, and GB2-FRB*-KKXX was trans-fected in HEK293 cells. The basal (popA) andAP21967-induced cell surface populations (popB) werealways determined by cell surface ELISA. The condi-tions were set up such that a 2-fold increase in the totalcell surface population of receptors was obtained aftertreatment with AP21967. This was confirmed by mea-suring the binding of [3H]-CGP54262A on intact cellsbefore (popA receptors 0.91�0.06 pmol/mg protein)and after (popA�popB receptors 1.80�0.07 pmol/mgprotein) AP21967 treatment. Of note, these expressionlevels are in line with those determined on culturedneurons (1.5 pmol/mg proteins; ref. 18), indicatingthat our experiments are performed under conditions

Figure 1. Schematic representation of themethod to study the association-dissociationoccurring within and between the complexes atthe cell surface. Interaction between a pool ofreceptors present at the cell surface (popA)could be assessed by FRET, while a second poolof receptors (popB) would be retained in theER. By adding a drug, the second pool ofreceptors would be targeted at the cell surface,and the potential reorganization between thetwo populations would be monitored by re-cording two FRET signals at two distinct wave-lengths.

3432 Vol. 26 August 2012 COMPS-AGRAR ET AL.The FASEB Journal � www.fasebj.org

consistent with native expression of the GABAB recep-tor. To measure FRET between the two populations,N-terminal ST-GB1 subunits were covalently labeledwith nonpermeant fluorescent benzyl-guanine deriva-tives carrying either the donor or the acceptor fluoro-phores compatible with TR-FRET (SNAP-Lumi4Tb andSNAP-Red). As previously reported, the conditionsused allowed a full and covalent labeling of all cellsurface STs with SNAP derivatives (10).

GABAB oligomers at the cell surface are stable

In a first experimental paradigm (Fig. 3A), popA ST-GB1subunits were labeled with the TR-FRET donor, SNAP-Lumi4Tb, for 1 h. Then, popB heterodimers were targetedto the cell surface, and the new ST-GB1 subunits were

labeled with the FRET acceptor SNAP-Red. We ensuredconditions to obtain a similar proportion of popA and popBat the cell surface after AP21967 treatment (Fig. 3B). Weanalyzed the distribution of popA and popB receptors in thecell population and found that 71 � 1% of transfected cellsexpressed both receptor populations (Fig. 3D and Supple-mental Fig. S3). The FRET between popA and popB GB1subunits (popA-popB FRET) was then recorded. The signalwas corrected for receptor internalization, recycling, andnew synthesis during the 3-h incubation, by subtracting thesignal measured on cells treated in parallel with culturemedium (DMEM) instead of AP21967. In these conditions,we measured a low but significant FRET signal betweenpopA and popB receptors (Fig. 3C). We compared thissignal to the maximal FRET signal (FRETmax) that would beobtained if all popA and popB receptors would be

Figure 2. Drug-inducible targeting system using FRB* domain and KKXX motif on GB2. A) Diagram of the drug-inducible targetingsystem applied to the GABAB receptor. The GABAB heterodimer is retained in the ER due to the interaction of KKXX retention motifof GB2 with COP1. After adding AP21967, the ternary complex FRB*-AP21967-FKBP is formed and masks the KKXX-motif, and theheterodimer is targeted to the cell surface. B) Topological organization of WT GB2 and the 4 FRB*-KKXX GB2 constructs (#1 to #4).VFT, Venus flytrap (GABAB extracellular domain); HD, heptahelical domain; CC, coiled-coil. C) Expression of GB2 constructsmeasured by anti-Flag ELISA on intact (white) or permeabilized (gray) transiently transfected HEK293 cells. D) Cell surfaceexpression of the GB2 constructs, measured by ELISA on intact cells after 3 h incubation with DMEM (white) or 1 �M AP21967(black). E) Cell surface expression of HA-GB1, measured by ELISA on intact cells also expressing Flag-GB2-FRB*-KKXX 2 after 3 hincubation with DMEM (white) or 1 �M AP21967 (black). Data are means � se of 3 independent experiments performed intriplicate. F, G) Evolution of the cell surface expression of GABAB heterodimer (HA-GB1/Flag-GB2-FRB*-KKXX), measured byanti-HA ELISA on intact cells relative to AP21967 concentration (F) and incubation time (G) in the presence of 1 �M AP21967. Dataare representative means � se of 3 independent experiments performed in triplicate.

3433GABAB RECEPTOR TETRAMERS ARE STABLE COMPLEXES

statistically and evenly distributed between the oligom-ers at the cell surface. We assessed this FRETmax on thesame cells through a unique labeling step with bothSNAP-Lumi4Tb and SNAP-Red after the popB recep-tors were targeted to the cell surface. The labelingconditions were such that any ST-GB1 was labeledeither by the donor or the acceptor fluorophore withan equiprobability (10). We found that the popA-popBFRET signal represents 29.1 � 3.5% of FRETmax (Fig.3C), indicating that a small proportion of GB1 subunits

already at the cell surface (popA) are in proximity withnewly targeted GB1 subunits (popB).

New oligomeric entities are formed after targeting ofnew receptors

To check whether the popA-popB FRET arises fromreceptor exchange or simple association between popAand popB oligomers, we designed a second experimen-tal paradigm in which two different FRET signals

Figure 3. Experimental paradigm 1 (association). A) Schematic representation of experiment principle. In step 1, popA(HA-ST-GB1/GB2 heterodimers) is labeled with SNAP-Lumi4Tb, while popB (HA-ST-GB1/GB2-FRB*-KKXX heterodimers) isretained in the ER. In step 2, AP21967 (or DMEM) is applied; popB reaches the cell surface and is labeled with SNAP-Red.Association between popA and popB heterodimers is monitored by measuring FRET between HA-ST-GB1 subunits. FRETmax isdetermined when both popA and popB are labeled with a mixture of SNAP-Lumi4Tb and SNAP-Red. B) Cell surface expressionof GABAB heterodimer, determined by anti-HA ELISA prior (popA; green) or after DMEM (popA; white) or AP21967(popA�popB; black) treatment. C) popA-popB FRET (light gray) and FRETmax (dark gray) as determined in the paradigm 1.Data are representative means � se of 3 independent experiments performed in triplicate. D) Confocal images of cells labeledwith SNAP-Red (popA) and then with SNAP-Green (popB) after fixation and permeabilization (left panel) or after AP21967treatment (right panel). Nucleus is shown in blue. Bar graph represents the quantification of transfected cells expressing eitherpopA and popB, popA only, or popB only; n � 577 cells analyzed (see also Supplemental Fig. S3).

3434 Vol. 26 August 2012 COMPS-AGRAR ET AL.The FASEB Journal � www.fasebj.org

(recorded at different wavelengths) were monitored atthe same time. Indeed, the FRET between GB1 sub-units within popA tetramers was measured with a redacceptor, while the association between popA GB1 andpopB GB1 was determined using a green acceptor(Fig. 4A). This is possible thanks to the peculiar spectralproperties of the donor, which emits at several wave-lengths, and the absence of any cross contaminationbetween the two FRETs (ref. 22 and Supplemental Fig.S1). In that setup, FRETmax should be defined throughthe random labeling of all ST-GB1 subunits (from popAand popB) with all 3 substrates, SNAP-Lumi4Tb, SNAP-Red, and SNAP-Green, at a 25:25:50 ratio. Because sucha ratio cannot be experimentally reached easily, wemeasured the signal recorded with SNAP-Lumi4Tb andSNAP-Green only at a 50:50 ratio. The FRET signalmeasured under these conditions is theoretically 2-foldhigher than that expected due to the 2-fold increase inthe donor labeling, such that FRETmax was then de-fined as half of that signal. In our conditions, theamount of ST-GB1 at the cell surface is doubled afterAP21967 treatment, as assessed by ELISA (Fig. 4B). Inagreement with the data obtained in experimentalparadigm 1, some ST-GB1 subunits from popB were inclose contact with popA ST-GB1 subunits, as shown bythe low but significant popA-popB FRET signal re-corded in the green channel (30.2�3.5% of FRETmax;Fig. 4D). However, we found that the red FRET (FRET

within popA) remained constant after targeting ofpopB to the cell surface, suggesting that popA tetram-ers did not dissociate (Fig. 4C). As a control, we verifiedthat the separation of the oligomers by expressing acompeting GB1 without ST resulted in a large decreasein red FRET signal (Supplemental Fig. S4). This find-ing further supports the suggestion that a small propor-tion of popA GB1 associates with popB receptors with-out a real exchange of subunits between oligomers atthe cell surface. Altogether, our data suggest thatGB1-GB1 interaction is stable once formed.

To make sure that the popA-popB FRET signaloriginates from a specific oligomeric organization ofGABAB rather than from a random clustering, weapplied our experimental paradigms to a model ex-pected to be a positive control of stability: the GABABheterodimer itself (i.e., between GB1 and GB2). In-deed, this heterodimer is stabilized via a coiled-coilinteraction between the C-terminal tails of the twosubunits (Fig. 5 and ref. 30). For those experiments,because we monitored an interaction between twodistinct subunits, we used a GB2 carrying a CT, anengineered ST that specifically reacts with benzyl-cyto-sine derivatives (CLIP substrates), instead of SNAPderivatives (24, 31). In the first paradigm, popA CT-GB2 subunits were labeled with the donor CLIP-Lumi4Tb while popA ST-GB1 subunits were madeinactive with unlabeled SNAP substrate. In the second

Figure 4. Experimental paradigm 2 (association and dissociation). A) Schematic representation of experiment principle. In step1, popA receptors are labeled with a mixture of SNAP-Lumi4Tb and SNAP-Red. In step 2, popB is labeled with SNAP-Green.Dissociation of popA tetramers is given by red FRET; association between popA and popB heterodimers is monitored bymeasuring green FRET. FRETmax is determined when both popA and popB are labeled with a mixture of SNAP-Lumi4Tb andSNAP-Green. B) Cell surface expression of GABAB heterodimer, determined by anti-HA-ELISA prior (popA; green) or afterDMEM (popA; white) or AP21967 (popA�popB; black) application. C) Red FRET intensities measured without (white) or afterAP21967 (black). D) Representation of green FRET intensities, popA-popB FRET (light gray) and FRETmax (dark gray). Dataare representative means � se of 3 independent experiments performed in triplicate.

3435GABAB RECEPTOR TETRAMERS ARE STABLE COMPLEXES

step, popB is targeted to the cell surface and the newST-GB1 subunits were labeled with the acceptor SNAP-Red (Fig. 5A). After AP21967 treatment (Fig. 5B), nosignificant FRET was measured between popA GB2 andpopB GB1 (1% of FRETmax), supporting the conclu-sion that no random clustering occurred when thesecond population was targeted to the plasma mem-brane (Fig. 5C). According to the second paradigm(described in Fig. 5D), the stability of the GB1-GB2interaction was confirmed, as no variation of the redFRET signal was observed after popB targeting (Fig.5E). In addition, no significant green FRET signal appeared(3% of FRETmax; Fig. 5F), further supporting the ab-sence of random GABAB heterodimer clustering.

In an effort to analyze the propensity of thereceptors to form new interactions, we calculated aprobability of labeling with each fluorophore accord-ing to the expression level of popA and popB recep-

tor subunits (see Supplemental Information). Usingthese results, we calculated the theoretical FRETvalue expected if a random repartition of the differ-ent populations occurred during the time frame ofthe experiment (resulting in an even distribution ofthe labeled subunits). This value has been set to100%, and each experimental FRET signal was calcu-lated as a percentage of that theoretical value. Thesecalculations support the formation of a small propor-tion of GABAB oligomers (30%) composed of popAand popB GB1 subunits, and this occurs without thedissociation of the preexisting popA oligomers(FRET within popA remains at 102 � 3% aftertargeting of popB). In contrast, and as expectedbased on the stable association of the heterodimerthrough the coiled-coil interaction, new assembly ofGABAB heterodimer made of popA GB2 and popBGB1 is very unlikely (3.2 � 0.2%).

Figure 5. GABAB receptor heterodimer stability. A) Schematic representation of paradigm 1 applied to GABAB heterodimer. Instep 1, CT-GB2 is specifically labeled with CLIP-Lumi4Tb, the SNAP-tag on GB1 is saturated by addition of unlabeled SNAPsubstrate. In step 2, GB1 from popB is labeled with SNAP-Red. Emergence of interaction between GB1 from popB and GB2 frompopA is given by red FRET. B) Cell surface expression of GABAB heterodimer, determined by ELISA prior (green) or afterDMEM (white) or AP21967 (black) application. C) popA-popB FRET (light gray) and FRETmax (dark gray) as determined inthe paradigm 1. D) Schematic representation of paradigm 2 applied to GABAB heterodimer. In step 1, popA GB1 is labeled withSNAP-Lumi4Tb and GB2 with CLIP-Red. In step 2, popB GB2 is labeled with CLIP-Green. Dissociation of popA heterodimersis given by red FRET; association between popA GB1 and popB GB2 is monitored by measuring green FRET. FRETmax isdetermined when both popA and popB are labeled with a mixture of SNAP-Lumi4Tb and CLIP-Green. E) Red FRET intensitiesmeasured without (white) or after AP21967 (black). F) Representation of green FRET intensities, popA-popB FRET (light gray)and FRETmax (dark gray) as determined in the paradigm 2. Data are representative means � se of 3 independent experimentsperformed in triplicate.

3436 Vol. 26 August 2012 COMPS-AGRAR ET AL.The FASEB Journal � www.fasebj.org

DISCUSSION

In the present study, we used an original method tostudy the stability of oligomeric GABAB receptorcomplexes. Indeed, we developed a new approachthat allows the rapid insertion of new receptors at thecell surface, and the examination of their ability toassociate with previously labeled cell surface recep-tors and disrupt preexisting GABAB oligomers. Ourdata show that the new GABAB receptors inserted atthe surface do not dissociate the preexisting oligom-ers within the 3-h time frame of the analysis, al-though the new GABAB heterodimers could clearlyassociate with preexisting ones.

A drug-inducible system that controls the secretion ofproteins retained intracellularly was previously devel-oped based on a self-associating mutant of FKBP thatclusters the proteins in the ER. Such clusters could bedissociated on binding of rapamycin or the higher-affinity ligand AP21998, allowing their traffic throughthe Golgi (32). This system was found optimal tocontrol the secretion of various proteins, but it hasbeen only occasionally used to control the cell surfacetargeting of membrane proteins (33, 34). Indeed, thissystem being based on the clustering of proteins, wewere afraid that this may not be optimal when used withself-associating proteins such as GPCRs, which couldeventually remain stable oligomers even after beingtargeted to the cell surface. We then designed analternative approach to control protein retention in theER. We added a COP1-binding motif at the C-terminalend of FRB* (28), and fused this construct to theC-terminal intracellular tail of GB2. Our idea was thatthe FKBP-FRB* interaction promoted by the dimerizerAP21967 would prevent COP1 interaction, allowing theprotein to process through the Golgi and reach the cellsurface. This approach worked properly, and allowedus to control in a reasonable time frame (within 3 h),the amount of GABAB receptors at the cell surface. Wealso found that it could be used to control the targetingof other plasma membrane proteins such as the �1-adrenergic receptor or cluster of differentiation 4(CD4) (Supplemental Fig. S5).

When associated with the cell surface covalent label-ing of proteins with TR-FRET compatible fluorophoresusing the ST (or CT) strategy (10, 24, 31, 35), thisallowed us to follow either the stability of preformedoligomers (popA), or the formation of new oligomerswhen new receptors (popB) are targeted to the plasmamembrane. Using a single donor, Lumi4Tb, compati-ble with two distinct acceptors (red and green), we haveeven been able to follow both processes at the sametime. Of note, this method is simple and can beperformed with a TR-FRET-compatible plate reader, ina 384-well plate format.

Surprisingly, while no dissociation of the preexistingpopA GABAB complexes was observed, the formationof new complexes containing popA and popB receptorscould be detected. Two hypotheses are compatible withthese observations: The first hypothesis is that the size

of the GABAB complexes increases according to the cellsurface receptor density. This would result in theappearance of a green FRET without changes in the redFRET. The second hypothesis is that each receptorpopulation contains a mixture of stable tetramers andstrict heterodimers, and after targeting of popB, recep-tors from each population could interact and form newstable tetramers. This would result in the measurementof a green FRET, while the red FRET would remainunchanged. In our previous study we favored the ideathat GABAB oligomers are limited to tetramers (10, 18).Indeed, although very efficient FRET can be measuredbetween GB1 subunits, no significant FRET can bemeasured between GB2 subunits, indicating that theseGB2 subunits remain at a large distance. In addition,the present data indicate that even though new GABABcomplexes can be detected between popA and popBheterodimers, no FRET signal could be measuredbetween popA GB1 and popB GB2 subunits. Of note,the physical properties of the fluorophores (R0) usedallow measurement of FRET between partners at adistance up to 75 Å, and the distance between GB1 andGB2 within oligomers is estimated to be at most 48 Å.Taken together, these results suggest an organizationlimited to tetramers with a fraction of ‘free’ het-erodimers at the cell surface rather than the formationin higher-ordered oligomers, but the question remainsopen, and a more direct demonstration is still awaited.

These data contrast with those recently reportedusing either FRAP or single-molecule fluorescence im-aging, which revealed a rapid exchange of subunitsbetween class A GPCR dimers (36–38). One may arguethat the new GABAB receptors targeted to the cellsurface (popB) are delivered in submembrane domainsdifferent from those that contain the preexistingGABAB receptors (popA), each being separated bymembrane diffusion barriers (39, 40), limiting proteinexchange. However, this possibility is very unlikely,since at least a detectable fraction of the popB recep-tors can associate with the preexisting ones, as revealedby popA-popB FRET resulting from a specific associa-tion of the GABAB heterodimers. Another limitationwould be that few cells expressed both popA and popBreceptors due to the transient transfection procedure,resulting in a reduced number of cells where thereceptor exchange could occur compared to totalnumber of cells used for FRET recording; however thiswas not the case, as 71 � 1% of the cells expressingpopA receptors also expressed popB receptors. Werecently reported that the large extracellular domain ofthe GB1 subunit, a domain absent in most class AGPCRs, is critical in the formation of GABAB tetramers.It is then likely that such GB1 interaction is responsiblefor the stability of the GABAB tetramers.

Our data indicate that GABAB heterodimers canform stable tetramers in transfected cells. AlthoughGABAB tetramers have already been observed in thebrain, whether these are also similarly stable remains tobe examined. Indeed, it is possible that associated pro-

3437GABAB RECEPTOR TETRAMERS ARE STABLE COMPLEXES

teins or regulation through the activity of specific signal-ing pathways can regulate this process in vivo.

The authors thank Drs. Sébastien Granier, DamienMaurel, and Gregory Stewart for constructive discussions.Ariad Pharmeceuticals, Inc. (Cambridge, MA, USA) grace-fully provided the AP21967. The FRET experiments havebeen performed using the ARPEGE (Pharmacology Screen-ing-Interactome) platform facility at the Institut de Gé-nomique Fonctionnelle (Montpellier, France). The au-thors thank the Montpellier RIO Imaging facility of theInstitut de Génétique Humaine (Montpellier, France).This work was supported by Centre National de la Recher-che Scientifique, Institut National de la Santé et de laRecherche Médicale, and Cisbio Bioassays, and by grantsfrom the French Ministry of Research, Agence Nationalede la Recherche (ANR-06-BLAN-0087 and ANR-09-BLAN-0272), and by an unrestricted grant from Senomyx (SanDiego, CA, USA). L.C.-A. was supported by a ConventionsIndustrielles de Formation par la Recherche fellowshipfrom Cisbio Bioassays and the French government. Authorcontributions: L.C.A. conceived, performed, and analyzedall the experiments and wrote the manuscript; J.K. partic-ipated in the experiments and their analysis and wrote themanuscript; C.B. initiated the validation of the drug-induced targeting system on CD4; E.T. and J.P.P. super-vised the work; and J.P.P. and J.K. finalized the manuscript.

REFERENCES

1. Bockaert, J., and Pin, J.-P. (1999) Molecular tinkering of G-pro-tein coupled receptors: an evolutionary success. EMBO J. 18,1723–1729

2. Bayburt, T. H., Leitz, A. J., Xie, G., Oprian, D. D., and Sligar,S. G. (2007) Transducin activation by nanoscale lipid bilayerscontaining one and two rhodopsins. J. Biol. Chem. 282, 14875–14881

3. Rasmussen, S. G., Devree, B. T., Zou, Y., Kruse, A. C., Chung,K. Y., Kobilka, T. S., Thian, F. S., Chae, P. S., Pardon, E.,Calinski, D., Mathiesen, J. M., Shah, S. T. A., Lyons, J. A.,Caffrey, M., Gellman, S. H., Steyaert, J., Shiniotis, G., Weis, W. I.,Sunahara, R. K., and Kobilka, B. K. (2011) Crystal structure ofthe beta(2) adrenergic receptor-Gs protein complex. Nature477, 549–555

4. Whorton, M. R., Bokoch, M. P., Rasmussen, S. G., Huang, B.,Zare, R. N., Kobilka, B., and Sunahara, R. K. (2007) A mono-meric G protein-coupled receptor isolated in a high-densitylipoprotein particle efficiently activates its G protein. Proc. Natl.Acad. Sci. U. S. A. 104, 7682–7687

5. Bouvier, M. (2001) Oligomerization of G-protein-coupled trans-mitter receptors. Nat. Rev. Neurosci. 2, 274–286

6. Gurevich, V. V., and Gurevich, E. V. (2008) GPCR monomersand oligomers: it takes all kinds. Trends Neurosci. 31, 74–81

7. Milligan, G., and Smith, N. J. (2007) Allosteric modulation ofheterodimeric G-protein-coupled receptors. Trends Pharmacol.Sci. 28, 615–620

8. Terrillon, S., and Bouvier, M. (2004) Roles of G-protein-coupledreceptor dimerization. EMBO Rep. 5, 30–34

9. Carriba, P., Navarro, G., Ciruela, F., Ferre, S., Casado, V., Agnati,L., Cortes, A., Mallol, J., Fuxe, K., Canela, E. I., Lluis, C., andFranco, R. (2008) Detection of heteromerization of more thantwo proteins by sequential BRET-FRET. Nat. Methods 5, 727–733

10. Maurel, D., Comps-Agrar, L., Brock, C., Rives, M.-L., Bourrier,E., Ayoub, M. A., Bazin, H., Tinel, N., Durroux, T., Prézeau, L.,Trinquet, E., and Pin, J.-P. (2008) Cell surface protein-proteininteraction analysis with combined time-resolved FRET andsnap-tag technologies: application to GPCR oligomerization.Nat. Methods 5, 561–567

11. Pisterzi, L. F., Jansma, D. B., Georgiou, J., Woodside, M. J.,Chou, J. T., Angers, S., Raicu, V., and Wells, J. W. (2010)Oligomeric size of the m2 muscarinic receptor in live cells as

determined by quantitative fluorescence resonance energytransfer. J. Biol. Chem. 285, 16723–16738

12. Galvez, T., Prézeau, L., Milioti, G., Franek, M., Joly, C., Froestl,W., Bettler, B., Bertrand, H.-O., Blahos, J., and Pin, J.-P. (2000)Mapping the agonist binding site of GABAB type 1 subunitsheds light on the activation process of GABAB receptors. J. Biol.Chem. 275, 41166–41174

13. Kniazeff, J., Galvez, T., Labesse, G., and Pin, J.-P. (2002) Noligand binding in the GB2 subunit of the GABAB receptor isrequired for activation and allosteric interaction between thesubunits. J. Neurosci. 22, 7352–7361

14. Duthey, B., Caudron, S., Perroy, J., Bettler, B., Fagni, L., Pin,J. P., and Prezeau, L. (2002) A single subunit (GB2) is requiredfor G-protein activation by the heterodimeric GABA(B) recep-tor. J. Biol. Chem. 277, 3236–3241

15. Galvez, T., Duthey, B., Kniazeff, J., Blahos, J., Rovelli, G., Bettler,B., Prézeau, L., and Pin, J.-P. (2001) Allosteric interactionsbetween GB1 and GB2 subunits are required for optimalGABAB receptor function. EMBO J. 20, 2152–2159

16. Margeta-Mitrovic, M., Jan, Y. N., and Jan, L. Y. (2000) Atrafficking checkpoint controls GABA(B) receptor het-erodimerization. Neuron 27, 97–106

17. Pagano, A., Rovelli, G., Mosbacher, J., Lohmann, T., Duthey, B.,Stauffer, D., Ristig, D., Schuler, V., Meigel, I., Lampert, C., Stein,T., Prezeau, L., Blahos, J., Pin, J., Froestl, W., Kuhn, R., Heid, J.,Kaupmann, K., and Bettler, B. (2001) C-terminal interaction isessential for surface trafficking but not for heteromeric assem-bly of GABA (b) receptors. J. Neurosci. 21, 1189–1202

18. Comps-Agrar, L., Kniazeff, J., Nørskov-Lauritsen, L., Maurel,D., Gassmann, M., Gregor, N., Prézeau, L., Bettler, B.,Durroux, T., Trinquet, E., and Pin, J.-P. (2011) The oligo-meric state sets GABAB receptor signalling efficacy. EMBO J.30, 2336 –2349

19. Schwenk, J., Metz, M., Zolles, G., Turecek, R., Fritzius, T., Bildl,W., Tarusawa, E., Kulik, A., Unger, A., Ivankova, K., Seddik, R.,Tiao, J. Y., Rajalu, M., Trojanova, J., Rohde, V., Gassmann, M.,Schulte, U., Fakler, B., and Bettler, B. (2010) Native GABA(B)receptors are heteromultimers with a family of auxiliary sub-units. Nature 465, 231–235

20. Bazin, H., Trinquet, E., and Mathis, G. (2002) Time resolvedamplification of cryptate emission: a versatile technology totrace biomolecular interactions. J. Biotechnol. 82, 233–250

21. Selvin, P. R. (2002) Principles and biophysical applications oflanthanide-based probes. Annu. Rev. Biophys. Biomol. Struct. 31,275–302

22. Pin, J.-P., Maurel, D., Comps-Agrar, L., Monnier, C., Rives, M.,Doumazane, E., Rondard, P., Durroux, T., Prézeau, L., andTrinquet, E. (2010) Time-resolved FRET approaches to studyGPCR complexes. In G Protein-Coupled Receptors: Structure, Signal-ing, and Physiology (Siehler, S., and Milligan, G., eds) pp. 67–89,Cambridge University Press, Cambridge, UK

23. Kniazeff, J., Saintot, P.-P., Goudet, C., Liu, J., Charnet, A.,Guillon, G., and Pin, J.-P. (2004) Locking the dimeric GABABG-protein coupled receptor in its active state. J. Neurosci. 24,370–377

24. Doumazane, E., Scholler, P., Zwier, J., Trinquet, E., Rondard, P.,and Pin, J.-P. (2011) A new approach to analyze cell surfaceprotein complexes reveals specific heterodimeric metabotropicglutamate receptors. FASEB J. 25, 66–77

25. Belshaw, P. J., Ho, S. N., Crabtree, G. R., and Schreiber, S. L.(1996) Controlling protein association and subcellular localiza-tion with a synthetic ligand that induces heterodimerization ofproteins. Proc. Natl. Acad. Sci. U. S. A. 93, 4604–4607

26. Pollock, R., Issner, R., Zoller, K., Natesan, S., Rivera, V. M., andClackson, T. (2000) Delivery of a stringent dimerizer-regulatedgene expression system in a single retroviral vector. Proc. Natl.Acad. Sci. U. S. A. 97, 13221–13226

27. Terrillon, S., and Bouvier, M. (2004) Receptor activity-indepen-dent recruitment of betaarrestin2 reveals specific signallingmodes. EMBO J. 23, 3950–3961

28. Bonifacino, J. S., and Lippincott-Schwartz, J. (2003) Coat pro-teins: shaping membrane transport. Nat. Rev. Mol. Cell. Biol. 4,409–414

29. Lee, M. C., Miller, E. A., Goldberg, J., Orci, L., and Schekman,R. (2004) Bi-directional protein transport between the ER andGolgi. Annu. Rev. Cell Dev. Biol. 20, 87–123

3438 Vol. 26 August 2012 COMPS-AGRAR ET AL.The FASEB Journal � www.fasebj.org

30. Calver, A. R., Robbins, M. J., Cosio, C., Rice, S. Q., Babbs, A. J.,Hirst, W. D., Boyfield, I., Wood, M. D., Russell, R. B., Price,G. W., Couve, A., Moss, S. J., and Pangalos, M. N. (2001) TheC-terminal domains of the GABA (b) receptor subunits mediateintracellular trafficking but are not required for receptor signal-ing. J. Neurosci. 21, 1203–1210

31. Gautier, A., Juillerat, A., Heinis, C., Correa, I. R., Jr., Kinder-mann, M., Beaufils, F., and Johnsson, K. (2008) An engineeredprotein tag for multiprotein labeling in living cells. Chem. Biol.15, 128–136

32. Rivera, V. M., Wang, X., Wardwell, S., Courage, N. L., Volchuk,A., Keenan, T., Holt, D. A., Gilman, M., Orci, L., Cerasoli, F., Jr.,Rothman, J. E., and Clackson, T. (2000) Regulation of proteinsecretion through controlled aggregation in the endoplasmicreticulum. Science 287, 826–830

33. Hansen, J. L., Theilade, J., Haunso, S., and Sheikh, S. P. (2004)Oligomerization of wild type and nonfunctional mutant angio-tensin II type I receptors inhibits galphaq protein signaling butnot ERK activation. J. Biol. Chem. 279, 24108–24115

34. Kaykas, A., Yang-Snyder, J., Heroux, M., Shah, K. V., Bouvier, M.,and Moon, R. T. (2004) Mutant Frizzled 4 associated withvitreoretinopathy traps wild-type Frizzled in the endoplasmicreticulum by oligomerization. Nat. Cell Biol. 6, 52–58

35. Gronemeyer, T., Chidley, C., Juillerat, A., Heinis, C., andJohnsson, K. (2006) Directed evolution of O6-alkylguanine-DNA alkyltransferase for applications in protein labeling. ProteinEng. Des. Sel. 19, 309–316

36. Dorsch, S., Klotz, K. N., Engelhardt, S., Lohse, M. J., andBunemann, M. (2009) Analysis of receptor oligomerization byFRAP microscopy. Nat. Methods 6, 225–230

37. Hern, J. A., Baig, A. H., Mashanov, G. I., Birdsall, B., Corrie,J. E., Lazareno, S., Molloy, J. E., and Birdsall, N. J. (2010)Formation and dissociation of M1 muscarinic receptordimers seen by total internal reflection fluorescence imagingof single molecules. Proc. Natl. Acad. Sci. U. S. A. 107,2693–2698

38. Kasai, R. S., Suzuki, K. G., Prossnitz, E. R., Koyama-Honda, I.,Nakada, C., Fujiwara, T. K., and Kusumi, A. (2011) Full charac-terization of GPCR monomer-dimer dynamic equilibrium bysingle molecule imaging. J. Cell Biol. 192, 463–480

39. Kusumi, A., Nakada, C., Ritchie, K., Murase, K., Suzuki, K.,Murakoshi, H., Kasai, R. S., Kondo, J., and Fujiwara, T. (2005)Paradigm shift of the plasma membrane concept from thetwo-dimensional continuum fluid to the partitioned fluid: high-speed single-molecule tracking of membrane molecules. Annu.Rev. Biophys. Biomol. Struct. 34, 351–378

40. Kusumi, A., Shirai, Y. M., Koyama-Honda, I., Suzuki, K. G.,and Fujiwara, T. K. (2010) Hierarchical organization of theplasma membrane: investigations by single-molecule trackingvs. fluorescence correlation spectroscopy. FEBS Lett. 584,1814 –1823

Received for publication January 23, 2012.Accepted for publication May 1, 2012.

3439GABAB RECEPTOR TETRAMERS ARE STABLE COMPLEXES