The N terminus of bradykinin when bound to the human bradykinin B2 receptor is adjacent to extracellular Cys20 and Cys277 in the receptor.

Chemical cross-linking combined with site-directed mutagenesis was used to evaluate the role of extracellular cysteines and their positions relative to the binding site for the agonist bradykinin (BK) in the human BK B2 receptor. All extracellular cysteines, Cys20, Cys103, Cys184, and Cys277, in the receptor were mutated to serines, and single and double mutants were transfected into COS-7 cells. The Ser20 and Ser277 single mutants and the Ser20/Ser277 double mutant bound [3H]BK and the antagonist [3H]NPC17731 with pharmacological profiles identical to the wild-type B2 receptor. In contrast, the Ser103 and Ser184 single mutants were unable to bind either of the two radioligands. However, these mutants were still expressed as determined by immunoblotting with anti-B2 receptor antibodies. Previous studies on the bovine B2 receptor showed that bifunctional reagents, which are reactive to amines at one end and to free sulfhydryls in the opposite end, cross-link the N terminus of receptor-bound BK to a sulfhydryl in the receptor (Herzig, M. C. S., and Leeb-Lundberg, L. M. F. (1995) J. Biol. Chem. 270, 20591-20598). Here, we show that m-maleimidobenzoyl-N-hydroxysuccinimide ester and 1,5-difluoro-2,4-dinitrobenzene cross-linked BK to the wild-type human B2 receptor and the Ser20 and Ser277 single mutant receptors, whereas these reagents were unable to cross-link BK to the Ser20/Ser277 double mutant. These results show that Cys103 and Cys184 are both required for expression of high affinity agonist and antagonist binding sites in the human B2 receptor, while Cys20 and Cys277 are not required. Furthermore, the results provide direct biochemical evidence that the N terminus of BK, when bound to the B2 receptor, is adjacent to Cys277 in extracellular domain 4 and Cys20 in extracellular domain 1 of the receptor.

BK 1 (Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg), a member of a family of potent vasoactive peptides called kinins, acts through BK B2 receptors (1,2). Cloning of the cDNA for the B2 receptor in human (3,4), rat (5), mouse (6), and rabbit (7) has revealed that this receptor belongs to the seven transmembrane-domain GPCR superfamily. These receptors have an extracellular N terminus named extracellular domain 1 (EC-I), an intracellular C terminus named intracellular domain 4 (IC-IV), and seven highly conserved transmembrane domains (TM-I-VII) interrupted by three intracellular loops (IC-I-III) and three extracellular loops (EC-II-IV) which alternate. The delineation of the domains in GPCR that are involved in the binding of ligands is a central question in the understanding of these receptors and crucial in the development of specific receptor antagonists. GPCRs for cationic amines bind agonists using residues located exclusively within transmembrane domains (8), while those for glycoproteins bind agonists using residues located primarily in the extracellular domains (9). On the other hand, the binding of peptide agonists to GPCRs appears to involve residues located in both extracellular and transmembrane domains (8,10,11).
Only limited progress has been made in delineating the binding site(s) for BK in the B2 receptor. Using site-directed mutagenesis, residues located near the extracellular surface in TM-VI were shown to be important for high affinity agonist binding (12,13). Furthermore, agonist binding was shown also to be critically dependent on specific charged residues in EC-IV (10). The involvement of extracellular residues in the binding of BK to the B2 receptor is not surprising considering the hydrophilic nature of this peptide. Interestingly, most mutations in the B2 receptor that affect agonist affinity only minimally affect antagonist affinity. Thus, the agonist and antagonist binding sites in the B2 receptor do not appear to be identical and may overlap only partially if at all.
Recently, we found that the N terminus of BK when bound to the B2 receptor in bovine myometrial membranes can be crosslinked by heterobifunctional cross-linkers to a sulfhydryl residue(s) in the receptor (14). The sequence of the bovine B2 receptor is unknown. However, the sequences of the human (3,4), rat (5), mouse (6), and rabbit (7) B2 receptors contain a conserved cysteine in each extracellular domain, as well as several cysteines in the transmembrane and intracellular domains. Considering that a sulfhydryl is located within the radii of the linker arms of the cross-linkers attached to the N terminus of the receptor-bound BK (14), we concluded that crosslinking studies on B2 receptors in which cysteines have been mutated to serines should enable us to identify directly the positioning of the N terminus of BK relative to specific cysteine residues in the receptor. In this report, we show the results of radioligand binding and cross-linking studies on WT human B2 receptors and receptors in which individual extracellular cysteines were replaced with serines.
Cell Culture-COS-7 cells (ATCC, Bethesda, MD) were seeded onto 10-cm dishes at a concentration of 8 ϫ 10 5 cells/plate and grown in growth medium (Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, glutamine, and 1% penicillin/streptomycin) for 24 h in 5% CO 2 at 37°C. A stable cell line expressing the human WT B2 receptor was prepared and maintained according to the following protocol. CHO cells were seeded into six-well plates at a density of 1 ϫ 10 5 cells/ml and allowed to grow for 24 h in complete medium (Ham's F-12 supplemented with 10% fetal bovine serum, glutamine, and 1% penicillin/streptomycin). B2 receptor cDNA (1 g/well) was then transfected into the cells utilizing Lipofectin as described by the manufacturer (Life Technologies, Inc.). Following transfection, stable cell lines were selected by continuing growth in complete medium supplemented with 500 g/ml G418. Monoclonal cell lines were established by serial dilution of cultures verified in radioligand binding assays to be expressing the B2 receptor. Stable cell lines were subsequently maintained in complete medium supplemented with G418.
Construction of Serine Mutants-The human B2 receptor cDNA was subcloned as a HindIII-XbaI fragment into the mutagenic vector pAL-TER-1 (Promega, Madison, WI). Single-stranded DNA was obtained and single amino acid mutations introduced as described in the manufacturer's protocol. Oligonucleotides designed to convert cysteine residues to serines had the following sequences: Ser 20 , 5Ј-CACTTGGGGG-CTTTTGCTCTGCGCAAAGGTCCC-3Ј; Ser 103 , 5Ј-GGCATTCACCACC-CGGGAGAGCGTCTC-3Ј; Ser 184 , 5Ј-GTAGCTGATGACGCTAGCGGT-GACGTT-3Ј; Ser 277 , 5Ј-CTCGTCCTGGCTGCTGGAGAGGAT-3Ј. Following verification of the mutations by restriction enzyme analysis, the receptor cDNAs were subcloned as a HindIII-XbaI fragment into the expression vector pCDNA3 (InVitrogen, San Diego, CA). The Ser 20 /Ser 277 double mutant was constructed by cutting the Ser 20 and Ser 277 single mutants with MroI and ligating the appropriate fragments containing the mutations together. Sequencing of each construct was then carried out utilizing a dideoxy sequencing kit (United States Biochemical Corp.).
Transient Transfection of COS-7 Cells-The appropriate plasmid DNA (17 g) was added to a tube containing 500 l of 250 mM CaCl 2 and mixed well. A 500-l aliquot of 2 ϫ balanced salt solution (50 mM BES, 280 mM NaCl, 1.5 mM Na 2 HPO 4 , pH 6.95) was added, and again the tubes were mixed well and allowed to stand at room temperature for 7 min. At 24 h following seeding of COS-7 cells in a 10-cm dish, the entire content of a tube was added to the cells, and the cells were incubated for an additional 18 h in 3% CO 2 at 37°C. The medium was then replaced with fresh growth medium, and the incubation was continued for 48 h in 5% CO 2 at 37°C. The growth medium was removed, and the cells were washed once with PBS without Ca 2ϩ and Mg 2ϩ . Cells were collected in PBS and centrifuged at 2,000 ϫ g for 10 min.
Membrane Preparation-Pellets were resuspended in buffer (25 mM TES, 1 mM 1,10-phenanthroline, pH 6.8) and homogenized using a Polytron at setting 5 for 10 s. Membranes were isolated by centrifugation at 48,000 ϫ g for 10 min at 4°C. The pellets were then resuspended in the above buffer containing, in addition, 0.1% bovine serum albumin and 0.014% bacitracin. Membrane preparations were aliquoted and stored at 1-3 mg of protein/ml at Ϫ80°C.
Bovine myometrial membranes were prepared by a procedure originally described by Frederick et al. (15) and subsequently modified by Herzig and Leeb-Lundberg (14).
Radioligand Binding-Radioligand binding assays were done essentially as described previously (14). Membrane preparations were thawed and diluted in binding buffer (25 mM TES, 0.5 mM EDTA, 1 mM CaCl 2 , 1 mM 1,10-phenanthroline, 0.014% bacitracin, 0.1% bovine serum albumin, pH 6.8) to give a signal of 1,500 -3,000 dpm/assay of specific [ 3 H]BK binding. This signal required 5-60 g of protein/ml of the WT and the various mutant receptors. Binding assays were initiated by addition of radioligand (50 l) with and without excess nonradiolabeled ligand to the receptor preparation (450 l). After incubation for 60 -90 min at 24°C, assays were terminated by dilution with 4 ml of ice-cold PBS and rapid vacuum filtration on Whatman GF/C filters previously soaked in 1% polyethyleneimine. The trapped membranes were then washed with an additional 2 ϫ 4 ml of ice-cold PBS. Filters were counted for radioactivity in a Beckman LS5000TD scintillation counter.
Protein Cross-linking-Membranes were thawed and diluted to 5-60 g/ml in cross-linking buffer (identical to binding buffer but with a pH ϭ 7.2) and allowed to bind ligand for 60 -90 min at 24°C as described above before addition of dimethyl sulfoxide (Յ2% final concentration) with and without cross-linker; the incubation was continued for an additional 10 min. The cross-linking reaction was then quenched by a 1:1 dilution with 2 M glycine in cross-linking buffer. In order to dissociate noncovalently bound radioligand, quenched samples were diluted 4-fold with buffer containing 1 M BK and 10 M Gpp(NH)p and incubated for Ն180 min at 24°C, followed by filtration on GF/C filters as described above. Cross-linked ligand was the amount of radioligand remaining after dissociation. Specific cross-linked ligand was crosslinked radioligand minus nonspecific binding (as determined in the presence of 1 M BK in the binding assay). Cross-linking efficiency was the specific cross-linking as a percent of total specific binding.
Immunoblotting-Membrane proteins (10 mg) were solubilized with 20 mM CHAPS, 20 mM PIPES, pH 6.8, including 1 mM phenylmethylsulfonyl fluoride, 2 M leupeptin, 2 M E64, and 5 mM EDTA for 20 min at 24°C. Acetone-precipitated proteins were dissolved in SDS-polyacrylamide gel electrophoresis sample buffer, boiled for 5 min, and subjected to SDS-polyacrylamide gel electrophoresis on 10% gels. Proteins were immunoblotted as described (16) using anti-B2 receptor antibodies. Immunoreactive bands were visualized with ECL according to procedures described by the supplier (Amersham Corp.).
Other Assays-Proteins were assayed using the method of Lowry (17).

Expression of WT and Mutant B2
Receptors-The expression of high affinity agonist and antagonist binding sites in the WT and serine mutant receptors was analyzed by radioligand binding using [ 3 H]BK and [ 3 H]NPC17731, respectively. As shown in Table I, both ligands bound with high affinities to the WT   (Table I). With each mutant, the higher ratio was due to an increase in the K D for agonist binding that was matched by a decrease in the K D for antagonist binding.

H]BK and [ 3 H]NPC17731 binding to WT and serine mutant B2 receptors in COS-7 cells
In contrast to the Ser 20 and Ser 277 mutants, the Ser 103 and Ser 184 single mutants were unable to bind either radioligand. To ensure that the Ser 103 and Ser 184 single mutant receptors were expressed, these mutants were probed by Western blot analysis with anti-B2 receptor antibodies. Fig. 1 shows that the Ser 103 (lane 3) and Ser 184 (lane 4) mutant receptors as well as the WT receptor (lane 2) were expressed. As controls, nontransfected COS-7 cells (Fig. 1, lane 1) and B2 receptors from human foreskin fibroblasts (lane 5) were also immunoblotted. These results show that mutation of Cys 20 and Cys 277 to serines in the B2 receptor results in a receptor with retained high affinity binding sites for agonists and antagonists, while mutation of Cys 103 and Cys 184 to serines completely interferes with the folding of the receptor to form proper binding sites for these classes of ligands.
Inclusion of 100 M Gpp(NH)p, a non-hydrolyzable analog of GTP, in the radioligand binding assay decreased the specific binding of [ 3 H]BK to WT, Ser 20 , Ser 277 , and Ser 20 /Ser 277 by 28 Ϯ 13%, 15 Ϯ 4%, 24 Ϯ 13%, and 41 Ϯ 4%, respectively (mean Ϯ S.E., n Ն 3). These results show that mutation of Cys 20 and Cys 277 to serines either individually or in combination does not interfere with the coupling of the B2 receptor to a G-protein.
Pharmacological Specificity of BK Binding to WT and Mutant B2 Receptors-In order to determine if the replacement of extracellular cysteines with serines in the B2 receptor alters the pharmacological profile of the receptor, the ability of various kinin agonists and a kinin antagonists to compete for [ 3 H]BK binding to the receptors was evaluated. As shown in Fig. 2A, [ 3 H]BK binding to the WT receptor was competed for by BK, a B2 receptor-specific agonist, and HOE140, a B2 receptor-specific antagonist, but not by the B1 receptor-specific agonist des-Arg 9 -BK. Fig. 2 (B-D) shows that the Ser 20 , Ser 277 , and Ser 20 /Ser 277 mutant receptors displayed pharmacological profiles virtually identical to the WT receptor. Thus, replacement of extracellular Cys 20 and Cys 277 with serines either individually or in combination in the human B2 receptor does not influence the pharmacological profile of the receptor.
Agonist Cross-linking to WT and Mutant B2 Receptors-We showed previously that the N terminus of BK when bound to the bovine B2 receptor can be cross-linked to a sulfhydryl residue in the receptor via the cross-linkers MBS and DFDNB (14). Here, we attempted to identify, using the human B2 receptor, the exact cysteine residue(s) to which BK can be cross-linked. The effectiveness of MBS in cross-linking BK to the receptors was assessed in [ 3 H]BK binding experiments (Fig. 3)   maximum cross-linking efficiencies of the Ser 20 and Ser 277 mutants were 20.6 Ϯ 5.3% and 16.3 Ϯ 3.0%, respectively. These results indicate that mutation of either Cys 20 or Cys 277 to a serine did not interfere with the ability of MBS to cross-link BK to the receptor. In contrast, no cross-linking of BK to the Ser 20 /Ser 277 mutant was observed. Fig. 4 shows that mutation of either of these cysteines to a serine also did not interfere with the ability of 1 mM DFDNB to cross-link BK, whereas, again, no significant cross-linking of BK to the Ser 20 /Ser 277 mutant was observed. These results show that either Cys 20 or Cys 277 can provide the sulfhydryl necessary for MBS and DFDNB to cross-link BK to the human B2 receptor and excludes any other sulfhydryl. In agreement with previous studies of the B2 receptor in bovine myometrial membranes (14), MBS was unable to cross-link the antagonist [ 3 H]NPC17731 when bound to either the wild-type or mutant human receptors (data not shown).

DISCUSSION
In this study, we analyzed the role of the four cysteines located on the extracellular surface of the BK B2 receptor and their positioning relative to the N terminus of BK when bound to the receptor. Our results show that Cys 103 in EC-II and Cys 184 in EC-III are essential for formation of high affinity agonist and antagonist binding sites in the receptor, while Cys 20 in the EC-I and Cys 277 in EC-IV are not essential. We further demonstrate that either Cys 20 or Cys 277 in the B2 receptor provides the sulfhydryl necessary for the bifunctional reagents MBS and DFDNB to cross-link receptor-bound BK to the receptor. Consequently, these residues must be located within the radius of the linker arms of MBS (9.9 Å) and DFDNB (3 Å) when attached to the N terminus of the receptorbound BK.
Two of the four extracellular cysteines in the B2 receptor, Cys 103 in EC-II and Cys 184 in EC-III (Fig. 5), are conserved in most members of the GPCR superfamily of which the sequence is known (8,18). Studies with two members, rhodopsin and the ␤-adrenergic receptor, have revealed that these two cysteines are required for proper expression of the receptor (19 -22). Most likely, these residues are linked in a disulfide bond that stabilizes the correctly folded conformation of the receptors. The Ser 103 and Ser 184 mutants of the B2 receptor were expressed in the membrane of COS-7 cells. However, these mutants did not express any high affinity binding sites for either the agonist BK or the antagonist NPC17731. Thus, in consensus with other members of the GPCR family, the B2 receptor appears to be absolutely dependent on these cysteines for expression of high affinity ligand binding sites. Conversely, Cys 20 in EC-I and Cys 277 in EC-IV are semi-conserved in the GPCR family and are found in the BK B2 receptor (3)(4)(5)(6), the BK B1 receptor (23,24), the neuropeptide Y1 receptor (25), the angiotensin II type 1 receptor (26), the endothelin receptors (27,28), and the interleukin 8 receptor (29). In the current secondary structure model of GPCR based originally on the crystal structure of bacteriorhodopsin (30,31) and since modified by the recent crystal structure of rhodopsin (32) and hydropathicity analysis of over 200 members of the receptor superfamily (33), the seven transmembrane helices are believed not to be in a linear array but rather to be folded back upon themselves juxtapositioning TM-I and TM-VII and bringing the EC-I in relatively close proximity to EC-IV (Fig. 5). Thus, it is reasonable to suspect that a disulfide bond may exist between the cysteines in EC-I and in EC-IV and may be required to stabilize the folding of these GPCRs. The Ser 20 and Ser 277 mutants of the B2 receptor were expressed and displayed pharmacological profiles identical to that of the WT receptor. Thus, a disulfide bond between Cys 20 and Cys 277 in the B2 receptor, if present, is not critical for expression of high affinity agonist and antagonist binding sites. Furthermore, these residues do not appear to be involved directly as determinants in either agonist or antagonist binding to this receptor. Dithiothreitol had little effect on the binding of BK to the B2 receptor in both bovine myometrial membranes (14) and intact rat myometrial cells. 2 Thus, even though disulfide bonds between some extracellular cysteines in the B2 receptor may be crucial during insertion of the receptor into the membrane for formation of proper binding sites, reducing such bonds in receptors already expressed in the membrane does not denature the agonist binding site. In contrast, expression of angiotensin II type 1 receptors in which the cysteines in EC-I and EC-IV are replaced with glycines resulted in a 10-fold decrease in angiotensin II affinity (34). Furthermore, the binding of angiotensin II to the WT receptor is sensitive to dithiothreitol. Thus, high affinity agonist binding to the angiotensin II type 1 receptor may be dependent on a disulfide between the cysteines in EC-I and EC-IV. The binding of losartan, a high affinity non-peptide angiotensin II type 1 receptor antagonist, was not sensitive to mutation of the above cysteines (34).
Interestingly, the affinity of agonist binding decreased and the affinity of antagonist binding increased upon mutation of Cys 20 and Cys 277 to serines either singly or in combination. According to a three-state model of agonist and antagonist binding to the B2 receptor that we proposed (35,36), these results suggest that these mutations favor the antagonist-stabilized inactive state of the receptor. This conclusion was supported by the increased Gpp(NH)p sensitivity of BK binding to at least one of the mutants, Ser 20 /Ser 277 . As the shift in agonist and antagonist affinity was observed with both the single and double mutants, it is possible that the normal equilibrium between conformational states of the receptor requires a disulfide bond between Cys 20 and Cys 277 .
The cross-linking studies presented here using the human B2 receptor are a direct extension of our previous studies using the B2 receptor in bovine myometrial membranes (14). In the previous studies, we showed that heterobifunctional reagents such as MBS, which are reactive to amines at one end and to free sulfhydryls in the opposite end, and DFDNB, which is usually considered to be homobifunctional in activity and structure, cross-link the N terminus of receptor-bound BK to a sulfhydryl in the receptor. Consequently, we concluded that the N terminus of BK when bound to the receptor is adjacent to a cysteine in the receptor. The agonist binding site on the B2 receptor was distinguished from the antagonist binding site by the fact that these cross-linking reagents were completely unable to cross-link either NPC17731 or HOE140, two high affinity receptor antagonists, to the receptor. Molecular modeling studies (37), which are supported by experimental site-directed mutagenesis studies (10), suggest that the N terminus of BK when bound to the receptor is extracellular. Consequently, four cysteine residues, Cys 20 , Cys 103 , Cys 184 , and Cys 277 , in the receptor are candidates for providing the sulfhydryls necessary for MBS and DFDNB to cross-link BK to the receptor. As described above, two of these cysteines, Cys 103 and Cys 184 , are presumed to be involved in a disulfide bond and, consequently, are not believed to be accessible for reaction with the crosslinkers. By identifying which, if any, of the above cysteines are critical for cross-linking, we would be able to identify which extracellular domains are adjacent to the BK N terminus. In this study, we show that the BK N terminus can be cross-linked by MBS and DFDNB to the receptor through both Cys 20 in the EC-I and Cys 277 in EC-IV. Indeed, only with the substitution of serine residues for both of these cysteine residues could the cross-linking of BK to the B2 receptor by MBS and DFDNB be prevented. These results exclude all other cysteines present in the receptor or any non-receptor cysteines in the membrane as anchors for the cross-linking. Therefore, we conclude that both EC-I and EC-IV of the receptor are positioned near the N terminus of BK when located in the agonist binding site. These biochemical results directly support the most recent structural model of GPCR described above in which the seven transmembrane helices are folded back upon themselves so that TM-I is adjacent to TM-VII. Indeed, without such an oval array of the seven transmembrane helices, it is difficult to envision how both EC-I and EC-IV could be within reach of the linker arm of the cross-linkers.
There is no immediate consensus for the exact location of the agonist and antagonist binding sites in the B2 receptor. However, several studies using different techniques including chemical cross-linking (14) and site-directed mutagenesis (10,12,13) show that these two sites are not identical and may be only partially overlapping. Mutagenesis studies using the rat B2 receptor have revealed some information regarding the possible location of the agonist binding site. Individual mutations of two conserved aspartate residues, Asp 268 and Asp 286 in EC-IV, which correspond to Asp 266 and Asp 284 in the human receptor (Fig. 5), to alanines resulted in a 17-and 25-fold decrease in BK affinity, respectively, and the double mutation resulted in a 500-fold decrease (10). These results are in agreement with a model of BK bound to the B2 receptor proposed by Kyle based on structural homology modeling, molecular modeling, and systematic conformational searching methods of BK and the receptor (37). In this model, Asp 268 and Asp 286 in the receptor interact electrostatically with the N-terminal amino group, the guanidinyl side chain, or both of Arg 1 in BK, a residue absolutely crucial for receptor binding. Immediately below EC-IV in TM-VI of the rat receptor residues crucial for high affinity agonist binding have also been identified. Replacement of Phe 261 with alanine decreased agonist affinity by 1,600-fold (12, 13), while replacement of Thr 265 with either alanine or valine decreased agonist affinity by 700-fold (12). A FIG. 5. Structural model of the BK B2 receptor. In a structural model of the B2 receptor, the receptor is viewed from the extracellular side of the membrane and folded so as to juxtapose TM-I and TM-VII as described for rhodopsin (32). Cys 20 , Cys 103 , Cys 184 , and Cys 277 are indicated by letter C and number, and Asp 266 and Asp 284 are indicated by letter D. The receptor N terminus is indicated by bold letter N. more modest decrease of 10-fold in the agonist affinity was observed when Gln 262 and Thr 267 were replaced with alanines (12). These studies suggest that residues located near the extracellular side in TM-VI and in EC-IV are parts of an agonist binding site in the B2 receptor. Anti-peptide antibodies developed against the C-terminal half of EC-IV of the human B2 receptor inhibited agonist binding providing further support for the involvement of this domain in this interaction. Interestingly, antibodies against the N-terminal half of EC-III inhibited both agonist and antagonist binding. Thus, determinants in EC-III and possibly in TM-IV and TM-V may be involved in binding both classes of ligands (16).
In conclusion, based on molecular modeling and site-directed mutagenesis, the consensus is evolving that regions in EC-IV and TM-VI are critical in the binding of agonists to the B2 receptor. However, these approaches clearly have limitation. The former approach is purely theoretical, while the latter approach may be either an effect of altering directly a ligand binding determinant in the receptor or an effect of a conformational change in the receptor, which is induced at a site distant to the binding site. Our approach using chemical cross-linking combined with site-directed mutagenesis probes residues that are located at a distance no greater than the linker arm of the cross-linking reagent from the bound ligand. The results that we have obtained indicate that the cysteines in EC-IV and EC-I of the B2 receptor can exist at a distance as short as 3 Å from the N terminus of the bound BK. Considering the current secondary structure model of GPCR, our results directly support the positioning of the N terminus of BK bound to the B2 receptor at the extracellular surface below EC-IV. Cys 277 is positioned 11 residues beyond Asp 266 and 7 residues before Asp 284 , residues believed to be located at the interface between TM-VI and TM-VII, respectively. This pair of acidic residues is also thought to participate in an electrostatic interaction(s) with the N-terminal Arg 1 in BK. The straight chain distance between Asp 266 and Cys 277 , which is about 40 Å, and between Asp 284 and Cys 277 , which is about 30 Å, indicate that EC-IV must be folded to align this cysteine and these aspartates with the BK N terminus. However, given the fact that both MBS and DFDNB can cross-link BK to these cysteines, there must be some flexibility either in the MBS linker arm or in EC-IV and EC-I. Interestingly, DFDNB cross-linking of BK to the Ser 277 mutant was consistently less than that to the Ser 20 mutant. Indeed, this result is not unexpected with this short crosslinker, since according to our model Cys 20 may be located slightly further away than Cys 277 from the BK N terminus. Finally, NPC17731 could not be cross-linked either to the WT or mutant receptors under any conditions. When taken into consideration with the lack of effect of mutation of the aspartates in EC-IV on NPC17731 binding (13), these results suggest that antagonist binding to the B2 receptor involves at least some determinants that are different from those involved in agonist binding.