Subunit composition and role of Na⁺,K⁺-ATPases in adrenal chromaffin cells

Abstract

Adrenal medullary (AM) cells are exposed to high concentrations of cortical hormones, one of which is a ouabain-like substance. Thus, the effects of ouabain on catecholamine secretion and distribution of Na⁺,K⁺-ATPase α and β subunits in rat and guinea-pig AM cells were examined using amperometry and immunological techniques. While exposure to 1 μm ouabain did not have a marked effect on resting secretion, it induced an increase in secretion due to mobilization of Ca²⁺ ions that were stored during a 4 min interval between muscarine applications. Immunocytochemistry revealed that Na⁺,K⁺-ATPase α1 subunit-like and β3 subunit-like immunoreactive (IR) materials were distributed ubiquitously at the cell periphery, whereas α2- and β2-like IR materials were present in restricted parts of the cell periphery. The α1 and α2 subunits were mainly immunoprecipitated from AM preparations by anti-β3 and anti-β2 antisera, respectively. Peripheral BODIPY-FL-InsP₃ binding sites were localized below membrane domains with α2- and β2-like IR materials. The results indicate that in AM cells, α1β3 isozymes of Na⁺,K⁺-ATPase were present ubiquitously in the plasma membrane, while α2β2 isozymes were in the membrane domain closely associated with peripheral Ca²⁺ store sites. This close association of the α2β2 isozyme with peripheral Ca²⁺ store sites may account for the facilitation of mobilization-dependent secretion in the presence of 1 μm ouabain.

Na⁺,K⁺-ATPase is a heterodimer consisting of one α and one β subunit; four α isoforms and three β isoforms are known (Blanco & Mercer, 1998). The α subunits are 112 kDa proteins and have ATPase activity with binding sites for Na⁺, K⁺, and ouabain. The expression of the α1, α2 and α3 subunits was found to be developmentally regulated in a tissue-specific manner (Wetzel et al. 1999), whereas the α4 subunit was expressed only in the testis (Shamraj & Lingrel, 1994). The α1, α2 and α3 subunits in the rodents can be divided into low (α1) and high affinity (α2 and α3) subunits, based on the affinity for ouabain (Blanco & Mercer, 1998). In arterial myocytes and astrocytes, the high and low affinity subunits were found to be distributed differently in the plasma membrane and suggested to have different functions (Juhaszova & Blaustein, 1997). On the other hand, the β subunits are glycoproteins with molecular weights of 40–60 kDa (Blanco & Mercer, 1998). This subunit is known to modify affinities for Na⁺, K⁺ and ouabain of the α subunits (Crambert et al. 2000). Furthermore, in autosomal dominant polycystic kidney disease, the Na⁺,K⁺-ATPase with the α1 isoform was mislocalized in the apical membrane domain, due to the dimer formation with the β2 subunit and the β1 subunit, the usual partner of α1 in the kidney, was confined to the endoplasmic reticulum (ER) (Wilson et al. 2000). Thus, β subunits may participate in some targeting of Na⁺,K⁺-ATPases to membrane domains. To elucidate the functional roles of specific heterodimeric Na⁺,K⁺-ATPases in cells, it is important to identify not only α isoforms, but also β isoforms in the heterodimers.

The Na⁺,K⁺-ATPase in adrenal medullary (AM) cells may be a target for an ouabain-like substance (OLS), which may be ouabain itself (Boulanger et al. 1993; Schneider et al. 1998) or its stereoisomer (Tymiak et al. 1993). First, high concentrations of OLS were detected in the adrenal gland of humans, cows and rats. Second, in rats, the OLS concentration in adrenal glands was highest among the peripheral tissues and brain, and that in the rat whole adrenal was 500-fold higher than in plasma (Hamlyn et al. 1991). Finally, angiotensin II enhanced secretion of OLS from bovine adrenal cortical cells in culture through the angiotensin II receptor type 2 (Laredo et al. 1997). It is generally thought that angiotensin II, which is generated in response to a decrease in renal blood flow and the subsequent secretion of renin, induces secretion of aldosterone with the consequent increase in blood volume. If angiotensin II induces secretion of OLS under such conditions, then catecholamine secretion from AM cells is probably enhanced through the inhibition of Na⁺,K⁺-ATPase. Catecholamine secreted in this way is expected to increase the cardiac output and renin secretion through β adrenergic receptors (Kern, 1993), thereby ameliorating the reduction in renal blood flow. In spite of the possible importance of secretion of catecholamine in response to OLS, the specifics of Na⁺,K⁺-ATPase isozymes present in AM cells remain unknown. The aims of the present experiment therefore are (i) to clarify the composition of Na⁺,K⁺-ATPases in AM cells, and (ii) to elucidate the functional implication of the presence of multiple isozymes.

Methods

Guinea-pigs and rats weighing 200–300 g were used. All experimental procedures involving animals were approved by the Institutional Animal Care and Use Committee of University of Occupational and Environmental Health.

Immunoblot

The animals were killed by cervical dislocation and the adrenal glands were excised and immediately put into ice-cold Ca²⁺-deficient balanced salt solution (in which 1.8 mm CaCl₂ was simply omitted from standard saline composition, see below). The adrenal cortex was removed from the adrenal gland using microscissors and forceps under stereoscopic observations. Adrenal medullae were homogenized in 10 volumes of 10 mm Tris-HCl, pH 7.4, 150 mm NaCl, and a proteinase inhibitor cocktail (1 mm 4-(2-aminoethyl)benzenesulphonyl fluoride, 0.3 μm aprotinin, 2 μm E-64, 1 mm EDTA, 2 μm leupeptin: Calbiochem). Homogenates were spun down at 2000 g for 10 min at 4°C, then the post-nucleus supernatants were mixed with equal volumes of the Laemmli sample buffer containing 250 mm Tris-HCl, pH 6.8, 4% SDS and 20% glycerol. Protein concentrations in samples were determined using a BSA protein assay kit (Pierce). Just before electrophoresis, 5% (v/v) 2-mercaptoethanol and 1% (w/v) bromophenol blue were added to the sample, and proteins were separated by 10% (w/v) SDS-PAGE, then transferred to a poly vinilidene difluoride (PVDF) membrane (Terry-Lorenzo et al. 2000). The membrane was blocked with 5% (w/v) fat-free powdered milk dissolved in 10 mm Tris-HCl, pH 7.4, and 150 mm NaCl and then incubated with a mouse anti-Na⁺,K⁺-ATPase α1 subunit monoclonal antibody (mAb) (Pietrini et al. 1992) (05-369; Upstate Biotechnology) (at a dilution of 1: 1000) for 1 h, a rabbit anti-α2 subunit Ab (Shyjan & Levenson, 1989) (06-168; Upstate) (1: 500) overnight, a rabbit anti-α2 Ab (anti-HERED Ab: Pressley, 1992) (1: 1000) for 1 h, a rabbit anti-α3 subunit Ab (Shyjan & Levenson, 1989) (06-172; Upstate) (1: 1000) overnight, anti-β1 mAb (05-382; Upstate) (1: 1000) for 1 h, rabbit anti-β2 Ab (06-171; Upstate) (1: 1000) overnight, or rabbit anti-β3 Ab (06-817; Upstate) (1: 1000) for 1 h. The α or β subunits were detected by incubating the membrane with a respective secondary Ab linked to horseradish peroxidase (Amersham), then with ECL-Plus (Amersham).

RT-PCR

Poly(A)⁺RNA was isolated from rat brain and adrenal medulla using the Micro-fast track kit (Invitrogen) according to the manufacture's instructions. The RNA yield was determined by measuring the specific absorption at 260 nm and the integrity of the RNA prepared was confirmed by agarose gel electrophoresis. Nucleotide sequences of rat Na⁺,K⁺-ATPase β1, β2 and β3 genes were obtained from BLAST (Accession numbers; NM_013003, NM_012507 and NM_012913). Nucleotide sequence of rat β-actin gene was also taken from BLAST (Accession numbers; NM_031144 and V01217). The nucleotide sequence of each primer for reverse transcription (RT) of Na⁺,K⁺-ATPase β-subunit genes was 5′-CCAATGTTCTCACCATACGC-3′ (β1; nucleotide number 1294–1313), 5′-GGCAGCGTTAATT-CGGC-3′ (β2; 1243–1259) and 5′-CCTGGTTCCATCAATTTGGC-3′ (β3; 894–913). These nucleotide sequences of RT-primers were derived from the sequences within the seventh (final) exon of β-subunit genes. A primer for RT of β-actin was 5′-GCTCAGTAACAGTCCGC-3′ (nucleotide number 3135–3151 of accession number V01217), based on the sequence of the 3′ untranslated region of the β-actin gene. Reaction of RT was performed in 10 μl of a reaction mixture containing 45 ng poly(A)⁺RNA, β-subunit and β-actin primer combinations (10 pmol each primer), 1 mm of dNTPs, 2.5 units of AMV reverse transcriptase, and RT buffers supplied with the kit (RNA LA PCR kit, Takara). RT was carried out for 25 min at 54°C, followed by an inactivation step at 90°C for 5 min. RT products were purified by using a micro spin column equipped with glass fibre matrix (Amersham), and purified DNA was suspended in 50 μl of water.

Nucleotide sequences of primers for PCR of β-subunit gene transcripts were 5′-GGAGTTTTTGGGCAGGACCGGTGG-3′ (β1 forward; nucleotide number 525–548), 5′-CCAGTGCACTGGACAGGTAGGACG-3′ (β1 reverse; 1083–1106), 5′-CCTCGTCTTCTATGGTTTCCTCACGGC-3′ (β2 forward; 593–619), 5′-CGTTCATGCTCTGGTTTGCCCCTGC-3′ (β2 reverse; 1029–1053), 5′-CGACCAGCGGAGAGTTTCTGGGG-3′ (β3 forward; 213–235) and 5′-GGAACTGACACGCACTATAGTCTGGACC-3′ (β3 reverse; 557–584). Primers for PCR of β-actin gene transcripts were 5′-CCTTCCTGGGTATGGAATCCTGTGGC-3′ (forward; nucleotide number 814–839 of accession number NM_031144) and 5′-CAGGAGGAGCAATGATCTTGATCTTCATGGTGC-3′ (reverse; 988–1020). The calculated sizes of RT-PCR products were 582 bp (β1 subunit), 461 bp (β2 subunit), 372 bp (β3 subunit) and 207 bp (β-actin). PCR was carried out with 1.25 μl of DNA template, 4 pmol of primers, 2 mm of dNTPs, 0.5 units of LA Taq, and PCR buffers supplied with the kit in a final volume of 20 μl. The PCR profile used started with an initial 2 min denaturation step at 94°C, followed by cycles of 30 s for denaturation at 94°C, 30 s of annealing at 62°C, and 75 s of extension at 72°C. To obtain maximum fidelity, hot-start conditions were applied, using ampliwax (Perkin-Elmer). In each PCR reaction, a 207 bp PCR product of β-actin mRNA was co-amplified and used as an internal standard. To ensure that β-actin mRNA would not reach the plateau phase earlier than the target gene, addition of β-actin primers was delayed, as described (Kinoshita et al. 1992). During the first six PCR cycles, only the target gene primers were present. After completion of the sixth elongation phase, the PCR reaction was halted and the reaction mixture was cooled to 4°C. Then, β-actin primers were added and the reaction was resumed, starting with a 2 min denaturation step at 94°C, followed by 23 cycles of the same PCR profile. Under these conditions both the target and the reference genes were amplified in the exponential range; thus, the relative amounts of β subunit mRNAs were measured as ratios of the β subunit to β-actin RT-PCR products. PCR products were separated by 2% agarose gel electrophoresis, stained with ethidium bromide and bands were scanned with FLA 2000 (Fujix), then quantified using NIH image (NIH, USA). Each experiment also contained two sets of negative controls containing RNA that had not been reverse transcribed or containing water instead of DNA. In all experiments the negative controls yielded no detectable products, indicating that the PCR products did not come from contaminating genomic DNA and that all reagents were free of target sequence contamination.

Immunoprecipitation

Brain (5 mg) and AM (2 mg) post-nuclear supernatants were solubilized in an immunoprecipitation buffer (10 μm deoxycholate, 150 mm NaCl, 10 mm Tris, pH 7.4, 1 mm PMSF, 10 μg ml⁻¹ leupeptin, 10 μg ml⁻¹ antipain and 10 μg ml⁻¹ pepstatin A) to bring the final protein concentration to 2 and 0.8 μg μl⁻¹, respectively. The sample was vortex mixed at 4°C for 15 min before sonication. The insoluble material was then pelleted by centrifugation at 12 000 g for 3 min at 4°C. The supernatant was removed, then pre-cleared with 50 μl of immobilized protein A (Amersham) for 15 min with vortex mixing at 4°C. The sample mixture was then centrifuged at 12 000 g for 3 min at 4°C. The supernatant was transferred to a fresh microcentrifuge tube, and 5 μg of anti-β1 Ab (06-170), anti-β2 Ab (06-171) or anti-β3 Ab (06-817) was added. The mixture was incubated on a rotating wheel for 3 h at 4°C. After this, 50 μl of immobilized protein A was added and mixed for an additional 4 h on a rotating wheel at 4°C. The mixture was then washed with 800 μl of washing buffer-1 (0.5% polyoxyethylene (9) octylphenyl ether (NP-40), 150 mm NaCl, 10 mm Tris, 1 mm EDTA, pH 7.5), and then centrifuged at 12 000 g for 3 min at 4°C. The wash procedure was repeated with buffer-1, and then three times with buffer-2 (0.5% NP-40, 500 mm NaCl, 10 mm Tris, 1 mm EDTA, pH 7.5). Immunoprecipitated Na⁺,K⁺-ATPase α subunits were dissociated from the protein A and Ab mixture by incubation in 45 μl of the Laemmli sample buffer for 30 min at 37°C. The samples were then subjected to SDS-PAGE, followed by Western blot analysis.

Immunocytochemistry

Adrenal medullae were cut into three to six pieces and incubated for 30 min with 0.25% collagenase dissolved in Ca²⁺-deficient solution. After incubation, tissues were washed in Ca²⁺-deficient solution, and placed in a dish with non-fluorescent glass at the central area (MatTek), then dissociated using fine needles. Dissociated AM cells were fixed in 2% paraformaldehyde in phosphate-buffered saline (PBS; pH 7.2) for 2 h at 4°C or in methanol at −18°C for 10 min, then pre-incubated in 5% goat or fetal bovine serum in PBS for 30 min (Inoue et al. 2000). For indirect immunofluorescence studies, cells were treated with the anti-α1 mAb (1: 100) overnight, the anti-α2 Ab (1: 200) for 2 days, the anti-α3 Ab (1: 250) for 2 days, the anti-β1 mAb (1: 100) overnight, the anti-β2 Ab (1: 100) overnight, the anti-β3 Ab (1: 100) overnight, an antiplasma membrane Ca²⁺ (PMCA) pump mAb (Borke et al. 1989) (804-049; Alexis) (1: 250) overnight, an anti-Na⁺–Ca²⁺ exchanger type 1 (NCX1) mAb (MAB1590; Chemicon; 1: 250) overnight, an anti-NCX1 mAb (Porzig et al. 1993) (R3F1; Swant; 1: 100) for 3 days, or a rabbit anti-calnexin Ab (SPA-860; StressGen; 1: 400) overnight. After the incubation, the cells were washed three times in PBS and then treated with a respective secondary Ab conjugated with fluorescein isothiocyanate (FITC) (goat anti-rabbit IgG; Cappel; 1: 25), rhodamine (goat anti-mouse IgG; Cappel; 1: 25), tetramethyl rhodamine isomer R (swine anti-rabbit IgG; DAKO; 1: 25), or Alexa 488 or 546 (goat anti-rabbit IgG; Molecular Probes; 1: 300). The fluorescence was observed using laser confocal microscopy (Zeiss LSM 410) (Inoue et al. 2000). The objective lens was an oil-immersion lens with a magnification of × 63 and a numerical aperture of 1.4. For FITC and Alexa 488, a 488 nm laser was used and 510–525 nm emission was observed, whereas for rhodamine and Alexa 546, a 543 nm laser was used and emission above 570 nm was observed. Fluorescence was observed with a full width at a half-maximum of 0.7 μm. Whole-cell images were acquired with illumination of the 488 nm laser and emission of all wavelengths, and the refractive index of the oil and immersion fluid were taken into account to measure cell size in a confocal image. Fluorescence images were quantitatively analysed using NIH image, and the cell shape or edge was determined by looking at staining intensity above background levels. To examine the specificity for the immunoreaction, the preparation was treated with a non-immune serum instead of a primary Ab, and almost no immunoreactivity was observed under the same conditions as used for a primary Ab.

Amperometry

Catecholamine secretion from dissociated guinea-pig AM cells was measured by amperometry, as described elsewhere (Inoue et al. 2002). After treatment with collagenase, the tissues of adrenal medullae were left in Ca²⁺-deficient balanced salt solution at 23–25°C until the experiments were begun. Prior to starting an experiment, one or two pieces of tissue were placed in the bath apparatus on an inverted microscope and AM cells were dissociated mechanically using fine needles. Chemicals were applied by addition to the bath. To measure catecholamine release from dissociated cells, a carbon-fibre electrode was carefully placed on the cell surface and +600 mV was applied to the electrode under voltage clamp conditions. The current due to oxidation of catecholamines at the tip of the electrode was stored on a data recorder and fed to a brush recorder after low pass filtering at 5 Hz. For quantitative analysis, the signals were low-pass filtered at 100 Hz and digitized at a sampling interval of 5 ms. The total charge of evoked currents was measured, and muscarine-induced secretion in the absence of external Ca²⁺ ions was obtained by subtraction of the secretion recorded before stimulation from that observed during 30–40 s stimulation. The standard saline contained (mm): 137 NaCl, 5.4 KCl, 1.8 CaCl₂, 0.5 MgCl₂, 0.53 NaH₂PO₄, 5 d-glucose, 5 Hepes, and 4 NaOH (pH 7.4). In nominally Ca²-free solution, 1.8 mm CaCl₂ in the standard saline was replaced with 3.6 mm MgCl₂. Data are expressed as means ±s.e.m.

Results

Identification of Na⁺,K⁺-ATPase α and β subunits

We first examined which α subunits were present in rat AM cells using immunoblotting. As shown in Fig. 1, the α1 subunit of approximately 110 kDa was detected in rat brain and AM homogenates and the amounts in the andrenal medulla was 47 ± 12% (n= 5) of that in the brain homogenate. In contrast, the amount of the α2 subunit detected immunologically with two different antisera in the andrenal medulla was 14 ± 3% (n= 7) of that in the brain. The α3 subunit of 110 kDa was detected in the brain, but not in the AM homogenate. In contrast to the α subunits, β subunits were present in several forms with molecular weights ranging from 44 to 52 kDa, attributable to N-glycosylation at several sites. An anti-β1 mAb recognized three bands with strong labelling of a 52 kDa protein in brain homogenates, whereas it mainly recognized a 46 kDa protein in the AM homogenate. An anti-β2 Ab labelled single proteins of 52 kDa in both homogenates. Compared with the anti-β1 and -β2 Abs, an anti-β3 Ab was less specific; it recognized not only three bands of 46, 49 and 52 kDa, but also bands with larger molecular masses in the brain homogenate, whereas the 52 kDa band was mainly labelled in the AM homogenate. Thus, the immunoblotting suggested the presence of the three β isoforms with different levels of glycosylation in the adrenal medulla.

Figure 1. Immunoblots for Na⁺,K⁺-ATPase α and β subunits of rat brain and adrenal medulla.

Immunoblots of adrenal medulla (A) and brain (B) homogenates with anti-α1 mAb, anti-α2 Ab, anti-α2 HERED Ab and anti-α3 Ab; anti-β1 mAb, anti-β2 Ab, and anti-β3 Ab. Arrowheads indicate β subunits with N-glycosylation. Proteins of 25 μg were separated in SDS-PAGE and transferred to a PVDF membrane. The membrane was treated with each Ab for 1 h or overnight (see Methods).

This notion was confirmed at the mRNA level (Fig. 2). The RT-PCR of brain mRNAs with probes specific for the β1, β2 and β3 subunits resulted in bands of about 600, 460 and 370 bp, respectively, values which correspond to those expected for the β subunits. The signal levels for the β1, β2 and β3 subunits in the brain, which were expressed as fractions of those for β-actin measured at the same time, were 4.46 ± 0.43 (n= 3), 3.49 ± 0.43 (n= 3) and 2.29 ± 0.21 (n= 3), respectively, whereas relative amounts of the mRNAs encoding for the β1, β2 and β3 subunits in the adrenal medulla were 3.24 ± 0.30, 0.60 ± 0.10 and 2.20 ± 0.08, respectively. Thus, the amount of β2 mRNA in the adrenal medulla was markedly smaller than those of the β1 and β3 mRNAs.

Figure 2. RT-PCR for Na⁺,K⁺⁺-ATPase β subunits in rat brain and adrenal medulla.

RNAs were extracted from rat brain and adrenal medulla and RT-PCR was performed with probes which were specific for each β subunit (see Methods).

The distribution of α and β subunits in AM cells was then examined using confocal microscopy and immunocytochemistry (Fig. 3). The α1 subunit-like immunoreactive (IR) materials were present at the whole cell periphery, whereas the α2 subunit-like immunoreactivity occurred at only part of the cell periphery. In contrast to the α1 and α2 subunits, α3 subunit-like immunoreactivity was not consistently observed. To evaluate quantitatively the distribution of α subunit-like IR materials, the length of the membrane area with immunoreactivity was measured and expressed as a fraction of cell perimeter. The α1-like IR materials were distributed at 100% of the cell perimeter, whereas the α2-like IR materials were present at 19%. This compares with the β2 subunit-like immunoreactivity occurring at only 23% of the cell perimeter, whereas the β3 subunit-like IR materials were distributed at the majority (80%) of cell periphery. The treatment of either paraformaldehyde or methanol-fixed AM cells with the anti-β1 mAb did not produce any IR materials.

Figure 3. Immunocytochemistry for Na⁺,K⁺-ATPase α and β subunits in rat adrenal medullary cells.

A, confocal microscopic images of rat adrenal medullary (AM) cells immunostained for α1, α2, α3, β1, β2 and β3 subunits. Brightness was enhanced to show non-specific bindings in the cytoplasm and consequently cell profile. B, ratio of the membrane with immunoreactive materials to the cell perimeter. The length of the membrane with immunoreactive materials, measured by using NIH image, was expressed as a fraction of the cell perimeter.

If the different distribution of Na⁺,K⁺-ATPase α isoforms in AM cells has functional implications, it should be conserved in other animals. Thus, we examined Na⁺,K⁺-ATPase α subunits in guinea-pig AM cells. In agreement with the rat cell, the α1 and α2, but not α3, subunits were present in guinea-pig AM cells, and the former was distributed at the whole cell periphery and the latter was located at part of the cell periphery (Fig. 4).

Figure 4. Presence of Na⁺,K⁺-ATPase α1 and α2 subunits in guinea-pig AM cells.

A, immunoblots for α1, α2 and α3 subunits of guinea-pig adrenal medulla (A) and brain (B). Proteins of 25 μg were loaded for each lane. B, confocal microscopic images of guinea-pig AM cells immunostained for α1, α2 and α3 subunits. Brightness was enhanced to see cell profile.

Immunoprecipitation

The immunocytochemical studies, taken together with immunoblottings, suggest that the α1 and α2 subunits form heterodimers with β3 and β2, respectively. In addition, based on previous findings (Schmalzing et al. 1997), the α1 subunit might also form a complex with β1 in AM cells. These possibilities were examined using immunoprecipitation (Fig. 5). Treatment of brain post-nuclear preparations with anti-β1 and anti-β3 Abs resulted in the precipitation of α1, but not of α2, and the amount of the α1 precipitated with the anti-β3 Ab was about 54 ± 6% (n= 3) of that with the anti-β1, whereas the α2 subunit was precipitated only with the anti-β2 Ab. Treatment of AM preparations with the anti-β2 Ab also induced the precipitation of the α2 subunit alone, and that with the anti-β1 and -β3 Abs resulted in the precipitation of α1, but not α2. The amount of the α1 precipitated with the anti-β3 Ab was 430 ± 53% of that with the anti-β1 Ab. The results suggest that in AM cells, the α1 and α2 subunits mainly form dimers with β3 and β2, respectively.

Figure 5. Immunoprecipitation of Na⁺,K⁺-ATPase α1 subunit with anti-β1 and anti-β3 Ab and of α2 subunit with anti-β2 Ab.

Upper and lower panels, immunoblots of precipitated proteins for Na⁺,K⁺-ATPase α1 and α2 subunits. Post-nuclear supernatants of rat brain (BR) and adrenal medulla (AM) were mixed with anti-β1, anti-β2 or anti-β3 Ab (see Methods).

Close apposition of α2 with Ca²⁺ store sites

Our previous findings (Inoue et al. 2003, 2004) indicate that Ca²⁺ store sites with inositol 1,4,5-trisphosphate (InsP₃) receptors and/or ryanodine receptors are located not only in the vicinity of the nucleus, but also beneath the plasma membrane. Thus, we explored the spatial relationship between the α2 isozyme and peripheral Ca² store sites. For the identification of Ca²⁺ store sites, we used BODIPY-FL-InsP₃ (Ozaki et al. 2000; Prestwich, 2004), which is visible as non-bleachable FITC-like fluorescence. Most of the BODIPY-FL-InsP₃ binding sites at the perinuclear area and at the cell periphery (which was defined as a site near the edge of a cell; Fig. 6A) were also stained with ER Tracker (Inoue et al. 2003) and with an Ab against calnexin (Fig. 6B), which is an integral protein of the ER membrane (Wada et al. 1991). These results indicate that BODIPY-FL-InsP₃ binding sites represent the ER. This notion was further supported by a three-dimensional organization of InsP₃ binding sites. The circular InsP₃ binding sites were located between the nucleus and the plasma membrane and appeared to be separated from each other (Fig. 6C), while the two binding sites at the cell periphery and next to the nucleus were not directly connected (Fig. 6D). However, when the fluorescence images in the five consecutive optical sections from Fig. 6C to Fig. 6D were projected to one plane (Fig. 6E), the binding sites in Fig. 6C turned out to be continuous with those in Fig. 6D, suggesting that BODIPY-FL-InsP₃ binding sites are distributed in a reticular pattern characteristic of the ER.

Figure 6. Close association of Na⁺,K⁺-ATPase α2 isozyme with peripheral Ca²⁺ store sites.

A and B, confocal microscopic images of 30 μm BODIPY-FL-InsP₃ bindings and calnexin-like immunoreacive (IR) materials in the same optical section of a rat AM cell, respectively. BODIPY-FL-InsP₃ binding sites and calnexin-like IR materials were visualized as FITC-like and rhodamine-like fluorescence, respectively. The asterisk in this and the following microscopic images represents the nucleus. C and D, confocal images of 30 μm BODIPY-FL-InsP₃ bindings in two different optical sections of one rat AM cell. E, projection image of the five consecutive images from C to D. Each of the five consecutive images, which were obtained at a step of 0.7 μm, was transformed to an image that was expressed in black or white, and all of the images were projected to one plane. The arrowheads indicate BODIPY-FL-InsP₃ binding sites. F and G, confocal images of 30 μm BODIPY-FL-InsP₃ bindings and α2-like IR materials in the same optical section of a guinea-pig AM cell, respectively. BODIPY-FL-InsP₃ binding sites and α2-like IR materials were visualized as FITC-like and rhodamine-like fluorescence, respectively. Images were a superimposition of cell profile (white) and fluorescent InsP₃ binding sites (red) or α2-like IR materials (green). H, a superimposition of F and G. I and J, confocal images of 30 μm BODIPY0FL-InsP₃ bindings and β2-like IR materials in the same optical section of a rat AM cell, respectively. Fluorescent InsP₃ binding sites and β2-like IR materials were visualized as FITC-like and rhodamine-like fluorescence, respectively. Images were a superimposition of cell profile (white) and fluorescent InsP₃ binding sites (reddish purple) or β2-like IR materials (green). K, superimposition of I and J. Arrowheads show close association of BODIPY-FL-InsP₃ binding sites with α2- (H) and β2-like (K) immunoreactivity in the plasma membrane. All the calibration bars are 5 μm.

Figure 6F illustrates the BODIPY-FL-InsP₃ binding sites distributed at the perinuclear region and cell periphery. In contrast, α2-like IR materials, observed as rhodamine-like fluorescence in the same optical section, were exclusively present at part of the cell periphery (Fig. 6G). The overlay (Fig. 6H) of these FITC and rhodamine images clearly indicates that the two main sites with α2-like IR materials were juxtaposed to peripheral BODIPY-FL-InsP₃ binding sites. The number of peripheral BODIPY-FL-InsP₃ binding sites examined in one optical section in each of 16 cells was 3.7 ± 0.5, whereas that of the binding sites closely associated with plasma membrane regions with α2-like IR materials (which were regarded as α2 regions if representing more than 1% of the cell perimeter) was 2.7 ± 0.4. Thus, 73% of peripheral BODIPY-FL-InsP₃ binding sites were located beneath α2 membrane regions. Since the α2 subunit in the AM cell was proposed to be present as an α2β2 heterodimer, colocalization of the α2 isozyme of Na⁺,K⁺-ATPases with peripheral BODIPY-InsP₃ sites was further examined with the anti-β2 Ab. Double staining with fluorescent InsP₃ (Fig. 6I) and anti-β2 Ab (Fig. 6J) clearly revealed that peripheral InsP₃ binding sites were located directly beneath the membrane domains with β2-like IR materials (Fig. 6K). The number of peripheral InsP₃ binding sites examined in one optical section in each of 15 cells was 4.9 ± 0.5, whereas that of the binding sites closely associated with β2-like IR materials in the membrane was 3.3 ± 0.3.

Since Na⁺–Ca²⁺ exchange (NCX) was proposed to be co-localized with the α2 subunit in the plasma membrane of astrocytes or the α3 subunit in arterial myocytes (Blaustein et al. 1998), the distribution of NCX was examined. Figure 7A shows immunostaining with a mAb (MAB 1590), which was obtained using the NCX1 isoform of the rabbit as an antigen. The entire cell periphery was immunoreactive, although the intensity of staining was low. Hence, we further examined the expression of NCX using another mAb (R3F1), which is known to recognize the NCX1 isoform of rat and dog, but also the NCX2 at higher concentrations. Again, faint immunostainings were observed along the entire cell periphery (Fig. 7B). We next examined the distribution of PMCA pumps with a mAb that recognizes all four isoforms of the PMCA pump (Fig. 7C). Consistent with the result in arterial myocytes (Blaustein et al. 1998), PMCA pump-like IR was distributed at nearly the entire cell periphery.

Figure 7. Na⁺–Ca²⁺ exchanger and plasma membrane Ca²⁺ pump are present ubiquitously in the plasma membrane.

Confocal microscopic images of guinea-pig AM cells immunostained with NCX1 mAb (MAB 1590), NCX1 mAb (R3F1) and PMCA pump mAb, respectively (see Methods). All calibration bars are 10 μm.

Effects of ouabain on secretion

In order to elucidate the role of the α2 subunit in the secretory function of AM cells, effects of 1 μm ouabain on the resting and agonist-evoked secretion were examined using amperometry. We employed a protocol for stimulation in which 10 μm muscarine was first applied for 30–40 s in the absence of external Ca²⁺, followed by 10 μm nicotine for about 60 s in the presence of external Ca²⁺, and finally 10 μm muscarine for 30–40 s in the absence of external Ca²⁺. The protocol was repeated at intervals of 4 min. Thus, the first muscarine-induced secretion in the protocol is due to Ca²⁺ ions that are stored during the unstimulated intervals, whereas the second muscarine-induced secretion is attributed to Ca²⁺ ions that are stored fully during the nicotine stimulation (we previously determined that 1 min stimulation with nicotine was sufficient to store Ca²⁺ ions fully). Figure 8A shows amperometric records of catecholamine secretion from a dissociated guinea-pig AM cell. Most of the cells examined exhibited little spontaneous secretion activity, as shown in Fig. 8Ad, and addition of 1 μm ouabain to a perfusate resulted in a marginal enhancement of the spontaneous secretion (Fig. 8Ae). Such measurable amperometric spikes, as were evoked by muscarine or nicotine, did not develop during exposure to 1 μm ouabain in 10 cells, but did in application of 100 or 200 μm with a latency of 20–30 s in three cells (not shown). In contrast, application of 1 μm ouabain induced a time-dependent enhancement of the secretion due to mobilization of Ca²⁺ stored during a 4 min interval between muscarine applications (partially stored Ca²⁺), but did not affect the secretion due to that of fully stored Ca²⁺. In 5 of 7 cells tested with the protocol, the secretion evoked by the first muscarine stimulation was enhanced on a second to fourth application after exposure to ouabain (note, in three cells, the enhancement occurred on the second application and thereafter; in one cell, on the third application and thereafter; in one cell, only on the fourth application). These results are summarized in Fig. 8B, which reveals that the secretion due to Ca²⁺ ions stored during intervals increased successively after exposure to ouabain, whereas that due to Ca²⁺ fully stored during the nicotine stimulation was not altered noticeably. To reduce variability of the partially stored Ca²⁺-dependent secretion, it was expressed as a fraction of the fully stored Ca²⁺-dependent secretion in the same cells. Figure 8C shows that the values expressed thus significantly increased in the fourth stimulation after exposure to ouabain, indicating that exposure to 1 μm ouabain results in facilitation of Ca²⁺ uptake into store sites with a marginal effect on resting secretion.

Figure 8. Effects of ouabain on catecholamine secretion.

A, amperometric records of catecholamine secretion from a dissociated guinea-pig AM cell before and after exposure to 1 μm ouabain. Chemicals were bath applied during the indicated periods (single bar for 10 μm muscarine (M); double bar for 10 μm nicotine (N); dotted line for ouabain (Ouab)). External 1.8 mm Ca²⁺ ions were replaced with 3.6 mm Mg²⁺ ions during the indicated period (interrupted line). The protocol, consisting of 30–40 s application of muscarine, about 60 s application of nicotine, and 30–40 s application of muscarine, was repeated every 4 min. Traces b and c were interrupted for about 4 min. Traces d and e were recorded before a and between b and c, respectively. The calibrations are the same in a to e. B, relative amount of secretion evoked by muscarine-induced Ca²⁺ mobilization before and after application of 1 μm ouabain. Open and hatched columns represent secretions due to Ca²⁺ mobilizations evoked by first and second muscarine stimulation in the protocol. Numbers in the abscissa mean number of protocol application after the exposure to ouabain. Amounts of secretion were expressed as a fraction of each secretion before exposure to ouabain (Con). C, amount of secretion in response to Ca²⁺ mobilization evoked by first muscarine stimulation in the protocol was expressed as a fraction of secretion evoked by second muscarine stimulation. Numbers in the abscissa have the same meaning as in B. Statistical significance was determined using a Dunnett's test. Data represent mean ±s.e.m. (n= 7).

Discussion

Composition of Na⁺,K⁺-ATPase

Na⁺,K⁺-ATPases present in all cells in the mammalian body except the testis are heterodimers consisting of each of three α and β isoforms. Thus, there are theoretically nine combinations of the α and β subunits. In fact, when combinations of individual α and β subunits were expressed in oocytes (Crambert et al. 2000) or Sf-9 insect cells (Blanco & Mercer, 1998), all possible combinations of α and β subunits were found to function. However, immunoprecipitation experiments suggested the presence of α2β2 and possibly α2β3, but not α1β2 in the brain, although three α isoforms were present (Gloor et al. 1990). Moreover, detailed histochemical studies of mouse retina revealed that one or two of α and β subunits each were selectively expressed in one type of cells in a cell-specific manner. Thus, the combination of α and β subunits is probably regulated transcriptionally and at the protein level. Because the selective ablation of α and β subunits resulted in death of infants or functional disturbance even if newborns can survive (Magyar et al. 1994; James et al. 1999), the combination of α and β subunits may be specific in vivo (Weber et al. 1998). However, there have been few reports in which the composition of Na⁺,K⁺-ATPases is investigated biochemically and morphologically in cells with physiological functions.

The present experiments demonstrated that the α1 isoform was present ubiquitously in the plasma membrane, as noted in arterial myocytes (Juhaszova & Blaustein, 1997), whereas the α2 subunit was distributed at part of the membrane. The α3 subunit was not immunodetected either in AM homogenates or in AM cells. Since intracellular concentrations of Na⁺ (5–20 mm; Sorimachi et al. 1994) in bovine AM cells approximated the EC₅₀ value (9–12 mm) of the α2 subunit for Na⁺ rather than 17–28 mm of the α3 subunit, the presence of α2, but not α3, in the AM cell would seem reasonable. On the other hand, isoforms of β subunits present in AM cells were not consistent when evaluated by immunoblotting, RT-PCR or immunocytochemistry. Immunoblotting revealed that in the adrenal medulla preparation all of the three isoforms were present with glycosylation levels that differed among the β subunits and between brain and AM. These findings agreed with the results obtained with RT-PCR, which showed that the levels of mRNAs encoding for β1 and β3 were comparable between adrenal medulla and brain, while the amount of the β2 mRNA in adrenal medulla was much smaller than that in the brain. Immunocytochemistry showed that β2-like IR materials were localized only to regions of the plasma membrane, whereas β3-like IR products were distributed throughout the majority of the cell membrane and in the cytoplasm. In contrast to the results with immunoblotting and RT-PCR, however, the β1 subunit could not be detected immunologically in either paraformaldehyde- or methanol-fixed AM cells. Because T-tubules of isolated rat ventricular myocytes were labelled with the same anti-β1 mAb (data not shown), this failure may not be due to the absence of reactivity of the mAb in immunocytochemistry, but to low expression levels of the β subunit. The amount of β1 mRNA in the adrenal medulla detected with RT-PCR might be at least in part ascribed to cells other than AM cells, such as endothelial or smooth muscle cells.

Immunocytochemistry suggested that the α1 subunits mainly form heterodimers with β3 subunits, while the α2 subunits heterodimerize with β2 subunits. This was confirmed by immunoprecipitation experiments. The α2 isoform was precipitated from brain and AM preparations only by the anti-β2 Ab. On the other hand, the α1 subunit was precipitated from both preparations by anti-β1 and anti-β3 Abs, and the amounts of the α1 precipitated with the anti-β3 Ab from brain and AM preparations were 54 and 430% that of the α1 with the anti-β1 Ab, respectively. These immunoprecipitation experiments indicated that the majority of α1 subunits in the brain and adrenal medulla formed dimers with β1 and β3 isoforms, respectively. The present experiments appear to be the first to demonstrate biochemically and morphologically the presence of α1β3 dimers in functioning cells.

Close association of α2 subunit with InsP₃R2

About 70% of peripheral BODIPY-FL-InsP₃ binding sites were juxtaposed to the membrane areas with α2-like or β2-like IR materials. This juxtaposition of the α2β2 isozyme with the InsP₃R-containing ER may account for the finding that exposure to 1 μm ouabain led to a time-dependent enhancement of catecholamine secretion in response to mobilization of partially stored Ca²⁺ ions, but induced a marginal effect on the resting secretion. The majority of this ouabain effect may be ascribed to inhibition of the α2 isoform of Na⁺,K⁺-ATPases and not to that of the α1, although a Na⁺,K⁺-ATPase with a low ouabain affinity in guinea-pig might have severalfold more sensitivity to ouabain than that in rat (Gao et al. 1995; Blanco & Mercer, 1998). Exposure to 1 μm ouabain had a marginal effect on resting secretion, whereas application of 100 or 200 μm ouabain enhanced secretion with a latency of 20–30 s, suggesting that the majority of Na⁺,K⁺-ATPases in guinea-pig AM cells are insensitive to 1 μm ouabain. The finding that exposure to 1 μm ouabain had a marked effect on catecholamine secretion in response to mobilization of partially stored Ca²⁺, but not on the resting secretion suggested that inhibition of the α2 subunit resulted in an increase in local [Ca²⁺] in the vicinity of Ca²⁺ store sites with the consequent facilitation of Ca²⁺ uptake, as was noted with inhibition of the α3 isoform in arterial smooth muscle (Arnon et al. 2000).

The inhibition of the Na⁺,K⁺-ATPase α2 subunit may result in an increase in local Na⁺ concentrations, subsequently hindering Ca²⁺ extrusion via nearby NCX. While the NCX in arterial myocytes was localized in the membrane near the sarcoplasmic reticulum (Blaustein et al. 1998), the NCX-like immunoreactivity in AM cells was distributed continuously along the cell periphery. This immunoreactivity may reflect the expression of NCX, since it was mainly confined to the cell periphery, and similar staining patterns were observed with two distinct mAbs against NCX1, but not with mouse non-immune serum. Moreover, the NCX1 gene was cloned from a bovine chromaffin cell cDNA library and the properties of expressed product of the gene were similar to those of native NCX activity in bovine chromaffin cells (Pan et al. 1998). The increase in [Ca²⁺] in the vicinity of peripheral Ca²⁺ storage would facilitate Ca²⁺ uptake, and the Ca²⁺ ions taken up at peripheral sites may diffuse to perinuclear store sites via an ER network (Terasaki et al. 1986; Park et al. 2000; Fig. 5E). This notion is consistent with the finding that extents of catecholamine secretion in response to mobilization of Ca²⁺ ions stored during 4 min intervals never became larger than those induced by mobilization of Ca²⁺ ions stored during nicotinic stimulation.

Physiological implication

The adrenal medulla in mammals is encapsulated by the adrenal cortex. Blood flows from the cortex to the medulla through the intra-adrenal portal vascular system, prior to emptying into the general circulation (Coupland, 1975). This vascular system exposes AM cells to cortical hormones at higher concentrations than that in the peripheral circulation (Unsicker et al. 1978). The OLS concentration in the whole rat adrenal gland was 500-fold higher than that in plasma (Hamlyn et al. 1991), whereas the concentration of ouabain in the adrenal venous system under resting conditions was not more than 5.7 times that (about 0.2 nm) in arterial plasma in awake dogs (Boulanger et al. 1993). Human plasma was reported to contain OLS at > 0.6 nm (Hamlyn et al. 1991). So far, there is no experimental evidence that the concentration of OLS in the adrenal venous system rises up to a submicromolar to micromolar level. However, since we currently have little information with regard to the conditions where OLS(s) is secreted from adrenal cortex, the possibility that submicromolar or micromolar concentrations of OLS(s) are secreted under some conditions cannot be excluded. It would be more reasonable to assume that the close association of the α2β2 isozyme with peripheral Ca²⁺ store sites has physiological implications. For example, plasma adrenaline was found to be elevated in parallel to an increase in plasma renin activity at 24 h after production of the experimental renal hypertension in rats (Chatelain et al. 1986); this increase in adrenaline secretion from AM cells might be mediated, at least in part, by angiotensin II-induced secretion of OLS (Laredo et al. 1997) and the subsequent inhibition of Na⁺,K⁺-ATPase α2 subunit with the consequent increase in Ca²⁺ mobilization. Since renin secretion from juxtaglomerular cells is facilitated by β receptor stimulation (Keeton & Campbell, 1980), there might be a positive feedback mechanism for renin secretion, in which renin, angiotensin II, OLS and catecholamine are involved.