Am J Physiol Cell Physiol 286: C495–C506, 2004.
First published November 5, 2003; 10.1152/ajpcell.00386.2003.
Gastric parietal cell secretory membrane contains PKA- and acid-activated
Kir2.1 K⫹ channels
Danuta H. Malinowska, Ann M. Sherry, Kirti P. Tewari, and John Cuppoletti
Department of Molecular and Cellular Physiology, University of
Cincinnati College of Medicine, Cincinnati, Ohio 45267-0576
Submitted 11 September 2003; accepted in final form 22 October 2003
H⫹-K⫹-ATPase; hydrogen chloride secretion; parietal cell K⫹ channel
of the H⫹-K⫹-ATPase and K⫹ and
Cl⫺ channels in the apical membrane of the gastric parietal cell
results in acid secretion (10, 19, 46, 50, 65). A Cl⫺ channel
identified to be ClC-2 by cloning and expression studies (35)
has been characterized in rabbit gastric membrane vesicles (6),
which also exhibit H⫹-K⫹-ATPase activity (10). Regulation of
ClC-2 by PKA and low extracellular pH is consistent with
regulation of gastric HCl secretion by cAMP (4) and with an
essential property of gastric apical membrane ion channels that
they must be able to function at very low extracellular (or
luminal) pH of 3 or lower. Both native and recombinant ClC-2
not only functioned at pH 3 but also were activated by low pH
on the luminal side of the membrane (6, 35, 55), mediated by
an extracellular pH sensor (56).
THE COORDINATED FUNCTION
Address for reprint requests and other correspondence: D. H. Malinowska,
Dept. of Molecular and Cellular Physiology, Univ. of Cincinnati College of
Medicine, PO Box 670576, Cincinnati, OH 45267-0576 (E-mail:
[email protected]).
http://www.ajpcell.org
Less is known of the K⫹ channel. Luminal K⫹ is required as
a substrate for the gastric H⫹-K⫹-ATPase. K⫹ movement
across the apical membrane was first suggested to occur
through K⫹ channels on the basis of experiments measuring
conductive K⫹ transport in isolated rabbit gastric vesicles,
using various approaches (8, 10, 46, 65), and on initial K⫹
channel current recordings in gastric vesicles (7). As with the
gastric Cl⫺ channel, gastric K⫹ channel properties must be
consistent with its physiological role and location. Minimally,
the channel must be able to function at low extracellular pH for
HCl secretion to occur. K⫹ channel activity has been measured
by patch clamp in the basolateral membrane of the Necturus
gastric oxyntic cell (57, 62) and the rabbit parietal cell (51), but
not in the apical membrane, since access to this membrane is
virtually impossible. Some initial tetraethylammonium (TEA)sensitive K⫹ channel current recordings from stimulated gastric membrane vesicles fused to lipid bilayers have been
reported (7). However, detailed molecular and functional studies
of gastric apical membrane K⫹ channels have been lacking. Two
studies recently appeared suggesting that KCNQ1 coupled with
KCNE2 or KCNE3 (20) or Kir4.1 (18) is the parietal cell apical
membrane K⫹ channel essential for gastric HCl secretion. Both
studies have some strong supporting data, but they are in conflict
with each other and both fail to prove an exclusive role for these
proteins in K⫹ transport at the apical membrane.
Studies of native gastric apical membrane proteins and their
regulatory mechanisms have utilized gastric H⫹-K⫹-ATPasecontaining membranes isolated from the gastric mucosae of
cimetidine-injected, nonsecreting (resting) and histamine-injected, secreting (stimulated) rabbits (6, 10). These resting and
stimulated gastric mucosal membranes exhibit conductive K⫹
transport (10, 46, 65), K⫹ channel activity (7), Cl⫺ channel
activity (6), H⫹-K⫹-ATPase activity (10), and HCl accumulation (10) consistent with the physiological state of the gastric
mucosae from which they were derived. In the present study,
Kir2.1 K⫹ channel mRNA was identified in rabbit gastric
parietal cells by in situ reverse transcriptase-polymerase chain
reaction (RT-PCR). Immunoconfocal microscopy of rabbit
gastric glands was used to localize Kir2.1 and compare this
with H⫹-K⫹-ATPase and ClC-2 Cl⫺ channel location. Some
properties of K⫹ channel function of native rabbit gastric
vesicles and recombinant Kir2.1 expressed in Chinese hamster
ovary (CHO) cells and in Xenopus laevis oocyte membranes
were characterized and compared. Regulation of native and
recombinant channel function was also examined. The data
suggest an important role for the Kir2.1 K⫹ channel in gastric
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.
0363-6143/04 $5.00 Copyright © 2004 the American Physiological Society
C495
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
Malinowska, Danuta H., Ann M. Sherry, Kirti P. Tewari, and
John Cuppoletti. Gastric parietal cell secretory membrane contains
PKA- and acid-activated Kir2.1 K⫹ channels. Am J Physiol Cell
Physiol 286: C495–C506, 2004. First published November 5, 2003;
10.1152/ajpcell.00386.2003.—Our objective was to identify and localize a K⫹ channel involved in gastric HCl secretion at the parietal
cell secretory membrane and to characterize and compare the functional properties of native and recombinant gastric K⫹ channels.
RT-PCR showed that mRNA for Kir2.1 was abundant in rabbit gastric
mucosa with lesser amounts of Kir4.1 and Kir7.1, relative to ␤-actin.
Kir2.1 mRNA was localized to parietal cells of rabbit gastric glands
by in situ RT-PCR. Resting and stimulated gastric vesicles contained
Kir2.1 by Western blot analysis at ⬃50 kDa as observed with in vitro
translation. Immunoconfocal microscopy showed that Kir2.1 was
present in parietal cells, where it colocalized with H⫹-K⫹-ATPase
and ClC-2 Cl⫺ channels. Function of native K⫹ channels in rabbit
resting and stimulated gastric mucosal vesicles was studied by reconstitution into planar lipid bilayers. Native gastric K⫹ channels exhibited a linear current-voltage relationship and a single-channel slope
conductance of ⬃11 pS in 400 mM K2SO4. Channel open probability
(Po) in stimulated vesicles was high, and that of resting vesicles was
low. Reduction of extracellular pH plus PKA treatment increased
resting channel Po to ⬃0.5 as measured in stimulated vesicles.
Full-length rabbit Kir2.1 was cloned. When stably expressed in
Chinese hamster ovary (CHO) cells, it was activated by reduced
extracellular pH and forskolin/IBMX with no effects observed in
nontransfected CHO cells. Cation selectivity was K⫹ ⫽ Rb⫹ ⬎⬎
Na⫹ ⫽ Cs⫹ ⫽ Li⫹ ⫽ NMDG⫹. These findings strongly suggest that
the Kir2.1 K⫹ channel may be involved in regulated gastric acid
secretion at the parietal cell secretory membrane.
C496
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
HCl secretion at the secretory membrane of the parietal cell. Some
of these findings have been reported in abstract form (9, 36).
MATERIALS AND METHODS
AJP-Cell Physiol • VOL
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
K⫹ channel screening. Poly(A)⫹ mRNA was isolated from frozen
rabbit gastric mucosa, large intestine, small intestine, and lung by
using Tri reagent followed by the Oligotex-dt mRNA procedure.
Isoform-specific PCR primers were designed from the known sequences of the inward rectifying Kir K⫹ channels 2.1 (rabbit; accession no. D21057), 4.1 (rabbit; accession no. AB005734), and 7.1
(human; accession no. AJ007557). Rabbit brain mRNA was used as a
positive control, and mRNA from three rabbit epithelial tissues (large
intestine, small intestine, and lung) was used to investigate tissue
distribution. First-strand cDNA was synthesized and amplified by
hot-start PCR, using TaKaRa ExTaq polymerase and Taqstart antibody. The same amount of cDNA was used for all the PCR reactions.
A 669-bp fragment of ␤-actin was amplified to control for relative
quantities of cDNA used. The PCR products were subcloned into
pCRII-TOPO (InVitrogen) and sequenced by the dideoxy chain termination method, using ␣-35S-labeled ATP and T7 DNA polymerase.
Cloning rabbit gastric Kir2.1. The full-length open reading frame
of the rabbit gastric Kir2.1 K⫹ channel was cloned from rabbit gastric
mucosal mRNA by using RT-PCR. First-strand cDNA was synthesized
and then amplified by using hot-start PCR with primers flanking the
coding region of the previously cloned rabbit heart Kir2.1 or RBHIK1
(30). The sense primer was 5⬘-GAAGCGATGGGCAGCGTGCG-3⬘,
and the antisense primer was 5⬘-GGAGTCGGTCAGTCATATCTCTGACTCTCGCCG-3⬘. PCR conditions consisted of 40 cycles of denaturing for 45 s at 94°C, annealing for 45 s at 64°C, and elongating for 2
min at 72°C. A 1,302-bp cDNA encoding the entire open reading frame
of rabbit gastric Kir2.1 was obtained and sequenced.
Direct in situ RT-PCR. Adult rabbit stomachs were removed and
washed in PBS. Gastric glands were isolated by collagenase digestion
(4), fixed in 4% paraformaldehyde, and paraffin embedded. Sections
(5 ␮m) were loaded onto slides, pretreated with proteinase K, and
acetylated before use (69). Sections were incubated with RNase-free
DNase I overnight at 37°C, heated for 2 min at 94°C to deactivate the
enzyme, washed with diethyl pyrocarbonate (DEPC)-treated water,
and placed in the ThermoSlide Block for the RT reaction. First-strand
cDNA synthesis was performed directly on the tissue sections (22, 42)
by using the antisense primer and SuperScript II RT. Sections were
rinsed with DEPC-treated water and placed again in the ThermoSlide
block for hot-start PCR amplification of an ⬃600-bp piece of the
3⬘-untranslated region (UTR) of the rabbit Kir2.1 K⫹ channel cDNA
using digoxigenin-11–2⬘-deoxyuridine-5⬘-triphosphate (DIG-11dUTP). The 3⬘-UTR sense primer was 5⬘-CTACACAGTCTGTTGGTCAGAGGCCCGAAACAG-3⬘, and the antisense primer was 5⬘CCCAGAGGCTACCATTTGGCTAGTGTTACATAATGC-3⬘. PCR
conditions consisted of 25 cycles of denaturing for 45 s at 94°C,
annealing for 45 s at 53°C, and elongating for 2 min at 72°C. PCR
cDNA products were detected as previously described (11). Briefly,
after PCR, the sections were rinsed with DEPC-treated water and
blocked with 150 mM NaCl-0.2% BSA at 50°C for 10 min. Sections
were incubated with alkaline phosphatase (AP)-conjugated anti-DIG
antibody (1:150) in blocking medium (0.1 M NaCl, 0.1 M Tris䡠HCl,
pH 7.4) at 37°C for 30 min. Sections were rinsed at room temperature
for 1 min in blocking medium, and the color reaction was developed
with 1:200 nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP) stock solution (18.75 mg/ml NBT and 9.4 mg/ml
BCIP) in blocking medium at 37°C for over 2 h. The following
negative controls were prepared in parallel to the test sample: RNasepretreated samples, no PCR primers, no ExTaq polymerase, and no
DIG-11-dUTP. No DNase pretreatment was the positive control.
Samples were photographed with a Nikon Microphot-Fxa microscope.
For identification of parietal cells in the gastric glands, an NBTsuccinic dehydrogenase-linked staining procedure was used.
Gastric membrane vesicle preparations. H⫹-K⫹-ATPase-containing membrane vesicles were purified by differential and sucrose
gradient fractionation of vesicles from the gastric mucosae of secreting (histamine ⫹ diphenhydramine injected) and nonsecreting (cimetidine injected) New Zealand White rabbits as previously described
(10) and are referred to as stimulated and resting membrane vesicles,
respectively.
Western blot analysis of Kir2.1 in gastric vesicles. Resting and
stimulated rabbit gastric vesicle proteins (10) were boiled for 30 s,
separated by SDS-PAGE, and transblotted to Hybond ECL nitrocellulose paper (Amersham). The nitrocellulose replicates were stained
with Blot FastStain and photographed before nonspecific sites were
blocked for 1 h at 22°C with TBS (20 mM Tris䡠HCl, 500 mM NaCl,
pH 7.5) containing 0.1% Tween 20, 5% blotting-grade nonfat dry milk
(Bio-Rad), and goat anti-rabbit Fab fragments (1:130). Nitrocellulose
replicates were incubated at room temperature with anti-Kir2.1 antibody (1:200) for 1 h, followed by horseradish peroxidase-conjugated
goat anti-rabbit antibody (1:1,000) for 1 h, both diluted in TBS-0.1%
Tween 20-1% milk. The ECL Western blotting analysis system from
Amersham was then used for detection and visualization on autographic film. For control, Kir2.1 antibody was blocked with its
antigenic peptide.
In vitro translation. For in vitro translation, the TNT T7 Coupled
Reticulocyte Lysate System (Promega) was used according to the
manufacturer’s instructions. Full-length cDNA for rabbit Kir2.1 was
used and [35S]methionine was used for detection. Reactions with or
and without cDNA and with and without microsomes were always
performed. Proteins were separated on mini-SDS PAGE gels, stained
with GelCode Blue Stain reagent, dried for 2 h, and then autoradiographed overnight.
Immunohistochemistry and confocal microscopy. Isolated, fixed,
and permeabilized adult rabbit gastric glands were used (26). Rabbit
gastric glands, isolated by collagenase digestion (4), were fixed for 20
min at 22°C in 4% paraformaldehyde. The glands were then permeabilized with 0.5% Triton X-100 in PBS for 20 min and washed three
times in PBS before settling onto poly-lysine-coated coverslips (26).
All reactions were carried out at room temperature. Glands were then
blocked (1% BSA, 0.05% Tween 20 in PBS) for 30 min and incubated
with rabbit anti-Kir2.1 antibody (1:200), chicken anti-ClC-2 antibody
(1:300), or anti-H⫹-K⫹-ATPase ␣-subunit antibody (1:200,000) for
1 h. Samples were washed five times for 5 min with PBS containing
0.05% Tween 20 and then incubated with 1:1,000 goat anti-rabbit IgG
(Kir2.1) labeled with Alexa Fluor 546 (red fluorescence) and goat
anti-chicken (ClC-2) or goat anti-mouse (H⫹-K⫹-ATPase ␣-subunit)
IgG labeled with Alexa Fluor 488 (green fluorescence). Controls
included blocking anti-Kir2.1 antibody with its antigenic peptide and
omitting the primary antibodies for Kir2.1, ClC-2, and H⫹-K⫹ATPase ␣-subunit. Images of the stained gastric glands were obtained
by using a laser scanning confocal microscope (LSM 510, Zeiss)
equipped with argon and helium-neon lasers, an Axioplan upright
microscope, and oil-immersion objectives. Digitized images were
processed using the LSM 510 software and Adobe Photoshop. Standard excitation/emission wavelengths for Alexa 546 and 488 were
used as previously described (54).
Bilayer electrophysiology. The procedures and equipment were as
described in detail previously (6, 35). Briefly, a GeneClamp 500
amplifier (Axon Instruments), a TL-1 DMA interface, and a Frequency Devices eight-pole Bessel function filter were used for these
studies. Routinely, 100-s channel recordings were obtained at set
holding potentials and filtered at 500 Hz. Multiple recordings (usually
3) and at least three reconstitutions were obtained in each condition.
Identification of events and determination of dwell times and amplitudes, as well as curve fitting, were carried out using the pCLAMP
program (versions 5.5 and 6). Lipid bilayers were formed across a
100-␮m-diameter aperture in the wall of a Delrin cup by using a 3:1
mixture of 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-L-serine] (POPS)
and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE), 30
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
AJP-Cell Physiol • VOL
expressed in Xenopus laevis oocytes as previously described for the
ClC-2 Cl⫺ channel (35, 55). Oocytes were isolated, injected with 10
ng of channel cRNA per oocyte, and left to express for 4 days. Oocyte
plasma membranes were then prepared (35), reconstituted into planar
lipid bilayers, and characterized electrophysiologically.
Materials. POPS and POPE were obtained from Avanti Polar
Lipids and dissolved in n-decane (Aldrich). HEPES, citric acid, PKA
catalytic subunit, and AP-conjugated goat anti-rabbit antibody were
from Sigma. Tri reagent was from Molecular Research Center, Oligotex-dt mRNA kit was from Qiagen, rabbit brain poly(A)⫹ mRNA
and Taqstart antibody were from Clontech, Taq polymerase was from
Pan Vera, the TA cloning kit was from Invitrogen, and T7 DNA
polymerase (Sequenase v.2) was from United States Biochemical. The
T7 mMessage mMachine kit was from Ambion. DIG-11-dUTP, APconjugated anti-DIG antibody, and NBT/BCIP were from Boehringer
Mannheim. Goat anti-rabbit Fab fragments were from Jackson ImmunoResearch; rabbit polyclonal antibody to Kir2.1 was from Alomone;
mouse anti-H⫹-K⫹-ATPase ␣-subunit antibody and Blot FastStain were
from Chemicon. The TNT T7 Coupled Reticulocyte Lysate System was
obtained from Promega, [35S]methionine was from Amersham, and
GelCode Blue Stain reagent was from Pierce. Alexa Fluor-labeled IgGs
were from Molecular Probes. The chicken anti-ClC-2 antibody (from Dr.
Carol Blaisdel) was made against a ClC-2 COOH-terminal fusion protein, and its specificity has been well characterized (38). All other
reagents were of the highest grade available.
RESULTS
K⫹ channel screening, epithelial tissue distribution, and
cloning from rabbit gastric mucosa. To investigate the molecular nature of the gastric K⫹ channel, we screened gastric
mRNA by RT-PCR for three different K⫹ channels: Kir2.1 (1,
30, 39, 44), 4.1 (2, 58), and 7.1 (33, 43), which appeared to
have one or more characteristics similar to those of native
gastric vesicles. Primers were designed to be isoform specific,
and ␤-actin was used as a control for the amount of cDNA
used. Brain mRNA was the positive control, because in other
species, brain has been found to have a major amount of all
three K⫹ channels (15, 30, 33, 43, 44, 58). Rabbit large
intestine, small intestine, and lung mRNAs were examined in
parallel. The results are shown in Fig. 1. Brain had large and
Fig. 1. Screening for K⫹ channels by RT-PCR. Isoform-specific primers were
used to amplify cDNA fragments of Kir2.1 (1,304 bp), Kir4.1 (635 bp), Kir7.1
(425 bp), and ␤-actin (669 bp) from brain, gastric mucosa, large (LG) intestine,
small (SM) intestine, and lung mRNA. Two controls were carried out: control
1 contained no mRNA in the RT reaction, and control 2 contained no DNA in
the PCR reaction. Shown are ethidium bromide-stained agarose gels (1.2% for
Kir2.1; 2% for Kir4.1, Kir7.1, and ␤-actin) of the PCR products.
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
and 10 mg/ml in n-decane, respectively. Except when specified differently, all of the solutions contained 2 mM MgSO4, 1 mM K-ATP (pH
7.4), and 10 mM K-EGTA (pH 7.4) and were buffered with either 10 mM
K-HEPES (pH 7.4) or 10 mM K-citrate (pH 3.0). A fire-polished glass
capillary tube was used to apply lipids to the aperture of the Delrin cup,
followed by vesicles (1–2 mg/ml protein) to the planar lipid bilayer.
Reconstitutions of the channels lasted from 10 min to several hours. Upon
vesicle fusion to the lipid bilayer, channel orientation has been previously
established such that the cis side corresponds to the cytosolic side of the
channel and the trans side corresponds to the extracellular or luminal side
(6). When specified, PKA catalytic subunit (50 U/ml) was added to the
cis side of the bilayer. pHcis was 7.4 in all experiments. The electrical
reference side of the bilayer was the trans side. In channel current
recordings, the zero current level, which corresponds to the closed state
of the channel, as well as the open state of the channel are indicated.
When two open-state current levels were observed that were an even
multiple of the lowest open-state current level, they were interpreted as
reflecting the activity of two channels present in the bilayer. All channel
open probability (Po) values reported have been corrected for the number
of channels present as previously described (37, 56). To increase the
chances of observing the correct number of channels in the bilayer, we
carried out multiple 100-s recordings in each reconstitution. For all the
gastric vesicle work, recordings from one, two, and three channels were
observed and used. The distribution of the observations was ⬃25% for
one channel, 50% for two channels, and 25% for three channels. This
distribution was similar regardless of physiological state of the vesicles.
The number of different membranes or reconstitutions of the channel (n)
on which current recordings were performed is indicated with each data
point. Current-voltage (I/V) relationships were fitted by linear regression.
Slope conductance was determined in symmetric salt solutions by linear
regression analysis of the I/V relationship. In cases where estimates of
errors in the reversal potentials (intercept of the x-axis) were required,
records from individual membranes or groups of recordings from several
membranes were analyzed separately, and reversal potential values were
obtained. The mean ⫾ SE for these values was then calculated. The K⫹
equilibrium potential was calculated using Nernst’s equation (29), and
activity coefficients were from Ref. 47.
Expression of Kir2.1 in CHO cells and patch-clamp measurements.
Kir2.1 cDNA was stably transfected into CHO cells by using Lipofectamine as previously described (60). Whole cell K⫹ currents were
measured as previously described (11, 60). Measurements started 50–100
ms after 1,500-ms voltage-clamp pulses began. These pulses occurred in
20-mV increments between ⫹40 and ⫺140 mV from a beginning
holding potential of ⫺30 mV. Patch pipettes (borosilicate glass) of 1–1.5
M⍀ of resistance were prepared by using a two-stage Narashige puller.
The bath solution contained 140 mM K-gluconate, 5 mM MgCl2, 10 mM
HEPES, 1 mM CaCl2, and 10 mM glucose (pH 7.3). The pipette solution
contained 140 mM K-gluconate, 5 mM MgCl2, 10 mM HEPES, 5 mM
EGTA, and 1 mM MgATP (pH 7.3). Data were acquired with an
Axopatch CV-4 head stage with a Digidata 1200 digitizer and an
Axopatch 1D amplifier. pCLAMP 6.04 software (Axon Instruments,
Foster City, CA), Lotus 123 (Microsoft), and Origin (Microcal) were
used for data analysis. For I/V curves, the currents were normalized to cell
capacitance (pA/pF). Statistical significance of the difference between
two means was determined with the Student’s t-test by using n as the
number of cells. To determine the cation selectivity, symmetric 140 mM
K-gluconate was used to start with, and then the bath solution was
changed to 28 mM K-gluconate plus 112 mM N-methyl-D-glucamine
(NMDG⫹)-Cl, NaCl, RbCl, CsCl, or LiCl. I/V curves were plotted, and
the reversal potentials were obtained.
Statistical analysis. Statistical significance of the difference between two means was determined by using the Student’s t-test to
compare the means of two small samples (n ⬍ 30) from normal
populations with unknown variances not assumed to be equal. We
used n in all calculations of SE and statistical significance.
Expression of Kir2.1 in Xenopus laevis oocytes. Kir2.1 K⫹ channel
cRNA was made by using the T7 mMessage mMachine kit and was
C497
C498
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
Fig. 2. Localization of Kir2.1 mRNA in gastric glands by
direct in situ RT-PCR. A: parietal cells (dark blue) were
identified by staining mitochondrial succinic dehydrogenase
with a nitro blue tetrazolium (NBT)-linked assay. B: direct
in situ RT-PCR for Kir2.1 was carried out on isolated
gastric glands: a, contained all of RT-PCR components; b–f,
controls: no Ex-Taq polymerase (b), no primers (c), no
digoxigenin-11–2⬘-deoxyuridine-5⬘-triphosphate
(DIGdUTP) (d), RNase-pretreated samples (e), and no DNase
pretreatment (f). Scale bars: 10 ␮m (A), 20 ␮m (B).
more intense staining throughout the glands, indicating amplification of genomic DNA as well as cDNA for Kir2.1 in all
cells (22, 42). Therefore, Kir2.1 mRNA was specifically localized to parietal cells of gastric glands.
Immunolocalization of the Kir2.1 K⫹ channel protein. If the
Kir2.1 K⫹ channel is involved in acid secretion, it is essential
to localize not only the mRNA but, in particular, the protein to
the parietal cell and to study how its localization compares with
the H⫹-K⫹-ATPase and with ClC-2 Cl⫺ channels known to be
present in the parietal cell apical secretory membrane (54).
Before immunomicroscopy of Kir2.1 in gastric glands was
carried out, Western blots of rabbit gastric vesicles were used
to investigate specificity of the polyclonal peptide anti-Kir2.1
antibody and to investigate whether Kir2.1 is present in both
resting and stimulated vesicles. The results are shown in Fig.
3A. An ⬃50-kDa protein band was evident in resting and
stimulated gastric vesicles, and this protein band was blocked
when the antibody was preincubated with its antigenic peptide.
When in vitro translation was performed with the Kir2.1 cDNA
(Fig. 3B), a major band of ⬃50 kDa was observed without
microsomes, the expected size for Kir2.1. With microsomes
present there did not appear to be a major increase in size,
suggesting low or no posttranslational modification by glycosylation in agreement with reports that Kir2.1 is not glycosylated (53). The Kir2.1 protein observed on Western blots would
therefore seem to be the unglycosylated form. The anti-Kir2.1
Fig. 3. Western blot of Kir2.1 in resting (R) and stimulated (S)
rabbit gastric vesicles and in vitro translation of Kir2.1. A:
rabbit gastric vesicle proteins separated by 10% SDS-PAGE
were transblotted to nitrocellulose, probed with anti-Kir2.1
antibody (1:200) for 1 h followed by horseradish peroxidaseconjugated goat anti-rabbit antibody (1:1,000) for 1 h, and
visualized using chemiluminescence. For control, Kir2.1 antibody was blocked with its antigenic peptide. Arrowheads indicate a specific protein of ⬃50 kDa detected by anti-Kir2.1
antibody and blocked by peptide. B: TNT T7 Coupled Reticulocyte Lysate System was used to transcribe rabbit Kir2.1
cDNA with [35S]methionine present. Proteins separated on
mini-SDS-PAGE gels were dried and autoradiographed. Reactions with (⫹) or without (⫺) microsomes are shown, and
reactions without cDNA were performed as control. Arrowhead
indicates that a major protein of ⬃50 kDa was made.
AJP-Cell Physiol • VOL
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
similar amounts of mRNA for all three K⫹ channels relative to
actin. The gastric mucosa had a major amount of Kir2.1
mRNA, with much less Kir4.1 and just a trace of Kir7.1
relative to actin. Large intestine, small intestine, and lung also
had major amounts of Kir2.1 mRNA, with differing amounts of
Kir4.1 and Kir7.1. The large intestine and lung had substantial
amounts of Kir4.1 mRNA, with much smaller amounts of
Kir7.1 mRNA, whereas the small intestine had the reverse.
Because Kir2.1 was the most abundant, the full-length cDNA
(2,285 bases with 1,281 bases open reading frame ⬅ 427
amino acids) was obtained by RT-PCR from rabbit gastric
mucosal mRNA and sequenced. Its nucleotide sequence was
identical to that cloned from rabbit heart (30).
Localization of Kir2.1 K⫹ channel mRNA. Isolated rabbit
gastric glands were used to localize Kir2.1 mRNA (Fig. 2) by
direct in situ RT-PCR. Staining mitochondria of gastric glands
with an NBT-succinic dehydrogenase-linked assay identified
the parietal cells (highly enriched in mitochondria) as dark blue
cells (Fig. 2A). This staining is similar to that observed when
glands are immunostained with the parietal cell-specific antiH⫹-K⫹-ATPase ␤-subunit antibody (5). The use of direct in
situ RT-PCR (Fig. 2Ba) showed that strong staining occurred
mainly in the parietal cells, which was specific upon comparison with controls. No ExTaq polymerase (Fig. 2Bb), no
primers (Fig. 2Bc), no DIG-11-dUTP (Fig. 2Bd), and RNasepretreated (Fig. 2Be) samples had little or no staining throughout the glands. No DNase pretreatment (Fig. 2Bf) resulted in
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
C499
Fig. 4. Immunohistochemical localization of Kir2.1 K⫹ channel protein in gastric glands. Fixed, permeabilized gastric
glands were incubated with Kir2.1 antibody or with Kir2.1
antibody that was preincubated with its antigenic peptide
(⫹ pep), followed by Alexa Fluor 546-labeled goat anti-rabbit
IgG (red fluorescence). Fluorescence and Nomarski images are
shown. Specific staining of parietal cells in the gastric glands is
evident, which was blocked by the peptide. Scale bar: 10 ␮m.
AJP-Cell Physiol • VOL
Effect of low pHtrans and PKA on K⫹ channel activity in
resting gastric vesicles. As shown in Fig. 7A, trace a, current
recordings were obtained with resting gastric membranes in
400 mM K2SO4 solutions, pHtrans 7.4 at ⫺80 mV, and the
Fig. 5. Immunohistochemical colocalization of Kir2.1 with H⫹-K⫹-ATPase
and with ClC-2. Fluorescence/Nomarski micrographs of gastric glands stained
for Kir2.1 K⫹ channels and H⫹-K⫹-ATPase (H/K) or Kir2.1 K⫹ channels and
ClC-2 Cl⫺ channels are shown. Kir2.1 was labeled with Alexa Fluor 546 (red),
and H⫹-K⫹-ATPase and ClC-2 were labeled with Alexa Fluor 488 (green).
Colocalization is indicated by yellow color. Kir2.1 K⫹ channels, H⫹-K⫹ATPase, and ClC-2 Cl⫺ channels colocalize in parietal cells. Scale bar: 10 ␮m.
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
antibody was then used to immunolocalize the channel in
gastric glands as shown in Fig. 4. Kir2.1 K⫹ channel protein
was localized mainly to parietal cells, and this staining was
absent when the antibody was blocked with the peptide antigen. Staining was also absent in the absence of primary
antibody and in the absence of secondary antibody (data not
shown). In Fig. 5, gastric glands were stained for Kir2.1 and
the H⫹-K⫹-ATPase or for Kir2.1 and ClC-2. Yellow fluorescence indicates colocalization of respective pairs of proteins.
Kir2.1 colocalized with not only the H⫹-K⫹-ATPase but also
with ClC-2, suggesting that Kir2.1 is present in the same
membranes as the H⫹-K⫹-ATPase and the ClC-2 Cl⫺ channel,
where it is physiologically relevant and essential for HCl
secretion.
Measurement of acid-activated K⫹ channel activity in native
stimulated gastric membranes: conductance, I/V relationship,
open probability, and ion selectivity. A K⫹ channel that is
involved in HCl secretion must be present and active in
stimulated gastric vesicles that accumulate HCl and have active
ClC-2 Cl⫺ channels. Minimally, this K⫹ channel must be able
to function at pHtrans 3. Therefore, to detect the relevant K⫹
channel, we reconstituted stimulated vesicles into a planar lipid
bilayer with symmetric 400 mM K2SO4 (800 mM K⫹) solutions and asymmetric pHcis 7.4 (intracellular) and pHtrans 3.0
(extracellular). Figure 6A shows current recordings obtained at
different holding potentials and corresponding amplitude histograms. A channel was evident and very active with open
probabilities around 0.5 irrespective of holding potential
(0.52 ⫾ 0.06, n ⫽ 6). The I/V relationship of the channel
currents at pHtrans 3.0 (Fig. 6B) was linear with a slope
conductance of 10.8 ⫾ 0.2 pS (n ⫽ 16). To determine the ion
selectivity, we recorded channel currents in stimulated gastric
membranes at pHtrans 3 with an asymmetric fivefold reduction
of the K2SO4 concentration on the trans side from 400 mM
K2SO4 to 80 mM K2SO4 (800/160 mM K⫹, cis:trans). The I/V
relationship was linear (Fig. 6B). There was a significant (P ⬍
0.01) leftward shift of the reversal potential from zero under
symmetric conditions to ⫺29.1 ⫾ 0.7 mV (n ⫽ 3) under
gradient conditions. This is as expected for a highly K⫹selective channel, because the K⫹ equilibrium potential in
these conditions was calculated to be ⫺29.2 mV using Nernst’s
equation (29) and known activity coefficients (47). Na⫹ was
virtually impermeant through the channel (data not shown), as
also shown in earlier work (10, 46). Regulation of this channel
was then investigated by using gastric vesicles isolated from
resting gastric mucosae.
C500
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
channel was virtually closed with a few short openings. This is
in contrast to the channels in stimulated gastric membranes that
are open with a Po of 0.4–0.5 (see Fig. 6A). In Fig. 7A, trace
b, pHtrans was reduced to 3.0 and further recordings were
obtained. The channel opened to a Po of about 0.20. PKA
AJP-Cell Physiol • VOL
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
Fig. 6. Channel current recordings and amplitude histograms at pHtrans 3.0
using native stimulated rabbit gastric membrane vesicles and current-voltage
(I/V) relationship, conductance, and ion selectivity. A: channel current recordings were obtained at the indicated holding potentials from stimulated membrane vesicles in symmetrical 400 mM K2SO4 solution that also contained 2
mM MgSO4, 1 mM K-ATP (pH 7.4), and 10 mM K-EGTA. pHcis was buffered
to pH 7.4 with 10 mM K-HEPES, and pHtrans was buffered to pH 3.0 with 10
mM K-citrate. Corresponding amplitude histograms of the open and closed
states of the channel are shown at right. c– indicates the closed state and –
alone indicates the open state of the channel. B: I/V relationship of channel
currents recorded from stimulated vesicles with symmetric 400 mM K2SO4,
pHcis 7.4 and pHtrans 3.0 (F) and under 5-fold gradient conditions of 800:160
mM K⫹ (cis:trans) with sulfate as the accompanying anion, pHcis 7.4 and
pHtrans 3.0 (Œ). Data are plotted as means ⫾ SE with the number of reconstitutions indicated in parentheses. Arrow indicates reversal potential.
catalytic subunit (50 U/ml) was then added to the cis side, and
in Fig. 7A, traces c and d show sequential recordings of the
channel. The channel opened further, with the Po increasing to
about 0.30 and then to about 0.50. Corresponding amplitude
histograms are shown to the right of these traces. Figure 7B
shows a summary of the data obtained from four to eight
experiments. At ⫺80 mV and pHtrans 7.4, resting Po was
virtually closed. Reducing pHtrans to 3.0 significantly increased
resting channel Po to 0.25 ⫾ 0.03 (n ⫽ 8; P ⬍ 0.001), and
addition of PKA to the cis side further significantly increased
it to 0.46 ⫾ 0.03 (n ⫽ 8; P ⬍ 0.01), similar to the Po of
channels in stimulated gastric vesicles at pHtrans 3.0 (0.53 ⫾
0.05, n ⫽ 5). Therefore, with low pHtrans and treatment with
PKAcis, K⫹ channels in resting vesicles appear to open to the
stimulated level. The I/V relationship of channels in resting
vesicles was linear and the slope conductance was 10.7 ⫾ 0.2
pS (n ⫽ 17) at both pHtrans 7.4 and 3.0 and with or without
treatment with PKAcis, similar to that measured for K⫹ channels in stimulated vesicles, supporting the view that the same
K⫹ channel is present in both resting and stimulated vesicles.
These findings suggest that K⫹ channel activity is regulated by
PKA phosphorylation and by low extracellular pH, both of
which activate the channel.
Functional expression of the recombinant Kir2.1 K⫹ channel. Kir2.1 was expressed in CHO cells to examine effects of
forskolin/IBMX and low extracellular pH. Figure 8A shows
typical whole cell currents obtained at extracellular (bath) pH
(pHo) of 7.4 and 6.0. Addition of 5 ␮M forskolin with 20 ␮M
IBMX resulted in a large increase in the current at pHo 7.4.
When the medium pHo was lowered to 6.0, the K⫹ current
increased, but only at negative holding potentials. Upon addition of forskolin/IBMX, the K⫹ current increased at both
negative and positive holding potentials. The data from three to
four experiments are plotted as I/V curves after the currents
were normalized to cell capacitance. At pHo 6, the K⫹ channel
clearly showed inward rectification, but upon addition of forskolin/IBMX, the I/V curve became virtually linear. Also
shown are the currents and I/V curves of nontransfected CHO
cells. No significant responses to either low extracellular pH or
forskolin/IBMX occurred. This finding indicates that the currents measured in the transfected CHO cells are due to Kir2.1.
Figure 8B summarizes the effect of forskolin/IBMX and pHo 6
on normalized channel conductance (at ⫺140 mV holding
potential) in Kir2.1-transfected and nontransfected CHO cells.
Significant increases in normalized conductances were observed with forskolin/IBMX at pHo 7.4 (P ⬍ 0.05), when pHo
was lowered to 6 (P ⬍ 0.02), and when lowered pHo was
followed by forskolin/IBMX addition (P ⬍ 0.05) in transfected, but not in nontransfected, cells. Therefore, Kir2.1 K⫹
channel activity appears to be regulated by PKA phosphorylation and low extracellular pH, as observed with native gastric
vesicles. To further characterize recombinant Kir2.1, we examined its cation selectivity, and the results are shown in Fig.
9. From the I/V curves, the cation selectivity of Kir2.1 was
shown to be K⫹ ⫽ Rb⫹ ⬎⬎ Na⫹ ⫽ Cs⫹ ⫽ Li⫹ ⫽ NMDG⫹.
Na⫹, Cs⫹, Li⫹, and NMDG⫹ were all less permeant through
Kir2.1, as previously reported for some of these cations (49,
68). Kir2.1 is highly K⫹ selective and virtually impermeant to
Na⫹, correlating well with the K⫹ channels in native gastric
vesicles and with previous studies (10, 46).
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
C501
Recombinant Kir2.1 was also expressed in Xenopus laevis
oocytes. Oocyte plasma membranes expressing Kir2.1 were
isolated, fused to planar lipid bilayers, and examined for K⫹
currents by using symmetric 400 mM K2SO4, pHtrans 3.0, and
PKAcis, conditions under which the gastric K⫹ channel is
optimally functional. Figure 10 shows current recordings obtained at different holding potentials with corresponding amplitude histograms (A) and the I/V relationship (B). Three very
active channels were evident. The Po was about 0.5–0.7 irrespective of holding potential. The I/V relationship was linear,
and the slope conductance was 11.6 ⫾ 2.5 pS (n ⫽ 26), which
is not significantly different from the conductance of native K⫹
channels present in resting and stimulated gastric vesicles. The
ability to function at pHtrans 3.0 is similar to native gastric K⫹
channels. However, the kinetics of channel opening and closing appear somewhat different from those observed with native
gastric vesicles.
DISCUSSION
The purpose of the present study was to identify and localize
K⫹ channels in the gastric mucosa, in particular, the parietal
cells, to characterize and compare electrophysiological properties of native (from resting and stimulated rabbit gastric
mucosae) and recombinant K⫹ channels, and to investigate
whether/how they are regulated. K⫹ channels in gastric mucosal membrane vesicles, which also exhibited H⫹-K⫹-ATPase
activity, were suggested to transport K⫹ from the cytosol to the
lumen of the gastric parietal cell to provide an essential
continuous supply of K⫹ to the H⫹-K⫹-ATPase during stimulated HCl secretion (8, 10, 46, 65). RT-PCR showed that
AJP-Cell Physiol • VOL
Kir2.1 K⫹ channel mRNA was the very abundant in rabbit
gastric mucosa compared with Kir4.1 and Kir7.1, which were
present at low levels relative to ␤-actin. These studies therefore
focused on Kir2.1. Kir2.1 mRNA was localized to parietal cells
of gastric glands by in situ RT-PCR. Immunoconfocal microscopy showed that Kir2.1 was localized to the parietal cell and
intracellular membranes. It also colocalized with H⫹-K⫹ATPase and ClC-2 in the parietal cell, strongly supporting the
suggestion that Kir2.1 is a candidate K⫹ channel involved in
HCl secretion at the parietal cell apical membrane.
Native gastric K⫹ channels present in gastric mucosal membrane vesicles described in the present studies have different
electrophysiological characteristics from those described in the
basolateral membrane of the Necturus oxyntic cell (57, 62) and
the rabbit parietal cell (51) as studied by patch clamp. However, these native gastric K⫹ channels showed many characteristics in common with Kir2.1. Both native and recombinant
channels were highly K⫹ selective and virtually impermeant to
Na⫹. Regulation of both channels was similar. They were both
activated by low extracellular pH and intracellular PKA. Slope
conductance of native K⫹ channels was similar to that of
Kir2.1 expressed in oocyte membranes, although channel kinetics appeared somewhat different.
In the experimental conditions used, rectification was not
observed in native K⫹ channels, although Kir2.1 has been
described as a strongly inwardly rectified channel that, in the
absence of Mg2⫹, becomes linear (1, 30, 34, 39, 44, 63, 67).
However, recently this rectification has been shown to be not
an intrinsic gating property of the channel as first thought but,
rather, due to a very high affinity of the channel for intracel-
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
Fig. 7. Effect of pHtrans (pHt) 3.0 and PKA on K⫹ channel current in resting (R) gastric vesicles. A: current recordings (with
corresponding amplitude histograms at right) obtained with resting gastric membranes in 400 mM K2SO4 solutions, pHcis and
pHtrans 7.4, at ⫺80 mV (a); after pHtrans was reduced to 3.0 (b); and after addition of PKA catalytic subunit (50 U/ml) to the cis
side (c and d). Traces c and d are sequential recordings. c– and dashed line indicate the closed state and – alone indicates the open
state of the channel. B: summary of data obtained. For comparison, the hatched bar shows the open probability (Po) of stimulated
(S) channels at pHtrans 3.0. *P ⬍ 0.001; #P ⬍ 0.01 compared with preceding Po. Data are plotted as means ⫾ SE with the number
of experiments shown in parentheses over each bar.
C502
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
Fig. 8. Effect of forskolin/IBMX (F/I) and extracellular pH
(pHo) 6.0 on whole cell K⫹ currents in Chinese hamster
ovary (CHO) cells expressing rabbit recombinant Kir2.1
K⫹ channels. Representative scans (left) and I/V curves
(right) in Kir2.1-transfected CHO cells (A) and nontransfected CHO cells (B) are shown. Currents were measured
before and after addition of forskolin (5 ␮M) and IBMX (20
␮M) at pHo 7.4 or after reduction of pHo to 6.0. For the I/V
curves, 3–11 experiments were combined and currents were
normalized to capacitance. Data for A and B at ⫺140-mV
holding potential are summarized in C. Data are plotted as
means ⫾ SE with no. of experiments indicated in parentheses. Vm, membrane potential. #P ⬍ 0.01; **P ⬍ 0.02;
* P ⬍ 0.05.
lular cations such as Mg2⫹, polyamines, and trace contaminants in organic buffers (e.g., HEPES) (23–25). High K⫹
concentrations also reduce rectification by displacing the cations, as do chelators such as EDTA, ATP, and EGTA, which
chelate Mg2⫹ to greater or lesser extents. Because experiments
with native channels were performed with intracellular ATP,
AJP-Cell Physiol • VOL
HEPES, EGTA, and Mg2⫹ present in Ca2⫹-free conditions and
a high K⫹ concentration of 800 mM, lack of rectification is not
surprising. I/V curves of recombinant Kir2.1 expressed in CHO
cells were also linear (without and with forskolin/IBMX) but
appeared inwardly rectified when extracellular pH was reduced
in the absence of forskolin/IBMX.
286 • MARCH 2004 •
www.ajpcell.org
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
Native gastric K⫹ channels and recombinant Kir2.1 showed
significant activation by low extracellular pH. As opposed to
voltage-gated K⫹ and Na⫹ channels, which are nonconductive
in their fully protonated state (27, 28), mouse Kir2.1 channel
remained highly permeable to K⫹ when fully extracellularly
protonated, although the conductance was 30–38% of the
initial level (48). However, caution is needed when comparing
channels of different species because a single amino acid
change can exert a major effect on function. It should be noted
that rabbit (and human) Kir2.1 have one additional externally
facing glutamic acid (E118) compared with mouse and rat and
that a single additional negatively charged glutamic acid
(E419) on the external surface of ClC-2 is the key to major
channel activation by low pH (56). Native K⫹ channels
(whether stimulated or resting) had a linear I/V relationship
AJP-Cell Physiol • VOL
with a slope conductance of 11 pS, and lowering pHo had no
significant difference on channel conductance. Because the
proteins of the parietal cell secretory membrane are in direct
contact with the primary gastric secretion HCl (6), at the very
least, the K⫹ channel must be functional at acid pH on the
extracellular (luminal or trans) side. With the use of native
stimulated vesicles, the K⫹ channel was able to function at low
pH. K⫹ channels of resting gastric vesicles at pHtrans 7.4 were
virtually closed (Po ⬇ 0.02) and acidic pHtrans activated the
channels to Po ⬇ 0.25.
Histamine activation of acid secretion is mediated by increased intracellular cAMP leading to PKA activation (4) and
an increase in intracellular Ca2⫹ (3). In the present studies K⫹
channel currents were routinely recorded in Ca2⫹-free conditions in the presence of EGTA, consistent with the established
findings that HCl accumulation by these vesicle preparations
does not require Ca2⫹ (10, 65) and indicating that these K⫹
channels are Ca2⫹ independent. PKA addition to the cis (intracellular) side of native resting K⫹ channels resulted in
activation of the Po to ⬃0.5, a level similar to that measured
using stimulated vesicles. Similar I/V characteristics and conductance of resting and stimulated K⫹ channels suggest that
the same channel is present in gastric mucosal membranes
from resting and stimulated gastric mucosae. This suggestion is
supported by the Western blotting of resting and stimulated
gastric vesicles, which both showed the presence of Kir2.1.
Resting K⫹ channel activation by PKA phosphorylation at the
cytosolic side and by low pH on the extracellular side is
consistent with the channel playing a physiological role in HCl
secretion. When Kir2.1 was expressed in CHO cells, significant increases in normalized currents and conductances were
observed with forskolin/IBMX at pHo 7.4, when pHo was
lowered to 6, and when lowered pHo was followed by forskolin/IBMX addition in transfected, but not in nontransfected,
cells. Thus, as observed with native gastric vesicles, Kir2.1 K⫹
channel activity is regulated by PKA phosphorylation and low
extracellular pH. PKA regulation of Kir2.1 has been controversial in studies by others, with effects varying from inhibition (64) to activation (17) and lack of effect (31). Further
studies are needed to fully understand the structural basis of
rectification and activation by extracellular pH and PKA.
Two other K⫹ channels (Kir4.1 and KCNQ1) have been
localized to the gastric parietal cell and shown to colocalize
with H⫹-K⫹-ATPase (18, 20). Fujita et al. (18) found Kir4.1,
Kir4.2, and Kir7.1 (but not Kir2.1) in rat gastric mucosa by
using RT-PCR. Kir4.1 was immunolocalized to the rabbit
gastric parietal cell, where it appeared to colocalize with the
␤-subunit of the H⫹-K⫹-ATPase. Although the data were not
shown, Fujita et al. stated that immunoreactivity of Kir4.2 was
localized to surface mucous cells and that Kir7.1 was absent
from the epithelium but was detected in “neural regions” of the
gastric mucosa. Others (13, 40) have shown that Kir7.1 is
located in the basolateral membrane in the kidney and is
thought to be involved with providing K⫹ to the Na⫹-K⫹ATPase. Functional data in the study by Fujita et al. were
minimal, limited to Ba2⫹ inhibition of [14C]aminopyrine uptake in intact isolated parietal cells, where effects on basolateral K⫹ channels were not eliminated or taken into consideration. Previous work (8) briefly reported that Ba2⫹ and TEA
inhibited K⫹-dependent Cl⫺ uptake by isolated stimulated
rabbit gastric vesicles. TEA inhibited [14C]aminopyrine uptake
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
Fig. 9. Cation selectivity of recombinant Kir2.1 expressed in CHO cells. I/V
curves for K⫹, Rb⫹, Na⫹, and N-methyl-D-glucamine (NMDG⫹) (A) and I/V
curves for K⫹, Cs⫹, Li⫹, and NMDG⫹ (B) are shown. Currents were
normalized to capacitance. Data are plotted as means ⫾ SE (n ⫽ 3). Insets
show reversal potentials (arrow). Both pipette and bath solutions contained 140
mM K-gluconate, 5 mM MgCl2, 10 mM HEPES, 5 mM EGTA, and 1 mM
MgATP (140K) at pH 7.3, initially. The bath solution was then changed to 28
mM K-gluconate plus 112 mM NMDG (K28NMDG112), NaCl (K28Na112),
RbCl (K28Rb112), CsCl (K28Cs112), or LiCl (K28Li112).
C503
C504
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
as well as H⫹-K⫹-ATPase-related ATP hydrolysis by stimulated rabbit parietal cells in which the basolateral membrane
was permeabilized by digitonin. NH4Cl overcame TEA inhibition of both K⫹-dependent Cl⫺ uptake and H⫹-K⫹-ATPaserelated ATP hydrolysis, indicating the presence of a K⫹ channel in parietal cell secretory membranes. However, Ba2⫹ and
TEA are general, nonspecific K⫹ channel inhibitors and do not
indicate the molecular identity of the channel. Fujita et al. also
showed that recombinant Kir4.1 K⫹ currents were not affected
by reduction of extracellular pH, a property essential for an
apical K⫹ channel. Detailed functional comparison between
Kir4.1 and native gastric K⫹ channel function was lacking.
Grahammer et al. (20) reported that KCNQ1, KCNE2, and
KCNE3 mRNA were present in human stomach and that
KCNQ1 localized in human parietal cells together with the
H⫹-K⫹-ATPase. Chromanol 293B, presumed by Grahammer
et al. to be a specific KCNQ1 inhibitor, inhibited acid secretion
in the rat perfused stomach, dog Heidenhain pouch, and intact
rabbit gastric glands. Coexpression of recombinant KCNQ1
with KCNE2, but not KCNE3, in COS cells resulted in extracellular acidic pH activation of K⫹ currents. Thus Grahammer
et al. suggested that KCNQ1, when coassembled with KCNE2
or KCNE3, is the K⫹ channel complex involved in K⫹ recycling at the secretory membrane essential for acid secretion.
Others have shown that KCNQ1/KCNE3 can be activated by
PKA (52), but this was not examined by Grahammer et al. In
contrast, H⫹ transport by H⫹-K⫹-ATPase-containing vesicles
from stimulated rabbits is insensitive to chromanol 293B, and
K⫹ channel currents measured in CHO cells expressing recomAJP-Cell Physiol • VOL
binant Kir2.1 were activated by 293B (12) with no effect
observed in nontransfected CHO cells. These findings strongly
argue against involvement of KCNQ1 in K⫹ recycling and
suggest that chromanol 293B has an alternative, unidentified
target in the parietal cell.
Nevertheless, it seems that Kir2.1, Kir4.1, and KCNQ1 are
all present in the gastric parietal cell and seem to colocalize
with H⫹-K⫹-ATPase and, in the case of Kir2.1, also with
ClC-2. K⫹ channel diversity is well established, and not only
have there been different channels from the same family
expressed in the same tissue (45) but also a single tissue, for
example, rat pituitary, has been reported to express transcripts
for 21 different Kv channel ␣-subunits [including Shakerrelated Kv, ether-à-go-go (eag), and KCNQ1–3], 9 Kir channel
␣-subunits, and 7 auxiliary subunits (e.g., Kv␤1–3, minK,
sulfonylurea receptor SUR1–2) (66). Such large diversity of
K⫹ channels has also been described in cardiac cells (41). The
clonal pituitary tumor cell line GH3/B6 has transcripts for 9 Kv,
6 eag, 3 KCNQ, and 5 Kir ␣-subunits and no auxiliary subunits
(66), suggesting that if the proteins are expressed, then these
channels are all present in the same cell. These channels cannot
be distinguished from each other on the basis of electrophysiology or pharmacology. Heterooligomeric assembly of different Kir K⫹ channels has been studied and shown to result in
different functional/electrophysiological properties compared
with homooligomeric assembly of four subunits, suggesting
interactions between different Kir K⫹ channel subunits (14, 16,
32, 41, 59, 67). Coexpression of auxiliary subunits with ␣-subunits can affect function and in some cases can be essential for
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
Fig. 10. Functional expression of Kir2.1 in Xenopus oocytes. Plasma membranes of oocytes expressing rabbit Kir2.1 were isolated, fused to planar lipid bilayers, and examined for K⫹ currents
using symmetric 400 mM K2SO4, pHtrans 3.0, and
PKAcis (50 U/ml), conditions under which the
gastric K⫹ channel is optimally functional. A:
current recordings and corresponding amplitude
histograms (right) obtained at the indicated holding potentials. Medium also contained 2 mM
MgSO4, 1 mM K-ATP (pH 7.4), 10 mM EGTA
(pH 7.4), and 10 mM K-citrate (pH 3). c– indicates the closed state and – alone indicates the
open state of the channel. B: I/V curve obtained
from data in A. Data are plotted as means ⫾ SE
with number of experiments indicated in parentheses.
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
ACKNOWLEDGMENTS
We thank Heidi Short and Elena Y. Kupert for technical help during the
early stages of these studies. We also thank Glenn Doerman for help with the
figures.
GRANTS
This work was supported by National Institute of Diabetes and Digestive
and Kidney Diseases Grant DK-43816 and National Heart, Lung, and Blood
Institute Grant HL-58399 (to J. Cuppoletti and D. H. Malinowska) and Army
Research Office Multidisciplinary Research Program of the University Research Initiative (ARO-MURI) Grant DAAD-19-02-1-0227 (to J. Cuppoletti).
AJP-Cell Physiol • VOL
REFERENCES
1. Aleksandrov A, Velimirovic B, and Clapham DE. Inward rectification
of the IRK1 K⫹ channel reconstituted in lipid bilayers. Biophys J 70:
2680–2687, 1996.
2. Bond CT, Pessia M, Xia XM, Lagrutta A, Kavanaugh MP, and
Adelman JP. Cloning and expression of a family of inward rectifier
potassium channels. Receptors Channels 2: 183–191, 1994.
3. Chew CS. Cholecystokinin, carbachol, gastrin, histamine, and forskolin
increase [Ca2⫹]i in gastric glands. Am J Physiol Gastrointest Liver Physiol
250: G814–G823, 1986.
4. Chew CS, Hersey SJ, Sachs G, and Berglindh T. Histamine responsiveness of isolated gastric glands. Am J Physiol Gastrointest Liver
Physiol 238: G221–G229, 1980.
5. Chow DC and Forte JG. Characterization of the ␤-subunit of the
H⫹-K⫹-ATPase using an inhibitory monoclonal antibody. Am J Physiol
Cell Physiol 265: C1562–C1570, 1993.
6. Cuppoletti J, Baker AM, and Malinowska DH. Cl⫺ channels of the
gastric parietal cell that are active at low pH. Am J Physiol Cell Physiol
264: C1609–C1618, 1993.
7. Cuppoletti J, Baker AM, and Malinowska DH. Acidophilic Cl⫺ and K⫹
channels of the gastric parietal cell: a new model of regulated HCl
secretion. In: Proceedings of NATO Advanced Research Workshop on
Molecular and Cellular Mechanisms of H⫹ Transport, edited by Hirst BH.
Berlin: Springer, 1994, vol. 89, p. 111–118.
8. Cuppoletti J and Malinowska DH. K⫹ channel in the gastric parietal cell
apical membrane (Abstract). Gastroenterology 94: A82, 1988.
9. Cuppoletti J and Malinowska DH. PKA-activated gastric K⫹ channel
that is active at low pH (Abstract). FASEB J 9: A3, 1995.
10. Cuppoletti J and Sachs G. Regulation of gastric acid secretion via
modulation of a chloride conductance. J Biol Chem 259: 14952–14959,
1984.
11. Cuppoletti J, Tewari KP, Sherry AM, Kupert EY, and Malinowska
DH. ClC-2 Cl⫺ channels in human lung epithelia: activation by arachidonic acid, amidation, and acid-activated omeprazole. Am J Physiol Cell
Physiol 281: C46–C54, 2001.
12. Cuppoletti J, Tewari KP, Sherry AM, and Malinowska DH. Kir2.1 K⫹
channels of the gastric parietal cell. In: Mechanisms and Consequences of
Proton Transport, edited by Urushidani T, Forte JG, and Sachs G. Boston,
MA: Kluwer Academic, 2002, p. 255–264.
13. Derst C, Hirsch JR, Preisig-Muller R, Wischmeyer E, Karschin A,
Doring F, Thomzig A, Veh RW, Schlatter E, Kummer W, and Daut J.
Cellular localization of the potassium channel Kir7.1 in guinea pig and
human kidney. Kidney Int 59: 2197–2205, 2001.
14. Derst C, Karschin C, Wischmeyer E, Hirsch JR, Preisig-Muller R,
Rajan S, Engel H, Grzeschik KH, Daut J, and Karschin A. Genetic and
functional linkage of Kir5.1 and Kir2.1 channel subunits. FEBS Lett 491:
305–311, 2001.
15. Doring F, Derst C, Wischmeyer E, Karschin C, Schneggenburger R,
Daut J, and Karschin A. The epithelial inward rectifier channel Kir7.1
displays unusual K⫹ permeation properties. J Neurosci 18: 8625–8636, 1998.
16. Fakler B, Bond CT, Adelman JP, and Ruppersberg JP. Heterooligomeric
assembly of inward-rectifier K⫹ channels from subunits of different subfamilies: Kir2.1 (IRK1) and Kir4.1 (BIR10). Pflügers Arch 433: 77–83, 1996.
17. Fakler B, Brandla U, Glowatzki E, Zenner HP, and Ruppersberg JP.
Kir2.1 inward rectifier K⫹ channels are regulated independently by protein
kinases and ATP hydrolysis. Neuron 13: 1413–1420, 1994.
18. Fujita A, Horio Y, Higashi K, Mouri T, Hata F, Takeguchi N, and
Kurachi Y. Specific localization of an inwardly rectifying K⫹ channel,
Kir4.1, at the apical membrane of rat gastric parietal cells; its possible
involvement in K⫹ recycling for the H⫹-K⫹-pump. J Physiol 540: 85–92,
2002.
19. Ganser AL and Forte JG. K⫹-stimulated ATPase in purified microsomes
of bullfrog oxyntic cells. Biochim Biophys Acta 307: 169–180, 1973.
20. Grahammer F, Herling AW, Lang HJ, Schmitt-Grass A, Wittekindt
OH, Nitschke R, Bleich M, Barhanin J, and Warth R. The cardiac K⫹
channel KCNQ1 is essential for gastric acid secretion. Gastroenterology
120: 1363–1371, 2001.
21. Gribble FM, Ashfield R, Ammala C, and Ashcroft FM. Properties of
cloned ATP-sensitive K⫹ currents expressed in Xenopus oocytes.
J Physiol 498: 87–98, 1997.
22. Gu J. In Situ Polymerase Chain Reaction and Related Technology.
Natick, MA: Eaton, 1995.
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
functional channel activity (e.g., Kir6.1 with SUR1) (21).
Kir2.1 is not affected by SUR1 (21). These types of interactions can result in K⫹ current inhibition, activation, altered
kinetics, rectification, and pH sensitivity; have been shown to
be specific and occur between small domains in the cytoplasmic region of the channels; and may depend in part on the ratio
of the amounts of various subunits present. Which K⫹ channel
subunits are present together in a single cell type is very
important and can have major functional consequences.
KCNE1 expressed with KCNQ1 increases the K⫹ current,
slows activation, and shifts voltage dependence toward more
positive potentials (52). In contrast, KCNE2 expressed with
KCNQ1 results in a greatly decreased K⫹ current and altered
gating properties compared with KCNQ1 alone (61), whereas
KCNE3 expressed with KCNQ1 results in a constitutively
open K⫹ channel that can be further activated by PKA (52).
KCNE3 suppresses KCNQ4 currents (52).
Whether Kir2.1, Kir4.1, and KCNQ1 (with KCNE2 or
KCNE3) coexist side by side in the gastric mucosal membranes
in parietal cells and are activated at appropriate times by
appropriate activators or whether they assemble and/or function cooperatively with each other or other subunits, including
auxiliary ones, remains to be examined. Such interactions
might explain the difference in the gating kinetics observed
between native gastric vesicles and Kir2.1 expressed in oocyte
membranes. Detailed functional studies and comparison of
behavior and kinetics of native gastric K⫹ channels with
recombinant K⫹ channels separately and coexpressed in CHO
cells might lead to a further understanding of how K⫹ is moved
from cytosol to lumen to enable acid secretion to occur.
Although care must be used in transposition of the results of
studies on isolated membranes to the native environment of the
cell, the present study provides electrophysiological, biochemical, and immunological evidence for a K⫹ channel, Kir2.1, in
gastric mucosal membranes with the required properties for
involvement in acid secretion and that colocalizes with both
H⫹-K⫹-ATPase and ClC-2 Cl⫺ channels. Similar in vivo
effects of histamine on a voltage-dependent rabbit gastric Cl⫺
channel, ClC-2, present in resting and stimulated gastric membranes, have been previously reported (6, 35). In these studies,
both native gastric resting and recombinant ClC-2 Cl⫺ channels are activated by PKA and low pH . Protein kinase action
and extracellular pH provide the means by which to control
both K⫹ and Cl⫺ efflux across the parietal cell secretory
membrane. The findings of the present studies strongly suggest
that Kir2.1 is present in gastric mucosal membranes of the
parietal cell and may be involved in acid secretion.
C505
C506
GASTRIC PARIETAL CELL APICAL K⫹ CHANNEL
AJP-Cell Physiol • VOL
47. Robinson RA and Stokes RH. The measurement and interpretation of
conductance, chemical potential, and diffusion in solutions of simple
electrolytes. In: Electrolyte Solutions (2nd ed.). London: Butterworth,
1970, appendix 8.10, Table 15, p. 501.
48. Sabirov RZ, Okada Y, and Oiki S. Two-sided action of protons on an
inward rectifier K⫹ channel (IRK1). Pflügers Arch 433: 428–434, 1997.
49. Sabirov RZ, Tominaga T, Miwa A, Okada Y, and Oiki S. A conserved
arginine residue in the pore region of an inward rectifier K channel (IRK1) as
an external barrier for cationic blockers. J Gen Physiol 110: 665–677, 1997.
50. Sachs G, Chang HH, Rabon E, Schackman R, Lewin M, and Saccomani G. A nonelectrogenic H⫹ pump in plasma membranes of hog
stomach. J Biol Chem 251: 7690–7698, 1976.
51. Sakai H, Okada Y, Morii M, and Takeguchi N. Anion and cation
channels in the basolateral membrane of rabbit parietal cells. Pflügers
Arch 414: 185–192, 1989.
52. Schroeder BC, Waldegger S, Fehr S, Bleich M, Warth R, Greger R,
and Jentsch TJ. A constitutively open potassium channel formed by
KCNQ1 and KCNE3. Nature 403: 196–199, 2000.
53. Schwalbe RA, Rudin A, Xia SL, and Wingo CS. Site-directed glycosylation tagging of functional Kir2.1 reveals that the putative pore-forming
segment is extracellular. J Biol Chem 277: 24382–24389, 2002.
54. Sherry AM, Malinowska DH, Morris RE, Ciraolo GM, and Cuppoletti J. Localization of ClC-2 Cl⫺ channels in rabbit gastric mucosa. Am J
Physiol Cell Physiol 280: C1599–C1606, 2001.
55. Sherry AM, Stroffekova K, Knapp LM, Kupert EY, Cuppoletti J, and
Malinowska DH. Characterization of the human pH- and PKA-activated
ClC-2G(2␣) Cl⫺ channel. Am J Physiol Cell Physiol 273: C384–C393, 1997.
56. Stroffekova K, Kupert EY, Malinowska DH, and Cuppoletti J. Identification of the pH sensor and activation by chemical modification of the
ClC-2G Cl⫺ channel. Am J Physiol Cell Physiol 275: C1113–C1123, 1998.
57. Supplisson S, Loo DDF, and Sachs G. Diversity of K⫹ channels in the
basolateral membrane of resting Necturus oxyntic cells. J Membr Biol
123: 209–221, 1991.
58. Takumi T, Ishii T, Horio Y, Morishige KI, Takahashi N, Yamada M,
Yamashita T, Kiyama H, Sohmiya K, Nakanishi S, and Kurachi Y. A
novel ATP-dependent inward rectifier potassium channel expressed predominantly in glial cells. J Biol Chem 270: 16339–16346, 1995.
59. Tanemoto M, Kittaka N, Inanobe A, and Kurachi Y. In vivo formation
of a proton-sensitive K⫹ channel by heteromeric subunit assembly of
Kir5.1 with Kir4.1. J Physiol 525: 587–592, 2000.
60. Tewari KP, Malinowska DH, Sherry AM, and Cuppoletti J. PKA and
arachidonic acid activation of human recombinant ClC-2 chloride channels. Am J Physiol Cell Physiol 279: C40–C50, 2000.
61. Tinel N, Diochot Borsotto M, Lazdunski M, and Barhanin J. KCNE2
confers background current characteristics to the cardiac KCNQ1 potassium channel. EMBO J 19: 6326–6330, 2000.
62. Ueda S, Loo DDF, and Sachs G. Regulation of K⫹ channels in the
basolateral membrane of Necturus oxyntic cells. J Membr Biol 97: 31–41,
1987.
63. Wible BA, Taglialatela M, Ficker E, and Brown AM. Gating of
inwardly rectifying K⫹ channels localized to a single negatively charged
residue. Nature 371: 246–249, 1994.
64. Wischmeyer E and Karschin A. Receptor stimulation causes slow
inhibition of IRK1 inwardly rectifying K⫹ channels by direct protein
kinase A-mediated phosphorylation. Proc Natl Acad Sci USA 93: 5819–
5823, 1996.
65. Wolosin JM and Forte JG. Stimulation of oxyntic cell triggers K⫹ and
Cl⫺ conductance in apical H⫹-K⫹-ATPase membranes. Am J Physiol
Gastrointest Liver Physiol 248: G595–G607, 1985.
66. Wulfsen I, Hauber HP, Schiemann D, Bauer CK, and Schwarz JR.
Expression of mRNA for voltage-dependent and inward rectifying K
channels in GH3/B6 cells and rat pituitary. J Neuroendocrinol 12: 263–
272, 2000.
67. Yang J, Jan YN, and Jan LY. Determination of the subunit stoichiometry of
an inwardly rectifying potassium channel. Neuron 15: 1441–1447, 1995.
68. Yang J, Yu M, Jan YN, and Jan LY. Stabilization of ion selectivity filter
by pore loop ion pairs in an inward rectifying potassium channel. Proc
Natl Acad Sci USA 94: 1568–1572, 1997.
69. Zsengeller ZK, Wert SE, Bachurski CJ, Kirwin KL, Trapnell BC, and
Whitsett JA. Recombinant adenoviral vector disrupts surfactant homeostasis in mouse lung. Hum Gene Ther 8: 1331–1344, 1997.
286 • MARCH 2004 •
www.ajpcell.org
Downloaded from http://ajpcell.physiology.org/ by 10.220.33.3 on July 31, 2017
23. Guo D and Lu Z. Pore block versus intrinsic gating in the mechanism of
inward rectification in strongly rectifying IRK1 channels. J Gen Physiol
116: 561–568, 2000.
24. Guo D and Lu Z. IRK1 inward rectifier K⫹ channels exhibit no intrinsic
rectification. J Gen Physiol 120: 539–551, 2002.
25. Guo D, Ramu Y, Klem AM, and Lu Z. Mechanism of rectification in
inward-rectifier K⫹ channels. J Gen Physiol 121: 261–275, 2003.
26. Hanzel DK, Urushidani T, Usinger WR, Smolka A, and Forte JG.
Immunological localization of an 80-kDa phosphoprotein to the apical
membrane of gastric parietal cells. Am J Physiol Gastrointest Liver
Physiol 256: G1082–G1089, 1989.
27. Hille B. Charges and potentials at the nerve surface. Divalent ions and pH.
J Gen Physiol 51: 221–236, 1968.
28. Hille B. Potassium channels in myelinated nerve. Selective permeability to
small cations. J Gen Physiol 61: 669–686, 1973.
29. Hille B. Selective permeability: independence. In: Ionic Channels of
Excitable Membranes (2nd ed.). Sunderland, MA: Sinauer, 1992, p.
337–361.
30. Ishii K, Yamagishi T, and Taira N. Cloning and functional expression of
a cardiac inward rectifier K⫹ channel. FEBS Lett 338: 107–111, 1994.
31. Kamouchi M, Van Den Bremt K, Eggermont J, Droogmans G, and Nilius
B. Modulation of inwardly rectifying potassium channels in cultured bovine
pulmonary artery endothelial cells. J Physiol 504.3: 545–556, 1997.
32. Konstas AA, Korbmacher C, and Tucker SJ. Identification of domains
that control the heteromeric assembly of Kir5.1/Kir4.0 potassium channels. Am J Physiol Cell Physiol 284: C910–C917, 2003.
33. Krapivinsky G, Medina I, Eng L, Krapivinsky L, Yang Y, and
Clapham DE. A novel inward rectifier K⫹ channel with unique pore
properties. Neuron 20: 995–1005, 1998.
34. Lopatin AN and Nichols CG. [K⫹] dependence of open-channel conductance in cloned inward rectifier potassium channels (IRK1, Kir2.1).
Biophys J 71: 682–694, 1996.
35. Malinowska DH, Kupert EY, Bahinsky A, Sherry AM, and Cuppoletti
J. Cloning, functional expression, and characterization of a PKA-activated
gastric Cl⫺ channel. Am J Physiol Cell Physiol 268: C191–C200, 1995.
36. Malinowska DH, Kupert EY, Sherry AM, and Cuppoletti J. A candidate K⫹ channel involved in gastric HCl secretion (Abstract). FASEB J 14:
A338, 2000.
37. Mathews CJ, Tabcharani JA, Chang XB, Jensen TJ, Riordan JR, and
Hanrahan JW. Dibasic protein kinase A sites regulate bursting rate and
nucleotide sensitivity of the cystic fibrosis transmembrane conductance
regulator chloride channel. J Physiol 508: 365–377, 1998.
38. Murray CB, Morales MM, Flotte TR, McGrath-Morrow SA, Guggino
WB, and Zeitlin PL. ClC-2: a developmentally dependent chloride
channel expressed in the fetal lung and down-regulated after birth. Am J
Respir Cell Mol Biol 12: 597–604, 1995.
39. Nagashima M, Ishii K, Tohse N, Taira N, and Yabu H. Unitary current
through the inward rectifier K⫹ channel cloned from rabbit heart— comparison with the native K⫹ channel. J Mol Cell Cardiol 28: 957–965, 1996.
40. Nakamura N, Suzuki Y, Sakuta H, Ookata K, Kawahara K, and
Hirose S. Inward rectifying K⫹ channel Kir7.1 is highly expressed in
thyroid follicular cells, intestinal epithelial cells and choroid plexus
epithelial cells: implication for a functional coupling with Na⫹,K⫹ATPase. Biochem J 342: 329–336, 1999.
41. Nerbonne JM, Nichols CG, Schwarz TL, and Escande D. Genetic
manipulation of cardiac K⫹ channel function in mice. Circ Res 89:
944–956, 2001.
42. Nuovo GJ. PCR In Situ Hybridization: Protocols and Applications (3rd
ed.). Philadelphia, PA: Lippincott-Raven, 1997.
43. Partiseti M, Collura V, Agnel M, Culouscou JM, and Graham D.
Cloning and characterization of a novel human inwardly rectifying potassium channel predominantly expressed in small intestine. FEBS Lett 434:
171–179, 1998.
44. Raab-Graham KF, Radeke CM, and Vandenberg CA. Molecular
cloning and expression of a human heart inward rectifier potassium
channel. Neuroreport 5: 2501–2505, 1994.
45. Raap M, Biedermann B, Braun P, Milenkovic I, Skatchkov SN,
Bringmann A, and Reichenbach A. Diversity of Kir channel subunit
mRNA expressed by retinal glial cells of the guinea-pig. Neuroreport 13:
1037–1040, 2002.
46. Reenstra WW and Forte JG. Characterization of K⫹ and Cl⫺ conductances in apical membrane vesicles from stimulated rabbit oxyntic cells.
Am J Physiol Gastrointest Liver Physiol 259: G850–G858, 1990.