Bradykinin-induced Internalization of the Human B2Receptor Requires Phosphorylation of Three Serine and Two Threonine Residues at Its Carboxyl Tail*

The binding of bradykinin (BK) to B2 receptor triggers the internalization of the agonist-receptor complex. To investigate the mechanisms and the receptor structures involved in this fundamental process of receptor regulation, the human B2 receptor was mutated within its cytoplasmic tail by complementary strategies of truncation, deletion, and amino acid substitution. Ligand binding, signal transduction, internalization as well as phosphorylation were studied for the mutated receptors expressed in COS, CHO, and HEK 293 cells. Truncation of 44 out of 55 amino acid residues of the receptor’s cytoplasmic tail corresponding to positions 321–364 did not alter the kinetics of BK binding and the receptor coupling to phospholipase C and phospholipase A2. By contrast, truncations after positions 320 and 334, deletions within the segment covering positions 335–351, as well as alanine substitution of serine and threonine residues within segment 335–351 diminished the internalization capacity of the mutant receptors. Mutants with a markedly reduced internalization potential failed to produce BK-induced receptor phosphorylation suggesting that phosphorylation may be involved in receptor internalization. The mutagenesis approaches converged at the conclusion that three serines in positions 339, 346, and 348 and two threonines in positions 342 and 345, contained in a sequence segment that is highly conserved between species, have a critical role in the ligand-dependent internalization and phosphorylation of kinin receptors and can intervene in these processes in an alternative manner. However, mutants lacking these residues were still sensitive to dominant-negative forms of β-arrestin and dynamin, suggesting the existence of additional receptor structure(s) involved in the receptor sequestration through clathrin-coated vesicles.

The capacity of most G protein-coupled receptors (GPCRs) 1 to respond to agonists becomes altered after previous exposure to their ligands, a phenomenon referred to as desensitization. Desensitization can be a consequence of receptor endocytosis (1) and may also result from receptor phosphorylation that prevents further activation of G proteins (2,3). Negative cooperativity in bradykinin (BK) binding due to receptor-receptor interaction in the cell membrane can also contribute to the desensitization of some receptors including the BK B 2 receptor (4). For few GPCRs, internalization appears to be a prerequisite for resensitization prior to receptor recycling to the plasma membrane (5). Internalization is believed to involve clathrincoated vesicles and/or caveolin-rich vesicles, and to result from an interaction of components of the endocytic machinery with specific motifs located in the cytoplasmic domains of the receptor (6). For many receptors, the carboxyl tail has been found to play a critical role in receptor internalization (2,3,7) but other critical motifs have been identified in the second or third intracellular loops (8,9) as well as in the seventh transmembrane domain (10).
The molecular mechanisms that trigger internalization of the agonist-occupied receptor have been studied for only a few receptors. Phosphorylation has been implicated in the desensitization process uncoupling the receptor from G proteins, and in the initiation of internalization sequestering the receptor from the cell surface (2,3). Indeed, mutations of potential phosphorylation sites of the GLP-1, muscarinic m 3 , and the cholecystokinin receptors have been shown to reduce their internalization (11)(12)(13). However, receptor internalization and desensitization are not always causally linked, e.g. mutations of certain GPCRs suppressed desensitization but not internalization (14,15). Conversely mutations in some GPCRs altered internalization but not desensitization (10,15). Clearly, the two processes can proceed independently although they may involve the same type of post-translational modifications such as phosphorylation.
Prototypical GPCRs that differ by their ligand-induced densensitization, phosphorylation, and internalization are the receptors for the vasoactive kinin peptides. Two subtypes of mammalian kinin receptors, B 1 and B 2 , have been recognized so far (16 -18). The B 2 receptor is responsible for most of the physiological actions of BK including vasodilation (19) through activation of G proteins (20) that stimulate the activity of phospholipase C (PLC) and phospholipase A 2 (PLA 2 ) and increase the cytoplasmic calcium concentration, [Ca 2ϩ ] i (21). More recently, other signaling pathways such as the mitogen-activated protein (MAP) kinase pathway were found to be triggered via the B 2 receptor (22,23). BK-induced B 2 receptor internalization has been reported in several cell systems including cultured human fibroblasts HF-15 (24,25) and CHO-K1 cells transfected with the human B 2 receptor cDNA (4), and we have recently reported that in HF-15 cells BK induces the phosphorylation of serine and threonine residues located in the COOH terminus of the B 2 receptor (21). The time course of BK-stimulated phosphorylation paralleled the kinetics of desensitization/resensitization and of internalization/recycling of the receptor, suggesting that these phenomena may be interrelated (21).
The present work was aimed at studying the role of the COOH-terminal tail of the human renal B 2 bradykinin receptor (referred as wild-type, B 2 wt) in BK binding, coupling to signaling pathways, internalization, and phosphorylation. For this purpose, we generated a series of human B 2 receptor mutants and transfected them into COS-7 cells, CHO-K1, and HEK 293 cells. Three complementary strategies were followed for mutagenesis, by generating mutants with truncated COOH-terminal tail, mutants with deletion of internal regions of the COOH-terminal tail, and mutants with serine and/or threonine residue(s) replaced by alanine. This allowed us to demonstrate that a cluster of three serine and two threonine residues located in the center portion of the COOH-terminal tail region that becomes phosphorylated in response to BK is involved in the internalization of the B 2 receptor. Co-transfection of the B 2 wt and its mutants with ␤-arrestin and dynamin mutants documented a role of clathrin-coated pit pathway in receptor internalization.

EXPERIMENTAL PROCEDURES
Construction of the Mutant Receptor cDNAs-Mutant cDNAs were constructed by using the previously cloned B 2 wt cDNA placed under the control of cytomegalovirus promoter into the eucaryotic expression vector pcDNA 3 (Invitrogen, Leek, Netherlands) (4) as a template in sitedirected mutagenesis using the Transformer TM Site-directed Mutagenesis Kit (2 nd version, CLONTECH, Palo Alto, CA). All mutations were sequenced using an Amplicycle TM sequencing kit (Perkin Elmer, Langen, Germany). Three truncated B 2 mutants, tR351, tI334, and tY320 were created by single nucleotide substitutions (indicated by bold face in the sequences given below) at the appropriate sites so as to create stop codons (underlined) terminating translation at positions 351, 334, and 320, respectively, of the protein sequence (numbering according to Hess et al. (17)). The nucleotide sequences of the oligonucleotide primers used were: 5Ј-CGTGGAACGCTAGATTCACAAACTG-3Ј for the mutant tR351, 5Ј-CAGAACCCATTTAGATGGAGAACTCC-3Ј for tI334, and 5Ј-GGAGGTGTACTAGGGAGTGTG-3Ј for tY320. Hence deletion mutants tR351, tI334, and tY320 lacked the terminal 13, 30, and 44 amino acid residues, respectively, of the human B 2 receptor. Deletions of regions within the COOH-terminal tail of the B 2 receptor were obtained by mutagenesis with the following oligonucleotide primers:  17,8, and 9 amino acid residues, respectively; their relative positions within the protein sequence are given in brackets. Point mutations were created by the same procedure to replace Ser and/or Thr residues in the COOH-terminal tail by Ala using the following oligonucleotide primers: 5Ј-GGAGATCCAGGCCGAGAGGAGG-3Ј  for the mutant T237A, 5Ј-CGAAAGAAGGCTTGGGAGGTG-3Ј for  S316A, 5Ј-GGCTGCAGGGCAGAACCC-3Ј for S331A, 5Ј-GATGGAGA-ACGCCATGGGCAC-3Ј for S339A, 5Ј-CTCCATGGGCGCACTGCGGA-C-3Ј for T342A, 5Ј-GCACACTGCGGGCCGCCATCTCCGTG-3Ј for  T345A/S346A, 5Ј-GACCTCCATCGCCGTGGAAC-3Ј for S348A, 5Ј-GG-AGAACGCCATGGGCGCACTGCGG-3Ј for S339A/T342A and 5Ј-GAA For immunoprecipitation and phosphorylation studies, transient  receptor expression was achieved by transfecting COS-7 cells grown in  6-well plates using the LipofectAMINE TM method (Life Technologies). Otherwise, transfections were performed with cells grown in T75 flasks (COS-7 cells) by using the DEAE-dextran method (26) and the cells were then subcultured into 24-well plates where binding, internalization, PLC, and PLA 2 activation experiments were performed or with cells directly grown in 24-well plates (HEK 293 and CHO-K1 cells) by using Superfect Transfection Reagent (Qiagen, Courtaboeuf, France). CHO-K1 cell clones that were selected with geneticin (0.75 g/ml) for stable expression and with [ 3 H]BK binding and coupling assays for receptor expression, were expanded to obtain the CHO-K1 cell lines used in some experiments. All cell types were maintained at 37°C in a humidified water-jacketed incubator with 5% CO 2 . Functional studies of the transfected cells were done at confluence. Control cells were created by transfecting the pcDNA 3 vector without insert.
Radioligand Binding and Internalization of [ 3 H]BK-Cells were incubated as described previously (4) (4). Cells were washed with HBSS before cell surfacebound radioligand was separated from internalized radioactivity by an acidic washing step (27) with 0.2 M acetic acid, 0.5 M NaCl, pH 2.5. To study the role of protein kinase C (PKC) activity, cells were preincubated with 10 Ϫ7 M phorbol 12-myristate 13-acetate (PMA) or 10 Ϫ7 M staurosporine (Sigma, St-Quentin-Fallavier, France) for 30 min before and during the incubation with [ 3 H]BK. To monitor the effect of phosphatase inhibitor, cells were incubated with 10 Ϫ7 M okadaic acid (Sigma) 5 min prior to [ 3 H]BK challenge, and internalization was followed for 10 min in the continuous presence of the phosphatase inhibitor. In some internalization experiments, cells were treated with hypertonic saline solution containing 0.4 M sucrose to study the contribution of clathrin-coated vesicle pathway (28). In addition, internalization was examined in cells co-transfected with the receptor and with rat ␤-arrestin cDNA or its inactive fragment (319 -418) mutant in pEGFP obtained from Bunnett and co-workers (29) and pcDNA 3 plasmids encoding for bovine ␤-arrestin fragment (319 -418) mutant and inactive human dynamin K44A both obtained from Benovic and co-workers (30). Co-transfection was done using 15 g of each plasmid. For determination of specific binding, each assay included measurements of nonspecific binding in the presence of 1000-fold excess of unlabeled BK; nonspecific binding was substracted from total binding determined in the absence of unlabeled BK. In each culture plate, the protein content was determined according to the manufacturer's instructions in three wells using BSA as the standard protein (Bio-Rad, Mü nchen, Germany).
Inositol Phosphate Production-Cells were loaded for 18 h at 37°C with 3 Ci/ml myo-[2-3 H]inositol (10 -20 Ci/mmol, Amersham International) added to the culture medium. The cells were washed twice with HBSS, preincubated for 10 min at 37°C with 10 mM LiCl in HBSS containing 0.1% BSA and the protease inhibitors used for binding assay, and stimulated for 15 min with BK at varying concentrations (10 Ϫ10 to 10 Ϫ7 M). To study the role of PKC, cells were treated with PMA (10 Ϫ7 M) or staurosporine (10 Ϫ7 M) for 30 min before and during the incubation with BK. Reactions were terminated by addition of 3% (w/v) ice-cold perchloric acid and total [ 3 H]inositol phosphate radioactivity was isolated using AG1-X8 anion exchange column chromatography (formate form, 100 -200 mesh, Bio-Rad) after the radioactivity contained in phospholipids had been extracted with chloroform (4). Results are expressed as the ratio between the radioactivity measured in the inositol phosphates and the total radioactivity incorporated into the overall compounds labeled with myo- Measurement of Phospholipase A 2 activity-Cells were labeled to equilibrium with 1 Ci/ml [ 3 H]arachidonic acid (150 -230 Ci/mmol; Amersham International) for 18 h. After washing steps to eliminate unbound radiolabel (4), phospholipase A 2 activation experiments were performed at 37°C for 10 min in HBSS containing 0.1% BSA, protease inhibitors, and the test compounds or vehicles. The medium radioactivity containing the released [ 3 H]arachidonic acid plus derived 3 H-labeled metabolites was counted and expressed in percent of the total radioactivity, i.e. the medium plus cell associated radioactivity. 35 S Labeling and Immunoprecipitation-Cells were washed twice with sulfur-free HEPES-buffered Dulbecco's modified Eagle's medium, incubated for 30 min at 37°C in the same medium, and labeled with 0.1 mCi/ml 35 S-labeled amino acids (Prox-mix TM , Amersham International) for 8 h (21). After three washes with 50 mM Tris, pH 7.5, 150 mM NaCl (Tris-buffered saline), cells were scraped into ice-cold lysis buffer containing 50 mM Tris, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% (w/v) Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 0.1 mM Pefabloc SC TM , and 10 g/ml each of 1,10-phenantroline, aprotinin, leupeptin, and pepstatin A, and incubated for 45 min at 4°C under gentle rocking. The resulting lysates were centrifuged for 10 min at 13,000 rpm and the supernatants were precleared with 50 l of Staphylococcus aureus cell suspension (Pansorbin TM , Calbiochem).
To immunoprecipitate the receptor, 2.5 l of antiserum AS346 (21) diluted in 0.1 ml of 5% BSA in lysis buffer was added. After 15 min at room temperature, 50 l of Pansorbin TM was added, and the suspension incubated for 10 min at room temperature. The precipitate was recovered by a centrifugation for 2 min at 6,000 rpm and washed three times with lysis buffer and once with water. SDS sample buffer (25 l) was added to the immunoprecipitate followed by a 15-min incubation at 45°C. Proteins were resolved by 10% SDS-polyacrylamide gel electrophoresis (PAGE) in the presence of 5 M urea. After fixation with 20% (w/v) trichloroacetic acid for 20 min, the gels were washed several times with water and subjected to fluorography using 15% (w/v) sodium salicylate as the fluorophor. For control, the antiserum was preadsorbed for 90 min at room temperature with the peptide antigen coupled to Affi-Gel 10 (Bio-Rad).
Receptor Phosphorylation-Cells were washed twice with phosphatefree HEPES-buffered Dulbecco's modified Eagle's medium, incubated for 1 h at 37°C, and labeled with 0.25 mCi/ml [ 32 P]orthophosphate (ICN) for 8 h in the same medium. After a 5-min exposure to 1 M BK or vehicle at 37°C, cells were scraped into 1 ml of ice-cold lysis buffer containing protease inhibitors (see above) and phosphatase inhibitors (50 mM sodium fluoride, 25 mM sodium pyrophosphate, 1 mM sodium orthovanadate). To study the role of PKC and the effect of phosphatase inhibitor, cells were exposed to 0.1 M PMA and 0.1 M okadaic acid, respectively, in the absence or presence of 1 M BK before the cell lysis. Solubilization, immunoprecipitation, and gel electrophoresis were carried out as above. Proteins labeled by [ 32 P]orthophosphate were revealed by autoradiography.

RESULTS
Our strategy to analyze the role of the COOH-terminal domain of the B 2 receptor in ligand binding, signal transduction, sequestration, and phosphorylation involved the production of mutants with progressive truncations of the COOH-terminal tail of the receptor, deletions of defined portions of the tail region, or exchange of single or multiple residues of serine and threonine against alanine at defined positions of the COOHterminal domain of the receptor.
Endogenous Receptor of Cells Used for Transfection-Initially we tested for the presence of endogenous B 2 receptor in the COS cells using mock-transfected cells and comparing them with transfected cells expressing the wild-type human B 2 receptor (COSB 2 wt). The receptor density measured by [ 3 H]BK binding and the BK-sensitive PLC activity determined by IPs production in mock-transfected COS cells represented Ͻ5% of the values measured in COSB 2 wt, whereas the ligand-induced [Ca 2ϩ ] i transients of mock-transfected cells were similar to those obtained with COSB 2 wt (data not shown). No PLA 2 activation was observed in mock-transfected or COSB 2 wt cells indicating the absence of this transduction pathway in COS-7 cells. Therefore we choose [ 3 H]BK binding and IPs production to exemplify the signaling properties of the recombinant human B 2 receptors expressed in COS-7 cells. No endogenous B 2 receptor was detected with [ 3 H]BK binding, PLC, or PLA 2 assays in the mock-transfected CHO-K1 and HEK 293 cells.
Generation of Mutant B 2 Receptors-To study the functional role of the COOH-terminal tail region corresponding to intracellular domain ID-4 of the B 2 wt we produced three truncation variants with progressive deletions of the COOH-terminal portion of the receptor. These truncation mutants are designated by "t" followed by the name (single letter code) and position of their carboxyl-terminal amino acid, i.e. tR351, tI334, and tY320; they lack the COOH-terminal 14, 30, or 44 residues of the native receptor ( Fig. 1). Three deletion mutants are designated by "del" followed by the positions delimiting the deleted region, i.e. del[335-351], del [335][336][337][338][339][340][341][342], and del[343-351] lacking 17, 8, and 9 residues, respectively, of the COOH-terminal tail region (Fig. 1). We also created 7 receptor mutants in which alanine (Ala) was substituted for serine (Ser) or threonine (Thr) residue(s), singly or combined (Table I). In the mutant B 2 ⌬ST we exchanged 5 Ser/Thr residues located in the center portion of the COOH-terminal domain, i.e. at positions 339, 342, 345, 346, and 348 ( Fig. 1). We have also constructed 3 mutants where Ser/Thr residues have been replaced by Ala in other regions of the receptor, i.e. T237A, S316A, and S331A (not shown). The rationale for the design of the various mutants is given below.  (Table I) and at 10 Ϫ11 M to 10 Ϫ7 M for CHO-K1 cells (Table II). The incubation was carried out for 6 h under equilibrium conditions at 4°C to suppress internalization. In COS cells ( Table I) the kinetics of [ 3 H]BK binding displayed negative cooperativity as previously observed for the human B 2 wt in CHO-K1 cells (4). The receptor density determined by Scatchard analysis (31) was similar for the B 2 wt and mutant receptors except for tY320 and B 2 ⌬ST which were 10-and 5-fold less expressed, respectively. The kinetics of [ 3 H]BK binding estimated from Hill analysis of the binding data (32) were similar for the various receptors since half-maximum binding (K D,app ) occurred within a narrow range (2.3 to 6.3 nM) of [ 3 H]BK concentrations. Regarding the signal transduction properties, the basal IPs production was similar for B 2 wt and the mutant receptors demonstrating that none of the mutations resulted in a constitutive activation of PLC. All mutant receptors were able to trigger PLC activation in response to 10 Ϫ11 to 10 Ϫ7 M BK and responded to BK with maximum stimulation (4.1-8.5-fold) and EC 50 values (0.2-0.8 nM) in the same order of magnitude as those observed with B 2 wt with the exception of tY320 which had a higher EC 50 value (2.2 Ϯ 0.9 nM). Then, we choose to stably express in CHO-K1 cells the mutant tY320 bearing the most pronounced structural change and also the largest deleted mutant, del[335-351] lacking the 5 Ser and Thr residues. The B 2 wt CHO cell line was obtained as described previously (4). In these cells (Table II), tY320 also exhibited a higher EC 50 value for PLC activation, a greater EC 50 value for PLA 2 activation and a lower affinity for [ 3 H]BK. By contrast, the mutant del[335-351] exhibited a slightly higher affinity for [ 3 H]BK and exhibited a 30-and 50-fold lower EC 50 value for PLC and PLA 2 activation, respectively, compared with B 2 wt (Table II). These data indicate that the COOH-terminal tail is not necessary for signal transduction but that its conformation influences the efficiency of BK binding and coupling to PLC and PLA 2 .

TABLE I [ 3 H]BK binding and BK-induced PLC activation characteristics of B 2 wt and mutant receptors in COS cells
In each binding experiment, up to eight mutants were studied together with the B 2 wt as control. Cells were used 72 h after transfection, and incubated at 4°C for 6 h with seven increasing concentrations of [ 3 H]BK (from 1.25 ϫ 10 Ϫ11 to 2.5 ϫ 10 Ϫ8 M), before determination of specific binding (see "Experimental Procedures"). Data were first plotted using Scatchard coordinates. Given the curvilinear character of the plots obtained, binding values at the three highest [ 3 H]BK concentrations were used to estimate the maximal binding capacity (B max ). All binding values were then plotted using Hill coordinates to estimate the [ 3 H]BK concentration (K Dapp ) corresponding to half-saturation of binding sites. In each PLC activation experiment out of the 14 performed, up to six mutants were studied together with the B 2 wt as control. Cells at 72 h after transfection were incubated at 37°C for 15 min with 10 Ϫ11 to 10 Ϫ7 M BK and 10 mM LiCl, before determination of IPs production (see "Experimental Procedures"). For each mutant receptor, results are the mean Ϯ S.E. of at least three independent experiments, each performed in triplicate. Because the segment of the COOH-terminal domain of the B 2 wt that appears to interfere with receptor internalization contains 5 residues of serine/threonine that may be critical for receptor sequestration we constructed mutant receptors where a single residue was changed to alanine, i.e. S339A, T342A, and S348A, or two adjacent residues were substituted at the same time, i.e. T345A/S346A. The internalization capacity was reduced in the order Thr 342 Ͼ Thr 345 , Ser 346 Ͼ Ser 339 Ͼ Ser 348 although the level (41.1 Ϯ 4.5% to 88.3 Ϯ 3.1% of B 2 wt) was not as dramatic as for del[335-351] and tI334 (24.7 and 35.1% of B 2 wt, respectively) (Fig. 2). By contrast, the simultaneous replacement of distant Ser/Thr residues, i.e. S339A/T342A and S339A/T345A/S346A, produced a markedly reduced internalization (27.6 Ϯ 2.6 and 32.4 Ϯ 1.8% of B 2 wt, respectively) similar to that observed for del[335-351] and tI334 (Fig. 2). This increment was not further increased by the substitution of

FIG. 2. [ 3 H]BK internalization in COS-7 cells transfected with B2wt and mutant receptors.
Cells (used 72 h after transfection) were incubated with 2 nM [ 3 H]BK at 37°C for the times indicated. Unbound radioactivity was removed at 4°C before the cell surface-associated and the internalized radioactivities were separated and quantitated as described under "Experimental Procedures." Specific binding was calculated as the difference of total binding and nonspecific binding in the presence of unlabeled BK (10 M).  (Fig. 2, note that this mutant B 2 ⌬ST represents a full-length receptor protein). Together these findings suggest that Ser and Thr residues within the 17-residue segment spanning positions 335 to 351 contribute differentially to receptor endocytosis and can very likely intervene in this process in an alternative manner. Mutations of Ser/Thr residues external to this segment, e.g. T237A, S316A, and S331A neither altered the binding capacity nor the internalization rate of the resultant receptor mutants (data not shown).

Internalization of [ 3 H]BK by Mutant B 2 Receptors Expressed in CHO-K1 and HEK 293
Cells-Additionally to their characterization in the COS-7 cells, the wild type and some mutant receptors were tested for their ability to internalize [ 3 H]BK when expressed in CHO-K1 cells (stably and transiently) and in HEK 293 (transiently). The results summarized in Fig. 3 demonstrated that in those cells the truncation mutant tY320 and the deletion mutant del[335-351] as well as the point mutation B 2 ⌬ST receptor all exhibited a markedly decreased ability to internalize [ 3 H]BK compared with the B 2 wt, like in COS-7 cells. However, the fraction of the [ 3 H]BK binding which was internalized differed slightly from one cell type to another. For the control B 2 wt, this fraction was of 72.6 Ϯ 1.5% in CHO-K1, 57. Ϯ 3.1% in HEK 293, and 60.3 Ϯ 2.5% in COS-7. These observations can be linked to the differential amount of GRKs and arrestins expressed in these cell types reported by Menard et al. (33). It should be pointed out that the deletion mutant del[335-351] exhibited in every cell type the same reduced internalization rate than B 2 ⌬ST supporting our hypothesis that phosphorylation of Ser and Thr within the segment 335-351 is crucial for an optimal internalization of the BK B 2 receptor.
Immunoprecipitation of Mutant Receptors-The above observations made it of interest to study the receptor phosphorylation, especially to demonstrate whether the Ser and Thr residues discussed above are implicated. To this end, transfected cells with the various truncated, deleted and point mutated receptor mutants were labeled with [ 35 S]methionine or [ 32 P]orthophosphate, and incubated with or without 1 M BK for 5 min prior to cell lysis. For immunoprecipitation of the receptor we applied antiserum AS346 which has been raised against a peptide derived from the COOH terminus of the B 2 wt (positions 329 to 364) (21). In 35 S labeling experiments (Fig.   4d), immunoprecipitation of B 2 wt resulted in a diffuse band of 60 -100 kDa that was superimposable with the immunoreactive band found in 32 P labeling experiments (Fig. 4, a-c). All mutant receptors including B 2 ⌬ST were readily immunoprecipitated by the antiserum except for deletion mutants del[343-351] and del [335][336][337][338][339][340][341][342][343][344][345][346][347][348][349][350][351] which were precipitated at a 3-4-fold lower efficiency, and truncation mutants tR351, tI334, and tY320 which failed to react. These findings localize major immunogenic epitope(s) recognized by antiserum AS346 to the extreme COOH-terminal receptor portion distal of residue Arg 351 .
Phosphorylation of Mutant Receptors-Under basal conditions, i.e. in the absence of BK, a significant phosphorylation was observed for B 2 wt and mutant receptors except for del [335][336][337][338][339][340][341][342][343][344][345][346][347][348][349][350][351], del [343][344][345][346][347][348][349][350][351], and S348A suggesting that Ser at position 348 may represent a major target site for ligandindependent phosphorylation. BK dramatically increased the phosphorylation level of B 2 wt as well as of T237A, S316A, and S331A where Ser and Thr residues external to the Gln 335 3 Arg 351 segment had been mutated (data not shown). In marked contrast, BK failed to increase the phosphorylation level of del[335-351], B 2 ⌬ST, and S339A/T345A/S346A (Fig. 4); these mutants are characterized by a markedly reduced internalization capacity (Figs. 2 and 3). In all mutants exhibiting an "intermediate" internalization phenotype such as del[335-342], del[343-351], S339A, T342A, S348A, S339A/T342A, and T345A/S346A BK was still able to increase the phosphorylation (Fig. 4). Together these results demonstrate that the Ser and Thr residues phosphorylated upon ligand stimulation are located within the Gln 335 3 Arg 351 segment. The data also point to a close relationship between internalization and phosphorylation implying that BK promotes receptor internalization by homologous phosphorylation. To further test this hypothesis, we investigate with the B 2 wt whether PKC is involved in the internalization and phosphorylation processes. It was apparent that [ 3 H]BK binding was not sensitive to application, before and during the ligand incubation, of 0.1 M PMA or staurosporine, an activator and inhibitor of PKC, respectively (not shown). Under the same conditions, there was also no change in [ 3 H]BK internalization and BK stimulation of PLC. Concurrently, an immunoprecipitation of the B 2 wt showed that, unlike BK, PMA did not increase the receptor phosphorylation (not shown). Thus, PKC regulation of B 2 wt was not detected in COS-7 cells like in CHO-K1 cells (4). As discussed elsewhere in detail (4), this is in contrast with the reported role of PKC in B 2 receptor modulation in other cell types (17, 34 -36) and can be explained by a low PKC content of COS-7 and CHO-K1 cells facing receptor overexpression.
Interestingly, the treatment of COSB 2 wt with 0.1 M okadaic acid (a potent phosphatase 2A/1 inhibitor) for 5 min prior to and during the 10-min incubation period with 2 nM [ 3 H]BK (Fig. 5) resulted in a considerable increase in the internalization rate: 64.6 Ϯ 6.7% of the total [ 3 H]BK bound was internalized under okadaic acid treatment compared with 34.4 Ϯ 2.2% under control conditions (p Ͻ 0.05). Okadaic acid also resulted in a marked increase in the phosphorylation level of B 2 wt in the presence of BK (Fig. 5). By contrast, the same treatment applied to the cells expressing the mutant B 2 ⌬ST had no effect on the [ 3 H]BK internalization rate and the phosphorylation level (Fig. 5). These findings further support the hypothesis of a close relationship between internalization rates and phosphorylation levels of the B 2 receptor.
Relationship between B 2 Receptor Phosphorylation and Internalization-It has been reported that sucrose inhibits receptor endocytosis through clathrin-coated vesicles and that phosphorylated Ser/Thr residues in the COOH-terminal tail region of the ␤2-adrenergic receptor couple to the endocytic machinery via the adaptor protein arrestin (5, 6, 33). We here wondered whether the critical residues in the Gln 335 3 Arg 351 segment are essential for receptor endocytosis through clathrin-coated vesicles. For this purpose, we exposed cells expressing the B 2 wt or the mutant receptor del[335-351] to a standard solution or to hypertonic solution containing 0.4 M sucrose for 30 min prior to and for 10 min after the addition of [ 3 H]BK, and measured the internalization of the radioligand. Both the B 2 wt as well as the mutated receptor showed reduction of [ 3 H]BK internalization by roughly 70% (not shown). The same results were obtained whether receptors were expressed in COS-7 cells or CHO-K1 cells. This suggests that the clathrin-mediated, sucrose-sensitive endocytosis of the bradykinin B 2 receptor might not require the Ser/Thr phosphorylation residues in the Gln 335 3 Arg 351 segment. To further document this hypothesis, we tested the ability of wild type and dominant-negative mutant ␤-arrestin and dynamin to influence the sequestration of B 2 wt and the residual sequestration of B 2 ⌬ST receptors in COS-7 cells. Fig. 6 shows that unlike for the ␤-adrenergic receptor (33), the co-transfection with ␤-arrestin did not increase the internalization of [ 3 H]BK by either the B 2 wt or B 2 ⌬ST receptors, but co-expression of ␤-arrestin (319 -418) or dynamin K44A mutants exhibited inhibited [ 3 H]BK endocytosis for both receptors. One possibility to explain the inability of the overexpression of wild type ␤-arrestin to increase significantly the internalization rate of both receptors, is that the endogenous proteins may be sufficient to mediate maximal internalization of these receptors. In any case these observations indicate that although clathrin-mediated endocytosis may be important for B 2 receptor internalization this mechanism does not exclusively involve the phosphorylation of Ser/Thr residues in the Gln 335 3 Arg 351 segment. DISCUSSION Post-stimulatory desensitization and cellular redistribution of GPCRs are important mechanisms that regulate the availability and signaling capacity of hormonal effector systems. For instance, ligand-induced internalization of the ␤-adrenergic receptor allows the assembly of a scaffold of signaling factors of the mitogen-activated protein kinase system on intracellular vesicles and thus connects this receptor to the mitogenic pathways of the cell (37). Accumulating evidence suggests that specific sequence motifs of the intracellular loops and/or reversible modifications such as acylation and phosphorylation of the cytoplasmic domains of the receptors play critical roles in these processes (38). However, the precise molecular mechanisms driving desensitization and internalization of GPCRs have often remained unknown.
The present work demonstrates that a short sequence segment covering positions 335 to 351 of the bradykinin B 2 receptor COOH-terminal tail plays a major role in the agonistinduced internalization and phosphorylation of the receptor, but is not necessary for PLC and PLA 2 activation. Five amino acid residues, namely Ser in positions 339, 346, and 348 and Thr in positions 342 and 345, were shown to be involved in these functions and their respective contribution was analyzed. Complementary approaches for receptor modifications, i.e. truncation, deletion, and mutation of single or multiple residues were used to get insights into the structure-function relationships within the COOH-terminal tail. Most of the mutations resulted in alteration in the receptor capacity to undergo internalization and these alterations were consistent in different cell types. A reduction was always observed, indicating that the COOH-terminal tail possesses positive rather than negative motifs for internalization; at maximum we observed almost 80% reduction in the receptor sequestration in COS-7 cells. This contrasts with the findings that the human ␤ 2 -adrenergic receptor truncated at position 365 was internalized to a greater extent than the wild type receptor in these cells (14). Our data show that the COOH-terminal tail contributes significantly to the internalization of the B 2 wt, although other structures such as the intracellular loops may also be involved, albeit with a reduced efficiency (39).
The maximum reduction in receptor internalization was obtained with three distinct types of mutations and the effect of these mutations was consistent in the three different cell types tested COS-7, CHO-K1, and HEK 293. Substitution of the 5 Ser/Thr residues in the Gln 335  ⌬ST. This indicates that the contribution of the non-mutated residues is smaller in the mutated receptors than in the intact receptor and suggests that the Ser/Thr residues of the Gln 335 3 Arg 351 segment do not operate in a simple additive manner but operate in coordination to optimize receptor sequestration.
The present study shows that B 2 receptor internalization is independent of the affinity for BK or of the efficiency of PLC and PLA 2 coupling as reported for the truncated human platelet activating factor receptor (40). The COOH-terminal tail segment downstream of Tyr 320 as well as the proximal residues Thr 237 , Ser 316 , and Ser 331 do not function in the coupling of the receptors to G proteins stimulating the PLC pathway. These observations are in agreement with those of Faussner et al. (7) who reported that truncations either at Gly 327 or Lys 315 diminished the capacity to internalize BK but left the capacity to activate PLC unchanged. These and our results are, however, inconsistent with a recent report (39) claiming compromised ligand internalization associated with reduced PLC activity for a rat B 2 receptor mutant devoid of the terminal 34 residues (corresponding to tR331 for the human receptor). Differences in species (rat versus man) and expression systems (Rat-1 versus COS-7, CHO-K1, and HEK 293) may help explain some of the observed discrepancies.
Our work provides strong evidence that phosphorylation is the triggering signal that enters the B 2 wt in the internalization process. Several lines of evidence support this notion: (i) BK stimulation induces internalization and homologous receptor   2 ⌬ST, and S339A/T345A/S346A that suppress homologous receptor phosphorylation impair receptor internalization in all cell types tested; (vi) mutation of three phosphorylation sites, i.e. Ser 339 , Thr 345 , and Ser 346 , suffices to abolish homologous receptor phosphorylation and to markedly reduce internalization capacity of the human B 2 receptor. Remarkably Ala substitution of 3 out of 14 potential phosphorylation sites (Ser, Thr, and Tyr) present in the 4 intracellular domains are sufficient to abolish ligand-induced phosphorylation and to markedly reduce the internalization capacity of the human B 2 receptor studied in COS-7 cells. This observation points to the fact that the remaining non-mutated Ser/Thr residues cannot rescue the altered phosphorylation and internalization phenotype. On the contrary, the contributions of the non-mutated residues to internalization appear to be curtailed in these mutants.
The question whether BK-sensitive phosphorylation sites can serve as recognition motifs for receptor interaction with clathrin-coated vesicles was not elucidated in the present work. Indeed, the mutant del[335-351] with reduced internalization capacity and lacking these sites still exhibited a sensitivity to the clathrin-coated vesicle disrupting sucrose (data not shown) and to the effect of co-expression of dominant-negative forms of ␤-arrestin and dynamin. The experiments presented here are the first to address the involvement of arrestins and dynamin in the internalization of the bradykinin B 2 receptor. The magnitude of the inhibition of the B 2 ⌬ST internalization induced by co-expression of dominant-negative forms of ␤-arrestin or dynamin in COS-7 cells was of the same extent as that observed for B 2 wt in the same cell line. This suggests that the B 2 ⌬ST receptor has a affinity for arrestins comparable to those of the B 2 wt receptor and its residual internalization still involves clathrin-coated vesicles. The internalization that was not sensitive to ␤-arrestin and dynamin mutants could reflect sequestration of the receptor in caveolae, like that described for the B 2 receptor in DDT1 MF-2 smooth muscle cells by de Weerd et al. (25), and in A431 cells by Haasemann et al. (41). Interestingly, the desensitization that we have previously characterized in the CHO-K1 cells expressing the B 2 wt (4) did not occur in COS-7 expressing either the B 2 wt or mutant receptors, probably because of a differential expression of the human B 2 receptor desensitization machinery between the two cell types. The lack of desensitization in COS-7 cells, together with the decreased internalization observed in some mutants, is, however, of interest because it indicates that receptor internalization is not a prerequisite for the desensitization, as reported by others (10).
In conclusion, this work has allowed us to establish that optimal internalization and phosphorylation of the B 2 wt require the integrity of Ser 339 , Ser 346 , Ser 348 , Thr 342 , and Thr 345 residues, located in a short segment of 17 residues (Gln 335 3 Arg 351 ) in the center portion of the COOH-terminal tail. However, the mutation of these residues is not sufficient to completely suppress the sequestration through clathrin-coated vesicles, suggesting that the internalization of the [ 3 H]BK-B 2 receptor complex proceeds from different mechanisms involving distinct receptor structures. The critical Ser/Thr residues are flanked by 2 acidic residues, Glu in positions 337 and 350 delimiting a core sequence of 14 residues, Glu 337 3 Glu 350 . Previous in vitro studies with the ␤-adrenergic receptor have pointed to the potential importance of acidic residues juxtaposed to Ser/Thr residues (42). Comparison of the human Glu 337 3 Glu 350 segment with the sequences of other vertebrate kinin receptors (17,18,43) demonstrates that this cassette is well conserved among the kinin receptors: 9 of the 14 residues (65%) are invariant including all serine and threonine residues (Ser 339 , Ser 346 , Ser 348 , Thr 342 , and Thr 345 ), a centrally located positive residue of arginine (Arg 344 ), and the proximal acidic residue (Glu 337 ) whereas the distal acidic residue is conserved (Glu 350 or Asp 350 ). By contrast the overall sequence identity of intracellular domain ID4 is poor (15/64 corresponding to 23%). Our notion that the Glu 337 3 Glu 350 cassette plays a central role for receptor phosphorylation and internalization is further strengthened by the finding that the replacement of COOH-terminal tail region of the wild-type human B 1 receptor which fails to undergo ligand-induced phosphorylation and internalization, by the homologous region of the B 2 receptor holding the Glu 337 3 Glu 350 cassette confers the capacity to the B 1 /B 2 receptor chimera for ligand-induced internalization (7) and phosphorylation. 2 One limitation of the present approach is highlighted by the finding of alternative phosphorylation of closely spaced Ser/Thr residues within the putative phosphorylation cassette. Hence we cannot draw firm conclusions as to relative importance, functional hierarchy, and/or sequential modification of the critical Ser/Thr residues. Studies aimed at the precise mapping of the phosphorylation sites in vivo and at the elucidation of the sequence of phosphorylation events in the human B 2 bradykinin receptor are underway.