Trypsin IV, a novel agonist of protease-activated receptors 2 and 4.

Certain serine proteases signal to cells by cleaving protease-activated receptors (PARs) and thereby regulate hemostasis, inflammation, pain and healing. However, in many tissues the proteases that activate PARs are unknown. Although pancreatic trypsin may be a physiological agonist of PAR(2) and PAR(4) in the small intestine and pancreas, these receptors are expressed by cells not normally exposed pancreatic trypsin. We investigated whether extrapancreatic forms of trypsin are PAR agonists. Epithelial cells lines from prostate, colon, and airway and human colonic mucosa expressed mRNA encoding PAR(2), trypsinogen IV, and enteropeptidase, which activates the zymogen. Immunoreactive trypsinogen IV was detected in vesicles in these cells. Trypsinogen IV was cloned from PC-3 cells and expressed in CHO cells, where it was also localized to cytoplasmic vesicles. We expressed trypsinogen IV with an N-terminal Igkappa signal peptide to direct constitutive secretion and allow enzymatic characterization. Treatment of conditioned medium with enteropeptidase reduced the apparent molecular mass of trypsinogen IV from 36 to 30 kDa and generated enzymatic activity, consistent with formation of trypsin IV. In contrast to pancreatic trypsin, trypsin IV was completely resistant to inhibition by polypeptide inhibitors. Exposure of cell lines expressing PAR(2) and PAR(4) to trypsin IV increased [Ca(2+)](i) and strongly desensitized cells to PAR agonists, whereas there were no responses in cells lacking these receptors. Thus, trypsin IV is a potential agonist of PAR(2) and PAR(4) in epithelial tissues where its resistance to endogenous trypsin inhibitors may permit prolonged signaling.

We investigated whether extrapancreatic trypsins are agonists of PAR 2 and PAR 4 . PAR 2 is present in the kidney, pancreas, stomach, intestine, airway, skin, bladder, and brain, where it is localized to epithelial and endothelial cells, myocytes, fibroblasts, immune cells, neurons, and glial cells (2,3,9,10,14). PAR 4 is expressed in the liver, small intestine, pancreas, lung, placenta, thyroid, and prostate, and in megakaryocytes (2,3,6,7,15). Pancreatic trypsin may be a physiologically relevant agonist of PAR 2 and PAR 4 in the lumen of the small intestine, since trypsin can signal to enterocytes by cleaving PAR 2 at the apical membrane (16), and in the inflamed pancreas, when there is premature activation of trypsinogen, which can signal to acinar or ductal cells (10,17). However, given the widespread distribution of PAR 2 and PAR 4 , it is unlikely that pancreatic trypsin is a physiologically relevant agonist in tissues other than the small intestine and pancreas.
Trypsinogens I, II, and mesotrypsinogen are the major trypsinogen genes in the human pancreas (18,19). These genes encode proteins with a signal sequence, an activation peptide and a catalytic domain. Enteropeptidase cleaves (2) at the DDDDK2I site of the zymogens to release active trypsins (20,21), which are highly homologous and share critical active site residues. Trypsinogens I and II are the principal forms of trypsin that are secreted in pancreatic juice, activated by enteropeptidase in the small intestine and participate in digestion. Mesotrypsinogen is a minor trypsinogen in human pancreatic juice (22). However, mesotrypsinogen is uniquely susceptible to activation by lysosomal cathepsin B and may thus be prematurely activated in the inflamed pancreas (23). Moreover, mesotrypsin is unaffected by polypeptide trypsin inhibitors such as soybean trypsin inhibitor (SBTI) and human pancreatic secretory trypsin inhibitor (hPSTI) (19) and efficiently degrades and thus inactivates these inhibitors (23). Trypsinogen IV was originally identified in human brain and may be a splice variant of mesotrypsinogen (24). Trypsinogen IV differs from mesotrypsinogen in exon 1 in that it lacks a recognizable signal sequence. However, trypsinogen IV contains three potential furin-processing sites (Arg-X-X-Arg2) and may be a secreted enzyme. Brain trypsinogen IV has been proposed to control expression of glial fibrillary acidic protein and to process the amyloid precursor protein, but its physiological function is unknown (25). Trypsinogens have also been identified in extrapancreatic tissues including endothelial cells, tumor cell lines and tissues, and neurons (26 -30). However, the properties and functions of these extrapancreatic trypsinogens are not fully understood. * This work was supported by National Institutes of Health Grants DK43207, DK57840 (to N. W. B.), and DK52388 (to E. F. G.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We characterized trypsinogens expressed in human colonic mucosa and extrapancreatic epithelial cell lines and investigated whether they are agonists of PARs. Our aims were to: (a) determine if epithelial cell lines and tissues express trypsinogens, enteropeptidase, and PARs; (b) clone, express, and characterize the major trypsinogens in epithelial cells; (c) investigate whether extrapancreatic trypsins activate PARs.

EXPERIMENTAL PROCEDURES
Reagents-Oligonucleotides were from Sigma-Genosys (The Woodlands, TX). Enteropeptidase and Tgo DNA polymerase were from Roche Applied Science. Restriction enzymes and T4 DNA ligase were from New England Biolabs (Beverly, MA). Mouse monoclonal antibody to Myc was from Invitrogen (Carlsbad, CA). A rabbit polyclonal antibody to rat trypsin was a gift from Dr. Charles Craik (UCSF), and a mouse monoclonal antibody to human trypsin was from Chemicon International (Temecula, CA). Goat antimouse or rabbit IgG conjugated to horseradish peroxidase or FITC were from Jackson ImmunoResearch Laboratories (West Grove, PA). Bovine pancreatic trypsin was from Worthington Biochemical Corporation (Lakewood, NJ), human thrombin was from Calbiochem (San Diego, CA). A peptide corresponding to the tethered ligand domain of human PAR 2 (SLIGKV-NH 2 ) and analogues of the tethered ligand of human PAR 1 (TFLLR-NH 2 ) and PAR 4 (AYPGKF-NH 2 ) were from Genemed Synthesis, Inc. (South San Francisco, CA). Bovine pancreatic trypsin inhibitor (BPTI), soybean trypsin inhibitor (SBTI), ␣1-antitrypsin inhibitor, benzamidine, pepstatin A, and tosyl-glycine-proline-arginine-p-nitroanilide were from Sigma Chemical Co. E-64D and PPACK were from Calbiochem (San Diego, CA). Leupeptin was from Roche Applied Science.
Trypsinogen IV with C-terminal Myc or N-terminal Ig was transiently expressed in CHO cells using LipofectAMINE 2000 (Invitrogen) to generate CHO-TIV-myc cells and CHO-Ig-TIV cells, respectively. In control experiments, CHO cells were transfected with vector without the trypsinogen IV insert (CHO-vc cells). Cell lysates were prepared by scraping cells into 50 mM Tris/HCl pH 7.8 followed by sonication and clearing of the lysate by centrifugation (1,000 ϫ g, 10 min). Conditioned medium was obtained by culturing cells in reduced-serum medium (OptiMEM, Invitrogen) for 24 h post-transfection. The conditioned medium was cleared by centrifugation and concentrated using spin concentrators (10 kDa exclusion, Vivascience Inc., Edgewood, NY). Protein concentrations were determined using the bicinchoninic acid assay, with BSA as standard (34).
Measurement of Ca 2ϩ Mobilization-Cells were incubated in HBSS, 0.1% BSA, 20 mM HEPES, pH 7.4, containing 5 M Fura-2/AM (Molecular Probes, Eugene, OR) for 25 min at 37°C. Coverslips were mounted in an open chamber (Warner Instruments Corporation, Hamden, CT, 37°C) on a Zeiss 100 TV inverted microscope, and observed with a Zeiss Fluar 40ϫ objective. Fluorescence was measured at 340 and 380 nm excitation and 510 nm emission in individual cells, using an ICCD video camera (Stanford Photonics, Stanford, CA) and a video microscopy acquisition program (Axon Instruments Inc., Union City, CA). The emission ratio of the fluorescence at the two excitation wavelengths, which is proportional to the [Ca 2ϩ ] i , was calculated. Agonists were directly added to the chamber (50 l each injection).
Enzymatic Assays-Conditioned medium (5 g of total protein) or trypsin (12 ng) was incubated in 200 l of 50 mM Tris/HCl, pH 7.8, 20 mM CaCl 2 containing 0.5 mM tosyl-glycine-proline-arginine-p-nitroanilide for 15 min at 37°C. Activity was monitored by measuring absorbance at 405 nm using a V max microplate reader (Molecular Devices Corporation, Sunnyvale, CA). Inhibitors were preincubated on ice with conditioned medium for 30 min before addition of substrate.

Expression of Trypsinogen, PAR 2 , and Enteropeptidase in
Epithelial Cell Lines-We used RT-PCR to examine whether epithelial cell lines express trypsinogen, the trypsinogen activator enteropeptidase and the trypsin receptor PAR 2 . PC-3, SW480, Caco2, A549, and NCM460 cells all expressed mRNA of the sizes predicted for trypsinogen (400 bp) and PAR 2 (525 bp), and all cell lines except for Caco2 and NCM460 also expressed mRNA for enteropeptidase (568 bp) ( Fig. 1A for PC-3; others not shown). The identity of PAR 2 , trypsinogens, and enteropeptidase was confirmed by sequencing (2-8 sequencing reactions per gene per cell line). The primers for trypsinogen would amplify all known trypsinogen gene products. However, we only identified sequences corresponding to trypsinogen IV or mesotrypsinogen in all cell lines, except in NCM460 cells where we also found trypsinogen I. These primers amplified products with sequences for trypsinogen I and II from the pancreatic cell line HPAC (not shown) and were thus able to amplify the other forms of trypsinogen. We were unable to amplify full-length products using mesotrypsinogen-specific primers in any of the cell lines. However, primers specific to trypsinogen IV amplified a partial clone of trypsinogen IV. Thus, the extrapancreatic epithelial cell lines derived from several organs (prostate, colon, airway) express mRNA encoding trypsinogen IV and PAR 2 , and most also express enteropeptidase.
Expression of Trypsinogens in Normal Human Colonic Mucosa-To confirm that trypsinogens are expressed in normal human colonic mucosa, total RNA was analyzed by RT-PCR. A PCR product of the predicted size (400 bp) was obtained using the primers for trypsinogens (not shown). Products were subcloned and their identity confirmed by sequencing (21 sequencing reactions, 2 patients). All products were found to be identical to either trypsinogen IV or mesotrypsinogen. Using primers to distinguish between trypsinogen IV and mesotrypsinogen, a 474-bp product was amplified and identified as trypsinogen IV by sequencing (9 sequencing reactions, 1 patient). Thus, normal human colonic mucosa, which is known to express PAR 2 (10), also expresses trypsinogen IV.
Localization of Trypsinogen Immunoreactivity in Epithelial Cell Lines-To confirm expression of trypsinogen at the protein level, PC-3, and SW480 cells were stained using a monoclonal antibody to human trypsin. This antibody would be expected to cross-react with all forms of trypsin in view of their high degree of homology. Immunoreactive trypsin was detected in vesicles in the cytoplasm of PC-3 and SW480 cells (Fig. 1, B and C). The vesicles were densely localized in a perinuclear region in both cell types, but some were also detected in close proximity to the plasma membrane. The pattern of staining resembled that for trypsin in secretory granules of HPAC (not shown) and pancreatic acinar cells (Bunnett). 2 Since PC-3 and SW480 cells only expressed trypsinogen IV mRNA, we deduce that this immunoreactivity represents trypsinogen IV protein.
Expression of Functional PAR 2 in Epithelial Cell Lines-To confirm that PC-3 and SW480 cells expressed functional PAR 2 , we measured Ca 2ϩ mobilization to bovine pancreatic trypsin and the PAR 2 -selective agonist SLIGKV-NH 2 , which corresponds to the tethered ligand domain of human PAR 2 . Both trypsin and SLIGKV-NH 2 stimulated a prompt increase in [Ca 2ϩ ] i in both cell lines (Fig. 2, A and B). The responses were concentration-dependent, although trypsin was more potent than SLIGKV-NH 2 . The reverse sequence of the tethered ligand peptide, VKGILS-NH 2 , which does not activate PAR 2 , had no effect on [Ca 2ϩ ] i . Thus, PC-3 and SW480 cells express functional PAR 2 . Trypsin (10 nM) and SLIGKV-NH 2 (100 M) also mobilized Ca 2ϩ in NCM460 cells, confirming expression of PAR 2 (not shown). However, we could not detect responses to the PAR 4 -selective agonist AYPGKF-NH 2 in PC-3, SW480, or NCM460 cells (not shown), suggesting that these cells do not express functional PAR 4 . We have previously shown that PAR 2 is expressed in A549 cells and Caco2 cells (10).
Cloning and Expression of Human Trypsinogen IV-We amplified a partial clone of trypsinogen IV from PC-3 cells using RT-PCR and completed the clone as described in the "Experimental Procedures." Sequence analysis confirmed that the clone was identical to the reported sequence of human trypsinogen IV (a form) (24).
We expressed trypsinogen IV with C-terminal Myc in CHO cells. It was not possible to generate a viable cell line that stably expressed this construct. However, when transiently expressed the cells were viable. We investigated the presence of trypsinogen IV in whole cell lysates and conditioned medium by immunoblotting for Myc. In CHO-TIV-myc cells, we detected trypsinogen IV in the cell lysate, with an apparent molecular mass of 39 kDa, but not in conditioned medium (Fig. 3A, lane  2). Exposure of cells to phorbol-12-myristate-13-acetate or forskolin (both 1 M, 60 min) (not shown) did not stimulate detectable release of trypsinogen IV into the medium as assessed by Western blotting, indicating that activation of protein kinase C or cAMP does not stimulate secretion of trypsinogen IV. Immunoreactive proteins were not detected in lysate or medium from CHO-vc cells (expressing vector without trypsinogen IV insert). Analysis by immunofluorescence using the myc antibody detected trypsinogen IV in granules within the cytoplasm of CHO-TIV-myc cells (Fig. 3B), similar to that observed using the antibody to human trypsin in epithelial cell lines (Fig. 1, B and C). Immunoreactive trypsinogen IV was detected in densely packed vesicles throughout the cytosol and in close proximity to the nucleus and plasma membrane. Staining was not detected in CHO-vc cells (Fig. 3C), confirming the specificity. The trypsin antibody similarly stained CHO-TIV-myc cells (Fig. 3D) but not CHO-vc cells (Fig. 3E). Thus, trypsinogen IV is found in vesicles of transfected cells but we were unable to detect spontaneous secretion.
Enzymatic Analysis of Expressed Trypsinogen IV-To obtain trypsinogen IV for enzymatic analysis, we engineered trypsinogen IV with a murine Ig signal sequence at the N terminus, which leads to spontaneous secretion of the protein into the medium. Expression was assessed by immunofluorescence and Western blotting using antibodies to rat trypsin. Trypsinogen IV immunoreactivity was observed in granules in CHO-Ig-TIV cells (Fig. 4A). Analysis of lysates of CHO-Ig-TIV cells identified low levels of an immunoreactive protein with an apparent mass of 43.5 kDa (Fig. 4B, lane 3).
We concentrated conditioned medium, and incubated the concentrate with enteropeptidase to activate the zymogen. Trypsinogen IV cleavage was assessed by immunoblotting. Trypsinogen IV was detected in the conditioned medium of CHO-Ig-TIV cells as a protein with an apparent molecular mass of 36 kDa (Fig. 4B, lane 4). This reduction in apparent mass is probably due to removal of the signal sequence. Treatment of the concentrate with enteropeptidase to activate the zymogen reduced the size of the secreted protein to 30 kDa, consistent with trypsinogen IV cleavage to trypsin IV (Fig. 4B,  lane 5). There was no detectable trypsinogen IV in the medium or lysate from CHO-vc cells (Fig. 4B, lanes 1 and 2).
We assessed the activation of trypsinogen IV using an enzy-2 N. W. Bunnett, unpublished data. matic assay. There was no detectable trypsin activity in conditioned medium from CHO-Ig-TIV or CHO-vc cells. Treatment of conditioned medium from CHO-Ig-TIV cells with enteropeptidase generated tryptic activity (29 nmol/min/mg total protein). However, there was no detectable activity in enteropeptidase-treated medium collected from CHO-vc cells. We compared the effects of various inhibitors on the activity of bovine pancreatic trypsin and trypsin IV. Pepstatin A (inhibits aspartyl proteases) and E-64D (inhibits cysteinyl proteases) had little or no effect on either forms of trypsin (Fig. 5). Synthetic, low molecular mass inhibitors such as leupeptin, PPACK, and benzamidine abolished or strongly inhibited the activity of both trypsin IV and bovine trypsin. High molecular mass polypeptide inhibitors such as SBTI, BPTI and ␣1-antitrypsin inhibitor abolished the activity of the bovine trypsin, but had no effect on the activity of trypsin IV. Thus, trypsin IV   is a serine protease that is resistant to polypeptide inhibitors when compared with other isoforms of trypsin.

FIG. 2. Effects of graded concentrations of PAR 2 agonists on [Ca 2؉ ] i PC-3 cells (A) and SW480 cells (B). Cells
Trypsin IV Activates PAR 2 and PAR 4 -To examine whether trypsin IV activates PAR 2 , we measured [Ca 2ϩ ] i in KNRK-PAR 2 cells. Enteropeptidase-activated conditioned medium from CHO-Ig-TIV cells stimulated Ca 2ϩ mobilization in KNRK-PAR 2 cells (Fig. 6A) but not in untransfected KNRK cells, which express PAR 2 at only very low levels (Fig. 6C). There was no response in KNRK-PAR 2 cells to enteropeptidase-activated conditioned medium from CHO-vc cells, or to conditioned medium from CHO-Ig-TIV cells that had not been treated with enteropeptidase. Thus, trypsin IV cleaves PAR 2 in KNRK-PAR 2 cells to activate the receptor and mobilize Ca 2ϩ .
To confirm that trypsin IV cleaves PAR 2 , we examined responses in KNRK-PAR 2 cells treated with bovine pancreatic trypsin to desensitize PAR 2 . Challenge of KNRK-PAR 2 cells with 10 nM pancreatic trypsin stimulated a prompt increase in [Ca 2ϩ ] i and caused strong desensitization to a second challenge with the trypsin 2 min later (Fig. 7A). Under these conditions of desensitization of PAR 2 by pancreatic trypsin, responses to trypsin IV were completely abolished. In a similar manner, trypsin IV stimulated an increase in [Ca 2ϩ ] i and caused strong desensitization to a second challenge with the trypsin IV 2 min later. This treatment also desensitized the response to pancreatic trypsin by 56%, compared with cells treated with medium from vector control cells (Fig. 7A). These results confirm that trypsin IV activates the same receptor as pancreatic trypsin in KNRK-PAR 2 cells. PAR 4 is a second receptor for trypsin. Therefore, we determined whether trypsin IV activates PAR 4 . Activated medium from CHO-Ig-TIV cells stimulated Ca 2ϩ mobilization in KNRK-PAR 4 cells (Fig. 6B). There was no response in KNRK-PAR 4 cells to enteropeptidase-treated medium from CHO-vc cells, or to untreated medium from CHO-Ig-TIV cells. Thus, trypsin IV cleaves PAR 4 in KNRK-PAR 4 cells to activate the receptor and mobilize Ca 2ϩ . Challenge of KNRK-PAR 4 cells with 1 units/ml thrombin stimulated a prompt increase in [Ca 2ϩ ] i and caused strong desensitization to a second challenge with the same agonist 2 min later (Fig. 7B). Trypsin IV also increased [Ca 2ϩ ] i and caused strong desensitization to a second challenge with the trypsin IV 2 min later. This treatment also desensitized the response to thrombin by 62%, compared with cells treated with medium from vector control cells (Fig. 7B). These results confirm that trypsin IV activates the same receptor as thrombin in KNRK-PAR 4 cells.
Pancreatic trypsin is able to cleave and activate PAR 1 , albeit at higher concentrations of the enzyme (4). To determine if trypsin IV activates PAR1, we exposed KNRK-PAR 1 cells to trypsin IV and measured [Ca 2ϩ ] i . Trypsin IV had no effect on [Ca 2ϩ ] i in these cells at concentrations that strongly activated PAR 2 and PAR 4 (Fig. 6D). In contrast, both thrombin (Fig. 6D) and an analogue of the PAR 1 activating peptide (TFLLR-NH 2 ) (not shown) increased [Ca 2ϩ ] i in these cells. Thus, trypsin IV is not an agonist of PAR 1 at the concentration used.
To examine whether trypsin IV can activate PARs in epithelial cells naturally expressing these receptors, we measured [Ca 2ϩ ] i in PC-3 and SW480 cells that express PAR 2 but not PAR 4 . Enteropeptidase-activated medium from CHO-Ig-TIV caused a prompt increase in [Ca 2ϩ ] i in PC-3 and SW480 cells (Fig. 8). In contrast, activated medium from CHO-vc did not affect [Ca 2ϩ ] i in either cell line (data not shown). These results confirm that trypsin IV is able to cleave and activate PAR 2 in cells that naturally express this receptor. DISCUSSION Our results show that trypsinogen IV is co-expressed with PAR 2 in extrapancreatic epithelial cell lines and in normal colonic epithelium. Enteropeptidase, which is co-expressed with trypsinogen IV, cleaves and activates the zymogen, and trypsin IV is resistant to polypeptide inhibitors of pancreatic trypsins. Trypsin IV cleaves and activates both PAR 2 and PAR 4 . Thus, trypsin IV may be a physiological agonist of PAR 2 and PAR 4 in epithelial tissues where it could activate these receptors in a paracrine or autocrine manner.
Characterization of Trypsinogen IV-We found that trypsinogen IV (a-form) is expressed in epithelial cells of the colon, prostate and lung. In support of these results, trypsinogen IV mRNA has been amplified from PC-3 cells (35). Although trypsinogen IV cDNA encodes a protein that lacks a recognizable signal peptide, there are potential cleavage sites for furin, which may process the full-length protein into secreted forms. Indeed, trypsinogen IV was detected in vesicles in transfected cell lines and in epithelial cells, and may be secreted from these vesicles. The regulation of secretion remains to be investigated.
Expression of trypsinogen IV with an Ig signal sequence resulted in the constitutive secretion and allowed collection of sufficient quantities of the zymogen for characterization. Enteropeptidase treatment of trypsinogen IV generated active trypsin IV with an apparent molecular mass of 30 kDa, higher than the predicted size of 24 kDa. The additional mass of the expressed protein does not represent glycosylation (data not shown), and at present we have no explanation for the decreased electrophoretic mobility. We found that trypsin IV is resistant to polypeptide inhibitors of pancreatic trypsins I and II. In support of these results, mesotrypsin, a variant of trypsin IV, is also resistant to polypeptide inhibitors (19, 23). Sequence alignments and crystallographic studies have revealed the basis for this resistance (36). A single base change from glycine to arginine in trypsin IV (Gly 198 in trypsinogens I and II, Arg 255 in trypsinogen IV) results in a steric interference between the arginine of the enzyme and the P2Ј side chain of the polypeptide inhibitor. Mesotrypsin is similarly resistant to SBTI and hPSTI by virtue of this Gly 3 Arg substitution since an R198G mutant is sensitive to polypeptide inhibitors (23). In addition, this substitution renders mesotrypsin more resistant to autocatalytic destruction (23). Moreover, mesotrypsin also degrades and inactivates polypeptide trypsin inhibitors such as SBTI and hPSTI (23). In view of the similarity between mesotrypsin and trypsin IV, it is probable that trypsin IV also degrades these inhibitors although we have not tested this possibility.
Trypsin IV as a Potential Agonist of PAR 2 and PAR 4 -Several observations suggest that trypsin IV is a potential agonist of PAR 2 and PAR 4 . First, enteropeptidase-treated conditioned medium from CHO-Ig-TIV cells, which contained trypsin IV, increased [Ca 2ϩ ] i in transfected cells expressing PAR 2 or PAR 4 . Conditioned medium from CHO-Ig-TIV cells that was not treated with enteropeptidase, and enteropeptidase-treated conditioned medium from CHO-vc cells were inactive. Thus, the stimulatory effect of conditioned medium is due to the presence of trypsin IV. Second, trypsin IV did not signal to cells lacking PAR 2 and PAR 4 , and responses to trypsin IV were abolished by desensitization of PAR 2 or PAR 4 , which confirms that re-sponses to trypsin IV are mediated by these receptors. Finally, trypsin IV also increased [Ca 2ϩ ] i in epithelial cells lines that naturally express PAR 2 , confirming that trypsin IV can signal to cells expressing this receptor at physiological levels. To our knowledge, this is the first report that trypsin IV is an agonist of PAR 2 and PAR 4 . We did not formally compare the potencies with which trypsin IV and pancreatic trypsins activate PARs due to the lack of pure enzyme. However, comparable activities of trypsin IV and pancreatic trypsin toward a synthetic substrate (Gly-Pro-Arg-pNA) generated similar Ca 2ϩ signals in cell lines. 2 In support of our results, other extrapancreatic trypsins have been reported to activate PAR 2 . Trypsin II from COLO-205 cells activates PAR 2 in transfected cells (37), and trypsin I from colon cancer cells lines also activates PAR 2 (38). However, in these studies the active enzyme was not expressed nor characterized, and the ability of the trypsins to activate other PARs was not examined.
Physiological Roles of Trypsins as PAR Agonists-Under physiological circumstances, pancreatic trypsins in the lumen of the small intestine can signal to enterocytes by cleaving PAR 2 (16). During intestinal inflammation, when there is a loss of tight junctions in the intestinal epithelium, trypsin from the lumen could penetrate tissues and activate PAR 2 and PAR 4 on other cell types (39). In pancreatitis, trypsinogens are prematurely activated in the pancreas and may signal to acinar and ductal epithelial cells by cleaving PAR 2 (10, 17). However, PAR 2 and PAR 4 are widely expressed in epithelial tissues not normally exposed to pancreatic trypsins (2,6,7,9,10,14,15). We hypothesize that trypsin IV released from epithelial cells could signal in a paracrine or autocrine manner by cleaving and activating PAR 2 and PAR 4 . Although most of the epithelial cell lines expressing trypsinogen IV also expressed enteropeptidase mRNA, and enteropeptidase can activate trypsinogen IV, the factors that regulate the release and activation of trypsinogen IV are unknown.
The observations that mesotrypsin is resistant to polypeptide inhibitors, can degrade these inhibitors and is also resistant to autocatalytic destruction by virtue of the Arg 3 Gly substitution has lead to new theories about the function of this enzyme (23). Under physiological conditions, mesotrypsin may degrade foods that are naturally rich in trypsin inhibitors. In the inflamed pancreas, mesotrypsin may degrade the naturally protective hPSTI and thus exacerbate pancreatitis. In view of its homology to trypsin IV, it is likely that mesotrypsin activates PAR 2 and PAR 4 in the pancreas, which may contribute to the genesis of inflammation. The possibility that trypsin IV also degrades endogenous trypsin inhibitors coupled with our observation that trypsin IV is resistant to these inhibitors may promote prolonged signaling of this enzyme in extrapancreatic tissues by cleavage and activation of PAR 2 and PAR 4 .
Trypsinogen IV was originally identified in human brain (24) and, if released and activated, trypsin IV could be an agonist of PARs in neurons and glial cells. PAR 2 and PAR 4 are expressed throughout the nervous system (40 -45), but the agonists of PARs in neural tissues are not characterized. Overexpression of trypsinogen IV in neurons in the mouse brain results in a marked up-regulation of expression of glial fibrillary acidic protein in astrocytes (25), which may be mediated by release of trypsinogen IV from neurons and subsequent signaling to astrocytes through PAR 2 or PAR 4 . FIG. 7. Desensitization of PAR 2 and PAR 4 . Conditioned medium from CHO-Ig-TIV cells was incubated with enteropeptidase to generate trypsin IV (TIV). Enteropeptidase-treated conditioned medium from CHO-vc cells was used as the vector control (vc). Effects of pancreatic trypsin (PT, 10 nM), thrombin (Th, 1 unit/ml), trypsin IV, and vector control on [Ca 2ϩ ] i were assessed in KNRK-PAR 2 cells (A) and KNRK-PAR 4 cells (B). Note that pancreatic trypsin desensitized responses to trypsin IV in KNRK-PAR 2 cells and that thrombin desensitized responses to trypsin IV in KNRK-PAR 4 cells. Trypsin IV also desensitized responses to pancreatic trypsin and thrombin when compared with cells treated with the vector control. Each trace is an average response from n Ͼ 4 cells and is representative of 3-4 different experiments. The decline in the 340/380 ratio at the point of injection is an artifact of the concentrated medium.
FIG. 8. Trypsin IV signaling to PC-3 and SW480 cells that naturally express PAR 2 . Note that trypsin IV (TIV) caused a prompt increase in [Ca 2ϩ ] i in both cell lines. Each trace is an average response from n Ͼ 4 cells and is representative of 3-4 different experiments. The decline in the 340/380 ratio at the point of injection is an artifact of the concentrated medium.
Trypsin IV is the latest member of a growing number of proteases that can activate PAR 2 and PAR 4 . Tryptase from mast cells activates PAR 2 (11,12) and tryptase may be a physiological agonist in cells that are in close proximity to mast cells and which could be exposed to high concentrations of this enzyme, such as sensory and enteric nerves (42,45,46). However, tryptase is several orders of magnitude less potent than trypsin for activation of PAR 2 , and its capacity to activate PAR 2 is diminished by glycosylation of the receptor (47,48). Coagulation factors VIIa and Xa may activate PAR 2 in endothelial cells, but responses require that VIIa is anchored to the cellsurface in the vicinity of the receptor by tissue factor (8). A solubilized form of membrane-type serine protease 1, a cellsurface protease that is co-expressed with PAR 2 in PC-3 cells and certain tissues, can also activate PAR 2 expressed in Xenopus oocytes (49), although it is unknown whether the membrane-anchored protease is a PAR 2 agonist under physiological circumstances. Thrombin is likely to be a physiological agonist of PAR 4 in platelets and endothelial cells, and cathepsin G could mediate neutrophil-induced platelet aggregation (6,7,13).
In summary, trypsinogen IV is expressed in epithelial cell lines and tissues that also express enteropeptidase and PAR 2 . Enteropeptidase activates trypsinogen IV, and trypsin IV is an agonist of PAR 2 and PAR 4 . Given its resistance to polypeptide inhibitors, trypsin IV could signal for prolonged periods by activating these receptors.