Claudin-8 Expression in Madin-Darby Canine Kidney Cells Augments the Paracellular Barrier to Cation Permeation
2003; Elsevier BV; Volume: 278; Issue: 19 Linguagem: Inglês
10.1074/jbc.m213286200
ISSN1083-351X
AutoresAlan S.L. Yu, Alissa H. Enck, Wayne I. Lencer, Eveline E. Schneeberger,
Tópico(s)Connexins and lens biology
ResumoClaudins are a family of integral membrane proteins of the tight junction that are thought to participate in the permeation of solutes across epithelia via the paracellular pathway. Claudin-8 is expressed in the distal renal tubule, which has a characteristically low passive permeability to monovalent cations. To test the hypothesis that claudin-8 plays a role in forming a tight paracellular barrier to cations, stably transfected Madin-Darby canine kidney II cell lines with inducible expression of claudin-8 were generated. Induction of claudin-8 expression was associated with down-regulation of endogenous claudin-2 protein. Other tight junction proteins were expressed and targeted normally, and the number of junctional strands was minimally altered. By Ussing chamber and radiotracer flux studies, claudin-8 expression was found to reduce paracellular permeability to monovalent inorganic and organic cations and to divalent cations but not to anions or neutral solutes. The size selectivity, charge dependence, and activation energy of paracellular cation permeation were all unchanged. These observations are consistent with a model in which claudin-2 encodes a highly cation-permeable channel, whereas claudin-8 acts primarily as a cation barrier. When exogenous claudin-8 is expressed, it replaces endogenous claudin-2, inserting in its place into existing tight junction strands, thereby reducing the apparent number of functional cation pores. Our findings suggest that claudin-8 plays an important role in the paracellular cation barrier of the distal renal tubule. Claudins are a family of integral membrane proteins of the tight junction that are thought to participate in the permeation of solutes across epithelia via the paracellular pathway. Claudin-8 is expressed in the distal renal tubule, which has a characteristically low passive permeability to monovalent cations. To test the hypothesis that claudin-8 plays a role in forming a tight paracellular barrier to cations, stably transfected Madin-Darby canine kidney II cell lines with inducible expression of claudin-8 were generated. Induction of claudin-8 expression was associated with down-regulation of endogenous claudin-2 protein. Other tight junction proteins were expressed and targeted normally, and the number of junctional strands was minimally altered. By Ussing chamber and radiotracer flux studies, claudin-8 expression was found to reduce paracellular permeability to monovalent inorganic and organic cations and to divalent cations but not to anions or neutral solutes. The size selectivity, charge dependence, and activation energy of paracellular cation permeation were all unchanged. These observations are consistent with a model in which claudin-2 encodes a highly cation-permeable channel, whereas claudin-8 acts primarily as a cation barrier. When exogenous claudin-8 is expressed, it replaces endogenous claudin-2, inserting in its place into existing tight junction strands, thereby reducing the apparent number of functional cation pores. Our findings suggest that claudin-8 plays an important role in the paracellular cation barrier of the distal renal tubule. Claudin-8 expression in Madin-Darby canine kidney cells augments the paracellular barrier to cation permeation. Vol. 278 (2003) 17350-17359Journal of Biological ChemistryVol. 279Issue 49PreviewPage 17356, Fig. 7: The permeability measurements shown in the original figure were 10-fold too high. The corrected figure is shown below: Full-Text PDF Open Access Madin-Darby canine kidney cells transepithelial resistance doxycycline Vectorial transport across epithelia can occur via the transcellular or paracellular route. The molecular basis of transcellular transport, mediated by proteins that catalyze transmembrane movement of solutes and water at the apical and basolateral surfaces, is well understood. By contrast, little is known about the mechanisms of paracellular transport. The tight junction, which is the most apical of the intercellular junctional complexes at the lateral membrane, is generally believed to be the rate-limiting step in paracellular transport. The tight junction is composed of a complex of multiple proteins, many of uncertain function (1Denker B.M. Nigam S.K. Am. J. Physiol. 1998; 274: F1-F9Crossref PubMed Google Scholar, 2Fanning A.S. Mitic L.L. Anderson J.M. J. Am. Soc. Nephrol. 1999; 10: 1337-1345PubMed Google Scholar). Most of these proteins are cytoplasmic in location, and several are associated with the cytoskeleton via the perijunctional actin ring. In recent years, a number of integral membrane proteins associated with the tight junction have been identified, including occludin (3Furuse M. Hirase T. Itoh M. Nagafuchi A. Yonemura S. Tsukita S. Tsukita S. J. Cell Biol. 1993; 123: 1777-1788Crossref PubMed Scopus (2141) Google Scholar), junction-associated membrane proteins 1 and 2 (4Martin-Padura I. Lostaglio S. Schneemann M. Williams L. Romano M. Fruscella P. Panzeri C. Stoppacciaro A. Ruco L. Villa A. Simmons D. Dejana E. J. Cell Biol. 1998; 142: 117-127Crossref PubMed Scopus (1157) Google Scholar, 5Cunningham S.A. Arrate M.P. Rodriguez J.M. Bjercke R.J. Vanderslice P. Morris A.P. Brock T.A. J. Biol. Chem. 2000; 275: 34750-34756Abstract Full Text Full Text PDF PubMed Scopus (132) Google Scholar), Coxsackie- and adenovirus-associated receptor (6Cohen C.J. Shieh J.T. Pickles R.J. Okegawa T. Hsieh J.T. Bergelson J.M. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 15191-15196Crossref PubMed Scopus (558) Google Scholar), 1G8 antigen (7Nasdala I. Wolburg-Buchholz K. Wolburg H. Kuhn A. Ebnet K. Brachtendorf G. Samulowitz U. Kuster B. Engelhardt B. Vestweber D. Butz S. J. Biol. Chem. 2002; 277: 16294-16303Abstract Full Text Full Text PDF PubMed Scopus (188) Google Scholar), and the claudins (8Furuse M. Sasaki H. Fujimoto K. Tsukita S. J. Cell Biol. 1998; 143: 391-401Crossref PubMed Scopus (794) Google Scholar). Integral membrane proteins are of particular interest because, by definition, their extracellular domains must protrude into the lateral intercellular space and, therefore, may potentially have direct contact with solutes as they permeate through the paracellular pathway. Claudins are the most heterogeneous of these integral membrane proteins and consist of a family of at least 20 homologous isoforms (9Tsukita S. Furuse M. J. Cell Biol. 2000; 149: 13-16Crossref PubMed Scopus (407) Google Scholar). Moreover, each isoform clearly exhibits a tissue-specific and segment-specific pattern of distribution in epithelia such as those of the gastrointestinal tract (10Rahner C. Mitic L.L. Anderson J.M. Gastroenterology. 2001; 120: 411-422Abstract Full Text Full Text PDF PubMed Scopus (449) Google Scholar) and renal tubule (11Yu A. Enck A. J. Am. Soc. Nephrol. 2001; 12 (abstr.): 45Google Scholar, 12Kiuchi-Saishin Y. Gotoh S. Furuse M. Takasuga A. Tano Y. Tsukita S. J. Am. Soc. Nephrol. 2002; 13: 875-886Crossref PubMed Google Scholar, 13Reyes J.L. Lamas M. Martin D. Del Carmen Namorado M. Islas S. Luna J. Tauc M. Gonzalez-Mariscal L. Kidney Int. 2002; 62: 476-487Abstract Full Text Full Text PDF PubMed Scopus (143) Google Scholar). Claudins are therefore leading candidates for the molecular determinants responsible for the variety of different paracellular permeability properties found in different epithelia. Several recent reports provide convincing evidence that claudins regulate paracellular permeability. Human mutations in claudin-16 cause failure of paracellular reabsorption of divalent cations in the thick ascending limb of the renal tubule, leading to familial hypercalciuric hypomagnesemia (14Simon D.B. Lu Y. Choate K.A. Velazquez H. Al-Sabban E. Praga M. Casari G. Bettinelli A. Colussi G. Rodriguez-Soriano J. McCredie D. Milford D. Sanjad S. Lifton R.P. Science. 1999; 285: 103-106Crossref PubMed Scopus (974) Google Scholar), and a preliminary report suggests that knockout mice lacking claudin-16 may have a similar phenotype (15Lu Y. Choate K.A. Wang T. Lifton R.P. J. Am. Soc. Nephrol. 2001; 12 (abstr.): 756Google Scholar). Both claudin-1-deficient mice (16Furuse M. Hata M. Furuse K. Yoshida Y. Haratake A. Sugitani Y. Noda T. Kubo A. Tsukita S. J. Cell Biol. 2002; 156: 1099-1111Crossref PubMed Scopus (1200) Google Scholar) and transgenic mice overexpressing claudin-6 (17Turksen K. Troy T.C. Development. 2002; 129: 1775-1784PubMed Google Scholar) exhibit abnormally high epidermal permeability to water. Overexpression of claudins 1, 4, and 15 in MDCK1 cells all increase transepithelial resistance (TER) (18Inai T. Kobayashi J. Shibata Y. Eur. J. Cell Biol. 1999; 78: 849-855Crossref PubMed Scopus (225) Google Scholar, 19McCarthy K.M. Francis S.A. McCormack J.M. Lai J. Rogers R.A. Skare I.B. Lynch R.D. Schneeberger E.E. J. Cell Sci. 2000; 113: 3387-3398PubMed Google Scholar, 20Van Itallie C. Rahner C. Anderson J.M. J. Clin. Invest. 2001; 107: 1319-1327Crossref PubMed Scopus (496) Google Scholar, 21Colegio O.R. Van Itallie C.M. McCrea H.J. Rahner C. Anderson J.M. Am. J. Physiol. Cell Physiol. 2002; 283: 142-147Crossref PubMed Scopus (440) Google Scholar), whereas overexpression of claudin-2 markedly decreases TER (22Furuse M. Furuse K. Sasaki H. Tsukita S. J. Cell Biol. 2001; 153: 263-272Crossref PubMed Scopus (628) Google Scholar) by selectively increasing permeability to cations (23Amasheh S. Meiri N. Gitter A.H. Schoneberg T. Mankertz J. Schulzke J.D. Fromm M. J. Cell Sci. 2002; 115: 4969-4976Crossref PubMed Scopus (636) Google Scholar). Furthermore, Anderson and co-workers (20Van Itallie C. Rahner C. Anderson J.M. J. Clin. Invest. 2001; 107: 1319-1327Crossref PubMed Scopus (496) Google Scholar, 21Colegio O.R. Van Itallie C.M. McCrea H.J. Rahner C. Anderson J.M. Am. J. Physiol. Cell Physiol. 2002; 283: 142-147Crossref PubMed Scopus (440) Google Scholar) elegantly show that, although overexpression of claudin-4 reduces paracellular monovalent cation permeability, this can be abolished by altering the net charge at the first extracellular loop by site-directed mutagenesis. Similarly, although claudin-15 alone does not alter the relative preference of the paracellular pathway between Na+ and Cl−, mutating anionic extracellular residues to cationic ones makes the paracellular pathway more anion-selective (21Colegio O.R. Van Itallie C.M. McCrea H.J. Rahner C. Anderson J.M. Am. J. Physiol. Cell Physiol. 2002; 283: 142-147Crossref PubMed Scopus (440) Google Scholar). Thus, the contention that claudins have a direct role in regulating the magnitude and nature of paracellular permeability is now irrefutable. However, the plethora of published claudin overexpression and gene ablation experiments has uncovered a fundamental paradox. Claudin overexpression can be associated with both an increase (17Turksen K. Troy T.C. Development. 2002; 129: 1775-1784PubMed Google Scholar, 22Furuse M. Furuse K. Sasaki H. Tsukita S. J. Cell Biol. 2001; 153: 263-272Crossref PubMed Scopus (628) Google Scholar, 23Amasheh S. Meiri N. Gitter A.H. Schoneberg T. Mankertz J. Schulzke J.D. Fromm M. J. Cell Sci. 2002; 115: 4969-4976Crossref PubMed Scopus (636) Google Scholar) and a decrease (18Inai T. Kobayashi J. Shibata Y. Eur. J. Cell Biol. 1999; 78: 849-855Crossref PubMed Scopus (225) Google Scholar, 19McCarthy K.M. Francis S.A. McCormack J.M. Lai J. Rogers R.A. Skare I.B. Lynch R.D. Schneeberger E.E. J. Cell Sci. 2000; 113: 3387-3398PubMed Google Scholar, 20Van Itallie C. Rahner C. Anderson J.M. J. Clin. Invest. 2001; 107: 1319-1327Crossref PubMed Scopus (496) Google Scholar, 21Colegio O.R. Van Itallie C.M. McCrea H.J. Rahner C. Anderson J.M. Am. J. Physiol. Cell Physiol. 2002; 283: 142-147Crossref PubMed Scopus (440) Google Scholar) in paracellular permeability; similarly, knockout or inactivating mutations of claudin genes can cause either an increase (16Furuse M. Hata M. Furuse K. Yoshida Y. Haratake A. Sugitani Y. Noda T. Kubo A. Tsukita S. J. Cell Biol. 2002; 156: 1099-1111Crossref PubMed Scopus (1200) Google Scholar) or a decrease (14Simon D.B. Lu Y. Choate K.A. Velazquez H. Al-Sabban E. Praga M. Casari G. Bettinelli A. Colussi G. Rodriguez-Soriano J. McCredie D. Milford D. Sanjad S. Lifton R.P. Science. 1999; 285: 103-106Crossref PubMed Scopus (974) Google Scholar, 15Lu Y. Choate K.A. Wang T. Lifton R.P. J. Am. Soc. Nephrol. 2001; 12 (abstr.): 756Google Scholar) in paracellular permeability. Thus, in the absence of suitable models to explain such contradictory observations, attempts to infer the permeability properties of specific claudin isoforms remain problematic. A possible solution emerges if one views each tight junction strand as a continuous string of protein molecules (such as claudins, occludin, or other proteins) arrayed side-by-side, rather like beads on a necklace, so as to form an uninterrupted seal (9Tsukita S. Furuse M. J. Cell Biol. 2000; 149: 13-16Crossref PubMed Scopus (407) Google Scholar). In such a model, claudins would play a bipartite role, acting as a barrier by sealing the gaps between protein particles, and simultaneously providing a channel through that barrier from the apical compartment to the lateral intercellular space. This may explain why the complete loss of a claudin isoform can have diametrically opposite consequences. In the absence of any changes in other tight junction proteins, gaps would appear in the continuous seal, and the paracellular route would become more permeable. Conversely, when other components of the tight junction become up-regulated and fill in for the absent claudins, the paracellular route may become less permeable. Similar considerations may govern the consequences of overexpression experiments. The resulting phenotype is complex and dependent on the properties both of the heterologously expressed claudin and of endogenous tight junction proteins. In this manuscript, we describe a new tissue culture model with overexpression of claudin-8. We chose to study claudin-8 because previous in situ hybridization (11Yu A. Enck A. J. Am. Soc. Nephrol. 2001; 12 (abstr.): 45Google Scholar) and immunohistochemical studies (12Kiuchi-Saishin Y. Gotoh S. Furuse M. Takasuga A. Tano Y. Tsukita S. J. Am. Soc. Nephrol. 2002; 13: 875-886Crossref PubMed Google Scholar) revealed that it is expressed primarily in the aldosterone-sensitive distal nephron. In these segments of the renal tubule, the paracellular barrier is particularly tight to monovalent cations and protects transtubular gradients up to 1000:1 for H+, 20:1 for K+, and 1:3 for Na+. We therefore tested the hypothesis that claudin-8 plays a role in impeding paracellular cation permeation. We show that the combination of careful biochemical, histological, and physiological studies constitute a powerful tool to interpret the phenotype of such overexpression experiments. We propose potential models to explain the findings of such studies and show data that indicate that claudin-8 acts as a nonspecific cation barrier. A 675-bp DNA fragment containing the mouse claudin-8 coding sequence except for the initiation codon was amplified by PCR from a cDNA clone (IMAGE ID: 162549) and inserted into a FLAG epitope tag shuttle vector based on pcDNA3 (24Dowland L.K. Luyckx V.A. Enck A.H. Leclercq B. Yu A.S.L. J. Biol. Chem. 2000; 275: 37765-37773Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar) so that the claudin-8 N terminus was fused in-frame with a sequence encoding MDYKDDDDKGS (the FLAG octapeptide tag is underlined followed by a glycine-serine linker encoding BamHI restriction site). The entire coding region was then excised and cloned into the retroviral Tet response vector, pRevTRE (Clontech, Palo Alto, CA) to obtain the plasmid, pRev-mCLDN8-NFL. Plasmid DNA was transfected by lipofection into the packaging cell line, PT67, a polyclonal culture of stable transfectants selected using hygromycin, and virus-containing supernatant collected from the growth medium. MDCK II TetOff cells expressing the tetracycline-regulated transactivator from a cytomegalovirus promoter (Clontech) were infected with viral supernatant in the presence of Polybrene. Stable transfectants were selected by growth in 0.3 mg/ml hygromycin. Clonal cell lines were isolated using plastic cloning rings and three clones (2, 4, and 9) with strong induction of claudin-8 expression (see “Results”) used for further studies. Cells were maintained in Dulbecco's modified Eagle's medium with 57 fetal bovine serum, 0.1 mg/ml G418, 0.3 mg/ml hygromycin, and 20 ng/ml doxycycline (Dox+). To induce claudin-8 expression, doxycycline was omitted from the culture medium starting from the day of plating (Dox−). Studies were generally performed after 4–6 days except where otherwise indicated. To detect tight junction protein expression by immunoblotting, cultured cells were first homogenized and fractionated by centrifugation at 100,000 × g as described previously (24Dowland L.K. Luyckx V.A. Enck A.H. Leclercq B. Yu A.S.L. J. Biol. Chem. 2000; 275: 37765-37773Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar). No claudin-8 was ever detectable in the soluble fraction (100,000 × g supernatant). The total amount of protein isolated from plates of confluent Dox+ and Dox− cells was similar, and the fractional yield of membrane protein (100,000 × g pellet) was roughly 157 in both. Aliquots of 25 ॖg of membrane protein were then electrophoresed on denaturing SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membrane, and Western blots were performed using the ECL chemiluminescence kit (Amersham Biosciences). Claudin-8 transgene expression was detected with the M2 anti-FLAG monoclonal antibody (Sigma) at a 1:400 dilution. Antibodies to ZO1, occludin, and claudins 1–4 were obtained from Zymed Laboratories Inc., San Francisco, CA and used at the concentrations recommended by the manufacturer. Rabbit anti-Coxsackie- and adenovirus-associated receptor antiserum (a kind gift of Dr. Christopher Cohen, Children's Hospital of Philadelphia, PA) was used at 1:5000 dilution. Western blots were digitized with a flatbed optical scanner. Gray scale values were logarithmically transformed to obtain uncalibrated optical density estimates, and individual bands were quantitated using NIH Image 1.61 software (rsb.info.nih.gov/ nih-image). To determine the localization of claudin-8 protein expression, cell lines grown on 12-mm Transwell-Clear polyester filters (1-cm2 growth area, 0.4-ॖm pore size; Corning Costar, Acton, MA) were fixed with 47 paraformaldehyde at 4 °C for 15 min, then indirect immunofluorescence staining was performed either with the M2 mouse antibody at 1:400 dilution or (for the double labeling study shown in Fig. 1C) with a rabbit anti-claudin-8 antibody (933) at 1:100 dilution, all in the presence of 0.37 Triton X-100 using protocols described previously (24Dowland L.K. Luyckx V.A. Enck A.H. Leclercq B. Yu A.S.L. J. Biol. Chem. 2000; 275: 37765-37773Abstract Full Text Full Text PDF PubMed Scopus (30) Google Scholar). Affinity-purified rabbit polyclonal antibody 933 was raised against the claudin-8 C-terminal peptide, CQRSFHAEKRSPSIYSKSQYV (25Morita K. Furuse M. Fujimoto K. Tsukita S. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 511-516Crossref PubMed Scopus (982) Google Scholar). Images were acquired at the University of Southern California Center for Liver Diseases using a Nikon PCM confocal microscopy system with argon and helium-neon lasers. Freeze fracture with or without immunolabeling was performed essentially as described previously (19McCarthy K.M. Francis S.A. McCormack J.M. Lai J. Rogers R.A. Skare I.B. Lynch R.D. Schneeberger E.E. J. Cell Sci. 2000; 113: 3387-3398PubMed Google Scholar). Confluent monolayers were fixed either in 27 glutaraldehyde at 4 °C for 30 min (no immunolabeling) or in 17 paraformaldehyde at 4 °C for 15 min (for immunolabeling). They were then rinsed in phosphate-buffered saline, scraped from the filter, and infiltrated with 257 glycerol in 0.1 m cacodylate buffer for 60 min at 4 °C. Cell pellets were frozen in liquid nitrogen slush, and freeze-fractured at −115 °C in a Balzers 400 freeze-fracture unit (Lichtenstein). Processing of the replicas for strand morphometry and immunolabeling was as described (26McCarthy K.M. Skare I.B. Stankewich M.C. J. Cell Sci. 1996; 109: 2287-2298Crossref PubMed Google Scholar), and the replicas were examined with a Philips 301 electron microscope. The number of parallel strands in the tight junctions was quantified as described previously (26McCarthy K.M. Skare I.B. Stankewich M.C. J. Cell Sci. 1996; 109: 2287-2298Crossref PubMed Google Scholar). Because the frequency distribution of strand number was non-Gaussian, the difference in the median number of strands in Dox+ and Dox− cells was assessed for statistical significance using the Wilcoxon ranked sum test. Immunolabeling of SDS-treated replicas was performed with the M2 anti-FLAG antibody (1:100) followed by goat anti-mouse gold (1:100). Cells were plated at confluent density (2 × 105 cells/cm2) onto Transwell filters loaded in 12-well plates. The resistance across each filter in culture media was measured at room temperature by immersing into the well chopstick-style Ag/AgCl electrodes attached to a Millicell-ERS voltohmeter (Millipore, Bedford, MA). TER was calculated by subtracting the resistance determined in blank filters from the resistance measured in filters with monolayers. Cells were plated at confluent density on 1 cm2 Snapwell polyester filters (Corning Costar) and grown for the indicated number of days. The filter rings were then detached and mounted in Ussing chambers that were incubated in Ringer solution (150 mm NaCl, 2 mmCaCl2, 1 mm MgCl2, 10 mm glucose, 10 mm HEPES, pH 7.4) at 37 °C and continuously bubbled with 957 O2, 57 CO2. The fluid volume on each side of the filter was 4 ml. Voltage-sensing electrodes consisting of Ag/AgCl pellets and current-passing electrodes of silver wire were connected by agar bridges containing 3m KCl and interfaced via head-stage amplifiers to a microcomputer-controlled voltage/current clamp (DM-MC6 and VCC-MC6, respectively; Physiologic Instruments, San Diego, CA). Voltage-sensing electrodes were matched to within 1 mV asymmetry and corrected by an offset-removal circuit. The voltage between the two compartments (values reported are referenced to the apical side) was monitored and recorded at 5-s intervals, whereas the current was continuously clamped to zero. This was done to minimize current-induced local changes of salt concentration in the unstirred layers, which can generate spontaneous potentials due to the transport-number effect (27Wedner H.J. Diamond J.M. J. Membr. Biol. 1969; 1: 92-108Crossref PubMed Scopus (61) Google Scholar). The voltage was first measured with blank filters in each combination of buffers to be used for the experiments. The values obtained, which were generally less than 1 mV in magnitude, represent the difference in junction potentials between the two voltage-sensing bridges summed with any potential that might exist across the filter membrane. These were subtracted from all subsequent measurements with filters containing attached cell monolayers to determine transepithelial voltage (Vt). The total resistance between apical and basal compartments was determined in Ringer at the start and at intervals throughout the experiment from the voltage evoked by a 5-ॖA bipolar current pulse. The background resistance determined with blank filters (62.0 ± 1.0 ohm, n = 6), representing the sum of the resistances of the filter, of the fluid in the chambers and of the current-passing electrodes and bridges was subtracted from the total resistance measured with filters containing attached cell monolayers to determine the TER and hence conductance (TER−1). TER was found to be maintained within 57 of its base-line value for at least 2 h, whereas the duration of an experiment was typically about 60–90 min. Dilution and biionic potentials were determined by replacing the solution of one compartment, generally the basal side to minimize flow disruption of the integrity of the monolayer, while keeping the other side bathed in Ringer. In selected cases the solution in the apical compartment was replaced instead of the basal side and showed identical results (see for example Fig. 5B). For 2:1 NaCl dilution potentials, the 150 mm NaCl in Ringer was replaced with 75 mm NaCl, and the osmolality was maintained with mannitol. For biionic potentials, 150 mm NaCl was replaced with 150 mm chloride salt of the indicated alkali metal or organic cation. Organic cations that were obtained as free amine compounds were titrated to neutrality with HCl to form the chloride salt. All organic cations used had a pKa greater than 9.0; therefore, they were all assumed to be completely protonated at pH 7.4. In biionic potential experiments using near-impermeant organic cations, contamination of the solution with residual traces of Na+could lead to a significant overestimation of the permeability (28Dwyer T.M. Adams D.J. Hille B. J. Gen. Physiol. 1980; 75: 469-492Crossref PubMed Scopus (271) Google Scholar). To exclude this, multiple washes with Na+-free, organic cation-containing buffer were performed. A “bracketed” protocol was used in which each measurement in asymmetrical salt solutions was both preceded and followed by measurements in symmetrical Ringer to control for any time-dependent variation in the properties of the monolayer and for “memory” effects on the liquid junction potential (29Barry P.H. Diamond J.M. J. Membr. Biol. 1970; 3: 93-122Crossref PubMed Scopus (161) Google Scholar). Transmonolayer [3H]mannitol, [4C]urea, and [45]Ca tracer flux studies were performed by a modification of the method described by McCarthy et al.(19McCarthy K.M. Francis S.A. McCormack J.M. Lai J. Rogers R.A. Skare I.B. Lynch R.D. Schneeberger E.E. J. Cell Sci. 2000; 113: 3387-3398PubMed Google Scholar). Studies were performed in 12-well plates of Transwell filters in culture medium that already contained 1.8 mmCa2+. For mannitol and urea assays, 1 mmunlabeled substrate was also added to the medium in both compartments. Flux studies were initiated by adding 1–4 ॖCi/ml of the appropriate radioisotope (specific activity, 1–2 Ci/mol) to one side (cis compartment) followed by incubation at 37 °C. At 30-min intervals, 100 ॖl of medium was collected from thetrans compartment for liquid scintillation counting and replaced with an equal volume of fresh medium at 37 °C. Tracer accumulation for all three solutes was found to be linear from 30 to 60 min, and this was used to determine the flux rate and, hence, total permeability, PTot. To correct for the effect of the filter membrane and adjacent unstirred layers, the tracer permeability across blank filters, PBlank, was determined concurrently. Transepithelial permeability,PTE, was then calculated from the following equation.PTE=[(1/PTot)−(1/PBlank)]−1Equation 1 Data are presented as the means ± S.E., where n indicates the number of monolayers from a single experiment. Differences between groups were assessed for statistical significance using the unpaired two-tailed Student's t test. Nonlinear regression analyses were performed by the Levenberg-Marquardt method using GraphPad Prism 3 software. All results shown are representative of at least three separate experiments unless otherwise indicated. The relative ionic permeabilities of the monolayers were calculated using the Goldman-Hodgkin-Katz equation. This is justified because the equation was shown to fit our data well (see “Results” and Fig. 5). The individual permeabilities to Na+(PNa) and Cl−(PCl) in symmetrical 150 mm NaCl were deduced from the method of Kimizuka and Koketsu (30Kimizuka H. Koketsu K. J. Theor. Biol. 1964; 6: 290-305Crossref PubMed Scopus (87) Google Scholar) using the following equations,PNa=G·(RT/F2)/(a(1+औ))Equation 2 PCl=PNa·औEquation 3 where G is the conductance per unit surface area,a is the NaCl activity, and औ is the ratio of the permeability of Cl− to that of Na+ as determined by the Goldman-Hodgkin-Katz equation. All calculations used activities rather than concentrations. The mean activity coefficient of each monovalent cation-halide salt was assumed to be the same as that of NaCl, and the anion and cation in each case were assumed to have the same activity coefficient (Guggenheim assumption). To investigate the role of claudin-8 in paracellular permeability, we first generated cell lines overexpressing the claudin-8 protein. We first ascertained by low stringency Northern blot analysis that the canine renal tubule epithelial cell line, MDCK II, does not express claudin-8 endogenously (data not shown). We then generated by retroviral transduction three MDCK II clonal cell lines with stable heterologous expression of N-terminal FLAG epitope-tagged claudin-8 under the control of the TetOff system. Claudin-8 expression was induced in these cells by omitting doxycycline from the medium (Dox−) and suppressed by adding doxycycline (Dox+). By Western blotting and by immunohistochemistry using the FLAG antibody, claudin-8 protein was found to be expressed in Dox− cells but was completely undetectable in Dox+ cells (Fig. 1,A and B). By immunohistochemistry, claudin-8 in confluent Dox− cell monolayers was expressed at the apical end of the lateral cell membrane, where the tight junction is located, and colocalized with the tight junction scaffolding protein, ZO1 (Fig.1C). On electron microscopic examination of freeze fracture images, the number of tight junction strands appeared similar (Fig.2, A and B). Careful quantitative morphometry (Fig. 2C) revealed a small but statistically significant increase in strand number (median 4 in Dox− cells, 3 in Dox+, p < 0.0001). Tight junction morphology was mostly normal, although we observed
Referência(s)