Identification of Phosphorylation Sites Regulating sst₃ Somatostatin Receptor Trafficking

Abstract

The human somatostatin receptor 3 (sst₃) is expressed in about 50% of all neuroendocrine tumors and hence a promising target for multireceptor somatostatin analogs. The sst₃ receptor is unique among ssts in that it exhibits a very long intracellular C-terminal tail containing a huge number of potential phosphate acceptor sites. Consequently, our knowledge about the functional role of the C-terminal tail in sst₃ receptor regulation is very limited. Here, we have generated a series of phosphorylation-deficient mutants that enabled us to determine crucial sites for its agonist-induced β-arrestin mobilization, internalization, and down-regulation. Based on this information, we generated phosphosite-specific antibodies for C-terminal Ser³³⁷/Thr³⁴¹, Thr³⁴⁸, and Ser³⁶¹ that enabled us to investigate the temporal patterns of sst₃ phosphorylation and dephosphorylation. We found that the endogenous ligand somatostatin induced a rapid and robust phosphorylation that was completely blocked by the sst₃ antagonist NVP-ACQ090. The stable somatostatin analogs pasireotide and octreotide promoted clearly less phosphorylation compared with somatostatin. We also show that sst₃ phosphorylation occurred within seconds to minutes, whereas dephosphorylation of the sst₃ receptor occurred at a considerable slower rate. In addition, we also identified G protein-coupled receptor kinases 2 and 3 and protein phosphatase 1α and 1β as key regulators of sst₃ phosphorylation and dephosphorylation, respectively. Thus, we here define the C-terminal phosphorylation motif of the human sst₃ receptor that regulates its agonist-promoted phosphorylation, β-arrestin recruitment, and internalization of this clinically relevant receptor.

Somatostatin-14 (SS-14) is a cyclic peptide that regulates many physiological functions, including the secretion of hormones such as GH, TSH, ACTH, insulin, and glucagon (1). SS-14 is the natural ligand of a family of 5 human G protein-coupled receptors (GPCRs) named somatostatin receptor (sst)₁–sst₅ (2–4). Because of its short half-life (1–3 min) in human plasma, the clinical utility of this peptide is limited. Therefore, metabolically stable somatostatin analogs have been developed (5–8). In clinical practice, octreotide and lanreotide are used as first-choice medical treatment of neuroendocrine tumors such as GH-secreting adenomas and carcinoids (6, 9). Octreotide and lanreotide bind with high subnanomolar affinity to sst₂ only, have moderate affinity to sst₃ and sst₅ and show very low or absent binding to sst₁ and sst₄ (10). More recently, the novel multireceptor somatostatin analog, pasireotide (formerly known as SOM230), has been synthesized (7). Pasireotide is a cyclohexapeptide, which binds with high affinity to all ssts except to sst₄ (8). Pasireotide has been approved for the treatment of Cushing syndrome and more recently for the treatment of acromegaly (6, 11, 12).

We have recently used phosphosite-specific antibodies to examine agonist-induced phosphorylation of the sst₂ and the sst₅ (13–16). For the sst₂ receptor, we found that SS-14 promotes the phosphorylation of at least 6 C-terminal serine and threonine residues of the sst2 receptor namely, S341, S343, T353, T354, T356, and T359 (13, 17, 18). This phosphorylation is mediated by GPCR kinase (GRK)2 and GRK3 and followed by rapid cointernalization of the receptor and β-arrestin into the same endocytic vesicles (13, 19). Dephosphorylation of sst₂ is initiated directly after receptor activation at or near the plasma membrane and is mediated by protein phosphatase 1 (PP1)β (20).

Although there are many studies examining the expression and signaling of sst₂ and sst₅, there is little knowledge about the functional role of sst₃ in human tumors. More recent studies suggest an elevated expression in diverse neuroendocrine-related malignancies such as pancreatic tumors, pheochromocytomas, paragangliomas, gonadotroph adenomas, and lung carcinoids (21–24). Albeit the growing interest in tumoral sst₃ receptors, there is still a lack of knowledge about sst₃ receptor regulation and signaling. Interestingly, among the sst subtypes only sst₂ and sst₃ have been suggested to promote apoptosis of tumor cells. Although the clinically value of somatostatin analogs is primarily based on potent antisecretory effects, other positive effects of somatostatin analogs such as tumor shrinkage are poorly understood but could be related in part to induction of tumor cell apoptosis.

In contrast to sst₂ and sst₅, our knowledge about the functional role of C-terminal phosphorylation of the human sst₃ receptor is limited. For the rat sst₃, it has been reported that agonist-dependent internalization relies critically on the presence of 4 C-terminal hydroxyl amino acids namely Ser³⁴¹, Ser³⁴⁶, Ser³⁵¹, and Thr³⁵⁷ (25). However, the C-terminal regions of the human and the rat sst₃ receptor exhibit strikingly different sequences. Nevertheless, it is believed that the presence of the intact C terminus is required for triggering SS-14-induced apoptosis via the sst₃ receptor (26).

In the present study, we have examined the primary structure of the human sst₃ C-terminal tail. In fact, this 102 amino acids long sequence contains 18 potential phosphate acceptor sites including 12 serine and 6 threonine residues. We have constructed a series of phosphorylation-deficient mutants and generated phosphosite-specific antibodies, which enabled us to provide direct evidence for agonist-selective phosphorylation of the human sst₃ receptor. Using these antibodies, we identified kinases and phosphatases required for agonist-dependent phosphorylation and dephosphorylation.

Materials and Methods

Antibodies and reagents

The human hemagglutinin (HA)-tagged sst₃ receptor as well as the 317–370 S/T-A, 4 S/T-A, 371–418 S/T-A, and K-R mutants were generated via artificial gene synthesis and cloned into pcDNA3.1 by imaGenes. Phosphosite-specific antibodies were custom generated by Genosphere Biotechnologies. The phosphosite-specific antibodies for the S337/T341-phosphorylated form of sst₃ were generated against the following sequence that contained a phosphorylated serine and threonine residue: CRRVR(pS)QEP(pT)VGPP. This sequence corresponds to 333–345 of the human sst₃. Phosphosite-specific antibodies for the T348-phosphorylated form of sst₃ were generated against the following sequence that contained a phosphorylated threonine residue: CGPPEK(pT)EEED. This sequence corresponds to 343–352 of the human sst₃. Phosphosite-specific antibodies for the S361-phosphorylated form of sst₃ were generated against the following sequence that contained a phosphorylated serine residue: CDGEE(pS)REGG. This sequence corresponds to 357–365 of the human sst₃. The specificity of the antisera was initially tested using dot-blot analysis. For subsequent analysis, antibodies were affinity purified against their immunizing peptide as well as against the nonphosphorylated peptide. The phosphorylation-independent rabbit monoclonal anti-sst₃ antibody UMB-5 and the rabbit polyclonal anti-HA antibody were generated and extensively characterized as previously described (17, 27, 28). Anti-GRK2 (sc-562), anti-GRK3 (sc-563), anti-GRK5 (sc-565), anti-GRK6 (sc-566), anti-PP1α (sc-6104), anti-PP1β (sc-6106), and anti-PP1γ (sc-6108) antibodies were obtained from Santa Cruz Biotechnology. SS-14 was obtained from Bachem. Phorbol 12-myristate 13-acetate (PMA) and forskolin were obtained from Sigma-Aldrich. Octreotide, pasireotide, and the sst₃ antagonist, NVP-ACQ090 (29), were kindly provided by Dr Herbert Schmid (Novartis). Somatoprim was obtained from DeveloGen. L-796/778 (30), a selective ligand for sst₃, was provided by Merck (please see Supplemental Table 1).

Cell culture and transfection

Human embryonic kidney (HEK)293 cells were obtained from the German Resource Centre for Biological Material (DSMZ). AtT20-D16v-F2 cells were obtained from American Type Tissue Culture Collection. HEK293 cells and AtT-20 cells were grown in DMEM supplemented with 10% fetal calf serum. Cells were transfected with plasmids encoding for HA-tagged wild-type or mutant sst₃ receptors using Lipofectamine 2000 according to the instructions of the manufacturer (Invitrogen). Stable transfectants were selected in the presence of 400-μg/mL G418. To increase the total number of AtT20 cells stably expressing sst₃ or phosphorylation-deficient mutant receptors, we used fluorescence-activated cell sorting (FACS). Trypsinized cells were washed with PBS and transferred into opti-MEM containing an A488-labeled anti-HA antibody at a dilution of 1:1000 (Sigma-Aldrich). After 30 minutes of preincubation at room temperature, cells were centrifuged and the cell pellet was resuspended in FACS buffer (2mM EDTA and 0.5% BSA in PBS). FACS was executed using a BD FACSAria III cell sorter. Approximately 1% of the positive cell population was sorted at an average purity of 85%. Sorted cells were then recultivated. To ensure similar expression levels of wild-type and mutant receptors, stable cells were characterized using radioligand-binding assays, Western blot analysis, surface ELISA, and immunocytochemistry as described previously (19, 31). The level of sst expression was between 800 and 1200 fmol/mg membrane protein for all experiments using stably transfected and between 1500 and 2000 fmol/mg membrane protein for all experiments using transiently transfected HEK293 cells (19, 31) (please see Supplemental Figure 1).

Western blot analysis

HEK293 cells or AtT-20 cells stably expressing HA-tagged human sst₃ receptors were plated onto 60-mm dishes and grown to 80% confluence. After indicated treatment with SS-14, L-796/778, octreotide, pasireotide, somatoprim, PMA, or forskolin, cells were lysed in detergent buffer (50mM Tris-HCl, [pH 7.4], 150mM NaCl, 5mM EDTA, 10mM NaF, 10mM disodium pyrophosphate, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS) in the presence of protease and phosphatase inhibitors Complete mini and PhosSTOP (Roche Diagnostics) and centrifuged at 16 000g for 20 minutes at 4°C. When indicated, cells were incubated with 10nM or 30nM calyculin A or with 5nM or 50nM okadaic acid for 10 minutes before agonist exposure. All phosphorylation and dephosphorylation assays were performed at physiologic temperature (37°C) or at room temperature (22°C) for the indicated time periods. Glycosylated proteins were partially enriched using wheat germ lectin-agarose beads as described (32–34). When indicated, ssts were isolated using anti-HA antibodies coupled to protein A agarose beads. Proteins were eluted from the beads using SDS-sample buffer for 15 minutes at 60°C and then resolved on 8% SDS-polyacrylamide gels. After electroblotting, membranes were incubated with phosphosite-specific antibodies anti-pS337/pT341, anti-pT348, or anti-pS361 at a concentration of 0.1 μg/mL followed by detection using enhanced chemiluminescence (Amersham). Blots were subsequently stripped and reprobed with anti-sst₃ antibody (UMB-5) or anti-HA antibody to confirm equal loading of the gels. Films exposed in the linear range were then densitized using ImageJ 1.45. Data were analyzed using GraphPad Prism 4.0 software.

Analysis of receptor internalization by confocal microscopy

HEK293 cells stably expressing HA-tagged human sst₃ receptors or its phosphorylation-deficient mutants were grown on poly-L-lysine-coated coverslips overnight. After the appropriate treatment with 1μM ligands at 37°C, cells were fixed with 4% paraformaldehyde and 0.2% picric acid in phosphate buffer (pH 6.9) for 30 minutes at room temperature and washed several times. Cells were permeabilized and then incubated with anti-HA antibody followed by Alexa Fluor 488-conjugated secondary antibody (Amersham). Specimens were mounted and examined using a Zeiss LSM510 META laser scanning confocal microscope (35).

Quantitative internalization assays

Receptor internalization was quantified using a linear surface receptor ELISA that has been characterized extensively (13, 31, 35, 36). Equal numbers of stably transfected HEK293 cells expressing HA-tagged human sst₃ receptors or its phosphorylation-deficient mutants were seeded onto poly-L-lysine-treated 24-well plates (200 000 cells per well). The next day, cells were preincubated with 1-μg/mL anti-HA antibody for 2 hours at 4°C. After the appropriate treatment with SS-14 or another agonist (1μM) at 37°C, cells were fixed and incubated with peroxidase-conjugated antirabbit antibody overnight. After washing, plates were developed with ABTS solution and analyzed at 405 nm using a microplate reader. Statistical analysis was carried out with unpaired t test. P < .05 was considered statistically significant.

β-Arrestin-EGFP mobilization assay

Untransfected HEK293 cells were seeded onto 35-mm glass-bottom culture dishes (Mattek). The next day, cells were transiently cotransfected with 0.2 μg β-arrestin-2-EGFP or β-arrestin-1-EGFP and 2μg human sst₃ receptor or the phosphorylation-deficient mutants per dish containing 200 000 cells using TurboFect. After 24 hours, cells were transferred onto a temperature-controlled microscope stage set at 37°C of a Zeiss LSM510 META laser scanning confocal microscope. Images were collected sequentially using single line excitation at 488 nm with 515- to 540-nm band pass emission filters. Saturating concentrations of agonists (1μM) were applied directly into the culture medium immediately after the initial image was taken.

Colocalization of receptor and β-arrestin-EGFP

Untransfected HEK293 cells were grown on poly-L-lysine-coated coverslips overnight. The next day, cells were transiently cotransfected with 0.2μg β-arrestin-2-EGFP or β-arrestin-1-EGFP, 1.2μg human sst₃ receptor or the phosphorylation-deficient mutants and 0.8μg GRK2 per dish using TurboFect. After 24 hours, cells were preincubated with 1μg/mL anti-HA antibody for 2 hours at 4°C. After the appropriate treatment with 1μM SS-14 at 37°C, cells were fixed with 4% paraformaldehyde and 0.2% picric acid in phosphate buffer (pH 6.9) for 30 minutes at room temperature and washed several times. Cells were permeabilized and then incubated with chicken anti-GFP antibody (Aves) followed by goat antichicken Alexa Fluor 488-conjugated secondary antibody (Life Technologies) and with donkey anti-rabbit Cy3-conjugated secondary antibody (Jackson ImmunoResearch) for receptor staining. Specimens were mounted and examined using a Zeiss LSM510 META laser scanning confocal microscope.

Membrane potential assay

AtT-20 cells were plated into poly-L-lysine covered 96-well plates and a number of 30 000 cells per well in a volume of 200 μL. Cells were kept in an incubator for 48 hours to reach a confluency of 80% before experiments. After removal of medium, cells were washed with Hanks' balanced salt solution, buffered with HEPES 20mM (pH 7.4), containing 1.3mM CaCl₂, 5.4mM KCl, 0.4mM K₂HPO₄, 0.5mM MgCl₂, 0.4mM MgSO₄, 136.9mM NaCl, 0.3mM Na₂HPO₄, 4.2mM NaHCO₃, and 5.5mM glucose. The membrane potential dye (FLIPR Membrane Potential kit BLUE; Molecular Devices) was reconstituted according to manufacturer's instructions. Cells were incubated in 90 μL of Hanks' balanced salt solution/HEPES (vehicle solution) and an equal volume of assay dye solution for 45 minutes at 37°C. Plates were assayed using a FlexStation 3 microplate reader (Molecular Devices). All measurements were conducted at 37°C. Compounds were injected in a volume of 20 μL in a 10-fold concentration referring to the final concentration in the well. A baseline was recorded before agonist injection for 60 seconds. An equal concentration of solvent was used as vehicle control. Cells were exposed to maximum concentration of 0.5% dimethyl sulfoxide (DMSO). For data analysis GraphPad Prism 6 and OriginPro were used using a 4-parameter nonlinear regression for fitting concentration-response curves.

Small interfering RNA (siRNA) silencing of gene expression

Chemically synthesized double-stranded siRNA duplexes (with 3′ dTdT overhangs) were purchased from QIAGEN for the following targets: GRK2 (5′-CCGGGAGATCTTCGACTCATA-3′ and 5′-AAGAAGTACGAGAAGCTGGAG-3′), GRK3 (5′-AAGCAAGCTGTAGAACACGTA-3′ and 5′-GCAGAAGTCGACAAATTTA-3′), GRK5 (5′-AGCGTCATAACTAGAACTGAA-3′ and 5′-AAGCCGTGCAAAGAACTCTTT-3′), GRK6 (5′-AACACCTTCAGGCAATACCGA-3′ and 5′-AACAGTAGGTTTGTAGTGAGC-3′), PP1α catalytic subunit (5′-AAGAGACGCTACAACATCAAA-3′), PP1β catalytic subunit (5′-TACGAGGATGTCGTCCAGGAA-3′ and 5′-GTTCGAGGCTTATGTATCA-3′), PP1γ catalytic subunit (5′-AACATCGACAGCATTATCCAA-3′ and 5′-AGAGGCAGTTGG TCACTCT-3′), and a nonsilencing RNA duplex (5′-GCTTAGGAGCATTAGTAAA-3′ or 5′-AAACTCTATCTGCACGCTGAC-3′). HEK293 cells were transfected with 100nM siRNA for single transfection or double transfection using HiPerFect (QIAGEN). Silencing was quantified by immunoblotting. All experiments showed protein levels reduced by more than or equal to 80% (20).

Results

We have recently shown that agonist-induced phosphorylation of sst₂ receptor occurs at a specific cluster of serine and threonine residues within the C-terminal tail of the receptor (13, 18, 37). Similar, we found that trafficking of the sst₅ receptor is regulated by phosphorylation of C-terminal T333 (14). Nevertheless, little is known about the functional role of the C terminus for sst₃ receptor regulation. This is in part due to presence of a huge number of potential phosphate acceptor sites within its very long C-terminal tail. To address this problem, we first generated 2 sst₃ receptor mutants, where either all 6 serine and threonine residues present in the proximal part (317–370) or all 12 serine and threonine residues present in the distal part (371–418) of the C terminus were mutated to alanine (Figure 1A). This approach should allow us to determine which part of the sst₃ C terminus is crucial for agonist-dependent arrestin recruitment and receptor internalization. Thus, we employed functional β-arrestin-1 and β-arrestin-2 conjugated to enhanced green fluorescent protein to visualize the translocation of arrestin to the plasma membrane in live HEK293 cells transfected with wild-type or mutant sst₃ receptors. In the absence of somatostatin, both arrestins were mostly distributed throughout the cytoplasm of the cells (Figure 1B, 0 min). The addition of saturating concentrations of agonist induced a small but clearly detectable redistribution of β-arrestin-1 and a very robust redistribution of β-arrestin-2 from the cytoplasm to the plasma membrane in cells transfected with wild-type sst₃ receptor or the 371–418 S/T-A mutant. In contrast, no such translocation was observed in cells transfected with the 317–370 S/T-A mutant. Similar effects were obtained after qualitative and quantitative analysis of receptor internalization (Figure 1, C and D). In fact, confocal microscopic images show that a 30-minute SS-14 exposure of the wild-type sst₃ receptor induced a very robust redistribution of the receptor from the plasma membrane into puntate-like structures within the cytososl (Figure 1C). Quantitative analysis using a biochemical ELISA indicated that approximately 65% of all surface receptors were internalized (Figure 1D). A robust receptor internalization was also observed with 371–418 S/T-A mutant with only few receptors remaining at the plasma membrane (Figure 1C). Quantitative analysis revealed that up to 40% of all surface receptors were internalized (Figure 1D). In contrast, in the 317–370 S/T-A mutant receptor endocytosis was nearly undetectable (<5% internalization) (Figure 1, C and D). Both octreotide and pasireotide induced a partial internalization (∼20%) of the wild-type sst₃ receptor and the 371–418 S/T-A mutant, but virtually no internalization of the 317–370 S/T-A mutant (Figure 1, C and D). These results clearly indicate that arrestin recruitment and internalization of the sst₃ receptor crucially depend on phosphorylation of the proximal part of its C terminus.

Figure 1. Construction of phosphorylation-deficient sst₃ receptor mutants.

A, Schematic representation of the human sst₃ receptor indicating all serine and threonine sites within the C-terminal tail (317–418). Phosphate acceptor sites mutated to alanine in the 317–370 S/T-A, and 371–418 S/T-A mutants are marked in bold. B, HEK293 cells were transiently transfected with either wild-type hsst₃, 317–370 S/T-A or 371–418 S/T-A and β-arrestin-2-EGFP or β-arrestin-1-EGFP. The distribution of β-arrestin-2 and β-arrestin-1 were visualized sequentially in the same live cells before (0 min) and after (5 and 15 min) the addition of 1μM SS-14 to the culture medium. Shown are representative images from 1 of 4 independent experiments performed in duplicate. Scale bar, 15 μm. C, HEK293 cells stably expressing wild-type hsst₃, 317–370 S/T-A, or 371–418 S/T-A were treated with 1μM somatostatin, octreotide, or pasireotide for 30 minutes. Cells were then fixed, stained with the anti-HA antibody, and examined by confocal microscopy. Shown are representative images from 1 of at least 3 independent experiments performed in duplicate. Scale bar, 15 μm. D, HEK293 cells stably expressing wild-type or mutant hsst₃ receptors were treated for 30 minutes with 1μM agonist. Receptor sequestration was measured by surface ELISA. Data represent receptor internalization in agonist-treated cells as compared with sister cultures receiving 30 minutes of vehicle (100%). Data are presented as mean ± SEM from at least 4 independent experiments performed in quadruplicate. Results were analyzed by unpaired t test vs wild type receiving 30 minutes of vehicle (*, P < .05).

We then aligned the proximal part of the sst₃ C terminus with the C-terminal domains of the well-characterized sst₂ and sst₅ receptors. Figure 2A depicts the amino acid sequences of all 3 receptors beginning with the NPXXY motif which is conserved in GPCRs of the rhodopsin subfamily and marks the end of the seventh transmembrane domain and the beginning of the C terminus. All known phosphorylation sites namely S341, S343, T353, T354, T356, and T359 in sst₂ and T333 and T347 in sst₅ are depicted in bold. Comparison of these sequences suggests that C-terminal phosphorylation of these receptors occurs preferentially in a certain distance from the plasma membrane. In the corresponding region of the sst₃ receptor, 4 potential phosphate acceptor sites namely S337, T341, T348, and S361 are present (Figure 2A). In order to assess their role in sst₃ regulation, we generated a receptor with combined mutation of only these 4 serine and threonine residues (4 S/T-A). Interestingly, analysis of the 4 S/T-A mutant revealed no mobilization of β-arrestin-1 or β-arrestin-2 and virtually no internalization (<5%) in response to SS-14 exposure (Figure 2, B–D).

Figure 2. Internalization and β-arrestin recruitment of the sst₃ receptor is regulated by 4 C-terminal serine and threonine residues.

A, Schematic representation of the human sst₂, sst₅, sst₃, and sst₃-4 S/T-A receptors beginning with the conserved NPXXY motif which marks the end of seventh transmembrane domain and the beginning of the C-terminal tail. Phosphate acceptor sites essential for the agonist-driven internalization of the human sst₂ and sst₅ receptors are marked in bold. Serine and threonine residues present in the corresponding region of the human sst₃ receptor were replaced by alanine resulting in a 4 S/T-A mutant. B, HEK293 cells were transiently transfected with either wild-type hsst₃ or 4 S/T-A and β-arrestin-2-EGFP or β-arrestin-1-EGFP. The distributions of β-arrestin-2 and β-arrestin-1 were visualized sequentially in the same live cells before (0 min) and after (5 and 15 min) the addition of 1μM SS-14 to the culture medium. Shown are representative images from 1 of 4 independent experiments performed in duplicate. Scale bar, 15 μm. C, HEK293 cells were transiently transfected with either wild-type hsst₃, 317–370 S/T-A, 371–418 S/T-A, or 4 S/T-A and β-arrestin-2-EGFP or β-arrestin-1-EGFP as well as GRK2. The next day, cells were treated with 1μM SS-14 for 0 and 2 minutes, fixed, stained, and examined by confocal microscopy. Shown are representative images from 1 of at least 3 independent experiments performed in duplicate. Scale bar, 15 μm. D, HEK293 cells stably expressing wild-type hsst₃ or 4 S/T-A were treated with 1μM somatostatin, octreotide, or pasireotide for 30 minutes. Cells were then fixed, stained with anti-HA antibody, and examined by confocal microscopy. Shown are representative images from 1 of at least 3 independent experiments performed in duplicate. Scale bar, 15 μm. E, HEK293 cells stably expressing wild-type hsst₃ or 4 S/T-A receptor were treated for 30 minutes with 1μM agonist. Receptor sequestration was measured by surface ELISA. Data represent receptor internalization in agonist-treated cells as compared with sister cultures receiving 30 minutes of vehicle (100%). Data are presented as mean ± SEM from at least 4 independent experiments performed in quadruplicate. Results were analyzed by unpaired t test vs wild type receiving 30 minutes of vehicle (*, P < .05).

Our results suggest that these 4 sites are most likely to undergo agonist-induced phosphorylation. Consequently, we generated phosphosite-specific antibodies to S337/T341, T348, and S361 (Figure 3A). These antisera were affinity-purified against their immunizing peptides and initially tested in dot-blot assays using the phospho- and the corresponding nonphospho-peptides (data not shown). All antibodies, which clearly detected their respective phospho-peptide and did not cross-react with the corresponding nonphospho-peptide were further characterized in Western blot analysis using HEK293 cells transiently transfected with the human sst₃ receptor and its phosphorylation-deficient mutants (Figure 3B). In Western blottings from cells expressing the wild-type sst₃ receptor, a band migrating at M_r 65 000–80 000 was detectable which corresponds to the expected molecular size of the sst₃ monomer (Figure 3B). The anti-pS337/pT341, anti-pT348 as well as anti-pS361 antibodies detected the phosphorylated form of the sst₃ receptor in SS-14-treated cells but not in untreated cells (Figure 3B). As expected, analysis of the 371–418 S/T-A mutant revealed a phosphorylation profile that was similar to that seen with the wild-type sst₃ receptor (Figure 3B). Moreover, no phosphorylation signal was detectable in the 317–370 S/T-A or the 4 S/T-A mutant (Figure 3B) indicating that the newly generated anti-pS337/pT341, anti-pT348, and anti-pS361 antibodies specifically detect the phosphorylated form of their targeted receptor. Moreover, the results show that all 4 sites including S337, T341, T348, and S361 undergo agonist-promoted phosphorylation in the presence of the endogenous ligand SS-14.

Figure 3. Generation of phosphosite-specific antibodies for the human sst₃ receptor.

A, Schematic representation of the human sst₃ receptor indicating all potential phosphate acceptor sites within the C-terminal tail in gray. Epitopes of the phosphosite-specific antibodies are depicted in black. B, HEK293 cells stably expressing wild-type hsst₃, 317–370 S/T-A, 4S/T-A, or 371–418 S/T-A were either not exposed or exposed to 1μM SS-14 for 15 minutes at 37°C. The levels of phosphorylated sst₃ receptor and phosphorylation-deficient mutants were then determined using the phosphosite-specific anti-pS337/pT341, anti-pT348, or anti-pS361 antibodies. Blots were subsequently stripped and reprobed with anti-HA antibody to confirm equal loading of the gels (anti-HA). Blots shown are representative of 3 independent experiments. The positions of molecular mass markers are indicated on the left (in kDa).

We then examined the ability of sst₃ and its phosphorylation-deficient mutants to activate G proteins. Similar to other sst subtypes, sst₃ is known to activate G protein-coupled inwardly rectifying potassium (GIRK) channels (38). Measuring GIRK channel activation is therefore a plausible surrogate parameter to assess an agonist-induced sst₃-mediated G protein downstream signal. We therefore adapted a fluorescence-based membrane potential assay protocol originally established by Vazquez et al (39) and Walsh (40) in AtT-20 cells. This mouse corticotrope cell line expresses abundantly sst₂ and sst₅ receptors (41, 42). In wild-type AtT-20 cells, the sst₃-selective agonist L-796/778 did not induce a significant change in fluorescence signal even in a saturating concentration of 10μM. However, in cells stably expressing hsst₃ L-796/778 induced a fast and robust decrease in fluorescence signal intensity (Figure 4A). This agonist-induced hyperpolarization was efficiently blocked by subsequent application of the sst₃-selective antagonist NVP-ACQ090 (Figure 4B). In wild-type AtT-20 cells, the SS-14-induced decrease of fluorescence was completely abrogated by pretreatment with pertussis toxin (PTX), which is indicative for a Gα_i protein-mediated process (Figure 4C). Tertiapin-Q (TPN-Q) significantly inhibited both the SS-14- and the L-796/778-induced hyperpolarization, which is indicative for a GIRK-mediated mechanism (Figure 4C). Interestingly, SS-14, octreotide and pasireotide induced a dose-dependent hyperpolarization in wild-type AtT-20 cells (Figure 4, D–F). After coexpression of sst₃, however, only SS-14 and pasireotide but not octreotide were able to promote a robust leftward shift of the dose-response curve (Figure 4, D–F).

Figure 4. Membrane potential measurements.

A, Wild-type AtT-20 cells (gray trace) and AtT-20 cells stably expressing hsst₃ (black trace) were exposed to a solution containing 10μM sst₃-selective agonist L-796/778 to yield a final in-well concentration of 1μM. Changes in fluorescence signal intensity detected for wild-type cells were 0.6 ± 0.7% (mean ± SEM, n = 3, ns), respectively, 17.2 ± 2.3% (mean ± SEM, n = 4; *, P < .05). Shown is 1 representative experiment executed as triplicate (mean ± SD). Data indicated are composed of 3 independent experiments executed as triplicate vs vehicle injection as duplicates. Vehicle induced background changes in membrane potential were subtracted from agonist-induced signals. All values indicated are interpreted at t_180sec. B, Inhibition of L-796/778-induced hyperpolarization by sst₃-selective antagonist NVP-ACQ090. AtT-20 cells stably expressing wild-type sst₃ were primarily exposed to 50μM L-796/778 at t_60sec for 180 seconds (black and gray trace) and subsequently exposed to either a 50μM solution of NVP-ACQ090 at t_240sec for another 180 seconds (black trace) or vehicle (gray trace). C, PTX and TPN-Q sensitivity of SS-14 and L-796/778-induced changes in membrane potential. Left panel, Exposure of wild-type AtT-20 cells with 10μM SS-14 induced a change of 29.2 ± 0.8% (mean ± SEM, n = 5) of the fluorescence signal (left bar). After pretreatment with 300-ng/mL PTX for 20 hours, signal magnitude significantly decreased to 1.8 ± 0.9% (mean ± SEM, n = 4; **, P < .01) after exposure to 10μM SS-14 (center bar) compared with untreated cells. In cells pretreated with 500nM TPN-Q for 5 minutes before agonist exposure (right bar), signal magnitude significantly decreased to 3.6 ± 0.9% (mean ± SEM, n = 5; *, P < .05). Right panel, AtT-20 cells stably expressing hsst₃ were challenged with 10μM 796/778 (left bar) to induce a significant change in fluorescence signal of 17.2 ± 0.6% (mean ± SEM, n = 4). L-796/778 induced a significant decrease of fluorescence intensity in cells pretreated with PTX (1.0 ± 0.8%, mean ± SEM, n = 4; *, P < .05) or TPN-Q (1.3 ± 0.7%, mean ± SEM, n = 4; *, P < .05) compared with the untreated cells. D, Dose-response curves of wild-type AtT-20 cells vs hsst₃-expressing cells stimulated with SS-14 in a concentration range from 1pM to 1μM. The EC₅₀ value obtained was 4.5 ± 0.5nM for wild-type cells and 1.7 ± 0.3nM for hsst₃-expressing cells. E, Dose-response curves of wild-type AtT-20 cells vs hsst₃-expressing cells stimulated with octreotide in a concentration range from 10pM to 5μM with EC₅₀ of 30.5 ± 4.6nM for wild-type cells and 30.5 ± 3.9nM for hsst₃-expressing cells. F, Dose-response curves of wild-type AtT-20 cells vs hsst₃-expressing cells stimulated with pasireotide in a concentration range from 1pM to 1μM with EC₅₀ of 10.4 ± 0.8nM for wild-type cells and 5.8 ± 1.3nM for hsst₃-expressing cells. EC₅₀ values are given as mean ± SD. Results were analyzed by unpaired t test vs wild type.

We then analyzed the capacity of clinically relevant somatostatin analogs to induce sst₃ phosphorylation using HEK293 cells stably expressing sst₃ receptor (Figure 5, left panel). In Western blottings from these cells, 2 broad bands were detectable migrating at M_r 65 000–80 000 and at M_r 130 000–160 000 corresponding to the sst₃ monomer and sst₃ homodimer, respectively. The anti-pS337/pT341, and the anti-pT348 antibodies detected the phosphorylated forms of sst₃ monomers and sst₃ dimers in SS-14-treated cells but not in untreated cells. Interestingly, SS-14 induced a very strong phosphorylation signal, whereas the stable somatostatin analogs pasireotide and octreotide promoted a much weaker although clearly detectable sst₃ phosphorylation (Figure 5). In contrast, neither somatoprim nor the selective sst₃ agonist L-796/778 were able to induce any detectable sst₃ phosphorylation (Figure 5). As expected, analysis of the 371–418 S/T-A mutant revealed a phosphorylation profile that was nearly indistinguishable from that observed for the wild-type sst₃ receptor (Figure 5, third panel). Moreover, no phosphorylation signal was detectable in the 317–370 S/T-A or the 4 S/T-A mutant (Figure 5, second and fourth panels) indicating that the newly generated anti-pS337/pT341 and anti-p348 antibodies specifically detect the phosphorylated form of their targeted receptor. In addition, the present results suggest an excellent correlation between sst₃ receptor phosphorylation and internalization (Figures 1, 2, and 5). We then assessed G protein signaling of all sst₃ mutants using the selective sst₃ agonist L-796/778 (Figure 5, bottom panel). Although we were unable to detect any differences in E_max values after exposure of saturating concentrations of L-796/778, we found robust differences in EC₅₀ values indicating a leftward shift of the dose-response curves in the phosphorylation-deficient 317–370 S/T-A and 4 S/T-A mutants (Figure 5, bottom panel).

Figure 5. Evaluation of agonist-dependent phosphorylation of phosphorylation-deficient human sst₃ receptor mutants.

HEK293 cells stably expressing wild-type hsst₃, 317–370 S/T-A, 4S/T-A, or 371–418 S/T-A were treated with 1μM SS-14, octreotide, pasireotide, somatoprim, or L-796/778 for 10 minutes at room temperature. The levels of phosphorylated sst₃ receptors were then determined using the phosphosite-specific anti-pS337/pT341 or anti-pT348 antibodies. Blots were subsequently stripped and reprobed with the phosphorylation-independent anti-HA antibody to confirm equal loading of the gels (anti-HA). Blots shown are representative of 3 independent experiments. The positions of molecular mass markers are indicated on the left (in kDa). Dose-response curves of AtT-20 cells stably expressing wild-type hsst₃, 317–370 S/T-A, 371–418 S/T-A, or 4S/T-A stimulated with sst₃-selective agonist L-796/778. Cells were stimulated with L-796/778 at concentrations ranging from 0.1nM to 5μM to yield a dose-response curve. EC₅₀ values are given as mean ± SD.

Next, we examined dose- and time-dependent sst₃ phosphorylation after exposure to various concentrations of SS-14 ranging from 10⁻⁹M to 10⁻⁵M. As depicted in Figure 6, left panel, a weak phosphorylation signal was first detectable at concentrations of SS-14 as low as 10nM. Phosphorylation reached a maximum at concentrations between 100nM and 1μM SS-14. In the presence of 1μM SS-14, sst₃ phosphorylation became detectable within 2 minutes even at room temperature (∼22°C) (Figure 6, second panel). Maximal sst₃ phosphorylation with no further increase was reached after 5 minutes of SS-14 exposure at room temperature (Figure 6, second panel). To examine the time course of sst₃ dephosphorylation, we first exposed the cells for 10 minutes to SS-14. The ligand was then washed out and cells were incubated in culture medium in the absence of agonist for different periods of time ranging from 2 to 90 minutes. Under these conditions, dephosphorylation of T348 occurred within 10–15 minutes, whereas dephosphorylation of S337/T341 required more than 30 minutes (Figure 6, third panel). We also observed that SS-14-induced phosphorylation of the sst₃ receptor was completely blocked by the specific sst₃ antagonist NVP-ACQ090 (Figure 6, fourth panel). This antagonist also inhibited SS-14-mediated arrestin recruitment and receptor internalization (data not shown). Phosphorylation of GPCRs can occur via specific GRKs or second messenger-activated kinases. We therefore treated sst₃-expressing HEK293 cells either with the protein kinase A activator forskolin (10μM) or the protein kinase C activator PMA (0.1μM) and examined sst₃ phosphorylation. We found that neither PMA nor forskolin produced any detectable phosphorylation of S337/T341 or T348 (Figure 6, right panel).

Figure 6. Agonist-induced phosphorylation and dephosphorylation of the human sst₃ receptor.

Left panel, HEK293 cells stably expressing the sst₃ receptor were either not exposed or exposed to the indicated concentrations of SS-14 for 10 minutes at room temperature. Second panel, cells were exposed to 1μM SS-14 at room temperature for the indicated time periods. Third panel, cells were exposed to 1μM SS-14 for 10 minutes, washed, and incubated in the absence of agonist for 0, 2, 5, 10, 15, 30, 60, or 90 minutes. Fourth panel left, Cells were incubated with the sst₃ antagonist NVP-ACQ090 for 10 minutes before SS-14 exposure at 37°C. Fourth panel right, Cells were exposed to either 1μM SS-14, 0.1μM PMA, or 10μM forskolin for 10 minutes at room temperature. Cells were lysed and immunoblotted with the phosphosite-specific antibodies. Blots were stripped and reprobed with the phosphorylation-independent anti-HA antibody to confirm equal loading of the gel (anti-HA). Representative results from 3 independent experiments are shown. The positions of molecular mass markers are indicated on the left (in kDa).

We then elucidated the question, which GRKs are involved in sst₃ phosphorylation. We used specific siRNAs to knockdown gene expression of GRK2, GRK3, GRK5, or GRK6. Under these conditions, we observed a robust inhibition of SS-14-induced sst₃ phosphorylation after knockdown of GRK2 or GRK3 (Figure 7). Quantitative analysis revealed that phosphorylation was reduced by approximately 40% after knockdown of GRK2 or GRK3. In contrast, no such inhibition was observed after knockdown of GRK5 and/or GRK6 (Figure 7).

Figure 7. GRK2 and GRK3 are responsible for agonist-induced phosphorylation of the human sst₃ receptor.

A, HEK293 cells stably expressing the sst₃ receptor were transfected with siRNA targeted to GRK2, GRK3, GRK5, and GRK6 or nonsilencing siRNA control (SCR) for 72 hours and then exposed to 1μM SS-14 for 10 minutes (room temperature). Cells were lysed and immunoblotted with anti-pS337/pT341 antibodies. Blots were stripped and reprobed with the phosphorylation-independent anti-HA antibody to confirm equal loading of the gels (anti-HA). Receptor monomers were quantified and expressed as percentage of maximal phosphorylation in SCR-transfected cells. Data correspond to the mean ± SEM from 3 independent experiments. Results were analyzed by two-way ANOVA followed by the Bonferroni post hoc test (*, P < .05). The arrow indicates the position of sst₃ monomers that were used for quantification. B and C, siRNA knockdown of GRK2, GRK3, GRK5, and GRK6 was confirmed by Western blotting. The positions of molecular mass markers are indicated on the left (in kDa).

Recently, we identified PP1 as GPCR phosphatase for the sst₂ and sst₅ using a combination of chemical inhibitors and siRNA knockdown screening (14, 20). We hence used the same approach to identify the GPCR phosphatase responsible for dephosphorylation of sst₃. When stably transfected HEK293 cells were exposed to increasing concentrations of protein phosphatase inhibitors, sst₃ dephosphorylation was inhibited in a dose-dependent manner only by calyculin A (Figure 8A) but not by okadaic acid (data not shown). Both calyculin A and okadaic acid can effectively block the protein phosphatases PP2, PP4, and PP5 (43). In contrast to okadaic acid, calyculin A is also a potent inhibitor of PP1 activity (43). Thus, the present data suggest that PP1 activity is required for the S337/T341 dephosphorylation. To date, the 3 distinct catalytic subunits α, β, and γ for PP1 are known. To elucidate which of these PP1 isoforms is involved in sst₃ dephosphorylation, we performed siRNA knockdown experiments. As depicted in Figure 8B, only a knockdown of PP1α and PP1β resulted in a robust inhibition of sst₃ dephosphorylation. These results indicate that PP1 is the GPCR phosphatase responsible for S337/T341 dephosphorylation of sst₃.

Figure 8. Calyculin A prevents sst₃ dephosphorylation and PP1α and PP1β catalyze this process.

A, HEK293 cells stably expressing sst₃ were treated with calyculin A for 10 minutes (37°C) at the indicated concentrations and then exposed to 1μM SS-14 for 10 minutes (37°C) in the presence of calyculin A. Cells were washed 3 times and then incubated for 0, 5, or 15 minutes in the presence of calyculin A but absence of agonist (room temperature). Data correspond to the mean ± SEM from 3 independent experiments. B, HEK293 cells stably expressing the sst₃ receptor were transfected with siRNA targeted to PP1α, PP1β, PP1γ, or nonsilencing siRNA control (SCR) for 72 hours and then exposed to 1μM SS-14 for 10 minutes. Cells were washed 3 times and then incubated for 0, 5, or 15 minutes in the absence of agonist. Cells were lysed and immunoblotted with anti-pS337/pT341 antibodies. Blots were stripped and reprobed with the phosphorylation-independent anti-HA antibody to confirm equal loading of the gels (anti-HA). Phosphorylation of monomers was quantified and expressed as percentage of maximal phosphorylation in SCR-transfected cells, which was set at 100%. Data correspond to the mean ± SEM from 4 independent experiments. Results were analyzed by two-way ANOVA followed by the Bonferroni post hoc test (*, P < .05). Note that 1) calyculin A inhibited sst₃ dephosphorylation in a dose-dependent manner and 2) transfection with PP1α and PP1β siRNA resulted in a significant inhibition of sst₃ dephosphorylation. The arrow indicates the position of sst₃ monomers which were used for quantification. C, siRNA knockdown of PP1 was confirmed by Western blotting using subunit-specific PP1 antibodies. The positions of molecular mass markers are indicated on the left (in kDa).

Given the differential trafficking of wild-type and mutant sst₃ receptors, we next assessed the effects of the ablation of specific phosphorylation sites on the down-regulation of the receptor proteins during prolonged SS-14 exposure by Western blotting. As shown in Figure 7, we did not observe any detectable change in the level of cellular 4 ST-A or 317–370 S/T-A mutants during the 24-hour treatment period. In contrast, the sst₃ receptor and the 371–418 S/T-A mutant underwent pronounced down-regulation, which became clearly detectable after prolonged agonist exposure (Figure 9). Furthermore a mutation of all 7 intracellular lysine residues which are potential acceptor sites for ubiquitination of the sst₃ receptor only delays down-regulation during the first 6 hours of SS-14 treatment. After 24 hours SS-14 the K-R mutant was degraded to virtually identical extend as the wild-type sst₃ receptor (Figure 9). These results suggest that rapid sst₃ down-regulation requires phosphorylation of S337, T341, T348, and S361, whereas ubiquitination appears to be not required.

Figure 9. Phosphorylation is required for sst₃ down-regulation.

HEK293 cells stably expressing the wild-type or mutant sst₃ receptors were treated with 1μM somatostatin for 0, 2, 4, 6, or 24 hours. Cells were lysed and immunoblotted with the phosphorylation-independent anti-HA antibodies. Blots were stripped and reprobed with the antitransferrin receptor antibody to confirm equal loading of the gel (TfR, transferrin receptor). Receptor monomers were quantified and expressed as percentage of untreated cells. Data correspond to the mean ± SEM from 5 independent experiments. Results were analyzed by two-way ANOVA followed by the Bonferroni post hoc test (*, P < .05). The positions of molecular mass markers are indicated on the left (in kDa).

Discussion

The human sst₃ receptor is unique among ssts in that it exhibits a very long intracellular C-terminal tail containing a large number of potential phosphate acceptor sites. Nevertheless, we show that the sst₃ receptor is regulated like a prototypical GPCR. In fact, we identify S337, T341, T348, and S361 as primary sites of agonist-induced phosphorylation mediated primarily by GRK2 and GRK3. These 4 sites are all located in the proximal part of the sst₃ C-terminal tail. Alignment of the phosphorylation motifs identified for the sst₂ and sst₅ receptors suggested that GRK2/3-mediated phosphorylation of these receptors occurs preferentially in a defined distance from the NPXXY motif. This knowledge led us to the identification of the sst3 phosphorylation motif with high precision and to successfully generate several novel phosphosite-specific antibodies. The C-terminal regions of the human and the rat sst₃ receptor exhibit strikingly different sequences. However, critical sites involved in agonist-induced phosphorylation have previously been identified in a similar region of the rat sst₃ receptor namely S341, S346, S351, and T357 (25). In contrast, phosphorylation of the proximal part of the sst₃ C-terminal tail or the third intracellular loop appears not to be required for its rapid agonist-dependent regulation.

Generation of these phosphosite-specific antibodies enabled us to demonstrate an excellent correlation between extent and temporal dynamics of C-terminal phosphorylation and the trafficking properties of the human sst₃ receptor. We also show that S337/T341 is rapidly phosphorylated and slowly dephosphorylated in a SS-14-dependent manner. In addition, we also show that mutation of these 4 serine and threonine residues is required and sufficient to completely block arrestin recruitment and receptor internalization. Moreover, we identify GRK2 and GRK3 as kinases responsible for phosphorylation and PP1α and PP1β as phosphatases responsible for dephosphorylation.

We have previously shown that analysis of agonist-induced phosphorylation using sst chimeras can provide valuable and previously unappreciated information about multireceptor somatostatin analogs currently under clinical and preclinical examination (18). In fact, pasireotide exhibited potent agonistic activity at the sst₅ receptor but only weak partial agonistic properties at sst₂ and sst₃ receptors. Consequently, pasireotide should be classified as sst₅-preferring ligand. Octreotide is a full agonist at the sst₂ receptor but exhibited partial agonistic activity at the sst₃ receptor and virtually no agonistic activity at the sst₅ receptor. Consequently, octreotide should be classified as sst₂-preferring ligand. In contrast, somatoprim is unique in that it was a potent agonist at both sst₂ and sst₅ receptors but no activity at the sst₃ receptor. The present results using our newly generated phosphosite specific sst₃ antibodies targeting directly the sst₃ receptor not only confirm these results but also show that these antibodies can be valuable tools for pharmacological characterization of novel somatostatin analogs.

The human sst₃ receptor is also unique among ssts in that it is rapidly down-regulated during prolonged agonist exposure (35). This rapid down-regulation was observed after exposure to both pasireotide and octreotide and, hence, appears to be independent of the nature of the peptide ligand (35). We have previously reported that the sst₃ receptor undergoes agonist-induced ubiquitination (19). To test whether this ubiquitination is a precondition for down-regulation, we mutated all 7 lysine residues. However, after 24 hours SS-14 the K-R mutant was degraded to virtually identical extend as the wild-type sst₃ receptor. Thus, rapid sst₃ down-regulation requires phosphorylation of S337, T341, T348, and S361 and subsequent internalization, whereas ubiquitination appears to be dispensable.

In many systems, rapid down-regulation of GPCRs during continued or repeated drug administration leads to development of tolerance. Based on our down-regulation assays, we would predict that an early loss of clinical response is more likely to occur in tumors with predominant sst₃ expression than in sst₂- or sst₅-expressing tumors. Thus, due to its rapid down-regulation, the sst₃ receptor seems to be a less favorable pharmacological target for long-term administration of somatostatin analogs. Nevertheless, we observed that overexpression of sst₃ in AtT-20 cells which endogenously express sst₂ and sst₅ receptors results in a leftward shift of the dose-response curve for pasireotide but not octreotide. Moreover, the sst₃ receptor is one of the few GPCRs, which can drive tumor cells apoptosis and the C-terminal tail seems to be required for this activity (3, 26, 44–46). Thus, the sst₃ receptor is a potential pharmacological target for diagnostic and acute therapeutic intervention using multireceptor somatostatin analogs (22–24, 47).

In conclusion, we provide direct evidence for agonist selective phosphorylation of C-terminal S337/T341, T348, and S361 of the human sst₃ receptor. In addition, we identify GRK2- and GRK3-mediated phosphorylation and PP1α- and PP1β-mediated dephosphorylation as key regulators of rapid phosphorylation and dephosphorylation of the human sst₃ receptor.

Funding Statement

This work was supported by the Deutsche Forschungsgemeinschaft Grant SCHU924/10–4, the Deutsche Krebshilfe Grant 109952, and a grant from the Dr. Robert Pfleger-Stiftung.