Jab1, a Novel Protease-activated Receptor-2 (PAR-2)-interacting Protein, Is Involved in PAR-2-induced Activation of Activator Protein-1*

Protease-activated receptor-2 (PAR-2), a G protein-coupled receptor for trypsin and tryptase, exerts important physiological and pathological functions in multiple systems. However, unlike PAR-1, the PAR-2-mediated intracellular signal transductions are hardly known. Here, using yeast two-hybrid screening with a human brain cDNA library, we identified an interacting partner of human PAR-2, the Jun activation domain-binding protein 1 (Jab1). The interaction was confirmed by glutathione S-transferase pull-down assays in vitro, and by co-immunoprecipitation assays in vivo. Jab1 was also shown to be colocalized with PAR-2 in both transfected HEK293 cells and in normal primary human astrocytes by double immunofluorescence staining. Further experiments demonstrated that multiple intracellular domains of PAR-2 are required for the interaction with Jab1. We then showed that agonist stimulation of PAR-2 disrupted the interaction, which could be prevented by the inhibitor of receptor endocytosis phenylarsine oxide, but not by the lysosomal protease inhibitor ZPAD. Importantly, we found that activation of PAR-2 induced the redistribution of Jab1 from the plasma membrane to the cytosol, but did not influence expression of Jab1. Furthermore, Jab1 mediated PAR-2-induced c-Jun activation, which was followed by increased activation of activator protein-1. Loss-of-function studies, using Jab1 small interfering RNA, demonstrated that Jab1 knockdown blocked PAR-2-induced activator protein-1 activation. Taken together, our data demonstrate that Jab1 is an important effector that mediates a novel signal transduction pathway for PAR-2-dependent gene expression.

signal transductions to the extracellular serine proteases trypsin, tryptase, and others during cell proliferation, pain, and inflammatory responses (1,2). Rather than mere ligand binding, which is responsible for receptor activation in most GPCRs, PAR-2 is activated by the newly exposed NH 2 terminus, which serves as a tethered ligand domain. This tethered ligand domain is generated by proteolytic cleavage of the extracellular NH 2 terminus of PAR-2 (3). Synthetic receptor-activating peptide, whose sequence is identical to that of the tethered ligand domain, could mimic the actions of trypsin and tryptase (2,3). Unlike PAR-1 whose signal transductions have been clarified in detail (4), only relatively few studies about signaling pathways mediated by PAR-2 were done. Nevertheless, PAR-2 activation has been found to induce the activation of the mitogen-activated protein kinase family, the extracellular signal regulated kinase, c-Jun NH 2 -terminal kinase, and p38 mitogenactivated protein kinase in multiple cell types (5,6). It has been reported that PAR-2 could further activate the downstream factors c-Fos, c-Jun, as well as the transcription factor nuclear factor B (5, 7), which may mediate PAR-2-induced expression and release of inflammatory factors.
PAR-2 was found to be abundantly expressed in the heart, lung, trachea, liver, pancreas, gut, kidney, prostate, skin, and brain (2,8,9). To date, most findings suggest that PAR-2 contributes to inflammation in the cardiovascular system, airway system, gastrointestinal system, skin, and peripheral nervous system (1,2,4,9). In the central nervous system, it was also reported that high concentrations of the PAR-2-activating peptide (AP) induced neuronal death in the rat hippocampus, suggesting that PAR-2 may be linked to neurodegeneration (10). Recent data obtained from PAR-2 knock-out mice indicated that the deficiency of the PAR-2 gene promoted acute focal ischemic brain injury (11). Although accumulating evidence indicates that PAR-2 is important in neuroinflammatory and neurodegenerative processes, the potential significance of PAR-2 in the central nervous system still remains elusive.
Jun activation domain-binding protein 1 (Jab1) was initially identified as a coactivator of c-Jun (12), and was later shown to be a component of the COP9 signalosome complex (13). The COP9 signalosome is a conserved multiprotein complex that has been found in plants, mammals, Drosophila, and the fission yeast (14). It consists of eight subunits, which exhibit significant similarity to the eight subunits of the lid of the 26 S proteasome (14,15). Previous work has already demonstrated that a number of diverse proteins interact with the subunits of the COP9 signalosome, especially with Jab1. Using the yeast two-hybrid system, it was found that Jab1 bound to the NH 2 -terminal activation domain of c-Jun and thereby activated c-Jun (12). Interestingly, Jab1 was shown to regulate the cell cycle by degrading the cyclin-dependent kinase inhibitor p27 Kip1 (16). Recently, Jab1 was found to be connected to the transmembrane protein integrin adhesion receptor LFA-1, and to mediate the activation of LFA-1-induced activator protein-1 (AP-1) (17). It was also reported that Jab1 interacts with the transcription factor hypoxiainducible factor-1␣, as well as the nuclear receptor progesterone receptor and the steroid receptor coactivator-1, which control the transcription and expression of a number of genes (18,19).
In the present study, we screened a human brain cDNA library using the yeast two-hybrid system, and identified Jab1 as an interacting partner protein of PAR-2. We found that Jab1 constitutively interacts with human PAR-2 in vitro and in vivo. Importantly, Jab1 potentiates PAR-2-induced c-Jun/AP-1 activation. Our data delineate a novel alternative signal transduction pathway for PAR-2-dependent gene expression.
For glutathione S-transferase (GST) pull-down assays, the GST cDNA containing the linker sequence at the NH 2 terminus was cloned into pVL1392 vector (Invitrogen), generating the C-terminal GST expression vector, pVL1392-GST. The cDNA fragments corresponding to the different regions of human PAR-2 ( Fig. 2B) were amplified by PCR and subcloned into pVL1392-GST. The full-length human Jab1 cDNA containing a consensus Kozak sequence upstream of the initiator ATG was amplified by RT-PCR and cloned into pcDNA3.1mycHis vector (Invitrogen).
For co-immunoprecipitation, PCR products of the full-length human PAR-2 with a consensus Kozak sequence upstream of the initiator ATG and the hemagglutinin epitope (YPYDVPDYA, HA) at the C terminus were cloned into pEAK10 vector (a gift from Dr. T. Koch, Institut für Pharmakologie und Toxikologie, Otto-von-Guericke-Universität Magdeburg).
Primer sequences are available on request. All of the DNA sequences of plasmid constructs were confirmed to be in-frame by ABI 310 sequencer.
Yeast Two-hybrid Screening Assays-Yeast two-hybrid screening was performed using the MATCHMAKER GAL4 Two-Hybrid System 3, according to the manufacturer's protocol and the Yeast Protocol Handbook (Clontech). Yeast strain AH109 cells were transformed with pGBKT7-hsPAR-2 plasmids by the lithium acetate method. The transformants were selected by growth on SD/-Trp media. Commercial pretransformed Y187 cells with the human brain MATCHMAKER cDNA library were incubated for 24 h with transformed AH109 for yeast mating. The entire mating culture was plated onto SD/-Trp-Leu-His media, and colonies were further spread onto SD/-Trp-Leu-His-Ade media. His ϩ Ade ϩ colonies were assayed for ␤-galactosidase activity by the LacZ colony lift filter assay. Plasmids were isolated from His ϩ Ade ϩ LacZ ϩ colonies and transformed into Escherichia coli XL 1-Blue cells for plasmid amplification. The rescued pACT2 plasmids containing the cDNA insert were sequenced, and analyzed with the program BLAST in the GenBank TM data base.
Normal primary human astrocytes were obtained from Cambrex Bio-Science Verviers SPRL (Verviers, Belgium). These astrocytes were established from normal human brain tissue. Normal primary human astrocytes were grown in the AGM Astrocyte Medium (Cambrex) at 37°C and 5% CO 2 . Normal primary human astrocyte cultures were used within 10 passages in the present study, because their characteristic properties are impaired with subsequent passages.
Cells (80% confluent) were transfected using DOTAP liposomal transfection reagent, according to the manufacturer's protocol (Roche Diagnostics). Briefly, 1 day before transfection the cells were plated on a 6-well plate. The following day, the subconfluent cells were transfected or co-transfected with the indicated plasmids using DOTAP. To generate the stable clone, the cells were selected with 500 g/ml of G418 (for HEK293-Jab1myc), 1 g/ml of puromycin (for HEK293-PAR-2-HA), or both (for HEK293-PAR-2-HAϩJab1myc).
PAR-2 Activation and Inhibitor Treatment-To activate PAR-2 receptor, cells were rinsed with Hank's solution, and stimulated in serum-free medium with bovine pancreatic trypsin (100 nM, Roche Diagnostics) as a physiological agonist or with PAR-2 AP SLIGKV-NH 2 (100 M, Bachem) as a specific agonist. To prevent receptor endocytosis or degradation, cells were pretreated with 80 M phenylarsine oxide (PAO; Calbiochem) for 15 min or with 100 M Z-Phe-Ala-diazomethylketone (ZPAD, Bachem) for 30 min prior to agonist stimulation. To inhibit protein synthesis, cells were pretreated with 70 M cycloheximide (Calbiochem) for 30 min. The inhibitors were included in medium throughout the experiments.
GST Pull-down Assays-The different GST fusion protein constructs were transfected into Spodoptera frugiperda (Sf9) cells, as described previously (20). The recombinant baculovirus was amplified and tested for the production of the fusion protein. Cells infected with recombinant baculovirus containing PAR-2-GST, PAR-2⌬IL12-GST, or PAR-2⌬IL3C-GST were lysed on ice in membrane fraction buffer 1 containing 50 mM Hepes, pH 8, 300 mM NaCl, 0.1 mM EDTA, 10 mM ␤-mercaptoethanol, and Protease Inhibitor Mixture (Roche Diagnostics, Germany, one tablet per 50 ml). After centrifugation for 10 min at 1,000 ϫ g and 4°C, the supernatant in membrane fraction buffer 1 was further centrifuged for 1 h at 100,000 ϫ g and 4°C. Afterward, pellets were resuspended in membrane fraction buffer 2 (50 mM Hepes, pH 8, 300 mM NaCl, 10 mM ␤-mercaptoethanol, 1% Brij 58, and Protease Inhibitor Mixture), stirred, and centrifuged again for 1 h at 100,000 ϫ g and 4°C. The resulting supernatant contained the membrane fraction of the recombinant fusion proteins. On the other hand, cells infected with recombinant baculovirus containing PAR-2IL3-GST or PAR-2C-GST were lysed in Sf9 lysis buffer containing 50 mM Tris/HCl, pH 7.5, 150 mM NaCl, 10 mM NaF, 1% Triton X-100, and Protease Inhibitor Mixture. After centrifugation for 5 min at 35,000 ϫ g and 4°C, the resulting supernatant contained the cytosolic fraction of the recombinant fusion proteins. The recombinant fusion proteins were purified by using glutathione-Sepharose beads (Amersham Biosciences). Equal amounts of GST and truncated PAR-2-GST fusion proteins immobilized on glutathione-Sepharose beads were incubated overnight at 4°C with the crude HEK293-Jab1myc cell extracts in HEK293 lysis buffer, containing 50 mM Tris/HCl, pH 7.5, 1 mM ␤-mercaptoethanol, 150 mM NaCl, 1% Igepal, and Protease Inhibitor Mixture. After washing three times with the HEK293 lysis buffer without protease inhibitor, the bound proteins were separated by SDS-PAGE, and immunoblotted with the anti-myc antibody (1:5000, Invitrogen).
Immunoprecipitation-Cells were rinsed with cold phosphate-buffered saline, and lysed in the HEK293 lysis buffer. Equal amounts of lysates were rotated with the anti-HA antibody (1:200, Sigma) or with the anti-PAR-2 (C-17) antibody (1:25, Santa Cruz) for 6 h, followed by incubating with protein A-Sepharose beads (Amersham Biosciences) or protein G-agarose beads (Sigma) overnight at 4°C. After washing three times with HEK293 lysis buffer without protease inhibitor, the bound beads were incubated in Laemmli buffer for 60 min on ice (for immunoprecipitating PAR-2), or boiled in Laemmli buffer for 5 min at 100°C (for immunoprecipitating PAR-2⌬IL12-HA), separated by SDS-PAGE, and immunoblotted with the anti-myc antibody (1:5000, Invitrogen) or with the anti-Jab1 antibody (1:1000, Santa Cruz). To confirm the specific immunoprecipitation by PAR-2, the membrane was stripped, blocked, and reprobed overnight with the anti-HA antibody (1:2000, Cell signaling technology) or with anti-PAR-2 antibody (C-17, 1:100).
Immunofluorescence Analysis-As described previously (21), cells were fixed with 4% paraformaldehyde for 20 min at room temperature, and permeabilized by incubation in phosphate-buffered saline with 0.2% Triton X-100 and 3% bovine serum albumin for 60 min. Then they were incubated overnight with mouse anti-myc antibody (2 g/ml, Invitrogen) and rabbit anti-HA antibody (4 g/ml, Sigma) or with mouse anti-Jab1 antibody (8 g/ml, Santa Cruz) and rabbit anti-HA antibody (4 g/ml, Sigma) at 4°C, washed, and incubated with Alexa Fluor 488 goat anti-mouse IgG antibody (20 g/ml, Molecular Probes) and Alexa Fluor 568 goat anti-rabbit IgG antibody (20 g/ml, Molecular Probes) for 120 min at room temperature in the dark. Mounted slides were observed with a LSM510 confocal laser scanning microscope (Carl Zeiss, Germany).
Reporter Gene Assays-Cells were transfected with an AP-1-driven chloramphenicol acetyltransferase (CAT) reporter plasmid (a gift from Dr. J. Kraus, Institut für Pharmakologie und Toxikologie, Otto-von-Guericke-Universität Magdeburg) using DOTAP, as described above. Six hours after transfection, the cells were treated with 100 M PAR-2 AP for 48 h in Dulbecco's modified Eagle's/Ham's F-12 medium supplemented with 1% fetal calf serum. The CAT activity in cell lysates was analyzed by enzyme-linked immunosorbent assay (Roche Diagnostics), and normalized to protein concentration. The empty CAT reporter vector was also transfected in parallel, as a negative control.
Small Interfering RNA (siRNA)-Human Jab1 siRNA was obtained from Santa Cruz Biotechnology, Inc. HEK293-PAR-2-HA cells were transfected with human Jab1 siRNA by using magnet-assisted transfection, according to the manufacturer's protocol (IBA GmbH, Germany). Non-silencing siRNA labeled with Alexa Fluor 488 served as a scrambled siRNA control (Qiagen). Jab1 knockdown was assessed by Western blot and RT-PCR at 48 h after transfection. Total RNA and subsequently cDNA were made, as described previously (21). Specific primers used were as follows: Jab1, forward 5Ј-CATATGAATACATGGCTGCA-3Ј and reverse 5Ј-GGCTTCTGACTGCTCTAAC-3Ј; glyceraldehyde-3phosphate dehydrogenase, as an internal control, forward 5Ј-TCC-AAAATCAAGTGGGGCGATGCT-3Ј and reverse 5Ј-ACCACCTG-GTGCTCAGTGTAGCCC-3Ј. Fragments were amplified by PCR for 15 min at 95°C, followed by repeated cycles of 30 s at 94°C, 90 s at 53°C, 60 s at 72°C, then a final 10-min extension at 72°C. The reaction products were analyzed by electrophoresis with 1% agarose gel containing ethidium bromide, and visualized by the Bio-Rad gel document system.
Statistical Analysis-Data were expressed as mean Ϯ S.E. Differences were examined by Student's t test between two groups or one-way analysis of variance within multiple groups. p Ͻ 0.05 was considered significant.

Jab1
Interacts with Human PAR-2 in Yeast-PAR-2, a GPCR, mediates intracellular signal transduction in response to stimulation with the extracellular proteases trypsin and tryptase. To identify intracellular proteins that are involved in PAR-2 signaling events, a yeast two-hybrid screening was performed. The full-length human PAR-2 cDNA was fused in-frame to the GAL4 DNA binding domain as bait. Using this bait, a pretransformed human brain cDNA library was screened as described under "Experimental Procedures." From 1.6 ϫ 10 6 clones screened that were grown on nutritionally deficient plates and that activated the ␤-galactosidase, 308 colonies were found positive for the selection markers histidine, adenine, and LacZ. Subsequent sequencing and BLAST analysis revealed 24 colonies to be truly positive. From two of these single colonies the cDNAs encoded amino acids (aa) 31-334 and 49 -334 of Jab1, respectively (Gen-Bank accession number BC001859) (22). Other candidate partners will be described elsewhere.
To verify the protein interaction in yeast, we transformed pGBKT7-hsPAR-2 into yeast Y187 cells and pACT2-Jab1 isolated from the library into yeast AH109 cells, and performed yeast two-hybrid tests. As expected, the transformants grew on SD/-Trp-Leu-His-Ade media, and were shown to be positive by the LacZ colony-lift filter assay. Therefore, these results suggest that Jab1 could interact with human PAR-2 in yeast.
PAR-2 Expression in Sf9 and HEK293 Cells-To identify proteinprotein interactions of PAR-2 with Jab1 in other systems, we engineered exogenous PAR-2 expression. We generated a novel baculovirus GST expression vector to facilitate the expression of the transmembrane protein tagged with GST in Sf9 cells. The full-length human PAR-2 cDNA was inserted at the NH 2 terminus of GST. After transfection in Sf9 cells using Lipofectin (Invitrogen), the recombinant baculovirus was amplified and tested for the production of the fusion protein. As shown in Fig. 1A, PAR-2-GST was strongly expressed with two major bands (ϳ60 -70 and ϳ140 -160 kDa) in Sf9 cells. The high molecular mass band might represent the PAR-2 receptor homodimer.
We also constructed a mammalian expression vector, pEAK10, inserted with the full-length PAR-2 cDNA fused with HA at the C terminus. This recombinant vector was used for stably transfecting HEK293 cells. As shown in Fig. 1B, PAR-2-HA expression was detected in the whole cell lysates from HEK293-PAR-2-HA cells, by using an antibody against HA. PAR-2-HA protein appeared as a smear band (mainly at ϳ37-70 kDa), which is consistent with previous reports by others (23,24). It is known that human PAR-2 possesses two potential N-linked glycosylation sites (23). Receptor glycosylation is important for PAR-2 expression and activation (23,25). To study whether the smear band of PAR-2 resulted from receptor N-glycosylation, whole cell lysates from HEK293-PAR-2-HA were incubated with peptide N-glycosidase F. As shown in Fig. 1C, peptide N-glycosidase F treatment significantly reduced the molecular mass of PAR-2-HA to ϳ32-44 kDa, suggesting that PAR-2-HA is a N-glycosylated glycoprotein. PAR-2 might also be O-glycosylated, because N-deglycosylated PAR-2 in Fig. 1C still appears as a smear band.
Multiple Intracellular Domains of PAR-2 Responsible for Interaction with Jab1-Next, we performed GST pull-down assays in vitro, to confirm interaction between PAR-2 and Jab1 by using biochemical methods. The PAR-2-GST fusion protein expressed in Sf9 cells was purified by glutathione-Sepharose beads. These beads with immobilized PAR-2-GST fusion protein were incubated overnight with whole cell lysates from HEK293-Jab1myc cells to examine the interaction of PAR-2-GST with Jab1myc. The interaction was detected by Western blot analysis using the anti-myc antibody. As shown in Fig. 2A, Jab1 specifically interacted with the full-length PAR-2-GST fusion protein (lane 3). To exclude the possible interaction of Jab1 with the GST tag protein, we, in parallel, incubated GST protein with the crude HEK293-Jab1myc cell lysates. Western blot analysis showed that Jab1 did not interact with GST protein alone ( Fig. 2A, lane 2). Aliquots from the HEK293-Jab1myc cell lysates, which were used for the pull-down assays also served as control for the Western blot ( Fig. 2A, lane 1).
To map the domains of PAR-2 responsible for interaction with Jab1, we generated a series of truncated PAR-2-GST fusion proteins, which are given schematically in Fig. 2B, below the full-length PAR-2-GST. These truncated fusion proteins were tested similarly for their capacities to bind to Jab1. As shown in Fig. 2A, PAR-2⌬IL3C-GST fusion protein (aa 1-245 of PAR-2) weakly interacted with Jab1 (lane 4), but a stronger interaction occurred with the PAR-2⌬IL12-GST fusion protein (aa 214 -397 of PAR-2) and Jab1 (lane 5), which was comparable with that between the full-length PAR-2-GST fusion protein and Jab1 (lane 3). These results suggest that the intracellular loop 3 and the carboxyl tail (C-tail) of PAR-2 were mainly responsible for the interaction with Jab1. Several previous studies have revealed that the C-tail of PAR-2 contributes to receptor trafficking and receptor-induced signal transduction (6). Interestingly, our studies here demonstrated that the PAR-2 C-tail alone (aa 346 -397 of PAR-2) weakly interacted with Jab1 ( Fig. 2A, lane 7). The interaction of PAR-2IL3-GST fusion protein (aa 243-311 of PAR-2) with Jab1 ( Fig. 2A, lane 6) was notably weaker than that between the PAR-2⌬IL12-GST fusion protein and Jab1 (Fig. 2A, lane 5), although it seemed to be stronger than that between the PAR-2C-GST fusion protein and Jab1 ( Fig. 2A, lane 7). Taken together, our GST pull-down findings imply that PAR-2 interacts with Jab1 through multiple intracellular domains.
Jab1 Interacts with PAR-2 in Vivo-To further determine whether Jab1 could interact with PAR-2 in vivo, we performed co-immunoprecipitation experiments. We generated a HEK293 cell line stably co-expressing PAR-2-HA and Jab1myc. Thus we had two different tags, the HA and the myc on the two proteins of interest. The whole cell lysates were immunoprecipitated by anti-HA antibody, and the interaction of PAR-2-HA with Jab1myc was examined by Western blot analysis using the anti-myc antibody. As shown in Fig. 3A, Jab1 was specifically co-immunoprecipitated by PAR-2-HA in the HEK293-PAR-2-HAϩJab1myc cells (lane 6), but not in the negative control wild-type HEK293 (lane 4) and HEK293-Jab1myc (lane 5) cells. The same lysates from the respective cells that were used for the immunoprecipitation (IP) experiments, served as control for the Western blot (lanes 1-3). These results suggest that Jab1 interacts with PAR-2. The nitrocellulose membrane was reprobed with the anti-HA antibody to confirm the specificity of the co-immunoprecipitation of PAR-2-HA.
We further asked whether interaction between endogenous Jab1 and native PAR-2 could occur. The Western blot analysis using the anti-Jab1 antibody showed that endogenous Jab1 was specifically co-immunoprecipitated from wild-type HEK293 cells by anti-PAR-2 antibody (Fig. 3B, lane 3), but not by the unrelated antibody against GST (Fig. 3B, lane 2). The cell lysate for the IP experiments is shown in lane 1. The membrane reprobed with the anti-PAR-2 antibody confirmed the specificity of the co-immunoprecipitation of PAR-2 (data not shown). These data demonstrated that Jab1 interacts with PAR-2 in vivo.
We next tested whether Jab1 could be immunoprecipitated by PAR-2⌬IL12-HA in vivo. The whole cell lysates were immunoprecipitated by anti-HA antibody, and the interaction of PAR-2⌬IL12-HA with Jab1myc was examined by Western blot analysis using the anti-myc antibody. As shown in Fig. 3C, Jab1 specifically associated with PAR-2⌬IL12-HA in the HEK293-PAR-2⌬IL12-HAϩJab1myc cells (lane 3), but not in the negative control wild-type HEK293 (lane 1) and HEK293-Jab1myc (lane 2) cells. The same lysates from the respective cells that were taken for the IP experiments served as control (lanes 4 -6). These results suggest that Jab1 interacts with the intracellular loop 3 and the C-tail of PAR-2, which is consistent with the GST pull-down data above in Fig. 2A. The membrane was reprobed with the anti-HA antibody to confirm the specificity of the co-immunoprecipitation of PAR-2⌬IL12-HA. A, the membrane fraction of PAR-2-GST was extracted from infected Sf9 cells, analyzed in 10% SDS-PAGE, and immunoblotted using a polyclonal anti-GST antibody. A representative blot from three independent experiments is given. The molecular mass marker is indicated on the right. Note: the band with the higher molecular mass might represent the PAR-2 receptor homodimer. B, whole cell lysates from HEK293-PAR-2-HA cells were analyzed in 10% SDS-PAGE, and immunoblotted using a monoclonal anti-HA (6E2) antibody. Experiments were repeated three times with comparable results. The molecular mass marker is indicated on the right. C, whole cell lysates (20 g) from HEK293-PAR-2-HA cells were incubated overnight with peptide N-glycosidase F (PNGase F, 500 units/reaction) at 4°C. Afterward, the reaction was stopped by dissolving in Laemmli buffer, then electrophoresed and immunoblotted by using a monoclonal anti-HA (6E2) antibody. Experiments were repeated three times with identical results. The molecular mass marker is indicated on the right. WB, Western blot. Colocalization of Jab1 with PAR-2 in Vivo-We next examined whether the interaction of Jab1 with PAR-2 is reflected by colocalization of the two proteins in vivo using double immunofluorescence staining. HEK293 cells were transiently co-transfected with pEAK10-PAR-2-HA and pcDNA-Jab1myc plasmids. 24 h after transfection, Jab1myc was shown to be diffusely distributed in the cytosol and nucleus, and was also detected at the plasma membrane (Fig. 4A, left top panel). PAR-2-HA, on the other side, was predominantly localized at the plasma membrane and in intracellular stores (Fig. 4A, middle top panel), where it colocalized with Jab1myc, given by the merge in the right top panel of Fig. 4A. In parallel, we co-stained wild-type HEK293 cells as negative control (Fig. 4A, lower panels).
We further asked whether the endogenous Jab1 protein could colocalize with PAR-2-HA. As shown in Fig. 4B, the endogenous Jab1 (left panel) was shown to have the same distribution pattern as the transfected protein (Fig. 4A, left top panel). Endogenous Jab1 was clearly colocalized with PAR-2-HA at the plasma membrane and in intracellular stores in the HEK293-PAR-2-HA cells (Fig. 4B, right panel). Jab1 Interacts with PAR-2 in Normal Primary Human Astrocytes-Next, we wanted to test the interaction of PAR-2 with Jab1 in brain cells. Human astrocytes were chosen as a cell model here, because astrocytes represent a critical cell type in brain and it has been shown that PARs mediate important functions in astrocytes (26 -28). The normal primary human astrocytes were shown by RT-PCR and calcium measurements to functionally express PAR-2 (data not shown). Thus, we performed the immunoprecipitation experiments to detect the interaction of Jab1 with PAR-2 in normal primary human astrocytes. The astrocyte lysates were immunoprecipitated by anti-PAR-2 antibody in the presence of protein G-agarose beads. The subsequent Western blot analysis (Fig. 5A) using the anti-Jab1 antibody demonstrated that Jab1 was specifically immunoprecipitated by anti-PAR-2 antibody (lane 2), but not by the antibody against GST (lane 3), indicating that Jab1 associates with PAR-2 in normal primary human astrocytes. The membrane reprobed with the anti-PAR-2 antibody confirmed the specificity of the co-immunoprecipitation of PAR-2 (data not shown).
We further studied the colocalization of Jab1 with PAR-2 in normal primary human astrocytes. Normal primary human astrocytes were transiently transfected with pEAK10-PAR-2-HA plasmids. At 24 h after transfection, cells were stained to detect the Jab1 and PAR-2-HA proteins by double immunofluorescence staining. As shown in Fig. 5B, Jab1 was mainly localized in the cytosol, as well as at the plasma membrane in normal primary human astrocytes (left panel). PAR-2-HA, on the other side, was predominantly localized at the plasma membrane (middle panel), where it colocalized with Jab1, shown by the merge in the right panel of Fig. 5B. Therefore, these data demonstrated that PAR-2 interacts with Jab1 in normal primary human astrocytes.
PAR-2 Activation Reduces Interaction with Jab1-To evaluate whether PAR-2 activation affects the interaction with Jab1, HEK293-PAR-2-HAϩJab1myc cells were treated with 100 nM trypsin for 2 and 30 min in serum-free medium. The whole cell lysates were immunopre-FIGURE 3. Interaction of Jab1 with PAR-2 in vivo. A, HEK293 cells were stably co-transfected with PAR-2-HA and Jab1myc or transfected with Jab1myc. The whole cell lysates from HEK293-PAR-2-HAϩJab1myc cells, as well as wild-type HEK293 cells (negative control) and HEK293-Jab1myc cells (negative control), were IP by anti-HA antibody in the presence of protein A-Sepharose beads and the immunocomplex was detected by Western blot (WB) using the anti-myc antibody. The IP was further confirmed by anti-HA antibody. B, the whole cell lysates from wild-type HEK293 cells were IP by anti-PAR-2 (C-17) antibody or by the unrelated antibody against GST (negative control) in the presence of protein G-agarose beads and the immunocomplex was detected by Western blot with the anti-Jab1 antibody. C, HEK293 cells were stably co-transfected with PAR-2⌬IL12-HA and Jab1myc or transfected with Jab1myc. The whole cell lysates from HEK293-PAR-2⌬IL12-HAϩJab1myc cells, as well as wild-type HEK293 cells (negative control) and HEK293-Jab1myc cells (negative control), were IP by anti-HA antibody in the presence of protein A-Sepharose beads and the immunocomplex was detected by Western blot using the anti-myc antibody. The IP was further confirmed by anti-HA antibody. Lysate shows 2% of the cell lysate used for the IP experiments. All experiments were repeated at least three times with identical results. The molecular mass marker is indicated on the right. Note, lower bands in the myc blot are unspecific staining by the myc antibody. cipitated by anti-HA antibody, and the immunocomplex was detected by Western blot analysis using the anti-myc antibody. As shown in Fig.  6A, the co-immunoprecipitation of Jab1 was slightly reduced at 2 min stimulation, compared with that in unstimulated cells (0 min). However, very little Jab1 protein was co-immunoprecipitated by PAR-2-HA at 30 min after stimulation, indicating that the interaction of Jab1 with PAR-2 was disrupted. The membrane reprobed with HA antibody revealed the specificity of the co-immunoprecipitation. To rule out any unspecific proteolytic activity of trypsin and to determine the specific activation of PAR-2, we treated the cells with the specific PAR-2 peptide agonist PAR-2 AP (100 M), which has been shown to specifically activate PAR-2 (29). PAR-2 activation by PAR-2 AP similarly reduced the interaction between Jab1 and PAR-2 at 2 min and abolished it at 30 min after stimulation (Fig. 6 B). The time course was similar to that observed with trypsin stimulation (Fig. 6A). These data obviously demonstrated that PAR-2 interacts with Jab1 in an agonist-dependent manner.
The activated PAR-2 receptor is internalized and rapidly sorted into lysosomes for degradation (30). To clarify whether the reduction of the interaction between Jab1 and PAR-2 results from protein dissociation or from PAR-2 degradation, we pretreated HEK293-PAR-2-HAϩJab1myc cells for 30 min with the lysosomal protease inhibitor ZPAD (100 M) (24) followed by 100 M PAR-2 AP stimulation for another 30 min. As shown in Fig. 7A, also in this case PAR-2 AP stimulation significantly reduced the interaction of PAR-2 with Jab1. However, ZPAD did not block the reduction of the interaction induced by PAR-2 activation. The data summarized in Fig. 7B demonstrate that PAR-2 activation induced a 64 Ϯ 2% loss of interaction of PAR-2 with Jab1 at 30 min after PAR-2 AP stimulation (p Ͻ 0.001), and ZPAD had no effect on this reduction (p Ͼ 0.05). The inhibitory effect of ZPAD was confirmed, as shown in Fig. 7C. Long-term stimulation (3 and 5 h) by PAR-2 AP time dependently led to PAR-2 degradation in the presence of the protein synthesis inhibitor cycloheximide (70 M). This apparently was prevented by the lysosomal protease inhibitor ZPAD (100 M). These results indicate that protein dissociation, but not PAR-2 degradation, leads to a loss of the interaction at 30 min after PAR-2 activation.
On the other side, we treated cells for 15 min with the inhibitor of receptor endocytosis phenylarsine oxide (80 M, PAO) (30) prior to PAR-2 AP stimulation. Interestingly, PAO partially inhibited the PAR-2-induced reduction of the interaction at 30 min (Fig. 7D). The summarized data demonstrate that PAR-2 activation led to only 33 Ϯ 3% loss of interaction of PAR-2 with Jab1 after pretreatment with PAO, and the reduction of the interaction was much less than that induced by PAR-2 AP alone (p Ͻ 0.001, Fig. 7E). Therefore, the dynamic interaction between PAR-2 and Jab1 was dependent on agonist stimulation and receptor endocytosis.
The Effect of PAR-2 Activation on Jab1 Expression and Distribution-We further examined whether Jab1 expression was upregulated by PAR-2 activation. HEK293-Jab1myc cells were treated with 100 nM trypsin for 0 -120 min in serum-free medium. Western blot analysis using the anti-Jab1 antibody demonstrated that the transfected (upper bands) and endogenous (lower bands) Jab1 expression were both not influenced by PAR-2 activation. The membrane reprobed with ␤-tubulin I confirmed the comparable loading for all lanes (Fig. 8).
We also studied the effect of PAR-2 activation on the distribution of endogenous Jab1 in HEK293-PAR-2-HA cells by immunofluorescence staining. As shown in Fig. 9, the endogenous Jab1 was localized in the cytosol and nucleus, as well as at the plasma membrane in unstimulated cells. After 10 min incubation with 100 nM trypsin or with 100 M PAR-2 AP, Jab1 was detected below the plasma membrane. After 30 min, Jab1 was accumulated in the cytosol and was not detected at the plasma membrane. Therefore, PAR-2 activation induced the rapid redistribution of endogenous Jab1. A, the whole cell lysates from normal primary human astrocytes were IP by anti-PAR-2 (C-17) antibody or by the unrelated antibody against GST (negative control) in the presence of protein G-agarose beads and the immunocomplex was detected by Western blot (WB) with the anti-Jab1 antibody. Lysate shows 2% of the cell lysate used for the IP experiments. A typical blot from three independent experiments is given. The molecular mass marker is indicated on the right. B, primary human astrocytes (NHA) were transiently transfected with PAR-2-HA. 24 h after transfection, cells were fixed, permeabilized, stained, and observed by a confocal microscope. Jab1 (green) was visualized by monoclonal anti-Jab1 antibody and Alexa Fluorா 488 goat anti-mouse IgG. PAR-2-HA (red) was visualized by polyclonal anti-HA antibody and Alexa Fluorா 568 goat anti-rabbit IgG. The overlay image (Merge, yellow) revealed the colocalization of Jab1 with PAR-2. All images are representative for three independent experiments. Scale bar, 10 m. Jab1 Mediates PAR-2-induced c-Jun Activation-Because Jab1 is known as a coactivator of c-Jun (12), here we studied whether PAR-2 activation increases endogenous c-Jun phosphorylation in HEK293-Jab1myc cells. The HEK293-Jab1myc cells and wild-type HEK293 cells were rinsed with Hank's solution, and treated with 100 M PAR-2 AP for 0 -2 h in serum-free medium. As shown in Fig. 10, A and B, PAR-2 AP stimulation time dependently increased c-Jun phosphorylation in HEK293-Jab1myc cells. The peak was obtained at 5 min, then the phosphorylation decreased gradually during 30 min. Further studies demonstrated that the increase in c-Jun phosphorylation induced by PAR-2 activation was significantly higher at 10 min in HEK293-Jab1myc cells than that in wildtype HEK293 cells (p Ͻ 0.01; Fig. 10C), although c-Jun was also activated by PAR-2 AP at 10 min in wild-type HEK293 cells (Fig. 10, A and C).
Our own observations demonstrated that PAR-1 activation could significantly up-regulate total c-Jun, besides c-Jun phosphorylation. 3 Therefore, here we investigated whether PAR-2 activation could also regulate the total c-Jun concentration via interaction with Jab1. As shown in Fig. 10, A, D, and E, a similar pattern was observed. Total c-Jun was slightly increased in a time-dependent manner by PAR-2 AP (100 M) stimulation in HEK293-Jab1myc cells, but not in wild-type HEK293 cells. Moreover, total c-Jun was significantly up-regulated at 10 min upon PAR-2 activation in HEK293-Jab1myc cells, compared with that in wild-type HEK293 cells (p Ͻ 0.05). These results suggest that the interaction of PAR-2 with Jab1 results in activation of c-Jun.
Jab1 Potentiates PAR-2-induced AP-1 Activation-The activated c-Jun translocates into the nucleus, and interacts with Fos and other Jun proteins to generate a dimer called the AP-1 complex. This AP-1 complex binds to a conserved DNA motif to regulate gene transcription and expression (31). Therefore, here we studied whether PAR-2 activation further regulated AP-1 activity via interaction with Jab1, by using an AP-1-driven CAT reporter gene. HEK293-Jab1myc cells and wild-type HEK293 cells were transiently transfected with pBL AP-1-CAT2 plasmids as well as the negative control pBL CAT2 plasmids. We treated these cells for 48 h with 100 M PAR-2 AP at 6 h after transfection. The CAT activity in cell lysates was measured by CAT enzyme-linked immunosorbent assay. As shown in Fig. 11, PAR-2 AP stimulation significantly increased the CAT activity by 36% in HEK293-Jab1myc cells  The whole cell lysates were IP by anti-HA antibody in the presence of protein A-Sepharose beads, and the immunocomplex was detected by Western blot (WB) using the anti-myc antibody. A and D, representative blots from three independent experiments. Lysate shows 0.5% of the cell lysate used for the IP experiments. The molecular mass marker is indicated on the right. B and E, the signal for the interaction was quantitated by analyzing the density by Quantity One software (Bio-Rad), and values were normalized to that in control cells. Mean Ϯ S.E., ***, p Ͻ 0.001, compared with control; ###, p Ͻ 0.001, compared with PAR-2 AP stimulation alone. C, the inhibitory effect of ZPAD on PAR-2 degradation. HEK293-PAR-2-HA cells were preincubated with cycloheximide (70 M) alone or together with ZPAD (100 M) for 30 min, followed by PAR-2 AP stimulation for 3 and 5 h. PAR-2-HA expression was determined by Western blot analysis. Representative blots from three independent experiments are given. ␤-Tubulin I detection served as loading control. HEK, wild-type HEK293 cells. The molecular mass marker is indicated on the right. HEK293-Jab1myc cells were treated with 100 nM trypsin for 0 -2 h in serum-free medium. Western blot analysis showed that trypsin stimulation did not influence the Jab1 expression in HEK293-Jab1myc cells. The upper bands represent the transfected Jab1myc protein, whereas the lower bands represent the endogenous Jab1 protein. ␤-Tubulin I served as loading control. Experiments were repeated three times with identical results. The molecular mass marker is indicated on the right. The endogenous Jab1 was stained using the monoclonal antibody against Jab1, and visualized with the Alexa Fluor 488 goat anti-mouse IgG by a confocal microscope. Areas marked by a square are enlarged and shown on the right. The endogenous Jab1 was translocated to below the plasma membrane at 10 min after PAR-2 agonist treatment, and was redistributed to the cytosol at 30 min. All images are representative for three independent experiments. Scale bar, 10 m.
(p Ͻ 0.05; middle columns in Fig. 11), indicating that PAR-2 activation increased the AP-1 activity in HEK293-Jab1myc cells. The CAT activity itself was not influenced by PAR-2 activation that was shown in cells transfected with control pBL CAT2 plasmid (p Ͼ 0.05). However, PAR-2 activation stimulated by PAR-2 AP did not induce AP-1 activation in wild-type HEK293 cells (left columns in Fig. 11). Therefore, these results suggest the existence of a physical and functional interaction between PAR-2 and Jab1. We then asked whether AP-1 activity could be further up-regulated in HEK293-Jab1myc cells transfected with PAR-2. As shown in the right columns in Fig. 11, the PAR-2-induced AP-1 activity was dramatically enhanced by 93% in HEK293-PAR-2-HAϩJab1myc cells (p Ͻ 0.001). Taken together, these findings clearly demonstrate that Jab1 mediates PAR-2-induced activation of AP-1.
To further confirm the specific function of Jab1 in PAR-2-induced AP-1 activation, we knocked down the endogenous Jab1 using Jab1 siRNA. Transfection of Jab1 siRNA (100 nM) significantly reduced the endogenous Jab1 expression at both mRNA and protein levels in HEK293-PAR-2-HA cells (Fig. 12, A and B). Scrambled siRNA did not affect Jab1 expression (Fig. 12, A and B), confirming the specificity of Jab1 siRNA. Next, we transfected HEK293-PAR-2-HA cells with Jab1 siRNA, followed after 24 h by transfection with the AP-1-driven CAT reporter plasmid. As shown in Fig. 12C, PAR-2 activation by PAR-2 AP significantly enhanced AP-1 activity by 67% in HEK293-PAR-2-HA cells. This stimulation was blocked by Jab1 siRNA, but not affected by scrambled siRNA transfection.   To determine the specificity of Jab1 in the PAR-2 signaling events, we stimulated HEK293-Jab1myc cells with the specific PAR-1 agonist TRag (10 M) (27,32) to examine AP-1 activity. PAR-1 activation failed to increase the AP-1 activity in HEK293-Jab1myc cells (data not shown). These results suggest that Jab1 is specifically involved in PAR-2-induced AP-1 activation.

DISCUSSION
PAR-2 mediates cellular signal transductions to the extracellular serine proteases trypsin, tryptase, and others, and exerts important physiological and pathological functions in multiple systems. In the present study, we for the first time show that Jab1 directly interacts with human PAR-2 receptor and potentiates PAR-2-induced c-Jun/AP-1 activation. GST pull-down findings clearly demonstrate that Jab1 mainly bound to the intracellular loop 3 and the C-tail of PAR-2. However, the intracellular loop 3 alone and the PAR-2 C-tail alone both weakly interacted with Jab1, indicating that multiple domains of PAR-2 are required to promote together the interaction with Jab1. The intracellular loops and the C-tail of PAR-2 are essential for the interaction with heterotrimeric G proteins and other signal effectors, as well as for receptor desensitization, internalization, and degradation, which lead to initiation or termination of PAR-2-mediated signal transduction. Therefore, Jab1 may be an important signal effector to mediate PAR-2-induced signal transduction. Several reports already demonstrated that the MPN (Mpr1p and Pad1p NH 2 -terminal) domain (aa 54 -190), the conserved functional domain of Jab1, mediated the interaction of Jab1 with DNA topoisomerase II␣ (33), or with macrophage migration inhibitory factor (34). In the present work, the truncated Jab1 proteins isolated from yeast (aa 31-334 and 49 -334) include this MPN domain, suggesting that the MPN domain might be responsible for the interaction between Jab1 and PAR-2.
The activated PAR-2 is internalized, rapidly translocated to endosomes, and then sorted into lysosomes. Our unpublished data 4 and others (37) demonstrated that the internalized PAR-2 was localized in the endosomes and partially in the lysosomes at 30 min after agonist stimulation. Here, we found that PAR-2 was almost completely dissociated from Jab1 at this time. However, the lysosomal protease inhibitor ZPAD did not prevent protein dissociation of PAR-2 from Jab1. In contrast, the inhibitor of receptor endocytosis PAO partially inhibited the reduction of protein interaction. Furthermore, immunofluorescence staining studies demonstrated that Jab1 was redistributed from the plasma membrane to the cytosol after PAR-2 activation. Therefore, these data support a model, which is given in Fig. 13, that Jab1 constitutively interacts with PAR-2 in the steady state. After PAR-2 activation, the receptor is internalized, which promotes dissociation of Jab1 from PAR-2. The internalized receptor is translocated to early endosomes, where Jab1 is completely dissociated from PAR-2. Subsequently, PAR-2 receptor is further sorted into lysosomes for degradation. On the other hand, the dissociated Jab1 acting as a signal messenger mediates the PAR-2 signaling to activate downstream factors.
In the present study, we found that PAR-2 activation induced c-Jun phosphorylation, and further activated the transcription factor AP-1.
Importantly, here we provide new insights that Jab1, as a protein partner of PAR-2, mediated PAR-2-induced c-Jun/AP-1 activation. It is well known that the c-Jun NH 2 -terminal kinase as a major protein kinase mediates c-Jun phosphorylation (38). Previous work already demonstrated that tryptase, acting on PAR-2, induced interleukin (IL)-6 and IL-8 release by activation of mitogen-activated protein kinase/AP-1 in human peripheral blood eosinophils, which may result in inflammatory responses (5). Recently, it has been shown that Jab1 is also able to phosphorylate c-Jun and to increase AP-1 activation (12,36). Jab1 is the fifth subunit of the COP9 signalosome, and all eight subunits of the COP9 signalosome form a complex to exert their biological functions in mammals (14). It was reported that the COP9 signalosome possesses the protein kinase activity, which is able to phosphorylate c-Jun, IB␣, and p105 (39). In addition, other studies also demonstrated that the COP9 signalosome associates with protein kinase CK2 and protein kinase D, which phosphorylate c-Jun and p53 (40). However, COP9 signalosomeinduced c-Jun phosphorylation is independent of c-Jun NH 2 -terminal kinase (41). Therefore, our data here demonstrated that Jab1-induced c-Jun/AP-1 activation might be a novel alternative signal transduction pathway for PAR-2, suggesting that Jab1 regulates the signaling of extracellular proteases trypsin, tryptase, and others to the nucleus.
Loss-of-function studies using specific Jab1 siRNA confirmed that Jab1 was involved in this novel signaling transduction pathway. This pathway put forward by us here might explain mechanisms of gene regulation by proteases, neurotransmitters, or hormones through their cognate GPCRs. On the other side, a previous report had shown that Jab1 is able to interact with an intracellular membrane protein, the precursor of lutropin/choriogonadotropin receptor (LHR), but not with the mature cell surface LHR (42). That study further found that Jab1 failed to potentiate the activity of AP-1 in the presence of overexpressed LHR. The explanation given by the authors was that the precursor of LHR competitively interacts with Jab1 to prevent the binding of c-Jun to Jab1. The precursor of LHR is different from the mature receptor, and it is impossible for the precursor of LHR to interact with receptor ligands and to further exert their actions, as does the mature LHR (43). Therefore, it is not surprising that Jab1 functions differently in both cases.
In this study, the reporter gene assay demonstrated that Jab1 specifically potentiated PAR-2-induced AP-1 activation. However, Jab1 failed to increase AP-1 activity induced by the PAR-1 agonist TRag, although it has been reported that thrombin, through PAR-1, could induce AP-1-4 W. Luo and G. Reiser, unpublished data. FIGURE 13. Model for Jab1-mediated signaling pathway for PAR-2. Jab1 constitutively interacts with PAR-2 in the steady state. After PAR-2 activation, the receptor is phosphorylated, and internalized in a clathrin-dependent manner, which promotes dissociation of Jab1 from PAR-2. The internalized receptor is translocated to early endosomes, where Jab1 is completely dissociated from PAR-2. Subsequently, PAR-2 receptor is further sorted into lysosomes for degradation. On the other hand, the dissociated Jab1 binds to c-Jun and thereby activates it. The activated c-Jun then translocates to the nucleus, and binds to AP-1 motif to initiate gene expression. mediated gene expression in 1321N1 astrocytoma cells (44). Therefore, other effectors, rather than Jab1, are involved in PAR-1-induced gene expression mediated by activated AP-1.
It is well known that PAR-2 is involved in inflammatory processes, because the proteases tryptase and trypsin are released during inflammation (1,45), and PAR-2 expression is up-regulated upon stimulation by some proinflammatory factors (46). Moreover, PAR-2 agonists caused an increase in IL-1␤, IL-6, IL-8, nitric oxide, prostaglandin E 2 , matrix metalloproteinase-9, and tumor necrosis factor-␣ productions in vivo and in vitro, which mediate the inflammatory reactions (5,(47)(48)(49). It has been reported that AP-1 regulates the gene transcription and expression of several proinflammatory factors, such as IL-1␤, IL-6, IL-8, and tumor necrosis factor-␣ (50). Therefore, our findings imply that Jab1 might mediate PAR-2-dependent inflammatory responses by regulating the release of proinflammatory factors.
In conclusion, we have found that Jab1 physically interacts with human PAR-2. PAR-2 activation induces the redistribution of Jab1 from the plasma membrane to the cytosol. Meanwhile, the internalization of the activated receptor promotes the dissociation of PAR-2 from Jab1. As a result, they are finally dissociated in the endosomes. Upon PAR-2 activation, Jab1 acting as a signal messenger potentiates PAR-2-induced c-Jun/AP-1 activation, which might mediate release of proinflammatory factors and trigger PAR-2-dependent inflammatory responses (Fig.  13). These data demonstrate a novel alternative signal transduction pathway for PAR-2-dependent gene expression.