INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CLC-K2, A KIDNEY-SPECIFIC member of the CLC chloride channel family (11), possesses ~80% amino acid identity with rat CLC-K1 (1, 12). RT-PCR using dissected nephron segments suggested that the localization of CLC-K2 was completely different from that of CLC-K1 (1), but the exact intrarenal localization of CLC-K2 has yet to be established. Previously, we determined the intrarenal and cellular localization of CLC-K1 in rat kidney by immunohistochemistry (10). CLC-K1 is exclusively localized to both the apical and basolateral plasma membranes of the ascending thin limb of Henle's loop (ATL). Moreover, since the functional characteristics of CLC-K1 expressed in Xenopus oocyte match those of chloride transport in ATL observed in in vitro perfusion experiments (10), CLC-K1 can be considered a major chloride channel mediating the transepithelial chloride transport in ATL. Considering its structural similarity to CLC-K1, it has been speculated that CLC-K2 shares a similar role as a route for transepithelial chloride transport in the thick ascending limb of Henle's loop (TAL) and collecting ducts (CD) (1). The presence of chloride channels in the basolateral surface of these nephron segments has been postulated by several physiological and biochemical studies (5-7, 14-16, 18, 19). Recently, Zimniak et al. (20) demonstrated that treatment of TAL cells with the antisense oligonucleotide for rabbit homolog of CLC-K2 decreased the reconstituted chloride channels from the cells, thus suggesting that CLC-K2 might provide the basolateral chloride conductance in TAL. Simon et al. (8) also recently reported that mutations in CLCNKB, the gene encoding the human homolog of rat CLC-K2 (9), led to Bartter's syndrome. This evidence strongly suggested that CLC-K2 had an important role in the transepithelial reabsorption of chloride ions. To understand how CLC-K2 is involved in the pathogenesis of Bartter's syndrome, the exact intrarenal localization of CLC-K2 must be determined. However, since CLC-K1 and CLC-K2 are highly homologous proteins, the generation of a specific antibody for each channel has been very difficult. In a report on the immunohistochemistry of rat CLC-K channels by Vandewalle et al. (13), the antiserum used recognized both CLC-K1 and -K2 and was not specific for CLC-K2. They showed the relatively wide distribution of immunoreactivity in the basolateral surfaces of the ATL and other distally located nephron segments. Recently, Winters et al. (17) reported the immunolocalization of rbClC-Ka, a gene that may be the rabbit homolog of rat CLC-K2. Their antiserum recognized the basolateral surface of the TAL and cytoplasm of intercalated cells in the cortical CD (CCD) (17). The antiserum in their study was raised against a 156-amino acid COOH-terminal fragment of rbClC-Ka protein, but they did not mention the specificity of the antiserum to rbClC-Ka against a rabbit homolog of rat CLC-K1 protein. Thus there were still wide discrepancies concerning the exact intrarenal localization of CLC-K2 between them. To overcome this situation, a highly specific probe for each channel must be prepared. Since the 3'-untranslated regions of these two channels are virtually not homologous and we were able to prepare a specific cRNA probe for each channel, we adopted in situ hybridization to precisely determine the sites of CLC-K2 expression in this study.

Our in situ hybridization study clearly demonstrated that CLC-K1 and CLC-K2 are localized in different areas of the kidney, and that the latter is most abundantly expressed in distal convoluted tubules (DCT), connecting tubules (CNT), and CCD. Moderate expression was also observed in the medullary TAL (MTAL).

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Animals. Male Wistar rats weighing ~150 g received water and standard rat chow for several days before the experiments. The rats were deeply anesthetized by an intraperitoneal injection of pentobarbital (50 µg/g) and transcardially perfused with a solution of 4% paraformaldehyde for in situ hybridization and immunohistochemistry.

In situ hybridization. After the perfusion with paraformaldehyde, the kidneys were placed in a solution of 4% paraformaldehyde for 2 h and 15% sucrose overnight. Cryostat sections (5 µm) were mounted on siliconized slides. Initially, the slides were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer for 20 min, washed with 0.1 M phosphate buffer, and treated with a solution of 10 µg/ml proteinase K in 50 mM Tris · HCl, pH 7.5, and 5 mM EDTA, pH 8.0, for 1 min at room temperature. Next, they were postfixed in 4% paraformaldehyde, treated with 0.25% acetic anhydride for 10 min, rinsed in phosphate buffer (28 mM NaH₂PO₄, 72 mM Na₂HPO₄), and dehydrated in increasing concentrations of ethanol.

In in situ hybridization using radiolabeled probes, tissue sections were hybridized for 24 h at 55°C in a buffer [50% deionized formamide, 10% dextran sulfate, 0.3 M NaCl, 1× Denhardt's solution, 20 mM Tris · HCl (pH 8.0), 5 mM EDTA (pH 8.0), 0.2% sarcosyl, 200 µg/ml salmon sperm DNA, and 500 µg/ml yeast tRNA] containing one of the ³⁵S-labeled RNA probes. The probe concentration was 1 × 10⁶ cpm/200 µl per slide. After hybridization, the sections were immersed in 5× SSC at 55°C and rinsed in 50% deionized formamide with 2× SSC at 65°C for 30 min. The sections were then rinsed with RNase buffer [0.5 M NaCl, 10 mM Tris · HCl (pH 8.0), 5 mM EDTA (pH 8.0)] for 10 min three times each at 37°C, incubated with 1 µg/ml RNase A in RNase buffer for 10 min at 37°C, washed in 50% formamide with 2× SSC at 50°C for 30 min, rinsed with 2× SSC and 0.1× SSC for 10 min each at room temperature, dehydrated in alcohol, and air dried. After the slides were initially exposed to X-ray film for 3 days to provide an indication of the intensity of the hybridization signal, they were coated with Kodak NTB-2 emulsion diluted 1:1 with water. The sections were exposed at 4°C for 2-3 wk in tightly sealed dark boxes, developed in Kodak D-19, fixed with photographic fixer, washed with water, and then counterstained with hematoxylin and eosin to allow morphological identification.

In in situ hybridization using digoxigenin-labeled riboprobes, slides were hybridized in a buffer [50% deionized formamide, 0.24% SDS, 0.3 M NaCl, 1 mM EDTA (pH 8.0), 10 mM Tris · HCl (pH 8.0), 0.7 mg/ml salmon sperm DNA, and 1 mg/ml yeast tRNA, 1× Denhardt's solution, and 10% dextran sulfate] at 42°C in a moist chamber for 16 h. Concentration of a digoxigenin-labeled sense or antisense probe was 8 ng/µl. After hybridization, the sections were immersed in 5× SSC at 42°C, rinsed in 50% deionized formamide with 2× SSC at 42°C for 30 min, and incubated with 1 µg/ml RNase A in RNase buffer [0.5 M NaCl, 10 mM Tris · HCl (pH 8.0)] for 30 min at 37°C. Washing procedures included a first washing step in RNase buffer for 10 min at 37°C, followed by three washes in 2× SSC, once at 42°C for 15 min and twice at room temperature for 10 min each. The slides were then equilibrated for 5 min in buffer 1 (100 mM maleate, 150 mM NaCl, pH 7.5) for immunohistochemical detection of digoxigenin-labeled probes, immersed in buffer 2 (buffer 1 containing 1% blocking reagent; Boehringer, Mannheim, Germany), and incubated for 30 min at room temperature. Blocking solution was drained from the slides and a polyclonal alkaline phosphate-coupled sheep anti-digoxigenin antibody (diluted 1:500 in buffer 1) was applied to the sections. The sections were then incubated in a moist chamber for 2 h at room temperature, washed twice for 2 min each in buffer 1, and equilibrated for 2 min in buffer 3 (100 mM Tris · HCl, 100 mM NaCl, and 50 mM MgCl₂, pH 9.5). For signal development, the slides were immersed in a substrate solution (buffer 3 containing 0.404 mM nitroblue tetrazolium chloride, 0.384 mM 5-bromo-4-chloro-3-indolyl phosphate, 4-toluidine salt, and 1 mM levamisol) at 37°C. Color reaction was observed under microscope and terminated by immersing the slides in TE buffer [10 mM Tris · HCl (pH 8.0), 1 mM EDTA]. After staining of nuclei with methyl green, sections were mounted in PBS-buffered glycerol.

To prepare the probes specific for rat CLC-K1 and -K2, the 3'-untranslated regions of rat CLC-K1 and -K2 were amplified by PCR using K1- and K2-specific primer sets. The primer sequences were as follows

The underscored sequences are EcoR I and BamH I sites introduced for subcloning. Nucleotide identity between K1 and K2 probes was less than 30%. PCR fragments of expected sizes (K1 for 220 bp and K2 for 159 bp) were subcloned into pSPORT1 (Life Technologies), and their sequences were verified. To make ³⁵S- or digoxigenin-labeled sense or antisense cRNA probes, in vitro transcription was performed using T7 RNA polymerase and SP6 RNA polymerase after linearizing by cutting with BamH I and EcoR I, respectively.

Immunohistochemistry. Immunohistochemistry of aquaporin-2 (AQP-2) water channel (2), Tamm-Horsfall glycoprotein (Biomedical Technologies, Stoughton, MA), thiazide-sensitive Na⁺-Cl cotransporter (TSC; the antiserum was a generous gift from Dr. S. C. Hebert in Vanderbilt University), and Na⁺/Ca²⁺ exchanger (the antiserum kindly provided by Dr. R. F. Reilly at Yale University) was performed in cryostat sections of rat kidney using a TSA-Indirect kit (NEN Life Science Products, Boston, MA).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In situ hybridization of rat CLC-K2. In film autoradiograms generated from sections hybridized with rat CLC-K1 and CLC-K2 antisense probes, the most intense CLC-K1 signals were localized in the inner medulla and the most intense CLC-K2 signals were localized in the superficial cortex (Fig. 1). There were also significant CLC-K2 signals in the outer medulla (Fig. 1), but there was virtually no signal of CLC-K2 expression in the inner medulla. No hybridization signal was detected using sense probes. As shown in Fig. 2, A and B, based on the hematoxylin-eosin staining, microscopic examination of emulsion-coated kidney sections identified a high grain density in the distal nephron segments in the cortex, but not in the proximal tubules. In the outer medulla, a moderate signal was also detected in the TAL (Fig. 2, C and D). In contrast to the inner medullary localization of CLC-K1, CLC-K2 was not present in the inner medulla. To identify the nephron segments where CLC-K2 is expressed, the signal of CLC-K2 mRNA in the in situ hybridization study was compared with the immunohistochemistry of several marker proteins. TSC was used to identify DCT, Na⁺/Ca²⁺ exchanger was used for CNT, AQP-2 water channel was used for CD, and Tamm-Horsfall glycoprotein was used as a marker of the TAL. In the superficial cortex, the tubules positive for TSC, Na⁺/Ca²⁺ exchanger, and AQP-2 water channel (indicated by white arrows, Fig. 3) were all positive for CLC-K2 expression (Fig. 3, A-F). Like ³⁵S-labeled sense probe, there was no signal when sense digoxigenin probe was used for hybridization (Fig. 3I), confirming that the signals detected by digoxigenin probe were also specific to CLC-K2. These results demonstrated that the most intense signals of CLC-K2 in the kidney were present in DCT, CNT, and CCD. The signal of CLC-K2 expression was faintly observed in the AQP-2-positive tubules in the inner cortex and the outer stripe of the outer medulla (Fig. 3, G and H), but almost absent in the inner stripe of the outer medulla and the inner medulla (data not shown). This indicated that the CLC-K2 expression in the CD significantly decreased as it entered into the medulla.

View larger version (51K): [in this window] [in a new window]   Fig. 1.   Film autoradiograms illustrating CLC-K1 and -K2 mRNA distribution in a rat kidney. The hybridization signal of rat CLC-K1 was detected in inner medulla. Strongest signals of rat CLC-K2 hybridization were detected in outer cortex and inner stripe of outer medulla.

View larger version (146K): [in this window] [in a new window]   Fig. 2.   In situ hybridization of rat CLC-K2 in cortex (A and B) and outer medulla (C and D) of rat kidney. Bright-field photomicrographs demonstrating the pattern of in situ hybridization of 35S-labeled rat CLC-K2 antisense probe. Note that signals in distal nephron segments in outer cortex are stronger than those in outer medulla. Magnifications: A and C, ×20; B, ×80; and D, ×100.

View larger version (113K): [in this window] [in a new window]   Fig. 3.   Sequential sections of paraformaldehyde-fixed superficial cortex hybridized with digoxigenin-labeled rat CLC-K2 antisense probe (A, C, and E) or immunostained with an anti-TSC antibody (a marker of distal convoluted tubules; B), an anti-Na+/Ca2+ exchanger antibody (a marker of connecting tubules; D), or an anti-AQP-2 water channel antibody (a marker of collecting ducts and connecting tubules; F). Sequential sections of inner cortex hybridized with digoxigenin-labeled rat CLC-K2 antisense probe (G) and immunostained with an anti-AQP-2 water channel antibody (H). I: hybridization with digoxigenin-labeled CLC-K2 sense probe. In each pair of sequential sections (A and B, C and D, E and F, G and H), the specific nephron segments identified in the immunostaining (B, D, F, and H) are marked with white arrows on in situ hybridization (A, C, E, and G). All the first antibodies and the second antibodies [horseradish peroxidase (HRP)-conjugated anti-rabbit IgG for the detection of TSC and AQP-2, and HRP-conjugated anti-guinea pig IgG for Na+/Ca2+ exchanger] were diluted to 1:300. AQP-2, aquaporin-2; TSC, thiazide-sensitive Na+-Cl cotransporter. Magnification, ×100.

In the inner stripe of the outer medulla, CLC-K2 expression matched the staining pattern of Tamm-Horsfall glycoprotein (Fig. 4, A-D), thereby confirming that the moderate expression of CLC-K2 in the outer medulla was present in the MTAL. In the cortex, the tubules positive for Tamm-Horsfall glycoprotein were faintly positive for CLC-K2 expression (Fig. 4, E and F), suggesting that the level of expression of CLC-K2 in the cortical TAL (CTAL) was significantly lower than that in the MTAL. This pattern of expression of CLC-K2 in the TAL is opposite to that in the CD.

View larger version (207K): [in this window] [in a new window]   Fig. 4.   In situ hybridization of rat CLC-K2 (A, C, E) and immunohistochemistry of Tamm-Horsfall glycoprotein (B, D, F) in sequential sections of rat kidney. A and B: inner stripe of outer medulla (×50). C and D: inner stripe of outer medulla (×100). E and F: cortex (×100). Tubules strongly positive for Tamm-Horsfall glycoprotein in inner stripe of outer medulla (B and D) were also moderately positive for rat CLC-K2 expression (A and C). In C, moderate expression of CLC-K2 was observed in the tubules positive for Tamm-Horsfall glycoprotein (indicated by white arrows). In contrast, rat CLC-K2 expression was faint in the tubules in the cortex (indicated by white arrows) (E) that were positive for Tamm-Horsfall glycoprotein (F).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We previously showed by RT-PCR using dissected nephron segments that the intrarenal localization of CLC-K2 is completely different from that of CLC-K1 (1). However, the technical difficulty of dissecting objective nephron segments without contamination by other nephron segments and the high sensitivity of the PCR method often resulted in false positive data or data that were not physiologically relevant to the localization of the message along the nephron segments. For example, CLC-K1 mRNA was detected in the CD in our initial report (12), but immunohistochemical findings clearly demonstrated that CLC-K1 protein was exclusively localized to the ATL (10). In a study using the RT-PCR technique, Kieferle et al. (3) also reported that rat CLC-K1 and CLC-K2 were present relatively broadly along the nephron segments and possibly colocalized in some nephron segments. In a recent RT-PCR analysis of rat CLC-K1 and -K2 by Vandewalle et al. (13), both clones were found to be widely expressed along the nephron segments. To precisely determine where CLC-K1 and CLC-K2 are expressed in the kidney, we prepared CLC-K1- and CLC-K2-specific riboprobes for in situ hybridization. The nucleotide sequences of coding regions of CLC-K1 and -K2 cDNAs are very homologous, so we chose the 3'-untranslated region of each cDNA to generate clone-specific riboprobes. Since the nucleotide identity between K1- and K2-specific probes was less than 30%, there was no cross-hybridization under the condition used in this study. Using these probes, our in situ hybridization study clearly demonstrated that although CLC-K1 and CLC-K2 are highly homologous proteins, they are not colocalized in the kidney. As in our previous RT-PCR study (1), the main sites of CLC-K2 expression were identified to be TAL and the distal nephron segments, including the DCT, CNT, and CCD. On the basis of the in situ hybridization results using ³⁵S-labeled cRNA probe, the intensity of the signal in the distal nephron segments was about an order of magnitude higher than that in the TAL (Fig. 2). Accordingly, the main site of CLC-K2 expression in the kidney is the distal nephron segments in the superficial cortex. This contradicts the result of our previous Northern blot (1), in which CLC-K2 expression was highest in the inner medulla, moderate in the outer medulla, and faint in the cortex. We reevaluated the previous record and came to the conclusion that the developed film had been mistakenly aligned with the blot. As a result, the lane of the inner medulla was labeled as the cortex, and that of the cortex was labeled as the inner medulla. We performed Northern blot again and confirmed that the signal of CLC-K2 in the kidney was highest in the cortex, moderate in the outer medulla, and faint in the inner medulla (data not shown), which is consistent with the present in situ hybridization data.

The expression of CLC-K2 mRNA was clearly detected in glomeruli using digoxigenin probe in the Bowman's capsule and epithelial cells (Fig. 3). The absence of signal in glomeruli using ³⁵S-labeled probe may be explained by the difference in sensitivity between ³⁵S-labeled and digoxigenin-labeled probes. The existence of CLC-K2 mRNA in glomeruli was consistent with our previous study, in which human CLC-K2 message was detected in glomeruli by the RT-PCR method (9). However, the physiological significance of CLC-K2 in the glomeruli remains to be determined. In contrast, previous physiological data suggested the existence of chloride conductance in the basolateral plasma membranes of the TAL and CD (4). Therefore, we speculate that CLC-K2 may be present in the basolateral plasma membranes of the DCT, CNT, CCD, and MTAL. This is in agreement with the evidence that the loss-of-function mutations of ClC-Kb (the human homolog of rat CLC-K2) lead to a loss of chloride ions from the body, i.e., Bartter's syndrome.

Although previous immunohistochemistry studies (13, 17) coincided with this in situ hybridization study in general, there are some discrepancies in the details. Winters et al. (17) reported that the rabbit homologue of CLC-K2 was present in the basolateral plasma membrane of the MTAL. In the cortex, intercalated cells of the CCD had cytosolic immunoreactivity. Our in situ hybridization detected strong signals of the CLC-K2 message in the principal cells of the CNT and CCD that far exceeded the signal detected in the MTAL. Regardless of whether the staining is basolateral or cytoplasmic, much stronger staining is expected in the cortex. This discrepancy could be partly due to the species difference between rat and rabbit. In a recent report on the immunohistochemistry of rat CLC-K channels by Vandewalle et al. (13), their antibody could recognize both CLC-K1 and -K2, since the antigen peptide, the carboxy-terminal end of CLC-K2, differed from that of CLC-K1 by only one amino acid. Accordingly, their immunostaining could not show where each channel was present in the kidney. Their RT-PCR analysis showed almost identical expression profiles of K1 and K2 along the nephron segments, suggesting that both clones could be colocalized. We clearly showed in this study that CLC-K1 is almost exclusively expressed in the inner medulla, further confirming our previous immunohistochemistry of CLC-K1 (10). In contrast to the inner medullary localization of CLC-K1, CLC-K2 is expressed in the cortex and outer medulla, but not in the inner medulla. This clearly showed that CLC-K1 and -K2 are not colocalized and may have different roles in different nephron segments in the kidney. Given that rat CLC-K1 is expressed only in the ATL (10), the staining in tubules other than the ATL in the study of Vandewalle et al. (13) should be the staining of CLC-K2. Although the staining of the DCT, CD, MTAL, and CTAL in their study mostly agrees with our present study, they clearly showed the staining of the inner medullary CD, where we could find no signal of CLC-K2 expression. Since the antibody in the study of Vandewalle et al. (13) could not be used for Western blot of kidney membrane, it is still possible that their antibody cross-reacted with proteins other than CLC-K channels in the kidney.

There has been no clear determination of localization of CLC-K1 and -K2 expression in the kidney. By clearly showing in this study that CLC-K1 is expressed only in the inner medulla, we could confirm our previous RT-PCR analysis and immunohistochemistry. We also were able to clearly determine the sites and the relative abundance of CLC-K2 expression along the nephron segments. CLC-K2 is moderately expressed in the MTAL and highly expressed in the DCT, CNT, and CCD. Based on this study and studies by Vandevalle et al. (13) and Simon et al. (8), it is highly conceivable that CLC-K2 is a chloride channel that serves as a route for transcellular chloride transport (an exit for chloride ions in the basolateral plasma membrane of the TAL, DCT, CNT, and CCD).