INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE MAMMALIAN INTESTINE displays axial heterogeneity of anion transport function along three anatomic axes: the proximal-distal organ axis from duodenum to colon, the villus-crypt epithelial axis from lumen to serosa, and the apical-basolateral cellular axis. Cl/HCO₃ exchange, Cl/OH exchange, and/or other anion exchange functions have been reported throughout the gut along all three axes (26, 27, 30, 32, 33, 41-43). As in other cells, Cl/HCO₃ exchange is thought to contribute to the housekeeping function of intracellular pH (pH_i) regulation. In addition, a major portion of intestinal Na⁺ reabsorption across the apical microvillar membranes is mediated via coupled function of Na⁺/H⁺ exchangers with Cl/HCO₃ and Cl/OH exchangers. Whereas the former have been defined as NHE3 and NHE2 (34, 46), the molecular identities of the polypeptides that mediate the anion exchange functions remain uncertain. Apical Cl/HCO₃ exchange has also been proposed to contribute to cystic fibrosis transmembrane conductance regulator-dependent cAMP-stimulated HCO₃ secretion in duodenum (16, 36) and elsewhere in the gut.

The localization of the anion exchanger 2 (AE2) Cl/HCO₃ exchanger, in particular, has been controversial. The first description of ileal AE2 polypeptide presented immunoblot evidence for an apical localization in rabbit (12). However, AE2 polypeptide has been localized to basolateral plasma membranes in gastric parietal cells (39), choroid plexus epithelium (4), and kidney tubular cells of rat (3) and mouse (40) and in many other epithelial tissues as well. Moreover, the properties of Cl/HCO₃ exchange in rabbit ileal basolateral membrane vesicles, but not in apical membrane vesicles (26), resemble those of recombinant AE2 expressed in Xenopus oocytes (18, 19, 47) and in transfected mammalian cells (20, 23).

The AE2 gene has been found to encode at least four transcripts (AE2a, AE2b, AE2c1, and AE2c2) generated from at least three promoters. These transcripts encode AE2 polypeptides with three distinct NH₂-terminal amino acid (aa) sequences, AE2a, AE2b, and AE2c (3, 40, 45). Although these sequence differences have been proposed to regulate distinct steady-state subcellular localizations (45), evidence is lacking for the moment. Similarly lacking is evidence that the alternative amino termini lead to variation in regulation of anion exchange activity in situ or in heterologous functional expression systems.

Because genetic experiments in mouse are likely to be helpful in the resolution of some of the above questions, we have characterized expression of AE2 mRNA and protein in the mouse intestine. The studies presented here indicate that mouse AE2 1) is expressed in enterocytes and enteric smooth muscle, 2) is in greater abundance in colon than in more proximal segments of intestine, 3) appears predominately in the AE2a isoform, and 4) that, as detected by antibodies to two epitopes, enterocyte AE2 is detected almost exclusively in the basolateral plasma membranes, with three exceptions. First, in ileal enterocytes fixed under mild conditions, the immunostaining pattern of one AE2 epitope was not polarized. Second, a different AE2 epitope was localized to the subapical region of duct cells at the mouths of duodenal and jejunal glands. Third, AE2-related immunoreactivity was also present in Paneth cells.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials. All reagents with no further specification in the text were purchased from Sigma (St. Louis, MO), Sigma-Aldrich, Fluka (Deisenhofen, Germany) or Merck (Darmstadt, Germany) at molecular biology grade or the highest grade available.

RT-PCR. Mouse stomach, duodenum, ileum, jejunum, colon, and kidney were resected, and the mucosa of the gastrointestinal organs was scraped off. Total cellular RNA was prepared from mucosal and submucosal tissue and kidney cortex with the use of guanidinium isothiocyanate and phenol-chloroform (Appligene, Heidelberg, Germany) extraction as described (11). RNA integrity was confirmed in all preparations by ethidium bromide staining and visualization of rRNA separated on glyoxal agarose gels.

Relative quantitation of AE2 polypeptide in intestinal tissues. Clarified NP-40 lysates of mouse duodenum, jejunum, ileum, and colon (<3 mg protein) were subjected to immunoprecipitation with 2-µl ascites containing monoclonal antibody to the COOH-terminal 12 residues of mouse AE2 (48), enough to precipitate the maximal precipitable amount of AE2 from NP-40 lysates of mouse stomach (3-5 mg protein, not shown).

Tissue preparation. Adult male CD1 mice were maintained on a standard diet with free access to water. Animals anesthetized with diethyl ether or methoxyflurane (MetoFane, Pittman-Moore) underwent cardiac perfusion for 2 min with 140 mM NaCl-20 mM sodium phosphate, pH 7.4 (PBS). Some mice were then perfusion-fixed with 2% paraformaldehyde-75 mM lysine-10 mM sodium periodate (PLP) as previously described (3, 4, 38-40). Other mice were perfusion-fixed with 2 or 3% paraformaldehyde alone. Perfused tissues were excised, cut into smaller lumen-exposed segments, further incubated in the same fixation media (overnight for PLP-fixed tissue and between 2 and 20 h for paraformaldehyde-fixed tissue), then stored until further use at 4°C in PBS containing 0.02% sodium azide.

Immunocytochemistry. Fixed tissue blocks were infiltrated with 30% sucrose in PBS-azide, frozen in liquid nitrogen, and sectioned at 5- to 7-µm thickness on a Reichert Frigocut cryostat. Tissue sections were placed on Superfrost/Plus microscope slides (Fisher) and stored in PBS-azide at 4°C until use.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

AE2 mRNA expression in mouse intestine. The AE2 gene is transcribed from at least three promoters, generating AE2a, AE2b, and AE2c transcripts (45). RT-PCR analysis of mouse colon RNA (Fig. 1) revealed the presence of AE2a (lane 1) and AE2b transcripts (lane 3). Both AE2c1 (RT-PCR band of 873 bp) and AE2c2 transcripts (band of 1161 bp) were present in mouse colon at very low levels (lane 5), in contrast to mouse kidney, which expressed only AE2c2 (40), and mouse stomach (lane 6), in which AE2c1 greatly exceeded AE2c2 (40). Figure 2 shows examples of semiquantitative RT-PCR analysis of AE2a, AE2b, and AE2c1 transcripts from mouse gastric mucosa RNA. Similar analysis comparing RNA from gastric mucosa with RNA from mouse intestinal tissues is presented in Fig. 3.

View larger version (81K): [in this window] [in a new window]   Fig. 1.   RT-PCR analysis of anion exchanger 2 (AE2) mRNA variants expressed in mouse colonic mucosa. A: ethidium bromide (EtBr)-stained AE2 amplification products (35 cycles) from mouse colonic mucosa RNA (lanes 1, 3, and 5) and from mouse gastric mucosa RNA (lanes 2, 4, and 6). Lane 5 is loaded with 8 reaction volume equivalents compared with lanes 1-4 and 6. Amplimers of 18S rRNA (lane 7, colon, and lane 8, stomach) underwent 24-cycle amplification. B: cDNAs in A gel were transferred to nylon and hybridized with a 32P-labeled internal AE2 oligonucleotide probe encoding a sequence present in all AE2 variants.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Semiquantitative RT-PCR analysis of AE2 mRNA variants in mouse gastric mucosa. Change in fluorescence intensity of EtBr-stained amplimers of AE2a (A), AE2b (B), and AE2c1 (C) as a function of amplification cycle number, compared in each case with that of 18S rRNA. ODI, integrated optical density units.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Comparison of relative expression levels of AE2a (A), AE2b (B), and AE2c1 (C) mRNA in mouse gastric and intestinal tissues, normalized to expression of 18S rRNA. Each bar represents the mean ± SE of at least 3 experiments.

AE2 polypeptide expression in mouse intestine. Anti-AE2 aa 1224-1237 displayed specificity as an AE2 immunoblot reagent (Fig. 4A) for both whole colon lysate (lanes 1 and 2) and for monoclonal anti-AE2 immunoprecipitate from the same lysate (lanes 3 and 4). Although the antibody used in these conditions did not precipitate all solubilized AE2 (lane 3 vs. lane 7), ECL signal from every tissue was nonetheless proportional to the input protein in the immunoprecipitation (Fig. 4A, bottom). NP-40 extraction of mouse intestinal tissues, with or without the further incubation associated with the immunoprecipitation procedure, led to SDS-resistant oligomerization of AE2. This finding suggests that the oligomeric state of intestinal AE2 resembles that of AE2 in pig gastric mucosa (5). With AE2 abundance in ileum valued at 1.0, AE2 relative abundance in colon (2.03 ± 0.11) was significantly higher (ANOVA with Tukey's all-pairs comparison) than abundances in jejunum (0.95 ± 0.11), in duodenum (0.91 ± 0.11; n = 3 for each), and in cecum (1.13; n = 1) (Fig. 4B). AE2a and AE2b polypeptide relative abundance in stomach was >7.0 (not shown). Anti-AE2 aa 109-122 was not usable as an immunoblotting reagent (3, 40).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Immunoblot analysis of AE2 polypeptide abundance in mouse intestine. A, top: specificity of AE2 detection by immunoblot with polyclonal anti-AE2 amino acid (aa) 1224-1237 in presence of irrelevant peptide (lanes 1, 3, and 5-7) or of peptide antigen (lanes 2 and 4). Lanes 1 and 2 were loaded with NP-40 lysate of whole colon (100 µg protein), lanes 3 and 4 with immunoprecipitate pellet obtained from 250 µg colon lysate protein with use of monoclonal anti-AE2 aa 1224-1237, and lane 7 with residual supernatant (60 µg). Lane 5 shows the precipitation pellet (from 250 µg protein) with use of nonimmune IgG and lane 6 with the residual supernatant (60 µg). A, bottom: polyclonal immunoblot of monoclonal AE2 immunoprecipitates from indicated amounts of NP-40 lysate (µg protein) from colon, ileum, jejunum, and duodenum; ** and * indicate AE2 dimer and monomer, respectively. B: ln/ln plot of pixel intensity (optical density, OD) of AE2 dimer vs. input protein (µg) for immunoprecipitation reaction, as portrayed in A, bottom. Immunoprecipitations were performed with detergent extract from membrane fractions of colon (), ileum (), jejunum (), and duodenum (). C: abundance of AE2 polypeptide in intestinal segments, normalized to AE2 abundance in ileum. Except for cecum (n = 1), data are depicted with mean values, 75%, and 95% confidence limits (n = 3).

Immunolocalization of AE2 polypeptide in intestine. Figure 5 confirms the specificity of anti-AE2 aa 1224-1237 as an immunocytochemical reagent in semithin cryosections of mouse colon. This antibody, after preincubation with the irrelevant peptide AE2 aa 109-122, detected not only AE2, evident in the mucosal enterocytes, but also crossreacted with AE1 in the trapped red blood cells (Fig. 5a), both of which signals are abolished in the presence of excess peptide antigen (Fig. 5b). However, whereas antibody preincubated with the crossreactive peptide AE1 917-929 completely lost reactivity with red blood cells, substantial immunoreactivity was retained with colonic mucosa (Fig. 5, c and d). Because this AE1 COOH-terminal peptide used in the preincubation also partially depletes immunoreactivity of this antibody tested against recombinant AE2, and because multiple red blood cell-reactive anti-AE1 antibodies did not immunostain colon (not shown), this reduced intensity signal is consistent with the presence of AE2 in colonic mucosa of mouse, as is also likely the case in rat (30).

View larger version (83K): [in this window] [in a new window]   Fig. 5.   Specificity of polyclonal anti-AE2 aa 1224-1237 as immunocytochemical reagent for colon. Anti-AE2 aa 1224-1237 previously depleted with agarose beads carrying either AE2 peptide aa 109-122 (a and b) or with AE1 COOH-terminal peptide aa 917-929 (c and d) were incubated with semithin sections of overnight 2% paraformaldehyde-75 mM lysine-10 mM sodium periodate (PLP)-fixed mouse colon in the presence of irrelevant peptide (a and c) or peptide antigen (b and d). rbc, Red blood cells. Bar = 50 µm.

View larger version (85K): [in this window] [in a new window]   Fig. 6.   Colocalization of 2 AE2 epitopes in mouse colon enterocytes. Anti-AE2 aa 109-122 (a, c, e, and g) or anti-AE2 aa 1224-1237 (b, d, f, and h) were incubated with cryosections of tissue fixed 2 h in 3% paraformaldehyde, in the presence of peptide antigen (c and d) or irrelevant peptide [a; b; c, inset (magnified from a); d, inset (magnified from b); e-h]. Distal colon (a-d), proximal colon (e and f), cecum (g and h) are shown. ap, Apical membrane; bl, basolateral membrane. Bar = 50 µm.

View larger version (98K): [in this window] [in a new window]   Fig. 7.   Comparison of AE2 expression in colon, ileum, and upper jejunum. Cryosections of colon (a), ileum (b), and duodenum (c) were immunostained with anti-AE2 aa 1224-1237. s, Luminal surface; cr, crypt(s); v, villus; gl, glands. Bar = 50 µm.

Exceptions to basolateral immunolocalization of intestinal AE2 polypeptide. The first two exceptions to basolateral localization of AE2, presented in Fig. 8, are restricted to the epitope detected by anti-AE2 aa 109-122. Figure 8, a, c, e, and g, show prominent staining of ileal Paneth cells in PLP-fixed tissue, whether Epon-embedded (Fig. 8, a and b) or frozen (Fig. 8, c and d). The AE2-related epitope was present in the plasma membrane but more abundantly in the large secretory granules. In ileum fixed 2 h with 2% paraformaldehyde, the Paneth cell staining was rather faint (Fig. 8e), but, after 24-h fixation, it was prominent in the granule content (Fig. 8g). Paneth cell immunostaining in all these preparations was competed completely by peptide antigen (Fig. 8, b, d, f, and h; Fig. 9, e and f).

View larger version (82K): [in this window] [in a new window]   Fig. 8.   Fixation condition dependence of AE2 epitope aa 109-122 in ileum. Etched Epon sections (1 µm, a and b) or 5-µm cryosections (c-h) from jejunum were fixed overnight (a-d) or fixed for 2 h (e and f) or 24 h (g and h) in 2% paraformaldehyde, in presence of irrelevant peptide (a, c, e, and g) or peptide antigen (b, d, f, and h); e and f are underexposed compared with other panels. sm, Smooth muscle. Bar = 50 µm.

View larger version (90K): [in this window] [in a new window]   Fig. 9.   AE2 epitope in Paneth cells. Cryosections of two jejunal crypts are shown, one viewed in phase (a) and the second stained with Dolichos biflorus agglutinin (b). Sequential sections of each crypt are immunostained with anti-AE2 aa 109-122 in presence of irrelevant peptide (c and d) or of peptide antigen (e and f). Bar = 50 µm.

View larger version (92K): [in this window] [in a new window]   Fig. 10.   AE2 epitopes in duodenum and upper jejunum. PLP-fixed cryosections of duodenum (sequential sections a and b) and upper jejunum crypt (c, e, and f) or villus (d) were immunostained with anti-AE2 aa 1224-1237 (a-d) or with anti-AE2 aa 109-122 (sequential sections e and f) in presence of irrelevant peptide (a, c, d, and e) or peptide antigen (b and f). Triple arrow in c demarcates basal surfaces of crypt or gland neck epithelial cells. Bar = 50 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have examined expression of AE2 in mouse intestine to provide a context in which to interpret studies of intestinal physiology in wild-type and mutant mouse strains. AE2 gene products were analyzed by RT-PCR, immunoblot, immunoprecipitation, and immunocytochemistry. Total AE2 mRNA was most abundant in colon and was of intermediate abundance in ileum and duodenum. In these regions, AE2a mRNA was approximately two- to threefold more abundant than AE2b with respect to rRNA in colon, ileum, and duodenum. Total AE2 mRNA was least abundant in jejunum, in which AE2a and AE2b were expressed at equal levels (Fig. 3).

AE2 polypeptide detected by semiquantitative immunoblot (AE2a + AE2b) migrated as an SDS-resistant dimer (5) and was approximately twofold more abundant in colon than in cecum, ileum, jejunum, or duodenum (Fig. 4). Because the AE2c mRNAs encode a polypeptide ~20 kDa shorter than those encoded by AE2a and AE2b mRNAs, the finding that the predominant gastric mucosal AE2 mRNA transcript is AE2c1 (Fig. 3) suggests that the reproducible observation of 165- and 145-kDa AE2 polypeptides in stomach of multiple species (39, 48, 49) reflects the translation of multiple AE2 transcripts more than it does proteolytic processing or degradation.

Immunolocalization studies with antibodies to two AE2 epitopes detected AE2 exclusively in basolateral membranes of colonic surface enterocytes, with low levels in colonic crypt enterocytes. In contrast, AE2 immunostaining intensity was comparable in villus and crypt enterocytes of ileum, jejunum, and duodenum. AE2 abundance in surface enterocytes declined from colon to duodenum. This basolateral localization conformed to the patterns previously detected for AE2 in other epithelial tissues (3, 4, 8, 14, 15, 29, 38-40). Basolateral AE2 likely contributes to pH_i regulation and to cell volume regulation during transepithelial reabsorption of NaCl, and potentially contributes in ileal crypts to Cl secretion by basolateral Cl loading.

AE2 or a protein carrying a closely related epitope was also detected in four locations previously undescribed or unsubstantiated. The first was in the mural smooth muscle of the gut, in both muscularis mucosae and in thicker outer layers. AE3, previously observed in vascular smooth muscle (7), is also present in intestinal smooth muscle (not shown). Although Cl/HCO₃ exchange is likely to contribute to regulation of smooth myocyte pH_i and excitability, regulation of pH_i in intestinal smooth muscle has been little studied.

The second novel location was in surface and internal membranes of Paneth cells in duodenum, jejunum, and ileum, detected by antibody to only one AE2 epitope, aa 109-122. Paneth cells secrete antibiotic secretagogue cryptdins, as well as other protein components of the innate immune system (28). Lectin costaining suggested that some of the AE2 epitope resided in granule membranes but could not rule out its presence in granule matrix. AE2 or a related protein in Paneth cells may facilitate protein exocytosis at the granular membrane or cell surface. However, despite the operational immunospecificity of Paneth cell AE2-related immunostaining, interpretation of the Paneth cell AE2 epitope requires caution. This is especially true in view of inability to stain Paneth cells with anti-AE2 aa 1224-1237 and the reputation of Paneth cells to stain nonspecifically with a wide range of in situ and immune reagents.

In lightly fixed tissue, anti-AE2 aa 109-122 detected immunostaining of both basolateral and apical membranes of ileal enterocytes in both villi and crypts. In both membranes the epitope appeared equally susceptible to paraformaldehyde concentration, exposure time, and postfixation storage. However, anti-AE2 aa 1224-1237 detected no apical AE2 immunostaining in mouse ileum, consistent with immunoblot analysis of rabbit ileum (32) and duodenum (33), in which AE2 detected by antibody to aa 1224-1237 was present only in basolateral membrane vesicles and not in apical vesicles.

Two previous studies have presented evidence of AE2 in apical membranes of polarized epithelial cells. In the first, guinea pig polyclonal antibody raised against a glutathione S-transferase fusion protein containing AE2 aa 396-499 detected a 160-kDa band in immunoblots of rabbit ileal brush-border membrane vesicles but not basolateral membrane vesicles (12). Neither specificity tests with recombinant AE3 nor use in immunocytochemical localization was reported. The AE2 region used as antigen consists of two portions: the COOH-terminal aa 433-499 share only 16% identity with rabbit AE3 (AF031650) but the internal aa 396-432 share 69% identity and 86% similarity with rabbit AE3. Transcripts and/or polypeptides encoding alternative AE3 transcripts are expressed throughout gut mucosae of rat (22), human (42), and mouse (unpublished observations).

In the second instance, a mouse monoclonal IgM raised against the species-specific human AE2 peptide sequence aa 882-895 (25), located in the largest exofacial loop between the fifth and sixth putative transmembrane spans and containing an N-glycosylation site at its NH₂-terminal end (48), immunostained apical canalicular membranes of hepatocytes and small bile ducts in human liver (24) and in both apical and basolateral membranes of human parotid gland interlobular ducts (44). Gastric parietal cells, choroid plexus (24), and salivary gland striated ducts (44) all exhibited basolateral immunostaining patterns with this antibody.

Unlike the apical AE2 epitope detectable only by anti-AE2 aa 109-122, the subapical/apical AE2 epitope detected in the necks of Brunner's gland-like structures in the duodunum and upper jejunum was detected only by anti-AE2 1224-1237. Brunner's glands have been thought to secrete HCO₃, although their abundance is not correlated with secretory rate (1). A subapical/apical pole AE2 or AE2-related polypeptide could mediate HCO₃ secretion, perhaps enhancing glandular secretion of lysozyme, defensins, and mucins. In contrast, the same antibody detected only basolateral AE2 immunostaining in the deep, branching portion of these glands, where it presumably contributes to Cl secretion.

This is the first subapical/apical pole distribution described for an epitope detected by anti-AE2 1224-1237. This epitope (as well as that detected only by anti-AE2 aa 109-122 in ileal apical membrane) may represent a novel AE2 or AE1 isoform, a product of a novel AE-related gene, or cell type-specific sorting or retention of either AE2a or AE2b. Study with additional antibodies to colinear AE2 epitopes will be required to resolve these possibilities.

Although some form of epitope-shielded AE2 polypeptide may plausibly contribute to apical Cl/HCO₃ exchange in enterocytes, rabbit ileal brush-border Cl/Cl exchange displays the lack of osmotic stimulation and shallow pH dependence (26) characteristic of recombinant AE3 expressed in Xenopus oocytes (unpublished observations). This pattern differs from the steep pH dependence and hyperosmotic activation of Cl/Cl exchange in gastric parietal cell basolateral vesicles (26) rich in AE2 polypeptide (5, 48, 49) and in Xenopus oocytes expressing AE2 (18, 19, 47).

A major portion of apical NaCl reabsorption in ileum and proximal colon is mediated by or requires the downregulated in adenoma (DRA) gene product (17, 37). Homozygosity for mutations in this protein is associated with congenital Cl-losing diarrhea. Although shown thus far only to mediate sulfate flux across the plasma membrane (9, 37), DRA likely mediates or potentiates Cl/HCO₃ or Cl/OH exchange. Homology of DRA with other putative sulfate transporters also suggests the possibility that one or more of those may also mediate or potentiate apical Cl/HCO₃ exchange in some portions of gut. However, the DRA-related transporter, Sat-1, prefers sulfate to Cl for exchange with bicarbonate or oxalate and is basolaterally, not apically, situated in renal proximal tubular epithelial cells (21).

Further understanding of the contribution of Cl/anion exchange to intestinal HCO₃ secretion and NaCl reabsorption will require identification and localization of additional AE anion exchangers, clarification of the function of putative anion exchangers from different gene families, and application of that knowledge to appropriate mouse knockout models. Development of pharmacological inhibitors of anion exchange more selective than those currently available (2) would accelerate the pace of progress.