INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

G PROTEIN-COUPLED RECEPTORS (GPCRs) are key components of the intercellular signaling networks tuning the homeostasis of multicellular organisms. The cellular action of GPCRs is tightly controlled on several levels of the signaling cascade. An initial agonist application often renders a cell insensitive to a second stimulus, a phenomenon termed desensitization or adaptation. Molecular studies have revealed multiple mechanisms contributing to the negative regulation of GPCR-derived signals (8, 22). First, synthesis and/or release of agonists are often restricted with respect to time and location. Once generated and competent to activate receptors, an agonist is often quickly removed from the extracellular space by endocytosis and/or degradation; these mechanisms ensure rapid signal attenuation at the prereceptor level (22). Second, receptor desensitization occurring during short-term (seconds to minutes) exposure of cell to agonists is mediated by uncoupling of activated receptors from G proteins, terminating signaling on the postreceptor level. Third, phosphorylation by second messenger-activated kinases such as protein kinases A and C, casein kinases, or specific G protein-coupled receptor kinases (GRKs) and subsequent binding of arrestins mediate rapid silencing of GPCRs at the receptor level (8, 22, 28, 39). On the cellular level, receptor sequestration depletes the plasma membrane of high-affinity receptors following ligand stimulation, thereby contributing to both desensitization and resensitization of signaling cascades via receptor recycling (42).

Whereas GPCR phosphorylation and sequestration are well-studied events, much less is known about the mechanisms of receptor downregulation reflecting the loss of receptors from the plasma membrane due to long-term exposure of cells to agonists (hours to days). Reduction of the receptor count per cell is either a result of enhanced degradation and/or of reduced protein synthesis (22). The relative lack of knowledge about the mechanisms underlying receptor downregulation is largely due to experimental limitations. For instance, the overexpression of recombinant epitope-tagged GPCRs often used to study receptor phosphorylation and endocytosis often masks physiologically relevant downregulation mechanisms. Strong viral promoters of generic expression vectors do not allow for transcriptional regulation of receptor synthesis, and the resultant excessively high concentrations of receptor mRNA and protein may well exceed the capacity of the cellular degradation machinery. Therefore, studies on the downregulation of GPCRs need to be done in a native cellular environment; however, the amount of endogenous GPCRs present in primary cells is often too low to allow the detailed analysis of dynamics of receptor expression, location, turnover, and activity.

A prototypical GPCR expressed in reasonable copy numbers by native cells and desensitizing on prolonged ligand stimulation is the human bradykinin B₂ receptor (B₂R). Bradykinin is a proinflammatory hormone mediating hypotension, edema formation, pain sensation, smooth muscle contraction, and cell growth (4). Its cognate receptor, B₂R, is well studied with respect to short-term desensitization resulting from ligand-induced receptor phosphorylation on serine and threonine residues in the COOH-terminal receptor domain (5, 7, 16). This posttranslational modification has also been shown to trigger endocytosis of B₂R (31). Whether the receptor is "channeled" into a clathrin-dependent pathway or whether it translocates into caveolin-rich compartments appears to be dependent on the cell type (3, 23, 31). Even though there is evidence that B₂R downregulation may occur, the mechanisms underlying such a phenomenon have not been studied in great detail (5, 35).

Here we have used specific mono- and polyclonal antibodies against the B₂R to analyze receptor downregulation in native human foreskin fibroblasts endogenously expressing the B₂R. Continuous stimulation with bradykinin decreased the number of specific binding sites and downregulated B₂R protein. This reduction in B₂R copy number was not due to transcriptional regulation because B₂R mRNA levels remained largely unchanged. Rather, de novo protein synthesis and receptor stability were reduced, suggesting translational and posttranslational control of cellular B₂R levels during prolonged agonist stimulation.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials. Reagents were obtained from the following manufacturers: Pro-mix ³⁵S in vivo cell labeling mix (>1,000 Ci/mmol), Hybond membranes and Megaprime DNA labeling kit from Amersham; sulfur-free DMEM medium from Applichem; bradykinin from Bachem; affinity-purified and preabsorbed anti-rabbit/mouse-IgG-antibody fragment [F(ab')₂] coupled to horseradish peroxidase from Biotrend; Nonidet P40 (NP-40) and Pansorbin from Calbiochem; [-³²P]dATP (>4,000 Ci/mmol) from ICN; protein ladder marker (10-200 kDa) from Life Technologies; [2,3-prolyl-3,4-³H]bradykinin (spec act 98 Ci/mmol) from NEN DuPont; aprotinin, Pefabloc SC, Rotiquant, and scintillation cocktail Rotiszint Eco Plus from Roth; GF 52 glassfiber filters from Schleicher & Schuell; leupeptin, pepstatin A and 1,10-phenanthroline from SERVA; and polyethylenimine [50% (wt/wt) aqueous solution] from Sigma. All other chemicals of analytic grade were from Applichem, Merck, Sigma, or Roth.

Peptide synthesis and production of monoclonal antibodies. A 36-amino acid peptide (dubbed CRS36) derived from the COOH terminally located intracellular domain (ID4) of the B₂R was synthesized by the solid phase method using fluorenylmethyloxycarbonyl chemistry. Peptide CRS36 was used in a nonconjugated form for immunization of six mice. Selection of positive clones, production of ascites, and characterization of the antibodies were done according to standard procedures (25). MBR1, a monoclonal antibody to the human bradykinin B₂ receptor, was isotyped IgG_2b. The production of antiserum -CRS36 (AS346) to peptide CRS36 of human B₂R has been detailed elsewhere (5, 6).

Radioligand binding studies. The binding activity of B₂ receptors was assayed on adherent cells or membrane preparations using [³H]bradykinin as detailed in former studies (1). Dissociation constants (K_D) for B₂R were calculated by Scatchard analysis with radioligand concentrations ranging from 10 pM to 20 nM using membranes prepared from HF5 cells (25 µg protein/measurement). When cells were preincubated with bradykinin, bound peptide was removed with 200 mM acetic acid, 500 mM NaCl, and 0.1% BSA, pH 2.8, for 5 min at 4°C ("acid stripping"). Cells washed three times with DMEM were used for [³H]bradykinin binding experiments.

³⁵S labeling and immunoprecipitation. HF-15 human foreskin fibroblasts cultured in DMEM with 10% FCS and used at passage 9-12 were grown on six-well dishes, washed twice with sulfur-free DMEM, and incubated for 30 min at 37°C in the same medium. For ³⁵S-labeled pulse-chase experiments, cells were labeled for 60 min with 0.33 mCi/ml [³⁵S]methionine-cysteine (Promix) in sulfur-free DMEM. Thereafter, radioactive medium was removed, and the cells were incubated with DMEM-10% FCS in the presence or absence of 1 µM bradykinin. To follow B₂R protein synthesis, we applied 0.1 mCi/ml ³⁵S-labeled Promix in sulfur-free DMEM in the absence or presence of 1 µM bradykinin. At the indicated time points, cells were washed three times with 50 mM Tris, 150 mM NaCl, pH 7.5 (TBS), lysed with 1% (mass/vol) NP-40, 0.5% (mass/vol) deoxycholate, and 0.1% (mass/vol) SDS in TBS, including protease inhibitors [1 mM Pefabloc and 10 µg/ml each of leupeptin, aprotinin, pepstatin A, and 1,10-phenanthroline (RIPA buffer)], and receptors were immunoprecipitated with 2.5 µl of antiserum AS346 (5). For control, preimmune serum or serum preabsorbed with peptide CRS36 was used. Proteins were separated by 10% SDS-PAGE and visualized by fluorography using 15% (mass/vol) sodium salicylate as the fluorophor.

Western blotting. HEK293T cells stably transfected with hemagglutinin (HA)-tagged B₂R or HF-15 cells were washed twice with TBS and lysed in RIPA buffer for 45 min at 4°C. B₂R was immunoprecipitated with 2.5 µl of a antiserum AS346 (anti-B₂R) or 4.5 µg of monoclonal antibody, and precipitates were separated by 10% SDS-PAGE. Proteins were transferred onto polyvinylidine difluoride or nitrocellulose membranes by semidry blotting for 45 min at 1.5 mA/cm² using 39 mM Tris, 48 mM glycine, and 1.3 mM SDS. Membranes were blocked either with 5% (mass/vol) fat-free milk powder in TBS with 0.1% (mass/vol) NP-40 or with 0.25% (wt/vol) gelatin in TBS, pH 7.5, including 5 mM EDTA and 0.05% (vol/vol) Triton-X 100 (blocking buffer). Membranes were then incubated with primary antibodies (0.1-2 µg/ml) diluted in corresponding blocking buffer for 2 h at room temperature. For protein visualization, an anti-mouse antibody coupled to horseradish peroxidase was diluted 1:1,000-1:5,000 in blocking buffer, followed by chemiluminescence detection.

RNA isolation and Northern blotting. Total RNA from HF-15 cells was isolated (13). Thirty micrograms of RNA were separated by denaturating agarose gel electrophoresis and transferred onto nylon membranes using 20× SSC buffer (0.3 M sodium acetate, 3 M NaCl, pH 7.0) for 12 h (34). RNA was fixed by 2 h incubation at 80°C; staining of the 28S and 18S ribosomal RNA with 0.2% methylene blue in 0.2 M sodium acetate was used to monitor equal sample loading. Hybridization of B₂R mRNA was performed in 0.5 M sodium phosphate, 7% SDS, 1 mM EDTA, pH 7.0, for 12 h at 68°C with a 1.2-kb B₂R cDNA generated from the full-length clone by random priming with [-³²P]dATP using the Megaprime kit of Amersham (18). Labeled RNA/DNA hybrids were analyzed by a PhosphorImager, and corresponding intensities were normalized for -actin controls.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Downregulation of bradykinin binding sites on HF-15 cells during prolonged agonist treatment. B₂R endogenously expressed by human HF-15 foreskin fibroblasts has been shown to inactivate by rapid desensitization paralleled by phosphorylation of a Ser-Thr cluster in the COOH-terminal receptor domain (5, 7). In a similar setting, B₂R-bradykinin complexes are rapidly internalized rendering HF-15 cells resistant to further agonist stimulation for a limited period of time (35). Therefore, we wondered whether the number of cell surface-exposed B₂R is progressively downregulated during long-term treatment of HF-15 cells with bradykinin. To this end we stimulated cells for up to 24 h with 1 µM bradykinin and followed the number of surface-exposed B₂R by a radioligand binding assay. To avoid interference with the radioligand, unlabeled bradykinin used to stimulate the cells was quantitatively removed by "acid stripping" before 5 nM [³H]bradykinin was added for 90 min at 4°C. The amount of cell-bound radioactivity was quantified in a -counter and compared with that of unstimulated cells, which had otherwise been treated identically. Our experiments revealed that during the first 6 h of agonist treatment the number of surface B₂R decreased only moderately to 75-80% of the control (100% at t = 0) and dropped to about 50% of control after 24 h (Fig. 1A).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Downregulation of [3H]bradykinin (BK) binding sites on HF-15 cells during prolonged agonist exposure. A: confluent HF-15 cells grown on 24-well plates were incubated in serum-free DMEM with 1 µM BK for the indicated time periods. To compensate for BK degradation, fresh BK was added every 6 h. Thereafter, BK was removed by "acid stripping" for 5 min at 4°C, and cell surface binding was measured using 5 nM of [3H]BK for 90 min at 4°C in the absence ("total binding") or presence of 10 µM unlabeled BK ("unspecific binding"). Specific binding calculated as the difference between total and unspecific binding was set 100% for nonstimulated cells (control). Mean values ± SD from four independent experiments are shown. B: for control cells and for cells pretreated 24 h with BK, the number of B2 receptor (B2R) in cellular membranes was determined by Scatchard analysis with radioligand concentrations ranging from 10 pM to 20 nM and constant competitor amounts of 10 µM BK to determine the unspecific binding. Protein concentrations in membrane preparations were estimated by the Bradford method. A representative experiment out of 3 is shown.

Protein levels of B₂ receptor during prolonged agonist treatment. Given the significant loss of bradykinin binding sites on HF-15 cells permanently exposed to agonist, we wondered whether total levels of B₂R protein would also be decreased under these conditions. Initially, we tried to detect B₂R protein solubilized from HF-15 cells by Western blotting with a polyclonal anti-B₂R antibody, which had performed well in immunoprecipitation experiments (5, 7, 43). Not unexpectedly, this approach failed to detect B₂R in HF-15 extracts, most likely because the absolute amount of B₂R applied per lane (up to 50 fmol receptor in 50 µg of total protein loaded) is well below the detection limit of our antiserum, previously judged to be 120 fmol B₂R protein (6). Alternatively, we employed a commercially available monoclonal antipeptide antibody to human B₂R (from Signal Transduction Laboratories; herein refered to as ST antibody). Using Sf9 or HEK293 cells highly overexpressing human B₂R, we were unable to find any specific bands in the Western blots with the ST antibody (data not shown). To test the capacity of the antibody for immunoprecipitation of B₂R, we applied ST to lysates from HEK293 cells massively overexpressing the HA-tagged human B₂R (~9 pmol of receptor per milligram of protein) but failed to produce specific bands (Fig. 2A). Under the same conditions, polyclonal antibody AS346 ("anti-B₂R") gave a major band of ~70 kDa, indicating that it efficiently precipitated human B₂R (Fig. 2A). This finding was confirmed by the application of a specific anti-HA antibody, which also precipitated the HA fusion protein with B₂R.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Reduction of B2R protein levels during continuous BK treatment. A and B: HEK293 cells expressing 9 pmol/mg of hemagglutiin (HA)-tagged B2R were lysed, and HA-B2R was immunoprecipitated from 1.5 mg protein with 2.5 µl AS346 ("anti-B2R"), 4.5 µg MBR1, 4.5 µg ST antibody, or 4.5 µg anti-HA antibody. c, Control. WB, Western blotting; IP, immunoprecipatation. Precipitates were separated by reducing 10% SDS-PAGE, transferred onto nitrocellulose membranes, and probed with 0.1 µg/ml anti-HA (A) or 0.5 µg/ml MBR1 (B) using the ECL method. C: HF-15 cells were lysed in RIPA buffer and about 4 mg protein was used for IP with preimmune serum, polyclonal antibody AS346 ("anti-B2R"), or AS346 preabsorbed with peptide CRS36. Precipitates were separated by 10% SDS-PAGE under nonreducing conditions to avoid interference of the precipitating antibody with the secondary antibody used for Western blotting. Proteins were transferred on polyvinylidine difluoride membranes and probed with 2 µg/ml of MBR1. A typical Western blot developed with the ECL method is shown. D: lysates from HF-15 cells pretreated for indicated time periods with 1 µM BK were subjected to same the IP and WB. A representative experiment out of 6 is shown.

Development of monoclonal antibodies to human B₂ receptor. To overcome the limitations with the available antibodies, we raised monoclonal antibodies to human B₂R. From multiple peptides covering various extra- and intracellular domains of human B₂R, peptide CRS36 corresponding to a continuous sequence portion of 36 residues at the tail domain (ID4) of B₂R consistently yielded high-titered sera against the receptor (data not shown). B cells of mice immunized with unconjugated CRS36 were immortalized, and monoclonal antibody MBR1 (IgG_2b) cross-reactive with human B₂R was generated. MBR1 was isolated from mouse ascites by affinity chromatography on protein G-Sepharose, and the purified antibody was characterized by Western blotting using lysates from Sf9 cells overexpressing human B₂R (6). Preliminary experiments indicated that MBR1 detected B₂R protein in Western blots more efficiently than any of our polyclonal antibodies. MBR1 also precipitated HA-tagged B₂R expressed in HEK293 cells, though less efficiently than polyclonal antibody AS346 ("anti-B₂R") or monoclonal anti-HA antibody (Fig. 2A). Using AS346 for immunoprecipitation and MBR1 (at 0.5 µg/ml) for Western blotting, we succeeded in detecting B₂R overexpressed in HEK293 cells (Fig. 2B) and also endogenous B₂R protein from HF-15 cells (Fig. 2C). HF-15 cell lysates gave the characteristic pattern for human B₂R, i.e., a major 68-kDa and a minor 52-kDa band (5, 7), whereas preimmune serum or preabsorbed immune serum failed to detect B₂R (Fig. 2C). Using HA-tagged B₂R, we observed a similar pattern for anti-HA, confirming the specificity of our detection system (Fig. 2B). Thus reporter antibody MBR1 in combination with precipitating antibody AS346 allowed us to detect even minute amounts of B₂R endogenously expressed by native cells, i.e., human foreskin fibroblasts. Under otherwise identical conditions, the ST antibody failed to produce specific bands with endogenous B₂R, even at significantly higher antibody concentrations (2 µg/ml).

Analysis of B₂ receptor mRNA levels by Northern blotting. One possible reason for the loss of binding sites and B₂R protein during long-term agonist exposure could be the attenuation of B₂R mRNA synthesis by transcriptional control mechanisms. We followed B₂R mRNA in HF-15 cells that were challenged with bradykinin for up to 24 h (Fig. 3A). Because all mammalian cells we have tested express low amounts of B₂R mRNA, we employed Sf9 insect cells for control. The unlabeled 1.2-kb B₂R cDNA fragment used to generate the radioactive probe served as a positive control for hybridizations (not shown). Because long-term treatment with bradykinin should not affect the concentrations of actin mRNA, we used a -actin probe as control and normalized B₂R mRNA levels to this reference. Over a 24-h period of bradykinin treatment, B₂R mRNA levels in HF-15 cells did not significantly change (Fig. 3A). Likewise, the ratio of B₂R mRNA versus -actin mRNA remained constant throughout the experiment (Fig. 3B). Therefore, the observed reduction of B₂R protein levels in HF-15 cells appears not to be due to reduced production or increased degradation of the corresponding mRNA.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Northern blots of B2R mRNA in HF-15 cells. A: total RNA (30 µg) isolated from HF-15 cells that had been treated for the indicated time periods with 1 µM bradykinin or from control Sf9 cells (c) was separated by denaturing agarose gel electrophoresis and transferred onto nylon membranes. Hybridizations were performed with a 1.2-kb B2R cDNA fragment labeled with [-32P]dATP by random priming. Unlabeled B2R cDNA served as a positive control (not shown). A typical hybridization experiment out of 6 is shown. B: labeled RNA-DNA hybrids were analyzed by a PhosphorImager, and the values were normalized for the levels of -actin; the ratio obtained in the absence of bradykinin was set 1 (outer left column). Mean values ± SD from 6 independent experiments are shown.

³⁵S-labeled pulse-chase experiments to follow B₂ receptor degradation. We next analyzed changes in B₂R protein stability as a potential cause for the reduced receptor count in the continuous presence of agonist. To this end we performed ³⁵S-labeled pulse-chase experiments by labeling cellular proteins with a relatively short pulse of [³⁵S]methionine-cysteine for 60 min and a subsequent chase for 24 h using ³⁵S-free medium with or without 1 µM bradykinin. We chose this method because of its enhanced sensitivity over Western blotting, which readily detects precursor (and possibly degradation) products of B₂R only in cells highly overexpressing the receptor. Immunoprecipitation of B₂R at the indicated time intervals allowed us to follow the fate of the ³⁵S-labeled receptor protein. SDS-PAGE of extracts from HF-15 cells immediately after the pulse (t = 0) showed a nonglycosylated 40-kDa B₂R precursor that is quantitatively converted into the glycosylated B₂R forms of 52 and 68 kDa within 3 h (Fig. 4A). Treatment of HF-15 cells with bradykinin for 3-24 h reduced the levels of ³⁵S-labeled B₂R by 20-40% as revealed by quantitative analysis of the radiograph using a PhosphorImager. The half-life of the "mature" B₂R forms of 52 and 68 kDa, which was ~9 h (range 8-10 h) in the absence of the ligand was reduced to 5 h (range 4-6 h) in the continuous presence of 1 µM bradykinin (Fig. 4B). Because the total amount of ³⁵S-labeled cellular protein did not differ in the absence or presence of ligand (data not shown), an unspecific or toxic effect of bradykinin treatment is unlikely to cause the observed loss of receptor protein. Rather, our experiments suggest that downregulation of B₂R is at least partially due to a reduced stability of B₂R proteins in the continuous presence of high bradykinin concentrations.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Half-life of B2R followed by 35S-labeled pulse-chase experiments. A: HF-15 cells grown on six-well dishes were pulse-labeled for 60 min with 0.33 mCi/ml [35S]methionine-cysteine (Pro-mix) in sulfur-free DMEM. Thereafter, radioactive medium was removed, and cells were incubated with DMEM-10% FCS in the absence (control) or presence of 1 µM BK. To compensate for BK degradation fresh BK was added every 6 h. After indicated periods, cells were lysed and B2R was immunoprecipitated with 2.5 µl of antiserum AS346 (anti-B2R). Proteins (~50-100 µg/lane) were separated by reducing 10% SDS-PAGE and visualized by fluorography. A representative fluorogram from 8 experiments is shown. B: quantitative analysis of A by PhosphorImager. Time courses are presented for control conditions (filled squares) and after BK treatment (open diamonds). The apparent half-life of B2R under the various conditions was determined graphically (broken lines).

Effect of bradykinin treatment on B₂ receptor biosynthesis. Because prolonged bradykinin exposure seems to affect B₂R stability on the protein level but not on the mRNA level, we wondered whether B₂R protein synthesis was also affected by long-term exposure to the agonist. To follow B₂R de novo protein synthesis, we modified the pulse experiment such that cells were labeled with ³⁵S-labeled amino acids for up to 24 h in the absence or continuous presence of 1 µM bradykinin. Immunoprecipitation of radiolabeled B₂R revealed that the continuous presence of bradykinin attenuates B₂R de novo synthesis at all investigated time intervals (Fig. 5A). Quantitative evaluation by PhosphorImager demonstrated that the amount of newly synthesized B₂R protein of treated fibroblasts is by 35-50% lower than that of untreated cells over the entire incubation period (Fig. 5B). Again, unspecific or toxic effects of bradykinin appear not to cause this phenomenon because incorporation of radioactive amino acids into the total cellular protein pools was identical in the presence or absence of the ligand (data not shown). Also enhanced receptor degradation may contribute to the observed phenomenon, although the increment due receptor proteolysis (~15% over 24 h) is much smaller than that contributed by the inhibition of de novo synthesis (up to 50% over 24 h). Thus long-term agonist treatment with bradykinin attenuates de novo synthesis of B₂R, suggesting a translational control of the B₂R levels.

View larger version (46K): [in this window] [in a new window]   Fig. 5.   Reduction of B2R de novo synthesis during prolonged agonist exposure. A: HF-15 cells grown on six-well dishes were incubated with 0.1 mCi/ml 35S in sulfur-free DMEM in the absence (control) or presence of 1 µM BK. To compensate for BK degradation fresh BK was added every 6 h. After the indicated time periods, cells were lysed and B2R was immunoprecipitated with 2.5 µl of antiserum AS346 (anti-B2R). Proteins (~50-100 µg/lane) were separated by reducing 10% SDS-PAGE and visualized by fluorography. A representative fluorogram of 4 experiments is shown. B: quantification of A by PhosphorImager analysis. Relative rates of B2R synthesis in the absence (solid bars) and presence (open bars) of 1 µM BK are given as means ± SD from 4 independent experiments. Significant differences between control and BK-treated cells were observed at 6, 12, and 24 h (P < 0.01, t-test).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Adaptation of GPCRs and their associated signaling systems to continuous agonist stimulation provides an important mechanism to protect cells from hyperresponsiveness to external signals. For instance many drugs targeted at GPCRs of neuronal cells are applied for weeks, months, or even for lifetime. Prototypes are serotonergic agonists such as buspiron used to relieve anxiety or dopaminergic effectors applied in Parkinson's disease. Amine uptake inhibitors commonly used for treatment of severe depressions increase serotonin and norepinephrine concentrations in synaptic clefts providing yet another example of sustained agonist exposure. Though the mechanisms of action of these drugs (and of some of their adverse effects) are not fully understood, it is well known that many, if not most, of them start showing their beneficial effects only after long-term application, i.e., after several days or even weeks. Thus adaptation and downregulation of corresponding GPCRs may be critically involved in the efficacy of these drugs (11, 17).

In the kininergic system, plasma kinin levels may significantly increase during antihypertensive therapy with angiotensin-converting enzyme (ACE) inhibitors. Because of their blood-pressure lowering effects, accumulated kinins contribute to the beneficial effects of ACE inhibitors, yet elevated kinin levels eventually lead to kinin receptor downregulation and reduced responsiveness to kinin peptides (40). Likewise, the use of long-lasting synthetic kinin agonists such as FR-190997 or RMP-7, which transiently open the blood-brain barrier and thereby facilitate the rapid diffusion of cytostatic drugs from the blood to brain tumors, may permanently stimulate and eventually downregulate their cognate receptors (2, 15). Also, small cell lung cancers can produce massive amounts of kinins stimulating the growth of cancer cells (37). Finally, hereditary forms of angioedema and the various forms of vasculitis (25) are characterized by the production of excessive amounts of kinins, which likely induce sustained downregulation of their receptors. Against this background, we considered studies with relatively high bradykinin concentrations (up to 1 µM) over a long period (up to 24 h) adequate to address the fate of kinin receptors during long-term agonist exposure of human foreskin fibroblasts. To counteract kinin degradation during prolonged incubation, e.g., due to cell-bound kininases, we added fresh bradykinin after every 6 h of incubation. Whereas we found that in vitro kinin generation, e.g., due to processing of FCS-borne kininogen, was almost negligible because most of the bovine kininogen was present in the kinin-free form (data not shown).

Our initial efforts focused on the development of a sensitive technique to detect B₂R protein endogenously expressed by native cells. To this end we used a polyclonal antibody for immunoprecipitation and a monoclonal antibody for immunoprinting of B₂R. Surprisingly, a commercial monoclonal antibody to B₂R, which has been widely used by many laboratories (12, 21, 41), failed to immunoprecipitate human B₂R and did not give specific signals in Western blots unless B₂R was highly enriched by immunoprecipitation from lysates of HEK cells massively overexpressing the receptor. This problem was overcome by the development of a monoclonal antibody, MBR1, which allowed us to follow the downregulation of endogenous B₂R protein. Long-term treatment with bradykinin decreased the number of specific binding sites as well as B₂R protein levels in foreskin fibroblasts. The reduction in B₂R count was not due to transcriptional regulation, because B₂R mRNA levels remained unchanged. Rather, receptor protein half-life and de novo protein synthesis were reduced suggesting translational and posttranslational control of cellular B₂R levels during chronic agonist stimulation.

An unanticipated finding of our studies was the apparent attenuation or even lack of B₂R sequestration during the first hours of agonist stimulation. This observation is in contrast with previous reports from many laboratories, including our own, in which receptor sequestration had been judged from radioligand internalization (7, 30, 31, 33). The seemingly conflicting results may be reconciled by a scenario where recycling of the internalized receptor to the cell surface is much faster than the intracellular degradation of the dissociated radioligand. Likewise, Lamb and co-workers (29) observed an apparent discrepancy between ligand internalization and B₂R sequestration. Thus radioligand internalization studies follow one-way dynamics of receptor trafficking, whereas our present studies address the actual amount of cell surface-exposed B₂R, which is balanced by receptor endocytosis, recycling, de novo synthesis, and downregulation.

The role of sequestration for downregulation of GPCRs is still obscure. Initial studies on the ₂-adrenergic receptor have indicated that endocytosis is crucial for receptor degradation, whereas a more recent report suggested downregulation to occur at the plasma membrane (20, 24). For the B₂R, sequestration has been studied using concanavalin A, an effective inhibitor of internalization; however, long-term toxicity of the drug did not allow to draw any firm conclusions from these experiments (data not shown). We anticipate that mutant B₂R with impaired internalization and/or phosphorylation features (7, 31) may help unravel the delicate role of these mechanisms for receptor targeting (10). Also the molecular determinants directing GPCRs for degradation or for recycling are largely unknown. First insights have been gained for the ₂-adrenergic receptor, where four residues at its extreme COOH-terminus engage in PDZ domain-mediated protein interactions that direct sequestrated receptors for recycling (19). Furthermore, a GPCR-associated sorting protein (GASP) has been identified, which binds to -opioid receptors prone to endocytosis and subsequent degradation but not to rapidly recycling µ-opioid receptors spared from downregulation (44). Disruption of receptor-GASP complex prevented -opioid receptor degradation and redirects it for recycling. Presently, the role of the tail domain of B₂R in orchestrating recycling and/or downregulation of the receptor during long-term agonist stimulation is unknown.

Two other mechanisms may partake in the downregulation of GPCRs, which are receptor ubiquitination and mRNA destabilization. Ubiquitination of yeast GPCRs has been known for a while (34), and more recently ubiquitination has been implicated in endocytosis and downregulation of the ₂-adrenergic receptor (38). For B₂R no systematic studies addressing the role of ubiquitination in receptor downregulation are available. Our present studies did not reveal the presence of high molecular weight B₂R conjugates that may indicate receptor ubiquitination. Alternatively, GPCR can be downregulated by the destabilization of their corresponding mRNAs. For the ₂-adrenergic receptor it has been shown that protein kinase A and mRNA binding proteins are involved in this process (9, 22, 32). For the B₂R, we did not find any changes in mRNA levels during long-term agonist treatment; however, we cannot exclude the possibility that even a modest change in mRNA levels, which would escape our detection system, may have a profound effect on receptor biosynthesis. Our ³⁵S-labeling experiments suggest that B₂R biosynthesis may be reduced during continuous stimulation with agonist. Proteins such as ferritin and 15-lipoxygenase are prone to a gene-specific regulation of their mRNA translation (14). Also the initiation of translation is controlled by kinase pathways downstream of B₂R and other GPCRs (27), and therefore sustained agonist exposure may modulate B₂R biosynthesis. Clearly, more in-depth analyses are required to elucidate the translational and posttranslational mechanisms mediating downregulation of bradykinin receptors in the continuous presence of the ligand.