Role of the C Terminus in Histamine H2 Receptor Signaling, Desensitization, and Agonist-induced Internalization*

To evaluate the role of the histamine H2 receptor C terminus in signaling, desensitization, and agonist-induced internalization, canine H2 receptors with truncated C termini were generated. Wild-type (WT) and truncated receptors were tagged at their N termini with a hemagglutinin (HA) epitope and expressed in COS7 cells. Most of the C-terminal intracellular tail could be truncated (51 of 70 residues, termed T308mutant) without loss of functions: cAMP production, tiotidine binding, and plasma membrane targeting. In fact, the T308 mutant produced more cAMP than the WT when cell-surface expression per cell was equivalent. Pretreatment of cells with 10−5 m histamine desensitized cAMP productions via WT and T308 receptors to similar extents. Incubation of cells expressing WT receptors with 10−5 m histamine reduced cell-surface anti-HA antibody binding by approximately 30% (by 30 min, t 1/2 ∼ 15 min), but did not affect the B max of tiotidine in membrane fractions, which represents total receptor amounts, suggesting that WT receptors were internalized from the cell surface. In contrast, no internalization was observed with T308 receptors following histamine treatment. A mutant with a deletion of the 30 C-terminal amino acids, termed T329, was functional but was as potent as the WT in terms of cAMP production. Apart from being desensitized by histamine, the internalization of the receptor was indistinguishable from that of the WT. Internalization was observed in the T320 but not in T313 mutant, narrowing the region involved in internalization to that between Glu314 and Asn320 (ETSLRSN). Of these seven residues, either Thr315, Ser316, or both, were replaced with Ala. Thr315 and Ser316 are conserved among species. The mutation at Thr315 (but not that at Ser316) abolished internalization. Taken together, these results demonstrate that Thr315 is involved in agonist-induced internalization. Furthermore, the finding that T308 receptors were desensitized in the absence of internalization suggests that internalization and desensitization are meditated by independent mechanisms.

As observed in a number of guanine nucleotide-binding protein-coupled receptors (GPCRs), 1 cAMP responses occurring via the histamine H2 receptor are rapidly desensitized after agonist stimulation (1)(2)(3)(4)(5). The H2 receptor also exhibits an agonist-induced internalization from the cell surface (2,6). However, the mechanisms underlying these phenomena have yet to be identified. Recently, a number of studies focusing on GPCRs have examined the role of the C terminus in agonistinduced desensitization and internalization, with varying results. For example, C-terminal truncation of the angiotensin II receptor inhibited agonist-induced internalization of this receptor, but not desensitization of the calcium response mediated via the receptor (7). On the other hand, C-terminal truncation of ␤ 2 -adrenergic, ␣ 1B -adrenergic, lutropin/choriogonadotropin, platelet-activating factor, and neurokinin-2 receptors resulted in impairment of homologous desensitization (8 -12). Thus, the C termini of GPCRs have functional importances, which may differ among receptors. The present study was designed to analyze the role of this important portion of the histamine H2 receptor in signaling, desensitization, and agonist-induced internalization. To this end, we constructed H2 receptor cDNAs, devoid of either the 71 or the 51 amino acids at the C terminus, and expressed these cDNAs in COS7 cells. Herein, we present evidence that the 51 C-terminal amino acids of the histamine H2 receptor are important for agonist-induced internalization, but not for either signaling or homologous desensitization of cAMP response, which occurs via this receptor. In addition, these amino acids exert inhibitory effects on cAMP production via the H2 receptor.
replaced with Ala (A 315 , A 316 , and A 315,316 ), were also constructed. The HA epitope (YPYDVPDYA) was inserted into the extracellular N termini of the wild-type (WT) and mutant receptors. After confirming the sequences, the cDNAs encoding the WT histamine H2 receptor and its truncated mutants with or without the HA-tag were subcloned into pCAGGS. COS7 cells were transfected with varying amounts of plasmid DNAs by the DEAE-dextran method.
Immunocytochemical Localization of the H2 Receptor in COS7 Cells-For immunofluorescence microscopy, COS7 cells were fixed in 3% formaldehyde/phosphate-buffered saline (PBS), scraped off the dish with a rubber blade, and embedded in 10% gelatin/PBS. Semithin frozen sections (1 m thick) were cut and incubated with anti-HA antibody or antibody against the C terminus of the canine H2 receptor (anti-H2R CT antibody) (2). These sections were then incubated with rhodamine-labeled affinity-purified goat anti-mouse or anti-rabbit IgG (Jackson Immunoresearch, West Grove, PA) (13).
Quantification of Cell-surface H2 Receptors-At 24 h post-transfection, COS7 cells expressing the WT and truncated H2 receptors with the HA epitope or parental COS7 cells were plated onto 24-well plates at a density of 10 5 cells/well. At 48 h post-transfection, cells were incubated in Hepes-tyrode's buffer (140 mM NaCl, 2.7 mM KCl, 1.8 mM CaCl 2 , 0.49 mM MgCl 2 , 0.37 mM NaH 2 PO 4 , 5.6 mM glucose, 25 mM Hepes, pH 7.4) containing 0.1% bovine serum albumin (BSA) at 37°C for 30 min. After incubation in the presence or absence of 10 Ϫ5 M histamine at 37°C for the indicated times, the cells were fixed with PBS containing 3% paraformaldehyde, washed again with PBS, and incubated with 5% skim milk/PBS for 1 h at room temperature. They were incubated with 200 l of 12CA5 (5 g/ml) in PBS for 2 h at room temperature, washed again with PBS, and incubated for another 1 h with 200 l of 125 I-labeled goat anti-mouse IgG (1:200 dilution) at room temperature. The wells were then washed twice with 5% skim milk/ PBS, twice with PBS, and three times with 0.05% Tween 20/PBS. Bound 125 I-labeled goat anti-mouse IgG was solubilized in 1% SDS, and radioactivities were determined in a ␥-counter. Specific bindings were determined by subtraction of the nonspecific bindings observed in parental COS7 cells.
Tiotidine Binding Assay-Tiotidine binding assays involving intact cells were performed as described previously (14). COS7 cells, grown in 24-well plates, were assayed at a density of 1 ϫ 10 5 cells/well. The cells were incubated for 2 h at 37°C in 200 l of Hepes-Tyrode's buffer containing 1 nM [ 3 H]tiotidine and various concentrations of unlabeled tiotidine (15). All samples were analyzed in triplicate. After incubation, cells were washed three times with ice-cold PBS, then removed from the wells in 0.1% SDS, and radioactivities were determined by liquid scintillation counting. Tiotidine bindings in membranes were assessed as described previously, with some modifications (2). Membrane fractions (200 g) from COS7 cells were incubated with 1 nM [ 3 H]tiotidine and varying concentrations of unlabeled tiotidine in 25 mM Hepes, 0.1% BSA, pH 7.4, in a final volume of 200 l at 37°C for 2 h. The binding reaction was terminated by filtration over Whatman GF/C glass fiber filters, followed by 10-ml washes with ice-cold incubation buffer. Radioactivity on the filters was determined by liquid scintillation counting. In both experiments, specific bindings were calculated by subtraction of the nonspecific bindings determined in the presence of 10 Ϫ4 M cimetidine. No specific binding was observed in either parental COS7 cells or COS7 cells transfected with the expression vector alone.
Measurement of cAMP Production-COS7 cells, plated onto 24-well plates, were assayed at a density of 1 ϫ 10 5 cells/well, as described previously (14). The cells were incubated for 30 min at 37°C in 450 l of Hepes-Tyrode's buffer containing 0.1% BSA and 0.1 mM 3-isobutyl-1-methylxanthine, after which 50 l of histamine solution was added to initiate the reaction. After 10 min of incubation at 37°C, the reaction was terminated by the addition of 500 l of 12% trichloroacetic acid. The samples were centrifuged for 5 min at 3000 ϫ g at 4°C. Following extraction of the supernatants three times with diethylether, cAMP contents in the samples were measured using a radioimmunoassay (16). Histamine-dependent and forskolin-dependent cAMP productions were determined by subtracting basal cAMP productions.

RESULTS
Expression of Wild-type and C Terminus-truncated Histamine H2 Receptors-To investigate the role of the cytoplasmic tail of the histamine H2 receptor in signaling and desensitization, we generated C terminus-truncated receptor cDNAs. Fig.  1 presents a schematic representation of the 359-amino acid canine histamine H2 receptor. Consensus sites for post-translational modifications, N-glycosylation, and palmitoylation are also indicated (17). We chose to delete 51 and 70 amino acids from the C-terminal tail, yielding two distinct mutant receptors, termed T 308 and T 289 , lacking 11 and 13 serine/threonine residues, respectively. The T 289 receptor was truncated at the beginning of the cytoplasmic C terminus and the T 308 receptor just distal to the putative palmitoylation site (Cys 305 ). All receptor constructs were tagged at their N termini with a 9-amino acid HA epitope, as shown in Fig. 1. As the HA-epitope is located in the extracellular region, the cell-surface receptor amount can be estimated utilizing the antibody against the HA-epitope, 12CA5. The cDNA constructs were subcloned into an expression vector, pCAGGS and transfected into COS7 cells. COS7 cells do not have detectable endogenous histamine H2 receptors, as demonstrated by the absence of histamine-dependent cAMP production and the absence of specific [ 3 H]tiotidine binding (data not shown), and are thus a suitable model system for these studies. At 48 h after transfection, the cells were subjected to immunocytochemistry using the anti-HA antibody or the anti-H2R CT antibody. Expression of the receptors is shown in Fig. 2. The WT, the HA-wild-type (HA-WT), and the HA-T 308 receptors were distributed in the plasma membranes, while the HA-T 289 receptor, the 70 C-terminal deleted amino acids of which include Cys 305 , a consensus site for palmitoylation, was distributed intracellularly as well as in the plasma membranes. Whether the different distribution of the HA-T 289 receptor was due to the loss of palmitoylation or to lack of the proximal amino acid residues from 290 to 307 was not determined. In either case, it can be concluded that the 51 C-terminal amino acid residues of the histamine H2 receptor, a large part of the C terminus, are not involved in trafficking of the H2 receptor to the plasma membrane. Similar results were obtained when various amounts of plasmid DNAs were used for transfection of COS7 cells (data not shown), indicating that the different distribution of the receptors is not a function of receptor numbers.
Tiotidine Binding of the WT and Truncated Receptors-To examine the effect of C-terminal truncation on ligand binding, intact COS7 cells on 24-well plates were subjected to binding of tiotidine (15), an H2 receptor antagonist, at 48 h post-transfection. The affinities of the WT and truncated receptors for tiotidine were determined by Scatchard plot analysis. K d values for tiotidine of the WT and T 308 receptors were comparable (Table  I). In addition, HA-tagging did not affect the affinity of these receptors for tiotidine (Table I) receptor induced by the deletion of amino acid residues from 290 to 307. Since tiotidine is membrane-permeable and is capable of binding to receptors distributed intracellularly, the inability of the T 289 mutant to bind to tiotidine is not due to the difference in the intracellular distribution of the receptor shown in Fig. 2.
cAMP Production via the WT and Truncated Receptors-To examine the role of the C terminus in H2 receptor signaling, we measured histamine-dependent cAMP productions via the WT and truncated receptors in COS7 cells. cAMP productions mediated by these receptors can be compared only when the amounts of cell-surface receptors are equivalent. Iida-Klein et al. (19) reported that in a COS7 transfection system using the DEAE-dextran method the amount of plasmid DNA used for transfection determined the number of cell-surface parathyroid hormone/parathyroid hormone-related protein receptors expressed per cell but not the percentage of transfected cells. We have obtained similar results, in that the amounts of plasmid DNA affected the number of H2 receptors expressed per cell and the percentage of transfected cells were constant even when the amount of plasmid DNA used was not equivalent (Table II). We transfected various amounts of plasmid DNA for HA-WT, HA-T 308 , and HA-T 289 receptors and measured the number of receptors expressed per cell. As shown in Fig. 3A, the amounts of cell-surface receptors were dependent on the amounts of plasmid DNA used for transfection and, in addition, differed markedly among the WT and mutant receptors even when equivalent amounts (molar) of plasmid DNAs were used for transfection. The number of cell-surface HA-WT receptors expressed was larger than that of HA-T 308 receptors when the equivalent amounts of plasmid DNAs were used for transfection. It is essential to study the functions of these receptors at similar expression level. Therefore, we measured histamine-dependent cAMP productions via the WT and mutant receptors at each transfection levels. This allowed comparison of the functions of these receptors. As shown in Fig. 3B, histamine-dependent cAMP productions were observed in HA-WT and HA-T 308 cells but not in HA-T 289 cells. Histamine-dependent productions of cAMP observed in HA-WT and HA-T 308 were inhibited by cimetidine but not by diphenhydramine (data not shown). Interestingly, as compared with HA-WT receptors, HA-T 308 receptors produced more cAMP than the WT receptor (Fig. 3B) if the numbers of cell-surface receptors per cell were equivalent. Therefore, the 51 C-terminal amino acids are not only essential for, but rather may actually inhibit, cAMP production mediated by the H2 receptor. It is also possible that inhibitory effects on the H2 receptor are exerted via the Cterminal region. Although less marked than those in HA-WT and HA-T 308 cells, specific cell-surface anti-HA antibody bindings were also observed in HA-T 289 cells (Fig. 3A). Thus, the absence of histamine-dependent cAMP production via T 289 re-    ceptors was not due to a lack of cell-surface T 289 receptors, but rather to the inability of the receptor itself to couple to G s . It is not clear whether this is due to the absence of the amino acid residues from 289 to 307, or to a conformational change induced by the deletion. Furthermore, the possible contributions of these amino acid residues to cAMP production could not be examined.
Desensitization of WT and T 308 Receptor-mediated cAMP Responses-The histamine H2 receptor reportedly shows homologous desensitization of the cAMP response following preincubation with histamine. We examined whether the C-terminal tail of the H2 receptor is involved in desensitization of the cAMP response, which occurs via this receptor. Since COS7 cells expressing the HA-T 289 mutant did not mediate cAMP production in response to histamine (Fig. 3B), this mutant was omitted from the experiment. COS7 cells were transfected with plasmid DNAs containing HA-WT (0.4 g/10-cm plate) or HA-T 308 (1.0 g/10-cm plate) receptors. Transfection with these DNA amounts resulted in similar expression levels of the HA-WT and HA-T 308 receptors (Fig. 3A). Preincubation of transfected COS7 cells with 10 Ϫ5 M histamine for 30 min led to reduced cAMP production in both the HA-WT and HA-T 308 cells as compared with untreated cells (Fig. 4A), indicating desensitization, while forskolin-dependent cAMP production was slightly increased (Fig. 4B). It is noteworthy that the transfection efficiency of the COS7 cell system is approximately 20% at best (Table II). Furthermore, the results shown in Fig.  4B do not necessarily mean that forskolin-dependent cAMP production in COS7 cells expressing the receptors was not altered. Therefore, we performed similar experiments in CHO cells stably expressing HA-WT or HA-T 308 receptors and found that histamine-dependent cAMP production via these receptors was reduced by preincubation with 10 Ϫ5 M histamine, while forskolin-dependent cAMP production was not (data not shown). Similar results were obtained following a 15-min preincubation with 10 Ϫ5 M histamine (data not shown). Taken together, these results suggest that the 51 amino acids of the C terminus of the histamine H2 receptor are not essential for desensitization of the receptor-mediated cAMP response.

Effect of C-terminal Truncation on Agonist-induced Internalization of the H2 Receptor-A number of GPCRs have been reported to be internalized into compartments inaccessible to
agonists acting at the plasma membrane. Agonist-induced internalizations of several other GPCRs have been detected by measuring the acid-resistant forms of radiolabeled agonists (7, 20 -23). However, this experiment cannot be performed for the H2 receptor because of the high levels of nonspecific uptake of histamine by cells. To circumvent this difficulty, we utilized the HA-tag, which is located at the N-terminal extracellular region, to quantify the cell-surface H2 receptor amount. Specific cellsurface bindings of anti-HA antibody were observed in COS7 cells expressing HA-WT and HA-T 308 receptors (Fig. 3A), but not in parental or mock transfected COS7 cells (data not shown). We examined the effect of incubation with 10 Ϫ5 M histamine on the cell-surface H2 receptor amount. As shown in Fig. 5A, in COS7 cells expressing HA-WT receptors the amount of cell-surface anti-HA antibody binding decreased by approximately 30% with a 30-min incubation with 10 Ϫ5 M histamine and the t 1/2 was approximately 15 min. However, neither the binding maximum (B max ) of tiotidine in membrane fractions (Fig. 5B), which represents the total receptor amount, nor the K d for tiotidine (data not shown) was decreased. These findings indicate that HA-WT receptors were internalized from the cell surface. Immunocytochemical examination after a 30-min incubation with histamine revealed the HA-WT receptors to be on the plasma membrane, possibly indicating that the receptors were in coated pits or caveola. In contrast, incubation of COS7 cells expressing HA-T 308 receptors with histamine affected neither cell-surface anti-HA antibody binding (Fig. 5A) nor the B max of tiotidine (Fig. 5B), indicating that the 51 Cterminal amino acids of the H2 receptor play a role in agonistinduced internalization of the receptor.
Effects of Truncation of 15 and 30 C-terminal Amino Acids on H2 Receptor Signaling, Desensitization, and Internalization-To define more clearly the region involved in internalization, two mutant receptors, T 329 and T 344 , with deletion of 30 and 15 C-terminal amino acids, respectively, were generated. These mutants were also tagged with the HA-epitope at their N termini and were expressed in COS7 cells. Both receptors were capable of binding tiotidine with affinities similar to that of the WT receptor (data not shown). In contrast to the finding in the HA-T 308 mutant receptor, cAMP productions via both receptors were comparable to those mediated by the HA-WT receptor provided that the amounts of cell-surface receptors per cell were equivalent (Fig. 6A). These observations indicate that these 30 amino acids of the C terminus are not involved in inhibition of cAMP production. In addition, cAMP productions via these receptors were desensitized by 10 Ϫ5 M histamine preincubation (data not shown). Finally, upon histamine exposure, HA-T 329 and HA-T 344 receptors showed internalization indistinguishable from that of the HA-WT receptor (Fig. 6B).
Identification of the Amino Acid Residues Involved in Internalization of the H2 Receptor-The results obtained indicate that the region responsible for agonist-induced internalization and inhibition of cAMP production resides between amino acid residues 308 and 328. To identify the amino acid residues involved in agonist-induced internalization, we generated additional mutants, HA-T 313 and HA-T 320 , with deletions of 39 and 46 C-terminal amino acids, respectively. Interestingly, agonist-induced internalization was observed in the HA-T 320 , but not in the HA-T 313 receptor (data not shown). Thus, the amino acid residues involved in agonist-induced internalization of the histamine H2 receptor are likely to reside between Glu 314 and Asn 320 (ETSLRSN). Of these seven amino acid residues, assuming that the hypothesis that serine and/or threonine phosphorylation is involved in agonist-induced internalization of the H2 receptor is correct, the likely candidates are Thr 315 and Ser 316 . This is because these two residues are conserved among species (Table III). To determine whether these amino acid residues are involved in agonist-induced internalization, we generated three mutant receptors, HA-A 315 , HA-A 316 , and HA-A 315,316 , in which either Thr 315 , Ser 316 , or both were replaced with Ala. As shown in Fig. 7, the HA-A 316 receptor internalization was indistinguishable from that of the WT, whereas those of the HA-A 315,316 and HA-A 315 were not. This finding clearly demonstrates that Thr 315 is involved in agonist-induced internalization of the histamine H2 receptor and that the inabilities of the HA-T 308 and HA-T 313 mutants to internalize were not due to nonspecific effects induced by the truncations.  The histamine H2 receptor plays a central role in acid production in gastric parietal cells via production of cAMP (24). Inappropriate gastric acid production can lead to the generation or exacerbation of peptic ulcer disorders. Tight control of signaling via the H2 receptor is thus vital for normal gastric function. It has been accepted that the histamine H2 receptor undergoes rapid desensitization of the cAMP response upon exposure to an agonist (1)(2)(3)(4)(5). Although homologous desensitization of agonist-induced signaling via GPCRs is an essentially ubiquitous phenomenon, the functional roles of the C termini vary among receptors. In a number of GPCRs, agonist-induced serine/threonine phosphorylation of the C terminus has been implicated in receptor desensitization (8 -10), while in others the C terminus has been suggested to be involved in agonistinduced internalization, but not in desensitization (20,25,26). To understand the as yet unidentified roles of the C terminus in the histamine H2 receptor, we generated two mutant receptors, T 289 and T 308 , which were truncated at the beginning of the cytoplasmic C-terminal portion and just distal to putative palmitoylation site Cys 305 , respectively. The WT and mutant receptors were tagged at their N termini with the HA epitope and expressed in COS7 cells.
One of the primary observations made in our present study was that the T 308 receptor with a truncation of 51 amino acids located distal to the putative palmitoylation site was desensitized upon histamine exposure. It is hypothesized that serine/ threonine phosphorylation is involved in agonist-induced desensitization of GPCRs. The T 308 mutant, which is missing 11 serine/threonine residues present in the C terminus of the histamine H2 receptor, did not show altered agonist-induced uncoupling of the H2 receptor from adenylyl cyclase. Therefore, it is likely that the serine/threonine residues, which may be phosphorylated as part of the desensitization process, reside in other intracellular regions or the proximal C-terminal residues, namely Thr 290 and Thr 297 . Possible contributions of this proximal carboxyl region could not be examined because T 289 receptors, which are devoid of all C-terminal amino acid residues, were nonfunctional (Table I, Fig. 3B). It is noteworthy that the lack of T 289 receptor activity is not due to the absence of cell-surface expression but rather to the receptor itself being devoid of functional activity (Fig. 3).
Second, the WT receptor was internalized upon histamine exposure while the T 308 mutant was not. In addition, the T 329 mutant with a deletion of 30 amino acids was internalized. Thus, the region involved in internalization of the histamine H2 receptor is present within the area including amino acid residues from 308 to 328. The observation that the T 320 , but not the T 313 receptor, was internalized further confined the amino acid residues involved in agonist-induced internalization. Finally, the finding that internalization occurred in the HA-A 316 receptor, but in neither the HA-A 315 nor the HA-A 315,316 receptor, identified the amino acid residue involved in agonist-induced internalization of the H2 receptor as Thr 315 . Our preliminary observations indicated that both the HA-WT and the HA-T 308 receptors were phosphorylated after histamine stimulation (data not shown). Thus, although we can speculate that phosphorylation does actually play a part in desensitization, we cannot determine whether it is involved in agonist-induced internalization of the H2 receptor. Recently, Barak et al. (27) reported that a tyrosine residue highly conserved in G-proteincoupled receptors and in the NPXXY motif is involved in sequestration of the ␤ 2 -adrenergic receptor. This tyrosine residue is conserved in the histamine H2 receptor (Tyr 288 ) (17,28,29). The T 308 receptor contains this tyrosine, such that the tyrosine residue Tyr 288 appears to be necessary for, but is not by itself sufficient to produce, internalization of the histamine H2 receptor. Furthermore, desensitization of the cAMP response occurring via the T 308 receptor, which was not internalized, indicates that internalization per se is not essential for agonistinduced desensitization of the cAMP response and that desensitization and internalization are mediated by independent mechanisms. Third, we have shown the T 308 receptor to be more active than the WT receptor in terms of adenylyl cyclase stimulation. In contrast, the T 329 and the T 344 receptors showed essentially the same levels of activity as the WT receptor. Thus, the 51 C-terminal amino acids of the histamine H2 receptor contain sequences that suppress the interaction of this receptor with G-proteins, and these sequences apparently lie within the region from amino acids 308 to 329. When these sequences are removed a more productive interaction takes place, and hormonal responsiveness upon truncation is enhanced. As desensitization still took place in the T 308 receptor, this inhibitory effect is unlikely to be related to desensitization.
In conclusion, we have shown in COS7 cells that the C terminus of the histamine H2 receptor is involved in agonistinduced internalization, but not in desensitization. Furthermore, the C terminus is likely to exert an inhibitory effect on cAMP production via the receptor. Since these experiments were performed using fibroblast cell lines transfected with histamine H2 receptor cDNAs, the results presented show that the processes involved in histamine-induced uncoupling of the histamine H2 receptor are not specific to gastric parietal cells. However, the extent of internalization observed in this system is minimal compared with those of other receptors. We cannot rule out the possibility that some other cofactor, which is not present in COS7, might function in histamine H2 receptor internalization. Future studies must be designed to identify the specific amino acid residues involved in agonist-induced desensitization.