Abstract
Upon sustained insult, kinins are released and many kinin responses, such as inflammatory pain, adapt from a B2 receptor (B2R) type in the acute phase to a B1 receptor (B1R) type in the chronic phase. In this study, we show that kinins modulate receptor endocytosis to rapidly decrease B2R and increase B1R on the cell surface. B2Rs, which require agonist for activity, are stable plasma membrane components without agonist but recruit β-arrestin 2, internalize in a clathrin-dependent manner, and recycle rapidly upon agonist treatment. In contrast, B1Rs, which are inducible and constitutively active, constitutively internalize without agonist via a clathrin-dependent pathway, do not recruit β-arrestin 2, bind G protein-coupled receptor sorting protein, and target lysosomes for degradation. Agonist delays B1R endocytosis, thus transiently stabilizing the receptor. Most of the receptor trafficking phenotypes are transplantable from one receptor to the other through exchange of the C-terminal receptor tails, indicating that the tails contain epitopes that are important for the binding of protein partners that participate in the endocytic and postendocytic receptor choices. It is noteworthy that the agonist delay of B1R endocytosis is not transplanted to the B2R via the B1R tail, suggesting that this property of the B1R requires another domain. These events provide a rapid kinin-dependent mechanism for 1) regulating the constitutive B1R activity and 2) shifting the balance of accessible receptors in favor of B1R.
Kinins are proinflammatory peptides that are formed extracellularly in response to pathological insult and act through two G protein-coupled receptors (GPCR), named B1 (B1R) and B2 (B2R) (Leeb-Lundberg et al., 2005). Through these receptors, kinins stimulate multiple inflammatory responses such as vasodilatation, vascular permeability, hyperalgesia, and pain. Sustained insult and various pathological conditions cause many kinin responses to adapt from a B2 type in the acute phase to a B1 type in the chronic phase (Dray and Perkins, 1993; Ongali et al., 2003).
GPCR signaling is tightly regulated at several steps in the receptor life cycle, ranging from receptor gene transcription to degradation of the mature receptor. At the cell surface, activation of the receptor by agonist is often followed by receptor phosphorylation via a G protein-coupled receptor kinase and subsequent arrestin binding to the receptor, which results in homologous desensitization (i.e., uncoupling of the receptor from the G protein and thereby from downstream signal transduction pathways) (Lefkowitz et al., 1998). Many GPCRs also subsequently undergo endocytosis through either clathrin-coated structures or lipid rafts, such as caveolae (Ferguson, 2001; Chini and Parenti, 2004). After endocytosis, GPCRs are either recycled back to the plasma membrane, leading to functional resensitization; remain internalized through storage, leading to desensitization; or are targeted to lysosomes for degradation, leading to functional down-regulation (von Zastrow, 2003). The postendocytic fate of a GPCR is determined via direct interactions of the receptor with various protein partners such as ERM-binding phosphoprotein 50/Na+-H+ exchanger regulatory factor, N-ethyl maleimide-sensitive factor, sorting nexin 1, GPCR-associated sorting protein (GASP), and lipid membrane components (Hall et al., 1998; Cao et al., 1999; Whistler et al., 2002; Xiang and Kobilka, 2003; Heydorn et al., 2004). The responsiveness of a GPCR signaling system is, to a large degree, determined by these regulatory events.
Kinin receptor adaptation is explained, at least in part, by the induction of B1R expression by pro-inflammatory cytokines such as interleukin-1β (Bachvarov et al., 1998), which is observed in cellular systems (Phagoo et al., 1999; Bengtsson et al., 2006). This induction may function to convert kinin signaling from a transient one to a sustained one because B1R is constitutively active (Leeb-Lundberg et al., 2001). Furthermore, B1R is not phosphorylated either basally or in response to agonist (Blaukat et al., 1999) and desensitizes only marginally upon further stimulation by the agonist (Mathis et al., 1996). On the other hand, B2R is subject to both basal and agonist-promoted phosphorylation (Blaukat et al., 1999, 2001) and desensitizes rapidly upon agonist stimulation (Mathis et al., 1996).
Kinins decrease the availability of cell surface B2R and increase the availability of B1R in some cell systems (Phagoo et al., 1999; Bengtsson et al., 2006), which may contribute to kinin receptor signal adaptation. The details of this regulation are not well understood, but some studies have suggested it may involve changes in receptor trafficking. Most trafficking studies of B2R and B1R have used bound radiolabeled agonist to monitor the receptors (Munoz and Leeb-Lundberg, 1992; Munoz et al., 1993; Faussner et al., 1998; Pizard et al., 1999; Lamb et al., 2001). In a few cases, receptors fused in their C-terminal domains to fluorescent proteins have also been used for this purpose (Bachvarov et al., 2001; Lamb et al., 2001; Sabourin et al., 2002). The former approach is restricted to monitoring only agonist-promoted changes, and is sensitive to changes in agonist affinity, whereas the latter approach requires receptor constructs drastically modified in their C-terminal domain, which can be critical for the postendocytic fate of the receptor (Cao et al., 1999; Finn and Whistler, 2001; Whistler et al., 2002; Gaborik and Hunyady, 2004; Heydorn et al., 2004; Bartlett et al., 2005). Nevertheless, these studies have proposed that B2R is rapidly and almost completely internalized by the agonist, whereas B1R is much less or differently affected by the agonist in this regard.
Our aim with this study was to use a representative cellular model with minimally modified receptors to determine the role of cellular trafficking in the rapid and inverse kinin regulation of B2R and B1R. HEK293 cells were chosen as a model system because this cell line is well characterized, easy to use and manipulate, has been shown to be a valuable tool in the trafficking studies of other model GPCR systems, and allows for the independent study of the trafficking of each receptor in the absence of regulation of de novo receptor synthesis.
Materials and Methods
Cell Culture and DNA Constructs. IMR90 human embryonic lung fibroblast cells (American Type Culture Collection, Manassas, VA) were grown in minimum essential medium containing 200 U/ml penicillin, 200 μg/ml streptomycin, 2 mM l-glutamine (Invitrogen, Carlsbad, CA), and 10% fetal bovine serum (FBS) (Hyclone Laboratories, Logan, UT) in 5% CO2 at 37°C. HEK293 cells (American Type Culture Collection) were grown in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen) supplemented with 10% FBS in 10% CO2 at 37°C. The human B2R and B1R cDNA were subcloned into a pcDNA3.1 vector containing a Zeocin selection marker. An N-terminal artificial signal sequence, as described previously (Whistler et al., 2002), and the FLAG sequence tag were added in series to make the B2R and B1R constructs named SFB2 and SFB1, respectively. Receptor constructs with exchanged C-terminal tails starting from the first amino acid after the NPXXY motif (B1CB2 and B2CB1) and B1R truncated after residue Phe319 (B1Stop320), which is located seven residues beyond the NPXXY motif, were created by PCR. GFP-β-arrestin 2 was kindly provided by Dr. Marc Caron (Duke University Medical Center, Durham, NC) (Barak et al., 1997). The cells were transfected using the calcium phosphate precipitate method. Single colonies were then chosen and propagated in the presence of selection-containing media to generate clonal stable cell lines.
Radioligand Binding Assay. Assays were performed on particulate preparations and on intact cells using [3H]BK and [3H]DAKD as described previously (Leeb et al., 1997; Phagoo et al., 1999). Pretreatment of intact cells with agonist at 37°C was followed by a 6-min wash with ice-cold 50 mM glycine-HCl, pH 3, and two washes in ice-cold phosphate-buffered saline (PBS) to remove the agonist.
Phosphoinositide Hydrolysis. Confluent cells were labeled with 0.4 μCi/well [myo-3H]inositol for 16 to 20 h in inositol-free DMEM containing 0.5% bovine serum albumin in 48-well dishes, washed in Leibovitz's L-15 medium, pH 7.4, containing 10 mM LiCl, and then stimulated with increasing concentrations of agonist for 30 min at 37°C. The cells were then lysed with 100 mM formic acid, and the lysate was mixed with anion exchange resin in a 48-well plate with a small hole in the bottom of each well. The resin was then washed sequentially with water and 60 mM ammonium formate by aspiration. Inositol phosphates were eluted from the resin with 1 M ammonium formate and counted for radioactivity in a scintillation counter (LS6000; Beckman Coulter, Fullerton, CA).
FACS Analysis. Confluent cells in 100-mm dishes were treated with and without agonist in DMEM and 10% FBS. The cells were then trypsinized and washed with ice-cold PBS plus Ca2+/Mg2+. Cells were then resuspended in 100 μl of ice-cold PBS plus Ca2+/Mg2+ with 50% FBS, 0.4 μg of M1 anti-FLAG antibody conjugated with APC using a commercial kit (Prozyme, San Leandro, CA), 1 μg of mouse IgG1 κ (MOPC 21), incubated for 20 min at 4°C, which was followed by two washes in PBS plus Ca2+/Mg2+. The cell pellet was finally resuspended in 1 ml of PBS plus Ca2+/Mg2+ with 50% FBS. Cells (∼20,000) were counted by FACS, and each receptor-positive cell line was gated against untransfected HEK293 cells to reduce background staining. Mean fluorescence was calculated for the gated signal-positive population of untreated and agonist-treated cells.
Biotinylation Protection Assay. Confluent cells were washed twice with ice-cold PBS and then incubated with 0.3 mg/ml disulfidecleavable sulfo-NHS-SS-biotin (Pierce) in PBS for 30 min at 4°C with gentle agitation. The cells were then washed twice with Tris-buffered saline (TBS) to quench the biotinylation reaction and returned to DMEM and 10% FBS for treatments. Cells were allowed to reequilibrate to growth conditions for 30 min at 37°C before further experimentation. Cells labeled Total and Strip in the figures were left on ice in TBS. At 37°C, cells were treated without (NT) or with 1 μM desArg10kallidin (DAKD) or 1 μM bradykinin (BK) (AG) for the indicated intervals and washed twice with ice-cold TBS. The remaining cell surface-biotinylated receptors on the cells together with those cells designated Strip were stripped with 50 mM glutathione, 0.3 M NaCl, 75 mM NaOH, and 1% FBS for 30 min at 4°C. The glutathione was then quenched using a 20-min wash with PBS containing 50 mM iodoacetamide and 1% bovine serum albumin. Cells were then extracted in 0.1% Triton X-100, 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 25 mM KCl containing a complete protease inhibitor cocktail (Roche), and cell debris was removed by centrifugation at 10,000g for 10 min at 4°C. Receptors were immunoprecipitated in the extraction buffer by incubating in anti-FLAG M2 affinity resin (Sigma) overnight at 4°C. The precipitates were washed extensively and sequentially in the extraction buffer and in 10 mM Tris-HCl, pH 7.4. The receptors were then deglycosylated for 2 h at 37°C with PNGase F (New England Biolabs, Ipswich, MA). Proteins were denatured in SDS sample buffer without reducing agent, fractionated by SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and the membrane was blocked for at least 1 h in TBS containing 0.1% Tween 20 and 10% nonfat milk. Biotinylated proteins were visualized by incubating with the Vectastain avidin-biotinylated enzyme complex immuno-peroxidase reagent (Vector Laboratories, Burlingame, CA) followed by development with ECL reagents according to the manufacturer's instructions (GE Healthcare, Little Chalfont, Buckinghamshire, UK).
Biotinylation Degradation Assay. Confluent cells were washed twice with ice-cold PBS and then incubated with 0.3 mg/ml disulfidecleavable sulfo-NHS-SS-biotin in PBS for 30 min at 4°C with gentle agitation. The cells were then washed twice with TBS to quench the biotinylation reaction and returned to DMEM and 10% FBS. Cells were then left on ice (Total) or returned to 37°C for 30 min after which they were incubated without (NT) or with 1 μM DAKD or BK (AG) for 120 min at 37°C. The receptors were then extracted, immunoprecipitated, and detected as described above under Biotinylation Protection Assay.
Fluorescence Microscopy. Cells were propagated in growth media on glass coverslips to 50% confluence and then treated in one of three ways. For whole-cell receptor distribution experiments, cells were incubated without and with agonist or antagonist for 30 min at 37°C. For cell surface receptor redistribution experiments, cells were incubated in media containing 3.5 μg/ml primary mouse anti-FLAG M1 monoclonal antibody (Sigma) for 30 min at 37°C to label surface receptors. The cells were then incubated without agonist or antagonist or a combination of the two for an additional 30 min at 37°C. For receptor recycling experiments, cells were incubated as described immediately above followed by stripping of remaining cell surface M1 antibody by depleting the media of Ca2+ with PBS containing 0.1% EDTA at room temperature. The cells were then further incubated without and with agonist or antagonist for an additional 30 min at 37°C. In all experiments, cells were then fixed using 3.5% formaldehyde in PBS and permeabilized using 0.1% Triton X-100. For visualization of receptors, fixed specimens were incubated with Alexa488-labeled anti-mouse IgG2b antibody (Invitrogen) to detect FLAG M1 antibodies bound to the tagged receptors. For colocalization of FLAG-tagged receptors with LAMP 1 and LAMP 2, primary H4A3 and H4B4 mouse IgG1 (Developmental Studies Hybridoma Data Bank) antibodies were incubated with secondary Alexa568-labeled anti-mouse IgG1 antibody (Invitrogen), to detect LAMP 1 and LAMP 2, along with secondary Alexa488-labeled antimouse IgG2b, to detect FLAG M1 antibodies bound to receptors. Images were collected by confocal microscopy using a Nikon Eclipse confocal microscope. Quantification of colocalized fluorescence was done using the Imaris software (Bitplane AG, Zurich, Switzerland).
β-Arrestin Recruitment. Cells stably expressing GFP-β-arrestin 2 were transiently transfected with SFB1, SFB2, SFB1CB2, or SFB2CB1. Transfected cells were plated on polylysine-treated glass coverslips in six-well dishes for 24 h before assay. The cells were then incubated in growth media without and with agonist for 5 min at 37°C. Cells were then fixed in PBS containing 3.5% formaldehyde for 30 min followed by primary staining with M1 anti-FLAG antibodies and secondary staining with Alexa568-labeled anti-mouse IgG1 antibody. Images were then collected by confocal microscopy using a confocal microscope (Eclipse; Nikon, Tokyo, Japan).
In Vitro Affinity Assay of GST Fusion Proteins and GASP. B1R and B2R receptor C-terminal tails starting after the NPXXY motif were cloned into the pGEX4t1 vector and C-terminal tail-GST fusion proteins were produced and attached to glutathione-Sepharose (Sigma, St. Louis, MO). Full-length GASP was in vitro-translated using the TnT quick-coupled transcription/translation systems (Promega, Madison, WI) with 35S-labeled methionine. Fusion proteins and radiolabeled probe were mixed in a wash solution containing 0.1% Triton X-100 and incubated for 1 h at room temperature. Glutathione resin was washed several times followed by denaturation by boiling in SDS sample buffer and proteins fractionated by SDS-polyacrylamide gel electrophoresis. Protein concentrations of receptor tails and controls were estimated by Coomassie staining of the gel, and pulled down probe was visualized by autoradiography.
Coimmunoprecipitation of Receptor and GASP. Confluent cells were washed twice with ice-cold PBS and lysed in 1 ml of lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 10 mM NaF, and 10 mM Na2HPO4) with complete protease inhibitor cocktail (Roche, Indianapolis, IN). Lysates were cleared by centrifugation at 14,000g for 15 min at 4°C. An aliquot of the cleared lysates was withdrawn from each sample to compare the GASP concentration by immunoblotting as described below. The rest of the cleared lysate was immunoprecipitated using 15 μl of M2 anti-FLAG affinity resin (Sigma) for 1 h at 4°C, washed extensively with 0.1% Triton X-100, 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 25 mM KCl containing a complete protease inhibitor cocktail (Roche), and deglycosylated with PNGase F (New England Biolabs) for 2 h. Precipitates were resolved ona4to 20% gradient Tris-HCl precast gel (Bio-Rad Laboratories, Hercules, CA) and transferred to a polyvinylidene difluoride membrane. The blots were cut below the 75-kDa marker band to separately immunoblot for either receptor (lower blot) or GASP (upper blot). GASP blots were incubated for 2 h at room temperature with rabbit anti-GASP (1:1000) and for 1 h at room temperature with HRP-conjugated anti-rabbit antibody (New England Biolabs) (1:4000), then visualized with ECL Plus (GE Healthcare). Receptor blots were incubated for 2 h with biotinylated FLAG M2 antibody (1:250) (Sigma) and then visualized with streptavidin overlay (Vectastain ABC reagents; Vector Laboratories) and ECL Plus.
Data Analysis. Data were analyzed by the Student's t test as indicated.
Results
Effect of Kinin Agonists on Cell Surface Availability of Kinin Receptors. Human embryonic lung fibroblast IMR90 cells express B2R and B1R at 1283 and 15 fmol/mg of protein, respectively (Phagoo et al., 1999). Treatment of these cells with a 0.1 μM concentration of the cognate B2R ligand BK at 37°C led within 30 min to a 78% decrease in the amount of B2R available on the cell surface, as determined by radioligand binding, and this effect was maintained at 120 min of treatment (Fig. 1A). In contrast, treatment with a 0.1 μM concentration of the cognate B1R ligand DAKD led within the same time period to an 81% increase in the number of available B1R, which was also maintained at 120 min of treatment (Fig. 1A). Thus, kinins rapidly and inversely regulate the cell surface availability of native B2R and B1Ras reported previously (Phagoo et al., 1999; Bengtsson et al., 2006).
To address the mechanism underlying this rapid inverse regulation, we generated HEK293 cell clones stably expressing FLAG-tagged human B2RorB1R at 1921 ± 496 and 343 ± 40 fmol/mg of protein (n = 3), respectively. These cells bound and responded to their respective agonist ligands with affinities and potencies similar to those of their wild-type counterparts (Table 1). FACS analysis with APC-conjugated M1 FLAG antibodies revealed that BK and DAKD treatment of these cells for 30 min at 37°C led to a 44% decrease and a 16% increase in the amount of B2R and B1R available at the cell surface, respectively (Fig. 1B). The lower percentage increase in B1R in the stable HEK293 cells compared with the IMR90 cells is probably due in part to the relatively higher amount of receptors expressed in the former cells. Nevertheless, stable HEK293 cells reflect the IMR90 cells concerning agonist-induced changes in surface receptor number. We thus used these cell lines to investigate the molecular mechanism whereby agonist ligand regulates these receptors.
Receptor-Mediated β-Arrestin Recruitment. β-Arrestin is a key GPCR regulatory protein that is rapidly recruited from the cytosol to the plasma membrane receptor upon agonist stimulation and involved in both receptor desensitization and internalization. Agonist stimulation of B2R led to β-arrestin 2 recruitment to the plasma membrane within 5 min (Fig. 2, B2 NT, AG). In contrast, no apparent effect on β-arrestin 2 distribution was observed either without or with agonist stimulation of B1R for the same time period (Fig. 2, B1 NT, AG). These results are consistent with the different agonist regulation of these receptors, where B2R is rapidly phosphorylated at specific serines and threonines in the C-terminal tail (Fig. 1C) (Blaukat et al., 1999, 2001) and desensitized by the agonist (Mathis et al., 1996), whereas B1R is apparently not phosphorylated (Fig. 1C) (Blaukat et al., 1999) and desensitizes very slowly in response to agonist (Mathis et al., 1996), as well as being constitutively active (Leeb-Lundberg et al., 2001).
Identification of the Receptor Domain Responsible for Differential Endocytic Trafficking of Receptors. The receptor C-terminal tail is important in the trafficking of many GPCRs. We thus examined whether the C-terminal tails of the B1R and the B2R conferred trafficking specificity (Fig. 1C). Receptors were generated in which the tail of the B1R was transplanted to the B2R (B2CB1) and the tail of the B2R was transplanted to the B1R (B1CB2). This exchange does not affect the agonist specificity of the receptors (Leeb-Lundberg et al., 2001; Kang and Leeb-Lundberg, 2002). We then examined the effects of the tail exchange on β-arrestin 2 recruitment and agonist-induced changes in cell surface receptor number. Substitution of the B2R tail in B1R created a receptor (B1CB2) that rapidly recruited β-arrestin 2 in response to DAKD (Fig. 2, B1CB2 NT, AG) in contrast to the B1R. These data suggest that the B2R tail contains an epitope important for β-arrestin recruitment. This epitope is most likely the phosphorylation of specific serines and threonines in the B2R tail, which is lacking in the B1R tail (Fig. 1C). Figure 1B shows that DAKD decreased the cell surface availability of B1CB2 by 20%, making it significantly different from B1R in which DAKD increased cell surface availability (p = 0.0003). Substitution of the B1R tail in B2R created a receptor (B2CB1) that was unable to recruit β-arrestin 2 in response to BK (Fig. 2, B2CB1 NT, AG). BK had no effect on the availability of B2CB1 making it significantly different from B2R in which BK decreased cell surface availability (p = 0.03). However, the C-terminal tail alone was not enough to transfer the B1R phenotype of agonist-promoted increase in receptor availability to the B2R.
Whole-Cell Distribution of Receptors. We next examined the distribution of the receptors and the tail swaps in the presence and absence of ligand. The whole-cell distribution of each receptor was analyzed by confocal fluorescence microscopy of formaldehyde-fixed and permeabilized cells. Under these conditions, the vast majority of B2R were located in the plasma membrane in the absence of agonist (Fig. 3A, B2 NT). On the other hand, treatment with 1 μM BK (B2 AG) for 30 min at 37°C promoted the redistribution of receptors to intracellular compartments, although some plasma membrane staining was still observed. Treatment with a 1 μM concentration of the B2R-specific antagonist icatibant alone had little effect on B2R distribution (data not shown). In contrast to B2R, a disperse pattern of B1R distribution was observed even in the absence of agonist, with receptors localized primarily intracellularly, in both a tubular network and in distinct puncta, and to a lesser degree in the plasma membrane (Fig. 3A, B1 NT). No obvious change in receptor distribution was observed after treatment with 1 μM DAKD (B1 AG) or the B1R-specific antagonist des-Arg10-icatibant (data not shown). Thus, the steady-state distribution of B2R and B1R are different, with a much greater intracellular pool of B1R. The tubular network may represent slow B1R maturation, whereas the distinct puncta may represent constitutive B1R endocytosis (see below). B1CB2 exhibited both basal and agonist-promoted distribution patterns (Fig. 3A, B1CB2 NT, AG) that were very similar to the B2R. Likewise, the B2CB1 distribution patterns in both the presence and absence of agonist (B2CB1 NT, AG) were very similar to B1R. Thus, the receptor C-terminal tails seem to be involved in both receptor maturation and endocytosis.
Constitutive and Kinin-Dependent Endocytosis of Receptors. To specifically address receptor endocytosis without the interference of receptor maturation, we used two assays that specifically monitor the fate of the plasma membrane pools of receptors. The first assay was an immunofluorescence confocal microscopy assay in which live cells are “fed” primary antibody to label surface receptors. The second assay was a quantitative biochemical assay, the biotinylation protection assay (BPA). The immunofluorescence assay takes advantage of the extracellular FLAG epitope tag on the receptors to monitor selectively the movement of cell surface receptors. Live cells were first incubated with M1 anti-FLAG antibody in the absence (NT) or presence of agonist (AG) for 30 min at 37°C. The cells were then fixed, permeabilized, and incubated with secondary antibody to detect receptor-antibody complexes. In the absence of agonist, B2R were located almost exclusively in the plasma membrane (Fig. 3B, B2 NT). Incubation with agonist at 1 μM promoted endocytosis of the B2R-antibody complexes (B2 AG), which was blocked by coincubation with a 5-fold excess (5 μM) of antagonist (data not shown). On the other hand, spontaneous agonist-independent endocytosis of the B1R-antibody complexes was found to occur (Fig. 3B, B1 NT). No apparent change in B1R distribution was observed when cells were treated with the agonist at 1 μM either without (B1 AG) or with a 5-fold excess (5 μM) of antagonist (data not shown). It is noteworthy that the tubular distribution of B1R that was seen in Fig. 3A was absent using this protocol, which strongly argues that the intracellular pool of B1R is a combination of both slow receptor maturation and constitutive receptor endocytosis.
In the BPA, cell surface receptors were first selectively labeled with a membrane-impermeable biotin analog. Internalized receptors were then selectively monitored by their protection from stripping by a membrane-impermeable reducing agent. This assay revealed internalization of B2R only in the presence of agonist (Fig. 4A, compare B2 NT 30 min and B2 AG 30 min), whereas B1R internalization occurred both in the absence and presence of agonist (Fig. 4A, compare B1 NT 30 min and B1 AG 30 min). These data confirm the results obtained using the immunofluorescence microscopy assay (Fig. 3). It is noteworthy that detailed densitometric analysis of the BPA revealed that significantly less B1R (43%, p = 0.02) was internalized in the presence of agonist than in the absence of agonist at 30 min (Fig. 4B). Thus, the B1R is subject to constitutive endocytosis but the agonist delays this event. In contrast, the B2R resides in the plasma membrane until it is exposed to agonist, at which point it rapidly internalizes. The relatively low amount of internalized B2R recovered in this assay may be explained in part by the extraction conditions used. However, other factors are probably also involved, because B1CB2, which behaved similarly to B2R, was significantly easier to extract.
B1CB2 was located exclusively in the plasma membrane in the absence of agonist, whereas incubation with agonist led to rapid endocytosis of the construct as evidenced by both the immunofluorescence assay (Fig. 3B, B1CB2 NT, AG) and BPA (Fig. 4, A and B). Thus, B1CB2 gains the endocytic features of B2R, which is consistent with introduction of both agonist-promoted β-arrestin recruitment (Fig. 2) and agonist-promoted decrease in the cell surface receptors (Fig. 1B). The immunofluorescence assay revealed that B2CB1 is constitutively internalized (Fig. 3B). This constitutive endocytosis, like that for the B1R, apparently proceeds by a β-arrestin 2-independent mechanism, because this receptor does not recruit β-arrestin 2 in either the presence or the absence of agonist (Fig. 2). The low level of expression of this construct in the stable cell lines prohibited BPA analysis. We were thus unable to determine whether agonist decreased receptor endocytosis of the B2CB1, as it does for the B1R (Fig. 4). However, the failure of BK to increase the number of accessible cell surface B2CB1 (Fig. 1B) suggests that the agonist does not delay the endocytosis of this receptor.
Inhibition of Receptor Endocytosis by Hyperosmotic Sucrose Treatment. Treatment of cells with hyperosmotic sucrose treatment has been shown to yield abnormal clathrin polymerization resulting in empty microcages in the membrane (Heuser and Anderson, 1989). Pretreatment of cells stably expressing B1R with 0.4 M sucrose for 60 min led to a 31% increase in the number of available B1R on the cell surface (Fig. 5A). Confocal microscopy of cells incubated with primary antibody during the last 30 min of the sucrose treatment showed that the increase in B1R was due to inhibition of constitutive receptor endocytosis (Fig. 5B). On the other hand, treatment with 5 mM methyl-β-cyclodextrin, which depletes the cells of cholesterol and perturbs lipid rafts, had no apparent effect on B1R endocytosis (data not shown). Hyperosmotic sucrose treatment had no effect on the number of cell surface B2R (Fig. 5A) or their localization in the plasma membrane (Fig. 5B). However, sucrose completely inhibited the agonist-promoted B2R endocytosis (Fig. 5B). Thus, both constitutive B1R and agonist-promoted B2Rendocytosis proceed through a clathrin-dependent pathway.
Postendocytic Sorting and Stability of Receptors. We next examined the postendocytic sorting of the B1R and B2R. First, a modification of the immunofluorescence assay described above in Fig. 3B was used to monitor the postendocytic trafficking profiles of the internalized receptors. To monitor receptor recycling, receptors were first allowed to internalize in the presence of the M1 anti-FLAG antibody and in the presence or absence of agonist for 30 min as described in Fig. 3B. Because the M1 antibody is dependent on calcium ions for binding to the FLAG epitope, any antibody remaining on the cell surface, either nonspecifically or receptor-bound, was then stripped by washing with EDTA, which also removes any residual agonist. As a result, only internalized receptors remain bound to the antibody. Any return of internalized receptor-antibody complexes to the plasma membrane can then be monitored as described previously by Finn and Whistler (2001) and Bartlett et al. (2005). According to this protocol, a protected pool of intracellular B2R-antibody complexes was observed primarily only after agonist treatment (Fig. 6, compare B2NT STRIP with B2 AG STRIP), whereas protected B1R-antibody complexes were observed both without and with agonist treatment (B1 NT STRIP, AG STRIP) confirming the significant constitutive endocytosis of B1R. The pool of B2R that was internalized by agonist rapidly recycled to the plasma membrane after further incubation in the absence of agonist for 30 min (Fig. 6, B2 AG STRIP NT). In contrast, the internalized pool of B1R remained intracellular after further incubation in the absence of agonist (B1 AG STRIP NT).
To determine whether B1R enters a degrading pathway, we first used the BPA, which selectively monitors the stability of internalized receptors. Clearly, the internalized B1R were significantly degraded between the 30- and 120-min time points in both the absence (Fig. 7, NT) (p < 0.05) and presence of agonist (AG) (p < 0.01). The degrading phenotype of B1R was further investigated by colocalization of receptors with LAMP 1 and LAMP 2, which are markers for the endstage compartments of the degrading pathway. B1R showed an overlap with LAMP 1 and LAMP 2 at 120 min both without (Fig. 8, B1 NT) (29%, n ≥25 cells) and with agonist treatment (B1 AG) (25%, n ≥25 cells) that was much higher than for B2R without (B2 NT) (0.01%, n ≥25 cells, p < 0.001) and with agonist (B2 AG) (0.66%, n ≥25 cells, p < 0.001) at this time point. Together these data demonstrate that the B2R is targeted to a recycling pathway, whereas the B1Ris targeted to a degrading pathway.
Because the C-terminal tails conferred the endocytic properties of the receptors, we next examined whether they were involved in postendocytic receptor sorting. The B1CB2 receptor recycled to the plasma membrane upon agonist removal (Fig. 6, B1CB2 AG STRIP NT) and exhibited limited overlap with LAMP 1 and 2 without (Fig. 8, B1CB2 NT) (0.14%, n ≥25 cells) and with agonist (B1CB2 AG) (0.57%, n ≥25 cells). Furthermore, a B1R truncated of its C-terminal tail (B1Stop320) internalized constitutively (Fig. 6, B1STOP320 NT) but did not recycle after endocytosis (B1STOP320 NT STRIP AG). In addition, the B2CB1 receptor did not recycle back to the plasma membrane (Fig. 6, B2CB1 AG STRIP NT) and showed overlap with the lysosomal markers without (Fig. 8, B2CB1 NT) (13%, n ≥25 cells, p < 0.001) and with agonist (B1CB2 AG) (20%, n ≥25 cells, p < 0.001) that was significantly higher than both B2R and B1CB2. Thus, the C-terminal tails of these receptors contain epitopes that participate in the postendocytic sorting of the receptors.
Interaction of Receptors with GASP. GASP is one candidate molecule responsible for preferentially targeting B1R to the lysosomal degrading pathway as opposed to the plasma membrane recycling pathway. We have previously established that this protein specifically interacts with the intracellular C-terminal tail of some GPCRs targeted for destruction by lysosomal proteolysis (Whistler et al., 2002; Bartlett et al., 2005). Because the B1R was degraded after its endocytosis whereas the B2R was not, we examined whether GASP binds specifically to the former receptor.
A difference in GASP1 binding to B1R and B2R was first evaluated using in vitro affinity assays with in vitro-translated GASP1 and GST fusion proteins corresponding to the C-terminal tails of B2R and B1R beyond the NPXXY motif. GASP1 showed at least a 5-fold higher affinity for the B1R tail than for the B2R tail (Fig. 9A). Furthermore, the binding of GASP1 to the B1R tail was directly dependent on the concentration of the tail in the in vitro affinity assays. The binding preference of GASP1 for B1R was also evaluated in vivo in HEK293 cells stably expressing each receptor. Immunoprecipitation of receptors with M1 anti-FLAG antibodies revealed that endogenous GASP1 bound to the B1R but not the B2R (Fig. 9B). Thus, GASP1 exhibits a clear preference for B1R over B2R. Substituting the B2R C-terminal tail in the B1CB2 receptor decreased the binding of GASP1 by 50% (Fig. 9B). Thus, GASP1 binding to B1R involves the C-terminal tail of the receptor, but another epitope(s) also seems to be involved. Such multiepitopal binding seems to common for several GPCR effector proteins, including β-arrestins.
Effect of Kinins on Total Receptor Stability. The above results suggest that kinins increase available cell surface B1R by delaying constitutive receptor endocytosis and consequently delay their degradation. In contrast, kinins acutely decrease available B2R by promoting endocytosis, but they do not promote down-regulation of this receptor because they enter an endocytic recycling pathway. Thus, kinins would be expected to increase the stability of the total cell surface pool of B1R. To examine this hypothesis, a biotin degradation assay was used, which monitors the stability of the total pool of mature cell surface receptors. At 120 min, 29 ± 4% of the B1R originating on the plasma membrane remained intact (Fig. 10, NT120min), whereas the presence of agonist significantly increased this amount to 56 ± 14% at this time point (p = 0.03) (AG120min). B1CB2 was more stable than B1R at 120 min, and agonist did not have any significant effect on the B1CB2 receptor stability. The stability of B2CB1 at 120 min was similar to B1R but agonist did not have any effect on the B2CB1 receptor stability, which is consistent with the failure of the B1R C-terminal tail to transplant the agonist delay of constitutive endocytosis to B2R. Agonist was unable to delay the degradation of B1Stop320. On the other hand, truncation of the B1R C-terminal tail led to a significant decrease in the degradation of the receptor. Thus, the B1R tail contains an epitope that participates in targeting the receptor for degradation but additional epitopes are likely to be involved as well.
Discussion
Here, we used a HEK293 cell model to elucidate the trafficking mechanism underlying the relatively rapid kininpromoted decrease of B2R and increase of B1R on the cell surface seen in native cells. In our model, B2R are relatively stable cell surface receptor components in the absence of agonist but translocate rapidly and in a clathrin-dependent manner to intracellular endocytic vesicles in response to agonist stimulation with the cognate agonist ligand BK and recycle back to the plasma membrane. In contrast, B1R are subject to constitutive endocytosis, also via a clathrin-dependent pathway, and target lysosomes for degradation. Binding of the cognate B1R agonist DAKD inhibits constitutive B1R endocytosis thereby decreasing the rate of spontaneous B1R clearance from the cell surface and delaying degradation. This regulation is also consistent with the agonist-dependent and constitutive activities of these receptors and the pattern of inflammatory kinin signaling during sustained pathological insult in vivo, which adapts from a B2-type to a B1-type during the course of the insult. Furthermore, delaying spontaneous endocytosis to temporarily increase the steady-state level of receptors on the cell surface presents a novel mechanism by which some agonists may act to promote GPCR signaling.
B2R is expressed almost exclusively in the plasma membrane, whereas B1R is retained to a large degree intracellularly. The tubular shape of the receptor-retaining intracellular structures suggests a membrane-bound localization of the receptor rather than a diffuse cytosolic localization, which presumably reflects the biosynthetic pathway. PNGase treatment of biotinylated cell surface B1R led to the conversion of higher molecular mass receptor species to a relatively homogeneous species of approximately 35 kDa consistent with the idea that the cell surface receptor is glycosylated as suggested previously (Fortin et al., 2006). The difference in distribution of B1R and B2R was apparently determined entirely by the C-terminal tail. Thus, the B2R C-terminal tail may be equipped with a plasma membrane targeting motif or the B1R tail with an intracellular retention motif. Whether or not the intracellularly retained B1R species can be recruited to the plasma membrane is unclear. However, considering the rapid spontaneous endocytosis of the receptor once it reaches the cell surface and the postendocytic targeting of the receptor for destruction described here, the retained receptor pool could possibly be viewed as a deposit to ensure that a steady flow of receptors reach maturation once expression of the receptor has been induced by a pro-inflammatory stimuli.
The rapid agonist-promoted recruitment of β-arrestin by B2R, clathrin-dependent endocytosis, and the ability of this receptor to recycle after agonist-promoted endocytosis are consistent with previous studies. Indeed, agonist exposure of the B2R leads, within minutes, to internalization of the receptor protein, which has been determined by immunoelectron microscopy (Haasemann et al., 1998), fluorescence microscopy (Bachvarov et al., 2001), the appearance of acidresistant radiolabeled agonist binding (Munoz and Leeb-Lundberg, 1992), and the reduction in the number of cell surface binding sites, which rapidly reappear on the cell surface upon agonist removal (Munoz et al., 1993). The fact that the internalized agonist reaches a steady-state level over time that is higher than can be explained by the total number of cell surface binding sites (Lamb et al., 2001) further supports our finding that the internalized B2R recycles to the plasma membrane.
Unlike the B2R, the plasma membrane B1R is unstable in that it is subject to high spontaneous endocytosis and degradation. This observation is consistent with the inducible expression and high agonist-independent activity reported for this receptor (Leeb-Lundberg et al., 2001). In other words, the functional properties of this receptor necessitate a tight regulation at every level from transcription, through maturation and signaling, and ending in lysosomal targeting and destruction. Our results are also consistent with the rapid and spontaneous clearance of the B1R response in rabbit aortic rings (Fortin et al., 2003). On the other hand, our observation apparently contradicts the results from the same group using a C-terminally GFP-tagged B1R, which showed neither spontaneous nor agonist-promoted internalization, even though the receptor was reported to relocate on the plasma membrane in response to agonist (Sabourin et al., 2002). It is important to realize that the C-terminal tail of the receptor has been shown to be crucial for the trafficking properties of a number of GPCR (von Zastrow, 2003; Heydorn et al., 2004; Terrillon et al., 2004), which may be particularly apparent with spontaneous endocytosis (Waldhoer et al., 2003).
Previous studies have shown that the kinetics and amount of agonist internalization of the B1R, as determined by the appearance of acid-resistant DAKD binding, is very slow and low, which is opposite to the B2R, through which acid-resistant BK binding occurs very rapidly (Faussner et al., 1998; Lamb et al., 2001). This, at first, would seem to support a model of a noninternalizing receptor. However, as we present here, agonist binding to the B1R in fact retards the rapid spontaneous internalization of this receptor, thereby temporarily stabilizing the receptor on the cell surface. The sucrose sensitivity of both constitutive and agonist-dependent B1R endocytosis suggests a common endocytic pathway under the two conditions but that the agonist promotes a receptor conformation that has lower affinity for the endocytic machinery. Thus, the combined effect of spontaneous surface clearance of the B1R and the retardation of the clearance by the agonist may then be viewed as a coincidence detector that would only allow optimal receptor signal transduction at the precise right moment.
The B2R phenotype of agonist-dependent β-arrestin 2 recruitment and sucrose-sensitive internalization support an internalization pathway well in line with a majority of the GPCR studied to date. On the other hand, the B1R still lacks an identified adaptor protein that links it to the internalization machinery. Several receptors, such as GluR1 (Dale et al., 2001) and PAR-1 (Paing et al., 2002) have been reported to internalize by a clathrin-dependent but β-arrestin-independent mechanism. The mechanisms underlying B1R internalization, therefore, need further investigation for us to understand the regulation of receptor level and hence tissue responsiveness to receptor agonists.
Because most of the trafficking phenotypes of B2R and B1R were transplantable by C-terminal tail exchange, this domain must be important for the unique trafficking protein interactions. Substitution of the B1R tail in B2R yielded a loss of β-arrestin 2 binding, which may be explained by the apparent absence of phosphorylation of this tail (Blaukat et al., 1999), in contrast to the B2R tail, which is phosphorylated (Blaukat et al., 2001). The importance of the B2R C-terminal tail for agonist-promoted endocytosis was confirmed when substituted in B1R. In contrast to B2R, the B1R tail conferred constitutive endocytosis to B2R, a process that must then be β-arrestin 2-independent. The B1R tail was unable to transplant the agonist delay in constitutive endocytosis to the B2R. However, B1R tail truncation also lost the agonist delay suggesting that this delay in endocytosis involves a fine interplay between the C-terminal tail and other domains in the receptor.
The receptor tails are also important for the postendocytic choices of the receptors. Lysosomal B1R targeting may depend in part on GASP binding, which showed higher affinity for the B1R tail than the B2R tail both in vitro and in vivo. These results are consistent with the fact that substitution of the B1R tail in B2R prevented the receptor from recycling and promoted degradation. Furthermore, truncation of the B1R tail decreased the rate of receptor degradation but did not abolish it, indicating that additional domains are involved. Therefore, the B2R tail either contains an epitope important for recycling or is able to mask an epitope in B1R that is important for degradation.
The combined events of kinin-promoted rapid B2R desensitization and internalization, together with the recruiting effect of kinins and their limited desensitizing potential on the induced B1R, provide a mechanism for the inverse kinin regulation of these receptors observed in cultured cells and agree well with the prevailing paradigm of an acute inflammatory B2-type response that over time shifts toward a chronic B1-type response. Chronic phase B1R signaling has attracted a lot of attention as a drug target for chronic pain management. Compounds have been, and are being, screened for their ability to block the ligand-independent signal transduction of this receptor, so called inverse agonists, but no candidate drugs are of yet available. Our data in the present study favor an increased rate of B1R clearance as a means to reduce B1R activity. Hence, we suggest screening for endocytic agonist properties rather than signal transduction inverse agonist properties as a novel approach to evaluate “antagonism” at B1R.
Acknowledgments
We thank Margareta Pusch and Joanna Daszkiewicz-Nilsson for expert technical assistance, and Dr. Marc Caron for kindly providing GFP-β-arrestin 2.
Footnotes
-
This work was supported in part by funds from the Swedish Research Council grant 15057, AlfredÖsterlunds Stiftelse, Torsten och Ragnar Söderbergs Stiftelser, the Faculty of Medicine at Lund University in part through the Vascular Wall and Blood and Defense prioritized programs, National Institutes of Health grant GM41659 (to L.M.F.L.-L.), and funds provided from the State of California for medical research on alcohol and substance abuse through the University of California San Francisco (to J.L.W.).
-
ABBREVIATIONS: GPCR, G protein-coupled receptors; B1R, B1 receptor; B2R, B2 receptor; GASP, GPCR-associated sorting protein; HEK, human embryonic kidney; FBS, fetal bovine serum; DMEM, Dulbecco's modified Eagle's medium; GFP, green fluorescent protein; PBS, phosphate-buffered saline; FACS, fluorescence-activated cell sorting; APC, allophycocyanin; TBS, Tris-buffered saline; DAKD, desArg10kallidin; BK, bradykinin; GST, glutathione transferase; BPA, biotinylation protection assay; PNGase, peptide N-glycosidase F.
- Received September 13, 2006.
- Accepted November 16, 2006.
- The American Society for Pharmacology and Experimental Therapeutics