Agonist-biased Trafficking of Somatostatin Receptor 2A in Enteric Neurons*

Background: Somatostatin regulates gut function via neuronal somatostatin receptors. Results: Somatostatin susceptibility to degradation by endosomal endothelin-converting enzyme 1 (ECE-1) defines receptor function. Conclusion: ECE-1 regulates the duration of somatostatin receptor signaling and trafficking. Significance: Therapeutic somatostatin analogs are ECE-1-resistant, which underlies their prolonged actions. Somatostatin (SST) 14 and SST 28 activate somatostatin 2A receptors (SSTR2A) on enteric neurons to control gut functions. SST analogs are treatments of neuroendocrine and bleeding disorders, cancer, and diarrhea, with gastrointestinal side effects of constipation, abdominal pain, and nausea. How endogenous agonists and drugs differentially regulate neuronal SSTR2A is unexplored. We evaluated SSTR2A trafficking in murine myenteric neurons and neuroendocrine AtT-20 cells by microscopy and determined whether agonist degradation by endosomal endothelin-converting enzyme 1 (ECE-1) controls SSTR2A trafficking and association with β-arrestins, key regulators of receptors. SST-14, SST-28, and peptide analogs (octreotide, lanreotide, and vapreotide) stimulated clathrin- and dynamin-mediated internalization of SSTR2A, which colocalized with ECE-1 in endosomes and the Golgi. After incubation with SST-14, SSTR2A recycled to the plasma membrane, which required active ECE-1 and an intact Golgi. SSTR2A activated by SST-28, octreotide, lanreotide, or vapreotide was retained within the Golgi and did not recycle. Although ECE-1 rapidly degraded SST-14, SST-28 was resistant to degradation, and ECE-1 did not degrade SST analogs. SST-14 and SST-28 induced transient interactions between SSTR2A and β-arrestins that were stabilized by an ECE-1 inhibitor. Octreotide induced sustained SSTR2A/β-arrestin interactions that were not regulated by ECE-1. Thus, when activated by SST-14, SSTR2A internalizes and recycles via the Golgi, which requires ECE-1 degradation of SST-14 and receptor dissociation from β-arrestins. After activation by ECE-1-resistant SST-28 and analogs, SSTR2A remains in endosomes because of sustained β-arrestin interactions. Therapeutic SST analogs are ECE-1-resistant and retain SSTR2A in endosomes, which may explain their long-lasting actions.

G protein-coupled receptors (GPCRs) 4 for peptide hormones and neurotransmitters control diverse pathophysiological processes and are a major therapeutic target. Diverse endogenous agonists and drugs can induce distinct conformations of the same receptor molecule, leading to divergent signaling and physiological outcomes. This "agonist-biased" signaling has been mostly examined in model systems treated with exogenous agonists. The ability of diverse endogenous peptides to differentially regulate receptors in the nervous system is largely unexplored.
GPCR signaling must be tightly controlled because dysregulation can cause disease. Signaling by bioactive peptides is controlled by mechanisms that regulate the activity of peptides and their receptors. Cell surface endopeptidases (e.g. neprilysin) degrade and inactivate peptides in the extracellular fluid and, thereby, attenuate their effects (1). ␤-arrestins interact with agonist-occupied receptors that uncouple receptors from heterotrimeric G proteins and terminate plasma membrane signaling. By recruiting receptors and signaling partners to endosomal signalosomes, ␤-arrestins also transmit signals from internalized receptors (2). Endosomal peptidases (e.g. endothelin-converting enzyme 1, ECE-1) degrade neuropeptides in endosomes, which disassembles signalosomes, terminates endosomal signaling, and promotes receptor recycling and resensitization of plasma membrane signaling (3). These processes are key regulators of peptide signaling. However, the capacity of these mechanisms to differentially regulate signaling by different endogenous or exogenous agonists of the same GPCR has not been examined.
We evaluated how endogenous and exogenous agonists differentially regulate somatostatin (SST) receptor 2A (SSTR2A). The cyclic peptides SST-14 and N-terminally extended SST-28 are endogenous agonists of SSTR2A that are produced by enteric neurons and enteroendocrine cells (4,5). SSTR2A is a major mediator of the pan-inhibitory actions of SST in the gastrointestinal tract. SSTR2A is prominently localized to enteric neurons (5,6), where activation stimulates K ϩ channels, leading to hyperpolarization and suppression of transit and secretion (7,8). SSTR2A is also an important therapeutic target. Metabolically stable SST analogs (e.g. octreotide) are effective treatments for acromegaly, diarrhea, bleeding disorders, and cancer, and radiolabeled agonists can be used to detect neuroendocrine tumors (4). However, these drugs have side effects related to their actions on the gastrointestinal tract, including constipation, abdominal cramps, and nausea (4).
SST-14 induces SSTR2A endocytosis, which depends on sustained interactions with ␤-arrestins, followed by recycling, which requires ECE-1-dependent SST-14 degradation in endosomes (9). However, the regulation of SSTR2A by different endogenous peptides and therapeutic agonists in functionally relevant cells has not been studied. By examining SSTR2A trafficking in primary enteric neurons and model cell lines, we report that endogenous and exogenous agonists, including clinically used analogs, differentially regulate SSTR2A. Differential regulation depends on agonist susceptibility to degradation by the endosomal peptidase ECE-1, which associates with SSTR2A and governs the duration of SSTR2A and ␤-arrestin interactions. Thus, although resistance to extracellular proteolysis is important for development of SST analogs, resistance to intracellular proteolysis is a key determinant of their longlasting therapeutic and side effects.
Statistical Analysis-Data are expressed as mean Ϯ S.E. and were analyzed using Student's t test or one-way analysis of variance and Dunnett's multiple comparison test. p Ͻ 0.05 was considered significant.

JOURNAL OF BIOLOGICAL CHEMISTRY 25691
To evaluate the mechanism of endocytosis, we treated neurons with hypertonic sucrose or dynasore, which, respectively, block clathrin-and dynamin-dependent endocytosis (16). Both treatments inhibited SSTR2A endocytosis stimulated by SST-14 (not shown) and octreotide (Fig. 3, A and B). Sucrose or dynasore had no effect on the subcellular localization of SSTR2A in unstimulated neurons (not shown). Thus, clathrin and dynamin mediate agonist-stimulated endocytosis of SSTR2A in myenteric neurons.
SSTR2A Recycling Requires an Intact Golgi Apparatus in Myenteric Neurons-Although SSTR2A is endocytosed to the trans-Golgi network (TGN) in cell lines and hippocampal neurons (18 -20), the importance of Golgi trafficking in enteric neurons is not known. After stimulation with SST-14, SST-28 (not shown), or octreotide (Fig. 5A), SSTR2A-IR trafficked to juxtanuclear structures that labeled with an antibody to the TGN (TGN-38). In cells treated with brefeldin A, the TGN was disrupted, and internalized SSTR2A-IR was not colocalized with TGN-38 (Fig. 5B). Brefeldin A did not affect SST-14-stimulated endocytosis of SSTR2A-IR but prevented recycling of SSTR2A-IR (Fig. 5, C and D; percent SSTR2A-IR at plasma membrane at 120 min, brefeldin A 52 Ϯ 2%, vehicle 75 Ϯ 4%, p Ͻ 0.0001). Thus, activated SSTR2A traffics to the TGN, and an intact TGN is necessary for receptor recycling in myenteric neurons.

SSTR2A Interacts with ECE-1 at the Cell Surface and in
Endosomes-The interaction between SSTR2A and the ECE-1a or ECE-1d isoforms was investigated using BRET. The SSTR2A-RLuc8 construct was coexpressed with YFP-tagged ECE-1a or ECE-1d in HEK293 cells, and the transfer of energy between these two proteins was measured upon addition of SST-14, SST-28, or octreotide (100 nM each). The BRET signal between ECE-1a and SSTR2A-RLuc8 decreased upon addition of all three ligands, suggesting an increase in the distance between the energy donor protein and the acceptor (Fig. 7A, p Ͻ 0.001 for 6 -20 min). In contrast, the BRET signal between ECE-1d and SSTR2A-RLuc8 increased upon addition of SST-14, SST-28, and octreotide, indicating an increased proximity between SSTR2A and ECE-1d (Fig. 7B, p Ͻ 0.01 for 9 -20 min). These results are in agreement with the observation that ECE-1a is primarily located at the plasma membrane. Thus, agonist-stimulated receptor internalization would be expected to reduce the SSTR2A-RLuc8 and ECE-1a-YFP BRET signal. Because ECE-1d is mainly located in endosomes, our results suggest that internalized SSTR2A traffics to endosomes, increasing the SSTR2A-RLuc8/ECE-1d-YFP BRET signal.
SST-28, Lanreotide, and Vapreotide Are Resistant to Degradation by ECE-1-ECE-1 degrades SST-14 but not octreotide at acidic endosomal pH. However, the kinetics and sites of hydrolysis have not been defined, and it is not known whether N-terminally extended SST-28 and peptide analogs are protected from degradation. To determine whether susceptibility to hydrolysis by ECE-1 could explain the differential effects of peptides on SSTR2A recycling, we compared peptide degradation by this enzyme. When SST-14 (250 M) was incubated with ECE-1 (200 nM) at endosomal pH 5.5, there was complete degradation after 120 min at 37°C (3.6 Ϯ 0.5% intact). Under the same conditions, SST-28 was degraded more slowly (54.9 Ϯ 3.0% intact, 120 min), and lanreotide and vapreotide were not degraded (lanreotide 100.4 Ϯ 1.4% intact, vapreotide 102.6 Ϯ 3.6% intact, 240 min; data not shown).
To define the kinetics and sequence of SST-14 and SST-28 hydrolysis, we monitored the appearance of products using the activity-based mass spectrometry proteinase activity labeling employing 18 O-enriched water (12). Newly formed cleavage products were identified by the incorporation of 18 O atoms into nascent C-terminal carboxylic groups, allowing for precise determination of the sequence of degradation. When SST-14 (100 M) was incubated with ECE-1 (100 nM) at endosomal pH 5.5, products were formed that were consistent with initial hydrolysis at Thr 10 -Phe 11 , followed by cleavage at Phe 6 -Phe 7 and then Asn 5 -Phe 6 (Fig. 8, A-C). ECE-1 cleaved SST-28 at the corresponding residues Thr 24 -Phe 25 and then Phe 20 -Phe 21 and   SEPTEMBER 6, 2013 • VOLUME 288 • NUMBER 36 Asn 19 -Phe 20 but at a much reduced rate compared with SST-14 (Fig. 8, D-G). Because all cleavage sites reside within the intramolecular disulfide bond, the cleavage products remained linked after the initial hydrolytic step, which resulted in a gain in mass of 18 Da corresponding to water incorporation. Reduction of disulfide bonds prior to analysis did not alter cleavage sites, but products were, as expected, not further linked. At extracellular pH 7.4, the same proteolytic processing of SST-14 occurred, but at a much reduced rate, and SST-28 processing at pH 7.4 was negligible (Fig. 8, B and F).

ECE-1 Regulates SSTR2A Trafficking in Neurons
Thus, although ECE-1 can rapidly degrade SST-14 at endosomal acidity, SST-28 is protected, and octapeptide SST analogs are completely resistant. These differences likely account for the differential effects of these agonists on SSTR2A recycling.
To assess the stability of interactions between SSTR2A and ␤-arrestin2, we either examined the decay in BRET signal in the continued presence of agonist over 10 min or measured the BRET signal at various times after agonist exposure and recovery of cells in agonist-free medium. In the presence of SST-14 or SST-28 (100 nM), the SSTR2A/␤-arrestin2 interaction declined within 5 min (Fig. 9D, left and center panels), whereas in the presence of octreotide (100 nM), the SSTR2A/␤-arrestin2 interaction was sustained for at least 10 min (right panel). The SSTR2A/␤-arrestin2 interaction declined to basal levels 50 min after agonist removal (SST-14, 0.033 Ϯ 0.004; SST-28 0.023 Ϯ 0.002, Fig. 9E, left and center panels). However, the BRET signal was sustained in octreotide-treated cells for at least 130 min after washing (0.184 Ϯ 0.015 at 50 min) (Fig. 9E, right panel). The ECE-1 inhibitor SM-19712 prevented the decay in BRET signal during incubation with SST-14 and SST-28 and attenuated the decline after agonist removal. SM-19712 did not affect the BRET signal during or after incubation with octreotide. Thus, by degrading internalized SST-14 and SST-28, ECE-1 controls the duration of SSTR2A and ␤-arrestin interactions. The sustained SSTR2A and ␤-arrestin interactions in octreotidetreated cells depends on the resistance of octreotide to degradation by ECE-1.

DISCUSSION
We report that diverse endogenous and therapeutic agonists differentially regulate SSTR2A in myenteric neurons. Although all agonists stimulate clathrin-and dynamin-dependent SSTR2A endocytosis and trafficking to the Golgi apparatus, they evoke distinct patterns of postendocytic receptor sorting. When activated by SST-14, SSTR2A transiently interacts with ␤-arrestins and rapidly recycles, but when activated by peptide analogs, SSTR2A shows sustained interactions with ␤-arrestins (peptide analogs) and recycles slowly (SST-28) or not at all (peptide analogs). These differences depend on agonist susceptibility to degradation by ECE-1. By degrading SST-14 in endosomes or the Golgi apparatus, ECE-1 promotes SSTR2A dissociation from ␤-arrestins and rapid receptor recycling. Agonists that are resistant to ECE-1 degradation promote stable interactions between SSTR2A and ␤-arrestins, which lead to intracellular retention of the receptor. This differential regulation of a key pan-inhibitory receptor may explain different physiological effects of endogenous agonists and could account for the longlasting therapeutic actions and side effects of clinically used agonists.
Agonist-biased Trafficking of SSTR2A-"Biased agonism," in which different agonists induce distinct conformations of the same receptor that lead to divergent signals and physiological FIGURE 7. SSTR2A interacts with the ECE-1a and ECE-1d isoforms. SSTR2A-RLuc8 and ECE-1a-YFP or ECE-1d-YFP were coexpressed in HEK cells. A, stimulation with SST-14, SST-28 and octreotide (100 nM) resulted in a significant reduction in BRET signal between SSTR2A-RLuc8 and the plasma membraneassociated ECE-1a-YFP isoform, indicating a loss of interaction between these two proteins. B, in contrast, agonist stimulation gave rise to a significant increase in BRET signal between SSTR2A-RLuc8 and the endosome-associated ECE-1d-YFP isoform, indicating increased interaction. **, p Ͻ 0.01; ***, p Ͻ 0.001.

ECE-1 Regulates SSTR2A Trafficking in Neurons
outcomes, is an emerging theme of GPCR activation that has been examined extensively for synthetic agonists of receptors expressed in model cell lines. By studying the effects of six endogenous and synthetic agonists of SSTR2A, we observed agonist-biased endocytosis and recycling of a GPCR in primary neurons.
Although all agonists stimulated endocytosis of SSTR2A, octreotide, and related octapeptide derivatives induced a greater degree of endocytosis than SST-14, SST-28, and L-054,264. One possible explanation for this difference is that the octapeptides are more potent agonists, although these agonists all activate SSTR2A with similar potencies (23,24). Alternatively, the octapeptide analogs may stabilize conformations that promote more sustained interactions of SSTR2A with ␤-arrestins, which couple GPCRs to clathrin and AP2 and, thereby, favor receptor endocytosis. Additional experiments are required to assess these possibilities.
There were marked differences in the postendocytic sorting of SSTR2A in neurons exposed to different agonists. Although SST-14 and L-054,264 induced transient internalization of SSTR2A and complete recycling to the plasma membrane within 60 -120 min, SST-28, octreotide, lanreotide, and vapreotide caused sustained SSTR2A retention in endosomes and the Golgi apparatus. This difference can be explained by the sensitivity of the different agonists to degradation by ECE-1. We observed that ECE-1 rapidly degrades SST-14 at endosomal acidity by cleaving at three sites within the dicysteine linked region, which would inactivate the peptide. An ECE-1-selective inhibitor or suppression of endosomal acidification (and, thus, ECE-1 activity) prevented SST-14-induced recycling of SSTR2A. As expected, ECE-1 inhibition did not affect SSTR2A recycling after treatment with L-054,264, a non-peptide that would not be degraded by ECE-1. ECE-1 degraded SST-28 at corresponding sites, but degradation was delayed, which indicates that the N-terminal extension protects the bioactive peptide from proteolytic inactivation. SST-28-activated SSTR2A recycled slowly in AtT-20 cells but not at all in neurons. The octapeptide analogs were completely resistant to ECE-1 degradation, and these caused sustained retention of SSTR2A in endosomes and the Golgi apparatus. These findings are consistent with our previous observations that ECE-1 degrades substance P, calcitonin gene-related peptide, and SST-14 in endosomes to promote receptor recycling and resensitization (3,9,13,25).
Our results demonstrate, for the first time, the close interaction between a GPCR and ECE-1. Moreover, we show that susceptibility of agonists to degradation by ECE-1 regulates the duration of interactions between SSTR2A and ␤-arrestins and that this process controls the rate of receptor recycling. The affinity of GPCR interactions with ␤-arrestins is a critical determinant of GPCR recycling. Most studies indicate that this interaction is an intrinsic property of the receptor that depends on the extent of phosphorylation of intracellular Ser and Thr residues by G protein-coupled receptor kinases that are required for ␤-arrestin binding. "Class A" GPCRs (e.g. ␤ 2 adrenergic receptor, neurokinin 3 receptor) have few phosphorylation sites, exhibit low affinity and transient interactions with ␤-ar-restin2, and rapidly recycle (26 -28). "Class B" GPCRs (e.g. neurokinin 1 receptor, neurotensin 1 receptor) are highly phosphorylated, interact with both ␤-arrestin1 and 2 with high affinity for prolonged periods in endosomes, and slowly recycle. We report that the susceptibility of the agonist to degradation by ECE-1 in endosomes is a second determinant of GPCR interactions with ␤-arrestins and the rate of receptor recycling. By using BRET to assess the duration of interaction between SSTR2A and ␤-arrestins, we found that although SST-14 and SST-28 cause transient interactions, metabolically stable octreotide causes a sustained interaction. The transient interaction after treatment with SST-14 and SST-28 was prolonged by an ECE-1 inhibitor, whereas the sustained interaction after octreotide was unaffected by ECE-1 inhibition. With time, the interaction between SSTR2A and ␤-arrestins waned even in octreotide-and ECE-1 inhibitor-treated cells, suggesting that additional mechanisms may also drive SSTR2A/␤-arrestin dissociation. These may include degradation of the peptide or receptor by other proteases or endosomal acidification that promotes ligand dissociation. Endosomal ECE-1, by cleaving SST-14 and SST-28, presumably accelerates this dissociation, freeing SSTR2A from ␤-arrestins and promoting receptor recycling.
Functional Relevance of Agonist-biased Trafficking of SSTR2A-GPCR trafficking and signaling are inextricably linked. In addi-tion to G protein-mediated signaling at the plasma membrane, internalized GPCRs can continue to signal by G protein-independent mechanisms (2). By uncoupling receptors from G proteins, ␤-arrestins terminate plasma membrane signaling. ␤-arrestins also recruit GPCRs and down-stream signaling partners, notably mitogen-activated protein kinases, to endosomes and, thereby, mediate sustained signals from internalized receptors. Dissociation from ␤-arrestins and recycling are then required for resensitization and maintenance of plasma membrane signaling. Our results show that the nature of the agonist and its susceptibility to degradation by intracellular ECE-1 determine the stability of SSTR2A/␤-arrestin interactions and, thus, the rate of SSTR2A recycling, with likely consequences for SSTR2A signaling. Those agonists that promote rapid SSTR2A recycling (SST-14, L-054,264) would facilitate a rapid resensitization of plasma membrane but transient intracellular signaling. Conversely, agonists that induce sustained retention of SSTR2A in endosomes and the Golgi apparatus (SST-28, peptide analogs) would favor sustained intracellular signaling and a prolonged suppression of plasma membrane signaling. Further studies are required to define the importance of agonist-dependent SSTR2A signaling from the plasma membrane and endosomes of enteric neurons. Although many GPCRs can signal from endosomes, GPCR signaling from the Golgi apparatus has not been examined in detail.
Octreotide is used clinically to treat endocrine tumors and the associated effects of altered pituitary hormone release. It is also used to inhibit the release of peptide hormones from the gut and endocrine pancreas. Octreotide is remarkably resistant to degradation by both extracellular and intracellular peptidases (29), a contributing factor to its therapeutic effectiveness. Our observation that clinically useful SSTR2A agonists are highly resistant to ECE-1 suggests that design of SSTR2A agonists with reduced ECE-1 susceptibility may lead to the development of better SSTR2A ligands for use in diagnostic imaging or treatment of neuroendocrine tumors (30) or controlling disorders such as acromegaly (31). Alternatively, agonists that are readily degraded by ECE-1 will enable more rapid receptor recycling and resensitization. We have also demonstrated clear differences in receptor recycling following SST-14 and SST-28 treatment that are attributable to their susceptibility to degradation by ECE-1. This has functional implications because SST-28 may have more prolonged actions or bias toward endosomal signaling (32).
In conclusion, this study identifies ECE-1 as a major regulator of recycling of internalized SSTR2A in myenteric neurons. Receptor recycling is highly dependent on the susceptibility of agonists to degradation by ECE-1, and recycling correlates with the duration of interaction between SSTR2A and ␤-arrestins. Therapeutic SST analogs are highly resistant to degradation by ECE-1, which may explain their long-lasting actions.