Differential Expression Patterns of Claudins, Tight Junction Membrane Proteins, in Mouse Nephron Segments

Antibodies

Rabbit anti-mouse polyclonal antibodies (pAb) to claudin-2, -3, -4, -5, -8, and -11 and rat anti-mouse occludin monoclonal antibody (mAb) were raised and characterized as described previously (11,14,18–20). Rabbit anti-rat aquaporin-2 (AQP2) pAb was kindly provided by Dr. Sei Sasaki (Tokyo Medical and Dental University, Tokyo, Japan). Rabbit anti-claudin-1 pAb, rabbit anti-Tamm-Horsfall glycoprotein (THP) pAb, and rabbit anti-chloride channel-K (ClC-K) pAb were purchased from Zymed Laboratories (South San Francisco, CA), Biomedical Technologies (Stoughton, MA), and Alomone Laboratories (Jerusalem, Israel), respectively. Rabbit anti-AQP1 pAb was purchased from Chemicon International (Temecula, CA).

For generation of anti-claudin-10 pAb in rabbits, a glutathione S-transferase (GST) fusion protein with the carboxy-terminal cytoplasmic domain of mouse claudin-10, which was produced in Escherichia coli (strain DH5a) and purified with glutathione-Sepharose 4B beads (Amersham-Pharmacia Biotech., Bucks, UK), was used as an antigen. For generation of anti-claudin-7, -12, -15, and -16 pAb, polypeptides corresponding to the respective carboxy-terminal cytoplasmic domains of mouse claudin-7, -12, and -15 and bovine claudin-16 (21) were synthesized and coupled, via cysteine, to keyhole limpet hemocyanin. These conjugated peptides were used as antigens in rabbits. The pAb were affinity-purified on nitrocellulose membranes with the corresponding GST fusion proteins. During the generation of pAb, animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Northern Blotting

Total RNA from mouse kidneys was isolated according to the method described by Chomczynski and Sacchi (22). Poly(A)⁺ RNA was obtained from total RNA by using oligo(dT)-cellulose beads (New England BioLabs, Beverly, MA). Aliquots of approximately 10 μg of poly(A)⁺ RNA were separated on 1% formaldehyde/agarose gels, transferred to positively charged nylon membranes, and crosslinked with ultraviolet light. Hybridization with digoxigenin (DIG)-labeled RNA probes was performed according to the protocol described by the manufacturer (Roche, Mannheim, Germany). RNA probes were DIG-labeled using a DIG RNA labeling kit (Roche).

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis and Immunoblotting

Lysates of E. coli expressing a maltose-binding protein fusion protein with the cytoplasmic portion of mouse claudin-10 or GST fusion proteins with the cytoplasmic portions of the other mouse claudin species were subjected to one-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (12.5%), according to the method described by Laemmli (23), and gels were stained with Coomassie Brilliant Blue R-250. For immunoblotting, proteins were electrophoretically transferred from gels to nitrocellulose membranes, which were then incubated with the first antibody. Bound antibodies were detected with biotinylated second antibodies and streptavidin-conjugated alkaline phosphatase (Amersham Corp., Arlington Heights, IL). Nitroblue tetrazolium and bromochloroindolyl phosphate were used as substrates for detection of alkaline phosphatase activity.

Immunofluorescence Microscopy

Kidneys were removed from mice and frozen with liquid nitrogen. For staining with anti-claudin-1 pAb, kidneys were fixed for at least 2 h with 4% paraformaldehyde in phosphate-buffered saline (PBS) (pH 7.2), immersed for 48 h in PBS containing 15% sucrose, and then frozen with liquid nitrogen. For TCA fixation (24), small pieces of kidney were soaked in ice-cold 10% TCA for 1 h, washed three times with PBS, and then frozen with liquid nitrogen. Pairs of serial frozen sections (approximately 10 μm thick) were cut with a cryostat, mounted on single glass slides, and air-dried. Sections were then fixed with 95% ethanol at 4°C for 30 min, followed by 100% acetone at room temperature for 1 min. After soaking in PBS containing 1% bovine serum albumin (and 0.2% Triton X-100 for claudin-1 staining only), one of the sections was incubated with one of the anti-claudin pAb (or preimmune serum) and the other was treated with one of the segment marker-specific pAb, in a moist chamber, for 30 min. For control experiments, anti-claudin pAb were preincubated with excess amounts of GST fusion proteins with the cytoplasmic domains of the respective claudins. Sections were then washed three times with PBS, followed by a 30-min incubation with secondary antibodies. Cy3-conjugated and FITC-conjugated secondary antibodies were used for the anti-claudin pAb-treated and anti-segment marker pAb-treated sections, respectively. After being washed with PBS, sections were embedded in 95% glycerol/PBS containing 0.1% p-phenylendiamine and 1% n-propyl gallate and were observed with a fluorescence microscope (Axiophot photomicroscope; Carl Zeis, Inc., Thornwood, NY) equipped with a cooled, charge-coupled device camera system.

Expression of Claudin Isotypes in the Kidney

We first examined the expression of claudin-1 to -16 in mouse kidneys by Northern blotting. When the poly(A)⁺ RNA isolated from the kidneys was probed with DIG-labeled mouse claudin-1 to -16 RNA, the expression of most claudin species was clearly detected, whereas claudin-6, -9, -13, and -14 were not detected (Figure 1A). Because claudin-6, -9, -13, and -14 mRNA were clearly detected with the same probes in other tissues, we concluded that the expression levels of claudin-6, -9, -13, and -14 were fairly low in mouse kidneys.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. (A) Northern blots of mouse claudin-1 to -16 (Cld-1 to -16) expression in the kidney. Claudin-6, -9, -13, and -14 were undetectable. Arrows, positions of individual claudin mRNA expected from Northern blotting analyses of other tissues, such as the liver. (B) Specificity of newly generated anti-claudin-10 and -16 polyclonal antibodies (pAb). Immunoblotting of total lysates of Escherichia coli expressing glutathione S-transferase (GST) (or maltose-binding protein [MAL]) fusion proteins with cytoplasmic domains of claudin-1 to -16 confirmed the specificities. CBB, Coomassie Brilliant Blue staining; anti-Cld-10 pAb, immunoblotting with anti-claudin-10 pAb; anti-Cld-16 pAb, immunoblotting with anti-claudin-16 pAb.

Next, to examine the distributions of individual claudins expressed in the kidney, we prepared antibodies specific for all of these claudins. Anti-claudin-1 pAb was purchased from Zymed Laboratories. pAb to claudin-2, -3, -4, -5, -8, and -11 were raised and characterized previously (11,14,18,19). We then produced GST fusion proteins or synthesized polypeptides from the cytoplasmic tails of the other claudins (claudin-7, -10, -12, -15, and -16) and raised pAb using those proteins or polypeptides as antigens. The pAb for claudin-10, -15, and -16 worked well for both immunoblotting and immunofluorescence microscopic analyses but, despite intensive efforts, we failed to obtain pAb for claudin-7 or -12 that could be used for immunofluorescence microscopy. Therefore, we have no information regarding the distributions of claudin-7 and -12 in nephrons. Furthermore, because claudin-5 and -15 were expressed specifically in endothelial cells of blood vessels, we do not discuss the expression or distribution of these claudins in the kidney in detail in this report. The specificities of newly generated anti-claudin-10 and -16 pAb were demonstrated by Western blotting analyses of total lysates of E. coli expressing GST fusion proteins with the cytoplasmic domains of claudin-1 to -16 (Figure 1B). As previously demonstrated (11,25), it was technically difficult to evaluate the specificities of anti-claudin pAb in Western blotting analyses of total lysates of whole organs, mainly because the levels of expressed claudins were fairly low.

Segment-Specific Expression of Claudin-8 in Nephrons

To identify individual nephron segments with immunofluorescence microscopy, we used antibodies to several segment markers (Figure 2). The proximal tubules and the thin descending limb of Henle were identified as AQP1-positive tubules in the cortex and the medulla, respectively (26). The thin and thick ascending limbs of Henle in the medulla were distinguished by the expression of ClC-K (27,28) and THP, respectively (29). The anti-ClC-K pAb used in this study recognized both ClC-K1 and ClC-K2, but in the medulla only the thin ascending limb of Henle was reported to express ClC-K1 (not ClC-K2) (27,28). Because the collecting tubules were reported to be positive for AQP2 (30), the distal tubules were identified as AQP1/2-double-negative tubules located around glomeruli.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Identification of nephron segments with fluorescence microscopy. (A) Nomenclature for nephron segments used in this study. (B) Segment markers. Frozen sections of mouse kidneys were immunofluorescently stained with pAb for each segment marker. The proximal tubules and the thin descending limbs of Henle were identified as aquaporin-1 (AQP1)-positive tubules in the cortex and medulla, respectively. The thin and thick ascending limbs of Henle in the medulla were distinguished by the expression of chloride channel-K1 (ClC-K1) and Tamm-Horsfall glycoprotein (THP), respectively. The distal tubules were identified as AQP1/2-double-negative tubules located around glomeruli (asterisk), and the collecting tubules were identified as AQP2-positive tubules. Arrows, tubules identified by segment markers. Bars, 20 μm.

We then immunofluorescently stained serial frozen sections from the mouse kidneys with pAb for each claudin species or with pAb for the segment markers. This method indicated that individual claudin species exhibit specific patterns of expression in nephron segments. The segment-specific expression of claudin-8 is demonstrated in Figure 3, as an example. In the cortex, claudin-8 was expressed and concentrated at junctional regions (probably TJ) in some but not all tubules. The AQP1-positive tubules lacked claudin-8 expression, whereas all of the AQP2-positive tubules expressed claudin-8, indicating that claudin-8 is expressed in collecting ducts but not in proximal tubules. The TJ-specific concentration of claudin-8 was also detected in the AQP1/2-double-negative tubules located around glomeruli, indicating the expression of claudin-8 in distal tubules. In the medulla, some tubules exhibited a concentration of claudin-8 at TJ. These tubules were ClC-K-positive but AQP1/THP-negative, leading to the conclusion that claudin-8 was expressed exclusively in the thin ascending limb of Henle in the medulla.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Segment-specific expression pattern of claudin-8 in nephrons. One of each pair of serial frozen sections was stained with pAb for segment markers in green (markers), and the other was stained with anti-claudin-8 pAb in red (claudin-8). These two images were then merged (merge). This analysis revealed that claudin-8 was concentrated at tight junctions (TJ) in epithelial cells of the thin ascending limb of Henle, the distal tubule, and the collecting duct. Arrows, tubules identified with segment markers; arrowhead, proximal tubule; double arrowhead, collecting duct; asterisks, glomeruli. Bars, 20 μm.

To evaluate these segment-specific staining patterns, we performed the following control staining experiments (Figure 4). First, frozen sections were double-stained with anti-occludin mAb and anti-claudin-8 pAb, to confirm that claudin-8 signals were derived from TJ. Second, sections were stained with preimmune serum. Third, sections were stained with anti-claudin-8 pAb in the presence of GST-claudin-8 or GST-claudin-5, to evaluate the specificity of the staining with anti-claudin-8 pAb. Using this type of analysis, with similar control staining experiments, we examined in detail the segment-specific expression of other claudin species in nephrons, as described below.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Specificities of anti-claudin-8 (anti-cld-8) pAb in immunofluorescence microscopic analyses. When the frozen sections were double-stained with anti-occludin monoclonal antibody (mAb) and anti-claudin-8 pAb, in some types of tubules (Figure 3) claudin-8 appeared to be concentrated in lines, which precisely coincided with the occludin-positive TJ. These claudin-8-positive lines were undetectable with preimmune serum or with anti-claudin-8 pAb pretreated with GST-claudin-8. Pretreatment with GST-claudin-5 did not affect the staining ability of anti-claudin-8 pAb. Bars, 20 μm.

Distinct Patterns of Expression of Claudins in Nephron Segments

Bowman’s Capsule.

To examine whether certain claudin species are expressed in Bowman’s capsule, we compared phase-contrast images of frozen sections with the corresponding immunofluorescence microscopic images obtained by using specific pAb. As demonstrated in Figure 5A, claudin-1 and -2 were concentrated at the borders of epithelial cells of Bowman’s capsule. We noted that the claudin-1 pAb (Zymed Laboratories) recognized not only claudin-1 but also claudin-3. To evaluate the staining of Bowman’s capsule with anti-claudin-1 pAb, we stained the kidneys of claudin-1-deficient mice (Furuse M, Hata M, Tsukita SH; unpublished data) with this antibody. In those mice, no signal was detected for Bowman’s capsule, indicating that claudin-1 is specifically expressed, together with claudin-2, in the capsule.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. (A) Claudins expressed in Bowman’s capsule. Claudin-1 and -2 were detected in Bowman’s capsule (arrows), which was identified by comparison of an immunofluorescence image (left) with the corresponding phase-contrast image (right). Asterisks, glomeruli. Bars, 20 μm. (B) Claudins expressed in the proximal tubule. Claudin-2, -10, and -11 were concentrated at TJ of epithelial cells in proximal tubules (red), which were identified as AQP1-positive tubules in the cortex (green). Bars, 20 μm.

Thin Descending Limb of Henle.

Only claudin-2 was detected in most of the AQP-1-positive tubules in the medulla, i.e., the thin descending limb of Henle (Figure 6). Under the fixation conditions used in this study, claudin-2 did not seem to be concentrated in TJ-like thin lines but was diffusely distributed in broad bands. This characteristic staining was absent with preimmune serum; very strangely, when sections were doubly stained with anti-occludin mAb and anti-claudin-2 pAb, not only claudin-2 but also occludin seemed to be diffusely distributed in broad bands. This staining became undetectable when anti-claudin-2 pAb was pretreated with GST-claudin-2, and staining was not observed in claudin-2-deficient mouse kidneys (data not shown). These findings suggested that the diffuse distribution of claudin-2 in the thin descending limb of Henle was specific.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. Claudins expressed in the thin descending limb of Henle. Only claudin-2 was detected in the thin descending limbs of Henle (red), which were identified as AQP1-positive tubules in the medulla (green). In these images, claudin-2 did not seem to be concentrated in TJ-like thin lines but was diffusely distributed in broad bands. Double staining with anti-occludin mAb and anti-claudin-2 pAb revealed that not only claudin-2 but also occludin seemed to be diffusely distributed in broad bands in the thin descending limb of Henle (arrows). This staining became undetectable when anti-claudin-2 pAb was pretreated with GST-claudin-2 (anti-cld-2 pAb + GST-cld-2, arrows). When the sections were fixed with TCA, similar diffuse claudin-2-staining was detected at low magnification (left), but close inspection (right, representing rectangle at left) revealed that individual sharp lines were intermingled within the claudin-2-positive zones (arrows). These lines were also Zomula Occludens-1-positive (occludin-positive) (arrows). Bars, 20 μm.

Then, to improve the preservation of fixed tubules, we applied the newly developed TCA fixation method (24) (Figure 6). Similar diffuse claudin-2 staining was detected at low magnification, but close inspection revealed that individual sharp lines were intermingled within the claudin-2-positive bands. These lines were also Zomula Occludens-1-positive (occludin-positive). Because the constituent cells in the thin descending limb of Henle are very flat and are interdigitated in a very complex manner (31), it is safe to conclude that claudin-2 was concentrated specifically at TJ in this segment.

Thin Ascending Limb of Henle.

In addition to claudin-8 (Figure 3), claudin-3 and -4 were concentrated at TJ of epithelial cells of ClC-K-positive tubules in the medulla, i.e., the thin ascending limb of Henle (Figure 7).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7. Claudins expressed in the thin ascending limb of Henle. In addition to claudin-8 (Figure 3), claudin-3 and -4 were concentrated at TJ of epithelial cells in the thin ascending limbs of Henle (red), which were identified as ClC-K-positive tubules (green). Bars, 20 μm.

Thick Ascending Limb of Henle.

Four distinct species of claudins, namely claudin-3, -10, -11, and -16, were detected at TJ of epithelial cells of THP-positive tubules in the medulla, i.e., the thick ascending limb of Henle (Figure 8). In the cortex of some species, including mice, THP has been reported to be expressed also in the distal tubules (32); however, among claudin-3, -10, -11, and -16, only claudin-3 staining was positive in the distal tubules, as demonstrated below. Because the frozen sections were not very thin, these claudin-positive lines frequently seemed to correspond to apical surfaces, but detailed double staining of transversely sectioned tubules with anti-occludin mAb clearly revealed that these lines corresponded to TJ, not to apical surfaces (Figure 8, insets).

• Download figure
• Open in new tab
• Download powerpoint

Figure 8. Claudins expressed in the thick ascending limb of Henle. Four distinct claudin species, namely claudin-3, -10, -11, and -16, were concentrated at TJ of epithelial cells in the thick ascending limbs of Henle (red), which were identified as THP-positive tubules (green). Insets, double immunostaining of transversely sectioned tubules with anti-occludin (occ) mAb and anti-claudin-10 (-11 or -16) (cld-10) pAb. The basal regions of these renal tubules were occasionally stained with anti-claudin-10 and -11 pAb but, as indicated by several control observations, this staining did not seem to be specific (insets). Bars, 20 μm.

Distal Tubule.

As demonstrated above, the distal tubules were identified as AQP1/2-double-negative tubules around glomeruli. In this type of tubule, claudin-3 (Figure 9A) and -8 (Figure 3) were expressed and localized at TJ.

• Download figure
• Open in new tab
• Download powerpoint

Figure 9. (A) Claudins expressed in the distal tubule. In addition to claudin-8 (Figure 3), claudin-3 was detected in TJ of distal tubules (red), which were identified as AQP1/2-double-negative tubules (arrows) located around glomeruli (asterisk). Double staining with anti-occludin mAb confirmed that these claudin-8 signals were derived from TJ (data not shown). Some signals seemed to be detected for glomeruli, including parietal epithelium, but control staining experiments (Figure 4) revealed that they were nonspecific. Bar, 20 μm. (B) Claudins expressed in the collecting duct. In addition to claudin-8 (Figure 3), claudin-3 and -4 were concentrated at TJ of epithelial cells in collecting ducts (red), which were identified as AQP2-positive tubules (green). Bars, 20 μm.

Collecting Duct.

In addition to claudin-8 (Figure 3), claudin-3 and -4 were clearly detected at TJ of epithelial cells of AQP2-positive tubules, i.e., collecting tubules (Figure 9B). These segment-specific expression patterns of claudins in nephrons are summarized in Figure 10.

• Download figure
• Open in new tab
• Download powerpoint

Figure 10. Summary of the segment-specific expression patterns of claudins in mouse nephrons. In the thin descending limb of Henle, some, but not all, AQP1-positive tubules were positive for claudin-2 (Cld-2). Similarly, from the thin ascending limb of Henle to the collecting duct, claudin-3 was detected in some, but not all, marker-positive tubules. These findings may be consistent with previous reports that the distal tubules, as well as the collecting tubules, are composed of several functionally distinct subsegments. In other cases, claudins were detected in all of the respective marker-positive tubules.