INTRODUCTION

Cellular responses to agonists of G-protein-coupled receptors are rapidly attenuated in the continuous presence of agonist and desensitize to repeated application of agonist (1, 2). With time between challenges, responses recover and cells resensitize. These are important processes, since they determine the ability of cells to respond to agonists; desensitization prevents the uncontrolled stimulation of cells, whereas resensitization allows cells to recover or maintain their responsiveness. Desensitization and resensitization are controlled at the level of the receptors and at more distal steps in the signaling pathway, but regulation of the receptors is of critical importance. The mechanisms may be distinct for different classes of G-protein-coupled receptors. Receptors for hormones and neurotransmitters, exemplified by the ₂AR¹ and the NK1-R, bind agonists reversibly and desensitize in a reversible manner. The reuse of these receptor molecules mediates resensitization of cellular responses (3, 4, 5). Receptors for proteases, such as the Th-R, are cleaved to expose a tethered ligand that binds and activates the receptor in an irreversible manner (6). This generates a quantum of signal that is rapidly attenuated or desensitized despite the irreversible mechanism of activation (7). Resensitization of cellular responses to thrombin requires the availability of new receptor molecules (8).

Uncoupling of receptors from G-proteins is a principal mechanism of desensitization (1, 2). For example, GRKs and second messenger kinases (PKA) phosphorylate the ₂-AR. -Arrestin binds to the phosphorylated receptor, disrupts interaction with G-proteins, and terminates the signal. The NK1-R and the Th-R both desensitize by this mechanism (5, 9, 10, 11). Agonist-induced endocytosis of receptors contributes to desensitization by depleting receptors from the plasma membrane, thereby making them inaccessible to agonists in the extracellular fluid (3, 4, 12).

Receptors for soluble ligands, including the ₂-AR and NK1-R, recycle after internalization (3, 4). Resensitization requires receptor internalization and processing (dissociation of agonist and -arrestins and receptor dephosphorylation) and receptor recycling (5, 13, 14). In contrast, the Th-R, which is irreversibly activated by cleavage, is internalized and degraded in lysosomes (8, 15, 16). Resensitization of responses to thrombin requires mobilization of the large stores of the Th-R in the Golgi apparatus.

PAR-2 is the second member of the thrombin family of proteinase-activated receptors (17, 18, 19, 20). Trypsin cleaves PAR-2 within its extracellular NH₂ terminus, exposing a tethered ligand that binds the cleaved receptor and activates phospholipase C ₁ (17, 20). PAR-2 is highly expressed in the pancreas and gastrointestinal tract, where it may be activated by pancreatic trypsin, and in certain tumor cells, where it may respond to tumor-associated trypsins (17, 20). In the small intestine, PAR-2 is expressed by surface epithelial cells and may mediate some of the effects of trypsin on gastrointestinal function (20). These include the inhibition of pancreatic secretion by a negative feedback mechanism (21) and regulation of prostaglandin generation and hormone secretion (22, 23, 24). Activation of PAR-2 also directly stimulates amylase secretion from pancreatic acini, stimulates contraction of gastric muscle, and inhibits tumor cell growth (20, 25). The mechanisms of desensitization and resensitization of these effects are completely unexplored but are likely to be of considerable importance in the intestine, where cells are episodically exposed to trypsin during feeding and fasting (24). Furthermore, the duration of the signal may be of importance in determining other effects of PAR-2, such as regulation of cell growth.

The aim of this investigation was to delineate the mechanisms of desensitization and resensitization of PAR-2 in transfected kidney epithelial cells and in hBRIE 380 cells, a highly differentiated cell line derived from the intestinal epithelium that naturally expresses PAR-2. We (a) characterized PAR-2-mediated Ca²⁺ mobilization, (b) examined cleavage of PAR-2 by trypsin, (c) investigated homologous desensitization to repeated activation of PAR-2, (d) determined the role of PKC in attenuating Ca²⁺ responses, (e) examined internalization and targeting of activated PAR-2, and (f) studied the mechanism of resensitization of Ca²⁺ mobilization.

EXPERIMENTAL PROCEDURES

Trypsin was purchased from Worthington Biochemical Co. (Freehold, NJ). Human AP (SLIGKV-NH₂) and rat AP (SLIGRL-NH₂) were synthesized by solid phase methods and purified by high pressure liquid chromatography. Human AP was used for all experiments with KNRK cells expressing human PAR-2 (19, 20); rat AP was used in all experiments with the rat hBRIE 380 cells (25). Lipofectin and G418 were from Life Technologies, Inc. Fura-2/AM and pleuronic were from Molecular Probes (Eugene, OR). Cycloheximide, PDBu, 4-phorbol 12,13-dicanoate, GF 109203X (bisindolylmaleimide I), bisindolylmaleimide V, forskolin, and H-89 were from Calbiochem. Bafilomycin A₁ was a gift from Dr. Jonathan R. Green (Ciba-Geigy Ltd., Basel, Switzerland). Amiloride, nifedipine, SK&F 96365, and thapsigargin were from Biomol Research Laboratories, Inc. (Plymouth Meeting, PA). Bradykinin was from Bachem Biosciences Inc. (King of Prussia, PA). Other reagents were from Sigma.

A rabbit polyclonal antibody to the hemagglutinin epitope (HA.11) and an antibody to the resident Golgi protein mannosidase II were from Berkeley Antibody Co. (Richmond, CA). A mouse monoclonal antibody (M1) to the Flag epitope was from International Biotechnologies, Inc. (New Haven, CT). A mouse monoclonal antibody to the transferrin receptor was a gift from Dr. Ian Trowbridge (The Salk Institute, San Diego, CA). A mouse monoclonal antibody (GM10) to membranes from insulin secretory granules, which recognizes lysosomes, was a gift from Dr. John Hutton (Cambridge, United Kingdom). Affinity-purified goat anti-rabbit or anti-mouse IgG coupled to fluorescein isothiocyanate or Texas Red were from Cappel Research Products (Durham, NC). ¹²⁵I-sheep anti-mouse IG (species-specific F(ab)₂)were from Amersham Corp.

We transfected cells with cDNA encoding human PAR-2 with an NH₂-terminal Flag epitope (DYKDDDDK). These cells were used for functional desensitization and resensitization experiments. We also transfected cells with this construct with a 12CA5 hemagglutinin epitope (YPYDVPDYA) at the intracellular COOH terminus, so that antibodies to this epitope could be used to localize PAR-2 and study receptor trafficking. In addition, we generated a cDNA encoding human PAR-2 with a proopiomelanocortin signal peptide, immediately followed by a Flag epitope amino-terminal to the trypsin cleavage site, and a 12CA5 epitope at the COOH terminus (see Fig. 5A). This construct would be expressed at the cell surface with an NH₂-terminal Flag sequence, so that loss of surface Flag immunoreactivity could be used to determine rates of cleavage and internalization of PAR-2. We verified that the epitope-labeled receptors were appropriately localized to the plasma membrane by immunofluorescence. Trypsin and AP stimulated Ca²⁺ mobilization in cells expressing these receptors, as they do in cells that naturally express PAR-2 (20), suggesting that these epitope tags do not influence signaling mechanisms of this receptor. In addition, the Flag epitope does not affect trafficking of other G-protein-coupled receptors (4).

KNRK cells were from the American Type Tissue Culture Collection (ATCC CRL 1569, Rockville, MD). cDNA encoding human PAR-2 alone or encoding epitope-labeled human PAR-2 was subcloned into the neomycin-resistant vector pcDNA3. KNRK cells were transfected with this cDNA by lipofection (20). Clones were selected in medium containing 800 µg/ml G418 and screened for uniformly high expression of PAR-2 by immunofluorescence using the HA.11 antiserum and by measuring trypsin-induced Ca²⁺ mobilization. Clonal lines were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 400 µg/ml G418. Cells were plated 24-48 h before experiments on poly-D-lysine-coated glass coverslips. hBRIE 380 cells were maintained in Iscove's modified Dulbecco's medium containing 10% fetal bovine serum and were plated on 1% Matrigel on glass coverslips for 48 h before use (26, 27).

Cells were washed with a physiological salt solution (in mM: 137 NaCl, 4.7 KCl, 0.56 MgCl₂, 2 CaCl₂, 1.0 Na₂HPO₄, 10 Hepes, 2.0 L-glutamine, and 5.5 D-glucose, pH 7.4) containing 0.1% bovine serum albumin. They were incubated in this solution with 2.5 µM Fura-2/AM and 0.2% pleuronic for 20 min at 37 °C, washed, and transferred to a spectrofluorometer containing physiological salt solution. Cells were challenged with 1 nM to 1 µM trypsin or 1-100 µM AP. Fluorescence was measured at 340 and 380 nm excitation and 510 nm emission, and the ratio of the fluorescence at the two excitation wavelengths, which is proportional to the [Ca²⁺]_i, was calculated.

To examine the contribution of extracellular Ca²⁺ to the change in [Ca²⁺]_i, cells were exposed to 100 nM trypsin or 100 µM AP in Ca²⁺-free solution. Ca²⁺ was added back to the assay solution 1-2 min later to give a final concentration of 2 mM. The channel in the plasma membrane that allowed entry of Ca²⁺ from the extracellular fluid was characterized by exposing cells to the Ca²⁺ channel blockers amiloride (10 µM, to inhibit N-type channels (28)), nifedipine (10 µM, to inhibit L-type channels (29)) or SK&F 96365 (10 µM, to inhibit the putative ``Ca²⁺ release-activated current'' or I_CRAC channel (30, 31)) 2 min before challenge with agonist.

To examine the role of PKC in regulating the Ca²⁺ response to trypsin and AP, cells were either treated with the selective PKC inhibitor GF 109203X (0.1-10 µM (32, 33)) 2-20 min before exposure to agonist or acutely exposed to the PKC activator PDBu (1 nM to 10 µM (33)) for 2 min before challenge with agonist. The time of treatment with GF 109203X (2 min or 20 min) did not significantly change the extent of its effect. In control experiments, cells were incubated with carrier, inactive bisindolylmaleimide V, or the inactive 4-phorbol-12,13-dicanoate. In some experiments, drugs were added at the peak of the Ca²⁺ response to trypsin or thapsigargin.

To study the role of PKA in regulating the Ca²⁺ response to trypsin and AP, cells were treated with either a PKA inhibitor H-89 (100 nM (34)) or with forskolin (10 µM) to activate adenylate cyclase and PKA 2 min before exposure to agonist. In control experiments, cells were incubated with carrier.

To examine desensitization and resensitization, cells were incubated with a desensitizing dose of 100 nM trypsin, 100 µM AP or carrier (control) for 1 min at 37 °C, washed, and then exposed to a test dose of 100 nM trypsin or 100 µM AP at various times after the desensitizing dose. The change from the base line immediately before agonist exposure was calculated. The response to the test dose of agonist was compared between agonist- and carrier-treated cells (set at 100%), and expressed as a percentage. Cells were also challenged with 1 µM bradykinin after trypsin or AP to verify that intracellular Ca²⁺ pools were intact.

To investigate the contribution of receptor internalization to desensitization, we treated cells with the endocytic inhibitor phenylarsine oxide, as we have previously described (4, 5). Cells were treated for 5 min with 80 µM phenylarsine oxide plus 5 µM -mercaptoethanol before [Ca²⁺]_i measurements (4, 5). To investigate the mechanisms of resensitization, cells were treated with cycloheximide, brefeldin A, or bafilomycin A₁. Cycloheximide is a protein synthesis inhibitor, and we have previously shown that incorporation of [³⁵S]methionine into KNRK cells is reduced by 95% by treatment with 70 µM cycloheximide (4). Brefeldin A causes disassembly of the Golgi apparatus and mixing with the endoplasmic reticulum and induces alterations in the morphological appearance of endosomes and lysosomes (35). Bafilomycin A₁ is an inhibitor of vacuolar-type H⁺-ATPase (36), and we have shown that 1 µM bafilomycin prevents vesicular acidification and NK1-R recycling in KNRK cells (4). Thus, cells were preincubated with 70 µM cycloheximide, 1 µM bafilomycin A₁, or 10 µg/ml brefeldin A for 30 min at 37 °C before exposure to agonists, and the drugs were included in all solutions during the experiment. In all experiments control cells were incubated with appropriate carrier solutions.

We quantified Flag immunoreactivity at the cell surface of KNRK-PAR-2 cells expressing PAR-2 with a NH₂-terminal Flag epitope (see Fig. 5A), as we have previously described for the NK1-R (4). Cells were placed in Dulbecco's modified Eagle's medium containing 0.1% bovine serum albumin and incubated with 10 nM trypsin, 100 µM AP, or carrier (control) for 2, 5, 10, or 30 min at 37 °C. Cells were washed and Flag antibody (0.1 µg/ml) was added for 60 min at 4 °C. Cells were washed, and incubated with ¹²⁵I-sheep anti-mouse IG (0.1 µCi/well) for 60 min at 4 °C. Cells were washed in PBS at 4 °C and lysed with 0.5 M NaOH. Radioactivity and protein content of the lysate were measured. Binding was normalized for protein content (expressed as cpm/mg of protein). Nonspecific binding to nontransfected KNRK was usually <5% of that measured in KNRK-PAR-2 cells. To investigate the contribution of receptor endocytosis to the loss of surface Flag immunoreactivity, we exposed cells to the endocytic inhibitors phenylarsine oxide or hypertonic sucrose (4, 5). Cells were pretreated for 5 min with 80 µM phenylarsine oxide or continuously incubated with 0.45 M sucrose.

Cells were placed in Dulbecco's modified Eagle's medium containing 0.1% bovine serum albumin at 37 °C. Trypsin (10-100 nM) or AP (10-100 µM) was added, and incubation continued for 5-120 min. Cells were fixed in 4% paraformaldehyde in 100 mM phosphate-buffered saline (pH 7.4) for 20 min at 4 °C. Cells were permeabilized by incubation in phosphate-buffered saline with 1% normal goat serum and 0.1% saponin for 3 periods of 5 min. They were incubated with primary antibodies overnight at 4 °C (anti-HA.11, 1:5000; anti-mannosidase II, 1:1,000; anti-transferrin receptor, 1:4000; GM10 1:6,000), washed, and incubated with fluorescent secondary antibodies (1:200) for 120 min at room temperature (4). Experiments were repeated at least three times with two coverslips per observation.

Slides were reviewed independently by two or three investigators to avoid bias. Cells were observed with an MRC 1000 laser-scanning confocal microscope (Bio-Rad Laboratories Inc., Hercules, CA) equipped with a krypton/argon laser and attached to a Zeiss Axiovert microscope. A Zeiss plan-Apochromat × 100 oil immersion objective with a numerical aperture of 1.4 (0.7) was used. Images were collected using an aperture of 2-5 mm and a zoom of 1-4. Typically, 10-20 optical sections were taken at 0.5-µm intervals through the cells. The resolution of the confocal microscope in the x-y axis was 170-200 nm, and in the z axis it was 230-400 nm. Images of 768 × 520 pixels were obtained. Images were processed using Adobe Photoshop 3.0 (Adobe Systems Inc., Mountain View, CA) and printed using a Fujix Pictrography 3000 printer.

Results are expressed as mean ± S.E. Differences between groups were examined by an analysis of variance and a Student-Newman-Keuls test or a Bonferroni t test. A p < 0.05 was considered significant.

RESULTS AND DISCUSSION

The mechanisms of desensitization and resensitization of PAR-2-mediated responses are completely unexplored but are of considerable general interest given the unusual mechanism of irreversible activation of this receptor. It is important to understand these processes since desensitization of PAR-2 will prevent the uncontrolled stimulation of cells by trypsin, and resensitization is necessary for cells to recover their ability to respond to trypsin. Thus, the physiological effects of trypsin that are mediated by PAR-2, which may include regulation of pancreatic secretion, hormone secretion, smooth muscle contraction, and growth (20, 25), are dependent on PAR-2 desensitization and resensitization. Furthermore, the duration of the signal to trypsin may be an important determinant of long-term effects of trypsin, including effects on cell growth (12). Therefore, we examined the mechanism of desensitization and resensitization of cells expressing PAR-2 to trypsin and AP.

We studied these processes in two epithelial cell lines, KNRK cells transfected with human PAR-2 and hBRIE 380 cells that naturally express PAR-2. KNRK cells were chosen because we have previously investigated the mechanisms of desensitization and resensitization of the NK1-R expressed in these cells (4, 5, 37), thus enabling us to compare our results with PAR-2 with those for a thoroughly studied neurotransmitter receptor. However, observations of transfected cells may be unique to the cell line or be artifacts of receptor overexpression. Thus, we also studied hBRIE 380 cells, a highly differentiated, polarized cell line derived from the intestinal epithelium (26, 27). This line naturally expresses PAR-2 (20) and closely resembles epithelial cells of the small intestine. From a physiological standpoint, it is important to study desensitization and resensitization of PAR-2 in intestinal epithelial cells, since they are episodically exposed to trypsin during feeding and fasting (24).

We examined the effects of PAR-2 activation by trypsin or AP on [Ca²⁺]_i, measured using Fura-2/AM. Trypsin and AP stimulated an increase in [Ca²⁺]_i in KNRK-PAR-2 cells (Fig. 1, A and B), although trypsin was 7,800-fold more potent than AP (EC₅₀ trypsin = 2.3 nM, AP = 18 µM). Similarly, trypsin and AP increased [Ca²⁺]_i in hBRIE 380 cells (Fig. 1, C and D) (EC₅₀ of trypsin = 25.3 nM, EC₅₀ of AP = 17.0 µM). For both cell lines, the Ca²⁺ response was biphasic, with a rapid increase in [Ca²⁺]_i, and a slower decline to a sustained plateau.

To determine which phase of this response was due to mobilization of Ca²⁺ from intracellular stores and which phase depended on entry of extracellular Ca²⁺, we loaded cells with Fura-2/AM and studied them in Ca²⁺-free medium. In KNRK-PAR-2 cells, 100 nM trypsin stimulated a prompt increase in [Ca²⁺]_i, which rapidly declined to basal levels (Fig. 2A). When Ca²⁺ was added back to the medium to a final concentration of 2 mM, 1 min after exposure to agonist, [Ca²⁺]_i again increased for an extended plateau phase (Fig. 2B). In hBRIE 380 cells, the response to 100 nM trypsin also rapidly declined in Ca²⁺-free medium. In contrast to KNRK cells, the plateau was not completely abolished but just diminished. When Ca²⁺ was replenished, [Ca²⁺]_i again increased to establish a plateau phase similar to control cells (Fig. 2, D and E). Similar results were obtained with 100 µM AP in both cell lines (not shown).

We used inhibitors to characterize the channel in the plasma membrane of KNRK-PAR-2 cells and hBRIE 380 cells that was responsible for the influx of extracellular Ca²⁺. Treatment of cells with 10 µM amiloride or 10 µM nifedipine to inhibit voltage-sensitive N- and L-type channels, respectively, had no effect on the influx of Ca²⁺ (not shown). However, treatment with 10 µM SK&F 96365, which blocks the I_CRAC channel, abolished the influx in both cell lines (Fig. 2, C and F).

The results show that the prompt increase in [Ca²⁺]_i that follows activation of PAR-2 is due to release of Ca²⁺ from intracellular stores, since it was still observed in the absence of extracellular Ca²⁺. This phase of the response is rapidly attenuated despite the irreversible nature of PAR-2 activation. The extent of the plateau phase and the mechanism of its generation are dependent on the cell type. In KNRK-PAR-2 cells the plateau phase depends completely on the influx of extracellular Ca²⁺, whereas in hBRIE 380 cells it was only diminished in the absence of extracellular Ca²⁺. This indicates that in hBRIE 380 cells, in addition to influx of extracellular Ca²⁺, continued Ca²⁺ mobilization from intracellular stores must contribute to the plateau phase. The influx of extracellular Ca²⁺ is triggered by depletion of the intracellular Ca²⁺ stores, which activates the ``Ca²⁺ release-activated current'' through the putative I_CRAC channel that is inhibited by SK&F 96365 (31). Ca²⁺ entry from the extracellular fluid in response to PAR-2 activation uses the SK&F 96365-sensitive I_CRAC channel in both cell lines. Thus, PAR-2 is similar to the neurokinin 1 and 2 receptors, which respond to agonists by both mobilizing intracellular Ca²⁺ and by opening a channel in the plasma membrane that is sensitive to SK&F 96395 (5). The influx of Ca²⁺ from extracellular fluid is an important component of signal transduction, since Ca²⁺ influx refills the intracellular stores, modulates the spatiotemporal pattern of propagation of Ca²⁺ waves, and is necessary for secretion of mediators from some cells. Furthermore, the duration of signal generation by growth factors may be a critical element in determining cell cycle reentry (12, 38). This may be important for PAR-2, since hBRIE 380 cells and A549 cells (lung adenocarcinoma), which naturally express PAR-2, exhibit prolonged Ca²⁺ responses to trypsin and AP, and because PAR-2 activation inhibits growth of A549 cells (20).

Desensitization is the diminution of a biological response to repeated exposure to agonist or to continued presence of agonist. Cellular responses to trypsin that are mediated by PAR-2 could desensitize by (a) enzymatic cleavage of the receptor, which ensures that a single receptor molecule, once cleaved, cannot be reactivated by trypsin; (b) receptor phosphorylation and uncoupling from G-proteins; and (c) inhibition of more distal steps of intracellular signaling. In contrast, AP activates PAR-2 without receptor cleavage, so that responses to AP could desensitize by uncoupling from G-proteins and inhibition of later stages of signaling. To compare desensitization of PAR-2 with trypsin and AP, we measured Ca²⁺ responses to repeated application of both trypsin and AP.

We examined desensitization of PAR-2 by exposing KNRK-PAR-2 cells or hBRIE 380 cells sequentially to 100 nM trypsin or 100 µM AP at 2-min intervals, without an intervening wash. In KNRK-PAR-2 cells, trypsin reduced or abolished the Ca²⁺ response to a second exposure to trypsin or AP given 2 min later (Fig. 3, A and B, Table I). AP substantially reduced, but did not abolish, the Ca²⁺ response to a second exposure to AP but did not affect the response to trypsin. In hBRIE 380 cells, trypsin abolished the response to a second exposure to trypsin and AP (Fig. 4, A and B, Table I). AP abolished the response to a second exposure to AP and markedly inhibited the response to trypsin. Thus, activation of PAR-2 by trypsin or AP induces homologous desensitization of the Ca²⁺ response. Trypsin causes a more substantial desensitization, especially in KNRK-PAR-2 cells.

Table I. Desensitization of the Ca2+ response of KNRK-PAR-2 cells and hBRIE 380 cells to trypsin, AP, and bradykinin Cells were challenged with 100 nM trypsin, 100 µM AP, or 1 µM bradykinin (first agonist) and then challenged with 100 nM trypsin, 100 µM AP, or 1 µM bradykinin (second agonist) 2 min later without an intervening wash. Values are expressed as the difference between the basal and the maximal [Ca2+]i (determined as the 340/380 nm ratio) and are the mean ± S.E. of three experiments. First agonist Response Second agonist Response KNRK-PAR-2 cells 100 nM trypsin 1.81  ± 0.21 100 nM trypsin NDa,b 100 nM trypsin 100 µM AP 0.20  ± 0.1c 100 nM trypsin 1 µM bradykinin 1.71  ± 0.19 100 µM AP 1.21  ± 0.11 100 nM trypsin 1.82  ± 0.10 100 µM AP 100 µM AP 0.40  ± 0.01c 100 µM AP 1 µM bradykinin 1.95  ± 0.14 1 µM bradykinin 2.25  ± 0.15 100 nM trypsin 0.66  ± 0.07a 1 µM bradykinin 100 µM AP 0.54  ± 0.06c 1 µM bradykinin 1 µM bradykinin NDd hBRIE 380 cells 100 nM trypsin 1.36  ± 0.05 100 nM trypsin NDb 100 nM trypsin 100 µM AP NDc 100 nM trypsin 1 µM bradykinin 0.08  ± 0.02d 100 µM AP 0.81  ± 0.07 100 nM trypsin 0.09  ± 0.01b 100 µM AP 100 µM AP NDc 100 µM AP 1 µM bradykinin 0.14  ± 0.02d 1 µM bradykinin 0.19  ± 0.01 100 nM trypsin 0.97  ± 0.18b 1 µM bradykinin 100 µM AP 0.64  ± 0.11 1 µM bradykinin 1 µM bradykinin NDd a  ND, not detectable. b  p < 0.05 for cells exposed to trypsin (second agonist) compared with trypsin (first agonist). c  p < 0.05 for cells exposed to AP (second agonist) compared with AP (first agonist). d  p < 0.05 for cells exposed to bradykinin (second agonist) compared with bradykinin (first agonist).

Bradykinin mobilizes Ca²⁺ in KNRK cells and hBRIE 380 cells. We investigated whether activation of PAR-2 with trypsin or AP caused heterologous desensitization of responses to bradykinin by exposing cells to 100 nM trypsin, 100 µM AP, or carrier (control) and then challenging them with a test dose of 1 µM bradykinin 2 min later. In both cell lines, trypsin and AP slightly reduced the response to bradykinin (Fig. 3, A and B, Fig. 4, A and C, Table I), although this was statistically significant only for the hBRIE 380 cells. Thus, activation of PAR-2 may cause heterologous desensitization of bradykinin responses in hBRIE 380 cells. However, the Ca²⁺ response to bradykinin in hBRIE 380 cells was <10% of that in KNRK-PAR-2 cells, suggesting that hBRIE 380 cells have fewer receptors. This may explain the difference in desensitization of the bradykinin response in these cells lines by trypsin and AP. The intracellular Ca²⁺ stores were not substantially depleted by trypsin or AP, since bradykinin was still able to mobilize Ca²⁺. This suggests that homologous desensitization of the Ca²⁺ response to trypsin and AP is not principally due to depletion of these pools, although some diminution of available stores could contribute to desensitization.

In a similar manner, we examined whether bradykinin desensitized subsequent Ca²⁺ responses to bradykinin, trypsin, or AP. In KNRK-PAR-2 cells and hBRIE 380 cells, 1 µM bradykinin abolished the response to 1 µM bradykinin, applied 2 min later (Fig. 3C, Fig. 4C, Table I). Bradykinin reduced, but did not abolish, the response to 100 nM trypsin or 100 µM AP. Therefore, the bradykinin response shows pronounced homologous desensitization and causes moderate heterologous desensitization of PAR-2.

Together, our results show that repeated application of trypsin or AP causes marked homologous desensitization of PAR-2. Desensitization mainly occurs due to specific alteration of the receptor rather than depletion of intracellular mediators, because bradykinin was still capable of mobilizing Ca²⁺. Trypsin causes a larger desensitization of PAR-2 than AP, although the magnitude of the Ca²⁺ response to both agonists is generally similar. This is expected since trypsin, but not AP, cleaves PAR-2, which ensures that a single receptor molecule, once cleaved, cannot respond again to trypsin. We determined the rate of this cleavage by measuring loss of Flag immunoreactivity from the cell surface.

We quantified Flag immunoreactivity at the surface of KNRK-PAR-2 cells expressing PAR-2 with an NH₂-terminal Flag epitope (Fig. 5A). We exposed cells to 10 nM trypsin, 100 µM AP or carrier (control) for 2, 5, 10, or 30 min, washed and incubated them at 4 °C with the Flag M1 antibody followed by a ¹²⁵I-sheep anti-mouse IG. Exposure to trypsin resulted in a rapid loss of surface Flag immunoreactivity, to only 72 ± 2% of that observed in control cells after 2 min and 36 ± 4.8% of controls at 30 min (Fig. 5B). In contrast, exposure to AP resulted in a far slower decline of surface immunoreactivity, to 80 ± 3.3% of controls at 30 min. Loss of surface Flag immunoreactivity after trypsin could be due to receptor cleavage and receptor internalization, whereas loss after AP would reflect receptor internalization only. We examined the relative contributions of receptor cleavage and receptor endocytosis to loss of surface Flag immunoreactivity using endocytic inhibitors. After 2 min of exposure to 10 nM trypsin, surface Flag immunoreactivity was 96 ± 17% in phenylarsine oxide-treated cells and 92 ± 0.2% in cells treated with hypertonic sucrose, compared with 75 ± 2% in controls. After 30 min of exposure to 10 nM trypsin, surface Flag immunoreactivity was 75 ± 8% in phenylarsine oxide-treated cells and 70 ± 9% in cells treated with hypertonic sucrose, compared with 42 ± 4% in controls. In contrast, phenylarsine oxide and hypertonic sucrose abolished the loss of surface Flag immunoreactivity to 100 µM AP at all times studied. We conclude that the loss of surface Flag immunoreactivity after trypsin is principally mediated by receptor cleavage and after AP is solely mediated by receptor internalization. We have previously shown that these treatments also prevent internalization of the NK1-R, examined in a similar manner in the same cell line (4).

Receptor endocytosis may contribute to desensitization by removing receptors from the plasma membrane to a compartment that is inaccessible to agonists in the extracellular fluid. We directly examined this issue by measuring desensitization of the Ca²⁺ response to sequential doses of 100 nM trypsin or 100 µM AP in KNRK-PAR-2 cells and hBRIE 380 cells treated with phenylarsine oxide to inhibit endocytosis. In phenylarsine oxide-treated cells, trypsin still reduced to undetectable levels the Ca²⁺ response to a second exposure to trypsin given 2 min later (Fig. 5C, Table II). Similarly, AP desensitized the response to a second AP challenge (Table II). Therefore, desensitization to trypsin is unaffected by inhibition of receptor endocytosis, and internalization is not the main mechanism of rapid desensitization of PAR-2. Furthermore, we observed that there was almost complete desensitization of PAR-2-mediated Ca²⁺ mobilization when the interval between sequential application of trypsin was 2 min (Fig. 3, Table I). At this time there was little internalization of PAR-2 in KNRK-PAR-2 cells (Fig. 5B), which also suggests that endocytosis is not the principal mechanism of desensitization. We have similarly shown that internalization of the NK1-R is not mediated by endocytosis, although this receptor is rapidly internalized after agonist binding (4, 5).

Table II. Effects of phenylarsine oxide on desensitization of the Ca2+ response of KNRK-PAR-2 cells and hBRIE 380 cells to trypsin and AP Cells were treated with phenylarsine oxide and challenged with 100 nM trypsin or 100 µM AP (first agonist), followed by 100 nM trypsin or 100 µM AP (second agonist) 2 min later without an intervening wash. Values are expressed as the difference between the basal and the maximal [Ca2+]i (determined as the 340/380 nm ratio) and are the mean ± S.E. of six experiments. First agonist Response Second agonist Response KNRK-PAR-2 cells 100 nM trypsin 1.73  ± 0.16 100 nM trypsin NDa 100 µM AP 0.89  ± 0.11 100 µM AP 0.30  ± 0.03 hBRIE 380 cells 100 nM trypsin 1.40  ± 0.08 100 nM trypsin ND 100 µM AP 1.58  ± 0.11 100 µM AP 0.01  ± 0.01 a  ND, not detectable.

Homologous desensitization of many G-protein-coupled receptors is mediated by receptor phosphorylation and uncoupling from G-proteins (1, 2). GRKs play an important role in this mechanism. GRKs phosphorylate agonist-occupied receptors, and -arrestin binds the phosphorylated receptor to disrupt the interaction between the receptor and G-proteins and cause receptor-specific, homologous desensitization. Effector kinases (PKC, PKA) participate in desensitization at multiple levels. First, they contribute to homologous desensitization by phosphorylating receptors that stimulate their activation. Second, they mediate heterologous desensitization by phosphorylating different receptors in an agonist-independent manner. Third, they exert negative feedback control of distal signaling mechanisms, thereby shutting down their activation.

PAR-2 is coupled to activation of phospholipase C ₁, which hydrolyzes phosphatidylinositol 4,5-biphosphate, generating inositol 1,4,5-triphosphate and diacylglycerol. Inositol 1,4,5-triphosphate stimulates the release of Ca²⁺ from intracellular stores, which induces the influx of extracellular calcium, whereas diacylglycerol activates PKC (17, 20, 39). PKC could mediate desensitization of PAR-2 by receptor phosphorylation and uncoupling from G-proteins or by negative feedback control of signaling mechanisms. We used PKC inhibitors and activators to investigate the contribution of PKC to desensitization of PAR-2-mediated Ca²⁺ mobilization.

To examine the effects of inhibition of PKC, KNRK-PAR-2 cells or hBRIE 380 cells were incubated with the PKC inhibitor GF 109203X, inactive bisindolylmaleimide V, or carrier (controls) for 20 min before challenge with 10-50 nM trypsin or 100 µM AP. In both cell lines, GF 109203X increased the magnitude of the Ca²⁺ response to trypsin and AP in a concentration-dependent manner (Fig. 6, A and B). The threshold was 1 µM, and the response was maximal at 10 µM GF 109203X, when it was up to 200% of the control. Thus, inhibition of PKC increases the size of the Ca²⁺ response to trypsin and AP in KNRK-PAR-2 cells and hBRIE 380 cells.

Acute treatment of cells with PDBu stimulates PKC. Therefore, we studied the effects of acute activation of PKC, by treating KNRK-PAR-2 cells or hBRIE 380 cells with PDBu, inactive 4-phorbol-12,13-dicanoate, or carrier (controls) for 2 min before challenge with 10-50 nM trypsin or 100 µM AP. In KNRK-PAR-2 cells and hBRIE 380 cells, PDBu inhibited the peak and plateau phase of the Ca²⁺ response to trypsin and AP in a concentration-related manner (Fig. 6, A and C). The threshold concentration was 10 nM, and the response was abolished by 10 µM PDBu. When cells were simultaneously exposed to 10 µM of the PKC activator PDBu and to 10 µM of the PKC inhibitor GF 109203X, the inhibitory effects of the activator were abolished (not shown). Therefore, this inhibition was caused by activation of PKC. We conclude that acute activation of PKC decreases the Ca²⁺ response to trypsin and AP in KNRK-PAR-2 cells and hBRIE 380 cells.

PKC is known to regulate the putative I_CRAC channel, which is also responsible for the influx of extracellular Ca²⁺ in KNRK-PAR-2 cells in response to trypsin. We examined whether PKC regulates this channel in KNRK-PAR-2 cells. Cells were exposed to 100 nM trypsin and then treated with 5 µM GF 109203X, 200 nM PDBu, or carrier (control) at the peak of the Ca²⁺ response. In control experiments, there was a distinct plateau phase in the response (Fig. 7A). This was enhanced by inhibition of PKC with GF 109203X but abolished by activation of PKC with PDBu (Fig. 7A). We conclude that the I_CRAC channel, which is activated by PAR-2, is regulated by PKC.

Thapsigargin inhibits the Ca²⁺-ATPase that is responsible for the reuptake of cytosolic Ca²⁺ by intracellular stores, resulting in an increase in [Ca²⁺]_i. Depletion of intracellular Ca²⁺ stores by thapsigargin also stimulates the influx of extracellular Ca²⁺ through the I_CRAC channel that is inhibited by SK&F 96365 (30). To determine whether the effects of PKC inhibitors and activators on PAR-2-mediated Ca²⁺ mobilization were due to a receptor-dependent or -independent mechanism, we investigated the effects of these drugs on the Ca²⁺ response to thapsigargin. Thapsigargin (1 µM) stimulated a prompt and sustained increase in [Ca²⁺]_i in KNRK-PAR-2 cells (Fig. 7B). The plateau phase of the response was absent in Ca²⁺-free medium but restored when Ca²⁺ was added to extracellular fluid (not shown), which indicates that the plateau is due to influx of extracellular Ca²⁺. This restoration was abolished by 1 µM SK&F 96365 (not shown), indicating that Ca²⁺ enters through the I_CRAC channel. In contrast to the response to trypsin, the peak increase in [Ca²⁺]_i induced by thapsigargin was unaffected by pretreatment with 5 µM GF 109203X or 200 nM PDBu (not shown). To examine the involvement of PKC in regulating the influx of extracellular Ca²⁺, we exposed KNRK-PAR-2 cells to 1 µM thapsigargin and treated cells with 5 µM GF 109203X or 200 nM PDBu at the peak of the Ca²⁺ response. In controls, the plateau (measured 200 s after the peak) was 73 ± 3% of the peak response to thapsigargin. In cells treated with GF 109203X the plateau was 111 ± 9% of the peak (p < 0.05 to controls), and in PDBu-treated cells the plateau was 63 ± 6% of the peak (not significantly different from controls) (Fig. 7B). Thus, PKC modulates I_CRAC regardless of whether it is activated by receptor-dependent (PAR-2) or -independent (thapsigargin) mechanisms.

PKA contributes to desensitization of receptors coupled to adenylate cyclase, such as the ₂AR. In response to low agonist concentrations, PKA phosphorylates the ₂AR at a site in the third intracellular loop that is essential for coupling to G_s (1, 2). PAR-2 also has a PKA consensus site in the third intracellular loop. Although PAR-2 is not known to be coupled to adenylate cyclase, we examined whether PKA could contribute to heterologous desensitization of PAR-2. Treatment of KNRK-PAR-2 cells or hBRIE 380 cells with 10 µM forskolin for 2 min, to activate PKA, had no effect on the Ca²⁺ response to 100 nM trypsin (not shown). Similarly, inhibition of PKA by treatment of cells with 100 nM H-89 for 2 min did not affect the trypsin response. Thus, we have no evidence that PKA mediates heterologous desensitization of PAR2.

Together, our results indicate that PKC plays an important role in regulating PAR-2-mediated Ca²⁺ mobilization. There is good precedent for PKC terminating signaling and mediating desensitization of G-protein-coupled receptors. Thus, acute treatment with phorbol esters inhibits agonist-stimulated Ca²⁺ mobilization and inositol 1,4,5-triphosphate generation mediated by other receptors (33, 40, 41, 42, 43), including the Th-R (44). Our finding that acute inhibition of PKC enhanced the Ca²⁺ response to trypsin is supported by the observation that PKC inhibitors enhance cellular responses to bradykinin and angiotensin II (40, 43). It is possible that inhibition of basal PKC activity alters the balance of phosphorylation and dephosphorylation of components of the signaling mechanism in favor of dephosphorylation, thereby enhancing second messenger generation. However, receptors are not generally the targets of basal phosphorylation, since the Th-R, the formylpeptide, and the C5a chemoattractant receptor are not phosphorylated without agonist stimulation (11, 12, 41). Alternatively, PKC inhibitors may enhance the Ca²⁺ response to trypsin by preventing the contribution of PKC to homologous desensitization of PAR-2.

PKC could participate in heterologous desensitization of PAR-2 if it is activated by a different receptor or in homologous desensitization of PAR-2 if it is directly activated by PAR-2. PKC may act at the level of the receptor or at more distal steps in the signaling pathway. Thus, PKC may phosphorylate PAR-2 and disrupt its interaction with G-proteins. Although studies of phosphorylation of PAR-2 have not yet been reported, human PAR-2 contains 10 Ser and Thr residues in the third intracellular loop and the COOH tail within PKC consensus sites, suggesting that this receptor may be a target for PKC. The closely related Th-R also has 7 Ser and Thr residues within PKC consensus sites in the third intracellular loop and the COOH-terminal tail. Furthermore, phorbol esters inhibit signaling and stimulate phosphorylation of the Th-R by activating PKC (11, 12, 44). However, GRKs also play a dominant role in homologous desensitization of the Th-R and phosphorylate the receptor faster and to a larger extent than PKC (11, 12). Thus, thrombin-stimulated phosphorylation of its receptor is not blocked by inhibition or down-regulation of PKC. Coexpression of GRK-3 with the Th-R inhibits receptor signaling by thrombin, which depends on the presence of potential phosphorylation sites in the COOH tail (11). The observation that GRK-3-insensitive mutants of the Th-R still undergo a rapid shut off suggests that the Th-R also relies on an additional desensitization mechanism, perhaps using PKC sites in the intracellular loops. Comparison of human Th-R and human PAR-2 reveals that the Th-R has 7 Ser and Thr residues in the COOH tail that are in GRK consensus sites, whereas PAR-2 has just one Ser residue in the third intracellular loop that is a potential target for a GRK. This observation, together with the preponderance of potential PKC sites in the COOH tail of PAR-2 and the marked effects of PKC activators and inhibitors on desensitization, suggests that PKC may play a more important role in homologous desensitization of PAR-2 than of the Th-R. However, it is likely that both GRKs and effector kinases contribute to homologous desensitization, possibly under different circumstances, as they do for the ₂-AR (1, 2).

PKC may also act on G-proteins, phospholipase C, or later steps in the signaling pathway to regulate PAR-2-mediated Ca²⁺ responses. Arguing that the G-protein is not the principal target of PKC is the finding that G_q-subunit, which at least partially transduces Th-R activation (45), is not phosphorylated in response to phorbol esters under conditions where the G_z-subunit is phosphorylated (46). Furthermore, we and others (42) found that the first phase of the Ca²⁺ response to thapsigargin, which acts distal to phospholipase C, is not sensitive to phorbol esters and is thus not regulated by PKC. Phospholipase C is a likely target for PKC, since phorbol esters inhibit both Ca²⁺ mobilization and formation of inositol 1,4,5-triphosphate, and phospholipase C is phosphorylated in response to activation of PKC by phorbol esters at a site that prevents interaction with G-proteins (47). Phorbol ester-induced heterologous internalization, recently demonstrated for the Th-R, may still be another mechanism of PKC-mediated desensitization (48).

We found that PKC also modulates the plateau phase of PAR-2-mediated Ca²⁺ response, which depends on the influx of extracellular Ca²⁺ through a capacitative mechanism. This was observed whether the Ca²⁺ influx was stimulated by PAR-2 or thapsigargin, which suggests that I_CRAC is directly regulated by PKC. Similarly, phorbol esters accelerate and PKC inhibitors delay the inactivation of the I_CRAC channel in basophilic leukemia cells (38). Ca²⁺ entry mediated by the adenosine A₃ receptor is also determined, at least in part, by PKC (38). Furthermore, the PKC isoenzyme is a specific negative regulator of Th-R-mediated Ca²⁺ entry in human megakaryoblastic HEL cells, since antisense cDNA to this isoenzyme promotes Th-R-mediated Ca²⁺ entry without altering release of Ca²⁺ from intracellular stores (49). Regulation of I_CRAC through PKC is of considerable functional importance given the importance of Ca²⁺ influx in intracellular signaling (12, 38).

Agonist-induced endocytosis contributes to desensitization of some receptors (2). Our experiments to measure surface Flag immunoreactivity (Fig. 5) indicate that PAR-2 internalizes after activation by trypsin or AP but that this is not the principal mechanism of rapid desensitization. However, endocytosis may contribute to a second phase of desensitization by depleting the plasma membrane of receptors that are accessible to hydrophilic agonists in the extracellular fluid. To examine this component of desensitization, we localized PAR-2 in KNRK-PAR-2 cells by immunofluorescence and confocal microscopy. Cells were incubated with 10 nM trypsin or 100 µM AP for 0-120 min and fixed, and PAR-2 was localized using the HA.11 antibody to the COOH-terminal 12CA5 epitope. Before exposure to agonists, PAR-2 was detected at the plasma membrane, in a prominent perinuclear store, and in small vesicles that were scattered throughout the cytoplasm (Fig. 8A). When cells were exposed to trypsin for 5 or 15 min, PAR-2 was also found in numerous superficial vesicles that were located between the plasma membrane and the nucleus, and there was a slight diminution of the intensity of labeling at the plasma membrane (Fig. 8B). After 30, 60, or 120 min, PAR-2 was in large, prominent vesicles that were scattered throughout the cell (Fig. 8C). In a similar manner, AP caused internalization of PAR-2 initially into small superficial vesicles and later into larger vesicles scattered throughout the cell (not shown). Thus, activated PAR-2 is rapidly internalized.

To identify the prominent perinuclear pool of PAR-2 in cells not previously exposed to trypsin, we simultaneously stained cells with the 12CA5 antibody and with an antibody to mannosidase II, to mark the Golgi apparatus. PAR-2 was detected in a perinuclear tubulovesicular network (Fig. 9A) of the same size, shape, and location as that stained by the mannosidase II antibody (Fig. 9B). This colocalization was confirmed by superimposition of optical sections, giving a yellow color at sites of co-localization (Fig. 9C). The small scattered vesicles that contained PAR-2 did not contain mannosidase II (Fig. 9, A-C). Treatment of cells with 10 µg/ml brefeldin A for 30 min disrupted the perinuclear Golgi pools of PAR-2 and mannosidase II, and both proteins were detected in numerous vesicles that were scattered throughout the cell (Fig. 9, D-F). PAR-2 and mannosidase II were not usually co-localized in vesicles after treatment of cells with brefeldin A. The reason for this is unknown but it suggests that brefeldin A has different effects on the distribution of PAR-2 and mannosidase II. Indeed, examination by immunoelectron microscopy indicates that brefeldin A differentially alters the distribution of proteins in the cis and trans Golgi cisternae (50). Thus, brefeldin A causes redistribution of mannosidase II from the cis and medial Golgi cisternae to cisternae and vesicular and reticular elements of the endoplasmic reticulum in multiple cell types, including rat kidney epithelial cells (50). In contrast, the distribution of glycoproteins (detected using wheat germ agglutinin) and resident membrane proteins of the trans Golgi cisternae is unaffected by brefeldin A. Similar studies by immunoelectron microscopy will be required to fully define the intracellular location of PAR-2 and further explore the effects of brefeldin A. 

To determine if the superficial vesicles containing PAR-2 that were observed immediately after treatment with trypsin or AP were early endosomes, we simultaneously localized PAR-2 and the transferrin receptor, an endosomal marker, 5 or 10 min after exposure to trypsin or AP. At this time, many of the superficial vesicles containing PAR-2 (Fig. 10A) were of the same size, shape, and location as vesicles that were stained for the transferrin receptor (Fig. 10B). Colocalization was confirmed by superimposition of confocal images (Fig. 10C). However, at later time points (30-60 min) there was minimal colocalization of PAR-2 and the transferrin receptor (not shown). Thus, PAR-2 initially enters an endosomal compartment but is then sorted into different organelles.

After irreversible cleavage and activation by trypsin, we expected that PAR-2 would not be reused but targeted to lysosomes for degradation. AP activates PAR-2 directly, without receptor cleavage, so that the receptor could be reused. To determine if PAR-2 is ultimately destined for lysosomes after irreversible and reversible activation, we simultaneously localized PAR-2 and GM10, a lysosomal marker, 30, 60, and 120 min after trypsin or AP treatment. Cells were treated with 10 mM NH₄Cl, to prevent lysosomal acidification and to inhibit degradation of PAR-2 and possible loss of receptor immunoreactivity. Large vesicles containing PAR-2 were scattered throughout the cytosol after exposure to trypsin or AP for 30, 60, or 120 min (Fig. 11, A and D). Many of these vesicles were lysosomes (Fig. 11, B and E). Close inspection of superimposed confocal images showed that whereas GM10 was located at the lysosomal membrane, PAR-2 was found in the center of the lysosomes (Fig. 11, C and F). Thus, PAR-2 is sorted to lysosomes after internalization, and this occurs whether the receptor is irreversibly activated by trypsin or reversibly activated by AP.

Together, our results show that PAR-2 internalizes after activation by trypsin or AP, rapidly enters an endosomal compartment, and then moves to lysosomes, where it is probably degraded. We do not know at present whether PAR-2 is similarly targeted in hBRIE 380 cells, since we lack antibodies to the native receptor. However, the Th-R is similarly targeted in transfected cells and in cells that naturally express this receptor (8, 15). Similarly, the Th-R is also internalized after activation by thrombin and AP, briefly resides in endosomes, and is then directed to lysosomes (8, 15, 16, 51). Although some cleaved Th-Rs do recycle, they are unable to respond again to thrombin and cannot contribute to resensitization (15). Thus, the targeting of the proteinase-activated receptors is distinctly different from that of the receptors for hormones and neurotransmitters. The ₂AR and NK1-R rapidly internalize after binding agonists into early endosomes and remain in endosomes until they recycle (3, 4). These receptors are rarely targeted to lysosomes. The explanation for this difference in behavior between the related receptors for neurotransmitters and proteases is unknown, but specific receptor motifs are believed to define the targeting of receptors within cells. For example, tyrosine-containing motifs in the COOH tail and transmembrane domains of G-protein-coupled receptors and single transmembrane receptors are required for endocytosis (52). PAR-2 also contains such conserved motifs. The identification of domains involved in intracellular sorting and recycling has received less attention. Recycling may be a default pathway for internalized receptors, since receptors and lipids recycle at similar rates (53), whereas targeting to lysosomes may require specific sorting motifs. We do not know the identity of these motifs in PAR-2, although this could be addressed using chimeras of PAR-2 and the recycling NK1-R. Since PAR-2 is targeted to lysosomes after cleavage with trypsin or AP, this targeting is independent of the mechanism of activation, and the extreme NH₂ terminus of PAR-2, which is removed by cleavage, is not involved in lysosomal sorting.

The difference in intracellular destination between internalized protease-activated receptors and neurotransmitter receptors suggests that they resensitize by distinctly different mechanisms.

Resensitization is important since it permits cells to recover or retain their ability to respond to trypsin. This may be necessary for the effects of trypsin on cell growth, which may require long term activation of PAR-2 (20). Therefore, we studied the time course of homologous desensitization and resensitization of PAR-2-mediated Ca²⁺ mobilization to repeated challenge with trypsin or AP.

We exposed cells to a desensitizing dose of 100 nM trypsin, 100 µM AP, or carrier (control) for 1 min, washed them, and then challenged them with a test dose of 100 nM trypsin or 100 µM AP at 2-60 min after the first exposure. In both KNRK-PAR-2 cells and hBRIE 380 cells, preexposure to trypsin reduced the response to a test dose of trypsin given 2 min after the first dose to ~15% of that observed in carrier-treated controls (Fig. 12, A and B). The response resensitized to control levels when the interval between trypsin doses was 30-60 min. Similarly, in both cell lines, the Ca²⁺ response to a test dose of AP desensitized when the interval between AP doses was 2-10 min, and resensitization was complete after 30-60 min. However, the magnitude of the desensitization to repeated exposure to trypsin was greater than desensitization to repeated exposure to AP.

We investigated the mechanism of resensitization of PAR-2 to repeated application of trypsin by using inhibitors of protein synthesis and receptor recycling and by disrupting Golgi pools of PAR-2. We exposed KNRK-PAR-2 cells to a desensitizing dose of 100 nM trypsin or carrier (control) for 1 min, washed them, and then challenged them with a test dose of 100 nM trypsin 90 min later. Some cells were exposed twice to desensitizing doses of 100 nM trypsin or carrier (control) for 1-min periods at intervals of 90 min and were then challenged with a test dose of 100 nM trypsin 90 min after the final exposure. The Ca²⁺ response to a test dose of trypsin resensitized to over 75% of that of controls by 90 min after exposure to one or two desensitizing doses of trypsin (Fig. 13A). To determine if the integrity of the stores of PAR-2 in the Golgi apparatus was important for resensitization, we incubated cells with 10 µg/ml brefeldin A, which we found to disrupt the stores of PAR-2 in KNRK cells (Fig. 9, D-F). Brefeldin A treatment slightly diminished resensitization to one desensitizing dose of trypsin but markedly inhibited resensitization to two doses of trypsin to less than 15% of controls (Fig. 13A). To determine if new protein synthesis was important for resensitization, we incubated cells with 70 µM cycloheximide, which we have shown inhibits protein synthesis in KNRK cells (4). Cycloheximide-treatment had no effect on resensitization to one desensitizing dose of trypsin but inhibited resensitization to two doses of trypsin to 50% of controls (Fig. 13A). We have previously shown that bafilomycin A₁, which prevents endosomal acidification and recycling of the NK1-R in KNRK cells, also inhibits resensitization of this receptor (4, 5). Therefore, to determine if endosomal acidification was important for PAR-2 resensitization, we incubated cells with 1 µM bafilomycin A₁. Bafilomycin A₁-treatment had no effect on resensitization to one or two desensitizing doses of trypsin (Fig. 13A).

In a similar manner, we examined resensitization of PAR-2 in hBRIE 380 cells after one or two desensitizing doses of trypsin. The Ca²⁺ response to a test dose of trypsin completely resensitized by 60 min after a single dose of trypsin but resensitized to only 50% of control levels by 60 min after exposure to two doses of trypsin, given at 60-min intervals (Fig. 13B). Brefeldin A had little effect on resensitization after one dose of trypsin but markedly inhibited resensitization after two doses of trypsin, as it did in KNRK-PAR-2 cells (Fig. 13B). Cycloheximide also reduced resensitization, especially after one dose of trypsin, and bafilomycin A₁ had little effect on resensitization.

Our results indicate that resensitization after one or two exposures to trypsin depends upon pools of presynthesized PAR-2, both in KNRK-PAR-2 cells and hBRIE 380 cells. After a single exposure, resensitization still occurred after disruption of the Golgi stores with brefeldin A, suggesting that the first-used stores of PAR-2 are not disrupted by brefeldin A. These first-used stores may represent a source of readily available PAR-2 present in a post-Golgi compartment. Indeed, PAR-2 was detected in small superficial vesicles in KNRK-PAR-2 cells (see Fig. 8A). However, resensitization after two exposures to trypsin was markedly attenuated by brefeldin A. This indicates that the first-used stores are depleted by the first dose of trypsin, and Golgi pools are the second-used stores of PAR-2 that are used for resensitization after two or more exposures to trypsin. Resensitization after two exposures to trypsin was attenuated by cycloheximide, suggesting that the continued synthesis of PAR-2 is required to replenish receptors after multiple rounds of receptor cleavage and resensitization in both cell lines. Cycloheximide also reduced resensitization after one exposure to trypsin in hBRIE 380 but not KNRK-PAR-2 cells, suggesting that new receptor synthesis is more important in hBRIE 380 cells. Another difference between the cell lines is that there was diminished resensitization of hBRIE 380 cells after two exposures to trypsin, which may reflect exhaustion of intracellular pools of PAR-2 or an inability of new PAR-2 synthesis to replenish the cleaved receptors. Endosomal acidification, which is critical for recycling and resensitization of neurotransmitter receptors (4, 5), is not required for resensitization of PAR-2, since bafilomycin did not diminish resensitization. Resensitization of responses to thrombin is associated with depletion of intracellular stores of the Th-R and also requires new protein synthesis (8). Therefore, like the Th-R, PAR-2 resensitizes by mobilizing intracellular pools of receptor and by synthesis of new receptors. In contrast, new protein synthesis is not necessary for resensitization of the NK1-R (54). Furthermore, there are minimal intracellular pools of the NK1-R in transfected KNRK cells or in endothelial cells (4, 37). This receptor resensitizes by recycling, since resensitization is inhibited by bafilomycin A₁ (5).

PAR-2 is expressed in high levels in the gastrointestinal tract and pancreas (17, 20). In the small intestine PAR-2 mRNA is expressed in surface epithelial cells, especially in the upper two-thirds of the villi. If PAR-2 is localized at the apical membrane of these cells, it may be activated by pancreatic trypsin in the intestinal lumen and could thus mediate some of the effects of trypsin in the small intestine. These include the negative feedback regulation of pancreatic exocrine secretion by luminal trypsin (21) and effects of trypsin on hormone secretion and prostaglandin generation (23, 24). Pancreatic trypsin is secreted in an episodic manner, with high levels after feeding and lower levels between meals. Thus, PAR-2 would be activated during the postprandial period by luminal trypsin. Our results in transfected KNRK cells and in hBRIE 380 cells, which closely resemble intestinal epithelial cells in vivo, indicate that PAR-2 would rapidly desensitize by a combination of (a) enzymatic cleavage, which would ensure irreversible activation of PAR-2, and (b) PKC-mediated mechanisms, which may include receptor phosphorylation and uncoupling, as well as effects on more distal components of the signaling pathway, which together would terminate the signal. Between meals, when trypsin secretion is low, epithelial cells would be able to replenish surface PAR-2 by mobilizing intracellular pools and by synthesis of new receptors. Elucidation of the mechanisms of desensitization and resensitization of PAR-2 in other tissues must await identification of the physiological activators and further characterization of its biological functions.