Nax is an atypical sodium channel that is assumed to be a descendant of the voltage-gated sodium channel family. Our recent studies on the Nax-gene-targeting mouse revealed that Nax channel is localized to the circumventricular organs (CVOs), the central loci for the salt and water homeostasis in mammals, where the Nax channel serves as a sodium-level sensor of the body fluid. To understand the cellular mechanism by which the information sensed by Nax channels is transferred to the activity of the organs, we dissected the subcellular localization of Nax in the present study. Double-immunostaining and immunoelectron microscopic analyses revealed that Nax is exclusively localized to perineuronal lamellate processes extended from ependymal cells and astrocytes in the organs. In addition, glial cells isolated from the subfornical organ, one of the CVOs, were sensitive to an increase in the extracellular sodium level, as analyzed by an ion-imaging method. These results suggest that glial cells bearing the Nax channel are the first to sense a physiological increase in the level of sodium in the body fluid, and they regulate the neural activity of the CVOs by enveloping neurons. Close communication between inexcitable glial cells and excitable neural cells thus appears to be the basis of the central control of the salt homeostasis.
- sodium sensor
- salt homeostasis
- glial sodium channel
- ependymal cell
- GABAergic neuron
- neuron-glia interaction
when the correct balance between water and sodium level in the body fluid is lost, terrestrial animals feel appetite or satiety for water or salt and show ingestive or aversive behaviors. To explain these strictly controlled animal behaviors, neurobiologists postulated the existence of both sodium-specific receptors and osmoreceptors to sense the body fluid condition in the brain (1, 4, 5, 24). However, the molecular entities of the osmoreceptor and sodium-specific receptor have been long unknown. We have recently found evidence that an atypical sodium channel, Nax, is the sodium-specific receptor in a set of studies on Nax-gene knock-out (Nax-KO) mice generated by insertion of the lacZ reporter gene in-frame (9, 10, 22). We first identified Nax-expressing loci in the central nervous system (CNS) and peripheral tissues by enzyme histochemistry of β-galactosidase, the gene product of lacZ, at the tissue level (22, 23). In the CNS, the four circumventricular organs (CVOs), the subfornical organ (SFO), the organum vasculosum of the lamina terminalis (OVLT), the median eminence (ME), and the posterior pituitary (PP), were distinctively positive for Nax expression. Among these organs, the SFO and OVLT are called the sensory CVOs (13), because they include cell bodies of neurons that are thought to be involved in osmotic and sodium sensing. After water deprivation, the Nax-KO mice showed hyperneural activity, as estimated by Fos immunoreactivity, in the SFO and OVLT compared with wild-type mice (22). Relevantly, Nax-KO mice did not stop salt intake under dehydrated conditions, whereas wild-type mice avoided salt ingestion (9, 22).
We subsequently demonstrated that the Nax channel is a concentration-sensitive sodium channel with a threshold value of ∼150 mM for the extracellular sodium ion concentration ([Na+]o) (10). We further demonstrated that the salt-aversive behavior occurs by direct infusion of a hypertonic sodium solution into the cerebral ventricle in wild-type mice, but not in Nax-KO mice (9). The behavioral phenotype of Nax-KO mice was completely recovered by site-directed transfer of the Nax gene into the SFO (9). Taken altogether, it is probable that the Nax channel is the brain sodium level sensor that was postulated to be present in the CVOs, the brain loci known to be involved in the salt/water-intake regulation.
As the next step, it is important to identify the cellular population expressing Nax channels in the CVOs, as a prerequisite to elucidation of cellular mechanisms for controlling the salt/water-intake behavior of mammals. Here, we show that the Nax channel is specifically expressed in perineuronal glial processes enveloping neuronal populations in the CVOs. These Nax-positive glial cells are sensitive to an increase in the extracellular sodium level. These findings indicate that the Nax-expressing glial cells are the primary site of sodium-level sensing, and a specific neuron-glia communication regulated by Nax channel is implicated in the ingestion control for the salt/water homeostasis.
MATERIALS AND METHODS
Animals and housing.
All the animal experiments were carried out according to the guidelines of the National Institute for Basic Biology (Okazaki, Japan). Male wild-type (C57BL/6J), homozygous Nax-KO (22), or heterozygous GAD67-GFP (Δneo) knockin mice (GAD-GFP) (21) at 8–24 wk of age were used for experiments [67-kDa isoform of glutamic acid decarboxylase (GAD67); green fluorescent protein (GFP)]. All mice were housed in plastic basket cages under a constant room temperature (23°C) with a 12:12-h light-dark cycle (incandescent lights on at 7:00 AM). They were kept in this condition with ad libitum access to water and food (0.32% sodium/Rodent Diet CE-2; CLEA Japan, Tokyo, Japan).
Mice were transcardially perfused under deep pentobarbital sodium anesthesia first with PBS (pH 7.4) and then with 4% paraformaldehyde and 0.5% glutaraldehyde in PBS. Brains were coronally cut with a microslicer (model VT1000S; Leica Microsystems, Wetzlar, Germany) at 50 μm. Single immunostaining for wild-type mice (n = 5) was performed with an affinity-purified rabbit polyclonal antibody to Nax (10), as described previously (19). No detergent was added to a series of staining solutions except for the blocking solution (Triton X-100, 0.05%), so as to protect the subcellular fine structures of tissues. A substrate for the peroxidase reaction, 3,3′-diaminobenzidine (DAB) was used. Double immunostaining for GAD-GFP mice (n = 3) was performed with anti-Nax and rat monoclonal anti-GFP antibodies (diluted at 1:1,000 with PBS; GF090R, Nacalai Tesque, Kyoto, Japan). The sections were first incubated with the blocking solution (Triton X-100, 0.05%) for 3 h at room temperature and then with anti-Nax antibody for 24 h at 4°C. After washing in PBS, the sections were incubated with species-specific biotinylated anti-rabbit IgG from donkey (diluted at 1:100 with PBS; RPN1004V1, Amersham Biosciences, Piscataway, NJ) for 3 h at room temperature, rinsed with PBS, and incubated with 1.4 nm streptavidin-conjugated gold particles (diluted at 1:100 with PBS; Nanoprobes, Yaphank, NY) for 3 h at room temperature. After being rinsed with PBS, the sections were postfixed with 1% glutaraldehyde in 0.1 M phosphate buffer (PB; pH 7.4) for 10 min and washed with distilled water. Subsequently, silver enhancement was carried out in the dark with HQ Silver kit (Nanoprobes). The sections were then incubated with anti-GFP antibody for 24 h at 4°C, rinsed with PBS, incubated with species-specific, horseradish peroxidase-conjugated anti-rat IgG (diluted at 1:100 with PBS; cat. no. AP183P, Chemicon, Temecula, CA) for 1 h at room temperature, rinsed with PBS, and processed for peroxidase reaction with DAB. The immunolabeled sections were further fixed in an osmium tetroxide solution, dehydrated, and embedded flat in Epon. Sections were first observed with a light microscope, and then ultrathin sections were prepared, stained with uranyl acetate, and finally examined with an electron microscope (model JEM-1200EX, JEOL, Tokyo, Japan). Nax-KO mice were used as negative controls for immunostaining with anti-Nax antibody.
Single immunohistochemistry with anti-Nax antibodies.
GAD-GFP (n = 4) mice were transcardially perfused under deep pentobarbital sodium anesthesia first with PBS and then with 3.5% formaldehyde in PBS. Tissue sections were prepared from the fixed brains and stained with anti-Nax antibody, as described above. Detergent (Triton X-100, 0.1%) was used in the blocking and first-antibody treatment steps. Immunoreactions were visualized with streptavidin-Texas Red (diluted at 1:100 with PBS; cat. no. RPN1233, Amersham Biosciences).
Under deep anesthesia with pentobarbital sodium, wild-type mice (n = 5) were perfused with heparinized PBS followed by 4% paraformaldehyde in 0.1 M PB. Brains were postfixed with 4% paraformaldehyde in 0.1 M PB (pH 7.4) for 24 h at 4°C. After several washes with PBS, brain blocks, including the SFO and OVLT, were cut on a vibratome (model DTK-1000 microslicer DSK, Kyoto, Japan) to obtain coronal sections (30 μm). The pituitary lobe was removed from the basicranium, and cut horizontally. The sections were incubated with 25 mM glycine in PBS for 30 min and pretreated with 5% normal goat serum (NGS) in PBS containing 0.3% Triton X-100 (PBST) overnight at 4°C. They were then incubated with anti-Nax antibody in PBST containing 5% NGS for 2–3 days at 4°C. The sections were then incubated with FITC-conjugated anti-rabbit IgG (10 μg/ml, Kirkegaard and Perry, Gaithersburg, MD) for 2 h. The sections were further incubated with guinea pig antibody against glia-specific glutamate transporter (GLAST) (diluted at 1:4,000, Chemicon) in PBST containing 5% NGS for 2 days at 4°C and treated with biotinylated anti-guinea pig (7.5 μg/ml, Vector Laboratories, Burlingame, CA) in PBST for 2 h, followed by streptavidin-Texas Red (20 μg/ml, Vector Laboratories) in PBST for 2 h. The sections were mounted on slide glasses and sealed with Vectashield (Vector Laboratories). Observations were made using a LSM510 laser-scanning confocal microscope (Carl Zeiss, Germany). The sequential scanning of the recording configuration for FITC and streptavidin- Texas Red was used to avoid bleed-through of FITC. Omission of the primary antibody did not produce any visible immunostaining in these preparations.
Triple immunostaining of dissociated SFO cells.
Adult wild-type mice (8–16 wk old: n = 10) were deeply anesthetized with ether and decapitated. SFOs were then dissected from brains. Dissociation of the SFO was performed as previously reported (9). Isolated cells were plated on coverslips (Matsunami Glass, Kishiwada, Japan) and placed within a humidity-controlled incubator gassed with 5% CO2 for 3 h at 37°C. Cells were then fixed with 10% formalin solution, incubated in 25 mM glycine in PBS for 30 min, and pretreated with 5% normal goat serum (NGS) in PBS containing 0.3% Triton X-100 (PBST) for 2 h at room temperature. They were then incubated with anti-MAP2 (goat; sc-5359; Santa Cruz Biotechnology, Santa Cruz, CA), anti-GLAST and anti-Nax antibodies in PBST containing 5% NGS for 2 h at room temperature. Cells were then incubated with Alexa Fluor 647-conjugated anti-rabbit IgG, Alexa Fluor 488-conjugated anti-guinea pig IgG, and Alexa Fluor 546-conjugated anti-goat IgG (diluted at 1:200; Molecular Probes, Eugene, OR) for 2 h.
Sodium-ion imaging of SFO cells.
Cells of SFO were obtained from brains of adult (8–16 wk old) wild-type mice, as described above. Cells were plated onto glass bottom dishes (Matsunami Glass) and placed within a humidity-controlled incubator gassed with 5% CO2 at 37°C for 3–8 h before measurements. Sodium ion imaging was performed as described previously (10). Briefly, cells were loaded in advance with 4 μM sodium-binding benzofuran isophthalate acetoxymethyl ester (SBFI/AM, Molecular Probes) plus 0.02% Pluronic F127 (Sigma) for 60 min at 37°C. During the recording, cells were continuously perfused at 1 ml/min with an isotonic solution containing (in mM): 135 NaCl, 5 KCl, 2.5 CaCl2 1 MgCl2, 20 HEPES, 10 glucose, and 10 NaOH, titrated to pH 7.3 with HCl (145 mM [Na+] solution). Ten minutes after the start of the recording, [Na+]o was increased to 170 mM by replacing the perfusion solution with additional NaCl (170 mM [Na+] solution). The fluorescent ratio was monitored at excitation wavelengths of 340 and 380 nm using a cooled charge-coupled device camera (model ORCA-ER; Hamamatsu Photonics, Hamamatsu, Japan) coupled to the microscope. Data were collected every 30 s and analyzed using AQUACOSMOS software (Hamamatsu Photonics). After the recording, cells were fixed with 10% formalin solution for 30 min and subjected to immunostaining with antibodies to Nax, GLAST, or glial fibrillary acidic protein (GFAP) (sc-9065; Santa Cruz Biotechnology).
Nax channel is coexpressed with a glial marker protein, GLAST, in the CVOs.
To identify Nax-expressing cells, we performed double-immunostaining using anti-Nax and GLAST antibodies. GLAST is one of the glutamate transporters distributed almost exclusively in the plasma membrane of glial cells (14, 18), and GLAST is the primary glutamate transporter in the CVOs (2). As shown in Fig. 1, the Nax channel was colocalized with GLAST in both the SFO and OVLT. In the OVLT, the expression of GLAST was almost consistent with the distribution of Nax (Fig. 1, D–F). However, in the SFO, GLAST distribution was slightly more abundant and homogeneous than distribution of the Nax channel (Fig. 1, A–C). This might be due to the difference in distribution of Nax channels and GLAST in glial processes or presence of a Nax-negative subpopulation in glial cells. When the organs were stained with an antibody to GFAP as a glial cell marker, Nax channel appeared not to be colocalized with GFAP (not shown). This is probably ascribed to the fact that GFAP is preferentially localized to cell bodies and primary shafts of glial cells, but not to their lamellate processes (25), where Nax channels are populated.
Nax channel is expressed in glial laminate processes.
To further confirm the expression of the Nax channel in glial cells of the SFO and OVLT, and to determine the subcellular localization of Nax channel in glial cells, we performed immunoelectron microscopic studies using Nax antibody. Optical microscopic immunostaining images of the SFO and OVLT used for electron microscopic studies are shown in Fig. 2, A and B and Fig. 3A, respectively. Specific immunopositive signals were observed throughout the SFO (Fig. 2A), but not in the fornix and choroid plexus. The most intensive signals appeared to be associated with the periphery of a subset of neural cell bodies and their principal processes (Fig. 2B). Ependymal cell layers in the SFO were also highly positive. In the OVLT, radially oriented dense fibrous structures extending from the ventricular surface and midline were markedly positive (Fig. 3A).
Figure 2, C–F show immunoelectron micrographs of the localization of Nax in the SFO. Here, immunopositive signals turned out to be localized mainly to thin lamellar processes (arrows) surrounding neuronal cell bodies and processes, and in some cases, synapses. These immunopositive thin lamellate processes were apparently extended from cell bodies of ependymal cells (Fig. 2, C and D) or astrocytes (Fig. 2, E and F). Nax channels thus appeared to be transported and localized to the lamellar processes and endfeet of ependymal cells and astrocytes surrounding neurons. Perineuronal immunopositive signals observed in the optical micrographs of the SFO (Fig. 2B) were therefore considered to have originated from the glial-cell enveloping around neurons. Endothelial cells of capillary vessels were free of signals (Fig. 2F). In the OVLT, similar features were observed in that endfeet of ependymal cells and astrocytes surrounding neural cell bodies and neurites were positive for Nax (Fig. 3, B and C). In the present study, we could not detect significant immunopositive signals in neuronal cell bodies and their processes, including synapses, in either the SFO or OVLT.
Glial cells isolated from the SFO are sensitive to the increase of the extracellular sodium level.
To examine the overlapping ratio between Nax-positive cells and GLAST-positive cells in the SFO, we performed triple immunostaining of dissociated SFO cells using antibodies to Nax, GLAST, and microtubule-associated protein 2 (MAP2) (Fig. 4A). MAP2 is a cytoskeleton member that is specifically expressed in neurons (3). Total numbers of cells were counted referring to the cellular nuclei visualized with a fluorescent dye, DAPI (4′,6-Diamidino-2-phenylindole dihydrochloride). Among 432 cells, 227 cells (52.5%) were positive for Nax, 283 cells (65.5%) were positive for GLAST, 142 cells (32.9%) were positive for MAP2, and 7 cells (1.6%) were negative for the three antibodies. Among 227 Nax-positive cells we examined, all were positive also for the glial cell marker GLAST, and none was positive for the neuronal cell marker MAP2.
By using intracellular ion-imaging analysis, we recently demonstrated that a subset of dissociated SFO cells show sodium influx in response to an increase of the extracellular sodium level (10). Such sodium-sensitivity was not observed in dissociated SFO cells derived from Nax-KO mice, but the cells transfected with the Nax-expression plasmid recovered the response, indicating that Nax is an extracellular sodium-level- sensitive sodium channel (10). In the present study, we examined whether the sodium-sensitive cells in the SFO are glial cells by examining the expression of the glial marker proteins GLAST and GFAP. Figure 4, B and C shows sodium-ion imaging using the dissociated SFO cells. As shown in our recent paper (10), a subset of the SFO cells are sensitive to a small increase of [Na+]o from the physiological level (145–170 mM). After the ion-imaging study, cells were stained with anti-Nax, anti-GLAST, or anti-GFAP antibodies. We found that all of the sodium-sensitive cells are Nax-immunopositive (Fig. 4, Ba, -b, -c, and C) and at the same time GLAST- (Fig. 4, Bd, -e, -f and C) and GFAP- (Fig. 4, Bg, -h, -i, and C) immunopositive, and that their intracellular Na concentrations were similarly elevated from 10 to 30 mM. Nax-, GLAST- and GFAP-negative cells were relatively small, and some of them extended neurite-like short processes (arrows in Fig. 4, Bc, -f, and -i). When similar experiments were performed using SFO cells derived from Nax-KO mice, no sodium-sensitive cells were observed (not shown). These results clearly indicate that sodium-sensing cells in the SFO are almost exclusively glial cells.
Nax-positive glial cells are associated with multiple neural populations in the SFO.
We next examined what types of neurons are surrounded by Nax-positive lamellate processes of glia. Numerous glutamatergic, serotonergic, GABAergic, and glycinergic fibers and their terminals have been identified by immunohistochemical techniques in the sensory CVOs (16). However, identified neuronal cell bodies containing neurotransmitters in the SFO are only GABAergic interneurons. Therefore, we addressed the relationship between GABAergic interneurons and Nax-positive glial cells using GAD-GFP knockin mice (21).
As shown in Fig. 5, A–C, some of GAD-positive neurons appeared to be surrounded by Nax-positive glial processes in the SFO. On the contrary, Nax-positive glial processes extended also to the area where GAD-positive neurons were not observed. This observation was confirmed by electron microscopic analysis (Fig. 5, G and H). GAD neurons were immunolabeled by the peroxidase method, and Nax-positive glial processes were labeled by the immunogold-silver method. GAD-positive neurons (GAD-N in Fig. 5G) were associated with glial laminate processes, which contain silver particles (see the inset of Fig. 5G). However, many GAD-negative neurons, which did not contain the reaction products of peroxidase (N in Fig. 5H), were also associated with glial laminate processes containing silver-particles (the inset of Fig. 5H).
In the OVLT region, GAD-positive neurons resided outside the OVLT in the dorsal area and did not overlap with the Nax-positive population at all (Fig. 5, D–F). Therefore, the possibility that Nax-positive glial cells envelop GABA neurons is low in the OVLT. These findings suggest that glial laminate processes expressing Nax channels contact various neurochemical circuitries in the SFO and OVLT, encompassing GABAergic interneurons in the SFO.
Beginning from our first report showing abnormal salt-intake behavior in Nax-KO mice under dehydrated conditions (22), we have shown that Nax is a sodium-level-sensitive sodium channel playing an essential role in sodium-level sensing in the CVOs and in the control of salt ingestion (9, 10). These studies on the whole proved Nax to be the molecular entity of the brain sodium sensor, which has long been hypothesized to be involved in water and salt homeostasis in mammals (5, 15). In this report, we demonstrate that the primary subcellular site of sensing the sodium level in body fluids is perineuronal glial processes in the CVOs. These findings indicate that control of neural activity by glial cells could be the main cellular mechanism underlying the sodium-sensing mechanism in the CVOs.
The CVOs are specialized brain structures, which are collectively so named because of their proximity to the ventricles of the brain (11). The five CVOs known in mammals are SFO, OVLT, ME, PP, and area postrema. In addition to the periventricular midline location in the brain, they have common features, such as an extensive vasculature lacking the blood-brain barrier, dense population of a variety of peptidergic factor receptors, and existence of atypical ependymal cells. Of these CVOs, only three (SFO, OVLT, and area postrema) include neuronal cell bodies with efferent fibers extending to many other brain regions and are termed the sensory CVOs (13). Because their neurons are directly exposed to the chemical environment of the circulation, the sensory CVOs have been thought to be the organs sensing various kinds of circulating information, including sodium concentration and osmotic pressure in body fluids (16).
The Nax-gene is expressed in four CVOs: SFO, OVLT, ME and PP (22). In the present study, using immunoelectron microscopy with anti-Nax antibody (Figs. 2 and 3), positive signals were not observed in neurons, including cell bodies and processes in the SFO and OVLT. When the dissociated SFO cells were stained with antibodies against Nax, GLAST, and MAP2, Nax and GLAST were almost coexpressed but none of the Nax-positive cells was positive for the neuronal marker protein MAP2 (Fig. 4A). In addition, the SFO cells responsive to the increase in the extracellular sodium level were revealed to be GLAST-positive when examined by sodium imaging and immunocytochemistry (Fig. 4, B and C). We confirmed that the sodium-responsive cells are also positive for GFAP, another glial cell marker (Fig. 4, B and C). Taken together with immunoelectron microscopic observation, it is most probable that the Nax channel is specifically expressed in astrocytes and ependymal cells of the sensory CVOs.
When we introduced Nax cDNA into the brain of the Nax-KO mice with an adenoviral expression vector, only those animals that received the transduction into the SFO regained the salt-avoiding behavior under dehydrated conditions (9). It is known that the adenovirus vectors have a much higher affinity to glial cells than neurons in the CNS (12). On injection into the SFO, a larger proportion of glial cells expressed the transgene (20). This also indicates that Nax expression in glial cells is requisite to the function of the SFO.
Recently, Grob et al. (8) reported that neurons in the median preoptic nucleus (MnPO) responded to a change in the extracellular sodium concentration. They also showed that the MnPO is positive for Nax expression by in situ hybridization. We detected X-gal staining in the medial preoptic area as well as the OVLT neighboring the MnPO, but not in the MnPO in Nax-KO mice (22). The reason for this discrepancy is not clear at present. Further investigation to identify Nax-expressing cells in these areas would be necessary.
Glial expression of Nax channel proteins was observed in not only the SFO and OVLT but also ME and PP. Ependymal cells and pituicytes (glial cells) were positive for X-gal staining in the ME and PP, respectively, in Nax-KO mice (22). This also supports the view that the major cellular components bearing Nax channels in the CVOs appear to be glial processes surrounding neuronal cell bodies, axons, dendrites, and their terminals. Ependymal cells and astrocytes may contribute to sensing the sodium level in the CSF and blood, respectively, because the former faces to the cerebral ventricle and the latter is located around the blood vessel.
It is also noteworthy that Nax-channel expression is confined to nonmyelinating Schwann cells single-enveloping the peripheral nerves (23), suggesting a common function of Nax channels in the central and peripheral glial cells. Furthermore, Nax transcription is induced in reactive astrocytes following excitotoxic lesion of neurons (7). This is reminiscent of the fact that Nax cDNA was originally cloned from cultured astrocytes and named the glial sodium channel (NaG) (6). Taken altogether, except for the peripheral ganglion neurons, such as dorsal root and trigeminal ganglion neurons (22) and alveolar type II cells in the lung (23), the major physiological role of the Nax channel appears to be in glia.
At present, it is not clear what kind of neuronal cells are surrounded by the Nax-positive glial processes in the CVOs. However, several lines of evidence support the idea that glial cells expressing Nax channels are not associated with a specific neuronal species bearing a particular neurotransmitter. As shown in Fig. 5, the Nax-positive glial processes associate with both GABAergic and non-GABAergic neurons in the two sensory CVOs, SFO and OVLT. The Nax-positive glial cells show homogeneous distribution in the PP, suggesting that they associate with both vasopressin- and oxytocin-nerve terminals. Similarly, they also show homogeneous distribution in the ME (22), suggesting that they associate with nerve terminals and fibers, including various kinds of neurotransmitters and hormones. Furthermore, Nax channel is expressed in nonmyelinating Schwann cells distributed throughout the whole body (23). Thus, it is presumable that the Nax channel in various glial cells regulates multiple types of neurochemical circuitries by enveloping them with perineuronal processes.
In summary, Nax channels populate perineuronal glial processes and thus, they appear to regulate neural activities through the glia-neuron communications in the sensory CVOs. Glial cells have long been considered to be inert partners of neurons in the CNS. However, it is becoming evident that glial cells are intimately involved in neuronal signaling (17). Our studies on knockout mice demonstrated that the Nax channel exerts inhibitory influences on neuronal activities in the SFO and OVLT, as judged from Fos immunoreactivity during dehydration (22). The Nax-channel function in the CVOs would provide a good system to study the molecular mechanisms for neuron-glia interactions.
This work was supported by grants-in-aids of the Ministry of Education, Culture, Sports, Science and Technology of Japan, and from the Japan Science and Technology Agency: Core Research for Evolutional Science and Technology (to M. Noda), Precursory Research for Embryonic Science and Technology (to E. Watanabe), and Solution Oriented Research for Science and Technology (to Y. Yanagawa). This work was also supported by The Salt Science Research Foundation (No. 0536), The Public Health Research Foundation, The Asahi Glass Foundation, and The Mitsubishi Foundation.
The authors are grateful to Kazumi Takeuchi, Tomomi Katayama, Chizu Egusa, and Mie Yasuda for technical assistance, and to Akiko Kodama for secretarial assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society