Kidney International, Vol. 62 (2002), pp. 476–487

The renal segmental distribution of claudins changes with development JOSE LUIS REYES, MONICA LAMAS, DOLORES MARTIN, MARIA DEL CARMEN NAMORADO, SOCORRO ISLAS, JOSE LUNA, MICHEL TAUC, and LORENZA GONZA´LEZ-MARISCAL Center for Research and Advanced Studies (CINVESTAV), Department of Physiology, Biophysics and Neurosciences, Me´xico City, Mexico, and Laboratoire de Physiologie Cellulaire et Moleculaire, Universite de Nice, Nice, France

The renal segmental distribution of claudins changes with development. Background. Permeability properties of mammalian nephron are tuned during postnatal maturation. The transepithelial electrical resistance (TER) and complexity of tight junctions (TJs) vary along the different tubular segments, suggesting that the molecules constituting this structure change. We studied the differential expression of occludin and several claudins in isolated renal tubules from newborn and adult rabbits. Methods. Isolated renal tubules from newborn and adult rabbits were processed for occludin, claudin-1 and claudin-2 immunofluorescence, and Western blot detection of claudin-1 and -2. Claudin-5 was detected in whole kidney frozen sections. RT-PCR from isolated tubules was performed for claudins-1 to -8. Results. Immunofluorescence revealed that occludin, claudin-1 and -2 were present at the cell boundaries at the neonatal stage of development. Claudin-1 was detected in the tighter segments of the nephron (distal and collecting duct), while claudin-2 was found in the leaky portions (proximal). Claudin 5 was found in the kidney vasculature. PCR amplification revealed the presence of claudins-1 to -4 in tubules of newborns. In adults, claudins-1, -2 and -4 were present in proximal, Henle’s loop and collecting segments; claudin-3 was in proximal and collecting tubules, while claudins-5 and -6 were absent from all tubular portions. Claudin-7 was restricted to proximal tubules, while claudin-8 was present in proximal and Henle’s segments. Conclusions. The pattern of occludin distribution is present from the neonatal age. Claudins-7 and -8 are up-regulated after birth. Each tubular segment expresses a peculiar set of claudins that might be responsible for the permeability properties of their TJs.

functions as a fence that maintains a polarized distribution of lipids and proteins between the apical and basolateral plasma membrane [1, 2]. Although the biochemical nature of TJs remained elusive for a long time, the discovery of several TJ and TJ-associated molecules presented the opportunity of studying the relationship between structure and function of these cell-cell junctions with improved precision. Using conventional freeze-fracture, tight junctions have been described as a series of strands and grooves that anastomose to form a meshwork that encircles the cells at the uppermost portion of the lateral membrane [3]. The recent development of sodium dodecyl sulfate– digested freeze-fracture replica labeling (SDS-FRL) allowed the immunocytochemical detection of the proteins constituting TJ strands [4]. This procedure thus permitted the identification of occludin and claudins for the first time as TJ filament components [5–7]. Occludin, with a molecular mass of ⬃65 kD, is comprised of four transmembrane domains, a long COOH-terminal cytoplasmic region, a short NH2-terminal cytoplasmic domain, two extracellular loops and one intracellular turn. Occludin is a functional component of TJs as (1) its overexpression in epithelial cells augments the transepithelial electrical resistance (TER) [8, 9]; (2) introduction of a carboxyl truncated occludin increases paracellular leakage [10]; (3) addition of a synthetic peptide homologous to the second extracellular loop of occludin down-regulates the TER [11]; and (4) introduction of terminally truncated occludin into epithelial cells alters the maintenance of the polarized distribution of lipids and proteins in the plasma membrane [5, 9]. Despite this evidence, the structure and function of TJs cannot be explained by occludin alone, as endothelial and Sertoli cells possess TJs that lack occludin [12, 13] and occludin knockout mice display TJs that do not appear to be morphologically or electrophysiologically altered [14]. These results predicted the existence of other proteins capable of forming TJ strands, even without occludin.

The tight junction (TJ) constitutes the main barrier in epithelia to the passive movement of electrolytes and macromolecules through the paracellular pathway. It also Key words: occludin, nephron permeability, tight junction, postnatal maturation, kidney tubules, growth and development. Received for publication July 25, 2001 and in revised form December 10, 2001 Accepted for publication March 22, 2002

 2002 by the International Society of Nephrology

476

Reyes et al: Renal segmental distribution of claudins

Such proteins, named claudins, have been identified in recent years and may constitute the barrier component of TJs. Claudins are 23 kD integral membrane proteins with four transmembrane domains, bear no sequence similarity to occludin, and when introduced into epithelial cells or fibroblasts, incorporate into preexisting TJs [5] or form well developed networks of strands similar to native TJs [15]. Today, 20 members of this family have been identified [16]. Among them, claudins-3, -4, -5 and -11 had been previously reported as rat ventral prostate-1 protein [17, 18], the receptor for Clostridium perfringens enterotoxin [19], the protein deleted in the Velo-Cardio-Facial syndrome [20], and an oligodendrocyte specific protein [21], respectively. More recently paracellin-1, whose mutations cause renal Mg2⫹ wasting in humans, was identified as claudin-16 [22]. The observation that this claudin is found exclusively in the TJ of the thick ascending limb of Henle, a region responsible for Mg2⫹ reabsorption, gave rise to the proposal that claudins may be selective paracellular channels [23]. In MadinDarby canine kidney (MDCK) cells, two strains that significantly differ in their TER also display a differential expression pattern of claudins. Claudin-1 and -4 are expressed in both tight (MDCK I) and leaky (MDCK II) strains, while the expression of claudin-2 is restricted to the leaky type. Conversion of the epithelia from tight to leaky was achieved by introducing claudin-2 into type I MDCK cells [24]. Furthermore, overexpression of claudin-1 in type II MDCK cells increases the TER fourfold [25]. Together these results suggest that a differential expression of claudins could be responsible for the wide range of permeabilities exhibited in epithelia. In the mammalian kidney the TER and the complexity of the TJ observed by freeze-fracture increase from the proximal to the collecting duct [26, 27]. For instance, while TER in the thick ascending limb of Henle’s has reported values of 25 and 34 ⍀ · cm2 [28, 29], it is of 150 ⍀ · cm2 in the distal tubule [30] and of 867 ⍀ · cm2 in the collecting duct [31]. Considering that epithelia constitute the barrier between the organisms and the environment and that at the proximal segment the epithelium is in contact with an isotonic fluid, similar to plasma, while at the collecting duct the tubular content is urine, the existence of a tighter barrier in the distal portions of the nephron became predictable. In agreement with those data, our previous study examining isolated renal tubules found that the expression of the TJ proteins zonula occludens-1 (ZO-1), ZO-2 and occludin paralleled the increase in junctional complexity found along the mammalian nephron [27]. Our current study initially explored whether the cell border distribution of occludin is already established in neonatal tubules. As no differences with adult tissues were found, we concentrated this study on the distribution of claudins along isolated renal segments derived

477

from adult and newborn animals. Using immunofluorescence and Western blot analyses, we found that the tighter segments of the mammalian nephron have claudin-1 at their TJs, while the leaky proximal segments lack this claudin and instead express claudin-2. This differential distribution of claudins-1 and -2 found in adult tubules is already present in newborn animals. By reverse transcription-polymerase chain reaction (RT-PCR) performed in isolated tubules, the mRNA expression of claudins-1 to -8 was explored. While claudins-1 to -4 are clearly expressed in newborn and adult derived tubules, claudins-7 and -8 are only detected in the latter. Claudins-5 and -6 were amplified from total kidney extracts, but remained undetectable in isolated renal segments, claudins-1 to -4 were expressed in both tight and leaky tubules, claudin-7 was only present in leaky segments, and claudin-8 was restricted to leaky and intermediately permeable renal portions. These findings show that renal tubules change the set of claudins that constitutes their TJs in accordance with their postnatal development, and demonstrate that segments with distinct permeability express a differential combination of claudins. METHODS Animals New Zealand male white rabbits (2.0 to 2.5 kg) were maintained on Diet Science (5321) and water ad libitum in our animal house (temperature 22 to 24⬚C, 50 to 55% humidity). Newborn rabbits (5 days old) were maintained with the dam until the day of the experiment. Care and handling of the animals were in accordance to internationally recommended procedures. Isolation of rabbit renal tubules Rabbit tubules were isolated by manual dissection as previously described [27]. Positive identification of different renal segments Nephron segments were identified according to their morphological appearance as previously described [27]. To further distinguish proximal from distal tubules, which sometimes are difficult to recognize after fixation, cytokeratin-8 immunostaining was performed. As found in earlier studies, while distal tubules show a positive staining, proximal tubules do not [32]. Kidney frozen sections Frozen sections from whole kidney were obtained as previously described by us [27]. Immunofluorescence The whole kidney frozen sections or the isolated tubules were fixed and processed for immunofluorescence as previously described [27]. The following primary anti-

478

Reyes et al: Renal segmental distribution of claudins

bodies were used: rabbit polyclonal anti-claudin-1 (Zymed 71-7800, Zymed, South San Francisco, CA, USA; dilution 15 ␮g/mL); claudin-2 (Zymed 51-6100; dilution 1 ␮g/mL); claudin-5 (Zymed 34-1600, dilution 20 ␮g/mL) occludin (Zymed 71-1500; 10 ␮g/mL) and mouse monoclonal antibodies (mAb) against cytokeratin no. 8 (Boehringer 1238-817; 5 ␮g/mL; Mannheim, Germany) and the luminal surface of intercalated cells (mAb 503). As secondary antibodies a FITC-conjugated goat anti-rabbit IgG (Zymed 65-6111; 6 ␮g/mL) or an anti-mouse IgG TRITC conjugate developed in goat (Sigma T-5393; 1:100, 7.5 ␮g/mL) were used. The frozen kidney sections were quenched for unspecific autofluorescence with an additional 10-minute treatment with Evan’s blue (0.02%, dilution 1:100). The fluorescence was examined using a confocal microscope (MRC-600; Bio-Rad Laboratory, Richmond, CA, USA) with a krypton argon laser. The images collected had an optical thickness of 1 micron for proximal, distal and collecting tubules. For Henle’s loop, due to its narrowness, 0.5-micron optical sections were performed. The images shown represent a projection of the sections made for each tubule. Percoll gradient method for isolation of rabbit tubules Isolation of renal tubules in a Percoll gradient was performed according to the technique reported by Vinay, Gougoux and Lemieux [33]. This procedure establishes a gradient of density in which the tissue separates into four distinct bands. The F4 band was significantly enriched in proximal tubules while the F1 band was markedly enriched in distal tubules, as assessed by microscopic observation of these tubular populations. Bands 2 and 3 have a mixture of nephron segments that include glomeruli, proximal and distal segments. Protein blotting and analysis Bands F1 and F4 from the Percoll gradient were washed twice with cold phosphate-buffered saline (PBS) and incubated for 30 minutes at 4⬚C in Triton X-100 lysis buffer (Tris-HCl 50 mmol/L, pH 7.4, NaCl 100 mmol/L, MgCl2 5 mmol/L, CaCl2 5 mmol/L, Triton X-100 1% and Nonidet P40 1%) with Complete娃 and phenylmethylsulfonyl fluoride (PMSF; 20 ␮g/mL), and sonicated three times for 30 seconds each in a high-intensity ultrasonic processor (Vibra cell; Sonics & Materials Inc., Danbury, CT, USA). Samples were centrifuged at 13,000 rpm at 4⬚C for ten minutes, and the supernatant was collected (soluble fraction). The pellet (insoluble fraction) was further lysed with RIPA buffer [40 mmol/L Tris-HCl pH 7.6, 150 mmol/L NaCl, 2 mmol/L ethylenediaminetetraacetic acid (EDTA), 10% glycerol, 1% Triton X-100, 0.5% sodium deoxycholate, 0.2% SDS]. After protein quantitation [34], samples were diluted (1:1) in treatment buffer (0.125 mol/L Tris-Cl, 4% SDS, 20% glycerol, 10% 2-mercaptoethanol, pH 6.8) and run in 8% polyacryl-

amide gels. Proteins in these gels were either stained for one hour with 0.25% Coomassie blue, or transferred to nitrocellulose papers (Hybond-C, RPN 303C; Amersham, Little Chalfont, UK), [35]. To identify the proteins, blotting was performed with rabbit polyclonal antibodies against claudin-1 (Zymed 71-7800; dilution 15 ␮g/mL); claudin-2 (Zymed 51-6100; dilution 1 ␮g/mL) and occludin (Zymed 71-1500; 10 ␮g/mL). Peroxidase-conjugated goat anti rabbit IgG (H⫹L) (Zymed 62-6120; dilution 1:2000, 0.38 ␮g/mL) was used as secondary antibody followed by a chemiluminescence detection system (ECL⫹PLUS, RPN 2132; Amersham). Nested PCR Kidney slices and manually dissected renal tubules were maintained in RNAlater (Cat 7021; Ambion, Austin, TX, USA) before RNA extraction with the TriPure Isolation Reagent (Cat. 1665157; Boehringer Mannheim) following the manufacturer’s instructions. Claudin cDNA was reverse transcribed from RNA using rTth DNA Polymerase (Cat. N808-0069; Perkin Elmer, Norwalk, CT, USA) with the 3⬘ claudin-1 primer CAGC CAAGGCCTGCATAGCCATGG for claudin-1, and the degenerate 3⬘ claudin primer AGSCRYCADSGG GTYRTAGAAGTC for claudins-2 to -8. A reverse transcription master mix was prepared by adding 2 ␮L of 10⫻ buffer (100 mmol/L Tris-HCl, pH 8.3, 900 mmol/L KCl), MnCl2 1 mmol/L, dNTPs (Cat 10297-018; Gibco BRL, Grand Island, NY, USA) 200 ␮mol/L each, 40 units of RNasin (N2111; Promega, Madison, WI, USA), 5 units of rTth and 0.75 ␮mol/L 3⬘-claudin primer. Reverse transcription was carried out in a final volume of 20 ␮L at 70⬚C for 15 minutes, and the reaction was stopped by placing the tubes on ice. Two consecutive rounds of PCR were performed. For the first PCR, 80 ␮L of PCR master mix for each sample were prepared by adding 65 ␮L H2O, 8 ␮L 10⫻ chelating buffer [50% (vol/vol) glycerol, 100 mmol/L Tris-HCl pH 8.3, 1 mol/L KCl, 0.5% (wt/vol) Tween 20, 7.5 mmol/L EGTA], 1.5 mmol/L MgCl2 and 0.15 ␮mol/L upstream 5⬘ claudin-1–specific primer or the 5⬘ claudins degenerated primer (RAGGGSCTSTGGATGDMST GYGTG). The following amplification protocol was performed: 3 minutes at 94⬚C followed by 10 cycles at 94⬚C for one minute, 43⬚C for one minute, 60⬚C for one minute, and subsequently 30 cycles at 94⬚C for one minute, 48⬚C for one minute, and 60⬚C for one minute. Five microliters of the first PCR served as a template for the second PCR that was carried out in a final volume of 25 ␮L and contained 10⫻ PCR buffer (KCl 500 mmol/L, MgCl2 15 mmol/L, Tris-HCl 100 mmol/L pH 9.0; Amersham Pharmacia Biotech), dNTPs 200 ␮mol/L (Gibco BRL), 2 units Taq DNA Polymerase (Cat. 27-0799-01; Amersham Pharmacia Biotech), and 50 pmol of each claudin-specific primer (Table 1). The following amplifi-

479

Reyes et al: Renal segmental distribution of claudins Table 1. Nucleotide sequences employed for PCR amplification of claudins-1 to -8 Claudin 1

2 3 4 5 6 7 8

Sequence

Nucleotide

Accession number

5⬘-AGCCAGGAGCCTCGCCCCGCAGCTGCA-3⬘ 5⬘-CAGCCAAGGCCTGCATAGCCATTG-3⬘ 5⬘-ATCGCAATCTTTGTGTCCACCATT-3⬘ 5⬘-ATTCTGTTTCCATACCATGCTGTG-3⬘ 5⬘-ATGTCCTCGCTGGCTTGTATTATCTCT-3⬘ 5⬘-GCCATGAAGATTCCAAGCAACTG-3⬘ 5⬘-CATCCTGCTGGCCGCCTTCG-3⬘ 5⬘-CCTGATGATGGTGTTGGCCGAC-3⬘ 5⬘-TCGTGGGTGCTCTGGGGATGCTT-3⬘ 5⬘-GCGGATGACGTTGTGAGCGGTC-3⬘ 5⬘-GCTGGCGCTGGTGGCACTCTTTGT-3⬘ 5⬘-GGCGAACCAGCAGAGCGGCAC-3⬘ 5⬘-ACCCTCCTCATTGTCCTGCTTGGC-3⬘ 5⬘-CCAGCAGACAGGAATGAGCGT-3⬘ 5⬘-CCTGGTGTTGGGCTTCTTAG-3⬘ 5⬘-TGACCAATCCAGGAACATGCTACCAA-3⬘ 5⬘-CGTCTTGGCTTTCTTGGCTTTCATG-3⬘ 5⬘-GGCAACCCAGCTGACAGG-3⬘

1–27 931–952 432–456 562–586 315–342 450–473 268–276 411–432 436–459 583–605 267–290 400–420 259–282 394–414 261–280 399–425 264–288 405–422

AF 072127

AF 072128 AF 087821 AF 087822 AF 087823 AF 087824 AF 087825 AF 087826

Table 2. Comparison of functional and biophysical parameters, in proximal and collecting tubules, that change with development Age Segment

Animal

Parameter

Neonatal

Adult

Reference

Proximal

Guinea pig

Net pressure for reabsorption mm Hg Urinary recovery of mannitol % Estimated pore radii of TJ ⫻10⫺10 m Hydrostatic hydraulic conductance ⫻10⫺3 mL · cm⫺2 · min⫺1 · cm H2O⫺1 Isotonic net fluid transport nL · mm⫺1 · min⫺1 Permeability of colloidal lanthanum and horseradish peroxidase Transtubular electrical resistance X · cm2 Hydraulic conductivity coefficient (Lp) 10⫺7 cm · S⫺1 atm⫺1

15.0 92 Higher by ⬃0.5 0.0367 ⫾ 0.0048

30.9 100

[53]

0.0052 ⫾ 0.002

[54]

Rabbit

Rat Collecting duct

Rabbit

cation protocol was performed: three minutes at 94⬚C followed by 45 cycles at 94⬚C for 30 seconds, 60⬚C for 30 seconds, 72⬚C for one minute; except for claudin-7, which required an annealing temperature of 50⬚C. Samples with no template served as negative controls before cDNA preparation, first PCR and second PCR. Reaction products were detected by agarose gel electrophoresis and ethidium bromide staining. RESULTS Occludin displays a continuous cell border pattern in rabbit collecting ducts Our previous work studied the distribution of the TJ proteins ZO-1, ZO-2 and occludin along different segments of the mammalian nephron [27]. The three proteins increase their expression from the proximal to the collecting segment, following the augmentation in both TER and TJ complexity. However, of the three proteins studied, occludin revealed the most pronounced change. Therefore, we decided to explore whether the expression pattern of this protein changed with maturation of the

0.26 Permeable 40 85 ⫾ 34

1.06 Non-permeable 200 10 ⫾ 7

[55] [56] [57] [58]

kidney. In several mammals, such as rat, rabbit and human, diverse functional parameters of the kidney depict postnatal maturation. Table 2 shows that certain physiological characteristics that reflect the permeability of the paracellular pathway, such as TER, mannitol fluxes, hydrostatic hydraulic conductance, permeability to paracellular tracers and isotonic net fluid transport, are different in newborn tubules than in the corresponding segments of the adult. In collecting ducts isolated from newborn rabbits, occludin formed continuous bands that surround the cellular borders (Fig. 1). It is noteworthy that this pattern was identical to the one we previously observed in adult tubules [27]. Nevertheless, other characteristics of occludin, such as its phosphorylation state or its response to external factors such as hormonal stimulation, were not explored and cannot be disregarded from influencing the changes during maturation. Immunofluorescence detection of claudin-1 reveals a junctional distribution in tight segments In recent years, other transmembrane TJ proteins named claudins have been identified [16, 36]. Therefore,

480

Reyes et al: Renal segmental distribution of claudins

Fig. 1. Occludin distributes along the cellular borders of a newborn collecting duct. The collecting duct was isolated from a 4-day-old rabbit. Bar length is equivalent to 20 ␮m in this and subsequent illustrations.

Fig. 3. Renal tubules isolated from newborn rabbits express claudin-1 in the same fashion as adults. (a, b) The tubule contains proximal and Henle’s loop segments. The latter can be distinguished by its width and cytokeratin-8 staining, which is absent from the proximal segment (b, arrow) and conspicuous in the Henle’s portion (b, arrowheads). While claudin-1 clearly labels the Henle’s segment (a, arrowheads), the signal at the proximal tubule is negligible (a, arrow). (c and d) The distal portion is positive for both cytokeratin-8 (d) and claudin-1 (c). (e) A clear chicken fence pattern is detected in the collecting tubule.

Fig. 2. A cell border pattern of claudin-1 is only found in the tight segments of the adult mammalian nephron. While claudin-1 is not detectable in a proximal tubule of an adult kidney (a), a diffuse expression is found at Henle’s loop (c). A clear claudin-1 signal is observed at the distal (d) and collector (e) tubules along the cellular borders (d). Distal and proximal tubules were further distinguished from each other by the positive cytokeratin-8 staining of the former ( f ) compared to the negative image observed in the latter (b).

we proceeded to analyze the tubular distribution of claudins-1 and -2 along the different nephron segments of adult and newborn rabbits. In adults, claudin-1 was not

detectable in proximal tubules (Fig. 2a), while a diffuse expression was found at Henle’s loop (Fig. 2c) and the distal segment (Fig. 2e). At the collecting tubule (Fig. 2d), clear and strong claudin-1 staining was detected at the cell boundaries. In addition to morphological identification, distal and proximal tubules were further distinguished from each other by the positive cytokeratin-8 staining of the former (Fig. 2f) and the negative image observed in the latter (Fig. 2b). Renal tubules isolated from newborn rabbits expressed claudin-1 in the same fashion as adults. Thus, Figure 3 shows that while claudin-1 labeled the loop of Henle’s segment (Fig. 3a, arrowheads), the signal at the proximal tubule was negligible (Fig. 3a, arrow). The distal portion (Figs. c, and d) was positive for both cytokeratin 8 (Fig. 3d) and claudin-1 (Fig. 3c). As observed in

Fig. 4. Claudin-1 distribution in principal and intercalated cells of the collecting duct shows a homogeneous pattern. (a) Claudin-1, homogeneously labels the cellular borders of a collecting segment (green). (b) Monoclonal antibody (mAb) 503 selectively stains the intercalated cells (red) of this segment. (c) A merged image shows the staining obtained with antibodies against claudin-1 (green) and intercalated cells (red). The arrows point at two intercalated cells (Reproduction of this figure in color was made possible by Bio-Rad Laboratories Mexico, S.A. de C.V.).

Fig. 9. In frozen kidney sections, claudin-5 is found at the cellular borders of vessels. (A) Claudin-5 stains the endothelia of the glomerular tuft and the vascular pole (arrow). (B) In a frozen section treated with Evan’s blue, the autofluorescence of the renal vasculature gives a strong red signal that clearly distinguishes endothelia (arrow) from the surrounding epithelial tubules. While in the latter no claudin-5 signal is detected, in the endothelia it is clearly detectable at the cellular borders (arrowheads) (Reproduction of this figure in color was made possible by Bio-Rad Laboratories Mexico, S.A. de C.V.).

482

Reyes et al: Renal segmental distribution of claudins

Fig. 5. Claudin-2 is found significantly only at the cell borders of leaky renal adult epithelia. Claudin-2 is shown by the clear cell border pattern at the proximal segment (a) and Henle’s loop (b). At a distal tubule (c), positively identified by its cytokeratin-8 staining (d), a strong claudin-2 signal is found at the cytoplasm together with a faint cell border staining. At the collecting tubule no defined pattern of claudin-2 staining is visible (e) while cytokeratin-8 is easily detectable ( f ).

the adult tissue, a clear “chicken fence” pattern was only detectable at the collecting tubule (Fig. 3e). Since collecting tubules have principal and intercalated cells, further experiments studied whether or not the latter depict the same pattern of claudin-1 fluorescence as principal cells. Figure 4 shows a collecting duct stained with a specific antibody directed against the luminal surface of intercalated cells (mAb 503) [37], where claudin-1 labeling was found around intercalated (red) and principal cells in a homogeneous pattern. The proportion of principal and intercalated cells in the collecting tubule is of 64 and 36%, respectively [38], however, the latter cells appear during postnatal development of the kidney. Since in newborn rabbit collecting ducts all cells resemble principal ones and intercalated cells have not been observed [39], we did not attempt to study the differential distribution of claudin-1 in newborn collecting ducts with the mAb 503. Claudin-2 is detected by immunofluorescence at the leaky cell borders of renal segments In contrast to the pattern depicted by claudin-1, claudin-2 was found only at the cell borders of leaky renal epithelia in adult and neonatal tubules. Figure 5 shows the clear cell border pattern of claudin-2 at the proximal adult segment (Fig. 5a) and Henle’s loop (Fig. 5b). At

Fig. 6. Claudin-2 in newborn rabbits is related to leaky epithelia. While the proximal segment displays a clear cell border pattern (a), only a faint signal is found at cell boundaries of Henle’s loop (b). (c) Identified by its positive cytokeratin-8 staining in the distal segment (d) claudin-2 gives a strong but diffuse labeling. In the collecting tubule (e) claudin-2 is barely detectable.

the distal tubule (Fig. 5c), identified by cytokeratin-8 staining (Fig. 5d), claudin-2 was detectable at the cytoplasm and had an almost negligible cell border pattern. At the collecting tubule only a diffuse claudin-2 staining was found (Fig. 5e), while cytokeratin-8 clearly delineated the cell border (Fig. 5f). In accordance, claudin-2 in newly born rabbits could be seen in a clear cell border pattern at the proximal segment (Fig. 6a). A weak signal was found at the cellular boundaries of Henle’s loop (Fig. 6b), while the distal segment (Fig. 6c), identified by cytokeratin-8 labeling (Fig. 6d), had a strong but diffuse pattern. In the collecting tubule (Fig. 6e), the claudin signal was almost indistinguishable. Western blot shows the differential presence of claudin-1 and -2 in the insoluble fractions of tight and leaky renal tubules Since our immunofluorescence observations mainly reflected the distribution pattern of claudin-1 and -2 and not their precise quantity, Western blot analyses of claudins were performed in tubule populations isolated by Percoll gradients. This procedure was chosen because microdissection of well-defined nephron segments provided only limited amounts of cell material. As previously reported [27], tubular isolation by Percoll gradi-

Reyes et al: Renal segmental distribution of claudins

483

Fig. 7. Western blot determination of claudin-1 and -2 in renal tubules isolated by a Percoll gradient. A representative example of a Western blot determination of claudin-1 (A) and claudin-2 (B) is shown above, while the quantitation of three independent experiments is shown below. In each enhanced chemiluminescent plate, the densitometric value obtained for the distal soluble fraction or the proximal soluble fraction of the newborn was taken as 1 to normalize the claudin-1 and claudin-2 experiments, respectively. The standard error is shown. In newly born and adult rabbits, the amount of claudin-1 present in the insoluble (I), junctional associated fraction, is higher in distal (D) than in proximal (P) tubules (newborn P ⬍ 0.005; adults P ⬍ 0.005). Claudin-1 is detected in the soluble fraction (S) of both proximal and distal segments. The amount of claudin-2 present in the insoluble proximal fraction is significantly higher than in the distal insoluble fraction in both newly born and adults (newly born P ⬍ 0.001; adults P ⬍ 0.005). Claudin-2 is present in similar amounts in the insoluble (䊏) and soluble (䊐) fractions of adult tubules and proximal segments of newborn animals.

ent provided pure preparations of proximal and distal segments, but not of Henle’s and collecting ducts. Therefore, the Western blot analyses were concentrated on the proximal and distal portions. Since our immunofluorescence observations had shown cell border localization of claudins together with a conspicuous and diffuse cytoplasmic staining, we proceeded to analyze Triton X-100 soluble and insoluble fractions, taking into account that proteins associated with the TJ were present in the insoluble fraction [40, 41]. The Western blot in Figure 7A displays a significantly higher amount of claudin-1 in distal (D) than in proximal (P) tubules derived from both adult and five-day-old rabbits in the insoluble fraction (I), confirming our immunofluorescence observations. Claudin-1 was present in the soluble fraction of distal and proximal tubules derived from both newborn and adult tissue. The amount of claudin-1 detected in soluble and insoluble fractions among each tubular segment was similar, with the exception of the proximal adult segment where almost no claudin-1 was detectable in the insoluble fraction. On the contrary, claudin-2 was clearly present in proximal tubules, both in the insoluble and soluble fractions, while in the distal segments only a barely detectable signal was found in the insoluble fractions (Fig. 7B).

A differential expression of claudins-1 to -8 mRNA found along the distinct renal segments and during postnatal development As previously stated, claudins comprise a family of up to 20 members. Initially our immunofluorescence and Western blot studies were concentrated on the distribution of claudin-1 and -2, since only antibodies against them were commercially available. However, as no change in claudin-1 and -2 expression was found that correlated with postnatal maturation of TJ permeability, we decided to explore the mRNA expression of claudins-1 to -8 in isolated renal segments derived from newborn and adult animals by using RT-PCR. When conventional PCR methods were employed claudins expression proved to be barely detectable (data not shown). Therefore, we optimized a nested PCR protocol for the claudin gene family that involved two consecutive rounds of PCR amplification. In the first round, a degenerated primer set designed from conserved sequences present in claudins-1 to -8 was used. This procedure increased the target concentration such that the second claudin PCR reaction could be performed with specific claudin primers to assay for the presence or absence of claudin-2 to -8 mRNAs within the tissue. As claudin-1 remained undetectable with this procedure, two sets of specific primers were used

484

Reyes et al: Renal segmental distribution of claudins

Fig. 8. A differential expression of claudins-1 to -8 mRNA is found along proximal, Henle’s and collector segments. (A) The RT-PCR shows a clear expression of claudins-1 to -8 in total kidney. (B) In the RT-PCR performed with isolated proximal tubules only claudins-1, -2, -3, -4, -7 and -8 are detected. In Henle’s preparation only claudins-1, -2, -4 and -8 are present, while with isolated collector tubules claudins-1, -2, -3 and -4 are clearly expressed and claudin-5 is barely detectable. (C) In the newborn proximal and collecting tubules only claudins-1 to -4 are conspicuously amplified.

for the nested amplification of claudin-1 (Table 1) To perform these experiments renal tubules were dissected and their identity determined by microscopic examination. As previously stated, proximal, Henle’s loop, and collecting duct tubules can be identified microscopically, while distal segments require the use of a fluorescent marker such as cytokeratin-8. In order to guarantee the purity of the RNA preparation, putative distal segments were excluded and around 50 proximal, Henle loops and collecting tubules were collected for RNA isolation. In newborn kidney the length of Henle’s segments that could be unequivocally identified was very small, and thus PCR amplification of claudins in this segment was not attempted because the amount of material was insufficient. Figure 8A shows that claudin-1 to -8 expression was clearly detectable in total kidney by the nested PCR procedure, confirming the efficiency of the primers employed. Figure 8B shows, to our knowledge for the first time, the differential pattern of claudin expression along the renal segments. While claudins-1, -2 and -4 were expressed in all of the studied segments from adult animals, specificity was given by the expression of claudin-3 at the proximal and collector segments, claudin-7 in the proximal tubules, and claudin-8 in proximal and Henle’s preparations. Claudin-6 was detectable in total kidney and absent in the isolated tubular RNA preparations. When the PCR experiments were performed with RNA obtained from newborn rabbits, claudins-5, -7 and -8 were not detectable, while claudins-1 to -4 were expressed in both proximal and collecting segments. Clau-

din-6, the mRNA of which has only been detected in embryonic tissue [42], was faintly amplified in newborn collector tubules in certain samples (data not shown). Claudin-5 was clearly detectable from RNA obtained from total kidney; however, it was absent in proximal and Henle’s segments. This is consistent with the observation that claudin-5 was mainly restricted to endothelial cells [43], which were absent in the isolated tubule preparation and present in the total kidney samples. To further explore this issue, whole kidney frozen sections were used to analyze the distribution of claudin-5. Figure 9a shows a conspicuous claudin-5 signal at the glomerular tuft and the vascular pole (arrow). No claudin-5 specific fluorescence could be detected at the cellular borders of tubules, while a clear signal was observed in the arterial endothelium (9b, arrowheads). Kidney vasculature could be unequivocally distinguished from the tubules due to the strong autofluorescence of complex molecules such as collagen, elastin, cholesterol and fluorescent lipids, present in the vessel walls [44], which after treatment with Evan’s blue showed a conspicuous red signal (9b, arrow). DISCUSSION Paracellular transport among different epithelia varies widely, giving rise to their characterization as tight or leaky. In their seminal work, Claude and Goodenough studied the fracture faces of TJ in tight and leaky epithelia, and found that the proximal tubules showed numerous intramembranous particles (IMP) scattered within the grooves of the E-face and short discontinuous strands in the P-face [26]. On the other hand, distal tubules presented continuous grooves mostly devoid of IMP in the E-face. The observation of the different structural appearance of TJ strands and grooves in tight and leaky epithelia prompted us to speculate whether it could be due to the different molecular composition of the strands. To study this issue, we analyzed the distribution of different TJ molecules in isolated renal tubules derived from newborn and adult animals. Proximal, distal and collecting tubules were isolated from the cortical region, while Henle’s loops were dissected from medullary sections. The difference between cortical and medullary segments was not explored, however, because medullary collecting tubules are exposed to steeper osmotic gradients than cortical segments [45], and so differences in claudin composition between cortical and papillary collecting tubules could be expected. Our previous study demonstrated that the amount of occludin and its cell border distribution increased along the mammalian nephron in accordance with TER values and TJ complexity [46]. Our current study demonstrates, to our knowledge for the first time, that the distribution of occludin along the cellular borders of tubular cells is already established in newborn animals. While there is

Reyes et al: Renal segmental distribution of claudins

485

Table 3. Classification of claudins in accordance with the tightness or developmental stage of the tissue in which the proteins are expressed Claudin 1 2 3 4 5 6 7 8

Tubular distribution Distal and collecting

Classification

Complementary supporting information

Tight

Transfected fibroblasts show conspicuous tight junction strands, with filaments associated to the P face [15] Proximal Leaky Transfected into fibroblasts forms TJ strands with discontinuous particles associated to the E face [15]; introduction of claudin-2 into MDCK I cells converts them from tight to leaky [24] Proximal, Henle and Promiscuous Homogeneously present in both crypts (low resistance) and villi (high resistance) of collecting intestinal epithelium [48] Proximal and collecting Promiscuous Not available None Endothelial leaky Constitutes TJ strands in endothelial cells [43] Newborn collecting Embryonic Absent in adult tissues and identified from an EST sequence derived from embryos [42] Proximal Leaky Not available Proximal and Henle Leaky Not available

We classified those claudins that were found both in tight and leaky epithelia as promiscuous. A summary of other reports that support our findings is included as complementary information.

only one gene for occludin, claudins comprise a multigene family consisting of over 20 members [16], and several organs, including kidney, express a wide variety of them [42]. Therefore, it was tempting to speculate that the occurrence of several claudins species in an organ such as kidney could correlate to the distinct paracellular permeability, electrical resistance and TJ pattern found at the different tubular segments and during postnatal maturation. We studied isolated renal segments from the viewpoint of claudins. The immunofluorescence appearance of claudin-1 in distal and collecting tubules, and of claudin-2 in proximal segments, strongly suggests the relationship of claudin-1 and -2 to tight and leaky epithelia, respectively. This observation was confirmed in the Western blot determination performed in tubules isolated in a Percoll gradient. The immunofluorescence differential distribution of claudins-1 and-2 was observed both in adult and newborn renal tissues. The Western blot analyses of samples separated in soluble and insoluble fractions revealed a more defined pattern of expression in adult than in newborn tissues. Thus, in the mature animal, claudin-1 becomes barely detectable in the insoluble fraction (associated to TJ) of proximal tubules and claudin-2 becomes scarcely noticeable even in the soluble fraction of adult distal segments. This observation might be related to the postnatal maturation of tubular function observed for several parameters (Table 1). The presence of both claudin-1 and -2 in the insoluble fraction, together with the observation of a diffuse cytosolic claudin immunostaining, suggested alternative roles for claudins that might be fulfilled in cell domains different from the TJ region. In this respect the analysis of claudin-1 expression in the epididymis revealed an appearance of this protein along the entire basolateral plasma membrane, suggesting a role for claudin-1 as an adhesion molecule [47]. A recent study has reported that claudin-3, -4 and -5 distribution is not restricted to the TJ microdomain [48]. The type of claudin that constitutes a TJ strand ap-

pears to be determinant to its appearance. For instance, in CNS myelin and Sertoli cells where the TJs are almost exclusively composed of claudin-11, the pattern of filaments is alike and instead of a mesh, are sets of parallel strands [49]. In fibroblasts transfected with claudin-2 [5] or claudin-5 [43], the appearance of the strands resembles the TJs present in non-neural endothelial cells [50, 51] and in leaky epithelia such as the proximal renal segment [26]. In contrast, the pattern of TJ strands found in fibroblasts transfected with claudin-1 [15] or claudin-3 [6] is similar to that of tight epithelia such as the distal and collecting segments of the nephron [26]. These results together with the recent observation that transfection of claudin-2 to MDCK I cells converts zonula occludens from tight into leaky strands [24] support the proposal that claudin-2 and claudin-1 are related to leaky and tight epithelia, respectively (Table 3). Distinct species of claudins have been shown to interact within and between TJ strands, except in some combinations [6]. This heterogeneous assembly might be responsible of the diversity of the structure and function of TJ strands. In the case of whole kidney extracts, the expression of mRNA for claudins-1, -2, -3, -4, -5, -7 and -8 has been revealed by Northern blot [5, 42]. To overcome the lack of commercial antibodies against claudins-3 to -8, we assessed the presence of their mRNA by RT-PCR performed in precisely identified isolated renal tubules. Additionally, to compare mRNA and protein expression RT-PCR amplifications of claudins-1 and -2 were performed. The mRNA for claudins-1, -2, -3 and -4 appears early in development and is conspicuously detected in newborn tissues. In fibroblasts transfected with claudin-3, the appearance of TJ strands and grooves resembles that of claudin-1 transfected cells [6], and its noticeable presence in proximal and collecting tubules of newborn and adult animals suggests a housekeeping role in strand framing. In agreement with this suggestion, this claudin has been found uniformly expressed in other organs such as liver, pancreas, gut and stomach

486

Reyes et al: Renal segmental distribution of claudins

[48]. Claudin-4 is specifically expressed in certain epithelia, such as kidney and lung, and is absent in others, such as liver, testis and spleen [48]. With regards to the expression of claudin-5, controversial results have been obtained: some authors describe it as an endothelial claudin [43], while others have observed it in certain epithelial cells such as enterocytes, stomach and pancreatic acinar cells [48]. Here we were able to amplify it by RT-PCR in whole kidney samples, where vessels are present, and yet claudin-5 remained undetectable in isolated newborn and adult tubules. These results, together with our immunofluorescence observations of frozen sections, indicate that in kidney claudin-5 is restricted to the vasculature. Claudin-6 was detected in whole kidney samples but not in isolated adult tubules. Since this claudin was not detected in several adult tissues examined by others and its EST sequences were identified from embryos [42], it is possible that the clear expression observed in our whole kidney experiments is due to stem cells, known to still be present in the adult organ [52]. The faint appearance of claudin-6 in certain newborn tubules (data not shown) and its disappearance from adult segments further suggest a developmentally regulated expression of this claudin. Northern blot analysis of claudins-7 and -8 suggest that these proteins are restricted to certain epithelia [42]. In isolated renal tubules these claudins could only be amplified from the leaky portions of the adult nephron (Table 3), and therefore, it is tempting to include them in the group of leaky claudins with claudins-2 and -5. These claudins could form aqueous pores with high conductance within TJ strands and thus their up-regulation would increase paracellular permeability. However, further work is needed to confirm this point. The presence of a specific claudin mRNA does not necessarily imply the expression of the protein, nor its location at the TJ region. For example, the differential expression of claudin-1 and -2 in tight and leaky renal segments found here with the immunofluorescence and Western blot approaches is not reflected in the mRNA results. However, these results open the possibility that claudin protein expression might be subjected to posttranscriptional control, a research area that remains unexplored. In summary, our results suggest that the tightness of an epithelium might not be determined simply by its total amount of claudins. Instead, the critical factor might be the combination and mixing ratios of the different claudin species within the individual TJs. The occurrence of a large number of claudins in the kidney might allow it to express the appropriate TJ junction permeability at each segment and developmental stage, which suggests that the up- or down-regulation of specific claudins in

response to physiological or pathological conditions might drastically change the tightness of the epithelium. ACKNOWLEDGMENTS This work was supported by grants 28083N, 37846-N, and G34511 from CONACYT. The authors thank Dr. Marcelino Cereijido from CINVESTAV, Me´xico, for his helpful comments during the preparation of this manuscript. Reproduction of Figures 4 and 9 in color was made possible by Bio-Rad Laboratories Mexico, S.A. de C.V. Reprint requests to Lorenza Gonza´lez Mariscal, Ph.D., Center for Research and Advanced Studies (CINVESTAV), Department of Physiology, Biophysics and Neurosciences, Ave. Politecnico Nacional 2508, Me´xico, D.F. 07000, Mexico. E-mail: [email protected]

REFERENCES 1. Le Grimellec C, Friedlander G, El Yandouzi EH, et al: Membrane fluidity and transport properties in epithelia. Kidney Int 42: 825–836, 1992 2. Cereijido M, Gonzalez-Mariscal L, Contreras RG, et al: The making of a tight junction. J Cell Sci (Suppl 17):127–132, 1993 3. Chalcroft JP, Bullivant S: An interpretation of liver cell membrane and junction structure based on observation of freeze-fracture replicas of both sides of the fracture. J Cell Biol 47:49–60, 1970 4. Fujimoto K: SDS-digested freeze-fracture replica labeling electron microscopy to study the two-dimensional distribution of integral membrane proteins and phospholipids in biomembranes: Practical procedure, interpretation and application. Histochem Cell Biol 107:87–96, 1997 5. Furuse M, Fujita K, Hiiragi T, et al: Claudin-1 and -2: Novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin. J Cell Biol 141:1539–1550, 1998 6. Furuse M, Sasaki H, Tsukita S: Manner of interaction of heterogeneous claudin species within and between tight junction strands. J Cell Biol 147:891–903, 1999 7. Furuse M, Fujimoto K, Sato N, et al: Overexpression of occludin, a tight junction-associated integral membrane protein, induces the formation of intracellular multilamellar bodies bearing tight junction-like structures. J Cell Sci 109(Pt 2):429–435, 1996 8. McCarthy KM, Skare IB, Stankewich MC, et al: Occludin is a functional component of the tight junction. J Cell Sci 109(Pt 9): 2287–2298, 1996 9. Balda MS, Whitney JA, Flores C, et al: Functional dissociation of paracellular permeability and transepithelial electrical resistance and disruption of the apical-basolateral intramembrane diffusion barrier by expression of a mutant tight junction membrane protein. J Cell Biol 134:1031–1049, 1996 10. Chen Y, Merzdorf C, Paul DL, et al: COOH terminus of occludin is required for tight junction barrier function in early Xenopus embryos. J Cell Biol 138:891–899, 1997 11. Wong V, Gumbiner BM: A synthetic peptide corresponding to the extracellular domain of occludin perturbs the tight junction permeability barrier. J Cell Biol 136:399–409, 1997 12. Moroi S, Saitou M, Fujimoto K, et al: Occludin is concentrated at tight junctions of mouse/rat but not human/guinea pig Sertoli cells in testes. Am J Physiol 274:C1708–C1717, 1998 13. Hirase T, Staddon JM, Saitou M, et al: Occludin as a possible determinant of tight junction permeability in endothelial cells. J Cell Sci 110(Pt 14):1603–1613, 1997 14. Saitou M, Furuse M, Sasaki H, et al: Complex phenotype of mice lacking occludin, a component of tight junction strands. Mol Biol Cell 11:4131–4142, 2000 15. Furuse M, Sasaki H, Fujimoto K, et al: A single gene product, claudin-1 or -2, reconstitutes tight junction strands and recruits occludin in fibroblasts. J Cell Biol 143:391–401, 1998 16. Tsukita S, Furuse M: Occludin and claudins in tight-junction

Reyes et al: Renal segmental distribution of claudins

17.

18.

19.

20.

21.

22.

23. 24.

25.

26. 27.

28. 29.

30. 31.

32.

33. 34.

35.

36.

37.

strands: leading or supporting players? Trends Cell Biol 9:268–273, 1999 Briehl MM, Miesfeld RL: Isolation and characterization of transcripts induced by androgen withdrawal and apoptotic cell death in the rat ventral prostate. Mol Endocrinol 5:1381–1388, 1991 Peacock RE, Keen TJ, Inglehearn CF: Analysis of a human gene homologous to rat ventral prostate.1 protein. Genomics 46:443– 449, 1997 Katahira J, Inoue N, Horiguchi Y, et al: Molecular cloning and functional characterization of the receptor for Clostridium perfringens enterotoxin. J Cell Biol 136:1239–1247, 1997 Sirotkin H, Morrow B, Saint-Jore B, et al: Identification, characterization, and precise mapping of a human gene encoding a novel membrane-spanning protein from the 22q11 region deleted in velocardio-facial syndrome. Genomics 42:245–251, 1997 Bronstein JM, Popper P, Micevych PE, et al: Isolation and characterization of a novel oligodendrocyte-specific protein. Neurology 47:772–778, 1996 Simon DB, Lu Y, Choate KA, et al: Paracellin-1, a renal tight junction protein required for paracellular Mg2⫹ resorption. Science 285:103–106, 1999 Wong V, Goodenough DA: Paracellular channels! Science 285: 62, 1999 Furuse M, Furuse K, Sasaki H, et al: Conversion of zonulae occludents from tight to leaky strand type by introducing claudin-2 into Madin-Darby canine kidney I cells. J Cell Biol 153:263–272, 2001 Inai T, Kobayashi J, Shibata Y: Claudin-1 contributes to the epithelial barrier function in MDCK cells. Eur J Cell Biol 78:849– 855, 1999 Claude P, Goodenough DA: Fracture faces of zonulae occludents from “tight” and “leaky” epithelia. J Cell Biol 58:390–400, 1973 Gonza´lez-Mariscal L, Namorado MC, Martin D, et al: Tight junction proteins ZO-1, ZO-2, and occludin along isolated renal tubules. Kidney Int 57:2386–2402, 2000 Burg MB, Green N: Function of the thick ascending limb of Henle’s loop. Am J Physiol 224:659–668, 1973 Greger R: Cation selectivity of the isolated perfused cortical thick ascending limb of Henle’s loop of rabbit kidney. Pflu¨gers Arch 390:30–37, 1981 Malnic G, Giebisch G: Some electrical properties of distal tubular epithelium in the rat. Am J Physiol 223:797–808, 1972 Helman SI, Grantham JJ, Burg MB: Effect of vasopressin on electrical resistance of renal cortical collecting tubules. Am J Physiol 220:1825–1832, 1971 Piepenhagen PA, Nelson WJ: Differential expression of cell-cell and cell-substratum adhesion proteins along the kidney nephron. Am J Physiol 269:C1433–C1449, 1995 Vinay P, Gougoux A, Lemieux G: Isolation of a pure suspension of rat proximal tubules. Am J Physiol 241:F403–F411, 1981 Bradford MM: A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proteindye binding. Anal Biochem 72:248–254, 1976 Towbin H, Staehelin T, Gordon J: Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76:4350–4354, 1979 Kubota K, Furuse M, Sasaki H, et al: Ca(2⫹)-independent celladhesion activity of claudins, a family of integral membrane proteins localized at tight junctions. Curr Biol 9:1035–1038, 1999 Tauc M, Koechlin N, Jamous M, et al: Principal and intercalated

38. 39. 40.

41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

52. 53. 54. 55. 56. 57. 58.

487

cells in primary cultures of rabbit renal collecting tubules revealed by monoclonal antibodies. Biol Cell 76:55–65, 1992 Emmons CL, Matsuzaki K, Stokes JB, et al: Axial heterogeneity of rabbit cortical collecting duct. Am J Physiol 260:F498–F505, 1991 Minuth WW: Neonatal rabbit kidney cortex in culture as tool for the study of collecting duct formation and nephron differentiation. Differentiation 36:12–22, 1987 Fanning AS, Jameson BJ, Jesaitis LA, et al: The tight junction protein ZO-1 establishes a link between the transmembrane protein occludin and the actin cytoskeleton. J Biol Chem 273:29745– 29753, 1998 Itoh M, Nagafuchi A, Moroi S, et al: Involvement of ZO-1 in cadherin-based cell adhesion through its direct binding to alpha catenin and actin filaments. J Cell Biol 138:181–192, 1997 Morita K, Furuse M, Fujimoto K, et al: Claudin multigene family encoding four-transmembrane domain protein components of tight junction strands. Proc Natl Acad Sci USA 96:511–516, 1999 Morita K, Sasaki H, Furuse M, et al: Endothelial claudin: Claudin-5/TMVCF constitutes tight junction strands in endothelial cells. J Cell Biol 147:185–194, 1999 Kingsley K, Carroll K, Huff JL, et al: Photobleaching of arterial autofluorescence for immunofluorescence applications. Biotechniques 30:794–797, 2001 Jamison RL, Lacy FB: Evidence for urinary dilution by the collecting tubule. Am J Physiol 223:898–902, 1972 Gonza´lez-Mariscal L, Namorado MC, Martin D, et al: Tight junction proteins ZO-1, ZO-2, and occludin along isolated renal tubules. Kidney Int 57:2386–2402, 2000 Gregory M, Dufresne J, Hermo L, et al: Claudin-1 is not restricted to tight junctions in the rat epididymis. Endocrinology 142:854–863, 2001 Rahner C, Mitic LL, Anderson JM: Heterogeneity in expression and subcellular localization of claudins 2, 3, 4, and 5 in the rat liver, pancreas, and gut. Gastroenterol 120:411–422, 2001 Gow A, Southwood CM, Li JS, et al: CNS myelin and sertoli cell tight junction strands are absent in Osp/claudin-11 null mice. Cell 99:649–659, 1999 Dempsey GP, Bullivant S, Watkins WB: Endothelial cell membranes: Polarity of particles as seen by freeze-fracturing. Science 179:190–192, 1973 Wolburg H, Neuhaus J, Kniesel U, et al: Modulation of tight junction structure in blood-brain barrier endothelial cells. Effects of tissue culture, second messengers and cocultured astrocytes. J Cell Sci 107(Pt 5):1347–1357, 1994 Slack JM: Stem cells in epithelial tissues. Science 287:1431–1433, 2000 Kaskel FJ, Kumar AM, Lockhart EA, et al: Factors affecting proximal tubular reabsorption during development. Am J Physiol 252:F188–F197, 1987 Horster M, Larsson L: Mechanisms of fluid absorption during proximal tubule development. Kidney Int 10:348–363, 1976 Larsson L, Horster M: Ultrastructure and net fluid transport in isolated perfused developing proximal tubules. J Ultrastruct Res 54:276–285, 1976 Larsson L: Ultrastructure and permeability of intercellular contacts of developing proximal tubule in rat kidney. J Ultrastruct Res 52:100–113, 1975 Reyes JL, Roch-Ramel F, Besseghir K: Net sodium and water movements in the newborn rabbit collecting tubule: Lack of modifications by indomethacin. Biol Neonate 51:212–216, 1987 Horster MF, Zink H: Functional differentiation of the medullary collecting tubule: Influence of vasopressin. Kidney Int 22:360–365, 1982