Serotonin 5-HT2C Receptor Homodimer Biogenesis in the Endoplasmic Reticulum

Dimerization is a common property of G-protein-coupled receptors (GPCR). While the formation of GPCR dimers/oligomers has been reported to play important roles in regulating receptor expression, ligand binding, and second messenger activation, less is known about how and where GPCR dimerization occurs. The present study was performed to identify the precise cellular compartment in which class A GPCR dimer/oligomer biogenesis occurs. We addressed this issue using confocal microscopy and fluorescence resonance energy transfer (FRET) to monitor GPCR proximity within discrete intracellular compartments of intact living cells. Time-lapse confocal imaging was used to follow CFP- and YFP-tagged serotonin 5-HT2C receptors during biosynthesis in the endoplasmic reticulum (ER), trafficking through the Golgi apparatus and subsequent expression on the plasma membrane. Real-time monitoring of FRET between CFP- and YFP-tagged 5-HT2C receptors was performed by acceptor photobleaching within discrete regions of the ER, Golgi, and plasma membrane. The FRET signal was dependent on the ratio of CFP- to YFP-tagged 5-HT2C receptors expressed in each region and was independent of receptor expression level, as predicted for proteins in a non-random, clustered distribution. FRET efficiencies measured in the ER, Golgi, and plasma membrane were similar. These experiments provide direct evidence for homodimerization/oligomerization of class A GPCR in the ER and Golgi of intact living cells, and suggest that dimer/oligomer formation is a naturally occurring step in 5-HT2C receptor maturation and processing.

G-protein-coupled receptors (GPCR) 2 are the largest family of receptor proteins and are targets for ϳ50% of all currently marketed pharmaceuticals. They are present in most all cells in the human body and are signaling proteins that transmit infor-mation across the plasma membrane in response to neurotransmitters, hormones, odorants, taste, and light. GPCR were originally thought to function as monomeric units. However, over the last decade many studies have provided evidence suggesting that GPCR form dimeric/oligomeric complexes (reviewed in Refs. 1 and 2). Recent studies using atomic force microscopy and resonance energy transfer techniques provide strong support in favor of GPCR dimer/oligomer formation (3)(4)(5)(6)(7). While dimerization has been reported to regulate ligand binding, signal transduction, and receptor trafficking (reviewed in Refs. 8 and 9), fundamental questions concerning the biogenesis of GPCR dimers/oligomers remain unanswered. For class A GPCR it is unclear as to where in the cell GPCR dimerization occurs. Somatostatin and gonadotropin-releasing hormone receptors have been reported to undergo ligand-induced dimerization on the plasma membrane (10,11). However, other class A GPCR have been reported to be constitutively dimerized on the plasma membrane in the absence of ligand (5,(12)(13)(14)(15)(16)(17)(18), suggesting that dimerization may occur within intracellular compartments prior to trafficking to the plasma membrane.
There is strong evidence in favor of class C GPCR dimer/ oligomer formation in the ER. For example, heterodimerization of GABA B R1 and GABA B R2 is essential for receptor trafficking to the plasma membrane and GABA-mediated signaling (19 -23). For class A GPCR, co-expression with ␤ 2 -adrenergic receptors has been reported to be necessary for expression of ␣ 1Dadrenergic receptors on the plasma membrane (24). Additional evidence supporting intracellular dimerization of class A GPCR is provided by studies showing that non-trafficking, mutant receptors decrease the plasma membrane expression of their wild-type counterparts (25)(26)(27)(28)(29).
Dimerization has been proposed as a general mechanism necessary for proper trafficking of class A GPCR to the plasma membrane (8,9,24). Until recently, experimental evidence in favor of this model has been limited to studies using immunoprecipitation of solubilized receptors prepared from whole cell lysates and confocal microscopy showing co-localization of fluorescent-tagged GPCR in endomembranes of fixed cells. More recently, this issue has been addressed using bioluminescence resonance energy transfer (BRET) following sucrose density gradient centrifugation. Positive BRET signals were detected in both plasma membrane and endomembrane-enriched fractions prepared from HEK293 cells expressing vasopressin or ␤ 2 -adrenergic receptors (14,30). While these results suggest that GPCR dimerization may occur within intracellular compartments, direct experimental evidence demonstrating the formation of class A GPCR homodimers in the ER and Golgi of intact living cells is still lacking.
The present study was performed to test the hypothesis that class A GPCR form dimers/oligomers in the ER and Golgi of intact living cells. Previously, we have shown that 5-HT 2C receptors are present as constitutive homodimers on the plasma membrane (15). Herein, we used a combination of biochemical and biophysical techniques to determine if 5-HT 2C receptors form homodimers in the ER prior to trafficking to the Golgi and plasma membrane. Immunoprecipitation with differentially tagged 5-HT 2C receptors revealed the presence of 5-HT 2C receptor bands from ER and plasma membrane, and confocal fluorescence imaging was used to quantify fluorescence resonance energy transfer (FRET) in the ER and Golgi of living HEK293 cells expressing fluorescent-tagged 5-HT 2C receptors. The confocal microscopy-based FRET approach allows specific intracellular compartments to be examined, without compromising cellular integrity, and real-time monitoring of GPCR dimerization in the ER and Golgi apparatus. These experiments are the first to use time-lapse confocal imaging of transfected HEK293 cells to provide real-time visualization of fluorescent-tagged 5-HT 2C receptors during biosynthesis in the ER, trafficking through the Golgi, and subsequent expression on the plasma membrane at physiologically relevant receptor expression levels. Our results demonstrate that class A GPCR can form homodimeric/oligomeric complexes in the ER and Golgi of intact living cells, indicating that dimerization/oligomerization may be a naturally occurring step in the maturation and processing of 5-HT 2C receptors.

MATERIALS AND METHODS
Cell Culture and Transfection-HEK293 cells from the American Type Culture Collection (ATCC) were cultured in Dulbecco's modified Eagle's medium (Cellgro) with 10% fetal bovine serum at 37°C, 5% CO 2 . HEK293 cells were plated at 5 ϫ 10 5 cells per well in 6-well plates containing polylysine-treated glass coverslips. Transfections were performed using Lipofectamine reagent (Invitrogen) according to the manufacturer's protocol. Cells were transfected for 5 h and cultured post-transfection in serum-free Dulbecco's modified Eagle's medium. All experiments were performed in serum-free culture media 12-36 h post-transfection.
Fusion Proteins-All studies were performed using the INI isoform of the human 5-HT 2C receptor. Cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) were attached to the C-terminal end of the 5-HT 2C or CCR5-chemokine receptor by ligation into pECFP-N1 or pEYFP-N1 vectors (Clontech) to create 5-HT 2C /CFP, 5-HT 2C /YFP, and CCR5/ YFP as previously described (15). The 5-HT 2C /HA fusion protein was created by PCR to add the 10 amino acid HA sequence to the C-terminal end of the receptor, as previously described (15). All constructs were confirmed by DNA sequencing (Center for Functional Genomics, Albany, NY). YFP and HA attached to the C terminus of the receptor had no effect on ligand binding or inositol phosphate production (15).
Confocal Microscopy-HEK293 cells (5 ϫ 10 5 cells/well) were transfected with 40 ng of the indicated plasmid DNAs and imaged live (at room temperature in HEPES-buffered Minimal Essential Medium without phenol, Cellgro) using a Zeiss LSM 510Meta confocal imaging system with a 30 milliwatt argon laser and a ϫ63 1.4 NA oil immersion objective at ϫ2 zoom. The pEYFP-ER plasmid (Clontech) encoding a YFP fusion protein containing the ER-targeting sequence of calreticulin with a KDEL retention sequence and the pEYFP-Golgi plasmid (Clontech) encoding a YFP fusion protein that specifically targets the trans-medial region of the Golgi apparatus were used as markers to visualize ER and Golgi, respectively, following excitation at 514 nm. Real-time imaging of CFP-and YFP-tagged 5-HT 2C receptors was performed 12-24-h post-transfection to visualize 5-HT 2C receptors following biosynthesis in the ER, trafficking through the Golgi, and subsequent expression on the plasma membrane. Cells co-expressing 5-HT 2C /CFP and 5-HT 2C /YFP were imaged following excitation at 458 nm. CFP and YFP fluorescence were separated using on-line finger printing and linear unmixing with the Zeiss Meta detector and Zeiss Aim Software, as previously described (15).
Fluorescence Resonance Energy Transfer-HEK293 cells, plated at 5 ϫ 10 5 cells per well in 6-well plates containing polylysine-treated glass coverslips, were co-transfected with 5-HT 2C /CFP and 5-HT 2C /YFP plasmid DNA in the following ratios: 1:1 (40 ng ϩ 40 ng); and 1:2 (20 ng ϩ 40 ng) using Lipofectamine (Invitrogen). For control experiments, cells were transfected with 5-HT 2C /CFP (40 ng) ϩ ER/YFP marker (40 ng) or 5-HT 2C /CFP (40 ng) ϩ CCR5/YFP (40 ng). Transfected cells were imaged live (at room temperature in HEPES-buffered Minimal Essential Medium without phenol) using a Zeiss LSM 510Meta confocal imaging system with a 30 milliwatt argon laser and a ϫ63 1.4 NA oil immersion objective at ϫ2 zoom. CFP and YFP emission spectra were collected (from cells expressing 5-HT 2C /CFP or 5-HT 2C /YFP alone) following excitation at 458 nm and were used as reference spectra for linear unmixing of CFP and YFP emission spectra in co-transfected cells using the Zeiss META detector and Zeiss AIM software, as previously described (15). FRET was measured by acceptor photobleaching (31), with the following modifications. Confocal microscopy was used to visualize a 2-m optical slice through the middle of live HEK293 cells expressing both 5-HT 2C /CFP and 5-HT 2C /YFP receptors in the ER, Golgi, and plasma membrane. Pre-bleach CFP and YFP images were collected simultaneously following excitation at 458 nm (38% laser intensity; detector gain ϭ 740). A selected region of interest was irradiated with the 514-nm laser line (100% intensity, 60 iterations, using a 458-nm/514-nm dual dichroic mirror) for 5-10 s to photobleach YFP. Post-bleach CFP and YFP images were collected simultaneously (at 458 nm) immediately following photobleaching. FRET was measured as an increase in CFP fluorescence intensity following YFP photobleaching. FRET efficiency was calculated as 100 ϫ [(CFP post-bleach Ϫ CFP prebleach)/CFP post-bleach] using the FRET Macro in the Zeiss Aim software package, taking into account CFP and YFP background noise in each channel. FRET efficiencies were calculated in cells co-expressing 5-HT 2C /CFP and 5-HT 2C /YFP, in cells expressing 5-HT 2C /CFP alone, and in cells co-expressing 5-HT 2C /CFP with ER/YFP or CCR5/YFP. For live cells expressing only 5-HT 2C /CFP the mean FRET efficiency was Ϫ1.3 Ϯ 0.7%, indicating that receptor migration within the region of interest during the 5-10-s photobleach period does not give rise to a false positive FRET signal.
Radioligand Binding Assay-HEK293 cells were plated and transfected as described above for the FRET experiments. Membranes were prepared from transfected cells and [ 3 H]mesulergine binding was performed as previously described (32). Protein was measured by BCA (Pierce). Data were analyzed using GraphPad Prism software.
Immunoprecipitation and PNGase Digest-HEK293 cells (4 ϫ 10 6 cells/100-mm dish) were pretreated with or without 1 M Brefeldin A (Sigma) for overnight prior to transfection with 1 g of 5-HT 2C /YFP and 1 g of 5-HT 2C /HA plasmid DNA using 20 l of Lipofectamine reagent (Invitrogen). As a control, separate dishes of cells were transfected independently with either 5-HT 2C /YFP or 5-HT 2C /HA, and the cells were mixed together after transfection. Transfected cells were washed with phosphate-buffered saline and scraped in 2 ml of phosphatebuffered saline and centrifuged at 1500 ϫ g for 5 min. Cells were resuspended in 0.5 ml of lysis buffer containing 50 mM Tris-HCl, pH 7.6, 1 mM EDTA, 10 l of protease inhibitor mixture (Sigma), sonicated for 30 s on ice, and centrifuged at 1500 ϫ g for 5 min. The supernatant was centrifuged at 21,000 ϫ g for 30 min at 4°C. The membrane pellet was resuspended in 0.2 ml of solubilization buffer (50 mM Tris-HCl, pH 7.6, 1 mM EDTA, 150 mM NaCl, 10 mM iodoacetamide, 5 l of protease inhibitor mixture, 10 mM CHAPS). The membrane proteins were solubilized for 60 min on ice and centrifuged at 21,000 ϫ g for 30 min at 4°C to pellet insoluble material. Solubilized membrane proteins were immunoprecipitated overnight with 10 l of HA(Y-11)-agarose (Santa Cruz Biotechnology) at 4°C. Samples were centrifuged at 4,000 ϫ g for 5 min, the pellet was washed twice with phosphate-buffered saline containing protease inhibitor mixture, resuspended in 50 l of denaturing Laemmli sample buffer. For the PNGase digest, solubilized membrane protein was diluted in 50 mM NaPO 4 (pH 7.5) and incubated with 2 l of PNGase (NewEngland Biolabs) for 3 h at 37°C. Samples were run on a Tris-HCl ready gel (Bio-Rad) at 100 V for 70 min under denaturing conditions. The gel was transferred to nitrocellulose (Bio-Rad), probed with GFP(B-2)-horseradish peroxidase antibody (Santa Cruz Biotechnology) diluted 1:3,000 and visualized by enhanced chemiluminescence (Amersham Biosciences).

Confocal Fluorescence Imaging of ER and Golgi Membranes
in Living Cells-HEK293 cells were transfected with plasmid DNA containing a mutant YFP protein engineered with an ERtargeting and retention sequence (ER-YFP) or with a YFP variant that specifically targets the trans-medial region of the Golgi apparatus (Golgi-YFP). Live cell confocal fluorescence imaging was performed to visualize ER and Golgi membranes in 2-m thick optical slices through the middle of HEK293 cells expressing the ER-YFP or Golgi-YFP marker. In HEK293 cells, ER membranes appear as a diffuse reticular network spreading outward from the nucleus throughout the cytosol toward the plasma membrane (Fig. 1A). In contrast, Golgi membranes are visualized as dense areas of intense fluorescence confined to a discrete perinuclear region of the cytosol (Fig. 1B).
Real-time Confocal Fluorescence Imaging of 5-HT 2C Receptors Trafficking from the ER to the Golgi and Plasma Membrane-Fluorescent proteins were used as markers to visualize 5-HT 2C receptors following biosynthesis in the ER and subsequent trafficking to the Golgi and plasma membrane. HEK293 cells cotransfected with 5-HT 2C /CFP and the ER/YFP marker showed co-localization of CFP and YFP fluorescence in the ER 12-h post-transfection (Fig. 1, C-E). Cells co-transfected with 5-HT 2C /CFP and the Golgi/YFP marker showed co-localization of CFP and YFP fluorescence in the Golgi 16-h post-transfection (Fig. 1, F-H). Twenty hours post-transfection, 5-HT 2C / CFP fluorescence was on the plasma membrane, and the Golgi/ YFP marker remained in the Golgi (Fig. 1, I-K). Based on these results, a time course was performed in cells co-transfected with 5-HT 2C /CFP and 5-HT 2C /YFP. The cells were imaged live using confocal fluorescence imaging at various time points post-transfection. Ten hours after the addition of transfection reagents, 5-HT 2C /CFP and 5-HT 2C /YFP receptor fluorescence began to emerge. 12-16-h post-transfection, confocal fluorescence imaging revealed many cells with a diffuse reticular pattern of fluorescence ( Fig. 2A), identical to HEK293 cells expressing the ER-YFP marker (Fig. 1A), demonstrating 5-HT 2C receptor expression in the ER. Confocal fluorescence imaging 16 -20-h post-transfection revealed many cells with a more clustered, dense pattern of labeling with a distinct perinuclear distribution (Fig. 2B), identical to HEK293 cells expressing the Golgi-YFP marker (Fig. 1B). By 20-h post-transfection, many cells were expressing 5-HT 2C receptors on the plasma membrane ( Fig. 2C). At this time, 5-HT 2C receptor expression appears to have reached a steady-state level with the majority of the receptor expressed on the plasma membrane and little intracellular fluorescence remaining. This phenomenon has been observed in every one of our transient transfections and in all of our stable cell lines. Whether this is because of mRNA stability or some other form of feedback regulation by the receptors once they reach the plasma membrane is unknown. While the time course for receptor expression is likely to vary for different receptors in different cell lines and different culture conditions, our results indicate that it is possible to monitor receptor trafficking through discrete intracellular compartments at early time points post-transfection.

5-HT 2C Receptor Homodimer/Oligomer Formation in the ER of Living
Cells-A confocal microscopy-based FRET method was used to monitor 5-HT 2C homodimers in the ER of living cells 12-16 h post-transfection. HEK293 cells co-transfected with 5-HT 2C /CFP and 5-HT 2C /YFP were visualized live by confocal fluorescence imaging using a Zeiss LSM-510 Meta laser scanning confocal microscope. CFP and YFP fluorescence were imaged simultaneously following excitation at 458 nm and subsequent linear unmixing of emission spectra using the Zeiss Meta detector (Fig. 3, A and B). CFP-and YFP-tagged 5-HT 2C receptors were co-localized in the ER of a living HEK293 cell (Fig. 3C). Acceptor photobleaching was used to measure FRET between 5-HT 2C /CFP and 5-HT 2C /YFP in the ER. This method involves the selective irradiation of YFP fluorescence using the 514-nm laser line. If 5-HT 2C /CFP (donor) and 5-HT 2C /YFP (acceptor) are within 10 nm of each other and the fluorophore dipoles are aligned, resonance energy can be transferred from CFP to YFP. If FRET occurs, then photobleaching of YFP fluorescence will result in enhanced CFP fluorescence, because of the dequenching of CFP following the removal of YFP. To perform acceptor photobleaching, a pre-bleach image was captured using the 458-nm laser line (Fig. 3, A-C). A region of the ER (marked by the white rectangle in Fig. 3E) was selectively irradiated using the 514-nm laser line. A 458/514-nm dual dichroic mirror was used to allow rapid, automated image acquisition immediately after YFP photobleaching (Fig. 3, D-F). An increase in CFP fluorescence was observed following YFP photobleaching (indicated by the arrow in Fig. 3D) and the FRET efficiency was 28.5%. Acceptor photobleaching experiments, as shown in Fig. 3, were performed on 50 living HEK293 cells co-expressing 5-HT 2C /CFP and 5-HT 2C /YFP. All FRET measurements were made 12-16 h after transfection, when 5-HT 2C receptor expression in the ER was visible. FRET efficiency was plotted versus the uD/A ratio (post-bleach CFP fluorescence/ pre-bleach YFP fluorescence). The amount of FRET observed within a given cell was dependent on the uD/A ratio (Fig. 4A). When FRET efficiency was plotted versus acceptor fluores-cence by uD/A ratio, there was no correlation between the amount of FRET observed and 5-HT 2C receptor expression level in the ER (Fig. 4B).

Real-time Visualization of 5-HT 2C Receptor Homodimers/ Oligomers Trafficking through the Golgi Apparatus of Living
Cells-Confocal imaging and acceptor photobleaching were used to visualize and quantify FRET between CFP-and YFPtagged 5-HT 2C receptors in the Golgi apparatus of living cells. HEK293 cells were co-transfected with 5-HT 2C /CFP and 5-HT 2C /YFP and imaged 16 -20 h following transfection. Differentially tagged 5-HT 2C receptors were clearly visible and colocalized within the Golgi apparatus (Fig. 5, A-C). A region of the Golgi (marked by the white rectangle in Fig. 5E) was selectively irradiated using the 514-nm laser line and a post-bleach image was captured (Fig. 5, D-E). An increase in CFP fluorescence was observed following YFP photobleaching (indicated by the arrow in Fig. 5D) and the FRET efficiency was 40.6%. Acceptor photobleaching experiments, as shown in Fig. 5, were performed on 50 living HEK293 cells co-expressing 5-HT 2C / CFP and 5-HT 2C /YFP. All FRET measurements were made 16 -20 h after transfection, when 5-HT 2C receptor expression in the Golgi was visible. FRET efficiency was plotted versus the uD/A ratio. The amount of FRET observed within a given cell was dependent on the uD/A ratio (Fig. 6A). FRET efficiency plotted versus acceptor fluorescence by uD/A ratio revealed no correlation between FRET efficiency and 5-HT 2C receptor expression (Fig. 6B). Fig. 3, the FRET efficiency for 5-HT 2C receptors expressed in the ER was 28.5%, while in Fig. 5 the reported FRET efficiency for 5-HT 2C receptors expressed in the Golgi was 40.6%. Comparing the FRET efficiencies from these two experiments would lead to the erroneous conclusion that FRET efficiency increases as 5-HT 2C receptors traffic from the ER to the Golgi. However, as shown in Figs. 4A and 6A FRET efficiency is dependent on the ratio of donor to acceptor expressed in a given cell. Therefore, a meaningful comparison of FRET efficiencies can only be made between cells with similar uD/A ratios. When the FRET data presented in Figs. 4A and 6A were divided into five separate groups, based on their uD/A ratios, the mean FRET efficiencies for each group were different (Table 2), again demonstrating the dependence of FRET efficiency on the uD/A ratio. However, the mean FRET efficiencies for ER and Golgi were not significantly different when compared within a given uD/A    SEPTEMBER 15, 2006 • VOLUME 281 • NUMBER 37 range, except for the 1.3-2.0 range where FRET efficiencies were less than 10%. FRET was measured on the plasma membrane of HEK293 cells co-expressing 5-HT 2C /CFP and 5-HT 2C /YFP, as described above for ER and Golgi ( Table 3). The overall mean uD/A ratios and FRET efficiencies were similar for CFP-and YFP-tagged 5-HT 2C receptors expressed in the ER, Golgi, and on the plasma membrane (Table 3).

5-HT 2C Receptor Expression
Level-Radioligand binding studies were performed in parallel with FRET experiments to monitor 5-HT 2C receptor expression levels in transfected cells used in the FRET assay. [ 3 H]Mesulergine binding yielded 5-HT 2C receptor expression levels of 6.9 Ϯ 0.6 pm/mg protein from three independent transfection experiments performed in parallel with transfections for the FRET experiments. The 5-HT 2C receptor expression levels in the transfected cells used for the FRET experiments are similar to endogenous 5-HT 2C receptor expression (10 pm/mg) in choroid plexus epithelial cells (33).
Immunoprecipitation and PNGase Digest-Immunoprecipitation of differentially tagged 5-HT 2C receptors was performed to confirm the FRET results by demonstrating a physical associa-tion between 5-HT 2C receptors. Solubilized membrane proteins from HEK293 cells separately transfected or co-transfected with 5-HT 2C /HA and 5-HT 2C /YFP were immunoprecipitated with HA-agarose. Following immunoprecipitation, the samples were denatured, separated on a 10% polyacrylamide gel, and immunoblotted with GFP-HRP antibody (Fig. 7A). Denaturation of the immunoprecipitate resulted in the appearance of GFP immunoreactive bands the predicted size of mature (fully glycosylated) 5-HT 2C /YFP (90 kDa) and immature 5-HT 2C / YFP (70 kDa). A PNGase digest was performed to confirm the identity of the 90-and 70-kDa bands as mature and immature forms of the 5-HT 2C receptor (Fig. 7B). The sample was digested with PNGase and run on a 7.5% acrylamide gel for better separation of the different glycoslyation states of the 5-HT 2C receptor. Digest with PNGase resulted in a single GFP immunoreactive band the predicted size of unglycosylated 5-HT 2C /YFP. When the cells were pretreated with Brefeldin A (to prevent receptor trafficking to the plasma membrane) and then immunoprecipitated, a GFP immunoreactive band the predicted size of unglycosylated/immature 5-HT 2C /YFP (70 kDa) appeared only in samples from cells co-transfected with 5-HT 2C /HA and 5-HT 2C /YFP (Fig. 7C).

DISCUSSION
Although GPCR are generally conceptualized as forming dimeric/oligomeric complexes, very little is understood about the molecular interactions between dimers/oligomers, the cellular compartment in which dimerization occurs, and how receptor dimerization regulates GPCR function. The present study was performed to identify the cellular compartment in which class A GPCR dimerization occurs. Our approach to understanding the biogenesis of GPCR dimers is unique in that we used a confocal microscopy-based FRET technique that allows the determination of protein:protein proximity within

TABLE 2 FRET efficiency by uD/A ratio for 5-HT 2C receptors expressed in the ER and Golgi
FRET efficiencies measured in the ER and Golgi of living HEK293 cells co-expressing 5-HT 2C /CFP (donor) and 5-HT 2C /YFP (acceptor) were divided into five groups based on the ratio of donor to acceptor (uD/A). Data represent the mean Ϯ S.E. for the number of cells indicated (n).

5-HT 2C Homodimers in the ER
specific intracellular compartments, such as ER and Golgi, of living cells. Time lapse confocal imaging of transfected HEK293 cells provided real-time visualization of fluorescent-tagged 5-HT 2C receptors during biosynthesis in the ER, trafficking through the Golgi, and subsequent expression on the plasma membrane.
FRET combined with confocal microscopy was used to determine the proximity of CFP-and YFP-tagged 5-HT 2C receptors within the ER and Golgi, and on the plasma membrane of living cells. Resonance energy can be transferred from CFP to YFP if they are within 10 nm of each other and their dipoles are appropriately aligned. Energy transfer from CFP (donor) to YFP (acceptor) results in the quenching of CFP fluorescence, an increase in YFP fluorescence, and can be measured as an increase in CFP fluorescence following selective irradiation of the YFP fluorophore. This method of measuring FRET is called acceptor photobleaching (31). Laser scanning confocal microscopy allows the photobleaching to be confined to very discrete intracellular regions. This minimizes the time required for irradiation of the acceptor fluorophore and makes the technique suitable for live cell imaging. A specific FRET signal resulting from proteins in a clustered distribution, such as dimers/oligomers, has been shown to be dependent on the ratio of donor to acceptor expressed in the cell, while FRET resulting from overexpression of randomly distributed proteins in close proximity to one another can be distinguished by a dependence on acceptor fluorescence or total protein expression level (34,35). In the present study, real-time FRET efficiencies measured for 5-HT 2C /CFP and 5-HT 2C /YFP fusion proteins in the ER and Golgi apparatus of living HEK293 cells were dependent on the ratio of donor to acceptor (uD/A) expressed within a given cell and independent of receptor expression level (measured as total acceptor fluorescence), as predicted for proteins in a non-random, clustered distribution (34,35). When FRET efficiencies were divided into five groups based on their uD/A ratios, the mean FRET efficiencies increased as uD/A decreased and were similar for ER and Golgi. In addition, the overall mean uD/A ratios and FRET efficiencies were similar for ER, Golgi, and plasma membrane. The similar mean FRET effi-ciencies determined for 5-HT 2C /CFP and 5-HT 2C /YFP fusion proteins in the ER, Golgi, and plasma membrane suggest that 5-HT 2C receptor homodimers/oligomers form in the ER and then traffic through the secretory pathway to the plasma membrane. The FRET results were confirmed by immunoprecipitation of both immature and mature glycosylated forms of the 5-HT 2C receptor.
Previous studies demonstrating the presence of constitutive dimers/oligomers on the plasma membrane (5,(12)(13)(14)(15)(16)(17)(18), studies using mutant receptors that retain their wild-type counterparts within intracellular compartments (25)(26)(27)(28)(29), and BRET studies using sub-cellular membrane fractions (14,30) suggest that GPCR dimers/oligomers may form prior to trafficking to the plasma membrane. The present study provides direct visualization of class A GPCR dimer/oligomer biogenesis occurring within discrete intracellular compartments (ER and Golgi) of living cells, and subsequent trafficking through the secretory pathway to the plasma membrane. This may be a naturally occurring step in GPCR biosynthesis, it may be required for trafficking out of the ER, and/or it may be essential for forming a functional signaling unit. At the present time the functional significance of this phenomenon remains unknown.
The obligatory dimerization of GABA B receptors critical for expression of a functional signaling unit, the intracellular retention of wild-type GPCR by their mutant counterparts, and the evidence provided by the current live-cell study indicate that dimerization may be a prerequisite for normal receptor trafficking and expression on the plasma membrane. Dimer/ oligomer formation following receptor biosynthesis may be necessary for passing ER quality control checkpoints that determine functionality. Proteins can be retained in the ER if they are not folded properly or assembled with other proteins critical for the formation of a functional complex. In the case of GABA B receptors, dimerization is essential for masking an ER retention motif located in the C-terminal region of the GABA B R1 receptor, allowing export of a functional receptor complex from the ER (22). It is also possible that dimerization in the ER may be a prerequisite for trafficking to the plasma membrane as dimers may represent the basic metabotropic signaling unit.
Studies involving the rhodopsin receptor support the hypothesis that the dimer may represent the basic signaling unit. Atomic force microscopy has been used to visualize rhodopsin receptors in native mammalian membranes as rows of dimeric complexes (6,7). Also, the distance between the ␣and ␥-subunits of a single heterotrimeric G-protein, which are the reported regions of contact with GPCR, is predicted to be too large to accommodate a single rhodopsin receptor (36,37). Studies using chemical cross-linking and purified leukotriene B4 receptors (LTB4) have demonstrated that an LTB4 homodimer forms a pentameric complex with a single heterotrimeric G-protein (38). We have recently reported that the predicted ligand:dimer:G-protein stoichiometry is 2:1:1 for 5-HT 2C receptors expressed in HEK293 cells (39). In addition, the D1-D2 dopamine receptor heterodimer has been reported to form a novel signaling complex in which ligand binding to both protomers results in G␣ q activation, but blockade of either protomer alone is sufficient to block signaling (40). The results of these experiments are consistent with a model in which class A GPCR dimers interact with a single G-protein and suggest that the dimer may represent the basic signaling unit. Elegant studies involving trafficking and non-trafficking wild-type and mutant class C metabotropic glutamate receptors have demonstrated that ligand binding to one protomer of the dimer can result in G-protein activation (41). Recent studies involving the D 2 -dopamine receptor have established that conformational changes occur at the dimer interface during receptor activation and inactivation (42), indicating that conformational changes at the dimer interface may provide a means for transactivation of G-proteins and may explain how heterodimerization between two different GPCR can produce a unique pharmacological profile.
In conclusion, our results indicate that class A GPCR dimer/ oligomer biogenesis occurs at an early time point during receptor biosynthesis and processing in the ER and Golgi. Live cell confocal microscopy combined with FRET provided real-time visualization of dimer/oligomer formation and trafficking through the secretory pathway to the plasma membrane, suggesting that homodimerization/oligomerization is a naturally occurring step in 5-HT 2C receptor maturation and processing. Future studies aimed at identifying the molecular interface between protomers of 5-HT 2C receptor homodimers, and subsequent disruption of dimer formation, will provide a model system which can be used for determining the importance of homodimerization in GPCR trafficking and signaling.