The known structure-activity relationship and docking models for peptide ligands of the bradykinin B2 receptor indicate a certain tolerance to N-terminal extension. We took advantage of this by generating two fluorescent bradykinin analogs containing 5(6)-carboxyfluorescein (CF) optionally used with the ε-aminocaproyl spacer condensed at the N terminus of the agonist. Pharmacological studies indicated that CF-bradykinin was virtually inactive as a B2 receptor ligand and agonist, whereas CF-ε-aminocaproyl-bradykinin (CF-εACA-BK) was 400- to 1000-fold less potent than bradykinin (competition of [3H]bradykinin binding to B2 receptors, contractility of the human isolated umbilical vein). Nevertheless, CF-εACA-BK (5 μM) was taken up by human embryonic kidney 293a cells expressing recombinant B2 receptors, but not by those cotreated with an antagonist or expressing a truncated receptor that is pharmacologically intact but not phosphorylable. A higher-affinity CF-conjugated peptide, the antagonist CF-εACA-d-Arg-[Hyp3,Igl5,d-Igl7,Oic8]-bradykinin (B-10380), labeled both intact and truncated receptor forms at the cell surface. The fluorescent agonist CF-εACA-BK was found in vesicles positive for β-arrestin1, Rab5, and Rab7, then apparently degraded as a function of time because the fluorescence was transferred from the vesicles to the cytosol in a vesicular-ATPase-dependent process (3 h). The ectopeptidase angiotensin-converting enzyme (ACE) is a major kininase. The binding affinity of CF-εACA-BK for this carboxydipeptidase is identical to that of bradykinin ([3H]enalaprilat displacement assay). Recombinant ACE is essentially a plasma membrane protein in CF-εACA-BK imaging of intact cells. Micromolar CF-εACA-BK is a probe for the two major physiological targets of bradykinin, the B2 receptor and ACE. As an agonist, it is subjected to β-arrestin-mediated endocytosis, trafficking, and subsequent ligand degradation.
Bradykinin is a blood-derived nonapeptide that stimulates the widely distributed B2 receptors and is degraded by a number of peptidases, among which is angiotensin-converting enzyme (ACE) (Leeb-Lundberg et al., 2005). It is well established that, after agonist stimulation, the bradykinin B2 receptor undergoes phosphorylation of serine residues (Blaukat et al., 1996, 1999), which constitute a recognizable site for the recruitment of the endocytic machinery by interaction with β-arrestins (Pizard et al., 1999; Hamdan et al., 2007). Endocytosis of the agonist-stimulated B2 receptors, like that of many G protein-coupled receptors that are G protein-coupled receptor kinase (GRK) substrates, involves the transport and subsequent degradation of the ligand bradykinin (Leeb-Lundberg et al., 2005). [3H]Bradykinin (10 nM) is transported at the sizeable rate of 758 fmol/(15 min) in 75-cm2 flasks by HEK 293 cells stably expressing a B2 receptor-green fluorescent protein (B2R-GFP) conjugate, not by control cells (Bawolak et al., 2007).
ACE is a bradykinin-inactivating ectoenzyme of prime importance in physiopathology and an important drug target (Leeb-Lundberg et al., 2005). Potentiation of endogenous kinins may be the basis of a fraction of the therapeutic and side effects of ACE inhibitors (Gainer et al., 1998; Bas et al., 2010; Schmidt et al., 2010). Thus, ACE represents a second molecular target of bradykinin of physiological importance.
In the present study, we verified whether novel fluorescent bradykinin analogs bind to the B2 receptors and to ACE, thus completing the full series of kinin receptor probes with an N-terminal extension that includes 5(6)-carboxyfluorescein (CF) (agonists and antagonists for both bradykinin B1 and B2 receptors; Bawolak et al., 2008, 2009b). The design of the novel fluorescent B2 receptor agonist peptides was based on the hypothesis of the tolerance of the bradykinin sequence for N-terminal extension, which has been tested. These ligands may support the study of receptor adaptation in an original manner. B2 receptor-mediated internalization of bradykinin should depend on receptor phosphorylation and combination with β-arrestins; the latter step has been verified in colocalization experiments (microscopy) and using a truncated version of myc-tagged rabbit B2 receptor (myc-B2R) devoid of the GRK substrate domain. Rab proteins are small GTPases that regulate vesicular trafficking (Grosshans et al., 2006). GTP-bound Rab5 is localized to both endocytic vesicles and early endosomes; later in trafficking, there is a Rab5 to Rab7 conversion at the surface of vesicles when the cargoes enter the late endosome compartment where degradation begins. A fluorescent bradykinin analog has been followed through these steps and was found to model peptide ligand degradation. Furthermore, imaging of ACE was attempted using a fluorescent bradykinin analog to verify whether the new analogs recognize additional molecular targets of bradykinin.
Materials and Methods
Drugs and Reagents.
Cell culture reagents were purchased from Invitrogen (Carlsbad, CA). The nonpeptide B2 receptor antagonist LF 16-0687 (1-[[2,4-dichloro-3-[[(2,4-dimethylquinolin-8-yl)oxy]methyl] phenyl]sulfonyl]-N-[3-[[4-(aminoiminomethyl)-phenyl]carbonylamino] propyl]-2(S)-pyrrolidinecarboxamide) (Pruneau et al., 1999) was a gift from Laboratoires Fournier (Daix, France). The fluorescent bradykinin analogs CF-ε-aminocaproyl-bradykinin (CF-εACA-BK) and CF-bradykinin were synthesized, purified, and characterized using general methods described previously (Gera et al., 1996). CF-εACA-d-Arg-[Hyp3,Igl5,d-Igl7,Oic8]-bradykinin (B-10380), a CF-conjugated peptide that is a B2 receptor antagonist, has been characterized previously (Bawolak et al., 2008). Bafilomycin A1 was purchased from LC Laboratories (Woburn, MA). All other drugs and reagents were purchased from Sigma-Aldrich (St. Louis, MO).
Generation of a Endocytosis-Resistant Truncated myc-B2R Construct.
In the present experiments, we mainly used a N-terminally myc-B2R that is functional because it binds [3H]bradykinin with an affinity identical to that of the wild-type receptor (Bawolak et al., 2007), binds the CF-conjugated B2 receptor antagonist B-10380 as well (Bawolak et al., 2008), and supports signaling and β-arrestin2 translocation to endosomes upon bradykinin stimulation (Bawolak et al., 2009a). The previously described construct myc-B2R (Bawolak et al., 2007) served as a template to generate an endocytosis-resistant version of the receptor. As is well established, the presence of a serine- and threonine-rich region in the C-terminal tail (Ser339, Thr342, Thr345, Ser346, Ser348) of the human bradykinin B2 receptor is necessary for the receptor ligand-induced phosphorylation (Blaukat et al., 2001), recruitment of β-arrestins, and subsequent endocytosis of the receptor (Pizard et al., 1999; Hamdan et al., 2007); this Ser/Thr-rich motif is identical in the rabbit B2 receptor sequence (Bachvarov et al., 1995). A truncation of the C terminus was obtained by mutating Glu340 in the myc-B2R sequence (numbering as in the rabbit wild-type receptor; Bachvarov et al., 1995) to the stop codon, thus eliminating the 28 C-terminal residues, a sequence that comprises the conserved Ser/Thr-rich domain that is the substrate of the GRKs (Leeb-Lundberg et al., 2005). The QuikChange site-directed mutagenesis method was used as directed (Stratagene, La Jolla, CA). Polymerase chain reaction was performed using the following sense and antisense primers, respectively: 5′–CAGGCGTAGAGCTCCATGGGC-3′ and 5′- GCCCATGGAGCTCTACGCCTG-3′, which contained a point mutation coding for a stop codon (boldface). The sequence of the constructions was verified (Plateforme de Séquençage et de Génotypage, Centre de Recherche du Centre Hospitalier de l'Université Laval, Quebec City, QC, Canada).
Cell Culture and Transfection.
A subclone of HEK 293 cells, called HEK 293a, originally obtained from Sigma-Aldrich was used in all experiments. This cell type was grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 1% l-glutamine, and 1% penicillin-streptomycin stock solutions (100×). HEK 293a cells were grown until they reached 70% confluence and were then transfected with one of the myc-B2R-coding vectors, the rabbit wild-type B2 or B1 receptor vector (Bachvarov et al., 1995; Larrivée et al., 2000) or with a vector coding for human ACE (coded in the peACE vector, gift from Prof. P. Corvol, INSERM U833, Paris, France; Wei et al., 1991), optionally cotransfected with β-arrestin1-cherryFP, or β-arrestin2-GFP, using the EX-Gen 500 transfection reagent (MBI Fermentas Inc., Flameborough, ON, Canada) as recommended by the manufacturer. β-Arrestin1-cherryFP and β-arrestin2-GFP fusion protein in pcDNA3 were kind gifts from Dr. J.-M. Beaulieu (Université Laval, Quebec, Canada) and Dr. M. Bouvier (Université de Montréal, Montréal, Canada; Bernier et al., 2004) respectively. C1-cherryFP, Rab5-GTP-locked-cherryFP and GFP-Rab7 were graciously given by Dr. M. J. Tremblay (Université Laval). Control cells were transfected with the empty vector (pCI-neo-myc) and, where indicated, with the vector coding for C1-cherryFP (an aid to localize transfected cells in microscopy).
HEK 293a cells transfected with myc-B2R or myc-B2Rtrunc were lysed and immunoblotted to detect the abundance of receptors using the anti-myc monoclonal antibody 4A6 (Millipore Corporation, Billerica, MA; dilution 1:1000) (technique described in Bawolak et al., 2007). Activation of ERK1/2 was determined by its phosphorylation state as described previously (Bawolak et al., 2007, 2009a; Morissette et al., 2007).
Binding Competition Assays.
The fusion protein myc-B2R has previously been evaluated as a valid construct for the binding of 3 nM [3H]bradykinin when transiently expressed by HEK 293a adherent cells (Bawolak et al., 2007). Application of the assay allowed the construction of a competition curve for fluorescent bradykinin analogs, with another for unlabeled bradykinin serving as a control. The rabbit wild-type B1 receptor was transiently expressed in the same cell type to detect any affinity of the fluorescent bradykinins for this receptor subtype by the displacement of the radioligand [3H]Lys-des-Arg9-bradykinin (1 nM; assay described by Bawolak et al., 2009b; the reference unlabeled competitor was Lys-des-Arg9-bradykinin). Likewise, [3H]enalaprilat binding to naturally expressed ACE in intact human umbilical vein endothelial cells has been previously characterized and shown to be displaced by ACE inhibitors (Morissette et al., 2008) and peptide substrates (Koumbadinga et al., 2010a). Competition curves for the displacement of 2 nM [3H]enalaprilat were established as described for bradykinin and analogs using primary cultures of endothelial cells (passages 4–5).
The anonymous use of human umbilical cords obtained after elective caesarean section deliveries was approved by a local institutional research ethics board. The preparation of the umbilical vein rings has been described previously (Bawolak et al., 2007). Tissues were randomly assigned to the agonists (bradykinin or a fluorescent analog), and cumulative concentration curves were constructed after a 3-h equilibration period.
Fluorescence Microscopy of CF-εACA-BK.
HEK 293a cells transfected with myc-B2R, myc-B2Rtrunc,and, optionally, with β-arrestin2-GFP were treated for 30 min with the agonist bradykinin, observed (1000×, epifluorescence with blue excitation and green emission filters), and photographed. Cells expressing myc-B2R or not were also treated with the fluorescent agonist CF-εACA-BK (concentration range from 0 to 10 μM) for 30 min at 37°C, and were then washed three times with Hank's balanced salt solution and observed (400×, phase contrast and epifluorescence). Microscopy was further pursued with a fixed concentration of CF-εACA-BK (5 μM) and additional incubation times (0–3 h, 37°C), as well as with optional pretreatment of myc-B2R-expressing cells with bafilomycin A1 or the B2 receptor antagonist LF 16-0687. Cells cotransfected with myc-B2R and β-arrestin1-cherryFP were also treated with CF-εACA-BK, using the same procedure described previously. Finally, the previously described B2 receptor fluorescent antagonist B-10380 (50 nM; Bawolak et al., 2008) was used to label myc-B2R at the cell surface. The cherryFP was routinely cotransfected with myc-B2R to validate transfection and localize cells of interest. Cell nuclei were counterstained with Hoechst 33258 (final concentration 25 μM in culture medium of intact cells, last 15 min of incubation at 37°C; Bawolak et al., 2009a). In another set of experiments, HEK 293a cells expressing human ACE were treated with CF-εACA-BK (5 μM, 30 min) at 22 or 37°C and observed. In addition, confocal sections of cells grown on a glass surface were also taken using a Volocity spinning disk confocal microscope (Improvision; PerkinElmer Life and Analytical Sciences, Waltham, MA) equipped with a Hamamatsu EMCCD digital camera (Hamamatsu Corporation, Bridgewater, NJ). Three-dimensional reconstitutions of cells from series of confocal images were made using FreeSFP software (http://www.svi.nl).
Photoshop software (version 6; Adobe Systems, Mountain View, CA) was used to quantify the endocytosis of fluorescent CF-εACA-BK into HEK 293a cells from the photographic record (1000×). The outline of each evaluated cell was manually drawn, and the median intensity of green pixels was recorded in the delimited surface area.
Numerical data were averaged for many cells, and the effects of cell treatment were statistically evaluated using nonparametric tests because of the non-normal distributions (Kruskall-Wallis test followed by Dunn's multiple comparison test for pairs of values).
The IC50 and EC50 values were obtained from averaged binding competition or concentration-effect curves, respectively, by interpolation from two points on each side of the half-maximal effect on a semilogarithmic scale. Numerical values are reported as means ± S.E.M.
Characterization of myc-B2Rtrunc.
myc-B2R represents a functional rabbit B2 receptor with the myc epitope fused at the N terminus and an affinity for [3H]bradykinin identical to that of wild-type rabbit B2 receptor (Bawolak et al., 2007). myc-B2Rtrunc, a truncated version without the 28 C-terminal residues that comprises the substrate domain of the GRKs (Leeb-Lundberg et al., 2005), was transiently expressed in HEK 293a cells and bound [3H]bradykinin with an affinity (KD 2.97 nM) similar to that of myc-B2R (2.42 nM) and had a slightly smaller cell surface density than that of the nontruncated receptor (Bmax 120.1 and 148.8 fmol/well, respectively; parameters derived from radioligand saturation curves; Fig. 1A). Untransfected cells specifically bound a low amount of [3H]bradykinin that may correspond to the small endogenous B2 receptor population in this cell type (Kramarenko et al., 2009). Anti-myc antibodies reveal the glycosylated myc-B2R construction as a wide band of ∼63 to 68 kDa (Bawolak et al., 2007), which was reproduced (Fig. 1B); the mass of myc-B2Rtrunc is predicted to lose ∼3 kDa, which may be compatible with the immunoblot pattern obtained in this experiment. Transiently expressed myc-B2Rtrunc supported acute bradykinin-induced ERK1/2 phosphorylation (Fig. 1C), just like myc-B2R (Bawolak et al., 2009a), and the competitive B2 receptor antagonist LF 16-0687 prevented this effect of bradykinin in both constructions.
When cotransfected with myc-B2R, β-arrestin2-GFP presents an even distribution in the cytosol of unstimulated cells, whereas the signal becomes condensed in endosomal structures, however, minimally at the cell surface when the cells are treated with the agonist bradykinin (100 nM, 30 min; Fig. 2). This translocation to intracellular particles of the previously diffuse signal was not observed in untransfected cells and in those expressing myc-B2Rtrunc, confirming the incompetence of the latter construction to recruit β-arrestin.
Pharmacology of Fluorescent Bradykinin Analogs.
Radioligand binding assay was conducted to evaluate the affinity of the novel fluorescent bradykinin analog CF-εACA-BK in HEK 293a cells transiently expressing myc-B2R. CF-εACA-BK displaced the specific binding of [3H]bradykinin with an IC50 of 800 nM, hence it was 400-fold less potent than bradykinin (IC50 of 2 nM) (Fig. 3A). CF-bradykinin was 3000-fold less potent than bradykinin in displacing [3H]bradykinin from myc-B2R. The human umbilical vein is a contractile bioassay for the endogenous B2 receptor and was exploited to further study the pharmacology of fluorescent bradykinin analogs. The fluorescent peptide was confirmed as a B2 receptor agonist. However, its EC50 value for the human receptor was >10 μM and only elicited 21.8% of the maximal contractile response at 6.3 μM concentration, which makes it approximately 1000-fold less potent than bradykinin (Fig. 3B). CF-bradykinin was inactive in this concentration range. Thus, the radioligand binding assay and the contractile assay both suggested an important loss of affinity of the fluorescent bradykinin analogs, probably caused by the N-terminal extension, but a partial protection by the presence of the spacer motif between bradykinin and the fluorophore. Further imaging of B2 receptors was based on the agonist CF-εACA-BK used at a pharmacologically active concentration range. CF-bradykinin and CF-εACA-BK in the concentration range 100 nM to 10 μM did not displace [3H]Lys-des-Arg9-bradykinin binding from rabbit recombinant B1 receptors (Fig. 3C).
Characterization and Endocytosis of the Fluorescent Bradykinin Analog CF-εACA-BK.
Figure 4 presents fluorescent labeling of untransfected HEK 293a cells or those expressing myc-B2R or its truncated form, when treated with increasing concentrations of CF-εACA-BK, ranging from 1 to 10 μM. Untransfected cells or untreated cells expressing one of the B2 receptor constructions did not exhibit any green autofluorescence under the employed settings (epifluorescence microscopy). The fluorescent peptide did not label untransfected cells or those expressing myc-B2Rtrunc. In myc-B2R-expressing cells, intracellular diffuse green fluorescent labeling was observed, as well as highly fluorescent intracellular small and dense structures. Intracellular condensed fluorescence of myc-B2R-expressing cells suggests the internalization of the fluorescent agonist into endosomes, in a receptor-dependent fashion, as would the natural agonist. The intensity of green fluorescent labeling was concentration-dependent and related to receptor occupancy in the radioligand binding competition assay. The higher concentrations (5 and 10 μM) lead to more intense labeling and better definition of the condensed fluorescent structures. Subsequent experiments were carried out using 5 μM as a concentration for imaging that supports a strong specific signal without noticeable nonspecific binding.
The three-dimensional rendition of the fluorescent B2 receptor ligands in whole cells that expressed the rabbit recombinant B2 receptor led to pronounced differences between the effect of the agonist (CF-εACA-BK; 5 μM) and the antagonist (B-10380; 50 nM) (Fig. 5, top). While the antagonist-bound receptors are uniformly distributed in the whole plasma membrane, without evidence for a polarized apical distribution of B2 receptors seen in some other epithelial cell types (Garcia-Perez and Smith, 1984), the endosomal sequestration of the fluorescent agonist (30 min treatment at 37°C followed by ample washings) is evidenced by the loss of the perceptible cell surface and signal fragmentation within the cytosol.
The endocytosis studies permitted by CF-εACA-BK were further pursued to assess the trafficking of the analog when bound to the receptor. Intact cells used for this series of experiments were routinely cotransfected with myc-B2R and cherryFP and counterstained with the nuclear dye Hoechst 33258 (Figs. 6 and 7). cherryFP red fluorescence was used as the predictor of the expression of the cotransfected receptors, and only red cells were included in the statistics related to the green fluorescence intensities in Fig. 6, right. This is validated by the positive and significant correlation between the intensity of the red fluorescence of cherryFP and that of CF-εACA-BK taken up in a series of cells cotransfected with myc-B2R (data not shown). Cells expressing myc-B2R were treated with CF-εACA-BK and observed after variable periods of incubation at 37°C (Fig. 6). When observation took place immediately after fluorescent agonist addition to the culture medium (conventionally indicated as 0 min), a very dim labeling appeared and endosomes were situated just below the cell membrane, whereas a 30-min incubation allowed an increased labeling of the cell cytosol and vivid fluorescence was observed in the endosomes. A 3-h incubation led to decreased condensed signal, whereas the cytosolic fluorescence actually increased. However, 1-h pretreatment with bafilomycin A1 (100 nM) reduced the diffusion of fluorescence from the endosomes to the cytosol. The statistics concerning green fluorescence intensity indicate a strong and significant reduction of median cell pixel in the presence of bafilomycin A1; this is because this measurement is very sensitive to the cytosolic background representing most pixels and less so to the peak fluorescence concentrated in dense vesicles that occupy a small surface area. In additional experiments presented in Fig. 6, cotreatment of myc-B2R-expressing cells with the B2R antagonist LF 16-0687 at equimolar concentration abated the uptake of CF-εACA-BK (5 μM, 30 min). Additional controls are not shown in Fig. 6, but were quantified for their green fluorescence and were not found to statistically differ from background fluorescence: cells that expressed B1 receptors with cherryFP or cherryFP alone, with or without CF-εACA treatment; other cells treated with bafilomycin A1 alone without CF-εACA (see Fig. 6 legend for numerical data).
Colocalization experiments are presented in Fig. 7. Cotransfection with β-arrestin1-cherryFP evidenced extensive translocation of the red fluorescence to CF-εACA-BK-containing endosomes (Fig. 7). myc-B2R-expressing cells cotransfected with Rab5-GTP-locked-cherryFP showed green fluorescence-containing, red fluorescence-labeled swollen endosomes. In control cells not treated with CF-εACA-BK, the red fluorescent protein distribution was similar to that illustrated, including the hollow giant endosomes that are typical of the activated Rab5 construction (Ceresa et al., 2001). To test whether the longer incubation period leads the fluorescent agonist into late endosome-lysosome compartments, the experiment was attempted in cells expressing cherryFP-Rab7 (Fig. 7, middle). As in previous experiments, there were few green endosomes 3 h after CF-εACA-BK application, but they seemed to be colocalized with Rab7-positive structures. At high magnification (1000×), CF-εACA-BK did not label cells that coexpressed myc-B2Rtrunc and cherryFP (Fig. 6). By contrast, the previously described B2 receptor fluorescent antagonist B-10380 (Bawolak et al., 2008) labeled the plasma membrane of HEK 293a cells expressing myc-B2Rtrunc (Fig. 7, bottom).
Imaging Angiotensin-Converting Enzyme.
CF-bradykinin retained ∼20% of bradykinin affinity at ACE, and CF-εACA-BK exhibited an affinity for ACE similar to that of bradykinin (Fig. 8; [3H]enalaprilat binding competition assay to natural ACE). These considerations support the possibility of using fluorescent bradykinins as ACE-imaging tools under conditions where the peptide associates with the catalytic site before the hydrolysis reaction takes place. In Fig. 9, it is shown that CF-εACA-BK at 5 μM stains intact HEK 293a cells that express recombinant human ACE (less efficiently at lower concentrations; data not shown). Cells that expressed only the tracking gene cherryFP did not take up CF-εACA-BK; further, enalaprilat cotreatment displaced the peptide-associated fluorescence from ACE (Fig. 9). Three-dimensional reconstitutions of the green fluorescence associated with intact HEK 293a cells expressing ACE were generated from serial confocal planes after treatments with CF-εACA-BK (5 μM) at 37°C (Fig. 5, bottom). CF-εACA-BK stained the surface of the cells in a uniform manner without evidence for significant endocytosis. The fluorescent bradykinin analog stained only weakly recombinant ACE at room temperature (data not shown), indicating that the presumably more rapid hydrolysis rate at 37°C is not detrimental to optimal imaging at this temperature. The fluorescent B2 receptor antagonist B-10380, with a Ki value in excess of 1 μM for the competition of [3H]enalaprilat (Fig. 8), at 50 nM did not label recombinant ACE expressed in HEK 293a cells (Fig. 9), whereas it labeled cell surface B2 receptors at this concentration (Fig. 5, top).
The structure-activity relationships of peptide ligands for both bradykinin receptor subtypes and their docking to these receptors have been well explored. Some of the findings can be summarized as follows: the C terminus of agonist or antagonist peptides plunges into the rosette formed by the transmembrane domains where it interacts with some of these domains, and the N terminus rather interacts with negatively charged residues of an extracellular loop (Leeb-Lundberg et al., 2005). Thus, coupling a cargo to the N terminus of the peptides may be feasible due to a likely tolerance for chemical extensions protruding out of the receptor into extracellular fluid; this is supported by some data for each receptor subtype. For instance, the natural sequence Met-Lys-bradykinin is only slightly less potent than bradykinin as an agonist of the recombinant human and mouse B2 receptors (Hess et al., 1994).
CF- εACA-BK completes the toolbox of CF-conjugated agonists and antagonists for both the B2 and the B1 receptors (Bawolak et al., 2008, 2009b). All such N-terminal extensions were detrimental for receptor affinity, but to a variable degree and with a significant effect of the animal species of origin for the receptor. The B2 receptor agonists described here are the ligands that suffered the most severe loss of affinity in the process of N-terminal extension with CF relative to the parent peptide. Thus, CF-bradykinin has little affinity for B2 receptors in the [3H]bradykinin binding competition assay, whereas the spacer alleviates in part this affinity problem in CF-εACA-BK that was approximately 400-fold less potent than bradykinin in the binding assay and ∼1000-fold less potent than bradykinin as a contractile agonist of the umbilical vein (Fig. 3). Unexpected steric hindrances between the exterior face of the receptor and the ligand N-terminal extension may have occurred, for instance, with the prominent N-linked sugar chains that make a large proportion of the B2 receptor weight and confer to it the heterogeneous appearance of one wide or multiple bands in immunoblots (Fig. 1B; Camponova et al., 2007). CF-εACA-BK did not label bradykinin B1 receptors. because the structural determinants of receptor subtype selectivity reside mainly in the ligand C terminus, which is identical in bradykinin and its fluorescent analog.
Despite the large loss of affinity relative to the parent peptide, the intrinsically fluorescent agonist CF-εACA-BK, used at 5 μM, was an important visualization tool of the endosomal lumen in microscopy when monitoring agonist-induced internalization of the B2 receptors in live cells and supported colocalization experiments with various proteins labeled with a red fluorophore (β-arrestin1, Rab5, Rab7). The fluorescent agonist (CF-εACA-BK) and antagonist (B-10380; Fig. 5) label cells in a strikingly different manner, the latter labeling the cell surface and the former associating only with intracellular vacuoles. These labeled structures are presumably the interior of endosomes, with significant colocalization with fluorescent β-arrestin1 and Rab5, but without significant labeling of the plasma membrane (Fig. 10, schematic representation). Intracellular particles where β-arrestin1-cherryFP and CF-εACA-BK colocalize undoubtedly correspond to agonist-receptor-arrestin trimolecular complexes (Fig. 7). The lack of plasma membrane labeling by the fluorescent agonist is possibly caused by the rapid dissociation of the low-affinity ligand upon cell washing and/or to the intracellular translocation of all surface receptors. The first possibility is more likely because the construction myc-B2Rtrunc, expressed at the cell surface, as shown by [3H]bradykinin binding (Fig. 1A) and labeling with the relatively high-affinity antagonist B-10380 (Fig. 7), did not retain any staining when sequentially treated with CF-εACA-BK and rinsed (Fig. 6).
The staining pattern of B2 receptors by CF-εACA-BK can be compared with the fluorescence associated with the receptors in HEK 293 cells stably expressing the B2R-GFP fusion protein that were resting or bradykinin-stimulated, respectively (Bachvarov et al., 2001; Bawolak et al., 2007, 2009a). Treatment with the nonfluorescent agonist bradykinin disrupts the smooth distribution of the receptor fluorescence at the plasma membrane, with labeling of endosomes by B2R-GFP, but variable residual signal from the cell surface in some cells may be caused by the high expression of B2R-GFP. In addition, various approaches indicate that the bradykinin-induced recycling of endocytosed B2 R-GFP at the cell surface is complete or nearly complete in 1 to 3 h (Bachvarov et al., 2001; Bawolak et al., 2007, 2009a), which also applies to naturally expressed receptors and is parallel to receptor dephosphorylation and dissociation from β-arrestin (Blaukat et al., 1996; Simaan et al., 2005). CF-εACA-BK imaging provides insight into another phenomenon, the intraendosomal degradation of the agonist: although endocytosed bradykinin is known to be extensively degraded (Munoz and Leeb-Lundberg, 1992), the inhibition of the cytosolic release of the fluorescence of CF-εACA-BK by bafilomycin A1 supports a role for the proton pump vacuolar-ATPase in the recycling of the B2 receptors. Thus, CF-εACA-BK has a certain instability, and its long-term persistence (e.g., in Rab7-positive late endosomes) concerns only a fraction of the endocytosed molecules (Fig. 10).
The agonist-stimulated B2 receptor is phosphorylated on conserved serine residues (Ser339, Ser346, Ser348) by GRKs, which allows the recruitment of the scaffold proteins β-arrestins, offering a docking structure for the endocytic machinery. Despite earlier conflicting evidence (Leeb-Lundberg et al., 2005), Hamdan et al. (2007) concluded that the endocytosis of the B2 receptor is mediated by clathrin-coated pits, based on the molecular interaction of β-arrestins and the AP-2 adaptor protein during the stimulation of this receptor. Consistent with previous work on the human B2 receptor, elimination of the phosphorylable domain in the C-terminal intracellular tail of myc-B2R rendered it incompetent for the recruitment of β-arrestin2 (Fig. 2), despite the conservation of its affinity and signaling (Fig. 1). myc-B2Rtrunc failed to support the endocytosis of CF-εACA, establishing the strict dependence of these transports on agonist-induced B2 receptor phosphorylation and interaction with β-arrestins. Thus, labeling of cells that express B2 receptors with CF-εACE-BK has an obligatory metabolic basis.
ACE is a bradykinin-inactivating ectoenzyme of prime importance in physiopathology and an important drug target. There is evidence that ACE is more densely expressed than B2 receptors in intact human umbilical vein endothelial cells, based on the comparison of [3H]enalaprilat and [3H]bradykinin binding sites in cell wells of identical size (Koumbadinga et al., 2010a, 2010b). Bradykinin has the highest affinity (Ki 525 nM) among physiological substrates (angiotensin I, acetyl-Ser-Asp-Lys-Pro) to displace the binding of the ACE inhibitor [3H]enalaprilat from naturally expressed ACE in human endothelial cells (Koumbadinga et al., 2010a). CF-bradykinin retained ∼20% of bradykinin affinity at ACE, and CF-εACA-BK exhibited an affinity for ACE similar to that of bradykinin (Fig. 9), showing again the utility of a spacer between the bradykinin sequence and the N-terminal CF moiety. It was verified that a fluorescent bradykinin can be used as an ACE imaging tool under conditions where the peptide associates with the catalytic site before the hydrolysis reaction takes place (Figs. 5 and 9): CF-εACA-BK stained mainly the surface of the cells expressing recombinant ACE in a uniform manner without evidence for important endocytosis, which is remarkable because the same peptide at the same concentration under the same incubation conditions is endocytosed in the same cell type that expresses recombinant B2 receptors (Fig. 5). Thus, the agonist CF-εACA-BK labels both ACE and the B2 receptors at similar micromolar concentration ranges; for the identification of naturally expressed B2 receptors in a context where ACE may be present, the fluorescent antagonist B-10380, with a Ki in excess of 1 μM for the competition of [3H]enalaprilat (Fig. 8), is theoretically superior, being selective for the receptors at 50 nM. Indeed, B-10380 at 50 nM did not label recombinant ACE expressed in HEK 293a cells (Fig. 9), whereas it labeled cell surface B2 receptors at this concentration (Fig. 5, top right). Alternatively, ACE or B2 receptor labeling by CF-εACA-BK can be differentiated in the presence of a B2 receptor antagonist or an ACE inhibitor to attenuate the confounding signal, as illustrated with recombinant receptors or peptidase in the present study.
The novel bradykinin analog CF-εACA-BK, used at micromolar concentrations, is a probe for the two major physiological targets of bradykinin, the B2 receptor and ACE. The labeling of the receptor-expressing cells is based on the β-arrestin-dependent endocytic capture of the peptide. CF-εACA-BK also models subsequent degradation of the endocytosed agonist.
Participated in research design: Gera, Bawolak, Lodge, and Marceau.
Conducted experiments: Gera, Bawolak, and Roy.
Contributed new reagents or analytic tools: Gera and Lodge.
Performed data analysis: Bawolak and Marceau.
Wrote or contributed to the writing of the manuscript: Gera, Bawolak, Lodge, and Marceau.
Other: Gera and Marceau acquired funding for the research.
We thank Drs. Marc Pouliot and Maria J. Fernandes (Centre Hospitalier Universitaire de Québec) for facilitating access to microscopic equipment and Gérémy A. Koumbadinga and Johanne Bouthillier for technical help.
This study was supported by the Canadian Institutes of Health Research [Grant MOP-93773]; and a Fonds de la Recherche en Santé du Québec Studentship Award (to M.T.B.).
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- angiotensin-converting enzyme
- green fluorescent protein
- B2 receptor-GFP
- 5(6) carboxyfluorescein
- [5(6) carboxyfluorescein]-ε-aminocaproyl-bradykinin
- cherry fluorescent protein
- extracellular signal-regulated kinases 1/2
- G protein-coupled receptor kinase
- myc-tagged rabbit B2 receptor
- truncated myc-B2R
- human embryonic kidney
- LF 16-0687
- Received November 9, 2010.
- Accepted January 3, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics