INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

PLASMALEMMAL Cl/ exchange activity contributes to regulation of intracellular pH (pH_i) and cell volume and to generation and maintenance of the transmembrane Cl gradient regulation in a wide variety of cell types. Polarized expression of Cl/ exchange activity in epithelial cells is thought to contribute to transepithelial transport of acid/base and volume equivalents (2).

Cl/ exchange activity has been measured throughout the length of the nephron, and different segments and cell types have been shown to express activity at the basolateral, apical, or both poles of the cell. mRNA transcripts of all three characterized AE anion exchanger genes, AE1, AE2, and AE3, are expressed in mammalian kidney (5, 9, 10). Among these mammalian transcripts, mRNAs encoding AE2 (3, 27) and AE1 (1, 9, 27) are the more abundant on Northern blots, whereas AE3 mRNA (5, 27) is detectable by reverse transcription-polymerase chain reaction (RT-PCR) but not easily by Northern blot.

An NH₂-terminally truncated AE1 polypeptide (9) has been localized in the kidney of many species, including rat (4) and mouse (8), to the basolateral surface of type A intercalated cells (IC) of both medullary collecting duct and cortical collecting duct (CCD) and of connecting segments (CNT). The source of this immunostaining pattern as AE1 has been more recently confirmed by documentation of its absence in type A IC of mice null for expression of the entire AE1 gene (30).

Molecular identification of the polypeptides responsible for Cl/ exchanger activities of other renal cells in situ, including that of the apical exchanger(s) of type B IC, has remained uncertain (1, 2, 4). Even the standard classification of IC as type A and type B may need revision to accommodate the growing evidence of increased heterogeneity displayed both in histochemical studies (4, 11, 32) and in ion transport studies of single cells (17, 39). AE1, AE2, and AE3 each mediate Cl/ exchange, and their anion specificities appear to be similar. However, anion transport by recombinant AE2 is regulated differently than that mediated by AE1, as described in distinct expression systems using different functional assays (21-23, 40). Thus it is likely that regulation of Cl/ exchange in different renal cell types will differ. RT-PCR analysis has already suggested that levels of AE1 and AE2 mRNA respond differently to identical stimuli in the intact animal (18). Immunolocalization of distinct AE anion exchanger isoforms in the kidney will contribute to the correlation of molecular structure with in vivo and in vitro function.

AE1 has been localized to basolateral membranes of type A IC in many species, consistent with the amply documented presence of Cl/ exchange in these cells (1, 17, 39). However, Cl/ exchange activity has also been measured in isolated perfused tubules from most nephron segments, as well as in primary cell cultures derived from many segments. In addition to participating in acid secretion by type A IC, Cl/ exchange participates in base secretion by type B IC (1, 17, 39), in volume regulation by cells of the medullary thick ascending limb (MTAL) (20), and in Cl secretion (26) and acid secretion (33) by the inner medullary collecting duct (IMCD). In addition, Cl/ exchange is present in nearly every renal epithelial cell in culture as part of the cellular "housekeeping function" of pH_i regulation (2).

The AE2 polypeptide has been detected by immunoblot in rat and mouse kidney, with higher levels per milligram of protein in medulla than in cortex (10). Antipeptide antibodies recognizing four distinct epitopes of AE2 have localized AE2 polypeptide by immunocytochemical techniques in transiently transfected cultured cells and in semithin sections of choroid plexus epithelial cells (6) and gastric parietal cells (35), the sites of greatest abundance of AE2 mRNA. Immunohistochemical detection of AE2 in stomach and choroid plexus correlated with AE2 polypeptide abundance on immunoblot in these tissues (6, 10, 35). However, immunolocalization of AE2 in kidney using the same methods and antibody reagents proved unsuccessful.

In recent years, epitope unmasking techniques have been introduced to enhance the sensitivity of antigen immunodetection in microscopic tissue sections and in fixed cells on coverslips. Recently, we reported a new addition to the repertoire of epitope unmasking techniques, that is, brief pretreatment with sodium dodecyl sulfate (SDS) of aldehyde-fixed tissue sections on slides. This procedure was useful for antibodies raised against numerous proteins and in several cases proved to be a requirement for immunocytochemical competence of the antibody. SDS pretreatment of fixed cells was also found greatly to enhance detection of basolateral AE2 anion exchanger in Madin-Darby canine kidney (MDCK) cells grown on cellulose nitrate supports (13).

We have now used the SDS epitope unmasking technique to detect the AE2 polypeptide in cryosections of rat kidney. AE2 was shown to be localized in basolateral membranes of tubular epithelial cells in all nephron segments beyond the proximal tubule. An additional AE2-related epitope was present in the Golgi apparatus of multiple cell types, most notably in epithelial cells of IMCD and MTAL.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Tissue preparation. Adult male or postpartum female Sprague-Dawley rats were maintained on a standard diet and had free access to water. Where noted, two animals were injected intraperitoneally with colchicine (0.5 mg/100 g body wt) 6 h before death as previously described (19). Animals anesthetized with Nembutal (65 mg/kg ip) were perfused via the left ventricle with Hanks' balanced solution, with drainage from the severed inferior vena cava, until the kidneys were thoroughly blanched. The rats were then perfusion-fixed with 2% paraformaldehyde/75 mM lysine/10 mM sodium periodate (PLP) as previously described (4, 19). Some rats were perfusion-fixed with 3% paraformaldehyde in 140 mM NaCl, 20 mM sodium phosphate, pH 7.4 [phosphate-buffered saline (PBS)]. PLP-perfused and paraformaldehyde-perfused kidneys were excised, cut into blocks of cortex, medullary outer stripe and inner stripe, or into larger coronal blocks, and further fixed in PLP overnight at 4°C. Fixed tissue blocks were washed four times with PBS, then stored at 4°C in PBS containing 0.02% sodium azide until further use.

Antibodies. Affinity-purified rabbit polyclonal anti-AE2 amino acids (aa) 1224-1237, directed against the COOH terminus of AE2, and affinity-purified rabbit polyclonal anti-AE2 aa 961-974 and 424-440 have been previously described (6, 35). Crude rabbit antiserum raised against mouse AE1 aa 917-929, directed against the COOH terminus of AE1, was prepared by the same methods and has been previously described as an immunoprecipitating reagent¹ (15). Mouse monoclonal antibody (MAb) to immunoglobulin G1 (IgG1), MAb 12B11, was raised against rat red blood cell ghosts stripped with 0.1 N NaOH and characterized as competent in immunoprecipitation and immunoblot assays using red cell AE1 and in immunocytochemical assays using red blood cells and (as shown here) type A renal IC. Secondary antibodies were Cy3-coupled donkey anti-rabbit Ig, and fluorescein-coupled or dichlorotriazinylamino-fluorescein-coupled goat anti-rabbit and goat anti-mouse Ig (Jackson Immunoresearch, West Grove, PA).

Immunofluorescence microscopy. Fixed tissue blocks were infiltrated with 30% sucrose in PBS, frozen in liquid nitrogen, and sectioned at 5-7 µm thickness on a Reichert-Jung Frigocut model 2300N cryostat. Some tissue blocks were sequentially infiltrated with 1.6 M and 2.3 M sucrose, prior to sectioning at 1-µm thickness on a Reichert-Jung Ultracut ultracryotome. Sections were placed on Superfrost/Plus Microscope Slides (Fisher) and stored in PBS/azide at 4°C until use or alternatively stored at 20°C for longer periods.

Indirect immunofluorescence was performed as previously described (4, 6, 32, 35). Sections were preincubated at room temperature in PBS for 10 min, in 1% bovine serum albumin in PBS for 15 min, then incubated at room temperature for 1-2 h with primary antibody as indicated. Some sections were subjected to a double-incubation procedure. Epitope unmasking with SDS was performed as previously reported (13). Cryosections of fixed tissue on slides were brought to room temperature, rehydrated in PBS for 5 min, then exposed to 1% SDS in PBS for 15 min, followed by three 5-min washes with PBS prior to incubation with primary antibody for 1 h. (SDS exposure for 10 or for 5 min was equally effective.) Peptide antigens were included in the incubation mix at 12 µg/ml unless otherwise noted. Irrelevant peptides were included at 12 µg/ml in all incubations designed to localize antigen. In some cases, consecutive sections were incubated with and without SDS treatment, to compare results on the same tubule segments.

Sections were then incubated for 1 h with fluorophore-conjugated secondary antibodies (10-15 µg/ml), again washed for three 5-min washes in PBS, and mounted in 50% glycerol in PBS, pH 7.5, containing 2% n-propyl-gallate as an antiquenching agent. Sections were examined and photographed with a Olympus BH-2 or a Nikon FXA epifluorescence photomicroscope, using Kodak TMAX 400 film push-processed to 1600 ASA.

Immunoperoxidase electron microscopy. PLP-fixed tissue was cryoprotected in 10% dimethyl sulfoxide for 1 h, then frozen in liquid N₂, and 30-µm cryosections were cut. Sections were incubated at 4°C for 10 min in PBS containing 0.05% saponin, then incubated overnight at 4°C with affinity-purified anti-AE2 aa 1224-1237 diluted 1:400 in PBS saponin. After six 10-min rinses in PBS saponin, sections were further incubated 6 h in biotin-coupled goat anti-rabbit IgG (Jackson) diluted 1:100. After six additional 10-min rinses in PBS saponin, sections were incubated overnight in PBS saponin containing avidin-biotin-horseradish peroxidase complex reagent (ABC, Vector Laboratories), then rinsed again for six 10-min rinses in PBS saponin and three 10-min rinses in PBS alone. The peroxidase reaction was initiated by addition of 6 µl of 30% H₂O₂ to 10 ml diaminobenzidine (DAB, 1 mg/ml). After 3 min, the reaction was stopped by removal of DAB solution, and the tissue was washed in PBS three times for 5 min each time, then fixed in 1% glutaraldehyde in PBS for 30 min. Tissues were then washed in PBS, postfixed 1 h in 1% osmium tetroxide, dehydrated in graded ethanol solutions, and embedded in LX-112 resin (Ladd Industries, Burlington, VT). Thin sections were cut and examined on a Philips CM10 electron microscope after heavy metal staining with uranyl acetate and lead citrate.

RT-PCR. Total RNA was prepared from freshly dissected rat kidney and rat stomach using the Qiagen RNeasy kit. RT was performed with the First Strand cDNA synthesis kit from Ambion. PCR was performed by the hot start procedure, using Taq DNA polymerase (Promega) in the supplier's recommended buffer. The forward (5') AE2a primer [nucleotide (nt) 18 through +8, rat numbering; Ref. 27], designed to be used for either rat or mouse, was of sequence 5' [AAGtGaTcA]GATTTGGCCATGAGCAG 3'. (Nucleotides within brackets consist of a three nucleotide spacer followed by a hexameric restriction site; lowercase letters indicate mismatches with the rat sequence introduced to create the restriction site.) The forward AE2b primer (nt 42-62; Ref. 38) was of sequence 5' CACTCCCGCAGGATGACTCAG 3'. The forward AE2c primer (nt 186-210; Ref. 38) was of sequence 5' CTGCAGTTTCAGAGTTCATTTCCAG 3'. The reverse (3') primer common to all AE2 isoforms (nt 1007-982; Ref. 27) was of sequence 5' [TGAGaaTtC]TGGTTTTTGTCCAACAG 3'. The resultant PCR fragments were of the following predicted lengths: AE2a, 1025 bp; AE2b, 977 bp; AE2c1, 451 bp; and AE2c2, 781 bp.

PCR mixes lacking only primers were preincubated at 82°C for 1 min, then primers were injected into the mix through oil. The complete reaction mixes were denatured for 2 min at 95°C, and then subjected to these cycle conditions: denaturation for 45 s at 94°C, annealing for 2 min at 60°C, and elongation for 2 min at 72°C. Final extension of 10 min at 72°C was terminated by rapid cooling to 4°C after 35 cycles (AE2) or 25 cycles [glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and -actin].

DNA transferred to nitrocellulose was hybridized to a ³²P-labeled internal oligonucleotide of sequence 5' CAGCACCTCCGTCGTCACCT 3' (nt 646-627) from a region present in all AE2 isoforms (38). Identity of PCR products was verified by DNA sequencing on an ABI 371 sequencer.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Immunostaining without epitope unmasking procedures. AE2 localization in rat kidney was examined by immunofluorescence microscopy using four different polyclonal antibodies and one MAb. None of the polyclonal affinity-purified antibodies directed against peptides encoding mouse AE2 aa 961-974, 835-846, 625-639, 424-440, 355-368, or 109-122 detected specific immunostaining on rat kidney cryosections, regardless of the fixation methods chosen (see METHODS). This was despite previous successful immunocytochemical localization of AE2 in gastric parietal cells and in choroid plexus the antibodies to aa 961-974, 424-440 (6, 35), and 109-122 (unpublished results). Similarly, although affinity-purified rabbit polyclonal anti-AE2 aa 1224-1237 successfully immunolocalized AE2 in these same tissues (6, 35), the reagent immunostained in perfused kidney only the cross-reactive AE1 present in the basolateral membranes of type A IC and the few retained red blood cells (32). Among the epitope unmasking techniques tried that did not elicit AE2 immunostaining with polyclonal anti-AE2 aa 1224-1237 were those of on-slide trypsinization of the fixed section, microwaving of the section, and treatment of the section with 10 mM NaOH or with a range of chaotropic agents and nondenaturing detergents.

However, one epitope unmasking technique greatly enhanced AE2 immunostaining in MDCK cells grown on permeable supports: treatment of the mounted, fixed cell monolayer with 1% SDS for 5-15 min (13). Similar on-slide treatment of thick cryosections of PLP-fixed rat kidney "unmasked" a remarkably enhanced immunostaining in various kidney cell types, as described below. Several control experiments, to be described below, demonstrated that this SDS-dependent immunostaining resulted from recognition of AE2 rather than of AE1.

AE2 in cortex. Figure 1 compares AE1 and AE2 localization in an SDS-treated section of rat kidney cortex. Figure 1a shows AE1 immunostaining in type A IC (arrows) of the CCD, as detected with the monoclonal anti-AE1 antibody, MAb 12B11. In Fig. 1b, polyclonal anti-AE2 1224-1237 detected in the same section not only AE1 in the same type A cells (arrows) but, in addition, moderate AE2 staining in principal cells (PC) and/or type B IC. Brighter AE2 immunostaining of variable intensity was present in the some but not all of the epithelial cells of the cortical thick ascending limb (CTAL). Staining of distal convoluted tubule (DCT) cells and CNT cells was weak or absent (not shown). S1/S2 proximal tubule autofluorescence (Fig. 1b, transverse tubular sections above CCD) often obscured a faint, diffuse AE2 immunostaining pattern, which could be competed by specific peptide (not shown). The SDS-elicited AE2 immunostaining shown in Fig. 1b and in subsequent figures was obtained in the presence of irrelevant peptide and, with the exception of PT, was completely (24 µg/ml) or nearly (12 µg/ml) abolished in the presence of peptide antigen (see below).

View larger version (95K): [in this window] [in a new window]   Fig. 1.   Comparison of AE1 and AE2 distribution in SDS-pretreated 6-µm cryosection of deep cortex of rat kidney. a: Monoclonal anti-AE1 antibody stains the basolateral membranes of type A intercalated cells (IC) in cortical collecting duct (CCD, arrows) but not type B IC and principal cells (PC). b: In the same section, anti-AE2 aa 1224-1237 reveals staining of basolateral membranes of epithelial cells of the cortical thick ascending limb (CTAL, asterisk). Although AE1 is still evident in basolateral membranes of type A IC (arrows), SDS pretreatment has revealed AE2 in the basolateral membrane of some adjacent CCD cells of other types. Diffuse proximal tubule staining was not competed by peptide antigen (not shown). Bar = 25 µm.

Brief SDS treatment of cryosections on slides was required for detection of the plasmalemmal AE2 COOH-terminal epitope by polyclonal anti-AE2 aa 1224-1237 (Fig. 1b). SDS pretreatment also enhanced AE1 immunostaining by anti-AE1 MAb 12B11 and by anti-AE1 aa 917-929 (not shown) but did not allow the anti-AE1 antibodies to detect AE2 in kidney (see below). Therefore, further AE2 immunolocalization studies in other sections of the kidney were carried out only with polyclonal anti-AE2 aa 1224-1237.²

The epithelial cells of the macula densa uniformly expressed AE2 in their basolateral membranes at higher levels than seen in the adjacent CTAL cells (Fig. 2). Within the glomeruli, there was also evident weaker but specific AE2 immunostaining, likely in mesangial cells and endothelial cells. Figure 2e shows (to the left of the macula densa) endothelial AE2 in an afferent arteriole. AE2 immunostaining beyond the macula densa decreased in the short terminal portion of CTAL and the DCT (not shown).

View larger version (87K): [in this window] [in a new window]   Fig. 2.   Five examples (a-e) of SDS-pretreated 1-µm cryosections of rat kidney cortex showing AE2 immunostaining of basolateral membranes of macula densa in transverse section. Adjacent CTAL cells are less intensely stained; e shows, in addition, an type A IC (arrow) and two red blood cells retained in the glomerulus. AE2 staining within the glomeruli is also competed by excess peptide antigen (12 µg/ml).

AE2 in outer stripe of outer medulla. In this kidney region, SDS pretreatment revealed basolateral localization of AE2 in MTAL cells with intensity stronger than in CTAL (Fig. 3A, note ""). AE2 was also present in PC of the outer medullary collecting duct (OMCD) (not shown). At the frontier between the outer stripe and the inner stripe (Fig. 3A), S3 proximal tubules showed little or no AE2 staining (Fig. 3, note "~"), whereas the descending thin limb (DTL) (Fig. 3, note asterisk) displayed an abrupt increase in staining in the basolateral membrane.

View larger version (97K): [in this window] [in a new window]   Fig. 3.   AE2 immunostaining in thin limbs (SDS-pretreated 5-µm cryosections). A: AE2 immunostaining in basolateral membranes of descending thin limb (*) at junction with S3 proximal tubule (~). Pair of tubules immediately above are medullary thick ascending limb (MTAL, ). B: AE2 immunostaining in basolateral membranes of ascending thin limb (asterisk) at junction with MTAL (~). Brightest staining at bottom left and top right is AE1 in type A IC of nearby outer medullary collecting ducts. Bar = 35 µm.

AE2 in inner stripe of outer medulla. Figure 4 shows double immunostaining for AE1 and AE2 in a thick cryosection of outer medullary inner stripe subjected to the SDS epitope unmasking treatment. As shown in Fig. 4A, the AE1-specific MAb immunostained the basolateral membranes of type A IC (arrows) but not PC (arrowheads). AE1 was also detected in red blood cells retained in the perfused kidney without staining the thick ascending limb. This immunostaining in both cell types was abolished by preincubation of the MAb in the presence of rat red blood cell ghosts (not shown).

View larger version (114K): [in this window] [in a new window]   Fig. 4.   Comparison of AE1 and AE2 distribution in SDS-pretreated 6-µm cryosection of inner stripe of outer medulla in rat kidney. A: monoclonal anti-AE1 antibody stains the basolateral membranes of type A IC in medullary collecting duct (arrows) but not PC (arrowheads) or cells of the MTAL. B: in the same section, anti-AE2 aa 1224-1237 reveals dense staining of basolateral membranes of MTAL epithelial cells. In collecting ducts, SDS pretreatment has revealed AE2 in the basolateral membrane of PC (arrowheads). AE1 is still evident in basolateral membranes of type A IC (arrows) but at reduced intensity as a result of the steric hindrance of the monoclonal antibody recognizing the same protein. Round white spots in top center are remnant red blood cells. Bar = 25 µm.

View larger version (95K): [in this window] [in a new window]   Fig. 5.   Comparison of AE1 and AE2 distribution in SDS-pretreated 6-µm transverse cryosection of upper portion of rat kidney inner medulla. b: The few type A IC remaining at this level of inner medullary collecting duct (IMCD) display basolateral AE1 (arrows), detected with monoclonal anti-AE1. a: In addition to enhanced staining of the same cells (arrows), the anti-AE2 aa 1224-1237 antibody detects basal (arrowhead) and basolateral staining of IMCD cells that are undetected by the specific anti-AE1 antibody (b). More faintly staining tubules in a include thin limbs. Bright spots in a show cross-reacting AE1 in red blood cells, also detected by the specific anti-AE1 antibody (b). Bar = 25 µm.

View larger version (71K): [in this window] [in a new window]   Fig. 6.   AE2 immunostaining in SDS-pretreated 6-µm longitudinal cryosection of a more distal portion of rat kidney inner medulla. IMCD epithelial cells of this junctional portion of the collecting duct uniformly display basolateral AE2. Fainter staining in the parallel tubules at top left represents thin descending limbs. Bright spots outside tubules are retained red blood cells with cross-reactive AE1. Bar = 25 µm.

View larger version (114K): [in this window] [in a new window]   Fig. 7.   Intracellular localization of an AE2-related epitope in 5-µm cryosections of outer medullary inner stripe (a and b, near sequential sections) or of the upper portion of inner medulla (c and d, nonsequential sections). Basolateral AE2 staining was evident in epithelial cells of MTAL (a) and of IMCD (c) and (more faintly) of thin limbs only following SDS pretreatment. In absence of SDS pretreatment, epithelial cells of MTAL (b) and of IMCD (d) both revealed a punctate pattern of intracellular staining. Intensely stained AE1 in type A IC and in red blood cells is evident in either condition (a-d). Bar = 50 µm.

View larger version (131K): [in this window] [in a new window]   View larger version (174K): [in this window] [in a new window]   View larger version (215K): [in this window] [in a new window]   Fig. 8.   Golgi localization of an AE2 epitope. a: Anti-AE2 aa 1224-1237 immunostaining in rat kidney inner medulla in SDS-untreated 6-µm cryosection. Note Golgi-like distribution in the nonerythroid cells located in the regions between the IMCDs (arrowhead), as well as in the epithelial cells of the IMCD. Bar = 25 µm. b: Electron microscopic immunoperoxidase localization of AE2 in rat kidney inner medullary interstitial cell. Note the electron-dense reaction product in the Golgi stacks (arrow) and in the centriole-like microtubule organizing centers (arrowheads). Bar = 1 µm. c: Electron microscopic immunoperoxidase localization of AE2 in rat kidney inner medullary endothelial cell adjacent to intraluminal erythrocyte. Bar = 2 µm. Note the electron-dense reaction product in the endothelial Golgi stacks (arrow) as well as in the red cell plasma membranes.

View larger version (73K): [in this window] [in a new window]   Fig. 9.   Effects of pretreatment of intact animals without (a and c) or with colchicine (b and d) on renal distribution of AE epitopes detected by anti-AE2 aa 1224-1237 antibody in 6-µm cryosections. Neither AE2 in IMCD or inner medullary thin limbs nor AE1 in type A IC (a) were altered in distribution by colchicine treatment (b). In contrast, the supranuclear distribution of the Golgi-associated AE2-like epitope of MTAL epithelial cells (c) was attenuated and redistributed by colchicine treatment to a more diffuse circumnuclear pattern (d). Bar = 50 µm.

View larger version (107K): [in this window] [in a new window]   Fig. 10.   Specificity of immunostaining with anti-AE2 aa 1224-1237 in mouse stomach in presence of irrelevant peptide (a), in presence of AE1 COOH-terminal peptide (b), in presence of AE3 COOH-terminal peptide (c), and in presence of the peptide antigen, AE2 COOH-terminal peptide (d). Bar = 50 µm.

View larger version (120K): [in this window] [in a new window]   Fig. 11.   Specificity of immunostaining in rat kidney cortex with anti-AE1 aa 917-929 (A-C) and with anti-AE2 aa 1224-1237 (D-L) with (D-I) or without SDS pretreatment (G-L). Antibody incubations were performed in presence of irrelevant peptide (top: A, D, G, and J), AE1 aa 917-929 antigen peptide (middle: B, E, H, and K), or AE2 aa 1224-1237 antigen peptide (bottom: C, F, I, and L). Only G, H, and I are sequential sections (gl, glomerulus). Scale bars: first 3 columns (A-I), 25 µm; right column (J-L), 20 µm.

View larger version (38K): [in this window] [in a new window]   Fig. 12.   A: reverse transcription-polymerase chain reaction (RT-PCR) phenotype of alternative AE2 transcripts in rat stomach (lanes 1-3) and in rat kidney (lanes 4-6) as detected by ethidium bromide fluorescence. PCR fragments are of the following lengths: lanes 1 and 4 (AE2a), 1,025 bp; lanes 2 and 5 (AE2b), 977 bp; lanes 3 and 6, 451 bp (AE2c1) and 781 bp (AE2c2). Lanes 5 and 6 were loaded with twice the volume of PCR product as loaded in lanes 1-4. B: Southern blot of AE2 PCR products from rat stomach (lanes 1-3) and rat kidney (lanes 4-6) transferred to nylon from gel of A and hybridized with a 32P-labeled common internal AE2 oligonucleotide probe as described in METHODS. Exposure time at room temperature for lanes 1-4 was 16 h and for lanes 5 and 6 was 48 h.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Specificity of the immunolocalization of the unmasked AE2 epitope. The three anti-AE2 antipeptide antibodies competent to immunolocalize AE2 in rat choroid plexus (6) and in rat stomach (35) did not detect AE2 in rat kidney processed by the same methods used for the earlier studies. However, application of a recent epitope-unmasking protocol, using SDS treatment of fixed cryosections prior to antibody incubation (13), allowed detection in rat kidney of a single AE2 epitope, the COOH-terminal aa 1224-1237. Other available AE2 peptide epitopes were not similarly "unmasked" in cryosections by SDS treatment.

Because the anti-AE2 antibody to the COOH-terminal aa 1224-1237 cross-reacted with the AE1 COOH-terminal epitope, it was necessary to document that the unmasked immunostaining observed in the kidney derived from AE2, and not from AE1. This was achieved by comparison of unmasked AE2 staining patterns with staining patterns of the anti-AE1 monoclonal antibody, MAb 12B11, that does not recognize AE2. In addition, to discriminate more precisely between the related COOH-terminal epitopes of AE1 and AE2, the immunostaining patterns and peptide competition specificities of antibodies directed against the respective COOH-terminal peptides of AE1 and AE2 were compared (Fig. 8). Thus it was shown that antibodies to two SDS-enhanced AE1 epitopes (this work), as well as two additional antibodies (4), revealed an immunostaining pattern restricted to type A IC basolateral membranes and to erythrocytes. In contrast, the immunostaining pattern of the unmasked AE2 epitope was competed completely by AE2 peptide antigen but only minimally by the related AE1 peptide. The minimal competition of AE2 immunostaining produced by the AE1 peptide was consistent with similar minimal competition of AE2 immunostaining in gastric parietal cells produced by the AE1 peptide (Fig. 7).

Rat kidney AE2 has been immunolocalized by detection of only a single epitope, albeit with multiple controls for immunospecificity. Therefore, this localization has been demonstrated less conclusively than the multi-epitope localizations of AE1 in rat kidney (4) and of AE2 in rat stomach (35) and rat choroid plexus (6). Nonetheless, four criteria of specificity support the validity of the localization of AE2 in rat kidney as reported above. First is the contrast in localization with multiple epitopes of AE1. Second is the isoform specificity of COOH-terminal peptide competition of the AE2 immunocytochemical signal. Third is the correlation between this isoform specificity of COOH-terminal peptide competition and that of the AE2 polypeptide detected on immunoblot analysis of rat kidney microsomes (10). Fourth is the general coincidence between the unmasked epitope and the localization of rat kidney AE2 mRNA in microdissected nephron segments subjected to RT-PCR (10). Moreover, the expression of AE2 in PC is consistent with the RT-PCR findings in immunodissected rabbit CCD cells by Fejes-Toth et al. (18). Figure 13 summarizes the immunocytochemically defined localization of AE2 polypeptide along the rat kidney nephron.

View larger version (19K): [in this window] [in a new window]   Fig. 13.   Schematic drawing of AE2 COOH-terminal epitope distribution along the nephron. Relative AE2 staining intensities are shown as +/-, +, and ++. The open double crosses and open triple crosses represent AE1 staining intensity. Type A IC of upper IMCD are not considered. DTL, thin descending limb; ATL, ascending thin limb; MD, macula densa; OMCD, outer medullary collecting duct.

Implications of the need for "unmasking" for visualization of the AE2 COOH-terminal epitope. Several epitopes of AE2 in gastric parietal cells and in choroid plexus are detectable by standard immunofluorescence methods, but visualization in kidney of even the most robust of these epitopes requires epitope unmasking.

Why might this be so? Some of this difference likely resides in simple variations in abundance: thus AE2 is more abundant in parietal cells and choroid plexus epithelium than in renal tubular cells by the criteria of immunoblot and mRNA level. A more speculative but attractive possibility is that the COOH-terminal amino acids of AE2 might be held in a different conformation or be "masked" in a tissue-specific manner. Such tissue-specific altered conformation could be achieved by interaction of AE2 with different sets of polypeptides or by differential covalent modification of AE2. Epitope-unmasking has been previously observed with the COOH-terminal epitope of the insulin-responsive glucose transporter, GLUT-4, in isolated adipocytes treated with insulin (34). More recently, in transgenic animals overexpressing GLUT-4 in skeletal muscle, "epitope unmasking" by acute insulin treatment of the animal prior to tissue fixation dramatically increased GLUT-4 detection by immunofluorescence microscopy in skeletal muscle, in parallel with increased glucose transport in T-tubules, whereas GLUT-4 detected by immunoblot remained constant (37).

Epitope masking has been suggested as an explanation for the absence of AE1 staining in apical membranes of type B IC in the rabbit CCD and in support of the proposal that AE1 serves as the apical Cl/ exchanger in these cells (1). The proposal is based on filter-lift membrane fractionation studies of polarized, functional CCD cells grown on permeable supports after enrichment for peanut lectin binding, as well as on considerations of variable lipid environments of apical and basolateral plasma membrane domains (1). However, none of the anti-AE1 or anti-AE2 antibodies tested on SDS-pretreated semithin sections of rat kidney revealed an apical pattern of immunostaining. Thus this particular form of epitope unmasking does not provide support for the hypothesis that AE1 mediates apical Cl/ exchange in type B IC or in any other renal tubular epithelial cell type.

Functional implications of the intrarenal distribution of plasmalemmal AE2: inner and outer medulla. The localization of AE2 to the basolateral membrane of IMCD epithelial cells supports the hypothesis, previously based on studies of immortalized IMCD cells in culture, that IMCD plays an important role in terminal urinary acidification (33). In addition, Cl/ exchange has been implicated in the recently reported Cl secretory function of the IMCD in at least one cultured cell model (26). AE2 is ideally suited to these functions in this region of the nephron. The extremes of luminal and interstitial acidification to which the IMCD epithelial cells can be subjected, especially during antidiuresis (25), should inhibit or abolish AE2-mediated anion exchange (40). However, two distinct regulatory properties of AE2 should allow continued function in the IMCD: stimulation of transport activity by elevated tonicity (22) and by elevated NH⁺₄ concentration (21).

The basolateral localization of abundant AE2 in the epithelial cells of the MTAL corresponds to the increased mRNA expression in this nephron segment (10) and to the presence of vacuolar H⁺-ATPase in the apical membrane of MTAL. AE2 is similarly well suited to functioning in the MTAL, not only because of the elevated tonicity and NH⁺₄ concentration to which this segment can also be exposed but also because AE2, in contrast to AE1, is capable of participating in the regulatory volume increase (23) thought to be required of MTAL cells to adapt to the fluctuating osmolar environment of the medulla (20).

Functional implications of the intrarenal distribution of plasmalemmal AE2: macula densa. The discovery of AE2 in the basolateral membrane of epithelial cells of the macula densa at higher levels than in cells of the surrounding CTAL suggests for it a possible role in tubuloglomerular feedback. The proposed mechanisms by which the macula densa transmits a signal to the juxtaglomerular mesangium reflecting the NaCl load in the lumen have all reflected the ability of luminal bumetanide to inhibit that signaling. Thus, in addition to provoking synthesis and release of first and second messenger molecules, the transepithelial delivery of chloride itself to the juxtaglomerular mesangium has been proposed as a signal. In the context of this proposal, basolateral AE2 could contribute to the regulation of pH_i, Cl concentration, and volume in the macula densa cell. On a much more speculative note, basolateral AE2 is strategically situated to mediate possible chloride reuptake from the juxtaglomerular mesangium as part of modulation or termination of the hypothesized extracellular chloride signal of the extraglomerular mesangium.

Implications of AE2 epitope in the Golgi apparatus. The COOH-terminal epitope of AE2 is present not only in the plasma membrane of some cells but also in the Golgi apparatus of a range of cell types. However, unlike the plasmalemmal AE2 COOH-terminal epitope that is unmasked by SDS, the Golgi epitope is evident in untreated sections and is destroyed by SDS.³ The Golgi epitope further distinguishes itself in being competed equally effectively by either AE2 COOH-terminal peptide or AE1 COOH-terminal peptide, at concentrations that display isoform specificity for their respective plasmalemmal epitopes. However, none of the tested antibodies raised against AE1 epitopes produced this pattern of Golgi staining. This difference in COOH-terminal epitope reactivity associated with subcellular distribution is reminiscent of that displayed by GLUT-4 in intracellular organelles and in plasmalemma in two distinct tissues (34, 37).

The AE2 Golgi epitope differs from the colchicine-resistant plasmalemmal epitope also in the susceptibility of its localization to 6 h in vivo exposure to colchicine. The Golgi apparatus in many cell types is disrupted by microtubule disruption, leading often to dispersal of fragmented Golgi cisternae throughout the cytoplasm (28). Yet another difference between the two epitopes is the ability to detect the Golgi epitope in glutaraldehyde-fixed tissue, allowing ultrastructural localization not yet achieved in kidney for the plasmalemmal AE2 epitope.

It is possible that a novel or a previously discovered isoform of AE2 or of AE1 contributes either to chloride or sulfate transport across the Golgi apparatus. It is also possible that such an AE isoform might contribute to anchoring the lipid bilayer of the organelle to elements of the organellar cytoskeleton. Recently, a novel isoform of -spectrin has been localized to the Golgi in skeletal muscle and in kidney (7, 16). In addition, one or more ankyrin isoforms may fulfill a connecting function between the proposed spectrin/actin cytoskeleton and integral proteins of the organellar lipid bilayer (31). Anti-AE2 aa 1224-1237 has also detected a Golgi distribution of immunofluorescent staining in cell lines derived from a normal and cystic human biliary epithelium (29) and from normal rat parotid duct (A. K. Stuart-Tilley, D. M. Jefferson, S. P. Soltoff, and S. L. Alper; unpublished results). In addition, the same antibody to mouse AE1 COOH-terminal aa 917-929 that immunostained Golgi-like structures in ROS osteosarcoma cells (24) also stains Golgi-like structures in immortalized epithelial cells (36) derived from mouse MTAL (Alper, unpublished results). Identification of the AE2-related protein of the Golgi apparatus will require additional experiments.