Carboxyl-Terminal Receptor Domains Control the Differential Dephosphorylation of Somatostatin Receptors by Protein Phosphatase 1 Isoforms

We have recently identified protein phosphatase 1β (PP1β) as G protein-coupled receptor (GPCR) phosphatase for the sst2 somatostatin receptor using siRNA knockdown screening. By contrast, for the sst5 somatostatin receptor we identified protein phosphatase 1γ (PP1γ) as GPCR phosphatase using the same approach. We have also shown that sst2 and sst5 receptors differ substantially in the temporal dynamics of their dephosphorylation and trafficking patterns. Whereas dephosphorylation and recycling of the sst2 receptor requires extended time periods of ∼30 min, dephosphorylation and recycling of the sst5 receptor is completed in less than 10 min. Here, we examined which receptor domains determine the selection of phosphatases for receptor dephosphorylation. We found that generation of tail-swap mutants between sst2 and sst5 was required and sufficient to reverse the patterns of dephosphorylation and trafficking of these two receptors. In fact, siRNA knockdown confirmed that the sst5 receptor carrying the sst2 tail is predominantly dephosphorylated by PP1β, whereas the sst2 receptor carrying the sst5 tail is predominantly dephosphorylated by PP1γ. Thus, the GPCR phosphatase responsible for dephosphorylation of individual somatostatin receptor subtypes is primarily determined by their different carboxyl-terminal receptor domains. This phosphatase specificity has in turn profound consequences for the dephosphorylation dynamics and trafficking patterns of GPCRs.


Introduction
The signaling output of G protein-coupled receptors (GPCRs) is desensitized by mechanisms involving phosphorylation, b-arrestin binding and internalization. GPCR signaling is resensitized by mechanisms involving dephosphorylation, but details about the phosphatases responsible are generally lacking. We and others have recently succeeded in identifying bona fide GPCR phosphatases for a number of receptors using a combined approach of phosphosite-specific antibodies and siRNA screening in HEK293 cells. First, we identified protein phosphatase 1b (PP1b) as GPCR phosphatase for the sst 2 somatostatin receptor [1]. Second, we identified PP1c as GPCR phosphatase for the m-opioid receptor and the sst 5 somatostatin receptor [2] [3]. Third, more recently Gehret and Hinkle identified PP1a as GPCR phosphatase for the thyrotropin-releasing hormone receptor [4]. All of the above observations were made in a similar cellular background. This suggests that a given GPCR may recruit its specific PP1 isoform for rapid dephosphorylation with remarkable selectivity. However, it is not known which GPCR domain directs the engagement of specific PP1 isoforms to the receptor.
Here, we have addressed this question using the closely-related sst 2 and sst 5 somatostatin receptors. The sst 2 and sst 5 receptors exhibit a high degree of homology in their transmembrane domains but exhibit divergent carboxyl-terminal tails. Both the sst 2 and the sst 5 receptor are pharmacological relevant targets for clinically-used drugs [5] [6] [7] [8] [9] but the two receptors exhibit strikingly different phosphorylation and trafficking patterns. The sst 2 receptor is a prototypical class B receptor that is phosphorylated at a cluster of at least six carboxyl-terminal serine and threonine residues upon agonist exposure. The sst 2 receptor than forms a stable complex with b-arrestin that co-internalize into the same endocytic vesicles. Consequently, the sst 2 receptor recycles slowly [1] [10] [11]. By contrast, the sst 5 receptor is a prototypical class A receptor in that its endocytosis is regulated by a single phosphorylation at T333. The sst 5 receptor then forms relatively unstable ß-arrestin complexes that dissociate at or near the plasma membrane. The receptor internalizes without ßarrestin and recycles rapidly [2] [12]. Here, we show that a tailswap mutation of sst 2 and sst 5 receptors is required and sufficient to reverse the patterns of dephosphorylation and trafficking of these two receptors.

Cell culture and transfection
Human embryonic kidney HEK293 cells were obtained from the German Resource Centre for Biological Material (DSMZ, Braunschweig, Germany). HEK293 cells were grown in DMEM supplemented with 10% fetal calf serum. Cells were transfected with plasmids using Lipofectamine 2000 according to the instructions of the manufacturer (Invitrogen, Carlsbad, CA). Stable transfectants were selected in the presence of 400 mg/ml G418. Stable cells were characterized using radioligand-binding assays, Western blot analysis, and immunocytochemistry as described previously. All receptors and chimeras tested were present at the cell surface, expressed similar amounts of receptor protein and had similar affinities for SS-14 as the wild-type receptors.

Analysis of receptor internalization by confocal microscopy
Cells were grown on poly-L-lysine-coated coverslips overnight. After treatment with 1 mM SS-14 for 0, 15 or 30 min at 37uC, cells were fixed with 4% paraformaldehyde and 0.2% picric acid in phosphate buffer (pH 6.9) for 30 min at room temperature and washed several times. Specimens were permeabilized and then incubated with anti-sst 2A {UMB-1} or anti-sst5 antibody {UMB-4} antibodies followed by Alexa488-conjugated secondary antibodies. Specimens were mounted and examined using a Zeiss LSM510 META laser scanning confocal microscope.

Quantification of receptor internalization by ELISA
Stably transfected HEK293 cells were seeded onto poly-Llysine-treated 24-well plates. The next day, cells were preincubated with 1 mg/ml anti-HA antibody for 2 h at 4uC. After the appropriate treatment with SS-14 (1 mM) for 30 min at 37uC, cells were fixed and incubated with peroxidase-conjugated anti-rabbit antibody overnight. After washing, plates were developed with ABTS solution and analyzed at 405 nm using a microplate reader.
b-Arrestin-EGFP mobilization assay HEK293 cells were seeded onto 35-mm glass-bottom culture dishes (Mattek, Ashland, MA). The next day, cells were transiently cotransfected with 0.2 mg b-arrestin-2-EGFP and 2 mg human or chimeric somatostatin receptor or with a mixture of 0.2 mg barrestin-2-EGFP, 0.8 mg GRK2 and 1.2 mg human/chimeric sst2 receptor per dish containing 200,000 cells using TurboFect TM (Fermentas) according to the instructions of the manufacturer. After 24 h, cells were transferred onto a temperature-controlled microscope stage set at 37uC of a Zeiss LSM510 META laser scanning confocal microscope. Images were collected sequentially using single line excitation at 488 nm with 515-540-nm band pass emission filters. Saturating concentrations of SS-14 (1 mM) were applied directly into the culture medium immediately after the initial image was taken.

Data Analysis
Data were analyzed using GraphPad Prism 4.0 software. Statistical analysis was carried out with Students t-test as well as with one-way or two-way ANOVA followed by the Bonferroni post-test. p-Values of ,0.05 were considered statistically significant.

Results
The sst 2 and sst 5 receptors exhibit a high degree of homology, yet these somatostatin receptors are dephosphorylated by different PP1 isoforms. To elucidate which receptor domains determine this remarkable phosphatase specificity, we first constructed tail-swap mutants of these two receptors. In initial studies, we confirmed that all receptors were expressed at similar levels on the cell surface, and exhibited similar binding properties. All four receptors also exhibited similar signaling properties determined as their ability to activate ERK in a pertussis toxin-sensitive manner (not shown). We then compared agonist-induced phosphorylation of the wild-type sst 2 receptor with that of the sst 5-2 receptor using phosphosite-specific antibodies for pS341/343, pT353/354 and pT356/359 (Figure 1, left panel). Phosphorylation at the three sites was not detectable in untreated cells. In the presence of SS-14 phosphorylation at all three sites became detectable within seconds of agonist exposure in both the sst 2 and the sst 5-2 receptor. Next, we compared agonist-induced phosphorylation of the wild-type sst 5 receptor with that of the sst 2-5 receptor using phosphosite-specific antibodies for pT333 and pT347.
Phosphorylation at T347 was already detectable in untreated cells for both receptors (not shown). By contrast, phosphorylation at T333 was not detectable in untreated cells. However, upon addition of SS-14 T333 phosphorylation occurred within a few seconds in both sst 5 and sst 2-5 receptors (Figure 1, right panel).
The sst 2 receptor and the sst 5 receptor dramatically differ in the extent of their agonist-induced internalization. In the presence of SS-14, nearly all cell surface sst 2 receptors are removed from the plasma membrane resulting in a ,80% loss of surface receptors after 30 min agonist exposure (Figure 2A, B). By contrast, the sst 5 shows only partial receptor internalization with a large proportion of receptors remaining at the plasma membrane resulting in a maximal internalization of ,25% after 30 min SS-14 treatment (Figure 2A, B). Interestingly, swapping the cytoplasmic tails completely reversed this trafficking pattern in that the sst 5-2 receptor revealed nearly complete (,70%) and the sst 2-5 receptor partial (,25%) endocytosis ( Figure 2).
We then employed functional b-arrestin-2 conjugated to enhanced green fluorescent protein (EGFP) to visualize the patterns of b-arrestin mobilization in live HEK293 cells. In the absence of agonist, b-arrestin-2-EGFP was uniformly distributed throughout the cytoplasm of the cells (Figure 3). The addition of saturating concentrations of SS-14 (1 mM) to the human sst 2 receptor induced a rapid redistribution of b-arrestin-2 from the cytoplasm to the plasma membrane resulting in robust fluorescent staining outlining the cell shape ( Figure 3A). Overexpression of GRK2 led to the formation of stable complexes between the sst 2 receptor and b-arrestin-2 that appeared as punctuate staining within the cytoplasm at later time points ( Figure 3B). The addition of 1 mM SS-14 to the human sst 5 receptor induced a redistribution of b-arrestin-2 from the cytoplasm to the plasma membrane that was less pronounced compared to that seen in sst 2 -expressing cells ( Figure 3A). Although over-expression of GRK2 clearly facilitated b-arrestin-2 recruitment in sst 5 -expressing cells at early time points, it did not lead to a redistribution of b-arrestin-2-EGFP into the cytosol at later time points ( Figure 3B). Again, swapping the cytoplasmic tails led to a complete reversal of the b-arrestin trafficking patterns of these two receptors ( Figure 3A, B).
The sst 2 receptor and the sst 5 receptor also dramatically differ in their patterns of dephosphorylation and recycling. Interestingly, dephosphorylation of individual sst 2 phosphate acceptor sites occurs with distinct temporal dynamics. Whereas T353/T354 dephosphorylation occurred rapidly (,5 min), T356/T359 dephosphorylation was delayed (,20 min) and S341/S343 dephosphorylation is only observed after extended SS-14 washout (,60 min) (Figure 4). When sst 5 -expressing cells were exposed to SS-14 for 5 min, washed and then incubated in agonist-free medium, T333 dephosphorylation occurred very rapidly (,2 min) ( Figure 4). In contrast, T347 phosphorylation was although to a lesser extent still detectable even after prolonged incubation in the absence of agonist (not shown). Analysis of chimeric receptors under identical conditions showed that transplantation of the sst 2 tail to the sst 5 receptor led to an sst 2 -like dephosphorylation profile Figure 1. Agonist-induced phosphorylation of sst 2 and sst 5 tail-swap mutants. Top, Schematic representation of the human wild-type sst 2 (depicted in grey) and human wild-type sst 5 receptors (depicted in black) and their corresponding tail-swap mutants. Phosphate acceptor sites targeted for the generation of phosphosite-specific antibodies are depicted as circles. Bottom, stably transfected HEK 293 cells were exposed to 1 mM SS-14 at room temperature for the indicated time periods. Cells were lysed and immunoblotted with the indicated phosphosite-specific antibodies. Blots were then stripped and reprobed with the phosphorylation-independent anti-sst 5  ( Figure 4). Conversely, transplantation of the sst 5 tail to the sst 2 receptor led to an sst 5 -like dephosphorylation profile (Figure 4).
Finally, we examined the PP1 specificity of the sst 2 receptor and the sst 5 receptor as well as their respective tail-swap mutants. To date, three distinct catalytic subunits a, b and c are known for PP1 [13]. To elucidate which of these PP1 isoforms is involved in sst 5 dephosphorylation, we performed siRNA knockdown experiments. As depicted in Figure 5, PP1b knockdown resulted in a robust inhibition of sst 2 dephosphorylation. In contrast, transfection of PP1a or PP1b siRNA did not result in a significant inhibition of sst 2 dephosphorylation ( Figure 5). For the sst 5 receptor only PP1c knockdown resulted in a detectable inhibition of its dephosphorylation, while transfection of PP1a or PP1b siRNA had no effect ( Figure 5). Interestingly, swapping the cytoplasmic tails conferred PP1c specificity to the sst 2 receptor and PP1b specificity to the sst 5 receptor ( Figure 5). These results suggest that PP1 specificity of individual somatostatin receptor subtypes is primarily determined by their different carboxylterminal receptor domains.

Discussion
Although the regulation of agonist-induced phosphorylation and internalization has been studied in detail for many GPCRs, the molecular mechanisms and functional consequences of receptor dephosphorylation are far from understood. We have recently observed that closely related somatostatin receptor subtypes can be dephosphorylated by distinct PP1 isoforms. However, it is not known which GPCR domain directs the engagement of specific PP1 isoforms to the receptor. The major  finding of this study is that carboxyl-terminal regions of different somatostatin receptor subtypes is a major determinants for their PP1 selectivity. This conclusion is based on the observation that transplantation of the sst 2 tail to the sst 5 receptor led to a predominant dephosphorylation by PP1b, whereas transplantation of the sst 5 tail to the sst 2 receptor led to a predominant dephosphorylation by PP1c. Moreover, swapping the cytoplasmic  were transfected with siRNA targeted to PP1a, PP1b, PP1c or non-silencing siRNA control (SCR) for 72 h and then exposed to 1 mM SS-14 for 5 min. Cells were washed three times and then incubated for 0, 15 or 30 min in the absence of agonist. Cells were lysed and immunoblotted with anti-pT356/T359 antibody {0522}. Blots were stripped and reprobed with the phosphorylation-independent anti-sst2 antibody {UMB-1} to confirm equal loading of the gels. (B) HEK 293 cells stably expressing the sst 5 receptor or the sst 2-5 tail-swap mutant were transfected with siRNA targeted to PP1a, PP1b, PP1c or non-silencing siRNA control (SCR) for 72 h and then exposed to 1 mM SS-14 for 5 min. Cells were washed three times and then incubated for 0, 2 or 5 min in the absence of agonist. Cells were lysed and immunoblotted with anti-pT333 {3567} antibody. Blots were stripped and reprobed with the phosphorylation-independent anti-sst5 antibody {UMB-4} to confirm equal loading of the gels. Receptor phosphorylation was quantified and expressed as percentage of maximal phosphorylation in SCR-transfected cells, which was set at 100%. Data correspond to the mean 6 SEM from three independent experiments. Results were analyzed by two-way ANOVA. (C) siRNA knockdown of PP1 was confirmed by Western blot using isoform-specific PP1 antibodies. The positions of molecular mass markers are indicated on the left (in kDa). doi:10.1371/journal.pone.0091526.g005 tails led to a complete reversal of the trafficking profiles of these two receptors.
The remarkable selectivity in the recruitment of specific PP1 catalytic subunits to individual somatostatin receptor subtypes is surprising. PP1 catalytic subunits bind to their regulatory subunits and some substrates in a mutually exclusive manner through a conserved RVxF motif. The three isoforms of the PP1 catalytic subunit share greater than 90% sequence identity, including the regions that interact with the RVxF sequence [13]. However, neither the human sst 2 nor the human sst 5 receptor has a potential PP1-binding motif in its carboxyl-terminal tail suggesting that somatostatin receptors do not bind to PP1 exclusively by the canonical RVxF motif. Instead, association of PP1 may occur directly through a noncanonical interaction or multiple weak interactions or indirectly via one or more regulatory subunits of PP1. Such targeting PP1 subunits are prime candidates to bring phosphatases in proximity to phosphorylated GPCRs. Nevertheless, the identity of the targeting PP1 subunits remains to be elucidated for both sst 2 and sst 5 .
Somatostatin receptor subtypes exhibit strikingly different barrestin trafficking patterns. The sst 2 receptor is a prototypical class B receptor that is phosphorylated at clusters of carboxylterminal serine and threonine residues. In turn the sst 2 receptor forms stable b-arrestin complexes, co-internalizes with b-arrestin and recycles slowly. By contrast, sst 5 is a prototypical class A receptor in that its endocytosis is driven by phosphorylation of a single threonine residue. The sst 5 receptor then forms unstable ßarrestin complexes, internalizes without b-arrestin and recycles rapidly. Thus, our finding that swapping the cytoplasmic tails led not only to reversal of the PP1 specificity but also to a reversal of the b-arrestin trafficking profiles of somatostatin receptors suggests a simple model in which fast recycling class A receptors are preferentially dephosphorylated by PP1c, whereas slow recycling class B receptors are preferentially dephosphorylated by PP1b. So far only few bona fide GPCR phosphatases have been identified [14] [15] [1] [2] [4]. However, it should be noted that this hypothesis is supported by our recent observation that the m-opioid receptor, which is a prototypical fast recycling class A receptor, is rapidly dephosphorylated by PP1c [3]. However, it should be noted that PP1a was identified as GPCR phosphatase for the thyrotropin-releasing hormone receptor [4], suggesting that different phosphatases can interact with different GPCRs to mediate their dephosphorylation. It is also possible that distinct phosphatase activities mediate dephosphorylation of plasma membrane receptors versus internalized receptors.
GPCR dephosphorylation has long been viewed as an unregulated process with little or no functional implications. Nevertheless, more recent evidence suggests that PP1ß-mediated dephosphorylation is involved in fine-tuning unconventional ßarrestin-dependent GPCR signaling. Indeed, inhibition of PP1ß expression results in a specific enhancement of sst 2 -driven ERK activation [1]. Given that arrestin-dependent signaling is initiated by binding to phosphorylated receptors, this finding suggests that PP1ß-mediated GPCR dephosphorylation limits b-arrestin-dependent signaling by disrupting the ß-arrestin-GPCR complex.
In conclusion, different GPCRs can recruit specific PP1 isoforms for their rapid dephosphorylation with remarkable selectivity. This GPCR phosphatase specificity is primarily determined by carboxyl-terminal receptor domains. Recruitment of different GPCR phosphatases has in turn profound consequences for the dephosphorylation dynamics and trafficking patterns of GPCRs.