0013-7227/03/$15.00/0
Printed in U.S.A.
Endocrinology 144(9):4087– 4096
Copyright © 2003 by The Endocrine Society
doi: 10.1210/en.2003-0418
Regulation of Water Absorption in the Frog Skins by
Two Vasotocin-Dependent Water-Channel Aquaporins,
AQP-h2 and AQP-h3
TAKAHIRO HASEGAWA, HARUNA TANII, MASAKAZU SUZUKI,
AND
SHIGEYASU TANAKA
Department of Biology, Faculty of Science, Shizuoka University, Shizuoka 422-8529, Japan
A new frog aquaporin (AQP) cDNA was cloned from a cDNA
library constructed from the ventral skin of the tree frog Hyla
japonica. This AQP (Hyla AQP-h2) consisted of 268 amino acid
residues with a high homology to mammalian AQP2. The predicted amino acid sequence contained the two conserved AsnPro-Ala motifs found in all the major intrinsic protein family
members and the putative six transmembrane domains. The
sequence also contained a mercurial compound: cysteine, one
potential N-glycosylation site at Asn-124, and a putative phosphorylation site recognized by protein kinase A at Ser-262. In
a swelling assay using Xenopus oocytes, AQP-h2 facilitated
water permeability, especially in response to cAMP. Expression of AQP-h2 mRNA was restricted to several tissues including the ventral skin, kidney, and urinary bladder; but with
A
MPHIBIANS DO NOT drink water through their
mouth. Instead, they possess a specialized region in
the ventral skin that, compared with that of other tetrapods,
is highly permeable to water and ions as well as to respiratory
gases (1–3). Water movement occurs across plasma membranes of various cells of animals, plants, and microorganisms through specialized water-channel proteins called
aquaporins (AQPs). Aquaporins form membrane pores selectively permeable to water and, isoform dependently, to
certain small solutes such as glycerol and urea. In mammals,
11 isoforms of AQP have been identified (AQP0 –AQP10;
Refs. 4 –7). Several AQP isoforms such as AQP1 display a
ubiquitous tissue distribution, whereas other AQP isoforms
display tissue-specific expression: for example, AQP0 [originally named major intrinsic protein (MIP) 26] in the eye lens
(8), as well as AQP2 in the apical plasma membrane (9) and
AQP3 in the basolateral membrane (10 –12) of the kidney
collecting duct. Accordingly, water channels have been
thought to exist in the ventral skin of amphibians. Indeed, for
a long time, frog skin, like the urinary bladder, has been used
as a useful model for investigating antidiuretic hormone
(ADH)-mediated regulation of transepithelial water permeability (13). Freeze fracture electron microscopical studies
have suggested that certain intramembrane particles in the
amphibian urinary bladder and skin may represent waterchannel proteins because of increasing water permeability
with ADH (14 –16). Recently, frog AQP cDNAs were cloned
Abbreviations: ADH, Antidiuretic hormone; AQP, aquaporin; DAPI,
4⬘,6-diamidino-2-phenylindole; DIG, digoxigenin; dNTP, deoxynucleotide triphosphate; MIP, major intrinsic protein; NPA, Asn-Pro-Ala; Pf,
water permeability; poly (A)⫹, polyadenylated.
immunofluorescence staining using an antipeptide antibody
(ST-140) against the AQP-h2 protein, immunopositive cells
were found only in the ventral skin and urinary bladder. In
the ventral pelvic skin, the label for AQP-h2 was localized in
the entire plasma membrane of the granular cells beneath the
outmost layer of the skin and in the basolateral membrane of
the granular cells in this layer. In response to vasotocin, however, the label for AQP-h2 became more intense in the apical
membrane in the granular cells of the outermost layer, similar
to the case for the earlier studied AQP-h3, which was specifically expressed in the ventral skin. Taken together, these
findings suggest that not only AQP-h3, but also AQP-h2 acts as
a regulator of the water balance in this frog. (Endocrinology
144: 4087– 4096, 2003)
from the urinary bladder of Rana esculenta (17) and Bufo
marinus (18), and they showed high homology to mammalian
AQP1. More recently, we cloned two types of cDNA encoding AQP from a cDNA library constructed from the ventral
skin of the tree frog, Hyla japonica: one (AQP-h1) was homologous to mammalian AQP1, and showed a ubiquitous
tissue distribution. The other (AQP-h3) displayed a specific
distribution in the ventral skin and had an amino acid sequence homologous to that of mammalian AQP2 (19).
Among mammalian AQPs, AQP2 is an ADH-dependent
AQP: in response to ADH, AQP2 moves from its cytoplasmic
pool to the apical plasma membrane, thereby causing water
to enter the cells (20, 21). Identification of the frog vasotocin
(counterpart of mammalian ADH)-dependent AQP is important to elucidate the regulatory mechanism of wateradaptation systems to maintain the water balance in
amphibians, but there is no evidence concerning a vasotocindependent AQP in frogs.
In this study, we identified a previously unrecognized
member of the amphibian AQP family, AQP-h2, in the ventral pelvic skin of the tree frog, and demonstrated that both
it and the previously identified AQP-h3 are vasotocinregulated AQP in amphibians.
Materials and Methods
Animals
Adult tree frogs (H. japonica) were captured in a field near our university, kept under laboratory conditions, and fed crickets. The ventral
pelvic skin was removed under anesthesia with MS 222 (Nacalai tesque,
Kyoto, Japan), and then processed for cDNA cloning. Similarly, several
tissues from the frogs were used for experiments on mRNA expression,
protein expression by Western blot analysis, and immunocytochemical
analysis.
4087
4088
Endocrinology, September 2003, 144(9):4087– 4096
Construction of the frog ventral skin cDNA library
Total RNA was prepared from 0.28 g of frog skin by using TRIZOL
RNA extraction reagent (Life Technologies, Rockville, MD). Then 11.52
␮g of polyadenylated [poly (A)⫹] RNA was selected from about 572 ␮g
of the total RNA by using Oligotex-dT30 super (Takara, Kyoto, Japan).
We constructed a ␭ZAP cDNA library (9.6 ⫻ 106 plaque-forming
units/␮g of arms) from the poly (A)⫹RNA by use of a ZAP express
cDNA synthesis kit and a Gigapack III Gold cloning kit (Stratagene, La
Jolla, CA), in accordance with the manufacturer’s instructions.
Oligonucleotide primers for PCR
Degenerate primers for the original amplification of frog AQP fragments
were designed based on the amino acid sequences around the two conserved Asn-Pro-Ala (NPA) boxes of MIP family aquaporins (20). The following primers were commercially synthesized (Life Technologies, Inc.): P1
(sense), 5⬘-AGCGGGG(CG)(CT)CAC(AC)T(CT)AACCC-3⬘; P2 (antisense),
5⬘-GG(AT)CC(AG)A(CT)CCA(AG)AAGA(CT)CCA-3⬘; P3 (antisense), 5⬘A(AG)(AG)GA(CG)C(GT)(GT)GC(AT)GG(AG)TTCAT-3⬘.
RT-PCR amplification
The skin poly (A)⫹RNA (0.5 ␮g) was heated at 65 C for 3 min and
cooled on ice. For cDNA synthesis, the denatured RNA was incubated
at 42 C for 1 h in a 20 ␮l-reaction buffer containing Rous-associated
virus 2 reverse transcriptase (9.9 U, Takara), 1 mm concentration of each
deoxynucleotide triphosphate (dNTP), 7.5 mm oligo-deoxythymidine
(19) primer, and ribonuclease inhibitor (20 U, Toyobo, Osaka, Japan) and
then at 52 C for 30 min. Using the reverse-transcribed first-strand cDNA,
we then performed PCR in 25 ␮l of Ex-Taq buffer containing a 0.2 mm
concentration of each dNTP, 1 mm primer P1, and 2 mm primer P2 along
with 0.625 U of Ex-Taq polymerase (Takara). Nested PCR amplification
was further performed using primers P1 and P3. The procedure of PCR
amplification consisted of an initial denaturation step of 95 C for 5 min,
followed by denaturation (94 C, 90 sec), annealing (50 C, 90 sec) and
extension (72 C, 150 sec) for 30 cycles in a thermal cycler (ASTEC,
Fukuoka, Japan). Amplified fragments were subcloned into pGEM-3z
vectors (Promega, Madison, WI), and sequenced. Because one clone with
its sequence corresponding to that of Bufo AQP-t2 (AAC69694) was
identified, we tentatively designated it as AQP-h2.
Screening of the frog skin cDNA library
A DNA probe, obtained from the first PCR product as described above,
was synthesized by using a digoxigenin (DIG)-High Prime kit (Roche
Molecular Biochemicals, Meylan, France) and used to screen the cDNA
library constructed from the pelvic skin in accordance with the manufacturer’s instructions. The membrane was hybridized with DIG-labeled
cDNA probes at 68 C overnight and washed twice in 1⫻ saline sodium
citrate/0.1% sodium dodecyl sulfate for 1 h at 50 C. After blocking, the
membrane was incubated with alkaline phosphatase-conjugated sheep
anti-DIG Fab antibody (Roche), reacted with 25 mm CSPD, disodium 3-(4methoxyspiro{1, 2-dioxetane-3, 2⬘-(5⬘-chloro)tricyclo[3.3.1.13.7]decan}-4yl)phenyl phosphate) chemiluminescence substrate (Tropix, Inc., PE
Applied Biosystems, Foster City, CA), and then juxtaposed to on Hyperfilm-ECL (Amersham Pharmacia Biotech, Buckinghamshire, UK).
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
of reaction buffer containing a 1 mm concentration of each dNTP, 9.9 U
of Rous-associated virus 2 reverse transcriptase (Takara), 20 U of ribonuclease inhibitor (Toyobo), 7.5 mm oligo-deoxythymidine (19) primer
(Life Technologies, Inc.). RT-PCR was performed basically by the same
method described above, by using homologous primers: P4 (sense),
5⬘-CTTCTGGATTGGACCTTTTG-3⬘ (609 – 628 bp); and P5 (antisense),
5⬘-GCATTAGAGCGTAGTAATCC-3⬘ (1027–1046 bp). The RT-PCR
products were analyzed on a 2% agarose gel containing ethidium bromide (0.5 ␮g/ml) with Marker 6 (␭/Sty1 digest; Wako Pure Chemicals,
Osaka, Japan) for molecular weight markers.
Southern blot analysis
To obtain cDNA probe, we performed the PCR using pBK-CMV
phagemid vector containing AQP-h2 cDNA, and primers P4 and P5 for
AQP-h2 as described above. DIG-labeled cDNAs were synthesized with
a DIG-High Prime kit (Roche). The membrane was subsequently hybridized with DIG-labeled cDNA probes, and then hybridization signals
were detected with CSPD on Hyperfilm-ECL after incubating with alkaline phosphatase-labeled anti-DIG antibody, as described above.
Osmotic water permeability of oocytes
cRNAs were prepared from linearized pBK-CMV phagemid vectors
containing the entire open reading frame of AQP-h1 or AQP-h2 with
XhoI (Takara) and transcribed/capped with T3 RNA polymerase (mCAP
RNA Capping kit, Stratagene). Stage V and VI Xenopus oocytes were
defolliculated by collagenase (1 mg/ml; Roche) and microinjected with
cRNAs (5 or 50 ng) or water. After a 3-d incubation in Barth’s buffer at
18 C, the oocytes were transferred from 200 mosmol to 70 mosmol
Barth’s buffer, and the osmotically elicited increase in volume was
monitored at 24 C under an Olympus BX50 microscope with a ⫻4
magnifying objective lens and a charge-coupled device camera connected to a computer. The coefficient of osmotic water permeability (Pf)
was calculated from the initial slope of oocyte swelling according to the
previous method (9, 22). In some experiments, HgCl2 was added to a
final concentration of 0.3 mm for 10 min. In other experiments, we treated
the oocytes with a mixture of cAMP (8-bromo-cAMP sodium salt; Sigma,
St. Louis, MO) and 1 mm 3-isobutyl-1-methylxanthine (Sigma) for 30 min
before the volume measurement. To confirm whether AQP-h2 protein
was expressed in Xenopus oocytes after the injection of AQP-h2 cRNA,
we evaluated AQP-h2 cRNA-injected or water-injected oocytes by Western blot analysis and immunostained as described below.
Antibody
Oligopeptide corresponding to the C-terminal amino acids 255–271
(ST-140:QEQPRRKSMELQTL) of the Hyla AQP-h2, with an aminoterminal cysteine residue, was synthesized with a Model 433A (PE
Applied Biosystems, Foster City, CA). The crude peptide was purified
by reverse-phase HPLC with a 0 – 60% linear gradient of CH3CN in 0.1%
trifluoroacetic acid. Purify of the peptide was confirmed by measuring
its molecular mass by mass spectrometry. The antibody was raised in a
guinea pig immunized with the ST-140 peptide coupled to Keyhole limpet
hemocyanin (Pierce, Rockford, IL) as described previously (23). Rabbit
anti-Hyla AQP-h3 serum was characterized previously (19).
DNA sequence analysis
Western blot analysis
The nucleotide sequence was analyzed by the dideoxy chain-termination method using a Thermo Sequenase Cycle Sequencing Kit (United
States Biochemical Corp., Cleveland, OH) and an Aloka DNA sequencing system [Model Lic-4200L(S), Aloka Co., Ltd., Tokyo, Japan].
The ventral pelvic skin, back skin, kidney, and urinary bladder taken
from the tree frogs were homogenized in cell lysis buffer [50 mm
Tris-HCl (pH 8.0), 0.15 m NaCl, 1% Triton X-100, 0.1 mg/ml phenylmethylsulfonyl fluoride, 1 ␮g/ml aprotinin] and centrifuged in a microcentrifuge for 5 min to remove insoluble materials. The proteins were
determined with a BCA Protein Assay Kit (Pierce, Rockford, IL). The
supernatant protein (10 ␮g) was denatured at 70 C for 10 min in denaturation buffer comprising 2% sodium dodecyl sulfate, 25 mm Tris-HCl
(pH 7.5), 25% glycerol, and 0.005% bromophenol blue, subjected to
electrophoresis on a 12% polyacrylamide gel; and then transferred to an
Immobilon-P membrane (Millipore, Tokyo, Japan). The proteins in the
membrane were reacted sequentially with anti-Hyla AQP-h2 serum diluted at 1:10,000, biotinylated antiguinea pig IgG (Jackson Immunoresearch, West Grove, PA), and streptavidin-conjugated horseradish
RT-PCR of Hyla tissues
The tissue expression of AQP-h2 mRNA was analyzed by RT-PCR.
TRIZOL reagent was used to prepare total RNA from various tree frog
tissues (ventral pelvic skin, dorsal skin, urinary bladder, kidneys, brain,
tongue, heart, lungs, liver, stomach, small intestine, large intestine, testes, ovaries, and blood cells). After treatment of 20 ␮g total RNA with
deoxyribonuclease I (4 U; Takara), a 10-␮g aliquot of the total RNA was
reverse transcribed at 42 C for 1 h and then at 52 C for 30 min in 20 ␮l
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
Endocrinology, September 2003, 144(9):4087– 4096 4089
peroxidase (DAKO Japan, Co., Ltd., Kyoto, Japan). The reaction product
on the membrane was visualized by using an ECL Western blot detection
kit (Amersham). As a control, the primary antibody was replaced with
anti-Hyla AQP-h2 serum preincubated with 10 ␮g/ml of the immunogen
peptide. To determine whether the immunoreactive proteins were glycosylated, we treated an extract from the urinary bladder for 24 h at 37 C
with peptide-N-glycosidase F (Daiichi Pure Chemicals, Tokyo, Japan)
before SDS-PAGE and Western blotting, in accordance with the manufacturer’s instructions.
AQP-h3 (64.4%; Ref. 19), rat AQP2 (59.1%; Ref. 9), mouse
AQP2 (59.1%; Ref. 25), and human (61.4%; Ref. 26) AQP2, but
less homology to Hyla AQP-h1 (45.0%; Ref. 19), rat AQP3
(32.6%; Ref. 11), mouse AQP3 (32.9%; Ref. 27), and human
AQP3 (31.9%; Ref. 28).
Experimental protocol for stimulation with vasotocin
To investigate the tissue distribution of Hyla AQP-h2
mRNA expression, we performed RT-PCR by using total
RNA from various tissues. AQP-h2 mRNA was observed in
the ventral skin, urinary bladder, kidney, tongue, liver, stomach, small and large intestine, testis, and ovary (Fig. 2). This
RT-PCR result was confirmed by Southern blot analysis (data
not shown).
The ventral skin was removed from each of 13 tree frogs and bathed
in frog Ringer’s solution. Under a stereomicroscope, a piece of each
ventral skin was divided into four small fragments (⬃5 ⫻ 5 mm), and
then two of the fragments were incubated with the Ringer’s solution in
the presence and two in the absence of 10⫺8 m [Arg (8)]-vasotocin
(Peptide Institute, Inc., Osaka, Japan) at 23 C for 20 min under 95% O2-5%
CO2 air. After incubation, the specimens were examined by immunofluorescence microscopy.
Distribution of Hyla AQP-h2 mRNA expression in
various tissues
Antibody specificity
Immunofluorescence
Several tissues including skin, kidney, and urinary bladder were
quickly removed, fixed overnight in periodate-lysine-paraformaldehyde fixative, dehydrated, and embedded in Paraplast. Three-micrometer sections were cut and mounted on gelatin-coated slides. The deparaffinized sections were rinsed with distilled water and PBS. For single
labeling of Hyla AQP-h2, immunofluorescence staining was performed
essentially as described previously (24). The sections were sequentially
incubated with 1% BSA-PBS, guinea pig anti-Hyla AQP-h2 serum
(1:5000), and indocarbocyanine-labeled donkey antiguinea pig IgG
(Jackson). For nuclear counterstaining, 4⬘,6-diamidino-2-phenylindole
(DAPI) was included in the secondary antibody solution. The sections
were finally washed with PBS and then mounted in PermaFluor (Immunon, Pittsburgh, PA). For double-immunofluorescence staining for
AQP-h2 and AQP-h3, sections were incubated with a mixture of guinea
pig anti-AQP-h2 (1:5,000) and rabbit anti-AQP-h3 (1:10,000), and then
reacted with a mixture of indocarbocyanine-labeled donkey antiguinea
pig IgG (1:400), fluorescein isothiocyanate-labeled donkey antirabbit
IgG (1:400), and DAPI. To check the specificity of the immunostaining,
we performed an absorption test by preincubating the anti-AQP-h2
serum with C-terminal peptide (10 ␮g/ml) of AQP-h2 protein used as
immunogen or with C-terminal peptide (10 ␮g/ml) of AQP-h3. Specimens were examined with an Olympus BX50 microscope equipped with
a BX-epifluorescence attachment (Olympus Optical Co., Tokyo, Japan).
To test the specificity of the anti-Hyla AQP-h2 toward the
Hyla skins, kidney, and urinary bladder, we conducted Western blot analysis of their extracts. In the extract of the ventral
pelvic skin, the antiserum detected a major band at 29.0 kDa
(Fig. 3A, lane 1). No band was detectable when the extracts
of dorsal skin and kidney were examined (Fig. 3A, lanes 2
and 3). Furthermore, immunopositive bands were seen at
29.0 kDa and at 42.5– 65.8 kDa in the case of the extract of
urinary bladder (Fig. 3A, lane 4). The immunopositive bands
described above were not detected when the antiserum was
preabsorbed with the peptide used as the immunogen (Fig.
3B). This evidence showed that these immunoreactive bands
were specific to the antiserum. To confirm that the immunoreactive bands for the urinary bladder were glycosylated,
we performed a digestion experiment using peptide-Nglycosidase F. After digestion, the stained smear band became a band of 29.0 kDa, suggesting that the bands of apparent higher molecular mass represented glycosylated
forms of the 29.0 kDa-Hyla AQP-h2 protein (Fig. 3C).
Localization of Hyla AQP-h2 in the epidermis
Results
cDNA cloning of Hyla AQP-h2
Figure 1A shows the full cDNA sequence of Hyla AQP-h2
and the deduced amino acids. The cDNA consisted of a
5⬘-untranslated region of 87 bp and a 3⬘-untranslated region
of 382 bp followed by a poly (A) tail. The open reading frame
encoded a protein of 268 amino acids with a relative molecular mass calculated to be 29,204 Da. Hydropathy analysis
predicted six transmembrane regions with an N terminus
and a C terminus localized in the cytoplasm, similar to other
MIP family members (Fig. 1B). There was one putative Nlinked glycosylation site at Asn-124, one protein kinase C
phosphorylation site at Ser-231, and one protein kinase A
phosphorylation site at Ser-262 in the AQP-h2 protein. The
amino acid sequence contained the conserved NPA motifs
found in all MIP family members, as well as a cysteine just
upstream from the second NPA motif, which positioning is
similar to that in Hyla AQP-h1 and AQP-h3 (19). Hyla
AQP-h2 had its highest amino acid sequence homology to
Bufo AQP-t2 (94.4%; AAC69094) and high homology to Hyla
The Hyla epidermis is organized into four successive layers: the stratum corneum, the granulosum, the spinosum,
and the germinativum. Each layer consists of two main cell
types: granular cells and mitochondria-rich cells. In some
sections, exocrine glands were observed. When the ventral
epidermis was stained by the immunofluorescence method,
AQP-h2 protein was detected in two or three sublayers of the
stratum granulosum, located just beneath the stratum corneum (Fig. 4, A and B). In the outermost sublayer of the
stratum granulosum, AQP-h2 was localized in the basolateral membrane (Fig. 4C). In one particular experiment, immunolabels were observed in the entire plasma membrane of
the granular cells of the lower granulosa sublayer. No signal
was found in other types of cells including the mitochondriarich cells and exocrine glandular cells in the pelvic skin (Fig.
4, A and B). To confirm the specificity of the staining, we
carried out a control experiment. Immunopositivity for
AQP-h2 in the granular cells was abolished when the antiserum was preincubated with the C-terminal peptide of Hyla
AQP-h2 protein used as the immunogen (Fig. 4D). On the
4090
Endocrinology, September 2003, 144(9):4087– 4096
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
FIG. 1. A, Nucleotide and deduced amino acid sequences of frog AQP-h2 cDNA. The predicted amino acid is shown below the nucleotide sequence
(DDBJ/EMBL/GenBank accession no. AB107014). The asterisk indicates the termination codon. Solid triangles indicate putative N-glycosylation sites. NPA motifs is boxed. The diamond, square, and open triangle indicate phosphorylation sites for protein kinase C and protein kinase
A, and mercurial-inhibition site, respectively. B, Kyte-Doolittle hydropathy profile (window 11) of the deduced AQP-h2 amino acid sequence.
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
Endocrinology, September 2003, 144(9):4087– 4096 4091
FIG. 2. Tissue expression of AQP-h2 mRNA by
RT-PCR. RT-PCR products obtained by using
primers as described in Materials and Methods
were separated on a 2% agarose gel and stained
with ethidium bromide. ⫹, Presence of the
mRNA; ⫺, absence of the mRNA.
FIG. 3. Characterization of anti-AQP-h2 serum by Western blot analysis. A, Immunoreactive bands are seen at 29.0 kDa in an extract of
ventral pelvic skin (lane 1). No bands are visible with the extract of
dorsal skin and kidney (lanes 2 and 3). For the extract of the urinary
bladder, immunoreactive bands are detected at 29.0 kDa and at 42.5–
65.8 kDa (lane 4). B, The membrane was immunostained with the
antiserum preabsorbed with the immunogen peptide (10 ␮g/ml). Immunoreactive bands were completely eliminated. C, Western blot
analysis of extracts before and after digestion of the extracts by peptide-N-glycosidase F. Specific bands with the extract of urinary bladder seen before digestion (lane 1) are replaced by a single band of 29.0
kDa, presumed to be the nonglycosylated form of Hyla AQP-h2, after
digestion (lane 2).
other hand, adsorption with the C-terminal peptide of
AQP-h3 did not affect the AQP-h2 labeling (data not shown).
When the urinary bladder was immunostained with the antiserum, the labeling was observed as a spot-like pattern in
the cytoplasm under the apical plasma membrane of the
granular cells (Fig. 4E). No labeling of the urinary bladder
was obtained when the antiserum was preabsorbed with its
corresponding immunogen (Fig. 4F). In addition, no positive
signal for AQP-h2 was detected in kidney (Fig. 4, G and H)
or in several other tissues, such as tongue, liver, stomach,
small and large intestines, testis, and ovary (data not shown).
Expression of Hyla AQP-h2 in Xenopus oocytes
Transmembrane water flow through the Hyla AQP-h2 was
evaluated by expression of the aquaporin in Xenopus oocytes.
After 3 d of incubation at 18 C, the oocytes were transferred
from isotonic (200 mosmol) to hypoosmotic (70 mosmol)
Barth’s solution. The swelling was monitored by using a
microscope with an attached charge-coupled device camera,
and the coefficient of osmotic Pf was calculated (Fig. 5, A and
B). Swelling of AQP-h2 cRNA-injected oocytes appeared
higher than that of water-injected oocytes, but no significant
difference was observed between them. However, when
cAMP was added to the medium containing the AQP-h2
cRNA-injected oocytes, the swelling of oocytes increased as
shown in Fig. 5A. The cAMP-stimulated water permeability
was completely inhibited by 0.3 mm HgCl2 (Fig. 5B). In addition, when the water-injected oocytes were stimulated with
cAMP, they showed no significant increase in the water
permeability. To evaluate whether AQP-h2 protein was expressed in the AQP-h2-injected oocytes, we performed a
Western blot analysis on an extract of the oocytes. As shown
in Fig. 5C, the AQP-h2 protein was detected as three bands
at 30 –32 kDa (lane 1), which were slightly higher in molecular weight than the AQP-h2 detected for the ventral skin
(lane 3) or urinary bladder (lane 4). In addition, no band was
detected for the extract of the water-injected oocytes (lane 2).
However, when sections of the AQP-h2 cRNA-injected oocytes were immunostained with anti-Hyla AQP-h2, a large
immunofluorescent mass was seen in the cytoplasm near the
nucleus, but not in the plasma membranes (Fig. 5D-a). The
immunopositive sites were completely eliminated by preabsorption of the antiserum with 10 ␮g/ml of the immunogen
peptide (Fig. 5D-b). When the oocytes were stimulated by
cAMP, the label for AQP-h2 was dispersed throughout the
cytoplasm, but was still not seen in the plasma membrane
(Fig. 5D-c). No immunolabel was found in water-injected
oocytes (Fig. 5D-d).
Responsiveness of AQP-h2 and AQP-3 proteins in the
ventral skin to vasotocin
To clarify whether AQP-h2 and AQP-h3 proteins are a
regulated AQP, we conducted an experiment in which the
ventral skin was stimulated with vasotocin (the nonmammalian vertebrate counterpart of vasopressin or ADH in
mammals). After fragments from the ventral skin had been
incubated in medium with or without vasotocin for 20 min,
we examined immunohistochemically the localization of
AQP-h2 and AQP-h3 proteins in the tissue. When the ventral
skin was incubated in the medium without vasotocin,
AQP-h2 was found in the basolateral membrane of the granular cells in the outermost sublayer of the stratum granulosum, located just beneath the stratum corneum (Fig. 6, A, C,
and D). On the other hand, when the skin fragments were
incubated in the medium with vasotocin, AQP-h2 was detected in the apical membrane in the granular cells of the
4092
Endocrinology, September 2003, 144(9):4087– 4096
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
FIG. 4. Immunofluorescence localization of AQP-h2 in the
ventral pelvic skin. Fluorescence image for AQP-h2 (A) and
corresponding Nomarski differential interference contrast
image (B) are shown. AQP-h2 is present in a few sublayers
of the stratum granulosum just beneath the stratum corneum. C, Enlarged view of AQP-h2-positive granular cells
in the stratum granulosum. The punctate label is seen in
the basolateral plasma membranes and in the cytoplasm.
D, No labeling is detected in any cells of the pelvic skin
when the anti-AQP-h2 was preabsorbed with the corresponding immunogen peptide. Nonspecific labels are seen
in the nucleus of the stratum corneum cells (arrows). E,
Fluorescence image showing the presence of AQP-h2 in the
urinary bladder. F, Nonspecific staining of the urinary
bladder with the antiserum absorbed with respective immunogen. G and H, Fluorescence images in the collecting
duct of the kidney. Background labeling is seen in a section
stained with anti-AQP-h2 (G) and in the one (H) reacted
with the antiserum absorbed with its respective immunogen. Nuclei are counterstained with DAPI (blue). Arrowhead, Mitochondria-rich cells, G, exocrine gland; L, lumen.
Bar: A, B, D–G, 10 ␮m; C, 10 ␮m.
outermost sublayer (Fig. 6, E, G, and H). We obtained similar
results for the AQP-h3 protein (Fig. 6, B–D and F–H). In the
granular cell sublayer beneath the outermost layer, the fluorescence for both AQP-h2 and AQP-h3 showed a pattern
similar to that found in the nonstimulated condition, although the staining intensity was slightly weaker (Fig. 6, A,
B, E, and F).
Discussion
The present report describes the full sequence of mRNA
encoding a vasotocin-dependent AQP from frog skin. The
AQP, AQP-h2, was structurally characterized by having two
NPA motifs and six putative transmembrane domains as
well as a cysteine at a mercurial sensitivity site just upstream
from the second NPA motif. From a homology analysis, the
deduced amino acid sequence of Hyla AQP-h2 showed a high
homology to mammalian AQP2. Furthermore, AQP-h2 had
a putative phosphorylation site by protein kinase A at Ser262, identical to that of mammalian AQP2 (29). The amino
acid sequence of AQP-h3, cloned in a previous study, has a
high homology with that of AQP-h2, and also contained a
putative phosphorylation site recognized by protein kinase
A (19). This phosphorylation plays an important role in the
sorting of mammalian AQP-h2 to the apical plasma membrane. Accordingly, these data suggest that AQP-h2 in addition to AQP-h3 is translocated from the cytoplasmic pools
to the apical plasma membrane by vasotocin.
In this study, we investigated the expression of AQP-h2
mRNA by using RT-PCR. AQP-h2 was expressed in several
tissues including vasotocin-dependent tissues, i.e. the ventral
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
Endocrinology, September 2003, 144(9):4087– 4096 4093
FIG. 5. Expression of AQP-h2 in Xenopus oocytes. A, Time course of the osmotic swelling. Oocytes were microinjected with water or cRNAs
encoding AQP-h2. Some of the AQP-h2-injected oocytes were incubated with no additive, some with cAMP alone, and some with cAMP plus
0.3 mM HgCl2. Some of the water-injected oocytes were incubated with cAMP. B, Osmotic Pf was calculated from the initial rate of oocyte swelling.
All data shown are the mean ⫾ SE of measurements from six to seven oocytes in each experimental group. *, P ⬍ 0.001 vs. water; #, P ⬍ 0.001
vs. AQP-h2 (⫹ cAMP). C, Western blot analysis of AQP-h2-injected oocytes with anti-AQP-h2. An immunoreactive band is seen at 30- to 32-kDa
band with an extract of the AQP-h2 injected oocytes (lane 1) prepared after complete swelling. In the water-injected oocytes, no bands are
discernable with the antibody (lane 2). In addition, the 29-kDa band is consistent with that detected when the extract of ventral skin (lane3)
or urinary bladder (lane 4) was examined. D, Immunofluorescence images for AQP-h2 protein in AQP-h2-injected oocytes. After complete
swelling of the oocytes, immunoreactive AQP-h2 is detected in punctate distribution in the cytoplasm near the nucleus, but not in the plasma
membrane (a). In the absorption test, the immunoreactivity observed with anti-AQP-h2 is nearly eliminated to the background level in the
AQP-h2-injected oocyte (b). When stimulated the AQP-h2-injected oocytes by cAMP, the immunoreactivity becomes dispersed throughout the
cytoplasm, but is still not seen in the plasma membrane (c). Similar to b, only a background level is seen in the water-injected oocyte reacted
with anti-AQP-h2 (d). Arrows indicate immunopositive labeling of AQP-h2. Bar, 50 ␮m.
4094
Endocrinology, September 2003, 144(9):4087– 4096
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
FIG. 6. Double-immunofluoresence micrographs showing
the granular cells in the outermost sublayer labeled for
AQP-h2 and AQP-h3 under nonstimulated and vasotocinstimulated conditions. Immunolabeling results for AQP-h2
(A) and AQP-h3 (B) under the nonstimulated condition are
shown. C, Merged image for AQP-h2 (red) and AQP-h3
(green); D, Nomarski image for A–C. Immunolabeling results for AQP-h2 (E) and AQP-h3 (F) in response to vasotocin. G, Merged image for AQP-h2 (red) and AQP-h3
(green); H, Nomarski image for E–G. Nuclei are counterstained with DAPI (blue). Arrowhead and arrows refer to
mitochondra-rich cell and nonspecific reaction, respectively. Bar: A–H, 10 ␮m.
skin, kidney, and urinary bladder. Taken together with the
data from sequence homology between AQP-h2 and mammalian AQPs, it is very likely that AQP-h2 in amphibians is
the counterpart of mammalian AQP2.
Western blot analysis and digestion experiments with peptide-N-glycosidase F of the extracts of Hyla skin and bladder
showed that Hyla AQP-h2 protein was present in nonglycosylated and glycosylated forms, and we clearly showed
that this antibody (ST-140) was specific for Hyla AQP-h2
protein.
The immunofluorescence experiments of our study revealed that AQP-h2 was specifically expressed in the granular cells of the ventral pelvic skin and in the urinary bladder.
This distribution of immunopositive sites was in good agreement with the results of the tissue distribution of mRNA
found by RT-PCR. However, no immunopositive cells were
found in the kidney, although AQP-h2 mRNA was detected
in the kidney by RT-PCR. Thus we assumed that AQP-h2 in
the Hyla kidney is scarcely translated, because AQP-h2 protein was not detected by either immunofluorescence or Western blotting analysis. However, experiments should be conducted under different physiological conditions to clarify
whether AQP-h2 protein is expressed in the kidney, because
expression of the protein may be detectable at a certain
condition.
In response to ADH or isoproterenol, the intramembrane
particles appear in the apical plasma membrane of the granular cells in the outermost granular layer of the skin, thereby
increasing water permeability; intramembrane particles
were therefore considered to be water channels (14, 30). It
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
was a challenging issue to reveal molecular characterization
of the intramembrane particles. In this study, we showed
their molecular identity.
In the present study, we demonstrated that AQP-h2 protein was localized in the basolateral plasma membrane of
granular cells in the outermost granular sublayer in situ. A
similar labeling pattern was obtained after skin had been
removed from the animal and incubated in the medium. In
subsequent study, we investigated the immunolocalization
of these AQPs in the frog ventral skin after vasotocin treatment. In the ventral skin after this treatment, signals for
AQP-h2 and AQP-h3 were enhanced in the apical plasma
membrane of the granular cells in the outermost granular
sublayer. This sublayer is referred to as the first reacting cell
layer and forms a continuous barrier between the outside
and inside of the body, its continuity being preserved by tight
junctions (14, 31–33). The tight junctions separate the plasma
membrane into two domains, i.e. the apical membrane and
basolateral membrane. In the frog skin, the tight junctions are
formed as the granular cells differentiate and move upward
to the surface. Accordingly, the reason why AQP-h2 or
AQP-h3 was found in the entire plasma membrane of the
granular cells beneath the outermost sublayer may be that
tight junction formation was yet incomplete between these
cells.
It is well known that, in mammals, AQP2 is involved in
reabsorption of water in the collecting duct of the kidney (9).
Mammalian AQP2 is expressed in the apical plasma membrane and cytoplasm just beneath the apical membrane (9,
34). In response to ADH, mammalian AQP2 is translocated
from the cytoplasmic pool to the apical membrane, thereby
raising the water permeability (34, 35); whereas in the nonstimulated condition, it is removed from the apical membrane by endocytosis, thereby decreasing water permeability
(35). From this evidence, it has been proved that mammalian
AQP2 is an ADH-regulated AQP. In the present study, we
obtained evidence that AQP-h2 as well as AQP-h3 migrate
to the basolateral plasma membrane in the nonstimulated
condition, whereas both were translocated to the apical
plasma membrane in response to vasotocin. Considering
these points together, we propose that both Hyla AQP-h2 and
AQP-h3 proteins are water-channel molecules that are regulated in response to vasotocin. However, the exact mechanism of sorting these AQP proteins to the apical or basolateral plasma membrane remains unclear at present.
Several lines of evidence showed that protein kinase A
phosphorylation at the consensus Ser-256 is necessary for
AQP2 to migrate from the intracellular membrane vesicles to
the apical plasma membrane (36, 37). Because there is a
protein kinase A site at Ser-262 in AQP-h2 and at Ser-255 in
AQP-h3 (19), phosphorylation in these sites may be necessary for translocating AQP-h2 or AQP-h3 protein to the apical plasma membrane.
In the osmotic water permeability experiments using Xenopus oocytes, we found weak water permeability in the
AQP-h2-injected oocytes. Concerning these oocytes, our
Western blot analysis and immunohistochemistry revealed
that AQP-h2 protein was not fully translocated to the plasma
membrane of the ooctyes, thereby causing low activity of
AQP-h2 in water permeability assay. A similar result was
Endocrinology, September 2003, 144(9):4087– 4096 4095
obtained for the AQP-h3-injected oocytes in a previous study
(19). On the other hand, in the presence of cAMP the water
permeability of AQP-h2-injected oocytes was greatly increased; and the immunoreactive AQP-2 was dispersed
throughout the cytoplasm, although we were not able to
detect a positive reaction in the plasma membrane. Because
cAMP enhanced the water permeability in the AQP-h2injected oocytes, a small amount of the protein, a degree that
is undetectable in this immunofluorescence study, may be
expressed in the plasma membrane of the oocytes. On the
other hand, in the Western blot analysis using the AQP-h2injected oocytes, we obtained immunopositive band with
slightly higher molecular size than the AQP-h2 detected for
the ventral skin. Consequently, such a difference in molecular sizes may be reflected on the disturbance of normal
intracellular traffic for AQP-h2 protein. However, it needs to
clarify the reason why sufficient amount of the AQP-h2 protein is not translocated to the plasma membrane in Xenopus
oocytes. Thus, our results suggest the possibility that vasotocin-regulated membrane trafficking mechanisms or unknown factors, which exist in the Hyla ventral skin cells, but
not in Xenopus oocytes, are necessary for the AQP-h2 protein
to be translocated to the plasma membrane.
Taken together, our data suggest that both AQP-h2 and
AQP-3 are vasotocin-regulated AQPs in amphibians, capable
of being translocated from the cytoplasmic pool to the apical
plasma membrane and thereby playing a role in water balance of the frog body.
Acknowledgments
Received April 3, 2003. Accepted May 14, 2003.
Address all correspondence and requests for reprints to: Dr. Shigeyasu Tanaka, Department of Biology, Faculty of Science, Shizuoka
University, Ohya 836, Shizuoka 422-8529, Japan. E-mail: sbstana@ipc.
shizuoka.ac.jp.
This work was supported in part by a grant-in-aid for scientific
research (to S.T.) from the Ministry of Education, Science, Sports, and
Culture of Japan.
References
1. Bentley PJ, Main AR 1972 Zonal differences in permeability of the skin of some
anuran Amphibia. Am J Physiol 223:361–363
2. Yorio T, Bentley PJ 1977 Asymmetrical permeability of the integument of tree
frogs (Hylidae). J Exp Biol 67:197–204
3. Bentley PJ, Yorio T 1979 Do frogs drink? J Exp Biol 79:41– 46
4. King LS, Agre P 1996 Pathophysiology of the aquaporin water channels. Annu
Rev Physiol 58:619 – 648
5. Yamamoto T, Sasaki S 1998 Aquaporins in the kidney: emerging new aspects.
Kidney Int 54:1041–1051
6. Ishibashi K, Kuwahara M, Sasaki S 2000 Molecular biology of aquaporins.
Rev Physiol Biochem Pharmacol 141:1–32
7. Hatakeyama S, Yoshida Y, Tani T, Koyama Y, Nihei K, Ohshiro K, Kamiie
JI, Yaoita E, Suda T, Hatakeyama K, Yamamoto T 2001 Cloning of a new
aquaporin (AQP10) abundantly expressed in duodenum and jejunum. Biochem Biophys Res Commun 287:814 – 819
8. Gorin MB, Yancey SB, Cline J, Revel JP, Horwitz J 1984 The major intrinsic
protein (MIP) of the bovine lens fiber membrane: characterization and structure based on cDNA cloning. Cell 39:49 –59
9. Fushimi K, Uchida S, Hara Y, Hirata Y, Marumo F, Sasaki S 1993 Cloning and
expression of apical membrane water channel of rat kidney collecting tubule.
Nature 361:549 –552
10. Echevarria M, Windhager EE, Tate SS, Frindt G 1994 Cloning and expression
of AQP3, a water channel from the medullary collecting duct of rat kidney. Proc
Natl Acad Sci USA 91:10997–11001
11. Ishibashi K, Sasaki S, Fushimi K, Uchida S, Kuwahara M, Saito H, Furukawa
T, Nakajima K, Yamaguchi Y, Gojobori T, Marumo F 1994 Molecular cloning
and expression of a member of the aquaporin family with permeability to
4096
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Endocrinology, September 2003, 144(9):4087– 4096
glycerol and urea in addition to water expressed at the basolateral membrane
of kidney collecting duct cells. Proc Natl Acad Sci USA 91:6269 – 627
Ma T, Frigeri A, Hasegawa H, Verkman AS 1994 Cloning of a water channel
homolog expressed in brain meningeal cells and kidney collecting duct that
functions as a stilbene-sensitive glycerol transporter. J Biol Chem 269:21845–
21849
Brown D, Katsura T, Kawashima M, Verkman AS, Sabolic I 1995 Cellular
distribution of the aquaporins: a family of water channel proteins. Histochem
Cell Biol 104:1–9
Brown D, Grosso A, DeSousa RC 1983 Correlation between water flow and
intramembrane particle aggregates in toad epidermis. Am J Physiol 245:C334 –
C342
Chevalier J, Bourguet J, Hugon JS 1974 Membrane associated particles: distribution in frog urinary bladder epithelium at rest and after oxytocin treatment. Cell Tissue Res 152:129 –140
Kachadorian WA, Wade JB, DiScala VA 1975 Vasopressin: induced structural
change in toad bladder luminal membrane. Science 190:67– 69
Abrami L, Simon M, Rousselet G, Berthonaud V, Buhler JM, Ripoche P 1994
Sequence and functional expression of an amphibian water channel, FA-CHIP:
a new member of the MIP family. Biochim Biophys Acta 1192:147–151
Ma T, Yang B, Verkman AS 1996 cDNA cloning of a functional water channel
from toad urinary bladder epithelium. Am J Physiol 271:C1699 –C1704
Tanii H, Hasegawa T, Hirakawa N, Suzuki M, Tanaka S 2002 Molecular and
cellular characterization of a water channel protein, AQP-h3, specifically expressed in the frog ventral skin. J Membrane Biol 188:43–53
Agre P, Brown D, Nielsen S 1995 Aquaporin water channels: unanswered
questions and unresolved controversies. Curr Opin Cell Biol 7:472– 483
Sasaki S, Ishibashi K, Marumo F 1998 Aquaporin-2 and -3: representatives of
two subgroups of the aquaporin family colocalized in the kidney collecting
duct. Annu Rev Physiol 60:199 –220
Zhang RB, Logee KA, Verkman AS 1990 Expression of mRNA coding for
kidney and red cell water channels in Xenopus oocytes. J Biol Chem 265:15375–
15378
Tanaka S, Nomizu M, Kurosumi K 1991 Intracellular sites of proteolytic
processing of pro-opiomelanocortin in melanotrophs and corticotrophs in the
rat pituitary. J Histochem Cytochem 39:809 – 821
Tanaka S, Yora T, Nakayama K, Inoue K, Kurosumi K 1997 Proteolytic
processing of pro-opiomelanocortin occurs in acidifying secretory granules of
AtT-20 cells. J Histochem Cytochem 45:425– 436
Hasegawa et al. • Vasotocin-Dependent AQPs in Frogs
25. Yang B, Ma T, Xu Z, Verkman AS 1999 cDNA and genomic cloning of mouse
aquaporin-2: functional analysis of an orthologous mutant causing nephrogenic diabetes insipidus. Genomics 57:79 – 83
26. Sasaki S, Fushimi K, Saito H, Saito F, Uchida S, Ishibashi K, Kuwahara M,
Ikeuchi T, Inui K, Nakajima K, Watanabe T, Marumo F 1994 Cloning, characterization, and chromosomal mapping of human aquaporin of collecting
duct. J Clin Invest 93:1250 –1256
27. Ma T, Song Y, Yang B, Gillespie A, Carlson EJ, Epstein CJ, Verkman AS 2000
Nephrogenic diabetes insipidus in mice lacking aquaporin-3 water channels.
Proc Natl Acad Sci USA 97:4386 – 4391
28. Ishibashi K, Sasaki S, Saito F, Ikeuchi T, Marumo F 1995 Structure and
chromosomal localization of a human water channel (AQP3) gene. Genomics
27:352–354
29. Brown D, Katsura T, Gustafson CE 1998 Cellular mechanisms of aquaporin
trafficking. Am J Physiol 275:F328 –F331
30. De Sousa RC, Grosso A 1982 Osmotic water flow across the abdominal skin
of the toad Bufo marinus: effect of vasopressin and isoprenaline. J Physiol
(Lond) 329:281–296
31. Voute CL, Ussing HH 1968 Some morphological aspects of active sodium
transport. The epithelium of the frog skin. J Cell Biol 36:625– 638
32. Farquhar MG, Palade GE 1965 Cell junctions in amphibian skin. J Cell Biol
26:263–291
33. Brown D, Ilic V 1979 Freeze fracture differences in plasma membranes of the
stratum corneum and replacement layer cells of amphibian epidermis. J Ultrastruct Res 67:55– 64
34. Nielsen S, DiGiovanni SR, Christensen EI, Knepper MA, Harris HW 1993
Cellular and subcellular immunolocalization of vasopressin-regulated water
channel in rat kidney. Proc Natl Acad Sci USA 90:11663–11667
35. Nielsen S, Chou CL, Marples D, Christensen EI, Kishore BK, Knepper MA
1995 Vasopressin increases water permeability of kidney collecting duct by
inducing translocation of aquaporin-CD water channels to plasma membrane.
Proc Natl Acad Sci USA 92:1013–1017
36. Kuwahara M, Fushimi K, Terada Y, Bai L, Marumo F, Sasaki S 1995 cAMPdependent phosphorylation stimulates water permeability of aquaporincollecting duct water channel protein expressed in Xenopus oocytes. J Biol
Chem 270:10384 –10387
37. Fushimi K, Sasaki S, Marumo F 1997 Phosphorylation of serine 256 is required
for cAMP-dependent regulatory exocytosis of the aquaporin-2 water channel.
J Biol Chem 272:14800 –14804