Endothelin-converting Enzyme 1 and β-Arrestins Exert Spatiotemporal Control of Substance P-induced Inflammatory Signals*

Background: Agonists stimulate the subcellular redistribution of receptors. The importance of trafficking for pathophysiological processes is unknown. Results: Substance P inflammatory signaling required receptor interaction with β-arrestins and endothelin-converting enzyme, which mediated endocytosis and recycling. Conclusion: Sustained substance P inflammatory signaling requires receptor recycling. Significance: Mechanisms that maintain receptors at the cell surface are necessary for sustained signaling by extracellular agonists. Although the intracellular trafficking of G protein-coupled receptors controls specific signaling events, it is unclear how the spatiotemporal control of signaling contributes to complex pathophysiological processes such as inflammation. By using bioluminescence resonance energy transfer and superresolution microscopy, we found that substance P (SP) induces the association of the neurokinin 1 receptor (NK1R) with two classes of proteins that regulate SP signaling from plasma and endosomal membranes: the scaffolding proteins β-arrestin (βARRs) 1 and 2 and the transmembrane metallopeptidases ECE-1c and ECE-1d. In HEK293 cells and non-transformed human colonocytes, we observed that G protein-coupled receptor kinase 2 and βARR1/2 terminate plasma membrane Ca2+ signaling and initiate receptor trafficking to endosomes that is necessary for sustained activation of ERKs in the nucleus. βARRs deliver the SP-NK1R endosomes, where ECE-1 associates with the complex, degrades SP, and allows the NK1R, freed from βARRs, to recycle. Thus, both ECE-1 and βARRs mediate the resensitization of NK1R Ca2+ signaling at the plasma membrane. Sustained exposure of colonocytes to SP activates NF-κB and stimulates IL-8 secretion. This proinflammatory signaling is unaffected by inhibition of the endosomal ERK pathway but is suppressed by ECE-1 inhibition or βARR2 knockdown. Inhibition of protein phosphatase 2A, which also contributes to sustained NK1R signaling at the plasma membrane, similarly attenuates IL-8 secretion. Thus, the primary function of βARRs and ECE-1 in SP-dependent inflammatory signaling is to promote resensitization, which allows the sustained NK1R signaling from the plasma membrane that drives inflammation.

Alterations in the subcellular distribution of G protein-coupled receptors (GPCRs) 4 are critically important for signal transduction. Receptors at the plasma membrane are necessary for detection of extracellular stimuli and for plasma membranedelimited signaling events, including coupling to heterotrimeric G proteins, activation of second messengers, and transactivation of receptor tyrosine kinases. Activated receptors translocate to endosomes, where they can signal by G proteindependent and -independent mechanisms (1,2). The movement of receptors back to the cell surface enables the recovery of plasma membrane signaling, whereas receptor trafficking to lysosomes and degradation terminates their ability to signal. The relevance of receptor trafficking to complex pathophysiological processes, such as inflammation, is unknown.
The interaction of GPCRs with ␤-arrestins (␤ARRs) at plasma and endosomal membranes regulates receptor signaling and trafficking (3). At the plasma membrane, G protein-coupled receptor kinases (GRKs) can phosphorylate activated receptors, thereby increasing receptor affinity for ␤ARRs. ␤ARRs disrupt receptor interaction with G proteins, which desensitizes signaling, and couple receptors to clathrin and AP2, which mediates receptor endocytosis. By recruiting downstream signaling proteins such as ERK to endosomes, ␤ARRs also transduce G protein-independent but ␤ARR-dependent signaling of internalized receptors (2). The dissociation of receptors from ␤ARRs is required for receptor recycling and for resensitization of plasma membrane signaling (4).
The availability of agonists at plasma and endosomal membranes also regulates receptor trafficking and signaling. In the case of receptors for neuropeptides, membrane metallopeptidases control the interaction of GPCRs with their agonists at the cell surface and in endosomes. The cell surface peptidase neprilysin degrades the proinflammatory neuropeptide substance P (SP) in the extracellular fluid, which restricts SP interaction with the neurokinin 1 receptor (NK 1 R) and attenuates SP-induced inflammation in multiple tissues (5,6). The endosomal peptidase endothelin-converting enzyme 1 (ECE-1) degrades SP in acidified early endosomes, which triggers dissociation of the SP-NK 1 R-␤ARR complex. This process induces NK 1 R recycling and resensitization of G protein-dependent plasma membrane signaling (7-9) but terminates ␤ARR-dependent endosomal ERK signaling (10).
The relative importance of plasma membrane and endosomal signaling in determining integrated cellular responses and pathophysiological outcomes is far from clear. ␤ARRs desensitize plasma membrane G protein-dependent signaling, mediate endosomal G protein-independent signaling, and facilitate the intracellular processing of receptors that is necessary for recycling and resensitization of plasma membrane signaling. ECE-1 promotes recycling and resensitization of plasma membrane signaling but terminates endosomal ␤ARR-dependent signaling. Therefore, is plasma membrane or endosomal signaling of primary importance, or is resensitization of cell surface signaling the key event in allowing sustained cellular responses?
To address these questions, we investigated the effects of altered expression and activity of ␤ARRs and ECE-1 on SP-dependent trafficking and inflammatory signaling of the NK 1 R. This strategy allowed us to study the relative contributions of plasma membrane and endosomal signaling events. Because there is increased mucosal expression of the NK 1 R in the inflamed intestine (11), we studied human colonocytes overexpressing the NK 1 R that respond to SP by release of NF-Bdriven proinflammatory cytokines, including IL-8 (12,13). We defined the importance of ␤ARRs and ECE-1 in regulating plasma membrane, endosomal, and inflammatory signaling. Our results show that the principal role of ␤ARRs and ECE-1 is to resensitize the plasma membrane NK 1 R signaling that is necessary for sustained SP-induced inflammation.
FRET-FRET was analyzed using a GE Healthcare INCell 2000 analyzer. For emission ratio analysis, cells were excited sequentially using a FITC filter (490/20) with emission measured using dsRed (605/52) and FITC (525/36) filters and a polychroic optimized for the FITC/dsRed filter pair (Quad4). Images of 14 wells were collected at 1-min intervals. HEK cells were equilibrated in Hanks' balanced salt solution at 37°C. Baseline emission ratio images were captured for 4 min. Cells were challenged with SP (1 nM) or vehicle, and images were captured for 20 min. At the end of every experiment, cells were stimulated with phorbol 12,13 dibutyrate (200 nM, 10 min, positive control) to generate a maximal increase in ERK phosphorylation, and images were captured for 4 min. Data were analyzed using ImageJ. Emission ratio image stacks (baseline, stimulated, positive control) were collated and aligned using a StackCreator script written by Cameron Nowell (Monash). Cells were selected, and fluorescence intensity was measured over the combined stack. Background intensity was subtracted, and FRET data were plotted as the change in the dsRed:FITC emission ratio relative to the baseline for each cell (F/F 0 ). Cells with Ͼ10% change in F/F 0 after stimulation with phorbol 12,13 dibutyrate were selected for analysis.
NF-B p65 Activation-NCM-NK 1 R cells were incubated with SP (10 nM) or vehicle (control) for 45 min, and phosphorylation of NF-B p65 was determined by Western blotting (13).
Statistics-Data are presented as mean Ϯ S.E. of triplicate observations from n Ͼ 3 experiments. Differences were assessed using Student's t test for two comparisons and one-or two-way analysis of variance followed by Bonferroni post test and Student-Newman-Keul test for multiple comparisons. p Ͻ 0.05 was considered significant.

SP Promotes NK 1 R Association with ␤ARRs and ECE-1 at the Plasma Membrane and in Endosomes of HEK293 Cells-We
analyzed the interactions between NK 1 R, ␤ARRs, and ECE-1 in defined cellular compartments using BRET. It was not possible to coexpress at adequate levels two BRET constructs in NCM460 colonocytes. Therefore, we examined the interactions between NK 1 R, ␤ARRs, or ECE-1 in HEK293 cells in which the NK 1 R could be coexpressed with ␤ARR or ECE-1 isoforms at suitable levels for BRET measurements. In HEK293 cells expressing NK 1 R-RLuc8 and ␤ARR1-YFP or ␤ARR2-YFP, SP stimulated a rapid (1-min), sustained (Ͼ10 min), and concentration-dependent increase in BRET between NK 1 R-RLuc8 and ␤ARR1-YFP or ␤ARR2-YFP (EC 50 for both isoforms, 2.5 nM) (Fig. 1, A and B).
To quantitatively assess SP-induced NK 1 R trafficking between the plasma membrane and early and recycling endosomes, we assessed BRET between NK 1 R-RLuc8 and Venustagged BRET acceptors that are localized to the plasma membrane (KRas) and early (Rab5a) or recycling (Rab11) endosomes. SP (1 nM and 10 nM) decreased NK 1 R-RLuc8 and KRas-Venus BRET (Fig. 1C) and increased NK 1 R-RLuc8 and Rab5a-Venus BRET (Fig. 1D) in a concentration-dependent manner. To assess NK 1 R trafficking after SP removal, cells were challenged with SP (1 nM, 10 min) and washed, and then BRET was measured 15-120 min after SP exposure. Cells were treated with cycloheximide (100 M) to prevent new receptor synthesis. After SP removal, NK 1 R-RLuc8 and KRas-Venus BRET continued to decrease, and NK 1 R-RLuc8 and Rab5a-Venus BRET continued to increase for 15 min (Fig. 1E). Thereafter, BRET between NK 1 R-RLuc8 and Rab11-Venus or KRas-Venus increased and NK 1 R-RLuc8 and Rab5a-Venus BRET declined. Thus, SP stimulates NK 1 R trafficking from the plasma membrane to early and then recycling endosomes, which deliver the NK 1 R back to the plasma membrane.
To determine whether the association of ECE-1 with the NK 1 R in endosomes controls the interactions between the NK 1 R and ␤ARRs, we examined the effects of the ECE-1 inhibitor SM-19712 on NK 1 R and ␤ARR BRET. SM-19712 caused a 2-fold increase in SP (1 nM)-induced BRET between NK 1 R-RLuc8 and ␤ARR1-YFP or ␤ARR2-YFP ( Fig. 2A). In contrast, the G␣ q inhibitor UBO-QIC and the PKC inhibitor GF109203X did not affect SP-induced BRET between NK 1 R-RLuc8 and ␤ARR1-YFP or ␤ARR2-YFP (Fig. 2B). These inhibitors also had no effect on SP-induced endocytosis of the NK1R, determined by measuring the decrease in NK 1 R-RLuc8 and KRas-Venus BRET in response to SP (1 nM) (Fig. 2C). Considered together, these results are consistent with the hypothesis that SP induces association of the NK 1 R with ECE-1, notably ECE-1d, in endosomes. Endosomal ECE-1 can degrade SP and, thereby, disrupt NK 1 R interaction with ␤ARRs. Furthermore, the NK 1 R/␤ARR interaction and NK 1 R trafficking to endosomes are independent of G␣ q -PKC activation.
␤ARRs and ECE-1 Regulate NK 1 R Trafficking in NCM460 Colonocytes-To ascertain whether ECE-1 and ␤ARRs regulate NK 1 R trafficking and signaling in cells that participate in SPmediated inflammation, we studied non-transformed human colonocytes (NCM460 cells) stably expressing the NK 1 R (NCM-NK 1 R cells). The NK 1 R is up-regulated in colonocytes in the inflamed intestine, and NCM460 cells are widely used to examined neuropeptide-stimulated inflammatory signaling (11). SP stimulated a small and variable increase in [Ca 2ϩ ] i in non-transfected NCM460 cells (data not shown), and, thus, the NK 1 R was overexpressed to facilitate studies of regulation and signaling. NCM-NK 1 R cells expressed endogenous ␤ARR1 and 2 mRNA (Fig. 3A), although ␤ARR1 immunoreactivity (IR) was  more abundant (Fig. 3B). NCM-NK 1 R cells also expressed mRNA for ECE-1a-d isoforms (Fig. 3C) and ECE-1-IR (Fig. 3D). Thus, NCM460 colonocytes endogenously express isoforms of ␤ARRs and ECE-1 that play a major role in regulating trafficking and signaling of the NK 1 R.
We incubated NCM-NK 1 R cells with Alexa-tagged SMSP, an NK 1 R agonist (21), to localize the NK 1 R. To determine the role of ␤ARRs in the endocytosis of Alexa-SMSP in NCM-NK 1 R cells, we transfected cells with ␤ARR1-GFP, ␤ARR2-GFP, or dominant negative ␤ARR1 319 -418 -GFP (23). Incubation of cells with Alexa-SMSP (10 nM) at 4°C induced translocation of both ␤ARR1-GFP and ␤ARR2-GFP to the plasma membrane (Fig. 4,  A and B). After subsequent incubation at 37°C for 30 min, ␤ARR1-GFP and ␤ARR2-GFP had trafficked with Alexa-SMSP to endosomes (Fig. 4, A and B). In contrast, Alexa-SMSP was not internalized in cells expressing ␤ARR1 319 -418 -GFP (Fig.  4C). Internalized Alexa-SMSP also colocalized with endogenous ␤ARR1-IR in endosomes (Fig. 4D). The localization and SP-induced redistribution of endogenous ␤ARR1-IR was completely consistent with the localization and SP-induced redistribution of ␤ARR1-GFP, and the ␤ARR1 antibody detected ␤ARR1 in Western blot analyses (not shown). These results suggest the specific detection of endogenous ␤ARR. Together, these results suggest that ␤ARRs mediate SP-induced endocytosis of the NK 1 R and traffic with SP/NK 1 R to endosomes in NCM-NK 1 R cells.

JOURNAL OF BIOLOGICAL CHEMISTRY 20287
␤ARRs and ECE-1 Control the Spatiotemporal Dynamics of Intracellular NK 1 R Signaling in NCM460 Colonocytes and HEK293 Cells-To precisely define the spatiotemporal regulation of ERK signaling and assess its regulation by ␤ARRs and ECE-1, we expressed, in NCM460 cells, the NK 1 R and the FRET biosensors CytoEKAR and NucEKAR, which permit the detection of ERK activity in the cytoplasm or nucleus, respectively (16). We also treated cells with ␤ARR siRNA or inhibitors of ECE-1 (SM-19712), G␣ q (UBO-QIC), PKC (GF109203X), or epidermal growth factor receptor (EGFR, AG1478) to evaluate the importance of these signal-ing pathways for SP-stimulated activation of ERK in the cytosol and nucleus.
In control experiments (no inhibitors), SP (1 nM) induced a rapid increase in cytosolic ERK phosphorylation within 5 min that declined to basal levels after 20 min (Fig. 6, A-C and G). In contrast, SP stimulated a gradual increase in nuclear ERK phosphorylation that was sustained for 20 min (Fig. 6, D-F and H).
To determine how the NK 1 R activates ERK in two spatially distinct pools with distinct kinetics, we assessed the effect of siRNA knockdown and of pharmacological inhibitors on SPinduced cytosolic and nuclear ERK phosphorylation. The use of siRNA allowed selective knockdown of ␤ARR1 and ␤ARR2 mRNA (Fig. 7, A and B), and ␤ARR1 and ␤ARR2 siRNA were combined for effective knockdown of both isoforms (Fig. 7C). ␤ARR1 and ␤ARR2 siRNA had no effect on SP-stimulated phosphorylation of cytosolic ERK (Fig. 6, A and G) but abolished SP-stimulated phosphorylation of nuclear ERK (Fig. 6, D  and H). The ECE-1 inhibitor SM-19712 did not affect SP-stimulated cytosolic ERK phosphorylation (Fig. 6, B and G) but increased nuclear ERK phosphorylation ϳ2-fold (Fig. 6, E and  H). Inhibitors of G␣ q (UBO-QIC), PKC (GF109203X), or EGFR (AG1478) abolished SP-induced cytosolic ERK phosphorylation (Fig. 6, C and G). Inhibition of G␣ q and PKC, but not EGFR, also prevented SP-induced nuclear ERK phosphorylation (Fig.  6, F and H).
To determine whether these effects were cell-type specific, we similarly examined SP-induced ERK signaling in HEK293 cells expressing NK 1 R and FRET biosensors. In HEK293 cells, and in NCM460 colonocytes, ␤ARR1 and ␤ARR2 siRNA blocked SP-stimulated phosphorylation of nuclear, but not cytosolic, ERK (Fig. 8, A, D, G, and H). SM-19712 magnified SP-stimulated phosphorylation of nuclear, but not cytosolic, ERK (Fig. 8, B, E, G, and H). UBO-QIC and GF-109203X both blocked SP stimulation of nuclear and cytosolic ERK, and AG1478 selectively prevented SP-induced stimulation of cytosolic ERK (Fig. 8, C, F, G, and H).
Considered together, the results of these experiments to examine mechanisms of ERK signaling support our previous studies showing that ␤ARRs can recruit the NK 1 R and upstream activators of ERK, including Src and MEKK, to endosomes (10,24). They are also consistent with the ability of ECE-1 to interact with the NK 1 R, degrade SP in endosomes, and disrupt the NK 1 R association with ␤ARR that is necessary for activation of nuclear ERK (9,10). Our results suggest that the transient cytosolic ERK phosphorylation mediated by the NK 1 R depends on activation of a G␣ q -PKC pathway leading to EGFR transactivation. The delayed and sustained phosphor-

Substance P Inflammatory Signaling
ylation of nuclear ERK may be mediated by G␣ q -PKC and ␤ARR pathways but is independent of EGFR. Although G␣ q and PKC inhibitors prevented nuclear ERK activation, they did not affect SP-induced recruitment of ␤ARR1 or ␤ARR2 to the NK 1 R (Fig. 2). Nor was there any effect of G␣ q or PKC inhibition on SP-induced endocytosis of the NK 1 R (Fig. 2). Thus, the G␣ q -PKC pathway appears to mediate both the sustained, delayed phosphorylation of nuclear ERK and the rapid, transient phosphorylation of cytosolic ERK. ␤ARRs may be required for increased nuclear ERK phosphorylation because they are necessary for internalization of the NK 1 R. Notably, these mechanisms operate in both NCM460 colonocytes and in HEK cells and are, thus, not cell type-selective. Future studies using overlapping genetic and pharmacological approaches to examine signaling by the endogenous NK 1 R are required to substantiate these suggestions.
␤ARRs and ECE-1 Promote Resensitization of NK 1 R Plasma Membrane Signaling in NCM460 Colonocytes-To define the mechanisms of resensitization of SP-induced Ca 2ϩ signaling, NCM-NK 1 R cells were preincubated with SP (10 nM) or vehicle (control) for 10 min, washed, recovered for 0 -150 min, and challenged with SP (10 nM) to assess resensitization. After incubation with SP for 10 min, the response to a second challenge with SP was desensitized (Fig. 9A). Complete resensitization occurred after recovery for 90 -150 min (Fig. 9B).
SP Induces the interaction between the NK 1 R and GRK2 (25), and GRK2 phosphorylates the NK 1 R and promotes interactions with ␤ARRs (17, 19). To investigate the contributions of GRK2 and ␤ARRs to NK 1 R resensitization, NCM-NK 1 R cells were preincubated with vehicle or 1 nM SP because GRKs and ␤ARRs preferentially mediate desensitization to submaximal agonist concentrations. The GRK2 inhibitor CMPD103A (26) increased the magnitude and duration of the initial response to SP (80 Ϯ 11% increase in maximal response compared with vehicle, p Ͻ 0.05, Fig. 9C) and inhibited resensitization of SPinduced Ca 2ϩ signaling (20 Ϯ 1.8% of control resensitization, 120-min recovery; p Ͻ 0.0001; Fig. 9, C and F). ␤ARR1 and ␤ARR2 siRNA caused a small but insignificant increase in the magnitude of the initial SP response (30 Ϯ 24% increase in maximal response compared with vehicle, p ϭ 0.33, Fig. 9D) and inhibited resensitization of SP-induced Ca 2ϩ signaling (43 Ϯ 1.8% of control resensitization; 120-min recovery; p Ͻ 0.0001; Fig. 9, D and F). The ECE-1 inhibitor SM-19712 did not affect the magnitude or desensitization of the first response to SP (10 nM) but did inhibit resensitization of SP (10 nM)-induced Ca 2ϩ signaling (16 Ϯ 6% of control resensitization; 90-min recovery; p Ͻ 0.001; Fig. 9, E and F).
Our results show that GRK2, ␤ARRs, and ECE-1 all promote resensitization of plasma membrane NK 1 R signaling in colonocytes. By phosphorylating the NK 1 R, GRK2 promotes the association with ␤ARRs, which deliver SP and the NK 1 R to endosomes containing ECE-1. ECE-1 can then degrade SP and disrupt the NK 1 R/␤ARR signalosome, which permits NK 1 R recycling.
In NCM-NK 1 R cells, SP-induced IL-8 expression involves the activation of PKC␦ and is blocked by rottlerin, which inhibits multiple kinases (12). Because SP causes ERK activation in NCM-NK 1 R cells by a ␤ARR-dependent mechanism that is terminated by endosomal ECE-1, ERK could also regulate IL-8 release. However, PD98059, an inhibitor of the MEK1-ERK pathway, did not affect SP-induced phosphorylation of NF-B p65 (Fig. 10B) or IL-8 release (Fig. 10E), whereas rottlerin blocked SP-induced phosphorylation of NF-B p65 (Fig. 10B), suggesting that ERK is not involved in SP-induced IL-8 expression in colonocytes.
Our results show that sustained SP inflammatory signaling requires the presence of the NK 1 R at the plasma membrane, which depends on the capacity of ECE-1 and ␤ARRs to promote recycling and resensitization of the NK 1 R. However, a substantial proportion of activated, phosphorylated NK 1 R remains at the plasma membrane, where protein phosphatase 2A (PP2A) dephosphorylates and resensitizes non-internalized NK 1 R (18). Fostriecin, an inhibitor of PP2A that blocks resensitization of non-internalized NK 1 R, strongly inhibited SP-stimulated IL-8 secretion (Fig. 10D). These findings reinforce the conclusion that sustained SP inflammatory signaling requires resensitization of NK 1 R at the plasma membrane rather than sustained endosomal NK 1 R signaling via ERK.

DISCUSSION
Our results provide insights into the relative importance of plasma membrane and endosomal signaling of the NK 1 R for SP-induced inflammation. We report that SP stimulates interactions between the NK 1 R and two classes of proteins that play a major role in regulating NK 1 R signaling at plasma and endosomal membranes: the scaffolding proteins ␤ARR1 and ␤ARR2 and the transmembrane endopeptidases ECE-1a through ECE-1d. ␤ARRs uncouple the NK 1 R from G proteins at the plasma membrane, which desensitizes plasma membrane Ca 2ϩ signal-ing, and mediate NK 1 R endocytosis, which is required for sustained activation of nuclear ERK. ECE-1d associates with SP/NK 1 R in endosomes and degrades SP, which terminates activation of nuclear ERK. By delivering SP/NK 1 R to endosomes, ␤ARRs facilitate ECE-1 degradation of SP. Our results suggest that both ␤ARRs and ECE-1 are necessary for resensitization of plasma membrane Ca 2ϩ signaling. The sustained stimulation of colonocytes with SP induces NF-B activation and IL-8 secretion. This inflammatory signaling does not require ERK activation and is independent of endosomal NK 1 R ERK signaling. However, ECE-1 inhibition and ␤ARR2 knockdown suppress SP-induced NF-B and IL-8 signaling, indicating that the principal function of ECE-1 and ␤ARR2 in colonocytes, rather than to participate in endosomal signaling, is to deliver resensitized NK 1 R to the plasma membrane, which is required for sustained SP inflammatory signaling. Consistent with this proposal is the finding that PP2A, which resensitizes non-internalizing NK 1 R at the plasma membrane, is also nec- essary for SP-induced IL-8 secretion. Considered together, our findings suggest that sustained inflammatory signaling of SP during inflammation requires efficient resensitization of the NK 1 R at the plasma membrane by ␤ARR2-, ECE-1-, and PP2Adependent mechanisms. Further studies using complimentary genetic and pharmacological approaches are required to define the importance of ␤ARR2, ECE-1, and PP2A for chronic inflammation in vivo.
Contribution of ECE-1 to SP Inflammatory Signaling-We show that ECE-1 resensitizes the NK 1 R at the plasma membrane, which is necessary for sustained SP inflammatory signaling. Although we did not determine the mechanism of ECE-1dependent resensitization of the NK 1 R in colonocytes, studies in other cell types indicate that ECE-1 degrades SP in acidic early endosomes and disassembles the SP-NK 1 R-␤ARR signalosome, allowing the NK 1 R, freed from ␤ARRs, to recycle and resensitize (8 -10). SP stimulated interactions between the NK 1 R and ECE-1c and ECE-1d, which constitutively internalize in the same endosomes as GPCRs (22). ECE-1 similarly interacts with somatostatin 2A receptors (15). Whether such interactions enhance neuropeptide degradation remains to be determined. Although ECE-1a and ECE1c were also detected at the plasma membrane, cell surface ECE-1 does not regulate SP signaling because ECE-1 does not degrade SP at an extracellular pH value (9). In contrast, neprilysin readily degrades SP at the cell surface and, thereby, inhibits the activation of cell surface NK 1 R (5, 6) and SP-induced colitis (27).
Our finding that ECE-1 inhibition attenuates SP-stimulated NF-B activation and IL-8 release from colonocytes implicates a primary role for plasma membrane signaling of the NK 1 R in SP-induced inflammation. By stabilizing SP/NK 1 R within endosomes, the ECE-1 inhibitor caused sustained activation of nuclear ERK in colonocytes and HEK cells, as in primary neurons (8,10). Although ERK mediates NF-B activation and IL-8 release in response to some proinflammatory stimulants in colonocytes (28), MEK inhibition did not affect SP-stimulated IL-8 secretion, which, instead, depends on PKC␦ and NF-B in NCM cells (12,13). The anti-inflammatory effects of an ECE-1 inhibitor are consistent with the inhibitory actions of ECE-1 blockers on resensitization of SP-stimulated plasma extravasation (7) and the protective effects of SM-19712 in colitis (29). Indeed, SP and the NK 1 R are key mediators of 2,4,6-trinitrobenzenesulfonic acid (TNBS) colitis, which depends on NK 1 Rdependent release of the neutrophil attractant IL-8 from colonocytes (11). Given the widespread proinflammatory actions of SP and the finding that the NK 1 R internalizes and recycles in many inflamed tissues (11), it is likely that ECE-1 plays a general role in the maintenance of proinflammatory SP signaling. Because ECE-1 degrades other proinflammatory neuropeptides in endosomes (22), ECE-1 inhibition could prevent sustained signaling of several neuropeptide mediators of inflammation.
Contribution of ␤ARR to SP Inflammatory Signaling-We report that SP stimulates translocation of ␤ARRs from the cytosol to the plasma membrane, followed by trafficking of SP, NK 1 R, and ␤ARR1/2 to early endosomes containing ECE-1a through ECE1d. Inhibition of GRK2 and disruption of ␤ARRs enhanced SP-mediated Ca 2ϩ signals and blocked NK 1 R endocytosis, consistent with the known role of GRKs and ␤ARRs in attenuating plasma membrane signaling and mediating NK 1 R endocytosis (17,24). ␤ARR knockdown also attenuated SP-induced activation of nuclear ERK and inhibited resensitization of Ca 2ϩ signaling. These findings supports the involvement of ␤ARRs in the assembly of MAPK signalosomes because of their role in endosomal trafficking of the NK 1 R (24). The results are consistent with receptor endocytosis, intracellular processing, and recycling as mechanisms of resensitization (19).
By selective siRNA knockdown, we probed the role of ␤ARR isoforms in SP-induced inflammatory signaling. Although knockdown of ␤ARR1 and ␤ARR2 was required to inhibit resensitization and internalization-dependent nuclear ERK signaling, ␤ARR2 siRNA alone attenuated SPstimulated activation of NF-B and IL-8 release, suggesting that ␤ARR2 promotes SP inflammatory signaling in colonocytes. ␤ARR2 also mediates lipopolysaccharide-stimulated IL-8 secretion (30). One explanation for the proinflammatory action of ␤ARR2 could be the ␤ARR2-dependent recruitment of regulators of NF-B activation to the NK 1 R. Similarly, ␤ARR2 recruits CARMA3 to the lysophosphatidic acid receptor to stimulate NF-B activation (31). Another explanation for the positive role of ␤ARR2 in SP inflammatory signaling may be related to the resensitization of plasma membrane signaling. By mediating endocytosis, ␤ARRs can promote ligand dissociation and receptor dephosphorylation in endosomes, which are required for recycling and resensitization. This possibility is supported by our observation that endosomal ECE-1 induces NK 1 R resensitization and inflammatory signaling. Our finding that ␤ARRs mediate SP-induced inflammation is supported by reports that ␤ARR1 deletion protects mice from colitis (32), which involves SP/NK 1 R interactions (11).
A limitation of our study is that it examines the regulation of the NK 1 R overexpressed in NCM460 colonocytes or HEK293 cells. Although we obtained similar findings with both cell lines, it will be important to confirm that similar mechanisms regulate the endogenous NK 1 R expressed in functionally relevant cell types and to use complimentary genetic and pharmacological approaches to define the mechanisms of sustained inflammatory signaling by SP and the NK 1 R.
In summary, our results suggest that the primary function of ␤ARRs and ECE-1 in SP-dependent inflammatory signaling is to maintain plasma membrane signaling through control of NK 1 R internalization, recycling, and resensitization. These processes may sustain NK 1 R-driven inflammatory signaling, an important component of the pathophysiology of widespread inflammatory disorders.