The N-terminal amino group of [Tyr8]bradykinin is bound adjacent to analogous amino acids of the human and rat B2 receptor.

To obtain data of the bradykinin B2 receptor's agonist binding site, we used a combined approach of affinity labeling and “immunoidentification” of receptor fragments generated by cyanogen bromide cleavage. Domain-specific antibodies to the various extracellular receptor domains were applied to detect receptor fragments with covalently attached [125I-Tyr8]bradykinin. As a cross-linker we used the homobifunctional reagent disuccinimidyl tartarate (DST), which reacts preferentially with primary amines. With this technique a [125I-Tyr8]bradykinin-labeled receptor fragment derived from the third extracellular domain was identified. The ε-amino group of lysine (Lys172) of the human B2 receptor provides the only primary amino group within this receptor fragment. This strongly suggests that DST attached the N-terminal amino group of [Tyr8]bradykinin to Lys172 of the human B2 receptor. Next we asked whether DST attaches [Tyr8]bradykinin to the analogous residue, Lys174 of the rat B2 receptor, which is 81% identical to the human B2 receptor, and we attempted to label the wild-type rat B2 receptor and a rat B2 receptor mutant where Lys174 had been exchanged for alanine. Affinity labeling of the wild-type rat B2 receptor worked efficiently, whereas DST did not attach detectable amounts of [125I-Tyr8]bradykinin to the K174A rat B2 receptor mutant. Taken together these observations indicate that the N-terminal amino group of [Tyr8]bradykinin is bound to analogous positions of the rat and of the human B2 receptor, i.e. [Tyr8]bradykinin's N terminus is bound adjacent to Lys172 of the human and Lys174 of the rat B2 receptor.

To obtain data of the bradykinin B 2 receptor's agonist binding site, we used a combined approach of affinity labeling and "immunoidentification" of receptor fragments generated by cyanogen bromide cleavage. Domainspecific antibodies to the various extracellular receptor domains were applied to detect receptor fragments with covalently attached [ 125 I-Tyr 8  The ligand binding sites of G protein-coupled receptors have been mapped using mutations and affinity labeling. These studies suggest that much of the ligand binding site of peptide hormone receptors such as the angiotensin AT 1 receptor (1), the neurokinin NK-1 receptor (2,3), the lutenizing hormone receptor (4), or the interleukin-8 receptor (5) is within the extracellular loop regions. Cloning of the cDNA coding for a rat (6) and a human (7) B 2 receptor revealed that the receptor for the peptide hormone bradykinin belongs to the family of G protein-coupled receptors. Mutation studies of the bradykinin B 2 receptor identified residues important for agonist binding on putative transmembrane regions TM6 (8) and at the top of TM6 and TM7 (9). Few of the mutations were made within the extracellular domains, and it is not clear whether they form a contact region for B 2 ligands. To assess the involvement of the extracellular domains in agonist binding without changing the receptor's primary structure, we previously produced and characterized six antisera to the predicted extracellular domains (EDs) 1 of the B 2 receptor (10). Antibodies to the amino-half of the third extracellular domain (ED3 N ) compete with bradykinin's binding and are agonists (10). To further the analysis of bradykinin's binding site(s), we combined affinity labeling and immunoidentification of [ 125 I-Tyr 8 ]bradykinin-labeled B 2 receptor fragments generated by cyanogen bromide cleavage to identify a region(s) where the agonist bradykinin contacts the B 2 receptor. We chose the amino group-specific homobifunctional cross-linker disuccinimidyl tartarate (DST) to localize the position of the N-terminal amino group of [ 125 I-Tyr 8 ] bradykinin when bound to the B 2 receptor. A combined approach of affinity labeling and identification of ligand-labeled receptor fragments has been extensively used to determine ligand binding site(s) of membrane receptors, e.g. of the GABA A receptor (11), of the renal V 2 vasopressin receptor (12), or of the nicotinic acetylcholine receptor (13). In this study we used this technique to localize a B 2 receptor region involved in the interaction with the agonist's N terminus. This should contribute to a refinement of the present model of the B 2 receptor (14), thereby facilitating the rational design of B 2 agonists or antagonists.
Cell Culture and Cell Transfection-HF-15 cells (15) expressing 0.5-1 pmol B 2 receptor/mg protein at passages 6 -11 were maintained in Dulbecco's modified Eagle's medium, and COS-1 cells were grown in RPMI 1640 medium. Both media were supplemented with 10% (v/v) fetal bovine serum, and cells were kept in a 5% CO 2 /95% air atmosphere. COS-1 cells were transfected with wild-type and K174A-mutated rat B 2 receptor cDNA by LipofectAMINE (Life Technologies, Inc.) according to the manufacturer's instructions. The rat B 2 receptor mutant was made by a modification of the polymerase chain reaction mutagenesis method (16) in a cassette encompassing the unique BglII and PvuII sites of the protein coding region. DNA was transiently expressed in COS-1 cells using the expression plasmid pSRF-159, which is a derivative of the SR␣ promotor vector pcDLSR␣296 (17). The mutagenesis cassette was completely sequenced to confirm the identity of the mutant.
Competition and Saturation Binding of Radiolabeled Ligands-Binding of [ 3 H]bradykinin and of 125 I-labeled antibodies to intact HF-15 or COS-1 cells were performed as described previously (10).
Radioiodination of Peptides-Peptides (1 g) dissolved in 100 l of phosphate-buffered saline were incubated with 2 mCi of carrier-free Na[ 125 I] on a solid phase of Iodogen (100 g/tube) for 10 min (18). Unreacted iodine was separated by anion exchange chromatography over Dowex AG1 ϫ 8.
Synthesis of Peptides and Production of Anti-peptide Antibodies-Synthesis of peptides and production and characterization of anti-peptide antibodies to the various extracellular domains of the B 2 receptor (see Fig. 1A) have been described previously (10). Antisera were tested for antigen specificity by the indirect enzyme-linked immunosorbent assay (19) using microtiter plates (Maxisorb, Nunc) coated with 2 g/ml of the peptide or of the conjugate.
Purification of Anti-peptide Antibodies by Affinity Chromatography-Peptides were covalently coupled to Affi-Gel 10 (5 mg/ml gel). The antiserum was applied (5 ml/ml gel) and incubated under gentle agitation for 12 h at 4°C. The affinity matrix was washed four times with phosphate-buffered saline, and bound antibodies were eluted with 0.2 M glycine, pH 2.5, and immediately neutralized with 1 M KOH. Antibodies were desalted and concentrated by a Centricon filtration unit, exclusion limit 30,000 Da. Purity and specificity of the antibodies was analyzed by SDS-PAGE and enzyme-linked immunosorbent assay.
Identification of B 2 Receptors in Immunoblots-Proteins were resolved by SDS-PAGE and transferred to nitrocellulose sheets using semi-dry blotting (20). The sheets were treated with 50 mM Tris, 0.2 M NaCl, pH 7.4 (buffer A), containing 5% (w/v) nonfat dry milk and 0.1% (w/v) Tween 20 for 1 h. Antisera were diluted 1:1000 in buffer A containing 2% (w/v) bovine serum albumin. After a 30-min incubation at 37°C, nitrocellulose was washed five times for 15 min each with buffer A and incubated for 30 min with peroxidase-labeled F(abЈ) 2 fragments of goat anti-rabbit antibody (Sigma, 1:5000). After extensive washing, bound antibody was visualized using the ECL chemiluminescence detection kit (Amersham Corp.).
Affinity Membrane preparation was done as described previously (10). For control a 1000fold molar excess of bradykinin or of the antagonist HOE140 was included. Prior to the addition of 1 mM DST, membranes of COS cells were washed twice with incubation buffer, whereas DST was directly added to solubilized HF-15 cells. After an additional incubation for 1 h on ice, the reaction was terminated by 1 M Tris, pH 8. Proteins of HF-15 cells were precipitated and desalted by 80% (v/v) acetone (21). For SDS-PAGE (22) proteins of HF-15 or of COS-1 cells were dissolved in SDS-sample buffer (2% (w/v) SDS, 5 mM EDTA, 5% (v/v) 2-mercaptoethanol, 20% (v/v) glycerol, 0.01% (w/v) bromphenol blue, 67.5 mM Tris-HCl, pH 6.7) and boiled for 5 min. After SDS-PAGE, B 2 receptor was visualized by autoradiography.

Affinity Chromatography of [ 125 I-Tyr 8 ]Bradykinin-labeled B 2 Receptors by
Anti-bradykinin Antibodies-Affinity-purified anti-bradykinin antibodies were covalently coupled to Affi-Gel 10 (15 mg/ml gel). After affinity labeling of solubilized B 2 receptors by [ 125 I-Tyr 8 ]bradykinin (see above), proteins were precipitated by 80% (v/v) acetone and redissolved in phosphate-buffered saline, 0.1% Nonidet P-40 (buffer B) including protease inhibitors. Gel filtration over Sephadex G-50 was performed to minimize contamination with free ligand. The flow-through was applied to the immunoaffinity matrix for an overnight incubation, the matrix was extensively washed with buffer B, and bound proteins were eluted with 0.2 M glycine, pH 2.5, supplemented with 10% (v/v) 1,4-dioxane. The eluted protein fraction was neutralized with 1 M Tris, pH 8.0, and concentrated and desalted over Centricon-30.
Immunoprecipitation of B 2 Receptor Fragments by 6 Different Domain-specific Antibodies-B 2 receptor fragments were identified by im-munoprecipitation with six different domain-specific antibodies. The antibodies were raised against peptides derived from the extracellular domains of the B 2 receptor (Ref. 10 and Fig. 1A). Each of the six different antibody fractions were bound to Protein A-Sepharose (50 l of gel, 5-10 g of affinity-purified IgG). The B 2 receptor fragments from the cyanogen bromide cleavage reaction were dissolved in 6 ml of HGNT buffer (20 mM HEPES, 150 mM NaCl, 10% glycerin, 20 mM NaH 2 PO 4 , 0.1% Triton X-100, pH 7.4), and 1 ml of the solution was used for immunoprecipitation with each individual antibody. After washing with HGNT buffer, the six different immunoprecipitated fragments were recovered by boiling of the Sepharose in SDS sample buffer. Proteins were separated by Tricine-SDS-PAGE (24) under reducing conditions, and [ 125 I-Tyr 8 ]bradykinin-labeled receptor fragments were identified by autoradiography of the gels for 3-6 days.

Affinity Labeling of the B 2 Receptor by [ 125 I-Tyr 8 ]Bradykinin-
We previously characterized domain-specific antibodies to the extracellular domains of the B 2 receptor. Antibodies to the amino half of ED3 N (Fig. 1, A and B) compete with bradykinin for binding to the B 2 receptor and are agonists (10). To further the analysis of the agonist's binding site(s), we used an affinity labeling-based approach. For affinity labeling, we chose [ 125 I-Tyr 8 ]bradykinin as B 2 agonist (Fig. 1B). This ligand binds to the B 2 receptor of HF-15 cells with a K D of 3 Ϯ 0.5 ϫ 10 Ϫ9 M. This value is close to the affinity of bradykinin, which binds to B 2 receptors of HF-15 cells with a K D value of 1.5 Ϯ 0.4 ϫ 10 Ϫ9 M. In contrast to [ 125 I]Tyr 0 bradykinin, which was used for previous studies (25), the N terminus of [Tyr 8 ]bradykinin is not modified.
[ 125 I-Tyr 8 ]bradykinin is an agonist, it stimulates similarly to bradykinin the phospholipase C pathway with an EC 50 of 0.8 Ϯ 0.2 ϫ 10 Ϫ9 M (not shown). This value was determined for the phospholipasc C-mediated rise in the intracellular free Ca 2ϩ concentration monitored by fura-2 (26). To detect a contact region between the agonist's N-terminal amino group and the B 2 receptor, we chose the homobifunctional amino group-specific crosslinker DST. Affinity labeling of the human B 2 receptor of HF-15 cells by [ 125 I-Tyr 8 ]bradykinin is shown in Fig. 2 (lane 1). The labeled B 2 receptor has a molecular mass of 69 Ϯ 4 kDa. This value is in good agreement with the molecular mass of the B 2 receptor determined previously (10). A 1000-fold molar excess of unlabeled [Tyr 8 ]bradykinin suppressed the labeling of the B 2 receptor (Fig. 2, lane 2 (Fig. 3A, lane 1). The anti-bradykinin antibodies stained a protein of 69 kDa. In some experiments a degradation product of 46 kDa was seen (not shown). Aggregation products of higher molecular mass are probably due to the high amount of B 2 receptor applied to each lane (10). For control, the affinity labeling was performed in the presence of 1, 10, 100, or 500 nM of the B 2 antagonist HOE140 (Fig. 3A,  lanes 2-5), which dose-dependently suppressed the subsequent staining with anti-bradykinin antibodies. This observation demonstrates that the anti-bradykinin antibodies cross-react with [Tyr 8 ]bradykinin-labeled B 2 receptors. The antibodies were used to enrich [ 125 I-Tyr 8 ]bradykinin-labeled B 2 receptors. The enriched B 2 receptor fraction was visualized by autoradiography after Tricine-SDS-PAGE (Fig. 3B, lane 1).
Chemical Cleavage of the B 2 Receptor by Cyanogen Bromide-To identify the region within the B 2 receptor to which [ 125 I-Tyr 8 ]bradykinin had been attached by DST, the affinityenriched B 2 receptor was chemically cleaved by cyanogen bro-mide. Cyanogen bromide specifically cleaves after methionine residues. The human B 2 receptor sequence contains 17 methionines, nine expected cyanogen bromide cleavage fragments vary between 2,000 and 12,000 Da, and four of these larger fragments cover the entire sequence of the putative extracellular domains (Fig. 1A). The cleaved B 2 receptor was separated by Tricine-SDS-PAGE, and autoradiography of the gel visualized [ 125 I]-labeled receptor fragments of about 6,000 Da and some aggregation products of 12,000 Da (Fig. 3B, lane 2). This experiment indicates that we efficiently cleaved the B 2 receptor, though we cannot completely rule out the possibility that some partial cleavage may have occurred.  Fig. 1A). Antibodies to a peptide ED1A (Fig. 1A) derived from the extended N-terminal region of the human B 2 receptor, which has been identified previously as the real start site of the human B 2 receptor protein (27) cells; they have been applied to enrich human and rat B 2 receptors for N-terminal protein sequencing (27). The affinities of the individual polyclonal domain-specific 125 I-labeled antibodies to the human B 2 receptor are similar, i.e. they vary between 30 and 40 nM (not shown). Cleavage sites of cyanogen bromide are not within the peptide sequences of ED1B, ED2, ED3 N , ED4 N , and ED4 C , which were used to raise antibodies ( Fig. 1A and Ref. 10). Hence immunoprecipitation is expected to be equally effective with these domain-specific antibodies. A cyanogen bromide cleavage site lies within ED3 C , and chemical cleavage may remove part of the epitope of the anti-ED3 C antibodies. This fact may reduce efficiency of immunoprecipitation with anti-ED3 C antibodies.

Immunoprecipitation of [ 125 I-Tyr 8 ]Bradykinin-labeled B 2 Receptor Fragments by Domain-specific Antibodies-After
After immunoprecipitation the immunoprecipitated proteins were separated by Tricine-SDS-PAGE under reducing conditions. Autoradiography of the gels visualized receptor fragments with covalently attached [ 125 I-Tyr 8 ]bradykinin (Fig. 4A,  lanes 3 and 4). For comparison, free [ 125 I-Tyr 8 ]bradykinin was included (Fig. 4A, lane 7). Antibodies to the third extracellular domain (anti-ED3 N and anti-ED3 C ) precipitated a 125 I-labeled receptor fragment with a molecular mass of 6 -8 kDa (Fig. 4A,  lanes 3 and 4), whereas antibodies to extracellular domains ED1B (ED1), ED2, ED4 N , and ED4 C did not enrich detectable amounts of [ 125 I-Tyr 8 ]bradykinin-labeled receptor fragments. The molecular mass of 6 -8 kDa of the labeled receptor fragment is in good agreement with the calculated molecular mass of the cyanogen bromide cleavage fragment derived from ED3 including the molecular mass of [ 125 I-Tyr 8 ]bradykinin. The amount of 125 I-labeled receptor fragments precipitated by ED3 C antibodies was less than the amount obtained with anti-ED3 N . Both antibodies bind to intact B 2 receptors with a similar affinity of 40 nM as determined with affinity-purified 125 Ilabeled antibodies (not shown). A cyanogen bromide cleavage site lies within the C-terminal portion of the peptide used to raise anti-ED3 C antibodies (Fig. 1A). Chemical cleavage may have removed part of the epitope of these antibodies, thereby explaining the lower efficiency in immunoprecipitation (see above). In addition to the receptor fragment of 6 -8 kDa, a 125 I-labeled fragment of about 2.5 kDa was seen in some experiments (Fig. 4A, lanes 3 and 4). This smaller fragment may represent a degradation product generated during the overnight incubation.
To control the specificity of immunoprecipitation, we performed the reaction in the presence of 10 M of the cognate peptides. The presence of the ED3 N or of the ED3 C peptide suppressed the immunoprecipitation with anti-ED3 N (Fig. 4B,  lane 2) or anti-ED3 C antibodies (Fig. 4B, lane 4), respectively. Together both experiments indicate that the N-terminal amino group of [ 125 I-Tyr 8 ]bradykinin was attached to an amino acid within the cyanogen bromide cleavage fragment containing ED3. For cross-linking we used the homobifunctional amino group-specific linker DST. Within ED3, Lys 172 is the only amino acid that provides a free primary amino group (Fig. 1B  and Ref. 7). This strongly suggests that the N-terminal amino group of [Tyr 8 ]bradykinin was attached by the amino groupspecific linker DST to the ⑀-amino group of Lys 172 .
Affinity Labeling Reaction Performed with the Wild-type Rat B 2 Receptor and with a K174A-mutated B 2 Receptor-The rat B 2 receptor protein sequence is 81% identical with the human B 2 receptor sequence (7). Therefore we asked whether DST attaches the N terminus of [Tyr 8 ]bradykinin to the analogous lysine within ED3 of the rat B 2 receptor, Lys 174 . We made a rat B 2 receptor mutant where Lys 174 was exchanged for alanine. The wild-type and the mutated B 2 receptor were transiently  (Fig. 5A, lane 2) B 2 receptor was confirmed in immunoblotting with anti-ED1B antibodies; anti-ED1B antibodies stained a protein of 55 Ϯ 4 kDa in COS-1 cells expressing wild-type or K174A-mutated B 2 receptor, whereas this protein was absent in mock-transfected COS-1 cells (Fig.  5A, lane 3). Differences in the apparent molecular mass determined after SDS-PAGE of B 2 receptors expressed in different cells (e.g. COS-1 cells, HF-15 cells) may be due to differential glycosylation (10). Intensities of the stained B 2 receptor bands were similar, indicating that equal amounts of wild-type and K174A-mutated B 2 receptors had been expressed in COS-1 cells. This was confirmed by binding studies with [ 3 H]bradykinin; COS-1 cells expressed 7.0 Ϯ 0.5 pmol/mg of protein of wild-type and 6.5 Ϯ 0.6 pmol/mg protein of K174A-mutated B 2 receptors.
We then performed affinity labeling of the wild-type B 2 receptor and of the K174A mutant by [ 125 I-Tyr 8 ]bradykinin. Proteins were separated by SDS-PAGE, and autoradiography of the gels visualized [ 125 I-Tyr 8 ]bradykinin-labeled B 2 receptors. Under the experimental conditions applied, DST did not attach detectable amounts of [ 125 I-Tyr 8 ]bradykinin to the K174A-mutated B 2 receptor (Fig. 5B, lane 3). In contrast, cross-linking of [ 125 I-Tyr 8 ]bradykinin to the wild-type receptor worked efficiently (Fig. 5B, lane 1). A protein of 56 Ϯ 3 kDa was labeled. The molecular mass of 56 kDa is in good agreement with the molecular mass of the B 2 receptor determined in immunoblotting with anti-ED1B antibodies (cf. Fig. 5A) The affinity labeling was B 2 receptor-specific because a 1000-fold molar excess of unlabeled ligand suppressed the labeling of the wild-type B 2 receptor (Fig. 5B, lane 2). This experiment with the wild-type and the K174A-mutated rat B 2 receptor strongly suggests that DST covalently attached the N-terminal amino group of [ 125 I-Tyr 8 ]bradykinin to the lysine within ED3, i.e. Lys 174 of the rat B 2 receptor.
With a combined approach of affinity labeling and immunoidentification of receptor fragments generated by cyanogen bromide cleavage, we identified Lys 172 within the amino half of ED3 (ED3 N ) of the human B 2 receptor to be located near the agonist's N-terminal amino group. This is in agreement with our previous finding that antibodies to ED3 N mutually compete with B 2 agonists for binding to the B 2 receptor (10). The lack to attach significant amounts of [ 125 I-Tyr 8 ]bradykinin to a rat B 2 receptor mutant where the analogous lysine Lys 174 had been replaced by alanine supports the observation made with the human B 2 receptor and indicates that [Tyr 8 ]bradykinin's N terminus is bound to the analogous position of the rat and of the human B 2 receptor. Previous work with heterobifunctional cross-linkers gave evidence that the bradykinin's N terminus was bound adjacent to a sulfhydryl group of the bovine B 2 receptor (28). Our data do not exclude such a possibility because (i) the cysteine providing the free sulfhydryl group is not identified, and due to the lack of a three-dimensional structure its potential distance to Lys 174 (or Lys 172 ) of the rat or human B 2 receptor is not determined, and (ii) we performed the crosslinking at 4°C and used the homobifunctional amino groupspecific cross-linker DST with a spacer of four carbon atoms (6.4 Å), whereas cross-linking of the bradykinin's N terminus to a sulfhydryl group of the B 2 receptor worked efficiently at room temperature with heterobifunctional linkers, e.g. m-maleimidobenzoyl-N-hydroxysuccinimide ester with a spacer arm between the two reactive groups of Ͼ9.9 Å (28).
A model of the rat B 2 receptor's agonist binding site suggests that the N-terminal amino and guanidino group of Arg 1 of bradykinin interact directly with negatively charged amino acids in extracellular domains ED3 and/or ED4 (14). This hypothesis was supported by rat B 2 receptor mutants. Replacing either Asp 268 or Asp 286 of ED4 with alanine reduces the affinity of bradykinin to the mutant receptors 19-or 28-fold, respectively (8), and a cluster mutation where the negatively charged amino acids of ED3, Asp 175 and Glu 178,179 , were exchanged for alanines resulted in a 12-fold loss in bradykinin affinity (8). But a direct interaction of bradykinin's N terminus with a distinct B 2 receptor domain had not been identified. Our data based on affinity labeling of the human and of the rat B 2 receptor give strong evidence that the N-terminal region of [Tyr 8 ]bradykinin is bound adjacent to Lys 172 or Lys 174 , respectively, of ED3. This finding is complementary to the observations made with anti-peptide antibodies, because domain-specific antibodies to ED3 N and bradykinin are mutually competitive in binding to the B 2 receptor (10). Furthermore our experiments with the human and rat B 2 receptor indicate that the location of [Tyr 8 ]bradykinin's N terminus is analoguous within the human and the rat B 2 receptor.
In addition to being a potential contact site, the interaction of the agonist bradykinin with ED3 N may be essential for the induction or stabilization of the active receptor conformation because anti-ED3 N antibodies are B 2 agonists (10). A refined model of the B 2 receptor's agonist binding site based on the finding that the N-terminal amino group of [Tyr 8 ]bradykinin is proximal to Lys 174 of the rat or Lys 172 of the human B 2 receptor may advance the rational design of B 2 agonists and antagonists (29) in the future.