Real Time Visualization of Agonist-mediated Redistribution and Internalization of a Green Fluorescent Protein-tagged Form of the Thyrotropin-releasing Hormone Receptor*

The long isoform of the rat thyrotropin-releasing hormone receptor (TRHR) was modified by the addition of a vesicular stomatitis virus (VSV) epitope tag and green fluorescent protein (GFP). VSV-TRHR-GFP bound TRH with affinity similar to that of the unmodified receptor and stimulated [3H]inositol phosphate production. A clone stably expressing VSV-TRHR-GFP at some 120,000 copies/cell was selected to visualize this receptor during cellular exposure to TRH. Internalization was detected within 3–5 min after treatment with 1 × 10−7 m TRH, with dramatic reductions in plasma membrane localization achieved within 10–15 min. The TRHR antagonist/inverse agonist chlordiazepoxide competitively inhibited internalization. Hyperosmotic sucrose inhibited internalization of VSV-TRHR-GFP, measured both by intact cell [3H]TRH binding studies and by confocal microscopy. Now TRH caused a redistribution of VSV-TRHR-GFP to highly punctate but plasma membrane-delineated foci. Pretreatment with the microtubule-disrupting agent nocodazole allowed internalization of the VSV-TRHR-GFP construct but only into vesicles that remained in close apposition to the plasma membrane. Covisualization of VSV-TRHR-GFP and Texas Red transferrin initially indicated entirely separate localizations. After exposure to TRH substantial amounts of VSV-TRHR-GFP were present in vesicles overlapping those containing Texas Red transferrin. Such results demonstrate the G protein-coupling capacity and provide real time visualization of the processes of internalization of a TRH-receptor-GFP construct in response to agonist.

The long isoform of the rat thyrotropin-releasing hormone receptor (TRHR) was modified by the addition of a vesicular stomatitis virus (VSV) epitope tag and green fluorescent protein (GFP). VSV-TRHR-GFP bound TRH with affinity similar to that of the unmodified receptor and stimulated [ 3 H]inositol phosphate production. A clone stably expressing VSV-TRHR-GFP at some 120,000 copies/cell was selected to visualize this receptor during cellular exposure to TRH. Internalization was detected within 3-5 min after treatment with 1 ؋ 10 ؊7 M TRH, with dramatic reductions in plasma membrane localization achieved within 10 -15 min. The TRHR antagonist/ inverse agonist chlordiazepoxide competitively inhibited internalization. Hyperosmotic sucrose inhibited internalization of VSV-TRHR-GFP, measured both by intact cell [ 3 H]TRH binding studies and by confocal microscopy. Now TRH caused a redistribution of VSV-TRHR-GFP to highly punctate but plasma membranedelineated foci. Pretreatment with the microtubuledisrupting agent nocodazole allowed internalization of the VSV-TRHR-GFP construct but only into vesicles that remained in close apposition to the plasma membrane. Covisualization of VSV-TRHR-GFP and Texas Red transferrin initially indicated entirely separate localizations. After exposure to TRH substantial amounts of VSV-TRHR-GFP were present in vesicles overlapping those containing Texas Red transferrin. Such results demonstrate the G protein-coupling capacity and provide real time visualization of the processes of internalization of a TRH-receptor-GFP construct in response to agonist.
Thyrotropin-releasing hormone (TRH) 1 is a hypothalamic tripeptide intimately involved in controlling the production of thyrotropin and prolactin from the anterior pituitary (1,2). TRH functions via binding to a seven-transmembrane element-G protein-coupled receptor (3), which, by interacting selectivity with G q and G 11 , causes activation of phospholipase C␤1 and the hydrolysis of phosphatidylinositol 4,5-bisphosphate (4 -7). As with other G protein-coupled receptors there has been great interest in the mechanisms of regulation of the TRH receptor (TRHR) (6 -11). One focus of studies on G protein-coupled receptors relates to processes contributing to desensitization, the phenomenon by which sustained exposure of a receptor to an agonist ligand results in a waning of cellular response to the ligand (12). Elements contributory to this phenomenon include receptor phosphorylation, which can either directly or indirectly result in a diminution of receptor-G protein interactions, internalization of the receptor and/or G protein, and alterations in cellular levels of these proteins (12).
During studies of the TRHR we have noted that both the rat TRHR and the G proteins G q ␣ and G 11 ␣ are internalized in response to agonist (13,14), although internalization of the G proteins proceeds substantially more slowly than the receptor (13,14). The G proteins are subsequently down-regulated (7).
The capacity of ligands to internalize G protein-coupled receptors has been studied using a wide range of approaches (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27), including the use of fluorescent ligands, antireceptor antibodies, and antibodies to epitope tags, which have been introduced to alter the sequence of a cDNA or DNA encoding the receptor. These approaches have been instrumental in providing key information in relation to receptor regulation. However, for many of these, analyses cannot be performed in real time and on intact cells because it is often necessary to fix the cell before visualization.
Recently, the green fluorescent protein (GFP) derived from Aequorea victoria has become a powerful adjunct to cell biological research as a molecular marker for gene expression (28 -30). Very recently it has begun to be applied for analysis of the subcellular distribution and regulation of G protein-coupled receptors (14,(31)(32)(33).
In the present study we utilize the stable expression of a chimeric protein in which a modified form of GFP (30) was attached to the COOH-terminal tail of the long isoform of the rat TRH receptor to analyze the processes contributing to agonist-mediated internalization and trafficking of this receptor. Construction of the VSV-TRHR-GFP Fusion Expression Construct-Production and subcloning of the vesicular stomatitis virus (VSV)tagged TRHR-GFP fusion protein were done in two separate steps. In the first step the coding sequence of the long isoform of the rat TRH receptor (3) was modified by polymerase chain reaction amplification. Using the amino-terminal primer 5Ј-AAAGCTAGCGCCACCATGTA-CACCGATATAGAGATGAACAGGCTGGGAAAGGAGAATGAAACC-GTCAGTGAACTGAAC-3Ј, an NheI restriction site and the VSV epitope (YTDIEMNRLGK) were introduced adjacent to the sequence of codon 2 of the TRHR. Using the COOH-terminal primer 5Ј-GCTATCTAGAGT-CAAAGCTTCTCCTGTTTGGCAGTCAAA-3Ј, an XbaI restriction site following the stop codon and a HindIII restriction site just in the front of stop codon were introduced. The HindIII restriction site changed the last four nucleotides of the TRHR coding sequence from 5Ј-AATA-3Ј to 5Ј-GCTT-3Ј and thus altered the last amino acid from Ile to Leu. The amplified fragment of VSV-TRHR digested with NheI and XbaI was ligated to pcDNA3.1(ϩ) expression vector (Invitrogen) digested with NheI and XbaI. The VSV-TRHR functionality was characterized by binding experiments and agonist-mediated inositol phosphate production (see "Results"). To obtain the VSV-TRHR-GFP fusion protein the coding sequence of a modified form of GFP (30) was amplified by polymerase chain reaction. Using the amino-terminal primer 5Ј-GAGA-AGCTTGGAGCTATGAGTAAAGGAGAAGAACTTTTCACT-3Ј, a Hin-dIII restriction site and a 2-amino acid spacer (Gly-Ala) were introduced in front of the initiator Met of GFP. Using the COOH-terminal primer 5Ј-TGCTCTAGATTATTTGTATAGTTCATCCATGCCATG-3Ј, an XbaI restriction site was introduced behind the stop codon of GFP. Finally, the VSV-TRHR construct in pcDNA3.1(ϩ) was digested with HindIII (to remove the stop codon from the VSV-TRHR coding sequence), and XbaI and was ligated together with the polymerase chain reaction product of GFP amplification which was digested with HindIII and XbaI. The open reading frame so produced represents the coding sequence of VSV-TRHR-GFP. This was sequenced fully before its expression and analysis.

Materials-All
Transient and Stable Transfection of HEK293 Cells-HEK293 cells were maintained in minimal essential medium (Sigma) supplemented with 0.292 g/liter L-glutamine and 10% newborn calf serum at 37°C. Cells were grown to 60 -80% confluence before transient transfection. Transfection was performed using LipofectAMINE reagent (Life Technology, Inc.) according to the manufacturer's instructions. To generate cell lines stably expressing VSV-TRHR-GFP 2 days after transfection cells were seeded/diluted and maintained in minimal essential medium supplemented with 1 mg/ml Geneticin (Life Technology, Inc.). Medium was replaced every 3 days with minimal essential medium containing 1 mg/ml Geneticin. Clonal expression was examined initially by fluorescence microscopy, and clones for further study were selected and expanded.
Inositol Phosphate Assays-Transiently transfected HEK293 cells (24 h after transfection) or cell lines stably expressing either the TRHR or VSV-TRHR-GFP were reseeded into 12-well plates and incubated for a further 24 h. They were labeled with [ 3 H]inositol (1 Ci/ml) in inositol-free Dulbecco's modified Eagle's medium supplemented with 2% dialyzed newborn calf serum and 1% glutamate for 24 h. On the day of the experiment cells were washed twice with Krebs-Ringer-Hepes-LiCl buffer (KRH/LiCl) (115 mM NaCl, 5 mM KCl, 15 mM LiCl, 1.2 mM MgSO 4 , 1.2 mM CaCl 2 , 20 mM Hepes, 1.2 mM Na 2 PO 4 , 10 mM glucose, 0.1% bovine serum albumin, pH 7.4), incubated for 10 min with KRH/ LiCl, and stimulation by varying concentrations of TRH was performed in the same buffer for 10 min. All manipulations were done at 37°C. Reactions were stopped by aspiration of the KRH/LiCl/TRH buffer, and the cells were lysed using 0.75 ml of 20 mM formic acid on ice (30 min). Supernatant fractions were centrifuged (14,000 ϫ g for 3 min). The following steps were carried out according to Ref. 34 Binding Experiments-For binding studies on intact cells, they were harvested using 0.5 mM EDTA in PBS, washed, and used directly for experiments. Membranes were prepared as described previously (6,13). Both types of experiment were performed with 10 nM [ 3 H]TRH and the addition of different concentrations of TRH at 4°C for 60 min. Free ligand was separated by vacuum filtration through GF/C filters followed by three washes (each 5 ml) with ice-cold buffer A (2 mM MgCl 2 , 40 mM Tris, pH 7.4). In the case of intact cell binding, cells were incubated in Krebs-Ringer-Hepes buffer (KRH) (130 mM NaCl, 5 mM KCl, 1.2 mM MgSO 4 , 1.2 mM CaCl 2 , 20 mM Hepes, 1.2 mM Na 2 PO 4 , 10 mM glucose, 0.1% bovine serum albumin, pH 7.4). Membrane binding was performed with 40 g of membrane protein in buffer A with the addition of 100 M p[NH]ppG to convert the TRHR into a single class of low affinity sites.
Confocal Laser Scanning Microscopy-Cells were observed using a laser scanning confocal microscope (Zeiss Axiovert 100) using a Zeiss Plan-Apo 63 ϫ 1.40 NA oil immersion objective, pinhole of 35, and electronic zoom 1 or 3. The GFP was exited using a 488 nm argon/ krypton laser and detected with 515-540 nm band pass filter. The Texas Red-modified transferrin was exited at 543 nm and detected with a long pass band filter 570 nm.
The images were manipulated with Zeiss LSM or MetaMorph software. Two different protocols for preparation of cells were used. When examining the time course of internalization, short time exposures to TRH, and the Texas Red transferrin colocalization studies, live cells were used. Cells were grown on glass coverslips and mounted on the imaging chamber. Cells were maintained in KRH buffer, and temperature was maintained at 37°C. Cell labeling by Texas Red transferrin was performed by incubation for 10 min in KRH buffer with 10 g/ml Texas Red transferrin, and after washing with KRH (three times) the cells were used for analysis. In other studies fixed cells were used. Cells on glass coverslips were washed with PBS and fixed for 20 min at room temperature using 4% paraformaldehyde in PBS and 5% sucrose, pH 7.2. After one wash with PBS coverslips were mounted on microscope slides with 40% glycerol in PBS.
Cytoskeletal Drugs-Cytochalasin B and nocodazole were dissolved as 500ϫ concentrated stock solutions in dimethyl sulfoxide and finally used in concentrations of 4 g/ml and 10 Ϫ5 M, respectively. Cytochalasin B was added 30 min and nocodazole 60 min before TRH treatment.
Measurement of TRHR Internalization-Cells were seeded to 12-well plates, and the amount of [ 3 H]TRH internalized was measured. On the day of the experiment, the medium was changed to serum-free minimal essential medium, and cells were equilibrated for 1 h in 5% CO 2 at 37°C. In some cases cells were treated with sucrose or with cytoskeletal modifying drugs before the addition of 10 nM [ 3 H]TRH. After incubation, plates were placed on ice, washed three times with 0.15 M NaCl, one time with 0.5 M NaCl and 0.2 M acetic acid, and finally one time with 0.15 M NaCl to remove excess and plasma membrane-bound [ 3 H]TRH. Cells were lysed with 1.5% SDS and 1.5% Triton X-100 and internalized radioactivity measured in this fraction. To estimate nonspecific binding, 10 Ϫ5 M TRH was added.

RESULTS
A cDNA encoding the long isoform of the rat TRHR was modified such that an 11-amino acid (YTDIEMNRLGK) VSV epitope tag was added to the NH 2 terminus of the encoded protein (VSV-TRHR). This construct was modified further by a polymerase chain reaction-based strategy such that a cDNA encoding a highly fluorescent and thermostabilized form of the A. victoria GFP (30) was linked in-frame with the VSV-TRHR cDNA to generate a VSV-TRHR-GFP cDNA. For convenience of construction this resulted in the alteration of the final amino acid of the TRHR from Ile to Leu and the incorporation of a 2-amino acid (Gly-Ala) linker between the two proteins now contained within the chimeric construct (Fig. 1).
Transient expression of the VSV-TRHR-GFP cDNA in HEK293 cells followed by fixation and imaging in a confocal microscope demonstrated expression of the construct. The expressed protein had a predominantly plasma membrane localization (14). To characterize the integrity and functionality of VSV-TRHR-GFP it was expressed transiently in HEK293 cells in parallel with both the unmodified TRHR and VSV-TRHR. The cells were labeled subsequently with [ 3 H]inositol (1 Ci/ ml, 24 h), and the capacity of varying concentrations of TRH to stimulate the generation of [ 3 H]inositol phosphates was measured. All three TRHR constructs allowed robust stimulation of [ 3 H]inositol phosphate production ( Fig. 2 with TRH displaying slightly greater potency at VSV-TRHR-GFP (EC 50 ϭ 3.2 ϫ 10 Ϫ9 M) compared with either VSV-TRHR (EC 50 ϭ 6.2 ϫ 10 Ϫ9

FIG. 1. Construction of a VSV-tagged rat TRHR-GFP chimera.
VSV-TRHR-GFP was constructed from the long isoform of the rat TRH receptor as detailed under "Experimental Procedures." The COOHterminal amino acid derived from the TRHR was altered from Ile to Leu, and a two-amino acid (Gly-Ala) linker was inserted between the two elements of the fusion protein to facilitate construction. Although not used specifically in the current study, an antibody to the VSV epitope tag permitted detection of the construct after both transient and stable expression. M) or TRHR (EC 50 ϭ 6.9 ϫ 10 Ϫ9 M). However, the fold stimulation of [ 3 H]inositol phosphate production over basal produced by VSV-TRHR-GFP (3.2-fold) was lower than that produced by either VSV-TRHR (6.7-fold) or TRHR (6.2-fold). The capacity of each construct to bind [ 3 H]TRH in membranes prepared from these transiently transfected cells and for this to be competed for by increasing concentrations of TRH allowed approximate assessment of the levels of expression and affinity for TRH by application of the formalisms of DeBlasi et al. (35) to the binding data. Over a series of such transfections the estimated K d for TRH at VSV-TRHR-GFP was 3.3 ϫ 10 Ϫ8 M, for VSV-TRHR it was 3.8 ϫ 10 Ϫ8 M, and for TRHR it was 4.9 ϫ 10 Ϫ8 M.
With clear demonstration that the VSV-TRHR-GFP construct was able both to bind TRH with high affinity and to activate G proteins and second messenger responses upon addition of agonist, this construct was expressed stably in HEK293 cells. A single clone, designated VTGP1, was selected for detailed analysis based on expression of the construct with an essentially homogeneous plasma membrane distribution (see later) and maintained contact inhibition of the cells (data not shown). Intact cell binding studies were performed on cells of clone VTGP1 (Fig. 3). These experiments indicated a K d for TRH of 3.8 Ϯ 1.2 ϫ 10 Ϫ8 M with expression of VSV-TRHR-GFP at 1.2 Ϯ 0.2 ϫ 10 5 copies/cell. Clone VTGP1 cells were visualized in a confocal microscope and then exposed to 1 ϫ 10 Ϫ7 M TRH to allow real time and direct visualization of VSV-TRHR-GFP (Fig. 4). Initially the construct was concentrated heavily at the plasma membrane of the cells (Fig. 4a). However, within 3-5 min distinct internalization was occurring (Fig. 4b). This became more pronounced such that within 10 -15 min the bulk of the VSV-TRHR-GFP had been relocated away from the plasma membrane (Fig. 4c), an effect that was maintained at subsequent times up to at least 30 min (Fig. 4d).
To establish that occupation of VSV-TRHR-GFP by TRH was required to produce internalization/sequestration we examined the concentration dependence and pharmacology of the effect. Cells were incubated with or without the low affinity TRHR antagonist/inverse agonist chlordiazepoxide (100 M) (36) in the presence of concentrations of TRH ranging from 5 ϫ 10 Ϫ10 M to 5 ϫ 10 Ϫ8 M (Fig. 5) for 30 min and fixed and visualized. Chlordiazepoxide alone did not alter the cellular distribution of VSV-TRHR-GFP (Fig. 5a). 5 ϫ 10 Ϫ10 M TRH produced little internalization of VSV-TRHR-GFP (data not shown); however, 5 ϫ 10 Ϫ9 M TRH caused strong internalization (Fig. 5b). Coadministration of chlordiazepoxide (100 M) along with 5 ϫ 10 Ϫ9 M TRH largely inhibited internalization (Fig. 5c), but 5 ϫ 10 Ϫ8 M TRH was able to overcome the blockade of internalization caused by the antagonist (Fig. 5d). The presence of hyperosmotic conditions has been established to block internalization processes that utilize clathrincoated pits and vesicles (37,38). To establish the concentration requirements for sucrose, intact clone VTGP1 cells in medium were exposed to 1 ϫ 10 Ϫ8 M [ 3 (Fig. 6). We wished to equate this biochemical measurement of VSV-TRHR-GFP internalization with direct visualization. Using the same assay conditions, TRH induced strong internalization of VSV-TRHR-GFP in a 30-min period in the absence of sucrose (Fig. 7, a and b). Dramatic differences were observed, however,

FIG. 2. VSV-TRHR-GFP stimulates inositol phosphate production in response to TRH with high potency. Each of TRHR (diamonds), VSV-TRHR (circles), and VSV-TRHR-GFP (squares) was expressed transiently in HEK293 cells. The cells were prelabeled with [ 3 H]inositol (1 Ci/ml, 24 h), and the capacity of varying concentrations of TRH to stimulate production of [ 3 H]inositol phosphates was assessed.
in the presence of sucrose and indeed between the two concentrations of sucrose used. In the presence of 0.4 M sucrose, little internalization was observed in response to TRH (Fig. 7c).
However, the VSV-TRHR-GFP was now distributed nonuniformly around the plasma membrane with clear punctate concentrations. In the presence of 0.3 M sucrose TRH produced a composite pattern in which a significant fraction of VSV-TRHR-GFP was internalized, but equally a significant fraction was located at punctate but plasma membrane-delineated sites (Fig. 7d).
Transferrin is internalized constitutively via clathrin-coated vesicles (39,40). Dual wavelength scanning allowed concurrent detection of plasma membrane-located VSV-TRHR-GFP and internalized Texas Red transferrin in live, untreated, clone VTGP1 cells. Application of TRH resulted in a time-dependent internalization of VSV-TRHR-GFP into vesicle populations that were either identical with or overlapped those containing Texas Red transferrin as assessed by the merging of the red and green color signals over a 6 -20-min period (Fig. 8).
Several studies have suggested roles for the cellular cytoskeletal architecture in internalization processes (41). Treatment of cells with cytochalasin B, an inhibitor of microfilament function (4 g/ml, 30 min), did not interfere with subsequent TRH-mediated internalization (Fig. 9a). By contrast, pretreatment with the microtubule-disrupting agent nocodazole (1 ϫ 10 Ϫ5 M, 60 min) resulted in TRH producing a punctate plasma membrane and periplasma membrane distribution pattern without resulting in the intense perinuclear, deep, cellular redistribution observed without pretreatment with nocodazole (Fig. 9b). The acid resistance of the bulk of the intact cell binding of [ 3 H]TRH after nocodazole treatment (Table I) demonstrated that the bulk of the observed VSV-TRHR-GFP was truly internalized rather than simply being bound to plasma membrane-localized TRHR which could be accessed by the low pH wash.
Very brief exposure (1 min) to TRH followed by washout of the ligand resulted in the appearance of a punctate plasma membrane staining and initiation of internalization (Fig. 10a). By following the same group of intact cells over an extended time period, even 40 min after the washout of TRH, substantial pools of internalized VSV-TRHR-GFP could be observed in many cells (Fig. 10b). Over this period substantial amounts of VSV-TRHR-GFP were recycled back to the plasma membrane. Furthermore, the plasma membrane distribution of remaining VSV-TRHR-GFP again became essentially homogeneous. A second period of exposure to TRH was then initiated (Fig. 10c). Rapid concentration of the plasma membrane VSV-TRHR-GFP into punctate structures was again observed, and further internalization of the VSV-TRHR-GFP was initiated, which could be most easily observed in the cells that had not retained heavy levels of internalized VSV-TRHR-GFP during the incubation phase in the absence of TRH (Fig. 10c). After sustained exposure to TRH, plasma membrane levels of VSV-TRHR-GFP became virtually undetectable (Fig. 10d).

DISCUSSION
The use of chimeric proteins containing modified forms of the GFP from A. victoria has recently provided a novel means to visualize the expression, distribution, and redistribution of proteins in response to stimulation in intact cells and in real time (28 -30). The use of this approach has recently been pioneered for G protein-coupled receptors in studies involving either transient or stable expression of both the ␤ 2 -adrenoreceptor (31,33) and the cholecystokinin receptor type A (32). In the current study we expand this to the use of a third G protein-coupled receptor, that for TRH.
One issue in any approach of this nature must be to demonstrate that the chimeric receptor-GFP maintains the pharmacological and signal transduction properties of the native receptor. This has, perhaps surprisingly, proved to be less of an issue than might be imagined after the addition of a 27-kDa polypeptide to the COOH-terminal tail of a receptor (32,33). Furthermore, GFP is not the only protein that can be attached to the COOH-terminal tail of a G protein-coupled receptor without disrupting function. We and others have produced fusion proteins in which the ␣ subunits of heterotrimeric G proteins have been attached directly to the COOH-terminal tail of various receptors (42)(43)(44)(45)(46). From the long isoform of the rat TRH receptor, we generated for these studies both an NH 2terminally epitope-tagged form of the receptor (VSV-TRHR) and subsequently the final chimeric protein (VSV-TRHR-GFP). Transient expression studies were used initially to demonstrate that the bulk of the expressed VSV-TRHR-GFP was targeted appropriately to the plasma membrane (14), that TRH bound to the construct effectively, and that the addition of TRH resulted in stimulation of second messenger production. In such transient assays, although clear plasma membrane tar-geting of the construct was observed (14), significant levels of fluorescence were also associated with perinuclear structures (14). It is unclear if this is simply a reflection of ongoing expression or whether this represents improperly folded or modified protein. Other studies have suggested that expression of the TRHR in HEK293 results in a largely intracellular localization of the protein (47). Attachment of the GFP to the rat TRH receptor did not reduce the affinity of binding of TRH or prevent G protein activation upon addition of TRH as judged by the capacity of the construct to stimulate [ 3 H]inositol phosphate generation.
Based both on the perinuclear localization of a significant fraction of VSV-TRHR-GFP after transient expression and obvious variation in expression levels in individual cells when using such an approach (data not shown) we generated cell lines stably expressing VSV-TRHR-GFP. Clone VTGP1 expressed VSV-TRHR-GFP to some 1.2 ϫ 10 5 copies/cell. This cell line allowed real time monitoring and visualization of alterations in VSV-TRHR-GFP distribution in response to ligand. Internalization was detectable with 3-5 min, virtually maximal within 10 -15 min, and was maintained up to at least 30 min. These results are consistent with related studies that have examined either the internalization of radiolabeled TRH (13,47) or have used epitope-tagged receptors that could be detected immunologically after cellular fixation (10,11,15). The TRHR has an extremely limited useful pharmacology, with the only widely available antagonist being chlordiazepoxide with K i in the region of 100 M. Despite this limitation we were able to demonstrate both concentration-dependent internalization of VSV-TRHR-GFP by TRH and that this was not mimicked by chlordiazepoxide. Furthermore, chlordiazepoxide blocked the effect of low but effective concentrations of TRH, and this could be overcome by higher concentrations of TRH. As such, VSV-TRHR-GFP internalization displayed a rational TRHR pharmacology.
Among the most interesting observations in the current study were the likely contributions of clathrin-coated vesicles and pits to agonist-induced internalization of VSV-TRHR-GFP.  Cellular shock with elevated osmolality is a well established means of interfering with this pathway (37,38) and has been used previously on studies of the TRH receptor (11,13). We demonstrated in initial studies that TRH bound to VSV-TRHR-GFP in intact cell binding assays with good affinity, which was at least as high as that to the isolated TRHR. The use of acid wash protocols to resolve plasma membrane-located and internalized TRHR is well established (11,13,36). We therefore used a combination of such a biochemical assay in concert with real time confocal visualization to explore this process in detail. The [ 3 H]TRH binding studies demonstrated that although 0.4 M sucrose was able to reduce VSV-TRHR-GFP internalization in response to 1 ϫ 10 Ϫ8 M TRH by some 80%, 0.3 M sucrose was less effective. These predictions were validated entirely in the cell biological examinations. Little internalization of VSV-TRHR-GFP could be observed after shock with 0.4 M sucrose. However, the pattern of VSV-TRHR-GFP fluorescence did not remain uniform around the plasma membrane. Instead, TRH now produced very clear punctate concentrations of VSV-TRHR-GFP at specific locations in the plasma membrane. It is tempting to assume that these represent clathrin-coated vesicles, but such definition will require detailed analysis, probably requiring electron microscopy for clear confirmation. After shock with 0.3 M sucrose and treatment with TRH a combination of punctate plasma membrane staining and internalization of VSV-TRHR-GFP was observed.
To explore the contribution of cytoskeletal elements to agonist-induced internalization of VSV-TRHR-GFP, clone VTGP1 cells were pretreated with either cytochalasin B or with nocodazole. Although cytochalasin B had no obvious effect on TRH-induced internalization of VSV-TRHR-GFP, nocodazole produced a striking effect. Punctate concentrations of VSV-TRHR-GFP were formed at and just below the plasma membrane, but further internalization was blocked. A growing literature has indicated that the movement of internalized proteins through early endosomes to late endosomes is dependent upon an integrity of the microtubular structure (48,49). By the use of nocodazole we appear to have trapped the internalized VSV-TRHR-GFP within such a vesicle population (Fig. 9b).
In an attempt to investigate the nature of the vesicular pool(s) into which the VSV-TRHR-GFP was internalized we performed covisualization studies with Texas Red transferrin. Transferrin is internalized constitutively, along with the transferrin receptor, into early endosomes through a recycling pool and then back to the plasma membrane (39,40). After preincubation of clone VTGP1 cells with Texas Red transferrin but in the absence of TRH, the VSV-TRHR-GFP and Texas Red transferrin could be shown to be compartmentalized separately. However, after exposures to TRH as brief as 6 -10 min the distribution of internalized VSV-TRHR-GFP could be shown to overlap strongly with that of Texas Red transferrin as demonstrated by the appearance of yellow color. It is not possible at this level of resolution to state categorically that this would indicate colocalization in the same vesicles, but clearly the vesicles containing the two proteins are at least in close proximity.
One feature of the expression of VSV-TRHR-GFP in the plasma membrane of clone VTGP1 cells was its uniformity of distribution in the absence of agonist. However, very early time points of agonist challenge as well as pharmacological interventions such as high sucrose concentration shocks resulted in the development of more localized patterns of VSV-TRHR-GFP localization. FIG. 10. Internalization, recycling, and further internalization of VSV-TRHR-GFP as TRH is applied, removed, and reapplied. Clone VTGP1 cells were treated with TRH (1 ϫ 10 Ϫ7 M) for 1 min. This was removed subsequently with extensive washing. Intact cells were then visualized to detect the cellular location of VSV-TRHR-GFP. After the time required for washing, the first exposure was at 4 min (panel a). Extensive internalization could already be observed as well as distinct plasma membrane clustering. Within 10 min, recycling of internalized VSV-TRHR-GFP to the plasma membrane was observed (not shown). 40 min after washout (panel b) a substantial amount of internalized VSV-TRHR-GFP could still be observed in many cells while the remaining plasma membrane located VSV-TRHR-GFP was once more distributed uniformly. TRH (1 ϫ 10 Ϫ7 M) was applied again; (panel c) within 2 min a new round of internalization was evident. With sustained exposure to TRH virtually all of the detectable VSV-TRHR-GFP was internalized within 20 min (panel d).
Considerable interest has been accorded not only to internalization of G protein-coupled receptors but also to their possible recycling to the plasma membrane. To examine aspects of this we challenged clone VTGP1 cells with TRH for only 1 min and then washed them extensively to remove the agonist. Clear development of punctate plasma membrane staining and internalization of VSV-TRHR-GFP was observed as rapidly as we could visualize the cells after this protocol (about 4 min). Subsequent visualization of the cells to 40 min after washout of TRH did not simply result in a recycling of the entire internalized population of VSV-TRHR-GFP back to the plasma membrane, although this seemed to be the case in certain cells. Two features, however, were obvious over the washout period. First, the remaining and recycled plasma membrane-located VSV-TRHR-GFP returned to a uniform distribution; and second, populations of the initially internalized VSV-TRHR-GFP remained internalized, unable to recycle to the plasma membrane. Because lysosomal compartments do not provide a general mechanism for proteins to be returned (50,51), it is likely that VSV-TRHR-GFP observed to be still internalized after maintenance of the cells in the absence of agonist for a substantial period represents protein committed to degradation and thus down-regulation.
These studies provide real time analysis of the processes involved in the clustering and internalization of a stably expressed G protein-coupled receptor in response to agonist and represent the first visualization of this process for the TRHR in intact, living cells. This could not have been achieved without the use of approaches derived from and dependent upon the use of GFP. Further development of such approaches and its use in concert with compartmental ablation strategies and electron microscopy are likely to provide detailed understanding of the trafficking of G protein-coupled receptors.