Subtype-specific trafficking of endothelin receptors.

We investigated the subcellular localization of two endothelin receptors (ET(A)R and ET(B)R). To visualize these receptors directly, the C terminus of each receptor was fused to the N terminus of enhanced green fluorescent protein (designated as ETR-EGFP). When transiently expressed in various mammalian cell lines, ET(A)R-EGFP was predominantly localized on the plasma membrane. By contrast, ET(B)R-EGFP was, independent of ligand stimulation, predominantly localized on the intracellular vesicular structures containing Lamp-1. Immunoblot analyses revealed that at steady state ET(B)R-EGFP was highly degraded, and its degradation was inhibited by bafilomycin A(1). Antibody uptake experiments suggested that the ET(B)R-EGFP molecules were internalized from the plasma membrane. It is therefore likely that ET(B)R is first transported to the plasma membrane and then internalized, irrespective of ligand stimulation, to lysosomes where it undergoes proteolytic degradation. Exchanging the C-terminal cytoplasmic tails of the two ETRs revealed that the cytoplasmic tail is responsible for both the intracellular localization and the degradation of the receptors. Deletion of the extreme C-terminal 35 amino acids from both receptors allowed the receptor proteins to localize predominantly in the intracellular vesicles and to degrade. These observations indicate that the cytoplasmic tail of ET(A)R determines its plasma membrane localization. Stimulation with endothelin-1 increased the amount of intact ETR-EGFP fusion proteins without increasing their de novo synthesis, suggesting that binding of endothelin-1 stabilizes the ETRs.

receptor (ET A R) and endothelin-B receptor (ET B R), both of which belong to the G protein-coupled receptor (GPCR) superfamily (3,4). ET A R exhibits higher affinity for ET-1 and ET-2 than ET-3 (3), whereas ET B R accepts all three endothelins with similar affinity (4). ET A R is predominantly coupled to G q and G s , whereas ET B R is predominantly coupled to G q and G i (5). Although these two receptors share approximately 55% overall amino acid sequence identity (Fig. 1A), several domains are less homologous to each other and contribute to the functional differences between these receptors. For example, transmembrane domains IV-VI of the ET B R determine its specificity to certain ligands, such as ET-3 and IRL1620 (6). On the other hand, transmembrane domains I-III and VII determine the ability of ET A R to bind its specific ligands, such as BQ 123 (6). Cytoplasmic loop II is essential for ET A R to couple to G s , whereas cytoplasmic loop III is necessary for ET B R to couple to G i (7).
In addition to the N-terminal region of the third cytoplasmic loop, the C-terminal cytoplasmic tail is the most divergent intracellular region between the two receptors ( Fig. 1, A and B). It has been suggested that the C-terminal cytoplasmic tail is responsible for ligand-dependent internalization and/or desensitization of the GPCRs. In this regard, the most characterized GPCR is ␤ 2 -adrenergic receptor. When it is stimulated with its agonist, the serine and threonine residues in the cytoplasmic tail are phosphorylated by G protein-coupled receptor kinases and cAMP-dependent kinase, followed by its association with ␤-arrestins, resulting in uncoupling to G proteins or desensitization. ␤-Arrestins also play a crucial role in the ligand-dependent internalization of this receptor, mediated by clathrin-coated vesicles. The internalized receptors are dephosphorylated at endosomes, resulting in their resensitization and recycling back to the plasma membrane (8)(9)(10). Although several other GPCRs, such as protease-activated receptors, are also internalized in response to the cleavage in their N-terminal region and activation by proteases, they are sorted to lysosomes instead of recycling to the plasma membrane and undergo proteolytic degradation in order to terminate the irreversible activation of the signal transduction (11)(12)(13)(14). It is the cytoplasmic tail that regulates the trafficking of these receptors (13).
It was recently shown that both ET A R and ET B R are phosphorylated by G protein-coupled receptor kinases and rapidly desensitized in an identical manner (15). However, truncation of the cytoplasmic tails of both ET A R and ET B R (lacking the last 36 and 40 amino acids, respectively) does not affect ligandinduced desensitization (16 -18), suggesting that the cytoplasmic tail of both endothelin receptors is dispensable in ligandinduced desensitization. In addition, although ligand-induced association of ␤-arrestins and internalization were observed in the case of ET A R (19,20), the role of the cytoplasmic tail in internalization is still unknown. In the present study, we therefore investigated the role of the C-terminal cytoplasmic tail of these receptors in intracellular trafficking. For this purpose, the C terminus of each ETR was fused with the N terminus of enhanced green fluorescent protein (EGFP) and tran-siently transfected in cell lines. We found that ET B R is sorted to lysosomes and undergoes proteolytic degradation independent of ligand stimulation, probably due to the absence of the sequence necessary for its anchoring to the plasma membrane. We also found that prolonged stimulation with ET-1 allows for stabilization of both receptors.

EXPERIMENTAL PROCEDURES
Plasmid Construction-The restriction maps of rat ET A R and ET B R cDNAs are shown in Fig. 1C. All mutations (Fig. 1C, underlined), as well as truncation of the C-terminal region of the receptors, were introduced by polymerase chain reaction using cDNAs for rat ET A R (21) and ET B R (4) as templates. The polymerase chain reaction products were subcloned into pCR2.1 TOPO TA cloning vector (Invitrogen) and sequenced by the dideoxy termination method using an autosequencer (Lichor). A BamHI site was introduced just before the termination codon of each clone and was used to insert it into pEGFP-N3 (CLON-TECH) in frame. To exchange the C-terminal cytoplasmic tail, an AsnI site was introduced at Ile 364 -Asn 365 of ET A R corresponding to Ile 380 -Asn 381 of ET B R. To exchange the third cytoplasmic loop, an RcaI site was introduced at Leu 294 -Met 295 -Thr 296 of ET B R, which corresponds to Leu 277 -Met 278 -Thr 279 of ET A R, and an Mlu NI site was introduced at Val 304 -Ala 305 -Lys 306 of ET A R, which corresponds to Val 320 -Ala 321 -Lys 322 of ET B R. All cDNAs of each ETR-EGFP fusion protein were subcloned between XhoI and NotI sites of pME 18Sf(Ϫ) vector.
Cell Culture and Transfection-L tk Ϫ cells obtained from Riken Cell Bank (Tsukuba, Japan) were maintained in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, Inc.) supplemented with 10% fetal calf serum. Transfection was performed by the DEAE-dextran method. HeLa cells were maintained in DMEM supplemented with 10% fetal calf serum. Transfection was performed using FuGENE 6 transfection reagent (Roche Molecular Biochemicals). Rat clone 9 cells were maintained in DMEM/F-12 medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum. Transfection was performed using FuGENE 6 transfection reagent.
Immunofluorescence Analysis-L tk Ϫ cells seeded on a 6-cm dish (7.5 ϫ 10 5 cells) were transfected with each plasmid. Six hours after transfection, the cells were reseeded onto 8-well Lab-Tek II chamber slides (Nunc). The cells were cultured for 48 h and then fixed with 4% paraformaldehyde in PBS for 1 h at room temperature.
For lysosomal staining, rat clone 9 cells or HeLa cells seeded directly on Lab-Tek II chamber slides (5.0 ϫ 10 3 cell/well for clone 9 and 1.0 ϫ 10 4 cells/well for HeLa cells) were transfected with each plasmid. Fortyeight hours after transfection, cells were fixed with cold methanol, permeabilized with 0.1% Triton X-100 in PBS, and incubated with rabbit polyclonal anti-rat Lamp-1 antibody (kindly provided by Dr. Kenji Akasaki, Fukuyama University, Japan) for clone 9 or mouse monoclonal anti-human Lamp-1 antibody (PharMingen) for HeLa cells for 1 h at room temperature, followed by labeling with Cy3-conjugated goat antibodies against rabbit or mouse IgG, respectively (Jackson ImmunoResearch Laboratories) for 1 h at room temperature.
For analysis of endocytosis of ET B R, clone 9 cells or HeLa cells transfected with the ET B R-EGFP plasmid were incubated with 5 g/ml rabbit anti-peptide antibody against the N-terminal region of human ET B R (amino acids 27-42) (IBL) for 24 h simultaneously with transfection. After washing, fixing with 4% paraformaldehyde in PBS, and permeabilization with 0.1% Triton X-100 in PBS, cells were stained with Cy3-conjugated goat antibody against rabbit IgG.
All cells were observed with a TCS4D laser-scanning confocal microscope (Leica).
Immunoblotting-L tk Ϫ cells on 10-cm dishes (1.5 ϫ 10 6 cells) were transfected with each plasmid. Forty-eight hours after transfection, cells were collected and lysed with SDS-PAGE sample buffer (10 7 cells/ ml), then subjected to SDS-PAGE, followed by electrical transfer to a PVDF membrane. EGFP or EGFP fusion proteins were detected with polyclonal anti-GFP antibody (CLONTECH), horseradish peroxidaselabeled goat anti-rabbit IgG, and enhanced chemiluminescence reagents (Amersham Pharmacia Biotech).
Northern Blot Analysis-Transfected L tk Ϫ cells on 10-cm dishes (1.5 ϫ 10 6 cells/dish) were cultured for 48 h, and cellular total RNA was extracted with ISOGEN (Nippon Gene). The extracted RNA (10 g) was separated by formaldehyde/agarose gel electrophoresis, transferred to a hybond N Nylon membrane (Amersham Pharmacia Biotech), and hybridized with rat ET A R, rat ET B R, or EGFP cDNA fragment as probes. Each of the probes was labeled with [ 32 P]dCTP (Amersham Pharmacia Biotech) by the random priming method.
Binding Experiments-Cells were subjected to competitive radioli-gand binding assay as described (21). All assays were started at 48 h after transfection. Transfected cells were seeded in 24-well plates at a density of 1.

Difference in the Subcellular Localization between ET A R and
ET B R-To visualize the intracellular localization of two subtypes of the endothelin receptor, ET A R and ET B R, the C terminus of each receptor was fused to the N terminus of EGFP (designated as ET A R-EGFP and ET B R-EGFP). When expressed in L tk Ϫ cells, both ET A R-EGFP and ET B R-EGFP were transcribed at comparable levels to wild-type ET A R and ET B R, respectively (Fig. 2, A and B, lower panels, lane 1, 2, 7, and 8), and had virtually the same binding characteristics as ET A R and ET B R, respectively (Fig. 2, C and D, and Table I). In addition, measurement of intracellular Ca 2ϩ mobilization in the transfected L tk Ϫ cells revealed that ET A R-EGFP and ET B R-EGFP could activate the intracellular signal transduction system to essentially the same extent as ET A R and ET B R, respectively (Fig. 2, E and F). To normalize transfection efficiency, ␣ 1B adrenergic receptor was cotransfected, and the effects of ET-1 (100 nM) were estimated as percent of maximal responses obtained by stimulation with noradrenaline (10 M). The values of intracellular Ca 2ϩ increment in ET A R, ET A R-  When expressed in L tk Ϫ cells, the ET A R-EGFP fusion protein was localized predominantly on the plasma membrane (Fig. 3A). By contrast, the ET B R-EGFP fusion protein was localized not only on the plasma membrane but also on intracellular vesicular structures (Fig. 3B), even in the absence of ligand stimulation. Essentially the same localization pattern of each fusion protein was observed using rat Clone 9 cells or human HeLa cells as the host (Fig. 3, I-L), excluding the possibility that the observations are cell line-specific. We previously observed that human ET B R-EGFP is localized on the intracellular vesicles, as well as on the plasma membrane, 2 confirming that the observation is a general phenomenon for ET B R across species.
We then analyzed the molecular forms of the fusion proteins by immunoblotting of the lysates from L tk Ϫ cells transfected with these constructs using polyclonal anti-GFP antibody. No band was detected with the anti-GFP antibody in the lysate from cells transfected with either ET A R or ET B R (Fig. 4, lane 1 and lane 8, respectively), and only one band at a molecular mass of 27 kDa was detected in the lysate from cells transfected with EGFP (Fig. 4, lane 7), demonstrating the specificity of the anti-GFP antibody. Under these experimental conditions, heterogeneous bands of ET A R-EGFP were detected between 75 and 100 kDa (Fig. 4, lane 2). The bands probably resulted from the heterogeneity in the protein glycosylation. On the other hand, ET B R-EGFP was highly degraded to smaller molecular

cells (I and J) or HeLa cells (K and L) were also transfected with ET A R-EGFP (I and K) or ET B R-EGFP (J and L).
Forty-eight hours after transfection, all cells were fixed with 4% paraformaldehyde in PBS for 1 h at room temperature, and fluorescence of EGFP was detected by laser-scanning confocal microscopy. Each of the results was reproduced at least three times. were extracted and subjected to SDS-PAGE followed by electrical transfer to polyvinylidene difluoride membrane and detected with polyclonal rabbit anti-GFP antibody. Each of the results was reproduced at least three times. forms (Fig. 4, lane 9). The most intense band detected at a molecular mass of 27 kDa may be the EGFP portion itself, which was generated by intracellular cleavage of the fusion protein and accumulated due to its unusually long half-life (approximately 24 h).

FIG. 4. Immunoblotting of lysates from L tk ؊ cells transfected with ET A R, ET B R, or their derivatives. Lysates from L tk Ϫ cells transfected with ET
Cytoplasmic Tail of ET A R Is Responsible for Its Localization to the Plasma Membrane-ET A R and ET B R are highly conserved in the intracellular regions, except for the N terminus of the third cytoplasmic loop and the C-terminal cytoplasmic tail (Fig. 1, A and B). In the case of other GPCRs, corresponding regions have been shown to be implicated in their desensitization and internalization. Therefore, it is reasonable to speculate that the sequences that determine the difference between the two receptors in the subcellular localization are within these regions. We therefore made chimeric constructs of ETR-EGFP having either the third cytoplasmic loop or the cytoplasmic tail, or both regions, derived from the other ETR subtype. When expressed in L tk Ϫ cells, the levels of mRNA of these ET A R-EGFP and ET B R-EGFP derivatives were approximately the same as those of wild-type ETR-EGFP (Fig. 2, A and B,  respectively). In addition, as summarized in Table I and Fig. 2, E and F, the binding property and signal transduction ability of these chimeric receptors were essentially the same as those of the wild-type receptors. These data confirm that these substitutions do not affect the function of ET A R and ET B R. Exchange of the third cytoplasmic loop of one receptor subtype with that of the other (ET A R (B CL3 )-EGFP and ET B R (A CL3 )-EGFP) did not alter either their subcellular localization (Fig. 3, E and F) or stability (Fig. 4 lanes 4 and 11) as compared with the wild-type receptors (Fig. 3, A and B, and Fig. 4, lanes 2 and 9). By contrast, exchange of the cytoplasmic tail (ET A R (B CT )-EGFP and ET B R (A CT )-EGFP) led the chimeric receptors to characteristics that are completely different from the wild-type receptors as follows: ET A R (B CT )-EGFP was localized on both the plasma membrane and vesicular structures (Fig. 3C) and was highly degraded (Fig. 4, lane 3), whereas ET B R (A CT )-EGFP was localized predominantly on the plasma membrane (Fig.  3D) and not so degraded (Fig. 4, lane 10). We obtained essentially the same results by exchanging both the third intracellular loop and the cytoplasmic tail (Fig. 4, lanes 5 and 12 and data not shown).
These observations suggest two alternative possibilities for the role of the cytoplasmic tail of the ETRs. One is that ET B R has a determinant for internalization and degradation in its cytoplasmic tail. The other is that the cytoplasmic tail of ET A R has a sequence responsible for its anchoring to the plasma membrane. To discriminate between these possibilities, we deleted the cytoplasmic tail of each receptor (ET A R (⌬CT)-EGFP and ET B R (⌬CT)-EGFP). Because recent reports had demonstrated that several Cys residues proximal to the membrane within the cytoplasmic tail (amino acids 385-388 in ET A R and 401-404 in ET B R) are palmitoylated and necessary for signal transduction (16,(22)(23)(24), we truncated ET A R and ET B R from the C terminus up to Ser 391 and Thr 406 , respectively (Fig. 1, A  and B). The deletion of the cytoplasmic tail led to localization to intracellular vesicular structures (Fig. 3, G and H) and complete degradation (Fig. 4, lanes 6 and 13) of both receptors. These data indicate that ET A R has a signal responsible for its anchoring to the plasma membrane within its cytoplasmic tail.
ET B R-EGFP Is Delivered to Lysosomes and Degraded-Because ET B R-EGFP was localized in intracellular vesicular structures and was highly degraded, it is reasonable to speculate that the structures to which ET B R-EGFP is delivered are lysosomes. To confirm this speculation, rat Clone 9 cells and HeLa cells transiently expressing ET B R-EGFP were stained  3 and 4). Twenty four hours after transfection, cells were treated with (lanes 2 and 4) or without (lanes 1 and 3) bafilomycin A 1 (100 nM) for 12 h, and the cell lysate was subjected to SDS-PAGE followed by immunoblotting using polyclonal anti-GFP antibody. Each of the results was reproduced at least three times.
with antibodies against a lysosomal marker, lysosome-associated membrane protein-1 (Lamp-1). As shown in Fig. 5, the signal of ET B R-EGFP was superimposed almost completely on the staining for Lamp-1 in both cell lines. To corroborate this observation, L tk Ϫ cells transfected with ET B R-EGFP were treated with bafilomycin A 1 , which is known to inhibit specifically vacuolar type H ϩ -ATPase and thereby inhibit lysosomal proteases by raising the pH of intracellular acidic compartments, including lysosomes (25). As shown in Fig. 6, the degradation of ET B R-EGFP was significantly suppressed by the treatment of the transfected L tk Ϫ cells with bafilomycin A 1 , whereas ET A R-EGFP was not affected by the treatment (Fig. 6,  lanes 1 and 2). These results indicate that ET B R is sorted to lysosomes and undergoes proteolytic degradation.
We needed to determine whether ET B R-EGFP is delivered directly from the Golgi to lysosomes or whether there is a mechanism whereby this receptor, once delivered to the plasma membrane, is then internalized to lysosomes. To find out, we examined uptake of an antibody against the 16 amino acids in the extracellular N-terminal domain of ET B R in ET B R-EGFPtransfected Clone 9 or HeLa cells. Only the cells with the fluorescence of ET B R-EGFP were recognized by the antibody, confirming its specificity. 3 As shown in Fig. 7, the antibody was internalized to intracellular vesicular structures that overlapped, although not completely, with the labeling of ET B R-EGFP. These observations indicate that at least some fraction of the ET B R molecules are constitutively internalized from the plasma membrane to lysosomes.
Endothelin Receptors Are Stabilized by Prolonged Stimulation with ET-1-We also investigated the effects of prolonged stimulation with the ligand ET-1 on the internalization and degradation of ETRs. Expression levels of mRNA for ET A R-EGFP or ET B R-EGFP were not altered by the stimulation with ET-1 for 24 h (Fig. 8B). However, such prolonged stimulation led to a significant increase in the total amount of both ET A R-EGFP and ET B R-EGFP proteins (49 and 131%, respectively) (Fig. 8A, lanes 1 and 2 and 5 and 6, respectively), although ET A R-EGFP as well as ET B R-EGFP was delivered to lysosomes under these conditions (Fig. 8C). Especially in the case of ET B R-EGFP, the amount of ϳ90and ϳ65-kDa species increased markedly. In the absence of ET-1, treatment of cells with cycloheximide (10 M) for 24 h reduced the amount of both ET A R-EGFP and ET B R-EGFP (55 and 58%, respectively) (Fig.  8A, lanes 1 and 3 and 5 and 7, respectively). However, ET-1 increased the amount of both receptors, even in these cycloheximide-treated cells (110 and 64%, respectively) (Fig. 8A, lane 3  and 4, and 7 and 8). These results suggest that prolonged stimulation with ET-1 increases the intact ET A R and ET B R not by enhancing the expression of these proteins but by stabilizing them.

DISCUSSION
In the present study, we expressed EGFP-tagged ETRs in various cell lines and examined their subcellular localization. We found that, at steady state, ET A R was mainly localized on the plasma membrane, whereas ET B R was localized not only on the plasma membrane but also on lysosomes, where it appeared to be degraded by lysosomal proteases. Furthermore, an antibody uptake experiment suggested that at least some fraction of ET B R molecules are internalized from the plasma membrane to lysosomes. These observations indicate that ET B R, once delivered to the plasma membrane, then constitutively internalizes to lysosomes for degradation, although we cannot exclude the possibility that some fractions of ET B R molecules are directly sorted from the Golgi to lysosomes. Nevertheless, we favor the former possibility because experiments using chimeric and truncated constructs suggest the presence of a specific signal for anchoring of ET A R to the plasma membrane but not for direct sorting of ET B R from the Golgi to lysosomes; ET B R bearing the cytoplasmic tail of ET A R was mainly localized on the plasma membrane, and not only ET A R with the ET B R cytoplasmic tail but also both receptors lacking the cytoplasmic tail are mainly localized on lysosomes. Although the physiological significance of the unique behavior of ET B R is currently unknown, its fast turnover is in line with previous reports consistently showing that expression of ET B R mRNA is up-or down-regulated by various stimuli in many cell types and tissues (26 -35).
Many GPCRs, such as AT 1A , TSH, CCKBR, NK-1, neurotensin, GRP-R, ␦-opioid, SSTR3, SSTR5, H 2 , TXA 2 R␤, and CCR2B receptors (36 -47), are known to be impaired in their internalization by truncation of the cytoplasmic tail, although the consensus sequence for the internalization remains to be identified. By contrast, as is the case for -opioid receptor (48) and SSTR2 (49), both ETRs lacking the cytoplasmic tail undergo constitutive internalization. Therefore, the determinant for internalization appears to be within a region/regions other than C-terminal 35 amino acids. In the case of muscarinic acetylcholine receptors, the third cytoplasmic loop has been shown to contribute to their internalization (50,51). Moreover, in the case of ␤ 2 -adrenergic receptor, its internalization depends not only on appropriate interactions of multiple molecular determinants within the cytoplasmic regions, including the first and and HeLa cells (right panels) were transfected with ET B R-EGFP for 24 h. Five g/ml of rabbit anti-peptide antibody against the N terminus of human ET B R (amino acids [27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42] was added during the transfection. Cell-bound and internalized antibody was labeled with Cy3-conjugated goat anti-rabbit IgG. ET B R-EGFP is shown in green. Antibody against the N terminus of human ET B R labeled with Cy3-conjugated goat antibody against rabbit IgG is shown in red. Colocalization of ET B R-EGFP and anti-N terminus antibody appears as yellow. Each of the results was reproduced at least three times. second cytoplasmic loops and the cytoplasmic tail, but also on conformational determinants that may influence their orientation (52). However, our experiments showed that neither the third cytoplasmic loop nor the cytoplasmic tail contributes to the ETR internalization. Further experiments will be required to determine the region(s) responsible for the internalization of the ETRs.
In contrast to ET B R that constitutively internalized to lysosomes, ET A R is internalized to lysosomes in an agonist-dependent manner. It is therefore likely that ET A R has a sequence within its cytoplasmic tail that prevents the receptor from constitutive internalization. However, we cannot rule out the possibility that the cytoplasmic tail of ET B R also contains a sequence that decelerates the internalization. Many GPCRs are known to recycle between the plasma membrane and intracellular compartments (8 -10). In the case of V 2 vasopressin receptor, the signal responsible for the recycling has been shown to be within its cytoplasmic tail (53). Receptors without such a motif once internalized seem to be sorted to lysosome (13). Neither ET A R nor ET B R seem to possess such a recycling motif, because both stay in lysosomes, once internalized. When continuously exposed to ET-1, both ET A R and ET B R were stabilized although they were localized to lysosomes. Especially in the case of ET B R, which is highly degraded at steady state, the amount of ϳ90and ϳ65-kDa species was signifi-cantly increased by agonist stimulation. Previous studies have reported that, in various tissues, ET B R exists not only as a 52-kDa intact species but also as a 34-kDa species that is generated through cleavage in the N-terminal extracellular domain by a metalloprotease(s) (54 -56). These two species have been shown to be able to bind ligands with high affinity (54). Taking into account the molecular mass of the EGFP portion (27 kDa) and heterogeneity in glycosylation, it is likely that the 90-and 65-kDa species of ET B R-EGFP detected in our experiments correspond to the 52-and 34-kDa species, respectively, detected in tissues. These ET B R species have been shown to form a very stable complex with ET in vitro, even in the presence of high concentrations (up to 2%) of SDS (56). Chun et al. (19) reported that cell surface ET A R that binds ET-1 undergoes internalization and also forms a stable complex with ET-1 for at least 2 h and that ET A R and bound ET-1 are localized in caveolae (57). These may be implicated in the production of the long lasting responses evoked by ET in vivo, such as contraction of smooth muscles and elevation of blood pressure (1). Although we did not show the localization of endothelin receptors in caveolae, our observations on the agonistdependent stabilization of both ET A R and ET B R make it tempting to speculate that ET evokes such long lasting cellular responses, at least in part, through stabilizing its receptors.  1-4) or ET B R-EGFP (lanes [5][6][7][8]. Twenty four hours after transfection, cells were treated with (lanes 2, 4, 6, and 8) or without (lanes 1, 3, 5, and 7) ET-1 (100 nM) for 24 h. Cycloheximide (10 M) was treated (lanes 3, 4, 7, and 8) simultaneously with ET-1. Each of the results was reproduced at least three times. B, Northern blot analysis of the cells expressing each endothelin receptor. Total RNA extracted from L tk Ϫ cells transiently transfected with expression constructs of ET A R-EGFP (lanes 1 and 2) or ET B R EGFP (lanes 3 and 4). Twenty four hours after transfection, cells were stimulated with (lanes 2 and 4) or without (lanes 1 and 3) ET-1 (100 nM) for a further 24 h. Each of the results was reproduced at least three times. C, lysosomal localization of endothelin receptors in rat Clone 9 cells stimulated with ET-1. Cells were transiently transfected with ET A R-EGFP (left panels) or ET B R-EGFP (right panels). Twenty four hours after transfection, cells were stimulated with ET-1 (100 nM) for a further 24 h, followed by immunofluorescent staining as described for Fig. 5.