Phosphorylation of Claudin-3 at Threonine 192 by cAMP-dependent Protein Kinase Regulates Tight Junction Barrier Function in Ovarian Cancer Cells
2005; Elsevier BV; Volume: 280; Issue: 28 Linguagem: Inglês
10.1074/jbc.m502003200
ISSN1083-351X
AutoresTheresa D’Souza, Rachana Agarwal, Patrice J. Morin,
Tópico(s)Caveolin-1 and cellular processes
ResumoClaudins are integral membrane proteins essential in the formation and function of tight junctions (TJs). Disruption of TJs, which have essential roles in cell permeability and polarity, is thought to contribute to epithelial tumorigenesis. Claudin-3 and -4 are frequently overexpressed in ovarian cancer, but the molecular pathways involved in the regulation of these proteins are unclear. Interestingly, several studies have demonstrated a role for phosphorylation in the regulation of TJ complexes, although evidence for claudin phosphorylation is scarce. Here, we showed that claudin-3 and -4 can be phosphorylated in ovarian cancer cells. In vitro phosphorylation assays using glutathione S-transferase fusion constructs demonstrated that the C terminus of claudin-3 is an excellent substrate for cAMP-dependent protein kinase (PKA). Using site-directed mutagenesis, we identified a PKA phosphorylation site at amino acid 192 in the C terminus of claudin-3. Overexpression of the protein containing a T192D mutation, mimicking the phosphorylated state, resulted in a decrease in TJ strength in ovarian cancer cell line OVCA433. Our results suggest that claudin-3 phosphorylation by PKA, a kinase frequently activated in ovarian cancer, may provide a mechanism for the disruption of TJs in this cancer. In addition, our findings may have general implications for the regulation of TJs in normal epithelial cells. Claudins are integral membrane proteins essential in the formation and function of tight junctions (TJs). Disruption of TJs, which have essential roles in cell permeability and polarity, is thought to contribute to epithelial tumorigenesis. Claudin-3 and -4 are frequently overexpressed in ovarian cancer, but the molecular pathways involved in the regulation of these proteins are unclear. Interestingly, several studies have demonstrated a role for phosphorylation in the regulation of TJ complexes, although evidence for claudin phosphorylation is scarce. Here, we showed that claudin-3 and -4 can be phosphorylated in ovarian cancer cells. In vitro phosphorylation assays using glutathione S-transferase fusion constructs demonstrated that the C terminus of claudin-3 is an excellent substrate for cAMP-dependent protein kinase (PKA). Using site-directed mutagenesis, we identified a PKA phosphorylation site at amino acid 192 in the C terminus of claudin-3. Overexpression of the protein containing a T192D mutation, mimicking the phosphorylated state, resulted in a decrease in TJ strength in ovarian cancer cell line OVCA433. Our results suggest that claudin-3 phosphorylation by PKA, a kinase frequently activated in ovarian cancer, may provide a mechanism for the disruption of TJs in this cancer. In addition, our findings may have general implications for the regulation of TJs in normal epithelial cells. Claudin proteins are a large family of transmembrane proteins essential in the formation and maintenance of tight junctions (TJs). 1The abbreviations used are: TJ, tight junction; FSK, forskolin; GST, glutathione S-transferase; PKA, cAMP-dependent protein kinase; PMA, phorbol 12-myristate 13-acetate; TER, transepithelial electrical resistance; PBS, phosphate-buffered saline; WT, wild type; DMSO, dimethyl sulfoxide; PKAc, PKA catalytic subunit. Tight junctions in epithelial and endothelial cells provide a selective barrier and establish cellular polarity (1Gonzalez-Mariscal L. Betanzos A. Nava P. Jaramillo B.E. Prog. Biophys. Mol. Biol. 2003; 81: 1-44Crossref PubMed Scopus (922) Google Scholar, 2Mitic L.L. Van Itallie C.M. Anderson J.M. Am. J. Physiol. Gastrointest. Liver Physiol. 2000; 279: G250-G254Crossref PubMed Google Scholar, 3Tsukita S. Furuse M. Itoh M. Nat. Rev. Mol. Cell. Biol. 2001; 2: 285-293Crossref PubMed Scopus (2041) Google Scholar, 4Matter K. Balda M.S. Nat. Rev. Mol. Cell. Biol. 2003; 4: 225-236Crossref PubMed Scopus (717) Google Scholar). These structures are typically lost in cancer, and this loss may contribute to the invasive and metastatic phenotype of tumor cells (5Langbein L. Pape U.F. Grund C. Kuhn C. Praetzel S. Moll I. Moll R. Franke W.W. Eur. J. Cell Biol. 2003; 82: 385-400Crossref PubMed Scopus (98) Google Scholar, 6Martin T.A. Jiang W.G. Histol. Histopathol. 2001; 16: 1183-1195PubMed Google Scholar, 7Itoh M. Bissell M.J. J. Mammary Gland Biol. Neoplasia. 2003; 8: 449-462Crossref PubMed Scopus (108) Google Scholar, 8Mullin J.M. Sci. STKE 2004. 2004; : pe2Google Scholar). Using serial analysis of gene expression, we have previously shown that claudin-3 and -4 are among the most highly up-regulated genes in ovarian cancer (9Hough C.D. Sherman-Baust C.A. Pizer E.S. Montz F.J. Im D.D. Rosenshein N.B. Cho K.R. Riggins G.J. Morin P.J. Cancer Res. 2000; 60: 6281-6287PubMed Google Scholar). The high expression of these claudins in ovarian cancer has been confirmed by our group and others using a variety of approaches such as microarrays, tissue arrays, and reverse transcription-PCR (10Rangel L.B. Agarwal R. D'Souza T. Pizer E.S. Alo P.L. Lancaster W.D. Gregoire L. Schwartz D.R. Cho K.R. Morin P.J. Clin. Cancer Res. 2003; 9: 2567-2575PubMed Google Scholar, 11Santin A.D. Zhan F. Bellone S. Palmieri M. Cane S. Bignotti E. Anfossi S. Gokden M. Dunn D. Roman J.J. O'Brien T.J. Tian E. Cannon M.J. Shaughnessy Jr., J. Pecorelli S. Int. J. Cancer. 2004; 112: 14-25Crossref PubMed Scopus (253) Google Scholar, 12Hibbs K. Skubitz K.M. Pambuccian S.E. Casey R.C. Burleson K.M. Oegema Jr., T.R. Thiele J.J. Grindle S.M. Bliss R.L. Skubitz A.P. Am. J. Pathol. 2004; 165: 397-414Abstract Full Text Full Text PDF PubMed Scopus (203) Google Scholar, 13Heinzelmann-Schwarz V.A. Gardiner-Garden M. Henshall S.M. Scurry J. Scolyer R.A. Davies M.J. Heinzelmann M. Kalish L.H. Bali A. Kench J.G. Edwards L.S. Vanden Bergh P.M. Hacker N.F. Sutherland R.L. O'Brien P.M. Clin. Cancer Res. 2004; 10: 4427-4436Crossref PubMed Scopus (177) Google Scholar, 14Lu K.H. Patterson A.P. Wang L. Marquez R.T. Atkinson E.N. Baggerly K.A. Ramoth L.R. Rosen D.G. Liu J. Hellstrom I. Smith D. Hartmann L. Fishman D. Berchuck A. Schmandt R. Whitaker R. Gershenson D.M. Mills G.B. Bast Jr., R.C. Clin. Cancer Res. 2004; 10: 3291-3300Crossref PubMed Scopus (377) Google Scholar). Considering that TJs are typically lost in neoplasia, these findings are somewhat surprising, and the exact role of these proteins in ovarian cancer remains unclear. There is, however, some evidence that claudin overexpression may actually lead to increased cell survival and invasion. 2R. Agarwal, T. D'Souza, and P. J. Morin, unpublished observations. A better understanding of the regulation of claudin proteins in ovarian cancer cells may help clarify their roles in this disease. Several lines of evidence suggest that TJs are involved in various cell signaling pathways, providing a possible link between cell polarity/cell contact and downstream signaling events (4Matter K. Balda M.S. Nat. Rev. Mol. Cell. Biol. 2003; 4: 225-236Crossref PubMed Scopus (717) Google Scholar, 15Schneeberger E.E. Lynch R.D. Am. J. Physiol. Cell Physiol. 2004; 286: C1213-C1228Crossref PubMed Scopus (1113) Google Scholar). Furthermore, the involvement of kinases in the biogenesis and regulation of several tight junction components has been established (16Kale G. Naren A.P. Sheth P. Rao R.K. Biochem. Biophys. Res. Commun. 2003; 302: 324-329Crossref PubMed Scopus (165) Google Scholar, 17Clarke H. Marano C.W. Peralta Soler A. Mullin J.M. Adv. Drug Delivery Rev. 2000; 41: 283-301Crossref PubMed Scopus (83) Google Scholar, 18Wang Y. Zhang J. Yi X.J. Yu F.S. Exp. Eye Res. 2004; 78: 125-136Crossref PubMed Scopus (103) Google Scholar, 19Eckert J.J. McCallum A. Mears A. Rumsby M.G. Cameron I.T. Fleming T.P. Reproduction. 2004; 127: 653-667Crossref PubMed Scopus (44) Google Scholar, 20Karczewski J. Groot J. Am. J. Physiol. Gastrointest. Liver Physiol. 2000; 279: G660-G665Crossref PubMed Google Scholar). However, only a few studies have clearly demonstrated phosphorylation of claudins. For example, phosphorylation of claudin-1 by mitogen-activated protein kinases (21Fujibe M. Chiba H. Kojima T. Soma T. Wada T. Yamashita T. Sawada N. Exp. Cell Res. 2004; 295: 36-47Crossref PubMed Scopus (121) Google Scholar) and protein kinase C (22Nunbhakdi-Craig V. Machleidt T. Ogris E. Bellotto D. White III, C.L. Sontag E. J. Cell Biol. 2002; 158: 967-978Crossref PubMed Scopus (223) Google Scholar), as well as phosphorylation of claudin-5 by cAMP-dependent protein kinase (PKA) (23Ishizaki T. Chiba H. Kojima T. Fujibe M. Soma T. Miyajima H. Nagasawa K. Wada I. Sawada N. Exp. Cell Res. 2003; 290: 275-288Crossref PubMed Scopus (164) Google Scholar, 24Soma T. Chiba H. Kato-Mori Y. Wada T. Yamashita T. Kojima T. Sawada N. Exp. Cell Res. 2004; 300: 202-212Crossref PubMed Scopus (104) Google Scholar), has been reported. Also, WNK4 kinase has been shown to phosphorylate claudin-3 and -4 (25Yamauchi K. Rai T. Kobayashi K. Sohara E. Suzuki T. Itoh T. Suda S. Hayama A. Sasaki S. Uchida S. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 4690-4694Crossref PubMed Scopus (228) Google Scholar). To address the issue of claudin regulation in ovarian cancer, we have investigated phosphorylation as a possible modulatory mechanism of claudin function. We first show that both claudin-3 and -4 are phosphorylated in ovarian cancer cells, although probably by different kinases. Indeed, whereas claudin-3 is a target of PKA, claudin-4 appears to be phosphorylated by protein kinase C. For this report, we have focused on PKA-mediated phosphorylation of claudin-3. We show that claudin-3 is phosphorylated at threonine 192 in the cytoplasmic C-terminal region. Mimicking the phosphorylated state by mutating this site to an aspartic acid decreases TJ strength in ovarian cancer cells. Our data provide the first evidence for the phosphorylation of claudin-3 and -4 in ovarian cancer cells and suggest that phosphorylation of claudin-3 by PKA, a kinase frequently activated in ovarian cancer, may provide a mechanism for the disruption of TJs in this cancer. Cell Culture—Ovarian cell lines BG-1, CAOV3, HEY, IGROV-1, UCI101, 2008, A2780, OVCA420, OVCA429, OVCA432, OVCA433, OVCAR-2, OVCAR-3, OVCAR-4, and OVCAR-5 were cultured in McCoy's 5A growth medium (Invitrogen) supplemented with 10% fetal bovine serum and antibiotics (100 units/ml penicillin and 100 μg/ml streptomycin). HOSE-B, an ovarian surface epithelial cell line immortalized with E6 and E7 (26Gregoire L. Munkarah A. Rabah R. Morris R.T. Lancaster W.D. In Vitro Cell. Dev. Biol. Anim. 1998; 34: 636-639Crossref PubMed Scopus (24) Google Scholar), was maintained in RPMI 1640 medium supplemented with 10% fetal calf serum and 300 μg/ml Geneticin (Invitrogen). HEK293 cells were maintained in McCoy's growth medium with the supplements described above. To establish stable clones of OVCA433, plasmid vectors were transfected using Lipofectamine plus reagent according to the manufacturer's protocol (Invitrogen). Clones were selected and maintained in McCoy's growth medium containing 500 μg/ml Geneticin. Antibodies—Rabbit polyclonal claudin-3 and mouse monoclonal claudin-4 were purchased from Zymed Laboratories Inc. Mouse monoclonal PKAc was part of a sampler kit obtained from BD Transduction Laboratories. Rabbit polyclonal PKAc α was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse monoclonal β-actin and glyceraldehyde-3-phosphate dehydrogenase were purchased from Abcam (Cambridge, UK). Peroxidase-linked donkey anti-rabbit immunoglobulin and sheep anti-mouse IgG horseradish antibodies were obtained from Amersham Biosciences. Alexa fluor antibodies were purchased from Molecular Probes. Immunohistochemistry—Expression of claudin-3 and PKAc α was assessed by immunohistochemical staining of normal ovarian tissue and an ovarian cancer tissue array constructed in our laboratory (10Rangel L.B. Agarwal R. D'Souza T. Pizer E.S. Alo P.L. Lancaster W.D. Gregoire L. Schwartz D.R. Cho K.R. Morin P.J. Clin. Cancer Res. 2003; 9: 2567-2575PubMed Google Scholar). Antigen-bound primary antibody was detected using streptavidin-biotin immunoperoxidase complex (Ultravision detection system; Labvision). For negative controls, absence of primary antibodies processed in parallel showed no positive staining. Images were acquired by Axiovision 3.1 software on a Zeiss Axiovert S100 microscope under ×40 objective lens (Carl Zeiss, Thornwood, NY). Immunoblotting—Total cell lysates were resolved by SDS-PAGE (Tris-glycine gels; Invitrogen) and transferred to polyvinylidene difluoride membranes (Millipore). Membranes were blocked with 5% nonfat dry milk in 10 mm Tris, pH 7.5, 100 mm NaCl, and 0.1% Tween 20 (v/v) and incubated overnight with primary antibody (claudin-3, 1:200; claudin-4, 1:250, PKAc, 1:1000; β-actin, 1:2000; glyceraldehyde-3-phosphate dehydrogenase, 1:4000). Following the washes in Tris-buffered saline, membranes were probed with horseradish peroxidase-conjugated antibody (anti-mouse or anti-rabbit, 1:10,000). ECL Western blotting detection kit (Amersham Biosciences) was used for visualization of the positive bands. In Vivo Phosphorylation and Immunoprecipitation—Ovarian cancer cell lines HEY, OVCA432, and OVCA433 were plated at 80–90% confluence in 60-mm dishes. Cells were serum-starved for 16 h and then labeled with [32P]orthophosphate (500 μCi/ml; MP Biomedicals, Irvine, CA) in phosphate-free, serum-free Dulbecco's modified Eagle's medium for 3 h at 37 °C. Cells were stimulated with forskolin (FSK; 30 μm; Sigma) and phorbol 12-myristate 13-acetate (PMA; 50 ng/ml; Sigma) for 20 min at 37 °C. When used, H89 (30 μm; Sigma) was added 30 min prior to the addition of FSK. HEK293 cells were plated at 70–80% confluence in 100-mm dishes and transfected with plasmid pCMV PKAc (Clontech), together with a plasmid encoding either human claudin-3 or the T192A mutant using FuGENE 6 transfection reagents. After 18 h of transfection, cells were labeled with [32P]orthophosphate (500 μCi/ml). Cells were washed in ice-cold phosphate-buffered saline (PBS) and incubated with 700 μl of modified radioimmune precipitation assay buffer (50 mm Tris-Cl, pH 7.4, 150 mm NaCl, 1% Nonidet P-40, 1.5 mm MgCl2, 10% glycerol, 1 mm EDTA, 25 mm NaF, 1 mm Na3VO4, Complete™ EDTA-free protease inhibitor mixture (Roche Applied Science) for 10 min at 4 °C. The lysate was cleared by centrifugation at 10,000 × g for 20 min. For immunoprecipitation, the supernatant was incubated overnight with 2 μg of primary antibody and 40 μl of protein A-Sepharose (Amersham Biosciences). Immune complexes were washed three times with the lysis buffer without inhibitors, with the final elution in 2× sample buffer, boiled for 5 min, and separated on SDS-PAGE. The gels were either dried, exposed to film for autoradiography, or transferred onto polyvinylidene difluoride membranes and immunoblotted as described above. Generation of Claudin-3 and -4 WT and Mutant Constructs—The coding sequence of human CLDN3 and CLDN4 was amplified by PCR from cDNA obtained from ovarian cancer cell line OVCAR5 and cloned into pCI-neo mammalian expression vector (Promega, Madison, WI) using the EcoRI and SalI restriction sites. DNA fragments of the GST fusion constructs were generated by PCR using the claudin plasmids described above. GST fusion proteins were constructed with the C-terminal areas of claudin-3 (amino acids 185–220) and claudin-4 (amino acids 181–209). The fragments were designed to contain EcoRI/SalI restriction sites allowing them to be subcloned in-frame with the GST fragment in pGEX-4T-2 vector (Amersham Biosciences). Point mutations of the serine or threonine residues in the consensus PKA site were introduced using the QuikChange site-directed mutagenesis kit (Stratagene). All constructs were confirmed by DNA sequencing. Expression and Purification of GST-claudin Fusion Proteins— Recombinant constructs of pGEX-CLDN3C, pGEX-CLDN3C-T192A, pGEX-CLDN3C-S199A, and pGEX-CLDN4C were transformed in BL-21 Escherichia coli (Stratagene) for protein expression. Purification was performed according to the manufacturer's protocol with some modifications. Briefly, cultures were grown overnight, and protein expression was initiated with 0.1 mm isopropyl β-d-thiogalactopyranoside at 30 °C for 3 h. Bacterial pellets obtained after centrifugation at 8000 rpm for 10 min at 4 °C were resuspended in cold PBS with Complete™ EDTA-free protease inhibitor (Roche Applied Science), 1 mm phenylmethylsulfonyl fluoride, and 1 mg/ml lysozyme. After sonication, the lysate was cleared by centrifugation at 10,000 rpm for 20 min. Fusion proteins were purified by affinity chromatograpy using glutathione-Sepharose 4B (Amersham Biosciences). The Sepharose beads were washed three times with PBS, and the fusion protein was eluted with glutathione Elution Buffer (50 mm Tris-Cl/5 mm reduced glutathione). In Vitro Phosphorylation—Purified GST, GST-fusion proteins, and immunoprecipitated claudin-3 (WT and T192A) proteins were incubated with kinase buffer containing 50 mm Tris-Cl, pH 7.5, 5.1 mm MgCl2, 100 μm ATP, and 1 μCi of [γ-33P]ATP. Catalytic subunit of PKA (5U; Promega) was added to the reaction to a final volume of 20 μl and incubated at 30 °C for 30 min. The reaction was stopped with 5× sample buffer. The proteins was separated on SDS-PAGE (14% Tris-glycine gels), dried, and exposed for autoradiography. Coomassie Blue staining and immunoblot were used to confirm equal loading of proteins. TER Measurements and Calcium Switch Assay—OVCA433 cells were seeded at a density of 5 × 105 on 12-mm polycarbonate transwell clear membranes (0.4-μm pore size; Costar, Cambridge, MA). TER was measured using a Millicell-ERS V-ohm meter (WPI, New Haven, CT). The values were normalized for the area of the filter and obtained after subtraction of blank values (filter and bath solution). For experiments with drugs, DMSO (vehicle) and FSK (30 μm) were added at time 0, with measurements taken every hour thereafter. For the calcium switch assay, the calcium in the medium was chelated using EGTA at a final concentration of 4 mm (27Citi S. J. Cell Biol. 1992; 117: 169-178Crossref PubMed Scopus (257) Google Scholar). Experiments were performed 7 days after seeding, after the TER had reached an optimal value. Cells were washed with Hanks' balanced salt solution and replaced with normal medium or low calcium medium. TER readings were taken for 4–5 h followed by replacement of low calcium medium with normal medium, and measurements were taken the following day. Permeability Assay—OVCA433 clones were plated at 5 × 105 cells/transwell and grown for 7 days to confluence and optimal TER development. Fluorescein isothiocyanate-dextran, with an average molecular mass of 4 kDa (Sigma-Aldrich), dissolved in McCoy's media to a concentration of 1 mg/ml was added to the upper chamber. The lower chamber was replaced with fresh media. At the 0, 4, and 24 h time points, 50-μl aliquots were collected from the lower chamber and assayed by luminescence spectrometer using excitation at 485 nm and emission at 530 nm (Cytofluor; PerSeptive Biosystems). The value for vector-transfected OVCA433 was used for normalization and set at 100%. Immunofluorescence—OVCA433 clones were washed with PBS, followed by fixation with cold methanol:acetone (1:1) for 4 min. OVCA433 cells treated with DMSO (vehicle), FSK (30 μm), or FSK+H89 were fixed in a similar fashion. Fixed cells were washed three times with PBS and blocked in 5% bovine serum albumin, followed by incubation with claudin-3 (1:100) at room temperature for 1 h. After three washes with PBS, the cells were stained with Alexa 594 goat anti-rabbit antibody for 1 h. Statistical Analysis—Data are expressed as means ± S.E. Statistical analysis was performed by using repeated measures analysis of variance and Tukey post hoc test (28Zar J.H. Biostatistcal Analysis. 4th Ed. Prentice Hall Inc., Upper Saddle River, NJ1999Google Scholar) (using SAS software version 9), with p < 0.05 considered statistically significant. Claudin-3 and -4 Are Phosphorylated in Ovarian Cancer Cell Lines—We first sought to determine whether claudin-3 and -4 could be phosphorylated in ovarian cancer cells. We used in vivo phosphorylation assays in which we stimulated metabolically labeled ovarian cancer cell lines with FSK and PMA, a PKA and a protein kinase C activator, respectively (Fig. 1A). Compared with unstimulated cells, claudin-3 phosphorylation was significantly increased in both OVCA432 and OVCA433 cells. In addition, some of the claudin-3 protein seems to be constitutively phosphorylated in the unstimulated OVCA432 cells. HEY cells, which do not express endogenous claudin-3, were included as a negative control, and the phosphorylated bands observed in these cells were nonspecific, with molecular masses slightly different from those of the claudin bands. Similar results were obtained with claudin-4 (Fig. 1B). Indeed, treatment of cells with the agonists induced phosphorylation of claudin-4 in both OVCA432 and OVCA433 cell lines. The increase in claudin-4 phosphorylation is particularly apparent in OVCA433 cells, in which phosphorylation is essentially absent without activation but highly elevated after FSK/PMA treatment. HEY cells were again used as negative control because they lack endogenous claudin-4 protein as well. To determine which of the agonists increased the phosphorylation of claudin-3 and -4 proteins, an in vivo phosphorylation assay was performed in OVCA433 cells in the presence of either FSK or PMA. Fig. 1C shows that claudin-3 was strongly phosphorylated following FSK treatment, whereas PMA increased the phosphorylation of claudin-4 in OVCA433 cells. Similar results were obtained in OVCA432 cells (data not shown). These results indicate that, in these cells, claudin-3 is phosphorylated by PKA, whereas claudin-4 is phosphorylated by protein kinase C. Moreover, FSK-dependent phosphorylation of claudin-3 was inhibited by H89, a PKA inhibitor, bringing it back to basal phosphorylation level (Fig. 1D). This experiment further strengthens our conclusion that claudin-3 is phosphorylated in vivo by PKA in these cells. Expression of PKA Catalytic and Claudin-3 Proteins in Ovarian Cancer Cell Lines and Cancer Tissues—Because of the findings above suggesting that claudin-3 is a likely target for PKA in ovarian cancer cells, we decided to examine endogenous protein levels of PKAc α in a panel of ovarian cancer cells. PKAc α was prominently present in all cell lines examined except HEY, which also fails to express claudin-3 (Fig. 2A). Combined expression of the catalytic subunit with the various regulatory subunits can generate various physiological consequences in different cells (29Tasken K. Aandahl E.M. Physiol. Rev. 2004; 84: 137-167Crossref PubMed Scopus (625) Google Scholar). Most of the cell lines in Fig. 2A also expressed regulatory subunits (data not shown), indicating the existence of functional PKA complexes in these cells. This is in agreement with previous reports showing a frequent activation of PKA in ovarian cancer (30McDaid H.M. Cairns M.T. Atkinson R.J. McAleer S. Harkin D.P. Gilmore P. Johnston P.G. Br. J. Cancer. 1999; 79: 933-939Crossref PubMed Scopus (60) Google Scholar, 31Alper O. Hacker N.F. Cho-Chung Y.S. Oncogene. 1999; 18: 4999-5004Crossref PubMed Scopus (28) Google Scholar). Claudin-3 expression was absent in non-malignant HOSE-B cells but found in most of the ovarian cancer lines. Overall, 11 of 16 cell lines expressed both claudin-3 and PKAc α. This is consistent with our earlier finding demonstrating expression of claudin-3 protein in the vast majority of ovarian cancers (10Rangel L.B. Agarwal R. D'Souza T. Pizer E.S. Alo P.L. Lancaster W.D. Gregoire L. Schwartz D.R. Cho K.R. Morin P.J. Clin. Cancer Res. 2003; 9: 2567-2575PubMed Google Scholar). Staining of normal ovarian tissue with PKAc α and claudin-3 indicated the presence of PKAc α in the epithelium and stroma of the tissue, but claudin-3 was clearly absent (Fig. 2B, a and d), similar to what we observed for the expression of these proteins in the HOSE-B non-maligant ovarian epithelial cells (Fig. 2A). Staining of our ovarian cancer tissue array with PKAc α antibody showed that this protein was present in all the ovarian cancer tissues and that the areas of PKAc staining overlapped with areas of claudin-3 staining (Fig. 2B, b, c, e, and f). These proteins were found both at the membrane and cytoplasm. The insets of Fig. 2B, c and f, which are shown at higher magnification, verify the localization of PKAc α and claudin-3. These results are consistent with our hypothesis that claudin-3 may be a target for PKA phosphorylation in ovarian cancer cells. Recombinant Claudin-3 Protein Is Phosphorylated in Vitro by PKA—We next wished to determine the exact PKA phosphorylation site(s) in claudin-3. We performed an in vitro phosphorylation assay with GST fusion proteins containing the C terminus of claudin-3 and -4 (Fig. 3A). PKA strongly phosphorylated the claudin-3 C-terminal region, whereas claudin-4 was very weakly phosphorylated compared with claudin-3 (Fig. 3B). The GST protein without claudin sequences was not phosphorylated in these assays. These in vitro findings are consistent with the in vivo observations, showing that claudin-3 is strongly phosphorylated by a PKA activator compared with claudin-4 (Fig. 1). To determine which of the serine/threonine residues in the C-terminal region of claudin-3 is phosphorylated, we performed site-directed mutagenesis on putative sites that met the minimal consensus sequence for PKA (R/KXT/S) (Fig. 3A). Fig. 3C shows that a single threonine to alanine substitution at position 192 of the C-terminal region of claudin-3 abolished the phosphorylation signal. Mutation at another putative phosphorylation residue (Ser-199), although leading to a slight reduction in signal, did not prevent strong phosphorylation of the claudin-3 C terminus. In addition, various GST fusions with other regions of claudin-3 (N-terminal, loop, and transmembrane regions) were not phosphorylated in these in vitro assays (data not shown). Claudin-3 Is Phosphorylated in Vitro and in Vivo by PKA at Thr-192—We next asked whether full-length claudin-3 could be phosphorylated in vitro and in vivo at Thr-192. Because a theoretical tryptic digest analysis of claudin-3 showed that the fragment containing Thr-192 would be out of the detection range for mass spectrometry, we resorted to genetic approaches to answer this question. Full-length claudin-3 WT and mutant (T192A) were overexpressed in two cell lines, HEY and HEK293, both of which were devoid of endogenous claudin-3. Immunoprecipitated proteins were incubated with PKAc along with radiolabeled ATP for a kinase assay. Immunoprecipitates of claudin-3 WT from both HEY and HEK293 exhibited phosphorylated claudin-3, whereas no phosphorylation was observed in cells expressing the mutant T192A (Fig. 3D). In addition, we co-expressed claudin-3 WT or its mutant T192A with PKAc in HEK293 cells. Fig. 3E shows that, in the presence of the active PKAc, claudin-3 WT was phosphorylated, whereas the mutant T192A was not. Interestingly, the phosphorylated claudin-3 was more prominent in the Nonidet P-40-soluble fraction than in the insoluble fraction (Fig. 3F). Claudin-3 Phosphorylation Affects Barrier Function—To evaluate the functional role of PKA phosphorylation in the barrier function of TJs, we overexpressed claudin-3 WT, mutant T192A (non-phosphorylatable), and mutant T192D (mimicking the phosphorylated state of the residue) in ovarian cancer cells. For this analysis, we chose the OVCA433 cell line because, among the ovarian cancer cell lines tested, these cells had the highest TJ function as evaluated by transepithelial electrical resistance (TER) measurements. Non-transfected cells, as well as cells expressing claudin-3 WT or the T192A mutant, developed relatively high TER over a period of ∼7 days (Fig. 4A). However, expression of the T192D mutant, mimicking the phosphorylated state, led to a significant inhibition of TER with at least a 6- or 4-fold reduction compared with the values of vector or claudin-3 WT, respectively (F3,16 = 14, p < 0.0001) at day 7, suggesting that the TJs could not form properly. The data shown here are from one set of OVCA433 clones, but comparable results were obtained with another series of clones, as well as with pools of transfectants (data not shown). We carried out similar experiments with Madin-Darby canine kidney I cells, which have high TER and are commonly used in TJ studies. Similar to what we observed for OVCA433 cells, expression of the mutant T192D in Madin-Darby canine kidney I cells decreased the TER 4- or 2-fold compared with vector or claudin-3 WT, respectively (data not shown). We also assessed permeability changes of OVCA433 stable clones using the paracellular tracer fluorescein isothiocyanate-dextran (4 kDa) across the established monolayers. Compared with vector-transfected cells, cells expressing claudin-3 T192D mutant had significantly increased paracellular flux of fluorescein isothiocyanate-dextran. Indeed, the transported dextran was increased by 30% after 4 h of incubation and by 150% after 24 h of incubation (Fig. 4B). The other two stable transformants expressing claudin-3 WT and mutant T192A exhibited permeability properties similar to the vector-transfected control. Recovery of Barrier Function is Normal in Claudin-3-overexpressing Cells after Ca2+ Switch—Removal of extracellular calcium results in a PKA-dependent disruption of TJs (32Klingler C. Kniesel U. Bamforth S.D. Wolburg H. Engelhardt B. Risau W. Histochem. Cell Biol. 2000; 113: 349-361Crossref PubMed Scopus (44) Google Scholar). PKA has also been shown to affect the resealing of TJs following the replenishment of calcium, possibly via vasodilator-stimulated phosphoprotein, a direct substrate for PKA, impl
Referência(s)