Internalization of the TXA2 Receptor a and b Isoforms ROLE OF THE DIFFERENTIALLY SPLICED COOH TERMINUS IN AGONIST-PROMOTED RECEPTOR INTERNALIZATION*

Thromboxane A2 (TXA2) potently stimulates platelet aggregation and smooth muscle constriction and is thought to play a role in myocardial infarction, atherosclerosis, and bronchial asthma. The TXA2 receptor (TXA2R) is a member of the G protein-coupled receptor family and is found as two alternatively spliced isoforms, a (343 residues) and b (407 residues), which share the first 328 residues. In the present report, we demonstrate by enzyme-linked immunosorbent assay and immunofluorescence microscopy that the TXA2Rb, but not the TXA2Ra, undergoes agonist-induced internalization when expressed in HEK293 cells as well as several other cell types. Various dominant negative mutants were used to demonstrate that the internalization of the TXA2Rb is dynamin-, GRK-, and arrestin-dependent in HEK293 cells, suggesting the involvement of receptor phosphorylation and clathrin-coated pits in this process. Interestingly, the agonist-stimulated internalization of both the a and b isoforms, but not of a mutant truncated after residue 328, can be promoted by overexpression of arrestin-3, identifying the C-tails of both receptors as necessary in arrestin-3 interaction. Simultaneous mutation of two dileucine motifs in the C-tail of TXA2Rb did not affect agonist-promoted internalization. Analysis of various C-tail deletion mutants revealed that a region between residues 355 and 366 of the TXA2Rb is essential for agonist-promoted internalization. These data demonstrate that alternative splicing of the TXA2R plays a critical role in regulating arrestin binding and subsequent receptor internalization.

Thromboxane A 2 (TXA 2 ) potently stimulates platelet aggregation and smooth muscle constriction and is thought to play a role in myocardial infarction, atherosclerosis, and bronchial asthma. The TXA 2 receptor (TXA 2 R) is a member of the G protein-coupled receptor family and is found as two alternatively spliced isoforms, ␣ (343 residues) and ␤ (407 residues), which share the first 328 residues. In the present report, we demonstrate by enzyme-linked immunosorbent assay and immunofluorescence microscopy that the TXA 2 R␤, but not the TXA 2 R␣, undergoes agonist-induced internalization when expressed in HEK293 cells as well as several other cell types. Various dominant negative mutants were used to demonstrate that the internalization of the TXA 2 R␤ is dynamin-, GRK-, and arrestin-dependent in HEK293 cells, suggesting the involvement of receptor phosphorylation and clathrin-coated pits in this process. Interestingly, the agonist-stimulated internalization of both the ␣ and ␤ isoforms, but not of a mutant truncated after residue 328, can be promoted by overexpression of arrestin-3, identifying the C-tails of both receptors as necessary in arrestin-3 interaction. Simultaneous mutation of two dileucine motifs in the C-tail of TXA 2 R␤ did not affect agonist-promoted internalization. Analysis of various C-tail deletion mutants revealed that a region between residues 355 and 366 of the TXA 2 R␤ is essential for agonist-promoted internalization. These data demonstrate that alternative splicing of the TXA 2 R plays a critical role in regulating arrestin binding and subsequent receptor internalization.
Thromboxane A 2 (TXA 2 ) 1 has a variety of pharmacologic effects which modulate the physiological responses of several cells and tissues (1). It is a product of the sequential metabolism of arachidonic acid by the cyclooxygenases and TXA 2 synthase (2). TXA 2 formation can result from activation of various cell types, including platelets, macrophages, and vascular smooth muscle cells (1). Binding of TXA 2 to its receptor (TXA 2 R) induces platelet aggregation, constriction of vascular and bronchiolar smooth muscle cells, as well as mitogenesis and hypertrophy of vascular smooth muscle cells. TXA 2 has been implicated in a wide variety of cardiovascular diseases (1).
While pharmacological studies have suggested the existence of TXA 2 R subtypes (3), the receptor appears to be encoded by a single gene that can be alternatively spliced in the carboxylterminal tail (C-tail) leading to two variants, TXA 2 R␣ and -␤, that share the first 328 amino acids. Complementary DNAs for the 343-amino acid TXA 2 R␣ were cloned from placental and megakaryocytic sources (4), whereas a cDNA for the 407-amino acid TXA 2 R␤ was isolated from a vascular endothelial library (5). The TXA 2 Rs have been shown to couple to the G proteins G q , G i2 , G 11 , G 12 , G 13 , G 16 , and an 85-kDa unidentified G protein, explaining the multiplicity of TXA 2 R-mediated signal transduction (6 -11). While no isoform-specific biological functions have been ascribed to either of the TXA 2 Rs, the different C-tails are likely to play a role in G protein-coupling specificity (12), and perhaps confer different desensitization characteristics. It has been suggested that homologous desensitization of the TXA 2 Rs in cells can be divided into two stages: the early stage involves uncoupling of receptors from G proteins, while a later stage involves a loss of receptor sites from the plasma membrane (13)(14)(15)(16). Recently, Habib et al. (17) showed that the dose and time dependence of agonist-induced receptor phosphorylation appeared similar for both isoforms in HEK293 cells stably overexpressing the TXA 2 Rs. However, phorbol esterpromoted activation of protein kinase C prevented the agonistmediated intracellular calcium rise in TXA 2 R␤-, but not TXA 2 R␣-, stably expressing CHO cells (18).
For many G protein-coupled receptors (GPCRs), the presence of Ser and Thr phosphorylation sites is critically important in receptor desensitization and internalization (see Ref. 19, and references therein). Homologous receptor desensitization appears to be primarily initiated by phosphorylation of the agonist-activated GPCR by a family of Ser/Thr kinases known as G protein-coupled receptor kinases (GRKs). This leads to high affinity binding of a second class of proteins referred to as arrestins (for review, see Refs. 19 -21), resulting in steric inhibition of G protein binding (22,23). Recently, non-visual arrestins were also found to promote receptor uptake into clathrin-coated pits. The non-visual arrestins, arrestin-2 and arrestin-3, 2 function as adaptor proteins to recruit receptors to coated pits by virtue of their ability to bind to both activated, phosphorylated GPCRs and clathrin (24). Studies of the m2 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Recipient of a postdoctoral fellowship from the Medical Research Council of Canada.
We were interested in determining if the different C-tails of the TXA 2 R␣ and -␤ could bestow distinct internalization characteristics to these receptor isoforms. Here, we report that the TXA 2 R␤, but not the TXA 2 R␣, undergoes agonist-induced internalization in several different cell types. The internalization of TXA 2 R␤ is inhibited by dominant negative forms of dynamin, GRK2, and arrestins, while internalization of both isoforms can be promoted by overexpression of GRK2 and arrestins. A small region which contains 3 serine residues in the C-tail of TXA 2 R␤ was identified as being essential for internalization. Taken together, these data suggest a role for receptor phosphorylation in TXA 2 R␤ agonist-induced internalization, in a process likely involving arrestins and clathrin-coated pits. Our results indicate that the C-tail of these two receptors is responsible for their different internalization properties.

EXPERIMENTAL PROCEDURES
Cell Culture and Transient Transfection-Human embryonic kidney cells (HEK293) and human epidermoid carcinoma cells (A431) were maintained in Dulbecco's modified Eagle's medium (DMEM, Life Technologies, Inc.) supplemented with 10% fetal bovine serum. MiaPACA (human pancreatic carcinoma cells) were grown in the same medium also containing 2.5% horse serum (Life Technologies, Inc.). Mouse lymphoid neoplasm cells (P388D1) were maintained in RPMI supplemented with 15% fetal bovine serum while CHO cells (Chinese hamster ovary cells) were grown in Ham's F-12 with 10% fetal bovine serum. All cells were kept in a humidified atmosphere of 95% air, 5% CO 2 at 37°C. Transfections were done with Fugene-6 (Boehringer Mannheim) according to the manufacturer's recommendations.
Construction of Epitope-tagged TXA 2 Receptors and Mutants-cDNAs for the TXA 2 R␣ and TXA 2 R␤ were obtained by reverse transcriptase-polymerase chain reaction using a pool of total RNA from HeLa, A431, U937, and MiaPACA cells. Primers used for TXA 2 R␣ were 5Ј-GGAATTCATGTGGC-CCAACGGCAGTTCCCTG-3Ј (TXAF) and 5Ј-CCGCTCGAGTCTTCCAAT-GTCTGCATGCCC-3Ј, while the primers TXAF and 5Ј-CCGCTCGAGCAT-TCAATCCTTTCTGGACAGAGC-3Ј(TXAB) were used for TXA 2 R␤. A pcDNA3 vector containing an HA epitope was constructed by annealing the two primers 5Ј-AGCTTCGATCGTCGACATGTACCCATACGATGTTCCA-GATTACGCTTCTAGAGGATCCCCGGGCGAGCTCG-3Ј and 5Ј-AATTCG-AGCTCGCCCGGGGATCCTCTAGAAGCGTAATCTGGAACATCGTATG-GGTACATGTCGACGATCGA-3Ј and ligating them in pcDNA3 digested with HindIII and EcoRI. The same strategy was employed to generate a pcDNA3FLAG vector using 5Ј-AGCTTGGGCACCATGAAGACGATCATC-GCCCTGAGCTACATCTTCTGCCTGGTGTTCGCCGACTACAAGGACG-ATGATGACACCG-3Ј and 5Ј-AATTCGGTGTCATCATCGTCCTTGTAGT-CGGCGAACACCAGGCAGAAGATGTAGCTCAGGGCGATGATCGTCTT-CATGGTGCCCA-3Ј as primers. The receptors were then subcloned in these vectors, yielding in-frame constructs expressing epitope-tagged receptor polymerase chain reaction. Mutated receptors were constructed by polymerase chain reaction using the Expand High Fidelity PCR System (Boehringer Mannheim) according to the manufacturer's recommendations, with the TXA 2 R␣ and TXA 2 R␤ cDNAs as templates. Mutants were created using the TXAF primer and the following corresponding primers: Radioligand Binding Assays-Competition binding curves were done on HEK293 cells expressing wild-type and mutant receptor species. Cells were harvested and washed twice in Buffer A (10 mM Hepes, pH 7.6, 129 mM NaCl, 8.9 mM NaHCO 3 , 0.8 mM KH 2 PO 4 , 0.8 mM MgCl 2 , 5.6 mM dextrose, 0.38% sodium citrate, pH 7.4, 5 mM EDTA, and 5 mM EGTA). Binding reactions were carried out on 5 ϫ 10 4 cells in a total volume of 0.25 ml in the same buffer with 10 nM [ 3 H]SQ29543 (a TXA 2 antagonist) and increasing concentrations of nonradioactive SQ29543 for 2 h at room temperature. Reactions were stopped by centrifugation at 14,000 rpm in a microcentrifuge for 1 min and the cell-associated radioactivity was measured by liquid scintillation.
Internalization Assays-For quantification of receptor internalization, ELISA assays were performed essentially as described by Daunt et al. (37). The cell lines were plated out at 6 ϫ 10 5 cells per 60-mm dish, transfected with 6 g of DNA and split after 24 h into 6 wells of 24-well tissue culture dishes previously coated with 0.1 mg/ml poly-L-lysine (Sigma). After another 24 h, the cells were washed once with PBS and incubated in DMEM at 37°C for several minutes. Then the TXA 2 agonist U46619 was added at a concentration of 100 nM in prewarmed DMEM to the wells. The cells were then incubated for various times at 37°C and reactions were stopped by removing the media and fixing the cells in 3.7% formaldehyde/TBS for 5 min at room temperature. The cells were then washed three times with TBS and nonspecific binding blocked with TBS containing 1% BSA for 45 min at room temperature. The first antibody (monoclonal HA 101R, Babco) was added at a dilution of 1:1000 in TBS/BSA for 1 h at room temperature. Three washes with TBS followed, and cells were briefly reblocked for 15 min at room temperature. Incubation with goat anti-mouse conjugated alkaline phosphatase (Bio-Rad) diluted 1:1000 in TBS/BSA was carried out for 1 h at room temperature. The cells were washed three times with TBS and a colorimetric alkaline phosphatase substrate was added. When the adequate color change was reached, 100-l samples were taken for colorimetric readings. Cells transfected with pcDNA3 were studied concurrently to determine background. All experiments were done in triplicate.
Inositol Phosphate Determination-HEK293 cells were seeded at a density of 80,000 cells per well of 12-well plates and transfected as described above with the WT or mutant receptors and labeled the following day for 18 -24 h with myo-[ 3 H]inositol at 4 Ci/ml in DMEM (high glucose, without inositol). After labeling, cells were washed once in phosphate-buffered saline (PBS) and incubated in prewarmed DMEM (high glucose, without inositol) containing 0.5% BSA, 20 mM Hepes, pH 7.5, and 20 mM LiCl for 10 min. Cells were then stimulated for 10 min with different concentrations of U46619. The reactions were terminated by removing the stimulation media and by the addition of 0.8 ml of 0.4 M perchloric acid. Samples were harvested in Eppendorf tubes, and a 0.5 volume of 0.72 N KOH, 0.6 M KHCO 3 was added. Tubes were vortexed and centrifuged for 5 min at 14,000 rpm in a microcentrifuge. Inositol phoshates were separated on Dowex AG 1-X8 columns. Total labeled inositol phosphates were then counted by liquid scintillation.
Immunofluorescence Microscopy-Cells (HEK293, P388D1, Mi-aPACA, A431, and CHO) were grown in 35-mm dishes on coverslips and then transfected as described above with 2 g of DNA/well. After 48 h, cells were incubated with Flag M1 antibody (1:500 dilution) for 1 h at 4°C in DMEM supplemented with 1% BSA and 1 mM CaCl 2 . Cells were washed twice with PBS containing 1 mM CaCl 2 , then treated with 100 nM U46619 for 1 h at 37°C in DMEM with 0.5% BSA, 20 mM Hepes, pH 7.4, 1 mM CaCl 2 . The cells were then fixed with 3.7% formaldehyde/PBS for 15 min at room temperature, washed with PBS/CaCl 2 , and permeabilized with 0.05% Triton X-100/PBS/CaCl 2 for 10 min at room temperature. Nonspecific binding was blocked with blotto (0.05% Triton X-100/PBS/CaCl 2 containing 5% nonfat dry milk) for 30 min at 37°C. Goat anti-mouse fluorescein isothiocyanate-conjugated secondary antibody (Molecular Probes) was then added at a dilution of 1:150 in blotto for 1 h at 37°C. The cells were then washed six times with permeabilization buffer with the last wash left at 37°C for 30 min. Finally, the cells were fixed with 3.7% formaldehyde as described. Coverslips were mounted using Slow-Fade mounting medium (Molecular Probes) and examined by microscopy on a Nikon Eclipse E800 fluorescence microscope using a Plan Fluor 60ϫ objective. Cells expressing the lowest levels of transfected proteins, but clearly above those of nonexpressing cells, were chosen for view. Images were collected using QED Camera software and processed with Adobe Photoshop v. 3.0.

RESULTS
A schematic representation of the carboxyl-terminal tails of the two isoforms of the human TXA 2 receptor is shown in Fig.  1. In order to investigate the internalization of the TXA 2 R␣ and -␤, epitope-tagged TXA 2 Rs were transiently expressed in HEK293 cells. Similar to previous reports (5,12,17), the TXA 2 R␣ and -␤ display similar binding affinities for the specific TXA 2 R antagonist SQ29543, with respective K d values of 11.2 Ϯ 1.4 and 12.4 Ϯ 1.8 nM, indicating that the C-tails do not affect the binding properties of these receptors. These observations were confirmed by evaluation of ligand binding to a mutant receptor truncated after residue 328 (R328Stop) (K d ϭ 12.9 Ϯ 1.7 nM). Addition of an HA or Flag epitope tag at the NH 2 terminus of the receptors also did not alter ligand affinities, nor did it effect the activation characteristics of the receptors as determined by their respective EC 50 values for inositol phosphate generation in response to agonist (data not shown).
The TXA 2 R expressing HEK293 cells were next assessed for agonist-promoted loss of cell surface TXA 2 Rs using an ELISA assay. A time course analysis of receptor internalization following stimulation with 100 nM of the agonist U46619 is illustrated in Fig. 2. While the TXA 2 R␣ does not undergo agoniststimulated internalization, ϳ40% of the TXA 2 R␤ are internalized after 2 h of agonist treatment, with a t1 ⁄2 of ϳ45 min. Stimulation of cells with U46619 concentrations ranging from 0.1 to 1000 nM all produced maximal effects which could be blocked by the specific antagonist SQ29543, while even 1 M U46619 treatment failed to induce TXA 2 R␣ internalization (data not shown). To confirm these results, we performed an additional series of studies using immunofluorescence microscopy to directly visualize internalization of the Flag-tagged TXA 2 receptors in HEK293 cells. Cells were incubated with the Flag antibody prior to agonist exposure so that only the trafficking of receptors initially present on the cell surface would be detected. After agonist treatment, cells were permeabilized and the Flag epitope was visualized with fluorescein-conjugated secondary antibody. As shown in Fig. 3A, the TXA 2 R␣ is found primarily at the cell surface even after the cells have been stimulated with agonist for 1 h. In contrast, agonist exposure of cells expressing the TXA 2 R␤ resulted in a striking redistribution of the receptors to intracellular compartments. These observations support the results obtained by ELISA analysis. Interestingly, there is also a significant amount of TXA 2 R␤ redistribution when the cells are incubated at 37°C in the absence of agonist compared with cells kept at 4°C (Fig.  3A). This suggests that the TXA 2 R␤ may also undergo some tonic internalization in the absence of agonist.
HEK293 cells may contain low levels of endogenous TXA 2 Rs (38) and thus would be expected to possess the necessary "cellular machinery" for TXA 2 R regulation. However, we also wanted to test whether these observations could be extended to other cell types. Thus, similar immunofluorescence microscopy studies were performed in P388D1 and A431 cells, which contain endogenous TXA 2 Rs, and MiaPACA and CHO cells, which lack endogenous TXA 2 Rs (data not shown). As depicted in Fig.  3B, the TXA 2 R␤ but not TXA 2 R␣ undergoes agonist-induced internalization in A431 cells transiently transfected with the epitope-tagged receptors. Comparable results were obtained in P388D1, MiaPACA, and CHO cells (data not shown). These results demonstrate that the differential internalization of TXA 2 R␣ and -␤ is a property of the receptors and that HEK293 cells are a good model for studying TXA 2 R trafficking. We next addressed whether the inability of the TXA 2 R␣ to internalize was due to low levels of proteins involved in receptor trafficking or a low affinity of the TXA 2 R␣ for such proteins. In recent studies, GRKs and arrestins have been shown to promote GPCR internalization (24, 25, 27-30, 39, 40). Indeed, it has been suggested that the extent of ␤ 2 -adrenergic receptor Single lines indicate homology between the C-tail of TXA 2 R␣ and residues 329 -344 of TXA 2 R␤. Double lines illustrate the sites where truncation mutations were made. Solid circles represent residues that were mutated to alanines.
internalization correlates with the intracellular levels of GRKs and arrestins (30). The ability of coexpressed GRK2, arrestin-2, and arrestin-3 to promote TXA 2 R␣ internalization after U46619 stimulation was thus evaluated (Fig. 4A). Coexpression of GRK2 resulted in a small increase in internalization (from ϳ1% in the absence to ϳ5% in the presence of GRK2), whereas co-transfection of either arrestin-2 or arrestin-3 significantly increased TXA 2 R␣ internalization (to ϳ22%). Coex-pression of GRK3, GRK5, or GRK6 also promoted only a modest increase in TXA 2 R␣ internalization (data not shown), indicating that GRKs are unlikely to be limiting in this process. Internalization of the TXA 2 R␤ was also promoted by coexpression of GRK2 and arrestins although the effects were modest (Fig. 4A). Time course experiments of receptor internalization were then performed in the presence of arrestin-3 coexpression (Fig. 4B). TXA 2 R␣ internalization in the presence of arrestin-3 was ϳ15% after 30 min of agonist exposure and reached a plateau of 20 -24% after 60 min. The rate and extent of TXA 2 R␤ internalization was also enhanced by arrestin-3 coexpression and reached levels of 50 -55% with a t1 ⁄2 of ϳ30 min. Since endogenous expression of arrestins in HEK293 cells was previously shown to be relatively high (30), our results suggest that the TXA 2 R␣ has a low affinity for arrestins. However, these results do not exclude the possibility that the two receptor isoforms differ in their ability to bind other proteins involved in receptor trafficking that might also contribute to their distinct internalization profiles.
Since the C-tail of the TXA 2 R␤ appears critical in the internalization of the receptor, we next tried to determine whether this function could be ascribed to any particular region of the tail. In addition to the 11 serine and 4 threonine residues serving as potential phosphorylation sites involved in internalization, the C-tail of the TXA 2 R␤ also contains other motifs that might contribute to this process. Among these are two dileucine motifs, Leu-386/387, and Leu-392/393. Dileucine motifs have been implicated in internalization of several receptors, including the ␤ 2 -adrenergic receptor (Ref. 44, and references therein). To address the potential role of dileucine motifs we constructed a mutant TXA 2 R␤ in which leucines 386, 387, 392, and 393 were simultaneously changed to alanines (LLLLA). In addition, 11 progressive deletion mutants of the TXA 2 R␤ C-tail were also generated (Fig. 1). All constructs were transfected in HEK293 cells with transfection conditions (i.e. amount of DNA) adjusted to give equivalent receptor expres-sion (ϳ1 pmol/mg of protein). Truncation of the TXA 2 R␤ C-tail had no effect on ligand binding (data not shown) and the R328Stop mutant had the same EC 50 for phosphatidylinositol turnover as the wild-type receptor (ϳ70 nM). The ability of the various receptors to internalize after 3 h of agonist exposure was then evaluated by ELISA analysis (Fig. 6). The LLLLA mutation had no apparent effect on U46619-induced internalization of TXA 2 R␤. Similarly, removal of up to 41 residues from the C-tail (G366Stop) had no effect on agonist-promoted internalization. However, truncation to residue 362 significantly reduced internalization while further deletion to residue 355 or beyond completely abolished TXA 2 R␤ internalization. Thus, the region found between residues 355 and 366 seems to play a critical role in TXA 2 R␤ internalization. These results, together with the potent inhibition of internalization caused by GRK2-K220R coexpression (Fig. 5), suggest that the serine residues within this domain may serve as GRK phosphorylation sites and prove critical in arrestin binding and agonist-promoted receptor internalization.
To further localize the arrestin-binding domain on TXA 2 R␤, we assessed whether arrestin-3 coexpression would enhance U46619-promoted internalization of the different truncation mutants. Arrestin-3 promoted the internalization of wild-type TXA 2 R␣ and -␤, as well as the S344Stop, S355Stop, G366Stop, and F380Stop mutants (Fig. 7). In contrast, the R328Stop mutant was not internalized even in the presence of arrestin-3. Since inclusion of various residues distal to residue 328 restores TXA 2 R␤ internalization and some of the arrestin-3 promotion, it seems likely that the C-tails of both TXA 2 receptor isoforms are required for the promotion of internalization by arrestin-3.

DISCUSSION
TXA 2 is a potent stimulator of platelet aggregation and smooth muscle constriction and is regarded as a mediator of myocardial infarction, atherosclerosis, and bronchial asthma (7). TXA 2 is synthesized by numerous cells in response to various physiological and pathological stimuli (7,45). It is then rapidly secreted and acts as a local hormone in the immediate vicinity of its site of production. Its wide spectrum of actions in the body are mediated by a GPCR that can be alternatively spliced in its C-tail to generate isoforms referred to as ␣ and ␤ (5, 12). It would seem likely that the formation and function of TXA 2 would be tightly regulated, given the critical role of this biological mediator. The effects of TXA 2 are regulated by its hydrolysis to inactive thromboxane B 2 as well as by homologous desensitization of its membrane receptor-mediated responses (17,46,47). The existence of agonist-induced receptor uncoupling and internalization mechanisms for the TXA 2 receptors might seem surprising when the natural agonist has such a short half-life (ϳ30 s). However, TXA 2 is released in large amounts over prolonged periods of time during certain vascular disorders, and could tonically stimulate target cells (16). Indeed, it has been postulated that TXA 2 levels sufficient to cause desensitization of the receptor may be maintained for long periods in vascular beds in which thrombosis and platelet activation are ongoing (16).
It was recently shown that both TXA 2 R␣ and TXA 2 R␤ undergo similar agonist-induced phosphorylation and desensitization in HEK293 cells (6,17). Desensitization of the Ca 2ϩ response after agonist stimulation was also similar for both receptor isoforms when overexpressed in CHO cells (18). However, PKC activation potently inhibited agonist-mediated intracellular Ca 2ϩ mobilization of TXA 2 R␤, but not TXA 2 R␣, indicating that different mechanisms of regulation of these receptor isoforms are likely to exist. Studies in CHRF-288 megakaryocytic (15), 1321N1 human astrocytoma (14), and rat glomerular mesangial (13) cells indicated that TXA 2 R desensitization involves initial uncoupling from the G protein although receptor internalization and degradation appear necessary for maximal desensitization.
The present study investigated the detailed mechanism for agonist-induced internalization of the TXA 2 receptors in HEK293 cells. Using expression of epitope-tagged versions of each receptor isoform in combination with ELISA analysis to measure the proportion of receptor remaining at the cell surface after agonist exposure, we have shown that the TXA 2 R␣ does not internalize even when incubated with high concentrations of U46619 for 3 h. In contrast, TXA 2 R␤ was observed to undergo agonist-promoted internalization with a plateau of ϳ40% after ϳ2-3 h of stimulation. Immunofluorescence microscopy confirmed the inability of TXA 2 R␣ to internalize, whereas the TXA 2 R␤ redistributed to intracellular punctate compartments, typical of early endosomes. Similar results were obtained in several other cell types, including two that express endogenous TXA 2 Rs (P388D1 and A431) and two that lack endogenous TXA 2 Rs (MiaPACA and CHO). These results confirm our observations in HEK293 cells and validate them as an appropriate model to study TXA 2 R trafficking.
Our data are in accordance with results obtained in K562, CHRF-288 megakaryocytic and rat glomerular mesangial cells which displayed a 50 -60% loss of endogenous TXA 2 R-binding sites after 3-6 h of agonist treatment (13,15,16). Internalized TXA 2 Rs were associated with light membrane fractions containing both microsomal and cytoplasmic enzyme markers (16), and treatment with endocytosis inhibitors prevented agonistinduced receptor internalization (15). Nonvisual arrestins can bind to agonist-activated phosphorylated GPCRs and promote their internalization by interacting not only with the GPCR but also with clathrin, the major protein component of the clathrinbased endocytic machinery (24).
Our data demonstrate that agonist-induced internalization of the TXA 2 R␤, like that of the ␤ 2 -adrenergic (24,27), follitropin (32), LH/CG (33), and the opioid (34) receptors, is mediated by both arrestin and clathrin-coated pits. Overexpression of dominant negative forms of arrestin-2, arrestin-3, GRK2, and dynamin inhibit the agonist-induced internalization of the TXA 2 R␤. While coexpression of wild-type GRK2 resulted in a small increase in internalization of both TXA 2 R isoforms, overexpression of arrestins significantly promoted internalization of these receptors suggesting that the arrestins are limiting in HEK293 cells were co-transfected with equal amounts of receptor and pcDNA3 or arrestin-3 DNA. Internalization was measured by ELISA analysis, following a 3-h stimulation with 100 nM U46619. Values shown represent the mean Ϯ S.E. of three to four independent experiments performed in triplicate. Data were analyzed by using the Student's t test (*, p Ͻ 0.05 compared with cells transfected with receptor alone). this process. Similar results were obtained for the LH/CG receptor where overexpression of arrestin-3 enhanced its internalization about 2-fold (33), whereas little or no effect on internalization of the ␤ 2 -adrenergic receptor was seen by overexpressing arrestins in HEK293 cells (27,30,35,48). It was reported that the LH/CG receptor had a slow rate of internalization (t1 ⁄2 ϳ 140 min), compared with that of the ␤ 2 -adrenergic receptor (t1 ⁄2 Ͻ 30 min), suggesting that the LH/CG receptor might have a low affinity for arrestins (33). While HEK293 cells express relatively high levels of arrestins (30), the TXA 2 R␤ internalizes slowly with a t1 ⁄2 ϳ 45 min, whereas TXA 2 R␣ shows no detectable internalization. Arrestin-2(319 -418) (35), arrestin-3(1-320), and arrestin-3(201-409) each inhibited ␤ 2 -adrenergic receptor internalization by ϳ40 -50%, while arrestin-3(284 -409) and arrestin-3(290 -409) inhibited internalization by ϳ70 and ϳ30%, respectively (36). Arrestin-2(319 -418) inhibited the internalization of the LH/CG receptor by ϳ30% (33). While our data for the TXA 2 R␤ generally corroborate these observations, we find greater inhibition of TXA 2 R␤ internalization by overexpression of arrestin-3(201-409) as compared with arrestin-3(284 -409), in contrast to the ␤ 2 -adrenergic receptor results (36). Although the mechanism of the dominant negative nature of arrestin-3(201-409) remains unclear (36), these disparities suggest that differences may emerge between the interaction of receptors with the different components of the clathrin-coated pit-mediated internalization pathway. The inability of dynamin-K44A or GRK2-K220R and arrestin-3(201-409) coexpression to completely inhibit agonist-induced internalization of the TXA 2 R␤ may be a result of the fact that these studies were performed using transient coexpression of receptor and dominant negative constructs. Alternatively, our results might suggest that additional non-GRK/arrestin or non-clathrin-mediated mechanisms of internalization may exist for TXA 2 R␤, as was suggested for other GPCRs (49), such as the m2 muscarinic (26), bradykinin B 2 (50), and angiotensin II type 1A (27) receptors.
The role of intracellular domains of GPCRs in triggering internalization has been studied by mutagenesis. Some segments in the intracellular loops (51-53) as well as in the C-tail domain of the yeast ␣-factor, thyrotropin, gastrin, parathyroid hormone, platelet-activating factor, m2 muscarinic, LH/CG, neurotensin, cholecystokinin, somatostatin type 5, thrombin, ␦ opioid, bradykinin B 2 , and angiotensin receptors were found to be specifically involved in the internalization process (54 -66), whereas deletion mutants of the m1 muscarinic receptor were internalized to a similar extent as the wild-type receptors (67). However, there is still a paucity of data regarding motifs known to target GPCRs to intracellular compartments upon agonist activation. Given the fact that internalization of the TXA 2 R in HEK293 cells seems largely determined by the C-tail of the ␤ isoform, we attempted to identify specific residues responsible for this effect. One plausible motif contained in the C-tail of the TXA 2 R␤ likely to be involved in internalization was two dileucines found at residues 386/387 and 392/393. Dileucine motifs bind AP1 and AP2 clathrin adaptor protein complexes (68) and have been implicated in internalization of several receptors, including the ␤ 2 -adrenergic receptor (44).
Our results indicate that the dileucine motifs are not required for agonist-stimulated internalization of the TXA 2 R␤. However, this does not exclude the possibility that these sequences might play important roles in other mechanisms of receptor regulation, such as receptor down-regulation or recycling. Deletion mutants were generated and a region corresponding to amino acids 355-366 (DSRASASRAAG) was demonstrated to be crucial in TXA 2 R␤ internalization. Significant receptor internalization was still present when the tail was shortened to residue 362, suggesting that the DSRASAS amino acid sequence contains essential determinants for this process. Since dominant negative GRK2 strongly inhibits TXA 2 R␤ internalization, this suggests that GRK-promoted phosphorylation of the receptor is important for its internalization. When taken together, these data argue for a possible role of the 3 serines found between residues 355 and 362 in receptor internalization, particularly Ser-357 which constitutes a good GRK phosphorylation site with an acidic residue preceeding it (69). Evidence is now accumulating that agonist-induced phosphorylation is likely important for the internalization of many GPCRs. A Ser/Thr-rich sequence was postulated to play an important role in internalization of the m1, m2, and m3 muscarinic cholinergic receptors (70). Moreover, either truncation or mutation of the Ser/Thr residues in the C-tail of the thrombin receptor inhibited both its agonist-induced phosphorylation and internalization (64). In addition, COOH-terminal deletions or point mutations of Ser/Thr residues in the COOH terminus of the ␦-opioid receptor significantly reduced agonist-induced internalization (65). Overexpression of dominant negative GRK2 also inhibited internalization of the ␤ 2 -adrenergic and m2 muscarinic acetylcholine receptors (25,28). It will be interesting to examine the desensitization of TXA 2 R mutants since it has been shown for other GPCRs that mutation of different clusters of Ser/Thr delineate distinct mechanisms with unique structural requirements that mediate receptor desensitization and internalization (71).
By observing the internalization of the TXA 2 R␤ deletion mutants, it can be hypothesized that some structural requirements need to be met for proper interaction with the endocytic machinery. Thus, the ability of each mutant to internalize in the presence of arrestin-3 was evaluated as a measure of interaction between the different receptor constructs and arrestin-3. Both isoforms of the receptor displayed enhanced internalization in the presence of coexpressed arrestin-3, while the R328Stop mutant showed no such effect, revealing the requirement of the different C-tails of both receptors to provide some interaction with arrestin-3. Interestingly, despite significant homology with the TXA 2 R␣ C-tail, agonist-induced internalization of the TXA 2 R␤ S344Stop mutant was promoted to a much lower extent by coexpression of arrestin-3, suggesting that residues specific to the TXA 2 R␣ C-tail confer distinct interaction properties with arrestin-3. However, some promotion of internalization by arrestin-3 was seen with the S344Stop, S355Stop, G366Stop, and F380Stop mutants, demonstrating the importance of residues distal to 328 in the TXA 2 R␤ C-tail in the interaction with arrestin-3. This region is likely to correspond to only one of several essential receptor domains involved in arrestin binding as suggested by the findings that interaction of arrestin-1 with rhodopsin, or arrestin-2 and arrestin-3 with the ␤ 2 -adrenergic and m2 muscarinic receptors, involves multiple contact sites that impart apparent positive cooperativity (72)(73)(74)(75)(76)(77). TXA 2 R␣ and TXA 2 R␤ constitute potential models to study factors affecting the binding affinity of arrestins to the receptor. In this regard, C-tail constructs of the two receptor isoforms will be useful in characterizing detailed interactions with arrestins, and to possibly identify additional proteins involved in their differential regulation.
Internalization has been postulated to play an important role in resensitization of GPCRs, enabling receptors to undergo dephosphorylation and subsequent recycling back to the cell surface (19). We show that the TXA 2 R␣ do not internalize upon agonist activation, contrary to the TXA 2 R␤. Intriguingly, it has been reported that the TXA 2 R␣ was subject to down-regulation while the TXA 2 R␤ was up-regulated, when CHO cells stably expressing TXA 2 Rs were stimulated for 24 h with agonist (18).
It is tempting to speculate that down-regulation of the TXA 2 R␣ is a mechanism to eliminate a receptor that cannot be resensitized, and that replenishment of the cell surface with functional TXA 2 R␣ would be achieved by other means, such as synthesis of new receptors. It will be interesting to evaluate these hypotheses in different cell types, and to correlate their physiological implications with the specific biological responses controlled by each receptor isoform as they become known.
In summary, our results demonstrate that the alternative splicing of the C-tail of the TXA 2 R generates isoforms of the receptor showing distinct internalization characteristics. The TXA 2 R␣ does not internalize after agonist stimulation, whereas TXA 2 R␤ internalization is dynamin-, GRK-, and arrestin-dependent. A region encompassing residues 355-366 of TXA 2 R␤ was shown to be essential for agonist-induced internalization, and possible receptor domains required for arrestin-3 interaction were mapped. We propose that agonist activation of the TXA 2 R␤ results in GRK-promoted phosphorylation of the receptor, increasing the affinity for arrestin binding, thereby targeting the receptor for clathrin-coated pit-mediated endocytosis. The TXA 2 R␣ and TXA 2 R␤ will constitute an interesting model for studying interactions with proteins of the endocytic machinery. Similar studies on other GPCRs with alternatively spliced C-tails, like the prostanoid EP3 (78) and FP (79) receptors, will be required to determine whether these findings also extend to other members of this receptor family.