INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

BRADYKININ (BK)-related peptides (the kinins) stimulate two homolog G protein-coupled receptors (GPCRs), the widely distributed and constitutively expressed B₂ receptor (B₂R) and the highly regulated B₁ receptor (B₁R) (17). Several findings support B₁R importance in late inflammatory events: it is selectively stimulated by a class of abundant class of kinin metabolites formed from the native kinins BK and Lys-BK by carboxypeptidases. Lys-des-Arg⁹-BK (des-Arg¹⁰-kallidin) is the optimal agonist sequence of the human and rabbit B₁Rs, and des-Arg⁹-BK is also a selective agonist, but of lower affinity in these species. B₁R gene expression is inducible under the influence of cytokines, some growth factors, and possibly direct noxious stimuli to cells (13, 17, 26). The regulation of the two receptor subtypes differs at the protein level: the B₁R is not importantly internalized after agonist stimulation relative to the B₂R (10, 33). Accordingly, the B₁R fails to undergo ligand-induced phosphorylation, whereas the B₂R is phosphorylated in comparative experiments based on Sf9 cells (6). Agonist-induced cellular redistribution of kinin receptors has been studied with fusion proteins composed of the rabbit B₁- or B₂Rs fused with green fluorescent protein (GFP) or its yellow color variant (YFP) (3, 12, 25). BK induces B₂R-GFP endocytosis in a recycling endosome compartment, with essentially complete subsequent reexpression at the cell surface (3). Lys-des-Arg⁹-BK induces a condensation of cell surface B₁R-YFP into aggregates that remain associated with the plasma membrane and that were identified as caveolae-related rafts; this redistribution was slowly reversible on washing at 37°C and distinct from endocytosis (25).

Cell surface proteins, including receptors, are susceptible to a turnover process. For GPCRs, there is evidence that an agonist-independent (tonic) endocytosis process may be mechanistically distinct from agonist-induced endocytosis (8, 21, 22, 28). We hypothesized that the rapidly induced B₁R is subjected to an accelerated cellular degradation when its synthesis subsides. This mechanism would contribute to termination of B₁R signaling as the inflammatory condition is being resolved. Relatively rapid clearance of B₁Rs in vivo is suggested in the lipopolysaccharide injection model in the rabbit: a general state of cardiovascular sensitivity to a B₁R agonist is apparent 5 or 20 h after administration (24) but not any more 48 h after a bolus injection of the bacterial substance (F. Marceau, unpublished data). Also, the local hyperalgesia produced by zymosan injection into the rat paw and mediated in part by the B₁R occurs in a narrow time window (5). In addition, there is a loss of binding sites corresponding to B₁Rs after cycloheximide treatment or serum deprivation in cultured rabbit smooth muscle cells (SMCs) (26). By contrast, the related and constitutive BK B₂R may be long lived. In the present experiments, we have exploited several previously characterized experimental systems that express natural or recombinant rabbit B₁- or B₂Rs to formally address these questions.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Drugs. HOE 140 (D-Arg[Hyp³,Thi⁵,D-Tic⁷,Oic⁸]-BK; icatibant), a documented antagonist peptide of the rabbit B₂R (12, 18), was a gift from Laboratoires Fournier (Daix, France). BK and des-Arg⁹-BK were purchased from Bachem Bioscience (King of Prussia, PA), human recombinant IL-1 was from R&D Systems (Minneapolis, MN), and the remaining drugs were from Sigma (St. Louis, MO).

Cells. The derivation of separate HEK 293 cell lines stably expressing B₂R-GFP or B₁R-YFP and some of their properties have been described elsewhere (3, 12, 25). These cells were used in binding assays (24-well plates), confocal microscopy (35-mm petri dishes), and immunoblots of receptors (based on anti-GFP antibodies, see Immunoblot of GFP-related proteins). Each receptor fusion protein is a high-affinity, functional receptor. HEK 293 cells transiently transfected with GFP- or YFP-expressing vectors (pEGFP-N3 or pEYFP-N1, respectively; Clontech) were used in a comparative immunoblot experiment.

Effect of drugs on subcellular distribution of B₁R-YFP or B₂R-GFP. Drugs known to inhibit translation (anisomycin) or lysosomal acidification (bafilomycin A₁) or an agonist (Lys-des-Arg⁹-BK) were added to the serum-containing culture medium of the stable transfectant HEK 293 cells expressing either B₁R-YFP or B₂R-GFP, and the subcellular fluorescence distribution was observed without fixation or drug washout with a Bio-Rad 1024 confocal microscope as a function of treatment duration (×60 objective with oil immersion; emission 488 nm).

Immunoblot of GFP-related proteins. This assay was applied to detect GFP-sized COOH-terminal metabolites in HEK 293 cells expressing the protein conjugates composed of rabbit kinin receptors fused to GFP or YFP. Monoclonal antibodies to GFP (clone JL-8, used at dilution 1:1,000) were purchased from Clontech; these antibodies recognize YFP equally well and exhibit an exceptionally low background in extract of nontransfected HEK 293 cells relative to other commercial anti-GFP antibodies for immunoblot applications based on total cell extracts (3, 12). The technique was applied as generally described elsewhere (3).

Binding assays. The clearance of kinin receptors was primarily evaluated with binding assays for the various forms of naturally expressed or recombinant B_1- or B₂Rs in intact cells (12-well plates for SMCs, 24-well plates for HEK 293 cells). Culture medium was replaced by FBS-free medium to which metabolic inhibitors that suppress surface protein synthesis or maturation were added (anisomycin and brefeldin A, respectively). Brefeldin A blocks one or more GTPases that are necessary for vesicle transit in the Golgi apparatus (11). It is assumed that the blockade of receptor synthesis or maturation will reveal the first-order rate of their degradation process (20). The concentrations of anisomycin (10 µM) and brefeldin A (18 µM) used are maximally effective in various in vitro systems (discussed in Refs. 2 and 9); these drugs were used in a time frame compatible with cell viability (6 h) and observed B₁R half-life. Time controls in these experiments were established in cells incubated in FBS-free medium. In some experiments, IL-1 (5 ng/ml) was added 4 h before treatment with metabolic inhibitors in the complete culture medium. Cells were incubated with a radioligand concentration sufficient to reveal a large fraction of the maximum binding capacity (B_max; 1 nM for the B₁R ligand [³H]Lys-des-Arg⁹-BK, 3 nM for the B₂R ligand BK; 90 and 80 Ci/mmol, respectively; both from Perkin Elmer Life Sciences) (3, 4, 12, 14, 25). The assays were performed as described for the B₁R (25) or the B₂R (12). A cold competing peptide (1 µM Lys-des-Arg⁹-BK or BK, respectively) was added to some cell wells to determine the nonspecific binding. More extensive characterization of radioligand binding to rabbit mesenteric artery SMCs was performed (saturation, pharmacological profile) to ascertain the identity of naturally expressed receptors.

Contractility assay. A local ethics committee approved the procedures using rabbits (New Zealand White, 1.5-2 kg; Charles River, St. Constant, Canada). Two rabbit blood vessels are extensively characterized as bioassays specific for each kinin receptor subtype: the aorta for the B₁R and the jugular vein for the B₂R (14, 23). We stimulated these assays with documented agonist drug concentrations and within time frames sufficient for equilibration and compatible with tissue viability (2-7.5 h for the aorta; 1-6 h for the jugular vein). Rabbit aortic rings were suspended under a tension of 2 g in 5-ml tissue baths containing oxygenated (95% O₂-5% CO₂) and warmed (37°C) Krebs solution as described previously (13). Rabbit jugular vein strips were prepared and mounted under a baseline tension of 1 g as described previously, except that captopril was omitted from the formulation of the Krebs medium (12). Contractility studies in the aortic preparation were based on the construction of cumulative concentration-response curves for des-Arg⁹-BK (a B₁R agonist in this tissue). These studies aimed to functionally evidence B₁R decay in a system in which there is an intense B₁R upregulation (the "postisolation" induction paradigm) (17). The spontaneous induction of the B₁R was allowed to proceed until 4.5 h after isolation in all tissues, with the construction of one des-Arg⁹-BK concentration-effect curve at 3.5 h (control values that were pooled for statistical analysis); protein translation or maturation was then stopped with anisomycin or brefeldin A, respectively, in some tissues. These drugs are documented to prevent completely the induction in isolated rabbit aortas if applied continuously from the beginning of the in vitro incubation (2, 9). In each tissue, a second concentration-response curve for des-Arg⁹-BK was constructed at 6.5 h. Contractile responses are expressed as percentage of an internal control, the maximal effect of phenylephrine (PE) recorded in each tissue at 1.5 h; the concentration-effect curve of PE was also recorded at 7.5 h to test for metabolic inhibitor specificity and tissue viability. A similar protocol has been applied to the rabbit jugular vein, a B₂R-expressing tissue for which BK is a stable contractile agent (18), in comparative experiments. Histamine was the reference contractile agent in the vein.

Data analysis. Results are expressed as means ± SE, and Mann-Whitney statistics were calculated with the InStat 2.0 computer program (GraphPad Software, San Diego, CA). The parameters of the Scatchard plots (binding data treatment) and first-order decay were calculated with a computer program (29).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of drugs on HEK 293 cells expressing fluorescent conjugates of rabbit kinin receptors. Treatment of cells stably expressing B₁R-YFP with the protein synthesis inhibitor anisomycin (4 h) produced a certain loss of plasma membrane fluorescence relative to control cells (confocal microscopy; Fig. 1, top). Cells expressing B₂R-GFP were not apparently influenced by anisomycin treatment (Fig. 1). To obtain more objective microscopic proof of fluorescent receptor turnover, we treated cells with bafilomycin A₁, a drug that blocks lysosomal proteases by raising the pH of lysosomes (1, 31). B₁R-YFP accumulated into relatively large cytosolic vacuoles in important quantities in bafilomycin-treated cells, whereas the drug effect was much less intense in cells expressing B₂R-GFP (Fig. 1). This kind of cellular redistribution of B₁R-YFP is different from the agonist-induced translocation: for comparison, the effect of 30-min exposure to the agonist Lys-des-Arg⁹-BK (100 nM) is shown in Fig. 1 [longer agonist treatment produces similar results, as previously described (25)].

View larger version (69K): [in this window] [in a new window]   Fig. 1.   Top: subcellular localization of fluorescent kinin receptors in HEK 293 cells stably expressing either kinin B1 receptor (B1R)-yellow fluorescent protein (YFP) conjugate or kinin B2 receptor (B2R)-green fluorescent protein (GFP) conjugate. The cells were transferred in serum-free medium optionally containing 1 of the indicated drugs and incubated at 37°C for 4 h (30 min for the B1R agonist; magnification ×575). The selected confocal planes are halfway to the thickness of most cells. Results were verified during at least 2 separate days of experiments in multiple microscopic fields. Bottom: immunoblot of GFP-related proteins present in total HEK 293 cell extract with anti-GFP monoclonal antibodies. Cells stably expressing either B2R-GFP (lane 3) or B1R-YFP (lane 5) were extracted, as well as nontransfected HEK 293 cells (lane 1) or cells transiently expressing GFP or YFP (lanes 2 and 4, respectively). Representative result of at least 3 experiments.

Immunoblot of GFP-related proteins from cells expressing B₂R-GFP or B₁R-YFP. Figure 1, bottom, shows that the anti-GFP monoclonal antibody used reacts minimally with a total extract of nontransfected HEK 293 cells; furthermore, it recognizes both GFP and YFP equally well as ~27-kDa bands. Cells stably expressing B₂R-GFP (a 101- to 105-kDa fusion protein; Refs. 3, 12) also contained a faint band identifying a COOH-terminal fragment of a size similar to that of GFP (Fig. 1). HEK 293 cells stably expressing B₁R-YFP exhibited a much more intense concentration of YFP-sized proteins (Fig. 1). The fusion protein B₁R-YFP cannot be identified with certainty in these immunoblots.

Clearance of kinin receptors addressed with radioligands and metabolic inhibitors. In this series of experiments, cells expressing natural or recombinant rabbit kinin receptors were exposed to metabolic inhibitors for up to 6 h to isolate the agonist-independent receptor clearance in the absence of new receptor synthesis. Anisomycin (10 µM) has been preferred to cycloheximide to inhibit translation because of a higher potency and stability in biological milieus. A series of experiments were based on two separate HEK 293 cell lines that stably express fluorescent conjugates of rabbit B₁- and B₂Rs; B₂R-GFP is expressed at a three- to fourfold higher density than B₁R-YFP in resting cells as assessed by the binding of their respective radioligands, [³H]BK and [³H]Lys-des-Arg⁹-BK (receptor density values per well cited in Fig. 2, corresponding approximately to 404 and 1,547 fmol/mg protein). The binding site density corresponding to B₁R-YFP decreased sharply in HEK 293 cells during the 6-h anisomycin treatment [extrapolated half-life (t_1/2) = 6.5 h]; that of the B₂R fusion protein was essentially constant (values not different from baseline density). The time control curves in Fig. 2 describe the binding site density in cells incubated in FBS-free medium from time 0, but without the metabolic inhibitors. Brefeldin A treatment (6 h) was at least as effective as anisomycin to decrease the density of surface binding sites in cells expressing B₁R-YFP (Fig. 2A; t_1/2 = 5.2 h) but decreased the density of B₂R-GFP by only ~15% (Fig. 2B). In cells expressing B₁R-YFP and exposed to metabolic inhibitors, the binding site decay rate is slower than expected from the first-order decay kinetics after 4 h. This may be due to the toxicity of the applied drugs over this relatively long time period.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Clearance of fluorescent forms of B1R and B2R stably expressed in HEK 293 cells as estimated by radioligand binding in cells treated with metabolic inhibitors. A: cells expressing B1R-YFP were treated with anisomycin (10 µM) or brefeldin A (18 µM) for the indicated time; control cells were transferred in FBS-free medium. The radioligand was [3H]Lys-des-Arg9-bradykinin (BK) (1 nM). B: cells expressing B2R-GFP were also submitted to the treatments with the 2 metabolic inhibitors. The radioligand was [3H]BK (3 nM). Results are means ± SE of n separate experiments. Average initial binding densities are indicated.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Clearance of naturally expressed B1Rs in rabbit aortic smooth muscle cells (SMCs) as estimated by radioligand binding in cells treated with metabolic inhibitors. A: cells treated in the regular culture medium were used. B: IL-1 (5 ng/ml) was added to the culture medium 4 h before cells were rinsed with serum-free medium and further incubated (2 h), optionally with metabolic inhibitor. The radioligand was [3H]Lys-des-Arg9-BK (1 nM). Results are means ± SE of n separate experiments. Average initial binding densities are indicated.

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Characterization of naturally expressed B1R in rabbit mesenteric artery SMCs. A: saturation of [3H]Lys-des-Arg9-BK binding. B: Scatchard plot derived from A. C: competition of the specific binding of [3H]Lys-des-Arg9-BK (1 nM) by a panel of cold peptides (each at 1 µM). D: mesenteric artery SMCs were treated with anisomycin or brefeldin A to evaluate the clearance of the B1R. The radioligand was [3H]Lys-des-Arg9-BK (1 nM). Results are means ± SE of n separate experiments. Average initial binding density is indicated.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Characterization of naturally expressed B2R in rabbit mesenteric artery SMCs. A: saturation of [3H]BK binding. B: Scatchard plot derived from A. C: competition of the specific binding of [3H]BK (3 nM) by a panel of cold peptides (each at 1 µM). D: mesenteric artery SMCs were treated with anisomycin or brefeldin A to evaluate the clearance of the B2R. The radioligand was [3H]BK (3 nM). Results are means ± SE of n separate experiments. Average initial binding density is indicated.

Kinin receptor function as modified by metabolic inhibitors in contractility assays based on rabbit blood vessels. The isolated rabbit aorta exhibits a progressively increasing maximal contractile response to the B₁R agonist des-Arg⁹-BK, as evidenced by comparing responses recorded at 3.5 h (pooled controls) and 6.5 h (control) after isolation (Fig. 6A). In some tissues, the metabolic inhibitors anisomycin or brefeldin A were applied from 4.5 h on. The maximal effect of the agonist recorded at 6.5 h was inferior to that recorded at 3.5 h in tissues treated with brefeldin A (P = 0.03, Mann-Whitney test) but not in those treated with anisomycin. The response to the -adrenoceptor agonist phenylephrine somewhat increased from the beginning (1.5 h) to the end (7.5 h) of the experiment, but the late response was not affected by the metabolic inhibitors (Fig. 6B).

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Change in the concentration-effect curve of the B1R agonist des-Arg9-BK as a function of incubation time and drug treatment in rabbit aortic rings. A: control concentration-effect curve was constructed for the peptide in each tissue at 3.5 h from the beginning of the in vitro incubation. A drug was then continuously applied from 4.5 h on, and the concentration-effect curve was constructed again for des-Arg9-BK at 6.5 h. Contractile responses are expressed as % of an internal control, the maximal effect of phenylephrine (PE) recorded in each tissue at 1.5 h. B: concentration-effect curve of PE was also recorded at 7.5 h to test for drug effect specificity and tissue viability. Values are means ± SE of n determinations.

View larger version (29K): [in this window] [in a new window]   Fig. 7.   Effect of brefeldin A or anisomycin on BK-induced contraction of the rabbit jugular vein. A: control concentration-effect curve was constructed for the peptide in each tissue 2 h from the beginning of the in vitro incubation. Brefeldin A (18 µM) or anisomycin (20 µM) was then continuously applied from 2.5 h on, and the concentration-effect curve was constructed again for BK at 5 h. Contractile responses are expressed as % of an internal control, the maximal effect of histamine recorded in each tissue at 1 h. B: concentration-effect curve of histamine was also recorded at 6 h to test for drug effect specificity and tissue viability. Values are means ± SE of n determinations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Various well-characterized experimental systems expressing rabbit kinin receptors have been exploited to test a novel hypothesis: the B₁R population, which is rapidly induced after tissue injury (17), could also be rapidly destroyed in an agonist-independent manner when inflammation subsides. A comparison with the structurally related BK B₂R has been made, because the latter subtype is constitutively expressed and could be more stable. While studying the properties of the two types of rabbit kinin receptor fused to fluorescent proteins (3, 12, 25), we repeatedly observed that immunoblots based on different anti-GFP antibodies revealed B₂R-GFP but little free GFP in a stable transfectant cell line expressing the B₂R construction. Conversely, immunoreactive B₁R-YFP was hardly detectable but YFP-sized proteins were abundant in two independent lines of HEK 293 cells stably expressing B₁R-YFP. The absolute abundance of B₁R-YFP (radioligand binding, membrane fluorescence intensity) was substantially less than that of B₂R-GFP in stably transfected HEK 293 cells, although the constructions are expressed under the control of the same promoter. The membrane fluorescence is interpreted as the fusion protein because GFP-related proteins dissolve in total cellular water. A possible model integrating these observations is that the B₁R construction is short lived relative to the B₂R-based fusion protein. Thus at equilibrium B₁R-YFP abundance at the cell surface would be inferior to that of B₂R-GFP because these proteins are cleared at a different rate, even if formed at a similar rate. GFP, a stable and compact globulin, has a long half-life in mammalian cells (7), and its accumulation is a sensitive indicator of partial receptor-GFP conjugate degradation (as previously observed with endothelin ET_B receptor-GFP, which is bound to lysosomal degradation in a largely agonist-independent manner) (1). We have validated that the presence of GFP-related proteins in HEK 293 cells is an indicator of B₂R-GFP degradation, because the treatment of intact cells with trypsin leads to GFP accumulation in 10 min (3). Thus immunoblot results constitute circumstantial evidence that B₁R-YFP is submitted to tonic, agonist-independent degradation.

Brefeldin A prevents the translocation of newly formed proteins from the endoplasmic reticulum to the Golgi apparatus (11) and is highly effective to prevent the de novo formation of B₁R in the rabbit aorta (2). Treatments with the metabolic inhibitors anisomycin and brefeldin A allowed us to observe the agonist-independent decay of either B₁R-YFP or wild-type B₁R as assessed with radioligand binding (Figs. 2-4). In comparative experiments, B₂R-GFPs or wild-type B₂Rs were more stable (Figs. 2 and 5). Comparison with naturally expressed receptors is important because conjugation with GFP-related proteins may theoretically induce conformational changes that could favor receptor degradation. The recorded or extrapolated t_1/2 values from both HEK 293 cells and SMCs show that GFP/YFP conjugation does not change the large stability difference between the kinin receptor subtypes. Moreover, the clearance difference between the wild-type B₁R and B₂R is maintained when receptors are expressed by the same cell type, the mesenteric artery SMCs. The B₁R may be further routed to lysosomal degradation, as suggested by the accumulation of intracellular fluorescence in vesicles in B₁R-YFP-expressing cells treated with bafilomycin A₁ (interrupted lysosomal function; Fig. 1).

We have further attempted to validate our findings in contractility assays based on freshly isolated rabbit blood vessels. A limitation of this comparison is the different identity of the tissues, with possible consequences such as differences in membrane composition and stability. Furthermore, a potential problem with the B₁R bioassay, the aorta, is the effect of time on the system. The progressive increase of the B₁R agonist maximal effect in the rabbit aorta has been found to be sizeable during the period 3.5-6.5 h (Fig. 6), because the system is subject to postisolation receptor synthesis (2, 26). Evidencing receptor degradation in this special experimental context, where a cohort of receptors is presumably maturing in the "secretory pathway" (reticulum-Golgi), is based on metabolic inhibitors. Anisomycin is just able to maintain the B₁R-mediated response at the preapplication level (Fig. 6), perhaps because the drug does not block receptor maturation and receptors emerging from the secretory pathway compensate for the lost ones. Brefeldin A is more effective, producing a significant loss of B₁R-mediated function over the 2-h observation period, because it may block B₁R maturation (Fig. 6). The drug acts with specificity, having no inhibitory action on phenylephrine-induced aortic contraction relative to similarly aged control preparations. Similarly, brefeldin A was more effective to decrease the density of B₁Rs than anisomycin in cultured SMCs only when the cells were pretreated with IL-1 (Fig. 3B), probably for the same reason (secretory pathway not immediately inhibited by anisomycin). Treatments with either brefeldin A or anisomycin do not influence the functional response to BK in the rabbit jugular vein (Fig. 7), supporting that the B₂R population is not rapidly eliminated from the preparation.

Structural determinants of B₁R antagonist-independent degradation remain to be identified. In some cases, separate motifs present in the COOH-terminal tail of some GPCRs determine agonist-induced endocytosis and tonic internalization (in receptors for thromboxane or thrombin) (21, 28). Whereas the first type of motif supports receptor phosphorylation, the second type does not in the two cited examples. The specificity of the present system may be that the spontaneously internalized B₁R undergoes minimal agonist-induced internalization and no phosphorylation (see introductory paragraphs). There is evidence that the phosphorylation motif in the delta-opioid receptor tail repels arrestin-3 specifically when not phosphorylated (32), suggesting that the presence of such a motif in the B₂R, and its absence in the B₁R, may be sufficient to account for kinin receptor turnover differences. The cytokine- and tissue injury-induced B₁R is subjected to an accelerated degradation, perhaps to adjust the receptor population to the inflammatory stimulus, relative to the B₂R, which may be more economically managed (constitutively expressed, completely recycled, long-lived at the protein level).