Mitochondria-rich (MR) cells, also known as flask cells, are characteristic of amphibian skin epithelium. The principal cells that form a functional syncytium are engaged in active transport of Na⁺, whereas the individually intercalated MR cells are the site of activated Cl^? conductance (G^Cl) and differ in their structure and composition from principal cells. Most tight epithelia are heterocellular and contain a small population of cells intercalated among principal epithelial cells. These cells carry out various transport functions in these tissues. Two cell models have been proposed to explain the functions of intercalated cells in mammalian kidney and carbonic anhydrase (CA-rich) cells in turtle urinary bladder, both involved in acid–base regulation. These models are based on physiological measurements and structural features: protons-secreting cells are designated as subtype, whereas ? subtype cells secrete bicarbonate. The operation of these models is based on the opposite location of H⁺-ATPase and the anion exchange enzyme band 3 in apical and basolateral membranes of cells. Amphibian MR cells execute several functions that vary among species, of which only the activated G^Cl has been firmly established in some species. These cells are stained by silver at their apical pole and contain high amounts of CA. Although H⁺-ATPase and band 3 have been demonstrated in amphibian MR cells, their functional significance remains unclear. Skin epithelium of Xenopus exhibits poor transport properties, and specific functions have not been assigned to their MR cells. However, cellular structure and density of MR cells change in response to long-term acclimation in high NaCl solutions and in brackish water. Therefore, we applied histochemical and immunohistochemical techniques to investigate the properties of MR cells in Xenopus skin epithelium to gain further insights into their possible functions. It was found that MR cells in Xenopus respond selectively to different ionic conditions, and that more than one subtype of MR cell can be distinguished morphologically.

Xenopus laevis (Daudin) of both sexes, weighing 35–45 g, were obtained from a local supplier in Israel. The animals were kept in the laboratory in tap water at room temp (approximately 21°C), and were not fed during experiments. Ionic acclimation was achieved by immersing the animals in 50 mmol/l solutions of NaCl, KCl or distilled water (DW) for 10 days or longer. Pieces of skin were obtained from anaesthetized animals (conducted according to the regulations of the local ethical committee), and were fixed in neutralized Bouin overnight for immunohistochemical staining. After dehydration in a series of increasing alcohol concentrations, samples were embedded in paraffin, and sectioned (25–30-?m thick) for confocal microcopy (see below). Alternatively, samples of a few mm² were fixed briefly in paraformaldehyde (1% in 0.1 M cacodylate buffer, pH 7.3), and were washed in buffer of 0°C and dehydrated in graded cold acetone and infiltrated overnight in TECHNOVIT 7100 (Kulzer glycol methacrylate embedding kits, Heraeus Kulzer, Germany;. Polymerization was performed at 6°C to preserve enzymatic activity. Sections (2-?m thick) were dry cut using glass knives, then stretched on a water drop and air-dried at room temperature.

Silver staining was performed on fresh skin pieces, as described earlier. Briefly, the skin pieces were incubated in 0.25% AgNO₃ for 2 min, washed and exposed for 30 min to strong illumination. Isolated epithelium after treatment with 1 mg/ml collagenase for 60–90 min was exposed to 0.05% methylene blue (MB) on the serosal side for 5–10 min, and was then analyzed microscopically.

Indirect immunohistochemical localization and SDS gel electrophoresis of polytron-homogenized collagenase-treated split epithelium were used to detect band 3 and H⁺-ATPase in skin epithelium. Anti-band 3 polyclonal antibody was raised in our laboratory against human erythrocytes band 3, and anti-70 kDa segment of H⁺-ATPase monoclonal antibody was obtained from AstraZeneca R&D (Molndal, Sweden; kindly supplied by Dr. J. Mattson). Thin sections were incubated with the primary antibody (diluted 1:50) overnight in the cold, washed with PBS the following morning and then incubated for 1 h with a fluorescence-conjugated (either green, FITC or red, TRITC) anti-rabbit antibody (Sigma, St. Louis MO, USA). Sections were visualized and subjected to image processing using a Nikon-based confocal microscope (BioRad, Herts, UK). Protein was quantified using the Bradford assay, separated on PAGE-SDS (10%) and transferred to nitrocellulose membranes. Western blots were performed by application of specific antibodies to visualize protein bands by an enhanced chemiluminescent reaction.

Carbohydrate components were stained by the periodic acid-silver methenamine (PASM) reaction using the Marinozzi procedure. Alkaline phosphatase activity (PALK, EC 3.1.3.1) was detected according to, using AS-BI naphthol phosphate as substrate and Fast Violet B as coupling agent (Burstone method). Preincubation of the slices with levamisole (10 mM) was used to verify the specificity of the enzymatic reaction. Malic dehydrogenase (MDH, EC 1.1.1.37) was demonstrated using the tetrazolium salt procedure as described by; in this reaction NAD is employed as coenzyme. The specificity of the reaction was controlled by incubation without substrate. Carbonic anydrase (CA, EC 4.2.1.1) activity was detected using the modified cobalt phosphate method, according to Hansson, modified by Waldeyer and Häusler. The incubation medium was freshly prepared and the slides were continuously dipped and lifted from the medium to allow the exchange of CO₂. Visualization of cobalt precipitates in the sections was enhanced by incubation in (NH₄)₂S. Acetazolamide (0.1 mM) was employed to inhibit CA activity. Mayer’s hematoxylin solution was used for nuclear staining.

Electrophysiological measurements were performed in Ussing-type chambers as described elsewhere. The skin (A=0.5 cm²) was continually perfused on both sides, and was clamped to 0 mV. Conductance was determined intermittently by application of 5 mV potential pulses across the skin every 90 s.

Student t-test was used for statistical analyses.

Over 35 specimens of X. laevis were used in the present study in various conditions. Animals were readily acclimated in distilled water (DW) and NaCl (50 mmol/l) for extended periods of time, but acclimation in KCl was not always successful, with almost 50% mortality. Only healthy animals were used in the experiments. Each test, either histochemical or immunocytochemical was repeated at least four times and number of sections are indicated below.

The size and density of silver-stained MR cells in the skin of Xenopus are shown in Table 1, and were not influenced greatly by the acclimation conditions. Fig. 1 shows two samples (of more than 10 that were used in each condition) of silver-stained MR cells in each of the acclimation conditions. The image shows that the apical aspect of MR cells is stained and well delineated, and its diameter was on average 7–9 ?m, except for cells in the KCl acclimated toads, where two distinct populations were observed: one population with an apical surface diameter of 5±1 ?m and the other with a diameter of 10±2 ?m (significantly different at p<0.05; n=10). The shape of MR cells was assessed in isolated epithelia, using methylene blue (MB) loading. MR cells exposed to MB on the serosal side accumulated the dye, and were then silver stained at the apical side. Fig. 2 shows three consecutive optical sections at 10 ?m intervals in the Z-direction. The small apical poles of the cells are present at the top (first section), and the larger cell bodies are shown in the third section. Reconstruction of MR cells from these sections yields a bottle-like structure, with a length of approximately 25 ?m, a diameter of approximately 20 ?m, and an apical external (silver-stained) diameter of approximately 7 ?m.