A Transplantable Phosphorylation Probe for Direct Assessment of G Protein-Coupled Receptor Activation

The newly developed multireceptor somatostatin analogs pasireotide (SOM230), octreotide and somatoprim (DG3173) have primarily been characterized according to their binding profiles. However, their ability to activate individual somatostatin receptor subtypes (sst) has not been directly assessed so far. Here, we transplanted the carboxyl-terminal phosphorylation motif of the sst2 receptor to other somatostatin receptors and assessed receptor activation using a set of three phosphosite-specific antibodies. Our comparative analysis revealed unexpected efficacy profiles for pasireotide, octreotide and somatoprim. Pasireotide was able to activate sst3 and sst5 receptors but was only a partial agonist at the sst2 receptor. Octreotide exhibited potent agonistic properties at the sst2 receptor but produced very little sst5 receptor activation. Like octreotide, somatoprim was a full agonist at the sst2 receptor. Unlike octreotide, somatoprim was also a potent agonist at the sst5 receptor. Together, we propose the application of a phosphorylation probe for direct assessment of G protein-coupled receptor activation and demonstrate its utility in the pharmacological characterization of novel somatostatin analogs.


Introduction
The development of novel multireceptor somatostatin analogs has primarily focused on the discovery of compounds with nanomolar binding affinities to more than one of the five somatostatin receptors (sst 1 -sst 5 ). It is not clear, however, whether these compounds exhibit full or partial agonistic properties at individual somatostatin receptor subtypes. This lack of knowledge is due to the limited availability of methods allowing a direct assessment of G protein-coupled receptor (GPCR) activation.
In clinical practice, octreotide and lanreotide are used as first choice medical treatment of neuroendocrine tumors such as GHsecreting adenomas and carcinoids [1,2]. Octreotide and lanreotide bind with high sub-nanomolar affinity to sst 2 only, have moderate affinity to sst 3 and sst 5 and show very low or absent binding to sst 1 and sst 4 . Recently, the novel multireceptor somatostatin analog, pasireotide (SOM230), has been synthesized [3]. Pasireotide is a cyclohexapeptide, which binds with high affinity to all somatostatin receptors except to sst 4 [4]. In contrast to octreotide, pasireotide exhibits particular high sub-nanomolar affinity to sst 5 [5]. Pasireotide is currently under clinical evaluation for treatment of acromegaly, Cushing's disease and octreotideresistant carcinoid tumors [6,7,8]. In addition to pasireotide, the novel pan-somatostatin analog somatoprim (DG3173) is currently under clinical and preclinical evaluation. Somatoprim exhibits a unique binding profile in that binds with high affinity to sst 2 , sst 4 and sst 5 but not to sst 1 or sst 3 .
We have recently uncovered agonist-selective and speciesspecific patterns of sst 2A receptor phosphorylation and trafficking [9]. Whereas octreotide, in a manner similar to that observed with somatostatin, stimulates the phosphorylation of a number of carboxyl-terminal phosphate acceptor sites in both rat and human sst 2 receptors, pasireotide fails to promote any detectable phosphorylation or internalization of the rat sst 2A receptor. In contrast, pasireotide is able to trigger a partial internalization of the human sst 2 receptor. At present it is unclear whether the agonist-selective regulation of the sst 2 receptor observed for pasireotide is a general property of all pan-somatostatin analogs, and whether such functional selectivity may exist for other clinically-relevant somatostatin receptors including sst 5 and sst 3 .
In the present study, we addressed this problem by using the carboxyl-terminal tail of the sst 2 receptor as transplantable phosphorylation probe to directly sense the activation of other somatostatin receptors. This approach was possible due to our recent success in generating a set of three phosphosite-specific antibodies for the sst 2 receptor which allowed us to determine distinct patterns of phosphorylation induced by different agonists. Our assay utilizes the unique ability of G protein-coupled receptor kinases (GRKs) to detect only active conformations of GPCRs. Different phosphorylation patterns may hence reflect distinct receptor conformations.

Generation of Mutant Somatostatin Receptors
A chimera of the human sst 5 receptor with the carboxylterminal tail of the human sst 2 receptor (hsst5-sst2CT) was generated by DNA synthesis by imaGenes (Berlin, Germany). A chimera of the rat sst 3 receptor with the carboxyl-terminal tail of the rat sst 2 receptor (rsst3-sst2ACT) was generated by exchange of the entire carboxyl-terminal tail using the FLS Motive present in the DNA sequence of both receptors in the seventh transmembrane domain. The fragments were cloned into pcDNA3.1(+) using HindIII and XbaI cloning sites.

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 Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum in a humidified atmosphere containing 10% CO 2 . Cells were transfected with plasmids encoding for wild-type or mutant somatostatin receptors using jetPEI TM according to the instructions of the manufacturer (Invitrogen, Carlsbad, CA). Stable transfectants were selected in the presence of 400 mg/ml G418. HEK293 cells stably expressing somatostatin receptors were characterized using radioligand-binding assays, Western blot analysis, and immunocytochemistry as described previously. The level of somatostatin receptor expression was ,900 fmol/mg membrane protein. All chimeras and mutants tested were present at the cell surface and expressed similar amounts of receptor protein. The IC 50 values of SS-14, octreotide and pasireotide for ligand binding affinities are given in Table S1.

Radioligand Binding Assay
Competition binding assays were performed on membrane preparations from stable transfected cells as described above. Cells were harvested into PBS and stored at -80uC. After thawing, cells were centrifuged at 20,0006g for 10 min at 4uC and then homogenized in lysis buffer (50 mM Tris-HCl, 3 mM EGTA, 5 mM EDTA, pH 7.4). Cell membranes were pelleted  in 96-well polypropylene plates in a final volume of 200 ml for 45 min at room temperature. Specific binding was calculated by subtracting non-specific binding -defined as that seen in the presence of 1 mM SS-14, octreotide, pasireotide or somatoprimfrom total binding obtained with radioligand alone. The incubation was terminated by addition of ice-cold buffer and rapid vacuum filtration through glass fiber filters presoaked in 0.3% polyethyleneimine using an Inotech cell harvester (Dittikon, Switzerland). Filters were rinsed twice with washing buffer and air-dried. Bound radioactivity was determined using a c-counter (COBRAII, Packard, USA). Data from ligand binding and IC 50 were analyzed by curve fitting using GraphPad Prism 4.0 software [15].

GTPcS Binding Assays
Cells were harvested and lysed as described above except that a lysis buffer containing 50 mM Tris, 10 mM EDTA and 1 mM EGTA (pH 7.4) was used. The resulting pellet was resuspended in assay buffer (20 mM HEPES, 100 mM NaCl, 10 mM MgCl 2 , pH 7.4). Aliquots containing 30 mg of protein were incubated with 3 mM GDP and 0.05 nM [ 35 S]GTPcS (Specific activity -43.3 TBq/mmol, PerkinElmer USA) in the presence or absence of either SS-14, octreotide, pasireotide or somatoprim in concentrations ranging from 10 212 to 10 26 M. Assays were carried out in a final volume of 1 ml for 30 min at 30uC under continuous agitation. Nonspecific binding was determined in the presence of 10 mM unlabeled GTPcS. The incubation was terminated by the addtion of ice-cold buffer and rapid vacuum filtration through glass fiber filters as described above. Filters were rinsed twice with washing buffer (50 mM Tris-HCl, pH 7.4) and dried. A scintillation mixture was added, and radioactivity was determined unsing a b-counter (1600 TR, Packard, USA) [15].

Results
We have recently shown that pasireotide exhibits partial agonistic properties at the sst 2 receptor. While binding with high affinity it triggers only a partial internalization of the human sst 2 receptor. To test the possibility that this behavior would be a general property of all multireceptor somatostatin analogs, we evaluated the internalization profile of somatoprim in comparison to pasireotide and octreotide. First, we examined human sst 2 , sst 3 and sst 5 receptors expressed in HEK293 cells by confocal microscopy revealing that in the absence of agonist all three somatostatin receptor subtypes were almost exclusively confined to the plasma membrane (Figure 1, left panel). As shown in Figure 1 (upper panel), octreotide and somatoprim were able to stimulate a robust endocytosis of human sst 2 receptors similar to that seen after SS-14 exposure. In contrast, a saturating concentration of pasireotide induced only a very limited internalization of sst 2 receptors. Conversely, examination of sst 3 -expressing cells revealed that pasireotide promoted a more pronounced receptor sequestration than octreotide, whereas somatoprim failed to stimulate any detectable sst 3 internalization (Figure 1, middle panel). When sst 5 -expressing cells were exposed to SS-14, octreotide, pasireotide or somatoprim only the endogenous ligand SS-14 was able to stimulate a clearly detectable receptor endocytosis (Figure 1,  lower panel).
Recently, we have generated a set of three phosphosite-specific antibodies, which allowed us to detect selectively the S341/S343-, the T353/T354-and the T356/T359-phosphorylated forms of the sst 2 receptor [16,17,18,19]. When HEK293 cells stably expressing the human sst 2 receptor were exposed for 5 min to SS-14, octreotide, pasireotide or somatoprim in concentrations ranging from 10 212 to 10 25 M, SS-14, octreotide and somatoprim were able to promote a robust dose-dependent phosphorylation of all three sites (Figure 2). In contrast, pasireotide stimulated only at saturating concentration a detectable phosphorylation of S341/ S343 and T356/T359 but not of T353/T354. Considering the high binding affinity of pasireotide this result was unexpected and indicates that, in contrast to octreotide and somatoprim, pasireotide is a partial agonist at the human sst 2 receptor. It also shows that the sst 2 receptor can exist in distinct active conformations, which favor different patterns of GRK-mediated phosphorylation. Thus, considerable differences may exist between the binding and efficacy profiles of pan-somatostatin analogs. It would therefore be desirable to know the patterns of phosphorylation induced by multireceptor ligands at the level of individual somatostatin receptors. However, at present phosphosite-specific antibodies are only available for the sst 2 receptor.
We therefore elucidated whether the carboxyl-terminal tail of the sst 2 receptor can be used as probe to sense the activation of other somatostatin receptors. Consequently, we transplanted the carboxyl-terminal phosphorylation motif of the sst 2 receptor to other clinically-relevant somatostatin receptors and assessed their patterns of activation using our set of three phosphosite-specific antibodies. Examination of cells expressing a sst 5 -sst 2 CT chimeric receptor revealed that SS-14 stimulated the most pronounced phosphorylation of all three sites (Figure 3 A). Pasireotide and somatoprim also promoted a robust phosphorylation of S341/ S343 and T356/T359 (Figure 3 A). In contrast, octreotide induced only at saturating concentration a detectable phosphorylation of S341/S343 and T356/T359 (Figure 3 A). Interestingly, similar results were obtained at the wild-type sst 5 receptor using a recently generated phosphosite-specific antibody to T333 validating our approach to study receptor activation (Figure 3 B). Analysis of ligand binding properties of the sst 5 -sst 2 CT chimera indicated that the transfer of the sst 2 carboxyl-terminal tail to sst 5 did not substantially affect the affinities for SS-14, octreotide, pasireotide or somatoprim (Table 1).
We then examined the capacity of these compounds to stimulate GTPcS binding in membrane preparations from the same cells (Figure 4). Unlike that seen in sst 2 receptor phosphorylation assays, pasireotide was able stimulate GTPcS binding to a similar degree as octreotide or somatoprim suggesting that pasireotide is a G protein-biased ligand. In contrast, octreotide stimulated GTPcS binding in both sst 5 -and sst 5 -sst 2 CT-expressing cells to a much lesser extend than pasireotide or somatoprim suggesting it is indeed a weak partial agonist at the sst 5 receptor. Again similar  results were obtained with the wild-type sst 5 and the sst 5 -sst 2 CT receptor.
To elucidate whether our approach can be used to directly assess the activation of a wide variety of G protein-coupled receptors, we next examined a sst 3 -sst 2 CT chimera. Examination of HEK293 cells stably expressing sst 3 -sst 2 CT receptors revealed that only pasireotide but not octreotide was able to promote a robust phosphorylation of S341/S343 and T356/T359 ( Figure 5). In contrast, somatoprim failed to induce any detectable phosphorylation. Pasireotide is less potent than octreotide in inducing internalization of the sst 2 receptor but more potent than octreotide in inducing internalization of the sst 3 receptor. Thus, the patterns of phosphorylation of the rsst 3 -sst 2 CT chimera correlates very well with pattern internalization of the wild-type sst 3 receptor (Figure 1). Nevertheless, it should be noted that these results were obtained with rat sst 3 -sst 2 CT receptor construct. Given the recently observed species differences for the sst 2 receptor, these results need to be reproduced with the human sst 3 receptor.

Discussion
The development of new drugs targeting GPCRs is primarily focused on the discovery of compounds with nanomolar and subnanomolar binding affinities. Then indirect methods mostly assessing G protein signaling are being used to determine whether a new compound is a full or partial agonist. Accumulating evidence suggests that more than one active conformation exists for many GPCRs and that many compounds selectively stimulate specific signaling pathways [20]. Thus, there is clearly a need for methods providing more direct information on receptor activation. However, structural information is only available for a few activated receptors, and none of these has been crystallized in more than one active conformation yet [21,22]. Determination of receptor activation using biophysical methods requires insertion of bulky fluorescent proteins into the receptor which may itself affect receptor activation [23].
In the present study, we have used the phosphorylation motif of the sst 2 receptor to probe GPCR activation. This approach was possible due to our recent success in generating a set of three phosphosite-specific antibodies for the sst 2 receptor which allowed us to determine distinct patterns of phosphorylation induced by different agonists. Given the unique ability of GRKs to detect only active GPCRs these distinct conformations may reflect different receptor conformations. However, phosphosite-specific antibodies are notoriously difficult to generate and are only available for a few receptors. We therefore elucidated whether the carboxyl-terminal tail of the sst 2 receptor can be used as probe to sense the activation of other somatostatin receptors. Perhaps the most convincing evidence that this might be a valid and useful approach comes from a sst 5 -sst 2 CT chimera. In fact, the results obtained with the sst 5 -sst 2 CT chimeric receptor and the wild-type sst 5 receptor were very similar. We have also confirmed that insertion of the sst 2 phosphorylation motif into other somatostatin receptors did not dramatically change their binding properties with regard to the compounds tested. In addition, construction of a sst 3 -sst 2 CT chimera was also successful indicating that this approach could be used to examine the activation of a wide variety of GPCRs. Our assay can be adapted to a quantitative ELISA method and thus be applied to screening of large numbers of ligands [24]. However, for other receptors it cannot be completely ruled out that transplantation of the sst 2 CT may alter receptor function. Therefore, a functional analysis of such chimeric receptors needs to be performed in each case.
Our study also yielded valuable and previously unappreciated information about the pan-somatostatin analogs currently under clinical and preclinical examination. Pasireotide exhibited potent agonistic activity at the sst 5 receptor but only weak partial agonistic properties at the sst 2 receptor. Consequently, pasireotide should be classified as sst 5 -preferring ligand. Octreotide is a full agonist at the sst 2 receptor but exhibited virtually no agonistic activity at the sst 5 receptor. Consequently, octreotide should be classified as sst 2 -preferring ligand. In contrast, somatoprim is unique in that it was a potent agonist at both sst 2 and sst 5 receptors. In fact, this may provide the molecular basis for the recent observation that somatoprim can inhibit GH release in cases, which did not respond to octreotide [25].
In conclusion, we describe the use of a phosphorylation probe for direct assessment of GPCR activation and demonstrate its utility in the pharmacological characterization of novel pansomatostatin analogs.

Supporting Information
Table S1 Ligand binding properties of rat, human and mutant somatostatin receptors. Ligand binding assays were carried out as described under ''Materials and Methods''. The half-maximal inhibitory concentrations (IC 50 ) were analyzed by nonlinear regression curve fitting using the computer program GraphPad Prism. Data are presented as the mean of three independent experiments performed in triplicate. (DOC)