Involvement of the Amino Terminus of the B2 Receptor in Agonist-induced Receptor Dimerization*

The mechanisms and the functional importance of G-protein-coupled receptor dimerization are poorly understood. We therefore analyzed dimerization of the bradykinin B2receptor. The binding of the agonist bradykinin to the B2receptor endogenously expressed on PC-12 cells led to the formation of receptor dimers, whereas the B2 antagonist HOE140 did not induce dimerization, suggesting that B2 receptor dimerization was linked to receptor activation. Addition of a peptide corresponding to the amino terminus of the receptor reduced the amount of detected B2 receptor dimers, whereas peptides derived from the extracellular loops had no effect. To further analyze the role of the amino terminus of the receptor in receptor dimerization, we created two different rat B2 receptor variants with truncated amino termini, B2 53 and B2 65, starting at amino acids 53 and 65. In contrast to the wild-type B2 receptor and to B2 53, bradykinin did not induce dimerization of the B2 65 receptor. Both receptor variants were similar to the wild-type B2 receptor with respect to agonist binding and signal generation. However, B2 65 was not phosphorylated, did not desensitize, and was not downregulated upon bradykinin stimulation. Likewise, antibodies directed to the amino terminus of the receptor partially reduced internalization of [3H]bradykinin on PC-12 cells. These findings suggest that the amino terminus of the B2 receptor is necessary for triggering agonist-induced B2 receptor dimerization, and receptor dimers are involved in receptor-mediated signal attenuation.

The mechanism and the function of agonist-induced receptor dimerization of G-protein-coupled receptors is still a matter of debate. For a variety of G-protein-coupled receptors, receptor dimers have been detected recently, e.g. for the ␤ 2 -adrenergic receptor (1), the dopamine D 2 receptor (2), the metabotropic glutamate receptor 5 (3), or the calcium sensing receptor (4). G-protein-coupled receptors seem to dimerize via two different mechanisms. Whereas dimerization of the dopamine D 2 and the ␤ 2 -adrenergic receptor occurs via transmembrane regions, the metabotropic glutamate receptor and the calcium sensing receptor dimerize via their large amino terminus and, to a lesser extent, via their transmembrane regions (1)(2)(3)(4). G-protein-coupled receptors are classified into three main families according to the structure and length of their amino terminus and the localization of the agonist binding site (5). The metabotropic glutamate receptor and the calcium sensing receptor are members of family 3 receptors. In family 3 receptors, the amino terminus is not only involved in dimerization but also essential for agonist binding. In contrast, adrenaline and dopamine, which are agonists on family 1 receptors, bind within the seven transmembrane helices (5). The transmembrane regions are also involved in dimerization of these receptors (1,2). Thus, the mechanism of receptor dimerization seems to be related to agonist binding. To further investigate this hypothesis, we analyzed dimerization of the bradykinin B 2 receptor. The B 2 receptor belongs also to family 1. However, receptors for catecholamines or dopamine belong to the 1a subfamily, whereas receptors for peptides like bradykinin are classified into the subfamily 1b (5). In contrast to family 1a receptors, 1b-type receptors are characterized by an agonist binding site within the amino terminus of the receptor and the extracellular loops (5). In accordance with its classification as a family 1b receptor, a binding site(s) for the agonist bradykinin was identified within the extracellular loop regions, e.g. the connecting loops between membrane domains IV and V and domains VI and VII (6 -9). Because the agonist binding site of 1b receptors differs from that of 1a family receptors, we asked whether the sites necessary for receptor dimerization were also different between members of these two subfamilies. Here, we report that binding of bradykinin to the B 2 receptor endogenously expressed on PC12 cells or transiently expressed in HEK293 cells led to the formation of receptor dimers/oligomers, which were irreversibly captured by covalent cross-linking. The amino terminus of the receptor seemed to be involved in receptor dimerization because a B 2 receptor variant lacking the entire amino terminus, B 2 65 was impaired in agonist-induced receptor dimerization. desensitization and down-regulation, the B 2 receptor was expressed at low levels (0.05-0.1 pmol/mg of protein) to avoid saturation of the downstream effector components of the cells. The amount of B 2 -cDNA used for these experiments was 0.1-0.2 g of DNA/10 6 cells, and the total amount of transfected DNA was held constant by plasmid pcDNA3. The transfection efficiency was usually 70 -90%, as determined by staining of the cells by X-Gal (5-bromo-4-chloro-3-indoyl-␤-D-galactopyranoside) after transfection with a plasmid encoding ␤-galactosidase.
Construction of Expression Vectors-The cDNAs coding for the rat B 2 receptor variants starting at amino acid 53 (B 2 53 ) and 65 (B 2 65 ) were constructed by polymerase chain reaction and subcloned into pcDNA3. An initiating ATG was introduced to start translation at the indicated amino acid. The identity of the constructs was confirmed by DNA sequencing.
Desensitization of the Bradykinin-induced Rise in [Ca 2ϩ ] i -Bradykinin-induced changes in the intracellular free Ca 2ϩ concentration, [Ca 2ϩ ] i , were determined in fura-2-loaded HEK293 cells expressing the different B 2 receptor variants. Cells (1 ϫ 10 6 cells/ml) suspended in incubation buffer (138 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.6 mM CaCl 2 , 1 g/liter glucose, 20 mM Na ϩ -HEPES, pH 7.3) were loaded with fura-2/AM as described previously (11). Desensitization of the B 2 receptor was initiated by stimulation of fura-2-loaded HEK293 cells with 1 M bradykinin for 5 min at 25°C. Subsequently, cells were washed three times with prewarmed incubation buffer and used for [Ca 2ϩ ] i determination. Changes in [Ca 2ϩ ] i are given as the ratio between 340 and 380 nm.
Down-regulation of the Bradykinin-stimulated Increase in Inositol Phosphate Levels-For determination of total inositol phosphate levels, HEK293 cells expressing the different B 2 receptor variants were seeded on six-well plates and labeled with myo-[2-3 H]inositol (2 Ci/ml; specific activity, 17 Ci/mmol) for 16 h in inositol-free RPMI medium containing 0.2% dialyzed fetal calf serum. Prior to the experiment, cells were washed with incubation buffer (138 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.6 mM CaCl 2 , 1 g/liter glucose, 20 mM Na ϩ -HEPES, pH 7.3), and stored for 5 min in incubation buffer with 10 mM LiCl. Then, the cells were placed at 37°C, and the experiment was started by addition of buffer or bradykinin as indicated. For B 2 receptor down-regulation, cells were incubated with 10 M bradykinin for 6 h and washed three times with incubation buffer before the experiment was started as described above by the addition of bradykinin (0.1 nM to 1 M). After 20 min, total inositol phosphates were extracted as described previously (12). B 2 Receptor Phosporylation-B 2 receptor phosphorylation was performed as described previously (13) with minor modifications. Briefly, B 2 receptor-enriched membranes from transiently transfected HEK293 cells were prepared at 4°C by sucrose density gradient centrifugation (14) in the presence of protease inhibitors (6). Membranes containing 3-5 pmol of B 2 receptor/100 g of protein were washed by 5 M urea and stored at Ϫ80°C. Recombinant G-protein-coupled receptor kinase 2 (GRK2) was purified from baculovirus-infected Sf9 cells as described (15). For B 2 receptor phosphorylation, reaction mixtures (50 ul) contained 20 nM B 2 receptor, 50 nM GRK2, 200 nM ␤␥ complexes purified from bovine brain (16), 5 mM MgCl 2 , 50 M [␥-32 P]ATP (NEN Life Science Products), and 20 mM Tris, pH 7.6. Incubations were carried out at 30°C for 25 min in the presence or absence of 1 M bradykinin. Reaction was stopped with an equal volume of 2ϫ SDS sample loading buffer, and samples were electrophoresed on SDS/10% polyacrylamide gels containing 6 M urea. Proteins were visualized by autoradiography.

Detection of Agonist-induced B 2 Receptor Dimerization on PC-12 Cells and on HEK293 Cells Expressing Different B 2 Receptor Variants-
Bradykinin-induced formation of receptor dimers on PC12 cells and on HEK293 cells expressing wild-type B 2 receptor, B 2 53 , and B 2 65 was detected similarly. Membranes of PC-12 or transfected HEK293 cells were prepared at 4°C in the presence of protease inhibitors as described previously (6). The membranes were incubated with 10 nM bradykinin or 10 nM HOE140 for 30 min at 25°C in incubation buffer (20 mM HEPES, 138 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , pH 7.4). For cross-linking, buffers were routinely degassed and handled under N 2 atmosphere. Cross-linking was initiated by the addition of 0.2 mM MBS and incubation for 15 min at room temperature. A new batch of MBS was used for each experiment to avoid inactivation of the cross-linker. Cross-linking was terminated by addition of 100 mM Tris, pH 8.0, and 10 mM dithiothreitol. Membranes were collected by centrifugation and solubilized in SDS sample loading buffer for 15 min at 25°C instead of boiling to avoid unspecific receptor aggregation. Solubilized proteins were separated by SDS-PAGE, including 6 M urea. Cross-linked bradykinin-B 2 receptor complexes were detected by antibradykinin antibodies (7). Similarly, HOE140-B 2 receptor complexes were detected by anti-HOE140 antibodies (6). To control B 2 receptor specificity, a 500-fold molar excess of HOE140 was used as a control for bradykinin cross-linking and a 500-fold molar excess of bradykinin as a control for HOE140 cross-linking. Bound antibodies were visualized with peroxidase-labeled secondary antibodies with enhanced chemiluminescence detection.
Effect of Peptides Corresponding to the Extracellular Domains of the Rat B 2 Receptor on B 2 Receptor Dimer Formation-PC-12 membranes containing 0.5-1 pmol of B 2 receptor/mg of protein were suspended in phosphate-buffered saline including protease inhibitors (6) and incubated without or with 10 M of a peptide corresponding to the sequence of the connecting loop between membrane domains II and III (ED2), IV and V (ED3 C ), and VI and VII (ED4 N ) (6) and to the amino terminus of the B 2 receptor (ED1) for 1 h at 4°C. The position of the peptides in the rat B 2 receptor and their synthesis has been described previously (6). To induce receptor dimerization, 10 nM [ 125 I]Tyr 8 bradykinin was added to the membranes for 30 min at 25°C. Preparation of [ 125 I]Tyr 8 bradykinin was performed as described previously (14), and cross-linking by 0.2 mM DFDNB was performed for 15 min at room temperature. The crosslinking reaction was terminated by 100 mM Tris, pH 8.0, and unbound ligand was removed by washing with the same buffer. The membranes were solubilized in SDS-sample buffer, and proteins were separated by 10% SDS-PAGE including 6 M urea. [ 125 I]Tyr 8 bradykinin-labeled B 2 receptors were visualized by autoradiography.

Effect of Site-directed Antibodies on the Internalization of [ 3 H]Bra
dykinin-PC-12 cells on six-well plates were incubated for 2 h at 4°C in the absence (control) or presence of 100 nM affinity-purified antibodies to the amino terminus of the B 2 receptor (␣-ED1) or to the connecting loop between membrane domains II and III (␣ -ED2), IV and V (␣-ED3 C ), and VI and VII (␣ -ED4 N ) in incubation buffer (138 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.6 mM CaCl 2 , 1 g/liter glucose, 20 mM Na ϩ -HEPES, pH 7.3). The generation of the antibodies has been performed essentially as described previously (6). Then, 5 nM [ 3 H]bradykinin was added and incubated for an additional 1 h at 4°C. To induce internalization, cells were shifted to 37°C. At the time points indicated, cells were washed three times with 0.2 M glycine, 0.5 M NaCl, pH 3.0, followed by three washing steps with incubation buffer. Then, cells were dissolved by 1 M NaOH, and internalized [ 3 H]bradykinin was determined in a ␤-counter.

Detection of Agonist-induced B 2 Receptor Dimerization on PC-12 Cells-
To determine a potential functional role of receptor dimerization, we attempted to detect B 2 receptor dimers in a system with endogenously expressed B 2 receptors, and we chose PC-12 cells (11). To detect bradykinin-stimulated receptor dimerization, we stimulated cell membranes by bradykinin at 25°C for 30 min and irreversibly captured receptor dimers by addition of the cross-linker MBS (0.2 mM). Cross-linked bradykinin-B 2 receptor complexes were detected by anti-bradykinin antibodies (7). And indeed, bradykinin binding to the B 2 receptor lead to the formation of receptor dimers and higher oligomerization states as shown in Fig. 1A, lane 1. The B 2 antagonist HOE140 did not significantly stimulate the formation of receptor aggregates when HOE140 was cross-linked to the B 2 receptor under similar conditions as described above and detected by anti-HOE140 antibodies (Fig. 1A, lane 3). To control B 2 receptor specificity, a 500-fold molar excess of HOE140 was used as a control for bradykinin cross-linking (Fig. 1A, lane 2) and a 500-fold molar excess of bradykinin as a control for HOE140 cross-linking (Fig. 1A, lane 4). The presence of a 500-fold molar excess of HOE140 or bradykinin abol-ished the specific labeling of the B 2 receptor. Thus, B 2 receptors had been labeled by bradykinin and by HOE140, but only bradykinin had induced the formation of receptor dimers/oligomers. This finding suggests that B 2 receptor dimerization/ oligomerization was dependent on the binding of agonist to the B 2 receptor.
Effect of Peptides Corresponding to the Extracellular Domains of the B 2 Receptor on B 2 Receptor Dimer Formation-We asked whether the extracellular domains of the B 2 receptor might be involved in B 2 receptor dimerization, and we analyzed the effect of peptides derived from the extracellular domains of the B 2 receptor. The position of the peptides in the rat B 2 receptor has been described previously (6). Membranes of PC-12 cells were preincubated without (Fig. 1B, lanes 4 and 5) or with 10 M of a peptide corresponding to the connecting loop between membrane domains II and III, ED2 (Fig. 1B, lane 1); to the connecting loop between membrane domains IV and V, ED3 C (Fig. 1B, lane 2); to the connecting loop between membrane domains VI and VII, ED4 N (Fig. 1B, lane 3); and to the amino terminus of the B 2 receptor, ED1 (Fig. 1C, lanes 1 and 2) (cf. Ref. 6 for sequence of the peptides). B 2 receptor monomers and dimers were identified by cross-linking of [ 125 I]Tyr 8 bradykinin to B 2 receptors on PC-12 cell membranes by 0.2 mM DFDNB. The presence of 10 M peptide corresponding to the extracellular loops of the B 2 receptor did not reduce the ratio between labeled B 2 receptor monomers versus dimers compared with the control without peptide (Fig. 1B, lanes 1-3  versus lane 4). Interestingly, no oligomeric states larger than dimers were found when we used the homobifunctional crosslinker DFDNB (Fig. 1B) compared with the heterobifunctional cross-linker MBS (cf. Fig. 1A). This finding may be due to different length and different specificity of the applied crosslinkers. The specificity of the cross-linking was confirmed by the finding that a 500-fold molar excess of unlabeled bradykinin suppressed the labeling of the B 2 receptor by [ 125 I]Tyr 8 bradykinin (Fig. 1B, lane 5). Together, these findings suggest that peptides corresponding to the extracellular loops of the B 2 receptor did not affect the capturing of receptor dimers by DFDNB.
In contrast, a peptide derived from the amino terminus of the B 2 receptor (ED1) significantly reduced the ratio between labeled B 2 receptor monomers versus dimers (Fig. 1C, lane 2) compared with the control without peptide (Fig. 1B, lane 4, and Fig. 1C, lane 1) or compared with the controls with peptides directed to the extracellular loops (Fig. 1B, lanes 1-3). The presence of a 500-fold molar excess of the B 2 antagonist HOE140 abolished the labeling by [ 125 I]Tyr 8 bradykinin, demonstrating B 2 receptor specificity of the signal(s) (Fig. 1C, lane  3). Together, these findings may suggest that the amino terminus of the B 2 receptor was involved in triggering B 2 receptor dimerization induced by bradykinin.
Expression of Amino-terminally Truncated B 2 Receptor Variants in HEK293 Cells-To further analyze the role of the amino terminus of the B 2 receptor in receptor dimerization, we created two rat B 2 receptor variants with truncated amino termini, B 2 53 , starting with amino acid 53 after the conserved cysteine 52 (17), and B 2 65 , lacking the entire amino terminus. The indicated positions of the amino acids are derived from the rat B 2 receptor sequence with extended amino terminus (17). The wild-type B 2 receptor and the two B 2 receptor variants were expressed in HEK293 cells. Expression of the B 2 receptor protein was detected in Western blot by anti-HOE140 antibodies after cross-linking of HOE140 to the B 2 receptor. The antibodies detected a major protein of 55 Ϯ 5 kDa in wild-type receptor-expressing cells (Fig. 2, A, lane 1, and B, lane 1), and proteins of 45 Ϯ 6 and 43 Ϯ 4 kDa were detected in cells expressing B 2 53 (Fig. 2B, lane 2) and B 2 65 (Fig. 2B, lane 3), respectively. No significant amounts of receptor dimers were detected (Fig. 2, A, lane 1, and B, lanes 1-3). The labeling of the B 2 receptors by HOE140 was suppressed when the cross-linking was performed in the presence of a 500-fold molar excess of bradykinin, as demonstrated for the wild-type B 2 receptor (Fig.  2A, lane 2). These findings demonstrate that binding of HOE140 to the different B 2 receptors expressed in HEK293 cells did not induce the formation of receptor dimers, although HOE140 was efficiently cross-linked to these receptors, revealing similar expression levels.

Agonist-induced Receptor Dimerization of B 2 Receptors
Expressed in HEK293 Cells-Next we asked whether bradykinin stimulated receptor dimerization of the wild-type B 2 receptor and the truncated B 2 receptor variants expressed on HEK293 cells. Similarly as on PC-12 cells, B 2 receptor dimers were irreversibly captured by MBS and detected by anti-bradykinin antibodies in Western blot. Bradykinin induced the formation of receptor dimers and higher oligomerization states on membranes of HEK293 cells expressing wild-type B 2 receptors (Fig.  2C, lane 1) and on membranes of cells expressing B 2 53 (Fig. 2C,  lane 2). Although no oligomeric states larger than dimers were found with B 2 53 , the ratio between B 2 receptor monomers versus dimers was similar for the wild-type B 2 receptor and for B 2 53 . In contrast, there were no significant amounts of receptor dimers detectable following stimulation of the B 2 65 variant by bradykinin (Fig. 2C, lane 3). Bradykinin was efficiently crosslinked to the monomeric form of B 2 65 (Fig. 2C, lane 3), confirm-ing that the truncation of further 12 amino acids in B 2 65 compared with B 2 53 did not reduce cross-linking efficiency (Fig. 2C,  lane 3 versus lane 2). A 500-fold molar excess of HOE140 suppressed the specific labeling of the wild-type B 2 (lanes 1-3), B 2 53 (lanes 4 -6), and B 2 65 (lanes 7 and 8) were phosphorylated by GRK2 (50 nM) in the absence (lanes 1, 4, and 7) or presence of 1 M bradykinin (lanes 2, 5, and 8) or desArg 10 kallidin (lanes 3 and 6) at 30°C for 25 min. Proteins were separated by SDS-PAGE (7.5%), and 32 P-labeled proteins were visualized by autoradiography. The autoradiography is representative of three independent experiments with similar results. B, desensitization of the bradykinin-induced increase in [Ca 2ϩ ] i . HEK293 cells (1 ϫ 10 6 cells/ml) expressing wild-type B 2 (panels 1 and 4) receptor variants. The concentration of bradykinin necessary to produce the half-maximum increase in inositol phosphate levels (EC 50 value) was 34 Ϯ 7 ϫ 10 Ϫ9 , 22 Ϯ 8 ϫ 10 Ϫ9 , and 38 Ϯ 6 ϫ 10 Ϫ9 M for the wild-type B 2 receptor, for B 2 53 , and for B 2 65 , respectively (cf. Fig. 3C). Thus, the amino terminus of the B 2 receptor is not essential for B 2 receptor activation.
Interrelationship between Agonist-induced Receptor Dimerization and Desensitization-B 2 receptor phosphorylation is the initial step in B 2 receptor desensitization (12,13). Therefore, phosphorylation of the B 2 receptor variants by GRK2 was analyzed. GRK2 phosphorylated the B 2 53 variant upon agonist stimulation (Fig. 3A, lane 5) similarly to the wild-type B 2 receptor (Fig. 3A, lane 2). No significant increase in B 2 receptor phosphorylation was detected after stimulation by a B 1 -specific agonist, desArg 10 kallidin (Fig. 3A, lanes 3 and 6) compared with the unstimulated control (Fig. 3A, lanes 1 and 4). In contrast, GRK2 did not significantly phosphorylate the B 2 65 variant upon bradykinin stimulation (Fig. 3A, lane 8) compared with the unstimulated control (Fig. 3A, lane 7). This experiment strongly suggests that agonist-induced phosphorylation of the B 2 65 receptor by GRK2 was impaired. Receptor phosphorylation is assumed to precede receptor desensitization (18,19). Therefore, we next asked whether desensitization of the B2 65 variant was also impaired. The desensitization of the bradykinin signal was analyzed for the bradykinin-stimulated rise in [Ca 2ϩ ] i on fura-2-loaded HEK293 cells expressing the different B 2 receptor variants. The bradykinin-induced rise in [Ca 2ϩ ] i was similar in wild-type B 2 receptor-expressing cells, in cells expressing B 2 53 , and in cells expressing B 2 65 (Fig. 3B, panels 1-3). To initiate receptor desensitization, the cells were preincubated for 5 min with 1 M bradykinin. After removal of bradykinin, the cells were given a second pulse of bradykinin. A second stimulation by bradykinin did not significantly increase [Ca 2ϩ ] i in wild-type B 2 receptorexpressing cells (Fig. 3B, panel 4) or in cells expressing the B 2 53 receptor (Fig. 3B, panel 5), indicating that the receptors were efficiently desensitized after prestimulation by 1 M bradykinin. In contrast, the B 2 65 receptor was not desensitized after the initial bradykinin stimulation, as revealed by a rise in [Ca 2ϩ ] i following a second pulse of bradykinin (Fig. 3B,  panel 6).
Down-regulation of the Different B 2 Receptor Variants-The B 2 65 receptor poorly dimerized, was not significantly phosphorylated, and did not desensitize upon bradykinin stimulation. Therefore, we asked whether the B 2 65 receptor could be down regulated. To this end, the B 2 receptor-expressing cells were prestimulated by 10 M bradykinin for 6 h. After washing, the bradykinin-stimulated increase in inositol phosphate levels was determined compared with the unstimulated control. After long-term bradykinin stimulation, the B 2 receptor-mediated signal was down-regulated on wild-type B 2 receptor-expressing cells and on B 2 53 expressing cells. The signal was reduced by 28 Ϯ 3 and 25 Ϯ 5%, respectively (Fig. 3C, panels 1 and 2).
In contrast, the B 2 receptor-mediated signal was not down regulated on B 2 65 -expressing cells (Fig. 3C, panel 3). Thus, a B 2 receptor variant lacking the entire amino terminus had lost its capability to undergo agonist-induced dimerization, phosphorylation, desensitization, and down-regulation.
Effect of Site-directed Antibodies on the Internalization of [ 3 H]Bradykinin on PC-12 Cells-Finally, we asked whether the amino terminus of the B 2 receptor was also involved in the signal attenuation process of a B 2 receptor expressed in a native system. To this end, we initially tested the effect of the peptide ED1 derived from the amino terminus of the B 2 receptor, which has been shown to reduce the ratio between detected B 2 receptor monomers versus dimers (cf. Fig. 1C). However, this peptide (10 M) had no significant effect on B 2 receptor phosphorylation and internalization under the conditions applied (not shown). Therefore, we next analyzed the effect of site-directed antibodies to the extracellular domains of the B 2 receptor. The generation, immunoselection, and characterization of these antibodies was performed as described previously (6). The site-directed antibodies used for this study did not affect the binding of [ 3 H]bradykinin or the activation of the B 2 receptor (6). The antibodies to the amino terminus of the B 2 receptor (␣-ED1) caused a significant reduction of internalized [ 3 H]bradykinin (Fig. 4, A and B) by 24 Ϯ 4%, whereas sitedirected antibodies to the extracellular loops (␣-ED2, ␣-ED3 C , and ␣-ED4 N ) had no effect on internalization of [ 3 H]bradykinin (Fig. 4A). Results obtained with the site-directed antibodies in phosphorylation studies were not consistent, probably due to the moderate effect of the antibodies (not shown). Nevertheless, our data indicate that the amino terminus of the B 2 receptor is also involved in B 2 receptor-mediated signal attenuation processes on PC-12 cells. DISCUSSION Although the detailed mechanism of receptor dimerization is far from being understood, different regions of G-protein-coupled receptors have been identified to be involved in receptor dimerization. Dimerization of the dopamine D 2 and the ␤ 2adrenergic receptor occur via transmembrane regions (1,2). In contrast, the metabotropic glutamate receptor and the calcium sensing receptor dimerize via their large extracellular amino terminus and to a lesser extent via their transmembrane regions (3,4). Although the B 2 receptor, as a family 1b receptor, lacks a large extracellular domain that is characteristic of family 3 receptors, agonist-induced B 2 receptor dimerization also needs the amino terminus of the receptor because (i) a peptide corresponding to the amino terminus of the B 2 receptor partially suppressed bradykinin-induced receptor dimerization on PC-12 cells, and (ii) a B 2 receptor variant lacking the entire amino terminus, B 2 65 , was impaired in agonist-induced receptor dimerization. The amino terminus, which was necessary for agonist-stimulated B 2 receptor dimerization, was not essential for agonist binding. Such a finding is in contrast to previous studies with the ␤ 2 -adrenergic, the dopamine, the calcium sensing, or the metabotropic glutamate receptor where the region(s) involved in receptor dimerization were overlapping with the agonist binding sites (1)(2)(3)(4). This potential mechanistic difference of B 2 receptor dimerization may be related to the predicted topology of the B 2 receptor, which is different from rhodopsin with five transmembrane domains instead of seven and two membrane reentrant segments, membrane domains I and II (20). Thus, membrane domain I, which follows the amino terminus, is not membrane-spanning.
In addition to the mechanism, the functional importance of receptor dimerization is not clear. Receptor dimerization of the ␤ 2 -adrenergic receptor seems to be related to receptor activation because the amount or receptor dimers increased upon agonist stimulation, and a peptide derived from transmembrane domain VI partially suppressed receptor dimerization and receptor activation (1). Several other lines of evidence suggest, however, that receptor dimerization does not seem to be essential for receptor activation. The stoichiometry of the interaction between the receptor, G-protein, and effector molecule is assumed to be 1:1:1 (21), and a single receptor with a single G␣ s protein fused to its carboxyl terminus is functionally active (22). Therefore, the potential role of receptor dimerization may lie apart from receptor activation. And indeed, a B 2 receptor variant (B 2 65 ) with impaired agonist-induced receptor dimerization was "normal" with respect to signal generation. In contrast, this B 2 receptor variant was defective in agonistinduced receptor phosphorylation, desensitization, and downregulation. Likewise, affinity-purified antibodies to the region that is necessary for receptor dimerization, the amino terminus of the B 2 receptor, partially reduced internalization of [ 3 H]bradykinin on PC-12 cells. Thus, agonist-induced dimerization of the B 2 receptor seems to follow receptor-induced G-protein activation and to precede receptor phosphorylation, desensitization, and down-regulation.
Twelve amino acids upstream of membrane domain I may be sufficient to support agonist-induced B 2 receptor dimerization because B 2 53 was still capable of dimerizing after bradykinin stimulation, whereas B 2 65 , lacking the entire amino terminus, did not significantly dimerize after bradykinin stimulation. Bradykinin-induced B 2 receptor dimerization did not depend on receptor overexpression because bradykinin-induced B 2 re-ceptor dimerization was also detected on PC-12 cells with endogenously expressed B 2 receptors. Thus, agonist-induced B 2 receptor dimerization via the amino terminus of the receptor is a characteristic of B 2 receptor signaling.
In addition to agonist-induced dimerization, G-protein-coupled receptors tend to dimerize under native conditions via hydrophobic interactions (7). Receptor clustering via hydrophobic interactions within the (trans)membrane domains may be supported by high receptor density. Future work will have to determine whether, in addition to agonist-induced receptor dimerization, which has been shown to regulate signal attenuation processes, receptor clustering under basal conditions represents an additional mechanism of regulating B 2 receptor activity.