INTRODUCTION

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 AT₁ 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) B₂ receptor revealed that the receptor for the peptide hormone bradykinin belongs to the family of G protein-coupled receptors. Mutation studies of the bradykinin B₂ 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 B₂ 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 B₂ 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 [¹²⁵I-Tyr⁸]bradykinin-labeled B₂ receptor fragments generated by cyanogen bromide cleavage to identify a region(s) where the agonist bradykinin contacts the B₂ receptor. We chose the amino group-specific homobifunctional cross-linker disuccinimidyl tartarate (DST) to localize the position of the N-terminal amino group of [¹²⁵I-Tyr⁸] bradykinin when bound to the B₂ 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 V₂ vasopressin receptor (12), or of the nicotinic acetylcholine receptor (13). In this study we used this technique to localize a B₂ receptor region involved in the interaction with the agonist's N terminus. This should contribute to a refinement of the present model of the B₂ receptor (14), thereby facilitating the rational design of B₂ agonists or antagonists.

EXPERIMENTAL PROCEDURES

Na[¹²⁵I] (17.5 Ci/mg) and the chemiluminescence detection kit (ECL) were from Amersham Corp.; [2,3-prolyl-3,4-³H]bradykinin (specific activity, 98 Ci/mmol) was from NEN DuPont; iodogen (1,3,4,6-tetrachloro-3-6-diphenyl-glycoluril) and DST were from Pierce; Sephadex G-50 and protein A-Sepharose were from Pharmacia Biotech Inc.; cyanogen bromide, Dowex AG1 × 8, and fluorescein isothiocyanate-conjugated goat anti-rabbit immunoglobulin were from Sigma; Centricon-30 filters were from Amicon; nitrocellulose sheets were from Schleicher & Schüll; Affi-Gel 10 was from Bio-Rad; LipofectAMINE was from Life Technologies, Inc. All other chemicals were of analytical grade.

HF-15 cells (15) expressing 0.5-1 pmol B₂ 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% CO₂/95% air atmosphere. COS-1 cells were transfected with wild-type and K174A-mutated rat B₂ receptor cDNA by LipofectAMINE (Life Technologies, Inc.) according to the manufacturer's instructions. The rat B₂ 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 pcDLSR296 (17). The mutagenesis cassette was completely sequenced to confirm the identity of the mutant.

Binding of [³H]bradykinin and of ¹²⁵I-labeled antibodies to intact HF-15 or COS-1 cells were performed as described previously (10).

Peptides (1 µg) dissolved in 100 µl of phosphate-buffered saline were incubated with 2 mCi of carrier-free Na[¹²⁵I] on a solid phase of Iodogen (100 µg/tube) for 10 min (18). Unreacted iodine was separated by anion exchange chromatography over Dowex AG1 × 8.

Synthesis of peptides and production and characterization of anti-peptide antibodies to the various extracellular domains of the B₂ 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 µg/ml of the peptide or of the conjugate.

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.

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(ab)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.).

For affinity labeling, B₂ receptors of HF-15 cells (50 pmol of receptor/100 mg of protein) were solubilized by 4 mM CHAPS, 20 mM PIPES, pH 7.4, including protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 µM leupeptin, 1 µM E64, 1 µM enalaprilate), and incubated for 60 min with 5 nM [¹²⁵I-Tyr⁸]bradykinin at 4 °C. Affinity labeling of B₂ receptors expressed in COS-1 cells was performed with membranes. Membrane preparation was done as described previously (10). For control a 1000-fold 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, B₂ receptor was visualized by autoradiography.

Affinity-purified anti-bradykinin antibodies were covalently coupled to Affi-Gel 10 (15 mg/ml gel). After affinity labeling of solubilized B₂ receptors by [¹²⁵I-Tyr⁸]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.

Affinity-purified [¹²⁵I-Tyr⁸]bradykinin-labeled B₂ receptors were chemically cleaved by cyanogen bromide as described previously (23).

B₂ receptor fragments were identified by immunoprecipitation with six different domain-specific antibodies. The antibodies were raised against peptides derived from the extracellular domains of the B₂ receptor (Ref. 10 and Fig. 1A). Each of the six different antibody fractions were bound to Protein A-Sepharose (50 µl of gel, 5-10 µg of affinity-purified IgG). The B₂ 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 NaH₂PO₄, 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 [¹²⁵I-Tyr⁸]bradykinin-labeled receptor fragments were identified by autoradiography of the gels for 3-6 days.

RESULTS AND DISCUSSION

We previously characterized domain-specific antibodies to the extracellular domains of the B₂ receptor. Antibodies to the amino half of ED3_N (Fig. 1, A and B) compete with bradykinin for binding to the B₂ 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 [¹²⁵I-Tyr⁸]bradykinin as B₂ agonist (Fig. 1B). This ligand binds to the B₂ receptor of HF-15 cells with a K_D of 3 ± 0.5 × 10⁹ M. This value is close to the affinity of bradykinin, which binds to B₂ receptors of HF-15 cells with a K_D value of 1.5 ± 0.4 × 10⁹ M. In contrast to [¹²⁵I]Tyr⁰bradykinin, which was used for previous studies (25), the N terminus of [Tyr⁸]bradykinin is not modified. [¹²⁵I-Tyr⁸]bradykinin is an agonist, it stimulates similarly to bradykinin the phospholipase C pathway with an EC₅₀ of 0.8 ± 0.2 × 10⁹ M (not shown). This value was determined for the phospholipasc C-mediated rise in the intracellular free Ca²⁺ concentration monitored by fura-2 (26). To detect a contact region between the agonist's N-terminal amino group and the B₂ receptor, we chose the homobifunctional amino group-specific cross-linker DST. Affinity labeling of the human B₂ receptor of HF-15 cells by [¹²⁵I-Tyr⁸]bradykinin is shown in Fig. 2 (lane 1). The labeled B₂ receptor has a molecular mass of 69 ± 4 kDa. This value is in good agreement with the molecular mass of the B₂ receptor determined previously (10). A 1000-fold molar excess of unlabeled [Tyr⁸]bradykinin suppressed the labeling of the B₂ receptor (Fig. 2, lane 2), indicating that we specifically labeled B₂ receptors.

To enrich [¹²⁵I-Tyr⁸]bradykinin-labeled B₂ receptor, we used affinity chromatography with anti-bradykinin antibodies. The antiserum cross-reacted with [Tyr⁸]bradykinin in enzyme-linked immunosorbent assay (not shown), its cross-reactivity with [Tyr⁸]bradykinin covalently attached to the B₂ receptor was verified in immunoblots of [Tyr⁸]bradykinin-labeled B₂ receptors (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 B₂ 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 B₂ 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 [Tyr⁸]bradykinin-labeled B₂ receptors. The antibodies were used to enrich [¹²⁵I-Tyr⁸]bradykinin-labeled B₂ receptors. The enriched B₂ receptor fraction was visualized by autoradiography after Tricine-SDS-PAGE (Fig. 3B, lane 1).

To identify the region within the B₂ receptor to which [¹²⁵I-Tyr⁸]bradykinin had been attached by DST, the affinity-enriched B₂ receptor was chemically cleaved by cyanogen bromide. Cyanogen bromide specifically cleaves after methionine residues. The human B₂ 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 B₂ receptor was separated by Tricine-SDS-PAGE, and autoradiography of the gel visualized [¹²⁵I]-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 B₂ receptor, though we cannot completely rule out the possibility that some partial cleavage may have occurred.

After chemical cleavage we identified the receptor fragments with covalently attached [¹²⁵I-Tyr⁸]bradykinin. To this end we immunoprecipitated the receptor fragments by six different domain-specific antibodies to the extracellular domains of the B₂ receptor (Ref. 10 and Fig. 1A). Antibodies to a peptide ED1A (Fig. 1A) derived from the extended N-terminal region of the human B₂ receptor, which has been identified previously as the real start site of the human B₂ receptor protein (27), have not been included in this study. The six different domain-specific antibodies immunoprecipitate the intact B₂ receptor of HF-15 cells; they have been applied to enrich human and rat B₂ receptors for N-terminal protein sequencing (27). The affinities of the individual polyclonal domain-specific ¹²⁵I-labeled antibodies to the human B₂ 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.

After immunoprecipitation the immunoprecipitated proteins were separated by Tricine-SDS-PAGE under reducing conditions. Autoradiography of the gels visualized receptor fragments with covalently attached [¹²⁵I-Tyr⁸]bradykinin (Fig. 4A, lanes 3 and 4). For comparison, free [¹²⁵I-Tyr⁸]bradykinin was included (Fig. 4A, lane 7). Antibodies to the third extracellular domain (anti-ED3_N and anti-ED3_C) precipitated a ¹²⁵I-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 [¹²⁵I-Tyr⁸]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 [¹²⁵I-Tyr⁸]bradykinin. The amount of ¹²⁵I-labeled receptor fragments precipitated by ED3_C antibodies was less than the amount obtained with anti-ED3_N. Both antibodies bind to intact B₂ receptors with a similar affinity of 40 nM as determined with affinity-purified ¹²⁵I-labeled 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 ¹²⁵I-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 µM 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 [¹²⁵I-Tyr⁸]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, Lys¹⁷² 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 [Tyr⁸]bradykinin was attached by the amino group-specific linker DST to the -amino group of Lys¹⁷².

The rat B₂ receptor protein sequence is 81% identical with the human B₂ receptor sequence (7). Therefore we asked whether DST attaches the N terminus of [Tyr⁸]bradykinin to the analogous lysine within ED3 of the rat B₂ receptor, Lys¹⁷⁴. We made a rat B₂ receptor mutant where Lys¹⁷⁴ was exchanged for alanine. The wild-type and the mutated B₂ receptor were transiently expressed in COS-1 cells. The mutation did not change the affinity of bradykinin and HOE140 compared with the wild-type B₂ receptor. The K_D values of bradykinin on both receptors were 0.5 ± 0.1 × 10⁹ M, and they were 0.4 ± 0.05 × 10⁹ M of HOE140 (not shown). Expression of the wild-type (Fig. 5A, lane 1) and of the mutated (Fig. 5A, lane 2) B₂ receptor was confirmed in immunoblotting with anti-ED1B antibodies; anti-ED1B antibodies stained a protein of 55 ± 4 kDa in COS-1 cells expressing wild-type or K174A-mutated B₂ 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 B₂ receptors expressed in different cells (e.g. COS-1 cells, HF-15 cells) may be due to differential glycosylation (10). Intensities of the stained B₂ receptor bands were similar, indicating that equal amounts of wild-type and K174A-mutated B₂ receptors had been expressed in COS-1 cells. This was confirmed by binding studies with [³H]bradykinin; COS-1 cells expressed 7.0 ± 0.5 pmol/mg of protein of wild-type and 6.5 ± 0.6 pmol/mg protein of K174A-mutated B₂ receptors.

We then performed affinity labeling of the wild-type B₂ receptor and of the K174A mutant by [¹²⁵I-Tyr⁸]bradykinin. Proteins were separated by SDS-PAGE, and autoradiography of the gels visualized [¹²⁵I-Tyr⁸]bradykinin-labeled B₂ receptors. Under the experimental conditions applied, DST did not attach detectable amounts of [¹²⁵I-Tyr⁸]bradykinin to the K174A-mutated B₂ receptor (Fig. 5B, lane 3). In contrast, cross-linking of [¹²⁵I-Tyr⁸]bradykinin to the wild-type receptor worked efficiently (Fig. 5B, lane 1). A protein of 56 ± 3 kDa was labeled. The molecular mass of 56 kDa is in good agreement with the molecular mass of the B₂ receptor determined in immunoblotting with anti-ED1B antibodies (cf. Fig. 5A) The affinity labeling was B₂ receptor-specific because a 1000-fold molar excess of unlabeled ligand suppressed the labeling of the wild-type B₂ receptor (Fig. 5B, lane 2). This experiment with the wild-type and the K174A-mutated rat B₂ receptor strongly suggests that DST covalently attached the N-terminal amino group of [¹²⁵I-Tyr⁸]bradykinin to the lysine within ED3, i.e. Lys¹⁷⁴ of the rat B₂ receptor.

With a combined approach of affinity labeling and immunoidentification of receptor fragments generated by cyanogen bromide cleavage, we identified Lys¹⁷² within the amino half of ED3 (ED3_N) of the human B₂ 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 B₂ agonists for binding to the B₂ receptor (10). The lack to attach significant amounts of [¹²⁵I-Tyr⁸]bradykinin to a rat B₂ receptor mutant where the analogous lysine Lys¹⁷⁴ had been replaced by alanine supports the observation made with the human B₂ receptor and indicates that [Tyr⁸]bradykinin's N terminus is bound to the analogous position of the rat and of the human B₂ 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 B₂ 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 Lys¹⁷⁴ (or Lys¹⁷²) of the rat or human B₂ receptor is not determined, and (ii) we performed the cross-linking at 4 °C and used the homobifunctional amino group-specific 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 B₂ 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 >9.9 Å (28).

A model of the rat B₂ receptor's agonist binding site suggests that the N-terminal amino and guanidino group of Arg¹ of bradykinin interact directly with negatively charged amino acids in extracellular domains ED3 and/or ED4 (14). This hypothesis was supported by rat B₂ receptor mutants. Replacing either Asp²⁶⁸ or Asp²⁸⁶ 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, Asp¹⁷⁵ 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 B₂ receptor domain had not been identified. Our data based on affinity labeling of the human and of the rat B₂ receptor give strong evidence that the N-terminal region of [Tyr⁸]bradykinin is bound adjacent to Lys¹⁷² or Lys¹⁷⁴, 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 B₂ receptor (10). Furthermore our experiments with the human and rat B₂ receptor indicate that the location of [Tyr⁸]bradykinin's N terminus is analoguous within the human and the rat B₂ 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 B₂ agonists (10). A refined model of the B₂ receptor's agonist binding site based on the finding that the N-terminal amino group of [Tyr⁸]bradykinin is proximal to Lys¹⁷⁴ of the rat or Lys¹⁷² of the human B₂ receptor may advance the rational design of B₂ agonists and antagonists (29) in the future.