Abstract

Activated G protein-coupled receptors (GPCRs) are phosphorylated and interact with β-arrestins, which mediate desensitization and endocytosis. Endothelin-converting enzyme-1 (ECE-1) degrades neuropeptides in endosomes and can promote recycling. Although endocytosis, dephosphorylation, and recycling are accepted mechanisms of receptor resensitization, a large proportion of desensitized receptors can remain at the cell surface. We investigated whether reactivation of noninternalized, desensitized (phosphorylated) receptors mediates resensitization of the substance P (SP) neurokinin 1 receptor (NK1R). Herein, we report a novel mechanism of resensitization by which protein phosphatase 2A (PP2A) is recruited to dephosphorylate noninternalized NK1R. A desensitizing concentration of SP reduced cell-surface SP binding sites by only 25%, and SP-induced Ca2+ signals were fully resensitized before cell-surface binding sites started to recover, suggesting resensitization of cell-surface-retained NK1R. SP induced association of β-arrestin1 and PP2A with noninternalized NK1R. β-Arrestin1 small interfering RNA knockdown prevented SP-induced association of cell-surface NK1R with PP2A, indicating that β-arrestin1 mediates this interaction. ECE-1 inhibition, by trapping β-arrestin1 in endosomes, also impeded SP-induced association of cell-surface NK1R with PP2A. Resensitization of NK1R signaling required both PP2A and ECE-1 activity. Thus, after stimulation with SP, PP2A interacts with noninternalized NK1R and mediates resensitization. PP2A interaction with NK1R requires β-arrestin1. ECE-1 promotes this process by releasing β-arrestin1 from NK1R in endosomes. These findings represent a novel mechanism of PP2A- and ECE-1-dependent resensitization of GPCRs.

  • neuropeptide
  • substance P
  • G protein-coupled receptor
  • endothelin-converting enzyme-1
  • β-arrestins

desensitization and resensitization of G protein-coupled receptors (GPCRs) are essential for preventing continuous activation and for regaining signaling capability. Substance P (SP) rapidly desensitizes the neurokinin 1 receptor (NK1R), a process involving receptor phosphorylation (34) and internalization (14). G protein-coupled receptor kinases-2 and -3 (GRK-2/-3) mediate SP-induced NK1R phosphorylation (17, 20). GRK-phosphorylated NK1R interacts with β-arrestins (βARRs), which sterically uncouple the receptor from G proteins to mediate homologous desensitization, and couple the receptor to clathrin and AP2 to mediate endocytosis (19). The second messenger-dependent protein kinase C (PKC) also phosphorylates the NK1R (34). PKC phosphorylation is sufficient to impair NK1R coupling to G proteins (11). However, this heterologous desensitization does not require agonist binding and does not induce receptor internalization.1

GPCRs can be classified according to their interaction with βARRs (22). Class “A” receptors [e.g., β2-adrenergic (β2AR), μ-opioid, neurokinin 3 receptors] show a preference for βARR2 over βARR1, interact transiently with βARR2 with low affinity, and rapidly recycle (22, 29). Class “B” receptors [e.g., NK1R, somatostatin 2A (sst2a), calcitonin receptor-like receptors] form high-affinity and sustained interactions with both isoforms of βARR and slowly recycle (22, 23, 33). Class B receptors have multiple S/T residues within the COOH-terminal domains that are potential sites of GRK phosphorylation, which confers high-affinity interactions with βARRs (21). As a class B receptor, the NK1R is sequestered with βARRs within endosomes for prolonged periods (19, 21, 22, 28, 29). We reported that the endosomal peptidase endothelin-converting enzyme-1 (ECE-1) plays a critical role in regulating interactions between NK1R and βARRs in endosomes. By degrading SP in acidified endosomes, ECE-1 promotes disassembly of the NK1R·βARR complex, allowing receptors to recycle and resensitize and βARRs to return to the cytosol (6, 10, 25).

Dephosphorylation is also a critical mechanism of GPCR resensitization. Shortly after stimulation, the phosphorylated β2AR appears in an endosomal vesicle fraction enriched with protein phosphatase type 2A (PP2A) activity (24). PP2A is a cytosolic enzyme that is a member of a diverse family of phospho-S- and phospho-T-specific enzymes ubiquitously expressed in eukaryotic cells (40). Dephosphorylation of the β2AR probably occurs in acidified vesicles, because neutralization with ammonium chloride prevents association of the receptor with PP2A, thereby preventing receptor dephosphorylation (16). βARRs may be critical in recruiting PP2A to GPCRs, since a proteomic-based study identified PP2A as an interaction partner of βARR2 (37). A βARR2·PP2A complex is also a signaling intermediate of the dopamine D2 receptor (3). Although phosphorylation-dependent desensitization and internalization of the NK1R have been thoroughly investigated, nothing is known about the protein phosphatases responsible for NK1R dephosphorylation and resensitization.

We examined the mechanisms of NK1R resensitization and the role of PP2A and ECE-1 in this process. We report the unexpected finding that, following treatment with SP, PP2A interacts with NK1R in a βARR1-dependent manner. PP2A mediates resensitization of NK1R, and ECE-1, by liberating βARR1 from endosomes, enhances this process. Our results represent a novel mechanism of βARR1, PP2A-, and ECE-1-mediated resensitization.

MATERIALS AND METHODS

Reagents.

Sources of most reagents have been described previously (23, 25, 28). Antibodies were from the following sources: monoclonal rat anti-human PP2A, rabbit anti-PP2A, and biotin-labeled goat anti-human ECE-1 from R&D Systems (Wiesbaden, Germany); rabbit anti-βARR1 from Abcam (München, Germany); mouse anti-βARR1 and mouse anti-PP2A catalytic subunit from BD Transduction Laboratories (San Jose, CA); rat high-affinity anti-hemagglutinin 11 (HA11) from Roche Applied Science (Indianapolis, IN); mouse anti-HA11 from Covance (Princeton, NJ); rabbit anti-NK1R 94168 (13). Duolink anti-mouse PLA probe plus, anti-rabbit PLA probe minus, and detection kit 563 were from Olink Bioscience (Uppsala, Sweden). GF 109203X was from AG Scientific (San Diego, CA). Other reagents were from Sigma Aldrich (St. Louis, MO).

cDNAs.

Flag-tagged rat NK1R has been described (35). The Flag epitope does not affect signaling, desensitization, or trafficking of NK1R (35). ECE-1(a-d) and βARR1-enhanced green fluorescent protein (EGFP) have been described (25, 28). Human PP2A-C with an NH2-terminal HA11-tag was a gift from Dr. Petra Knaus (Freie Universität Berlin, Germany).

Cell lines.

Generation and maintenance of human embryonic kidney 293 (HEK) FLP cells (Invitrogen, Carlsbad, CA) and KNRK (sarcoma virus-transformed rat kidney epithelial) cells stably expressing rat NK1R have been described (8, 9, 25). HEK 293 or KNRK cells were transiently transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's guidelines, and cells were studied 48–72 h later.

Small interfering RNA.

The small interfering RNA (siRNA) sequence targeting human βARR1 was 5′-AAAGCCUUCUGCGCGGAGAAU-3′ corresponding to positions 439–459 relative to the start codon, and nonsilencing RNA control was 5′-AAUUCUCCGAACGUGUCACGU-3′ (30). HEK-NK1R cells were transfected with siRNA as described (15).

3H-SP degradation.

HEK-NK1R cells were incubated with [propyl-2,4-3,4(n)-3H]SP [3H-SP, 100,000 counts per minute (cpm), 0.33 ml HBSS-0.1% BSA] and 1 or 10 nM SP (10 min, 37°C). Cells were washed once with HBSS-acetic acid pH 4.75 to remove noninternalized, cell-surface-bound SP and recovered for 0–30 min. The supernatant was collected, acidified by addition of trifluoroacetic acid, and analyzed by HPLC. Cells were lysed in HPLC buffer (water containing 0.1% trifluoroacetic acid) and analyzed by HPLC (23, 25).

3H-SP binding.

HEK-NK1R cells were incubated with SP (1 or 10 nM, 10 min, 37°C, HBSS-BSA), washed, and recovered for 0–60 min. Cells were set on ice and washed with HBSS-acetic acid pH 4.75 to remove noninternalized, cell-surface-bound SP. Surface binding was determined by incubating cells with 100,000 cpm 3H-SP and 0.01 nM SP (0.33 ml, 90 min, 4°C). Nonspecific binding was measured with 100,000 cpm 3H-SP in the presence of 1 μM SP.

Measurement of [Ca2+]i.

[Ca2+]i was measured using Fura-2AM (26). To assess desensitization and resensitization, cells were exposed to SP (1 or 10 nM, 10 min, 37°C) or vehicle, washed, and challenged again with SP (1 or 10 nM).

Immunofluorescence.

HEK-NK1R cells transiently transfected with human HA-PP2A and/or rat βARR1-GFP were incubated in DMEM-0.1% BSA with 10 nM SP (0–10 min, 37°C), washed, and recovered (0–60 min, 37°C). KNRK-NK1R cells transiently transfected with human ECE-1a-d and rat βARR1-GFP were incubated with 10 nM SP (0–10 min, 37°C), washed, and recovered (0–30 min, 37°C). Cells were fixed and incubated with primary antibodies: mouse or rat anti-HA11 (1:500–1,000), rabbit anti-NK1R (1:500), biotinylated goat anti-ECE-1 antibody (1:500). Fluorescent secondary antibodies or streptavidin were used for localization (23).

Confocal microscopy and image analysis.

Cells were observed by confocal microscopy (23). Intensity scans were analyzed using ImageJ (National Institutes of Health, Bethesda, MD). To analyze the subcellular distribution of PP2A and βARR1, plasma membrane fluorescence was calculated by subtracting cytoplasmic pixel counts from total cell pixel counts.

In situ proximity ligation assay.

Duolink in situ proximity ligation assays (PLA) using anti-mouse PLA probe plus, anti-rabbit PLA probe minus, and detection kit 563 were performed according to the manufacturer's instructions (Olink Bioscience, Uppsala, Sweden). Cells were grown on poly-d-lysine-coated 12-mm glass coverslips. After treatments, cells were fixed in 4% paraformaldehyde in PBS for 20 min and permeabilized using 0.1% Triton X-100 in PBS for 4 min. Cells were blocked with 1% normal goat serum in PBS (blocking buffer) for 60 min at room temperature and incubated with primary antibodies in blocking buffer overnight at 4°C. Primary antibodies used were rabbit anti-NK1R antibody 94168 (1:2,000) and either mouse anti-βARR1 (1:1,000) or mouse anti-PP2A (1:50). Unbound antibody was removed by washing with PBS. Cells were incubated in a humidified chamber at 37°C with 40 μl reaction volume per coverslip and dilution ratios recommended by the manufacturer. Cells were incubated with PLA probes diluted in blocking buffer for 120 min and washed with PBS. Cells were then incubated with hybridization mixture for 15 min, washed with Tris-buffered saline containing 0.05% Tween 20 (TBS-T), and incubated with ligation mixture for 15 min. Cells were washed with TBS-T, incubated with amplification mixture for 90 min, washed with TBS-T, and incubated with detection probes for 60 min. Finally, cells were washed with 2× saline-sodium citrate buffer (SSC), 1× SSC, 0.2× SSC, 0.02× SSC, and 70% ethanol. Coverslips were air-dried and mounted on slides with Vectashield containing DAPI (Vector Laboratories, Burlingame, CA). Images of PLA signals were collected at a zoom of 1 using the 543 nm laser, and 5 optical sections were taken at intervals of 0.32 μm. The number of PLA signals per cell was counted using the BlobFinder software (1) (http://www.cb.uu.se/∼amin/BlobFinder/). The cell average analysis mode was used, and errors in delineation of nuclei were manually corrected.

Immunoprecipitation and Western blotting.

To immunoprecipitate cell-surface Flag-NK1R, HEK-NK1R cells were incubated with mouse M2 Flag antibody (1:250, 90 min, 4°C), washed, and lysed in RIPA buffer. To immunoprecipitate PP2A, HEK-NK1R cells were lysed and incubated with anti-PP2A (1:2,000). Lysates were incubated with protein A/G beads, and immune complexes were fractionated by SDS-PAGE (12.5%) (23). Membranes were incubated with anti-PP2A (1:200) or anti-βARR1 (1:1,000), followed by goat anti-rabbit antibody coupled to peroxidase (1:2,000). Flag-NK1R was detected using a biotinylated M2 Flag antibody (1:200) and streptavidin-coupled peroxidase (1:2,000). Antibodies were detected by chemiluminescence.

Inhibitor treatments.

Cells were incubated with SM-19712 (10 μM), fostriecin (300 nM), okadaic acid (10 nM), GF 109203X (GFX, 1 μM), or vehicle.

Statistical analysis.

Data are presented as means ± SE. Differences were assessed by ANOVA and a Student-Newman-Keuls test or Student's t-test. P < 0.05 was considered significant.

RESULTS

NK1R resensitization and recovery of cell-surface SP binding sites.

Resensitization of many GPCRs is thought to occur via a recycling-based mechanism comprising three basic steps: 1) βARR-mediated internalization of the ligand-bound, phosphorylated receptor; 2) release of ligand, dissociation of the receptor from βARRs, and receptor dephosphorylation in endosomes; and 3) recycling of the receptor to the cell surface. We compared the timing of the reappearance of cell-surface 3H-SP binding sites to the resensitization of SP-induced Ca2+ signaling in HEK-NK1R cells, and examined the role of ECE-1 in these processes. HEK-NK1R cells were preincubated with SP (10 nM, 10 min) or vehicle, washed, and recovered in SP-free medium for 0–60 min. Cells were then either acid-washed (to remove surface-bound SP) and incubated with 3H-SP to determine cell-surface SP binding sites, or challenged again with SP (10 nM) to determine the change in [Ca2+]i and thereby assess resensitization. SP induced a rapid loss of surface binding sites (25% reduction, 10 min) that was sustained for 60 min (Fig. 1A), and caused complete desensitization of SP-induced Ca2+ signals that fully resensitized after 60 min (Fig. 1B). The ECE-1 inhibitor SM-19712 did not affect the reappearance of SP binding sites (Fig. 1A) but significantly inhibited resensitization of SP-induced Ca2+ transients at 30 or 60 min (60 min, % resensitization: control, 91 ± 7; SM-19712, 23 ± 9; Fig. 1B). Thus, after incubation with 10 nM SP, the NK1R resensitizes by an ECE-1-dependent mechanism that does not rely on recovery of cell-surface SP binding sites.

Fig. 1.

Endothelin-converting enzyme-1 (ECE-1) contribution to neurokinin 1 receptor (NK1R) recycling and resensitization. A and B: human embryonic kidney (HEK)-NK1R cells were incubated with substance P (SP; 10 nM, 10 min) or vehicle, washed, and recovered (0–60 min). Cells were acid-washed to remove noninternalized SP, and surface binding was determined using [propyl-2,4-3,4(n)-3H]SP (3H-SP) (A). Cells were challenged with 10 nM SP at 0–60 min recovery, and intracellular calcium concentration ([Ca2+]i) was measured (B). Cell-surface 3H-SP-binding sites (A) did not recover, whereas SP-stimulated Ca2+ signals fully resensitized (B). Inhibition of ECE-1 with SM-19712 prevented resensitization but did not affect recycling, following challenge with 10 nM SP. C: HEK-NK1R cells were incubated with SP (1 nM, 10 min) or vehicle before measurement of [Ca2+]i in response to a second challenge with SP (1 nM, no wash between challenges) to assess desensitization. D: HEK-NK1R cells were incubated with SP (1 nM, 10 min) or vehicle, washed, and recovered (5 min) before measurement of [Ca2+]i in response to a second challenge with SP (1 nM) to assess resensitization. Inhibition of ECE-1 with SM-19712 did not affect the desensitization or rapid resensitization following challenge with 1 nM SP. In A, SEs are within symbols. (A, n = 3; B, n = 4 experiments in triplicate; C and D, n = 3 experiments in triplicate.) *P < 0.05.

The intracellular trafficking pathway of the NK1R depends on the concentration of SP (26). A high SP concentration (10 nM) induces translocation of the NK1R and βARRs to perinuclear endosomes and slow recycling to the cell surface. Conversely, a low SP concentration (1 nM) induces translocation of the NK1R and βARRs to endosomes located just beneath the plasma membrane and rapid recycling to the cell surface. We therefore investigated whether ECE-1 also regulates NK1R resensitization following stimulation with a low concentration of SP (1 nM). HEK-NK1R cells were stimulated with SP (1 nM, 10 min) or vehicle (control), then the magnitude of the Ca2+ response to a second challenge with SP (1 nM, no wash between first and second SP challenges) was determined to assess desensitization. Stimulation with 1 nM SP (10 min) caused only ∼50% desensitization of SP-induced Ca2+ signals in both control and SM-19712-treated cells (% unstimulated cells: control, 54 ± 6; SM-19712, 44 ± 3; Fig. 1C). Furthermore, when cells were washed and allowed to recover in SP-free medium between challenges, the SP-induced Ca2+ signals were almost completely resensitized after 5 min in the presence or absence of SM-19712 (5 min, % resensitization; control, 97 ± 4; SM-19712, 88 ± 4; Fig. 1D). Thus, after incubation with 1 nM SP, the NK1R resensitizes rapidly in an ECE-1-independent manner.

The lack of recovery of cell-surface binding sites for 60 min after SP stimulation (Fig. 1A) suggests that internalized NK1R does not recycle within this time period. However, recycling is a dynamic process and the lack of recovery of cell-surface binding sites could also be attributed to rapid recycling and reinternalization of the NK1R, resulting in no change in the overall number of cell-surface binding sites. To investigate this possibility, we examined the effect of SP concentration on the loss and reappearance of cell-surface binding sites and the degradation of cell-associated SP. HEK-NK1R cells were incubated with 1 nM or 10 nM SP (0–10 min), washed and incubated in SP-free medium (0–10 min). Cell-surface 3H-SP binding sites were determined throughout the experiment. SP at 1 nM or 10 nM SP caused a rapid and equivalent loss of cell-surface 3H-SP binding sites (∼30% reduction at 2.5 min; Fig. 2A). In cells incubated with 10 nM SP, binding sites did not recover during the experiment. In contrast, in cells incubated with 1 nM SP, the recovery of cell-surface 3H-SP binding sites was rapid (94 ± 1.7% of control, 10 min recovery, Fig. 2A). Interestingly, in cells stimulated with 1 nM SP, the recovery of cell-surface binding sites was almost complete in the continued presence of ligand (Fig. 2A, 0–10 min). This observation indicates that recycled NK1R did not reinternalize but remained at the cell surface. We then examined the rate of degradation of different concentrations of SP by HEK-NK1R cells. HEK-NK1R cells were incubated with 3H-SP and 1 or 10 nM of unlabeled SP (0–10 min), acid-washed, and incubated in SP-free medium (0–30 min). At various times, cells were lysed and lysates and medium were analyzed by HPLC to assess SP degradation. Internalized 3H-SP was rapidly degraded, and the rate of degradation was the same for both 1 nM and 10 nM SP concentrations (1 and 10 nM SP, <15% intact at 30 min recovery; Fig. 2B). Intracellular metabolites coeluted with SP[1–6] and SP[1–9], with SP[1–6] the major product and SP[1–9] an intermediate (Fig. 2C). SP[1–6] but not SP[1–9] was slowly released into the medium (Fig. 2D).

Fig. 2.

Agonist-induced loss of cell-surface binding sites and degradation of internalized 3H-SP. A: HEK-NK1R cells were incubated with SP (1 or 10 nM, 10 min), washed, and recovered (0–10 min). Cells were acid-washed, and cell-surface 3H-SP-binding sites were determined (n = 3). BD: HEK-NK1R cells were incubated with 3H-SP and 1 or 10 nM unlabeled SP (10 min), acid-washed, and recovered (0–30 min). HPLC analysis of lysed cells (B and C) or medium (D) indicated that internalized SP was rapidly degraded, regardless of concentration. SP[1–6] and SP[1–9] accumulated in cells, and SP[1–6] was released to medium. Representative results are shown for n = 3 experiments. *P < 0.05.

The observations that recycled NK1R does not reinternalize after stimulation with 1 nM SP (Fig. 2A) and that the uptake and degradation kinetics of 1 nM and 10 nM SP are similar (Fig. 2B) suggest that the NK1R also does not reinternalize in cells stimulated with 10 nM SP. Thus, the lack of recovery of cell-surface SP binding sites for 60 min after stimulation with 10 nM SP (Fig. 1A) is probably attributable to intracellular retention of internalized NK1R. The resensitization of SP-induced Ca2+ signals observed at 30–60 min poststimulation with 10 nM SP (Fig. 1B) therefore occurs before receptor recycling to the cell surface. These findings suggest that reactivation of NK1R located at the plasma membrane is the primary mechanism of regaining SP sensitivity.

SP-dependent association of NK1R with PP2A and βARR1.

The protein phosphatases PP2A, PP2Cα, and PP2Cβ are interaction partners of βARRs (37). To examine whether SP induces colocalization of PP2A and βARR1, HEK-NK1R cells expressing HA-PP2A and βARR1-GFP were stimulated with SP (10 nM, 0–10 min), and PP2A and βARR1 were localized by confocal microscopy. In unstimulated cells, PP2A and βARR1 were similarly distributed throughout the cytosol (arrows) and NK1R was at the plasma membrane (arrowheads) (Figs. 3A and 5A). SP (2 or 10 min) caused translocation of both PP2A and βARR1 to the plasma membrane to colocalize with NK1R (Figs. 3A and 5A, arrowheads). We quantified changes in the subcellular localization of βARR1 and PP2A by determining the proportion of the total cellular pixels that were present at the plasma membrane. The percentage of βARR1-positive pixels at the plasma membrane increased from 19 ± 2% at 0 min to 48 ± 4% at 10 min (Fig. 3B). Similarly, the percentage of PP2A-positive pixels was 20 ± 2% at 0 min and increased to 29 ± 1% at 10 min (Fig. 3B). To determine whether PP2A and βARR1 interact, we immunoprecipitated PP2A and probed blots for βARR1. In unstimulated cells, PP2A was associated with βARR1, and SP (10 nM, 10 min) enhanced this association by 117% (Fig. 3C). Thus, SP induces membrane translocation and interaction of PP2A and βARR1.

Fig. 3.

SP-induced association of NK1R, β-arrestin-1 (βARR1), and protein phosphatase 2A (PP2A). A and B: HEK-NK1R cells transfected with βARR1-green fluorescent protein (GFP) and hemagglutinin (HA)-PP2A were incubated with SP (10 nM, 0–10 min). SP induced redistribution of βARR1-GFP and HA-PP2A from the cytosol (arrows) to the plasma membrane (arrowheads) (A). Quantification indicated that SP significantly increased plasma membrane βARR1 and PP2A (B). C: HEK-NK1R cells were incubated with SP. Endogenous PP2A was immunoprecipitated (IP) and Western blotted (WB) for endogenous βARR1. D and E: HEK-NK1R cells were incubated with SP (10 nM, 0–10 min), and association of NK1R with endogenous βARR1 or PP2A was determined by proximity ligation assay (PLA). The total number of PLA signals (red) was divided by the number of nuclei (blue) to calculate PLA signals per cell. SP significantly increased PLA signals for NK1R with βARR1 and PP2A. F and G: controls for PLA. Cells were incubated with SP (10 nM, 10 min), and association of NK1R with endogenous βARR1 or PP2A was determined by PLA. In HEK-NK1R cells (F), there were no detected PLA signals when primary antibodies were omitted. In HEK-FLP cells lacking NK1R (G), there were no PLA signals for NK1R and βARR1 (top) nor NK1R and PP2A (bottom). Scale, 10 μm. (n = 3–5 experiments, >10 images analyzed per experiment.) *P < 0.05.

To examine SP-induced interactions between NK1R and PP2A at the single cell level, we used an in situ PLA, which enables localization of protein-protein interactions at the cellular level (31). If two proteins are in close proximity (<40 nm), short DNA strands attached to secondary antibodies are ligated, and the DNA can be amplified by rolling circle amplification and detected using fluorescent oligonucleotides. Since NK1R is known to interact with βARRs (28), we assessed the utility of the PLA by studying SP-induced interaction between NK1R and endogenous βARR1 in HEK-NK1R cells. In unstimulated HEK-NK1R cells, there were few PLA signals per cell, indicating low basal interaction between NK1R and βARR1 (Fig. 3D, arrows). SP (10 nM, 10 min) increased the number of PLA signals per cell by 1.8 ± 0.1-fold (Fig. 3D). We similarly examined interactions between NK1R and endogenous PP2A. In unstimulated cells, there were few PLA signals, and SP increased PLA signals by 1.7 ± 0.2-fold (Fig. 3E). Thus, SP induces interaction of NK1R with βARR1 and PP2A. PLA signals were not detected when primary antibodies were omitted (Fig. 3F) or in HEK cells lacking NK1R (Fig. 3G), confirming specificity.

ECE-1 regulation of association of βARR1 and PP2A with cell-surface NK1R.

By selectively immunoprecipitating the cell-surface NK1R, we determined whether βARR1 and PP2A associate with NK1R that remains at the plasma membrane after stimulation with SP. HEK-NK1R cells were stimulated with SP (10 nM, 10 min) or vehicle, washed, and recovered in SP-free medium (30 min). Noninternalized NK1R was immunoprecipitated by incubating intact, nonpermeabilized cells with an antibody to the extracellular NH2-terminal Flag-epitope (Fig. 4A), and blots were probed for endogenous βARR1 and PP2A. In unstimulated cells, low levels of βARR1 and PP2A were associated with noninternalized NK1R (Fig. 4, B and C), consistent with the PLA (Fig. 3, D and E). After stimulation with SP for 10 min and recovery in SP-free medium for 30 min, NK1R·βARR1 association increased by approximately sevenfold (Fig. 4B, black bars) and NK1R·PP2A association increased by approximately fourfold (Fig. 4C, black bars). The ECE-1 inhibitor SM-19712 suppressed or abolished these interactions (Fig. 4, B and C, red bars). We also used immunofluorescence and confocal microscopy to investigate the role of ECE-1 in the association of cell-surface NK1R with βARR1-GFP and HA-PP2A. In HEK-NK1R cells stimulated with 10 nM SP for 10 min followed by recovery in SP-free medium for 60 min, βARR1 and PP2A were colocalized with the NK1R at the plasma membrane (Fig. 5B, arrowheads). The ECE-1 inhibitor SM-19712 prevented this colocalization by causing βARR1 retention in endosomes (Fig. 5B, arrows), as described (25). Thus, SP induces association of βARR1 and PP2A with noninternalized cell-surface NK1R, and ECE-1 promotes this association.

Fig. 4.

ECE-1-dependent interactions of cell-surface NK1R with βARR1 and PP2A. A. Schematic showing method used to specifically immunoprecipitate only cell-surface NK1R. Intact cells were incubated with antibody to extracellular Flag at 4°C, before cell lysis and isolation of NK1R immune complexes using protein A/G beads. BD: HEK-NK1R cells were incubated with SP (10 nM, 10 min) or vehicle, washed, and recovered. Cell-surface NK1R was immunoprecipitated and Western blotted, and blots were probed for endogenous βARR1 or PP2A, or for NK1R. Stimulation with SP significantly increased interaction of surface NK1R with βARR1 (B) and PP2A (C). Inhibition of ECE-1 with SM-19712 inhibited these interactions. βARR1 small interfering RNA (siRNA) knockdown (D) abolished interactions. (n = 3 experiments.) E: siRNA to βARR1 reduced expression of βARR1 in HEK-NK1R cells, determined by Western blotting compared with control siRNA. *P < 0.05.

Fig. 5.

SP-induced trafficking of βARR1, PP2A, and NK1R. HEK-NK1R cells expressing βARR1-GFP and HA-PP2A were incubated with SP (10 nM, 0–10 min), washed, and recovered for 10–60 min. A: in unstimulated cells, βARR1 and PP2A were similarly distributed in the cytosol (arrows) and NK1R was at the plasma membrane (arrowheads). After 10 min with SP, βARR1 and PP2A were detected with NK1R at the plasma membrane (arrowheads). B: after 10 min recovery, βARR1 colocalized with NK1R in endosomes (arrows), although some NK1R remained at the plasma membrane. After 60 min recovery, βARR1, PP2A, and NK1R were detected at the plasma membrane (arrowheads). Inhibition of ECE-1 with SM-19712 caused retention of βARR1 and NK1R in endosomes (arrows), preventing accumulation of βARR1 and PP2A at the plasma membrane. Scale, 10 μm.

βARR1-dependent PP2A·NK1R interaction.

Since PP2A interacts with βARR1 (Fig. 3C), we determined, using immunoprecipitation of noninternalized NK1R and Western blotting (Fig. 4A), whether βARR1 mediates the SP-induced PP2A·cell-surface NK1R interaction. To do so, we disrupted βARR1 by siRNA knockdown, which achieved a ∼65% reduction of βARR1 levels compared with control cells (βARR2 levels were not tested; Fig. 4E). βARR1 knockdown strongly inhibited SP-induced association of cell-surface NK1R with βARR1 and PP2A (Fig. 4D). Thus, βARR1 mediates SP-induced interaction of noninternalized NK1R with PP2A.

PP2A-mediated resensitization of NK1R signaling.

After stimulation with SP, PP2A associates with noninternalized NK1R at the cell surface, where PP2A may dephosphorylate receptors and mediate resensitization. Therefore, we determined whether inhibition of PP2A with the selective inhibitor fostriecin (3) or with okadaic acid prevents the resensitization of NK1R signaling. The inhibitor concentrations used (fostriecin, 300 nM; okadaic acid, 10 nM) were 100-fold the IC50 values of the inhibitors (7, 36). HEK-NK1R cells preincubated with fostriecin, okadaic acid, or vehicle (control) were stimulated with SP (10 nM, 0–10 min), washed, and incubated in SP-free medium (0–60 min). The magnitude of the Ca2+ response to a second challenge with SP (10 nM) was determined to assess resensitization. Fostriecin and okadaic acid both inhibited responses to the second SP challenge at 30 and 60 min (% resensitization, 30 min: control 46 ± 3; fostriecin 29 ± 5; okadaic acid 28 ± 4; Fig. 6). Thus, the phosphatase activity of PP2A is required for the resensitization of NK1R signaling.

Fig. 6.

Role of PP2A in NK1R resensitization. HEK-NK1R cells were incubated with SP (10 nM, 10 min) or vehicle, washed, and challenged with SP (10 nM) at 30 or 60 min recovery. PP2A inhibitors fostriecin and okadaic acid inhibited resensitization compared with vehicle. (n = 4–8 experiments in triplicate.) *P < 0.05.

ECE-1-mediated translocation of βARR1 from endosomes.

Activation of class B GPCRs (e.g., NK1R) causes sustained sequestration of βARRs with receptors in endosomes (28). To examine the effect of ECE-1 expression on localization of βARRs, KNRK-NK1R cells [which lack endogenous ECE-1 (25)] expressing βARR1-GFP and ECE-1a-d or empty vector (control) were stimulated with SP (10 nM, 10 min), washed, and recovered in SP-free medium (0–30 min) (Figs. 7 and 8). ECE-1 and βARR1-GFP were localized by immunofluorescence and confocal microscopy. In cells lacking ECE-1, SP induced translocation of βARR1-GFP to endosomes after 10 min, and βARR1-GFP remained sequestered in endosomes for at least 30 min (Fig. 7A). In cells expressing ECE-1a-d, βARR1-GFP was detected in endosomes and at the plasma membrane at 10 and 30 min (Fig. 7B, ECE-1a, and Fig. 8, ECE-1a-d). Fluorescence intensity scans revealed an approximately fivefold increase in the intensity of βARR1-GFP at the plasma membrane at 10 and 30 min recovery in ECE-1a-d cells compared with cells lacking ECE-1 (Fig. 7, C and D). Conversely, ECE-1 inhibition prevented membrane translocation of βARR1-GFP at 60 min recovery in HEK-NK1R cells (Fig. 5B). Thus, ECE-1, by degrading SP in endosomes, liberates βARR1 from endosomally sequestered NK1R and allows βARR1 to translocate to the plasma membrane, where it may promote PP2A interaction with the NK1R.

Fig. 7.

ECE-1-dependent translocation of βARR1 from endosomes. KNRK-NK1R cells were transfected with βARR1-GFP and ECE-1a or vector. Cells were stimulated with SP (10 nM, 0–10 min), washed, and recovered. A: in cells lacking ECE-1, SP induced sequestration of βARR1 in endosomes. B: in cells expressing ECE-1a, βARR1 was sequestered in endosomes at 10 min (arrows) but was also at the plasma membrane at 30 min recovery (arrowheads). C: line scans of βARR1-GFP intensity (B) illustrate translocation of βARR1 from the cytosol (0 min) to the plasma membrane (30 min recovery). D: quantification shows that overexpression of ECE-1a-d enhances plasma membrane levels of βARR1 at 10 and 30 min recovery. Scale, 10 μm. (n = 3 experiments, >10 cells scanned per treatment.) *P < 0.05.

Fig. 8.

Effect of ECE-1 expression on βARR1 translocation. KNRK-NK1R cells were transfected with βARR1-GFP and ECE-1a-d or vector. Cells were stimulated with SP (10 nM, 0–10 min), washed, and recovered for 30 min. In cells expressing ECE-1a, -b, -c, or -d, βARR1 was sequestered in endosomes (arrows) but was also at the plasma membrane (arrowheads). In contrast, in cells lacking ECE-1, βARR1 was only detected in endosomes at all time points after stimulation with SP (see Fig. 7A). Scale, 10 μm.

PKC-dependent regulation of βARR and NK1R signaling.

The second messenger-dependent PKC can phosphorylate the NK1R to regulate SP-induced Ca2+ mobilization (11, 34). Furthermore, βARRs can interact with PKC-phosphorylated GPCRs (18). Thus, PKC-mediated phosphorylation of the NK1R may regulate βARR recruitment to the plasma membrane and desensitization/resensitization of SP-induced Ca2+ signals. We first examined the effect of the PKC inhibitor GFX on SP-induced βARR1 recruitment to the plasma membrane (Fig. 9A). The concentration of GFX used in experiments (1 μM) was 100-fold the IC50 value of the inhibitor (32). In unstimulated cells treated with vehicle or GFX, βARR1 was distributed throughout the cytosol (Fig. 9A, arrows). SP (10 nM, 2 min) caused βARR1 translocation to the plasma membrane, where it colocalized with the NK1R, in both control and GFX-treated cells (Fig. 9A, arrowheads). Thus, PKC inhibition does not prevent SP-induced translocation of βARR1 to the plasma membrane, most likely because NK1R phosphorylation by GRKs can also mediate receptor interaction with βARRs (17, 20).

Fig. 9.

Role of PKC in βARR1 trafficking and NK1R desensitization/resensitization. A: HEK-NK1R cells transfected with βARR1-GFP were preincubated with the PKC inhibitor GF 109203X (GFX) or vehicle (control) before stimulation with SP (10 nM, 0–2 min). SP induced redistribution of βARR1-GFP from the cytosol (arrows) to the plasma membrane (arrowheads) in both vehicle- and GFX-treated cells. Scale, 10 μm. B and C: cells were preincubated with GFX or vehicle (control) and challenged with SP (10 nM), and [Ca2+]i was measured. D and E: HEK-NK1R cells preincubated with GFX or vehicle (control) were stimulated with SP (10 nM, 10 min) or vehicle, and washed. Then [Ca2+]i in response to a second challenge with SP (10 nM) was measured either immediately (D) or following recovery in SP-free medium (60 min) (E). Inhibition of PKC with GFX caused a greater and more sustained Ca2+ response to a single challenge with SP and suppressed resensitization of NK1R signaling. (B, average of triplicate observations, representative of n = 3 experiments; C and D, n = 3 experiments in triplicate, E, n = 5 experiments in triplicate.) *P < 0.05.

We then determined the effect of PKC inhibition on desensitization of NK1R signaling. The Ca2+ response to a single challenge with SP (10 nM) was greater and more sustained in cells treated with the PKC inhibitor GFX than in control cells (Fig. 9B). Inhibition of PKC increased the magnitude of the Ca2+ response by 120 ± 3% (Fig. 9C), indicating that PKC phosphorylation attenuates the Ca2+ response to a single challenge with SP. We then examined the role of PKC in desensitization to repeated challenges with SP. HEK-NK1R cells were incubated with SP (10 nM, 10 min) or vehicle, washed, and then challenged immediately with a second dose of SP (10 nM). The magnitude of the Ca2+ response to the second challenge with SP was determined to assess desensitization. In both vehicle- and GFX-treated cells, 10 nM SP stimulation for 10 min caused complete desensitization of SP-induced Ca2+ signals (% unstimulated cells: vehicle-treated cells, 4.4 ± 0.5; GFX-treated cells, 3.2 ± 0.4; Fig. 9D). Together, these results indicate that PKC phosphorylation contributes to attenuation of NK1R Ca2+ signaling in response to a single challenge with SP (Fig. 9, B and C) but does not contribute to desensitization of Ca2+ responses to repeated challenges with SP (Fig. 9D).

Finally, we examined the effect of PKC inhibition on the resensitization of SP-induced Ca2+ signals. HEK-NK1R cells were stimulated with SP (10 nM, 10 min) or vehicle, washed, and incubated in SP-free medium (0–60 min). The magnitude of the Ca2+ response to a second challenge with SP (10 nM) was determined to assess resensitization. The PKC inhibitor GFX significantly inhibited resensitization of SP-induced Ca2+ signals at 60 min (60 min, % resensitization: control, 54 ± 5; GFX, 39 ± 5; Fig. 9E). Thus, PKC phosphorylation is required for resensitization of NK1R signaling.

DISCUSSION

Our results provide evidence for a novel mechanism of resensitization of the NK1R. After stimulation with SP, a fraction of activated, GRK-phosphorylated and desensitized NK1R internalizes as a SP·NK1R·βARR1 complex, leaving most phosphorylated, desensitized NK1R at the plasma membrane. PP2A translocates to noninternalized NK1R, where PP2A may dephosphorylate and resensitize the receptor. ECE-1 promotes the interaction between PP2A and noninternalized NK1R to enhance resensitization, possibly by degrading SP in endosomes and liberating endosomally sequestered βARR1, which mediates interactions between noninternalized NK1R and PP2A. These findings represent a novel mechanism of GPCR resensitization.

Stimulation with a high concentration of SP (10 nM), which fully desensitized the NK1R, reduced cell-surface SP binding sites by ∼25%, consistent with surface retention of immunoreactive NK1R (14). Thus, in SP-treated cells, most NK1R remains at the cell surface and is desensitized. Since homologous desensitization mediated by GRK-dependent phosphorylation leads to βARR-dependent endocytosis of the NK1R (19, 20), noninternalized receptors may desensitize by a heterologous mechanism involving PKC-mediated NK1R phosphorylation (11, 34). SP-induced Ca2+ signaling was fully resensitized before cell-surface SP binding sites started to recover, indicating the necessity of reactivation of NK1R located at the plasma membrane for resensitization. Furthermore, in cells treated with a low concentration of SP (1 nM), NK1R rapidly recycled but did not reinternalize despite the availability of ligand, indicating that recycled NK1R returned to the plasma membrane in an inactive form. This observation suggests that recycled NK1R requires reactivation at the plasma membrane similar to noninternalized receptors. Bennett et al. (4) have reported similar findings in Chinese hamster ovary (CHO) cells stably expressing the NK1R. They demonstrated that resensitization occurs before NK1R recycling using measurements of SP-induced Ca2+ elevations, and radioactive and fluorescent ligand binding techniques. Furthermore, in cells treated with concanavalin A to inhibit receptor internalization, resensitization of NK1R responsiveness to SP still occurred, albeit to a lesser extent than in control cells. In agreement with our results, this study therefore indicates that NK1R localized at the cell surface can be reactivated and that resensitization occurs before NK1R recycling.

Our results suggest that dephosphorylation of surface-retained NK1Rs by PP2A is a major mechanism of resensitization. PP2A can dephosphorylate endocytosed GPCRs, such as the β2AR, to promote their rapid resensitization (16, 24, 38). However, we observed that SP causes PP2A trafficking to the plasma membrane. Using confocal microscopy, PLA, and coimmunoprecipitation of cell-surface NK1R, we found that βARR1 and PP2A interact with NK1R that is retained at the plasma membrane. SP stimulated these interactions, whereas an ECE-1 inhibitor impeded interactions, indicating a requirement for active ECE-1. The interaction of NK1R with PP2A depends on βARR1, since it was suppressed by βARR1 depletion. Inhibition of PP2A or ECE-1 suppressed resensitization of SP-induced Ca2+ signals, confirming the role of PP2A and ECE-1 in NK1R resensitization. However, the effects of the PP2A inhibitors fostriecin and okadaic acid on NK1R resensitization were quite modest. Because the inhibitors were used at concentrations that were 100-fold their IC50 values, it is unlikely that the modest effects of the inhibitors were due to incomplete inhibition of PP2A. Thus, additional mechanisms may also contribute to NK1R resensitization. Since GPCRs can signal via G proteins in endosomes (5), one possibility is that internalized SP could activate endosomally retained NK1R. However, this mechanism is unlikely to mediate resensitization after repeated challenge with SP because this peptide is unable to internalize without receptor binding (14). Notably, PP2A and ECE-1 inhibition suppressed NK1R resensitization at times when there was no recovery of cell-surface SP binding sites. Thus, our results suggest that SP induces interaction of PP2A with NK1R at the plasma membrane in a βARR1-dependent manner to promote dephosphorylation of noninternalized NK1R, which mediates resensitization.

We hypothesize that ECE-1 contributes to resensitization by degrading SP in endosomes to liberate endosomally sequestered βARR1. We recently reported that ECE-1 frees βARR1 from internalized NK1R (25). Surprisingly, we observed that liberated βARR1 was localized at the plasma membrane where it interacted with the NK1R, determined by coimmunoprecipitation with cell-surface NK1R. The crucial role of βARR1 in desensitization and endocytosis of GPCRs is well understood (39), and ECE-1-mediated dissociation of internalized receptors from βARR1 can allow for recycling (25). Less is known about the function of βARR1 in resensitization, and the precise role of βARR1 located at the plasma membrane after receptor endocytosis is unknown. Since the appearance of βARR1 at the plasma membrane coincides with the time course of NK1R resensitization, it is possible that βARR1-associated proteins, such as PP2A, mediate reactivation of NK1R at the cell surface. In this mechanism, resensitization of cell-surface-retained NK1R would depend to some extent on receptor internalization/trafficking to endosomes because ECE-1-mediated degradation of SP in endosomes would be required to liberate βARR1. Interestingly, the study by Bennett et al. (4) showed that, although resensitization of the NK1R occurs before receptor recycling, receptor internalization and endosomal acidification are required for complete resensitization. These results fit with our hypothesis that ECE-1 contributes to resensitization by liberating endosomally sequestered βARR1, since disruption of SP·NK1R·βARR1 internalization or prevention of endosomal acidification would impede this process and thereby inhibit resensitization (25). Thus, although resensitization occurs before recycling of internalized NK1R back to the cell surface, it is still dependent on mechanisms associated with the recycling process.

Further studies are required to define the mechanism by which βARR1 returns to the plasma membrane. One possibility is that βARR1, freed from NK1R in endosomes by the action of ECE-1, interacts with noninternalized phosphorylated NK1R at the cell surface. Noninternalized receptors are likely desensitized by a heterologous mechanism involving PKC-mediated NK1R phosphorylation (11, 34) and βARRs are capable of binding to PKC-phosphorylated GPCRs (18). We observed that PKC phosphorylation attenuates Ca2+ responses to a single challenge with SP. Furthermore, inhibition of PKC suppressed resensitization of SP-induced Ca2+ signaling, indicating that PKC phosphorylation promotes resensitization of the NK1R. Our interpretation of these results is that PKC phosphorylates the NK1R in response to SP and that these PKC phosphorylation sites may recruit βARRs to noninternalized NK1R to mediate resensitization. However, further studies are required to confirm this suggestion.

We also demonstrated that βARR1 mediates the SP-induced association of PP2A with noninternalized NK1R, since βARR1 depletion diminishes this interaction. However, it is possible that other proteins also regulate the interaction of PP2A with cell-surface NK1R, such as the A-kinase anchoring proteins (AKAPs) that regulate signaling of other GPCRs and are capable of binding phosphatases (2). Additional studies are required to examine the stoichiometry of NK1R associations with βARRs and PP2A, and to determine whether these proteins interact directly or with other intermediate binding proteins.

A limitation of our experiments is that we were unable to directly determine whether βARR1 is necessary for NK1R resensitization. A restriction of using βARR1 knockdown to study the contribution of βARR1 to PP2A-mediated NK1R resensitization is that this treatment amplifies signaling by inhibiting NK1R desensitization and endocytosis (19). These effects alone can influence the kinetics of NK1R resensitization, complicating interpretation of the contribution of βARR1 to PP2A-mediated mechanisms. Furthermore, since βARR2 is also a binding partner of PP2A (37), it is possible that βARR2 could also contribute to PP2A-mediated NK1R resensitization, but we did not address this possibility in our current study. Another limitation of this study is that we could not unequivocally demonstrate that resensitization was mediated by noninternalized receptors, which would necessitate studies of noninternalizing mutated receptors. Although internalization-defective NK1R mutants have been described (11), these mutants lack phosphorylation sites that could be regulated by PP2A, which would preclude examination of the role of PP2A in resensitization. We did not examine the effects of pharmacological inhibitors of receptor endocytosis or recycling on resensitization, since these treatments can block βARR recruitment and ECE-1 activity, which contribute to the mechanism that we describe. It will also be important to confirm that PP2A controls NK1R resensitization in cells that naturally express NK1R, although ECE-1 controls resensitization of endogenous NK1R in endothelial cells (6).

This mechanism of PP2A-mediated resensitization of noninternalized NK1R may be applicable to other GPCRs. βARR1 is present at the plasma membrane after stimulation of cells expressing the class B sst2a receptor with the ECE-1 substrate somatostatin-14 (27). In contrast, stimulation with octreotide, an ECE-1-resistant agonist, did not allow the release of βARR1-GFP from the internalized receptor (27). These data strengthen the notion that ECE-1-mediated degradation is essential for the release of βARR1. Inhibition of agonist-induced internalization does not impede resensitization of the dopamine receptor 1 (12), supporting the concept of resensitization mediated by reactivation of cell-surface receptors. Thus, PP2A may mediate resensitization of GPCRs through dephosphorylation of receptors at the plasma membrane, including the NK1R.

GRANTS

This work was supported by Interdisziplinares Zentrum fur Klinische Forschung (IZKF) Münster (STEI2/076/06), SFB293 (A14), DFG STE 1014 2-2 (to M. Steinhoff), University of Münster (to D. Roosterman), DFG (Br1589/8-1), IZKF (Bra1/001/08) (to E. Brand), British Heart Foundation (BHF) FS/08/017/25027 (to G. S. Cottrell), and National Institutes of Health (DK39957, DK43207, and DK57840) (to N. W. Bunnett).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

Footnotes

  • * J. E. Murphy and D. Roosterman are joint first authors; N. W. Bunnett and M. Steinhoff are joint senior authors.

  • 1 This article is the topic of an Editorial Focus by Enrique Rozengurt (27a).

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 27a.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
  34. 33.
  35. 34.
  36. 35.
  37. 36.
  38. 37.
  39. 38.
  40. 39.
  41. 40.
View Abstract