Two types of aquaporin 5 (AQP5) genes (aqp-xt5a and aqp-xt5b) were identified in the genome of Xenopus tropicalis by synteny comparison and molecular phylogenetic analysis. When the frogs were in water, AQP-xt5a mRNA was expressed in the skin and urinary bladder. The expression of AQP-xt5a mRNA was significantly increased in dehydrated frogs. AQP-xt5b mRNA was also detected in the skin and increased in response to dehydration. Additionally, AQP-xt5b mRNA began to be slightly expressed in the lung and stomach after dehydration. For the pelvic skin of hydrated frogs, immunofluorescence staining localized AQP-xt5a and AQP-xt5b to the cytoplasm of secretory cells of the granular glands and the apical plasma membrane of secretory cells of the small granular glands, respectively. After dehydration, the locations of both AQPs in their respective glands did not change, but AQP-xt5a was visualized in the cytoplasm of secretory cells of the small granular glands. For the urinary bladder, AQP-xt5a was observed in the apical plasma membrane and cytoplasm of a number of granular cells under normal hydration. After dehydration, AQP-xt5a was found in the apical membrane and cytoplasm of most granular cells. Injection of vasotocin into hydrated frogs did not induce these changes in the localization of AQP-xt5a in the small granular glands and urinary bladder, however. The results suggest that AQP-xt5a might be involved in water reabsorption from the urinary bladder during dehydration, whereas AQP-xt5b might play a role in water secretion from the small granular gland.
water channels, called aquaporins (AQPs), play important roles in water homeostasis and various physiological processes. AQPs are a class of integral membrane proteins that form a selective water pore in the plasma membrane of various cells of animals, plants, and microorganisms (2, 47). Thirteen isoforms of AQPs (AQP0–AQP12) are identified in mammals and categorized into three subfamilies: classical AQPs, aquaglyceroporins, and unorthodox AQPs (22). Classical AQPs, e.g., AQP2 and AQP5, conduct only water, whereas aquaglyceroporins, e.g., AQP3 and AQP7, transport not only water but also small uncharged solutes, such as glycerol and urea. Unorthodox AQPs, comprising AQP11 and AQP12, are so called because of their deviated amino acid sequences. On the other hand, anurans possess additional anuran-specific AQPs: i.e., AQPa1 and AQPa2 (The letter “a” represents “anuran”) (39, 41). AQPa2 is further subdivided into urinary bladder type (type 2) and ventral skin-type (type 3) (40, 41). Each AQP shows unique localization in the tissues and cells, therein permitting specific water influx and efflux. AQP5, a classical AQP, is expressed in various mammalian exocrine glands, including the sweat gland, lacrimal gland, salivary gland, and pyloric gland (42). At the cellular level, AQP5 is detected in either the apical or apical/basolateral membranes of the particular glandular cells, and plays a role in the secretion of water (42).
For anuran amphibians, an AQP5 homologue, AQP-x5, was identified in Xenopus laevis, and localized to the skin glands (26). The major skin glands of amphibians are divided into two types: the granular gland (also called the serous gland or poison gland) and the mucous gland (33, 45). The granular glands protect the body by producing a wide variety of biologically active substances, e.g., toxic amines, antimicrobial peptides, and toxic alkaloids (7, 13, 31). On the other hand, the mucous glands produce a mixture of glycopeptides, such as neutral, sialic, and sulfated mucins (11, 12, 38). The water included in the secreted fluid, especially from the mucous glands, aids in maintenance of a moist skin, cutaneous gas exchange, and thermoregulation (6, 28). Another two glands are observed in the X. laevis skin: the small granular (granulated) gland and the NP gland in the nuptial pad of the male forelimb (17). Immunohistochemistry and immunoelectron microscopy with an antibody raised against AQP-x5 revealed that AQP-x5 is present in the apical plasma membrane of acinar cells of the small granular glands (30).
Whereas AQP-x5 is the only AQP5 homologue whose cDNA was obtained from X. laevis, molecular phylogenetic analysis previously indicated the presence of two types of aqp5 in Xenopus tropicalis (40) (Ensembl, http://www.ensembl.org/index.html). In the present study, we first cloned cDNA encoding an anuran homologue of AQP5 in X. tropicalis (AQP-xt5a: Ensembl ENSXETG00000024580) because the nucleotide sequence of mRNA or EST was not determined for this gene. The relationships among AQP5 homologues were then examined utilizing the sequence data of 95 AQPs from mammals to fish. Immunohistochemical analysis was further carried out to elucidate the physiological roles of AQP-xt5a and AQP-xt5b (ENSXETG00000020388; accession no. NM_001015749; 25) under hydrated and dehydrated conditions.
MATERIALS AND METHODS
Animals, sampling, and plasma analysis.
Male tropical clawed frogs, X. tropicalis, were kept in freshwater at 25°C, and fed commercial trout pellets. The animals were 4.5–6.0 cm long (body length) and weighed 11–26 g. Before sampling, the animals were anesthetized with ethyl m-aminobenzoate methanesulfonate (Nacalai tesque, Kyoto, Japan). The blood was collected from the heart, and the plasma was separated by centrifugation and stored at −25°C. Plasma osmolality was measured by freeze point depression osmometry (Vogel OM815, Giessen, Germany), and plasma sodium concentration was determined by atomic absorption spectrophotometry (Hitachi Z5300, Tokyo, Japan). The urinary bladder was dissected out, immediately frozen in liquid nitrogen, and used for cDNA cloning experiments. Additionally, various tissues, such as the kidney, ventral hindlimb skin, ventral pelvic skin, and urinary bladder, were sampled for reverse transcription (RT)-PCR analysis and immunohistochemistry. All animal experiments were carried out in compliance with the Guide for Care and Use of Laboratory Animals of Shizuoka University.
Cloning and sequence analysis of AQP-xt5a cDNA.
Total RNA was extracted from the urinary bladder with TRIzol reagent (Life Technologies, Carlsbad, CA), and 10 μg of total RNA was incubated in 100 μl buffer, including 40 units RNase inhibitor (Promega, Madison, WI) and 2 units deoxyribonuclease I (Takara, Kyoto, Japan), at 37°C for 30 min. The total RNA (5 μg) was then reverse-transcribed in 20 μl of buffer containing 0.5 mM dNTP, 10 mM DTT, 500 ng of oligo-dT19 primer (Operon, Tokyo, Japan), 40 units RNase inhibitor (Promega), and 200 units Moloney-murine leukemia virus reverse transcriptase (Life Technologies). AQP-xt5a cDNA was then amplified in 100 μl reaction mixture with 4 μl cDNA product, 0.2 mM dNTP, 2.4 units Prime STAR HS DNA polymerase (Takara), 1 μM AQP-xt5a primer I, 5′-ATGAAGAGGGAACTTTGCTCC, and 1 μM AQP-xt5a primer II, 5′-CTACTGGGGGGCACATTTTTTATG, using a Program Temp Control System, PC-708 (Astec, Fukuoka, Japan). The PCR program consisted of 1 denaturation cycle of 95°C for 0.5 min, and 30 amplification cycles of 94°C for 0.5 min, 55°C for 0.5 min, and 72°C for 1 min. The PCR product was separated by electrophoresis, and a major band was subcloned into Sma I-cut pGEM-3Z vector (Promega). Sequencing reactions were conducted with a thermo sequenase cycle sequencing kit (Affymetrix, Santa Clara, CA), and nucleotide sequences were determined using a Li-Cor automated DNA sequencer model 4200L-2G (Li-Cor, Lincoln, NE). The sequence data were analyzed using Genetyx, ver. 8 (Genetyx, Tokyo, Japan) and TMHMM server v.2.0 (http://www.cbs.dtu.dk/services/TMHMM/).
The amino acid sequences of 95 AQPs from mammals to fish were aligned using Clustal W (44), and alignment parameters were set according to an instruction manual by Hall (18). An unrooted tree was inferred by the neighbor-joining (NJ) method (36) in the PAUP program ver. 4.0 beta 10 (Sinauer Associates, Sunderland, MA). The mean character difference was utilized to estimate the evolutionary distance. Confidence in the NJ tree was assessed with 10,000 bootstrap replications (14) and utilized to construct a 50% majority-rule consensus tree.
Swelling assay with X. laevis oocytes.
The coding region of X. tropicalis AQP-xt5a cDNA was synthesized by PCR using the total RNA from the urinary bladder, PrimeSTAR, and primers: 5′-GCCACCATGAAGAGGGAACTTTGCTC and 5′-(T)25 CTACTGGGGGGCACATTTTTTATG, as described above. The amplified cDNA was subcloned into the pGEM-3Z vector (Promega). AQP-xt5a cRNA was capped and transcribed from the AQP-xt5a cDNA/pGEM-3Z vector that was linearized with Kpn I (Takara), using mCAP RNA capping kit (Stratagene, La Jolla, CA). For the swelling assay, X. laevis oocytes (stages V and VI) were defolliculated for 1 h with 1 mg/ml collagenase B (Roche Diagnostics, Tokyo, Japan) in sterile OR2 solution [100 mM NaCl, 2 mM KCl, 2 mM MgCl2, and 5 mM Tris-HCl (pH7.5)]. Isolated oocytes were microinjected with 50 nl of either cRNA (1 μg/μl) or MilliQ water, and then incubated in Barth's buffer [8 mM NaCl, 1 mM KCl, 1.7 mM MgSO4, 0.33 mM Ca (NO3)2, 0.41 mM CaCl2, 2.4 mM NaHCO3, 10 mM Tris-HCl (pH 7.6), 10 μg/ml penicillin, and 10 μg/ml streptomycin, 200 mOsm] at 18°C for 3 days. After incubation, the oocytes were transferred from 200 mOsm to 70 mOsm Barth's buffer without antibiotics. The osmotically elicited increase in volume was monitored at 24°C under an Olympus BX50 microscope with a charge-coupled device camera connected to a computer (Olympus, Tokyo, Japan). The coefficient of osmotic water permeability (Pf) was calculated from the initial slope of oocyte swelling (48). In the inhibition experiment, AQP cRNA-injected oocytes were incubated with 0.3 mM HgCl2 for 10 min before transfer to 70 mOsm Barth's buffer. Further, to examine the recovery of the HgCl2-induced inhibition, AQP cRNA-injected oocytes were incubated in 5 mM 2-mercaptoethanol for 10 min following the incubation in 0.3 mM HgCl2.
Dehydration and administration of [Arg (8)]-vasotocin.
Clawed frogs were separated into a hydrated control group or a dehydrated group. The dehydrated group was kept in a container with water at a depth of ∼1 cm at 30°C for 1 day and then at a depth of ∼0.5 cm at 30°C for 1 day, before transfer into a container without water except for a moist sponge attached to the underside of a stainless-steel wire mesh floor (day 0). Five dehydrated frogs were kept in the container at 30°C, and sampled on day 7. The control group was kept in a container with water at a depth of ∼10 cm at 30°C, and five hydrated frogs were sampled on day 7. In the hydrated group, another five frogs were injected intraperitoneally with 50 μl of saline (0.65% NaCl) (Otsuka, Tokyo, Japan) containing 10−6 M [Arg (8)]-vasotocin (AVT) (Peptide Institute, Osaka, Japan), and sampled after 20 min. No food was given to frogs during the course of the experiments. Relative body weight (RBW) was calculated for each group on the following formula: RBW = mean body weight on day 0 or day 7/mean body weight on day 0.
Tissue distribution and quantification of AQP mRNAs.
Total RNA was extracted from various tissues of clawed frogs, using TRIzol reagent (Life Technologies). After treatment with deoxyribonuclease I (Takara), 5 μg total RNA was subjected to RT-PCR, as described above. Specific primers were used for detection of AQP-xt5a mRNA, AQP-xt5b mRNA (NM_001015749), or AQP3 mRNA (NM_001016845): AQP-xt5a primer III, 5′-CGTCGTATCACCAAACGTCAG, AQP-xt5a primer IV, 5′-ATCGCTCCAGCTTTCTTCTTCC, AQP-xt5b primer I, 5′-CACTATAGCATTTCTGATTGG, AQP-xt5b primer II, 5′-AAGAGAGATGCCAGAATTCC, AQP3 primer I, ATTCCTGACTGTCAATCTGG, and AQP3 primer II, TATAAGTGTCAGATGCTCCG. As an endogenous control, β-actin cDNA (BC068217) was also amplified with specific primers: β-actin primer I, 5′-ACTTGACCTGACAGACTACC, and β-actin primer II, 5′-CAGTATTGGCATAGAGGTCC. An aliquot of 10 μl of each amplified product was electrophoresed through an ethidium bromide-stained 2% agarose gel, and photographed with a FAS-III digital camera system (Nippon Genetics, Tokyo, Japan). End-point quantitative RT-PCR was carried out using AQP-xt5a primer III and AQP-xt5a primer IV to assess the AQP-xt5a mRNA levels. The mRNA levels of AQP-xt5b, AQP3, and β-actin were quantified with the above specific primers. The cycle numbers were chosen within the exponential and parallel amplification range: 31 for AQP-xt5a, 30 for AQP-xt5b, 28 for AQP3, and 25 for β-actin. The amount of each molecule was estimated from the band intensity based on the standard curve generated by serial dilution of cDNA/pGEM-3Z vector. The intensity of the amplified bands was quantified using ImageJ software (National Institutes of Health, Bethesda, MD), and the ratio of AQP-xt5 or AQP3 to β-actin transcripts was calculated for normalization.
An oligopeptide corresponding to the C-terminal amino acid residues 244–255 (ST-200: EEESWSDQQDNC; Fig. 1) of X. tropicalis AQP-xt5a was synthesized and coupled to keyhole limpet hemocyanin (Protein Purify, Maebashi, Japan). Antibodies were raised in a rabbit or guinea pig immunized with the ST-200 peptide coupled to keyhole limpet hemocyanin, as described previously (43). Rabbit or guinea pig anti-peptide antibodies were previously generated for X. laevis AQP5 homologue, AQP-x5(b) (26), and X. laevis AQP3, AQP-x3BL (30), bullfrog vacuolar H+-ATPase (V-ATPase) E-subunit (46).
Western blot analysis.
After swelling assay, the microinjected oocytes were homogenized in cell lysis buffer [150 mM NaCl, 0.1 mg/ml PMSF, 1 mg/ml aprotinin, and 50 mM Tris·HCl (pH 8.0)], and centrifuged at 12,000 rpm for 10 min to remove insoluble materials. After addition of the equal volume of 6% SDS solution [6% sodium dodecyl sulfate, 22.4% glycerol, 10% 2-mercaptoethanol, 140 mM Tris·HCl (pH 6.8), and 0.02% bromophenol blue], the supernatant proteins (10 μg) were denatured at 37°C for 1 h and electrophoresed on a 12% SDS-polyacrylamide gel. The proteins were then transferred to an Immobilon-P membrane (Millipore, Tokyo, Japan) by electroblotting at 70 V for 1 h. The membrane was blocked in Block ace (Dainippon Sumitomo Pharma, Osaka, Japan). After a washing with TBS [0.1 M Tris·HCl (pH 8.0), 0.5 M NaCl, 0.1% polyoxyethylene (20) sorbitan monolaurate], the membrane was incubated with anti-AQP-xt5a serum diluted at 1:8,000. After a washing with TBS, the membrane was incubated with biotinylated goat anti-rabbit IgG antibody (Dako, Tokyo, Japan), and streptavidin-conjugated horseradish peroxidase (Dako). After a washing with TBS, the reaction products were detected by use of ECL Western blot analysis detection reagents (GE Healthcare, Buckinghamshire, UK). To assess the specificity of anti-AQP-xt5a antibody, an absorption test was performed by preincubating anti-AQP-xt5a with the antigen peptide (ST-200) (10 μg/ml). Further, to determine whether the immunoreactive proteins were glycosylated, the extracts from the cRNA-injected oocytes were treated with peptide-N-glycosidase F (Daiichi Pure Chemicals, Tokyo, Japan) at 37°C for 1 h, and subjected to SDS-PAGE.
Immunohistochemistry and immunofluorescence quantification.
The urinary bladder and skin were fixed in periodate-lysine-paraformaldehyde fixative overnight at 4°C, dehydrated, and embedded in Paraplast plus (McCormick Scientific, St. Louis, MO). Thin sections (4 μm) were cut and mounted on gelatin-coated slides. They were then deparaffinized and rinsed with distilled water and PBS before incubation with 1% BSA-PBS for 30 min. For immunofluorescence staining, sections were first covered with rabbit anti-AQP-xt5a antibody at a 1:2,000 dilution for 16 h. After rinsing in PBS, the sections were covered with a mixture of indocarbocyanine (Cy3)-labeled affinity-purified donkey anti-rabbit IgG antibody (Jackson Immunoresearch, West Grove, PA) at a 1:400 dilution and 1 μg/ml 4′, 6-diamidino-2-phenylindole (DAPI) for 2 h. DAPI was included for nuclear counterstaining. The sections were finally rinsed with PBS and mounted with PermaFluor (Immunon, Pittsburgh, PA). The specificity of immunostaining was assessed using an absorption test by preincubating the anti-AQP-xt5a antibody with the antigen oligopeptide (10 μg/ml). For double-immunofluorescence staining, sections were incubated with a mixture of rabbit anti-AQP-xt5a antibody at a 1:2,000 dilution and guinea pig anti-AQP-x5(b) antibody at a 1:2,000 dilution, guinea pig anti-AQP-x3BL antibody at a 1:2,000 dilution, or guinea pig anti-bullfrog V-ATPase E-subunit antibody at a 1:2,000 dilution, followed by incubation with a mixture of Cy3-labeled donkey anti-rabbit IgG at a 1:400 dilution (Jackson Immunoresearch), Alexa Fluor 488-labeled goat anti-guinea pig IgG at a 1:200 dilution (Molecular Probes, Eugene, OR), and DAPI. Specimens were observed with an Olympus BX61 microscope equipped with epifluorescence and Nomarski differential interference-contrast optics. Immunolabels for AQP-xt5a were quantified in the urinary bladder epithelial cells. Fluorescent images were captured in randomly selected fields, and the mean immunofluorescence intensity was measured with ImageJ (National Institutes of Health). Nonspecific fluorescence was recorded from the urinary bladder submucosal cells and subtracted from the measurements as the background signal.
Results were presented as the means ± SE. Statistical significance between two groups was analyzed by Student's t-test because the variances were equal (F test). Multiple comparisons were made by the Steel-Dwass test, using PASW Statistics 18 software (IBM, Armonk, NY) since the variances were heterogeneous (Bartlett's test). Statistical significance was set at P < 0.05 or P < 0.01.
Characterization of AQP-xt5 and synteny analysis.
Using RT-PCR and subsequent molecular cloning experiments, we obtained a cDNA fragment encoding AQP-xt5a. An open reading frame of the cDNA is composed of 867 nucleotides, and AQP-xt5a was predicted to consist of 289 amino acid residues (AB795995), as annotated in Ensembl (ENSXETG00000024580) (Fig. 1). As in other classical AQPs, certain features were conserved in AQP-xt5a; these were a pair of canonical Asn-Pro-Ala motifs, and amino acid residues important for the aromatic/arginine restriction filter, i.e., Phe-49, His-173, Cys-182, and Arg-188, (8) (Fig. 1). Cys-182 is considered to be also the binding site for mercury, which inhibits water permeation through AQP-xt5a (20). In addition, three PKC phosphorylation sites were predicted at Ser-152, Ser-227, and Ser-232 (Fig. 1). One PKA phosphorylation site was also predicted at Ser-259 (Fig. 1), but no N-linked glycosylation site was detected in AQP-xt5a. Similar characteristics were also detected in AQP-xt5b (NM_001015749) (Fig. 1). To verify the phylogenetic position of AQP-xt5a and AQP-xt5b, an NJ tree was constructed using 95 AQP proteins of vertebrates from fish to mammals, and a part of this tree is shown in Fig. 2. In the AQP5 cluster, AQP-xt5a belonged to the same subcluster as human AQP5, chicken AQP5, and Anole lizard AQP5, whereas AQP-xt5b formed another subcluster with X. laevis AQP-x5(b) and Bufo marinus AQP-t4. The NJ tree further showed that AQP5 has close relationships to kidney-type AQP2. Additionally, AQP5 has a higher similarity to AQP2 in its gene structure and location in the genome. aqp-xt5a, aqp-xt5b, mammalian AQP5, and AQP2 all comprise a similar four exon-three intron structure (Ensembl) (data not shown). In the genome of X. tropicalis, aqp-xt5a and aqp-xt5b are located between faim2 encoding Fas apoptotic inhibitory molecule 2 and racgap1 for Rac GTPase-activating protein 1, together with urinary bladder-type aqpa2 and ventral skin-type aqpa2 (Ensembl) (Fig. 3). Likewise, AQP5 and AQP2 are sited between FAIM2 and RACGAP1 in mammalian genomes (Ensembl) (Fig. 3).
Expression of AQP-xt5a in X. laevis oocytes.
Transmembrane water flow through AQP-xt5a was evaluated by a swelling assay using X. laevis oocytes. After 3 days of incubation at 18°C, the oocytes were injected with AQP-xt5a cRNA or water, and transferred from isotonic (200 mOsm) to hypo-osmotic (70 mOsm) Barth's solution. The oocytes injected with AQP-xt5a cRNA swelled more rapidly than those injected with water (Fig. 4A), and the Pf of AQP-xt5a was ∼4-fold greater than the control (Fig. 4B). The enhanced water permeability was significantly inhibited by the treatment with 0.3 mM HgCl2, but this decrease was recovered after the additional exposure to 5 mM 2-mercaptoethanol (Fig. 4B), confirming that the increase in water permeability of cRNA-injected oocytes was mediated by the expressed AQP-xt5a protein. To confirm the proper expression of AQP-xt5a in the cRNA-injected oocytes, immunofluorescence staining was carried out using an anti-AQP-xt5a antibody. Immunopositive labels were detected along the plasma membrane and cytoplasm near the plasma membrane of the AQP-xt5a cRNA-injected oocytes (Fig. 4C, a and b). The immunoreactivity was eliminated by preabsorption of the antiserum with 10 μg/ml of the immunogen peptide (Fig. 4C, c). No labels were observed in the water-injected oocytes (Fig. 4C, d). Western blot analysis of the extracts from AQP-xt5a cRNA-injected oocytes detected a band at ∼29 kDa, close to the molecular mass (28.8 kDa) predicted from the deduced amino acid sequence of AQP-xt5a. Another band was also observed at about 44 kDa (Fig. 4D, lane 1). Neither band was found when anti-AQP-xt5a was preabsorbed with the immunogen peptide (Fig. 4D, lane 2), suggesting that the bands of 29 kDa and 44 kDa may represent AQP-xt5a and a modified form, respectively. However, the 44-kDa band was not shifted after treatment with peptide-N-glycosidase F (data not shown).
Physiological conditions of X. tropicalis.
The body weight of X. tropicalis was significantly decreased after 7 days of dehydration (Fig. 5A). Plasma osmolality was significantly increased after dehydration (Fig. 5B), but plasma sodium concentration showed no difference between hydrated and dehydrated frogs (Fig. 5C).
Gene expression of AQP-xt5 in tissues.
Tissue distribution of AQP-xt5a mRNA and AQP-xt5b mRNA was examined by RT-PCR. For the hydrated frogs, amplified bands for both AQP-xt5 mRNAs were detected in the ventral hindlimb skin, ventral pelvic skin, pectoral skin, and dorsal skin (Fig. 6A). A band for AQP-xt5a mRNA was also seen in the urinary bladder. In the end-point RT-PCR, all the bands appeared to be intensified in the dehydrated frogs (Fig. 6B). Additionally, a weak band for AQP-xt5b mRNA was detected in the lung and stomach of the dehydrated frogs (Fig. 6B). Quantitative RT-PCR using β-actin mRNA as an endogenous reference showed that the relative expression levels of AQP-xt5a mRNA increased significantly (P < 0.05) in the ventral hindlimb skin, ventral pelvic skin, dorsal skin, and urinary bladder of the dehydrated frogs, compared with the hydrated frogs (Fig. 7, A–D). The relative expression of AQP-xt5b mRNA also increased significantly (P < 0.01) in the hindlimb skin of the dehydrated frogs (Fig. 7E).
Gene expression was further examined for AQP3. RT-PCR analysis showed AQP3 mRNA to be expressed in all the tissues examined, including the skin and urinary bladder (Fig. 6). The AQP3 mRNA expression appeared to be enhanced after dehydration (Fig. 6B), and a significant increase (P < 0.05) was detected in the urinary bladder (Fig. 7F).
Localization of AQP-xt5 in the skin and urinary bladder.
Mallory's triple staining showed that the skin consisted of the epidermis and the dermis that was further divided into the spongy dermis and compact dermis (Fig. 8A), as in other anurans (10, 27, 35). Three types of skin glands, i.e., mucous gland, granular gland, and small granular gland, were observed in the skin of X. tropicalis (Fig. 8A). The acini of the mucous gland and small granular gland were stained light blue, while the acinus of the granular gland was stained red, like those in the skin of X. laevis (17). The acinus of each gland was located in the spongy dermis.
Immunofluorescence staining localized labels for AQP-xt5a along the peripheral area of the acinus of the granular gland in both the dorsal and ventral skins of hydrated frogs (Fig. 8, B, E, and F). For the granular gland, the secretory compartment of the acinus comprises a syncytium (9). Therefore, AQP-xt5 seems to be present in the peripheral cytoplasm of the secretory syncytium of the granular gland. No labels were observed in the mucous gland or small granular gland of hydrated frogs (Fig. 8, B, E, and F). The specificity of the immunoreaction was confirmed by the antibody absorption tests using anti-AQP-xt5a antibody preabsorbed with the antigen (Fig. 8, C and D). On the other hand, immunoreactive AQP-xt5b and AQP3 were localized along the apical plasma membrane and basolateral membrane of secretory cells of the small granular gland of hydrated frogs, respectively (data not shown), like AQP-x5(b) and AQP3 in X. laevis kept in the water (26, 40).
After dehydration, immunopositive labels for AQP-xt5a became detectable not only in the granular gland, but also in the small granular gland. High-magnification images revealed that labels for AQP-xt5a were located over the cytoplasm of acinar cells in the small granular gland (Fig. 8, G and H). The specific immunoreactivity was again confirmed by the antibody absorption tests (data not shown). Double-immunofluorescence staining was carried out to elucidate the detailed distribution of AQPs in the acinar cells of the small granular gland. Staining against AQP-xt5a and V-ATPase E-subunit, a marker molecule for mitochondrion-rich cells (46), depicted AQP-xt5a in secretory cells of the small granular gland of dehydrated frogs, but not in the mitochondrion-rich cells (Fig. 9A). It was further shown that AQP-xt5a, AQP-xt5b, and AQP3 (AQP-xt3BL), resided at the cytoplasm, apical plasma membrane, and basolateral plasma membrane of secretory cells of the small granular gland of dehydrated frogs, respectively (Fig. 9, B–D). AQP-xt5b was not found in the apical plasma membrane of flattened cells located in the intermediate region between the acinus and excretory duct of mucous glands, unlike X. laevis AQP-x5(b) (26, 30). These findings are summarized in Fig. 12A.
In addition to the skin glands, AQP-xt5a was observed in the urinary bladder. Immunofluorescence staining revealed AQP-xt5a in a number of the luminal epithelial cells under the hydrated condition, specifically at the apical plasma membrane and cytoplasm (Fig. 10, A–D). Double-immunofluorescence staining against AQP-xt5a and V-ATPase E-subunit showed that AQP-xt5a was not expressed in the mitochondrion-rich cells of hydrated frogs (Fig. 10, A and B). On the other hand, AQP3 immunoreactivity was detected along the basolateral membrane of these AQP-xt5a-positive cells (Fig. 10, C and D). Because AQP3 is expressed in the granular cells of the luminal epithelium (30, 40, 41), AQP-xt5a-positive cells are considered to be the granular cells. After dehydration, the number of AQP-xt5a-positive cells increased, and labels were observed in the apical membrane and cytoplasm of most granular cells (Fig. 10, E–G). Furthermore, labeling was increased significantly (P < 0.05) in the apical membrane (Fig. 11). Dehydration did not induce expression of AQP-xt5a in the mitochondrion-rich cells, however (Fig. 10, E and F). As in hydration, AQP3 was localized along the basolateral membrane of AQP-xt5a-positive cells, but its immunoreactivity tended to be increased (Fig. 10, G and H).
The effect of AVT on the localization of AQP-xt5a was examined in the hydrated frogs. Twenty minutes after injection, AVT did not induce the expression of AQP-xt5a in the secretory cells of the small granular glands or in most granular cells of the urinary bladder, such as was seen following dehydration (data not shown). In addition, the AVT injection scarcely increased the trafficking of AQP-xt5a into the apical membrane in the urinary bladder granular cells (Fig. 11).
In the present study, we have demonstrated the presence of two types of AQP5, AQP-xt5a and AQP-xt5b, in X. tropicalis. Although aqp-xt5b (ENSXETG00000020388) is annotated as aqp2 in Ensembl, we consider that this gene encodes an AQP5 homolog, for the following reasons. First, our molecular phylogenetic analysis assigned this gene to the AQP5 cluster. Secondly, in Ensembl, the gene order of aqp2 (aqp-xt5b) and aqp5 (aqp-xt5a) in X. tropicalis genome is inverted, compared with that of mammalian AQP2 and AQP5, in spite of the same orientation of transcription (Ensembl) (Fig. 3). Third, if this gene is aqp2, it should be expressed in the kidney, as in mammals (37), quails (32), and tree frogs (34). However, RT-PCR analysis did not detect the expression of this gene in the kidney, or rather it was expressed in the skin glands. Given these lines of evidence, it is more plausible that the current gene arrangement of aqp-xt5a (aqp5) and aqp-xt5b (aqp2) had been caused by a deletion of aqp2 and a tandem gene duplication of aqp-xt5.
The small granular gland was first reported in X. laevis (17), but little is known about its function. In the present study, we found small granular glands in the skin of X. tropicalis, and localized immunoreactive AQP-xt5b and AQP3 to the apical plasma membrane and basolateral membrane, respectively, of their secretory cells. In addition, we found that the subcellular distribution of both AQPs did not change when the frogs were dehydrated. These findings suggest that water might be secreted from the small granular gland through AQP-xt5b and AQP3 not only in hydrated frogs but also in dehydrated frogs. When the frogs dwell in water, this secretion might contribute to the disposal of excessive fluid entering the body osmotically. In some natural habitats, X. tropicalis migrates to the river banks during the dry season and hide in holes, under flat stones, or under roots (Amphibiaweb, http://amphibiaweb.org/). When the frogs are, thus, dehydrated on the land, the water secreted from the small granular gland could provide the body surface with moisture and help decrease body temperature via the endothermic effect of evaporation. The present immunohistochemistry further revealed the occurrence of AQP-xt5a in the cytoplasm of secretory cells of the small granular gland under the dehydrated condition. However, the physiological role of this AQP is obscure.
In the present study, AQP-xt5a, but not AQP-xt5b, was detected in the urinary bladder. This is the first finding of the expression of vertebrate AQP5 in this tissue. The urinary bladder is an important osmoregulatory organ for many adult anurans, and water is reabsorbed from the urine stored in the urinary bladder to compensate for the water loss when on dry land (4, 23). The urinary bladder has a tight epithelium that shows a low water permeability in the basal state but allows enhanced water permeability in response to antidiuretic hormone (ADH), especially vasotocin (AVT) (1, 3, 15). On the other hand, the urinary bladder of aquatic Xenopus was regarded as an exception because ADH failed to increase its water permeability (1, 3). The present immunostaining revealed that AQP-xt5a was located in the apical plasma membrane of granular cells, but the number of such cells was limited when the frogs were kept in the water. In addition, AVT injection did not increase the number of AQP-xt5a-expressing granular cells, and scarcely facilitated the translocation of AQP-xt5a to the apical membrane. These findings suggest that the failure of AVT to increase water permeability of Xenopus urinary bladder might be partially due to insufficient expression of AQP5a. When Xenopus inhabits water, this low expression might be important to prevent water reentry from the urinary bladder into the body.
By contrast, after dehydration, AQP-xt5a appeared along the apical plasma membrane of most granular cells, while AQP3 was detected along their basolateral membrane. These findings suggest that when exposed to dehydration, X. tropicalis could reabsorb water from the urinary bladder via AQP-xt5a and AQP3. As mentioned above, the frogs migrate onto the land during the dry season (Amphibiaweb). Hence, water reabsorption from the urinary bladder might have an essential role in the survival of this species in the dry season.
Our RT-PCR analysis revealed an increase in the expression of both AQP-xt5a and AQP3 mRNAs in the urinary bladder after dehydration. Serosal hypertonicity is reported to increase water permeability of ranid and bufonid urinary bladder (19, 24). Additionally, adrenal steroids enhance the vasopressin-stimulated water flow across toad urinary bladder (49). Therefore, it is possible that these factors might be involved in inducing the gene expression of AQP-xt5a and AQP3, although the actual molecular mechanisms are not known.
Previously, we reported that in a terrestrial species, Hyla japonica, urinary bladder-type AQPa2 (AQP-h2) was expressed in most granular cells, translocated into the apical membrane in response to AVT, and could cooperate with basolateral AQP3 for water reabsorption from the urinary bladder (40, 41). For X. tropicalis, we have recently found that urinary bladder-type AQPa2 (AQP-xt2) was also expressed in the granular cells (unpublished observation), and the physiological role of AQP-xt2 is currently under investigation.
Perspectives and Significance
Amphibian water economy has been extensively investigated in association with their adaptation to diverse ecological environments. When on land, many anurans absorb water through the posteroventral skin and reabsorb water from the urinary bladder (5, 23, 29). Thus far, Xenopus has been studied as a representative of aquatic anurans that do not show these responses (3). However, we have here reported the first evidence to suggest that even in Xenopus, the urinary bladder could reabsorb water under dehydrating conditions, which mimic the natural environment in the dry season. In addition, it is also the first finding that AQP5(a) could be involved in this water reabsorption. AQP5a and/or AQP5b are likely to exist not only in aquatic Xenopus but also in semiaquatic and terrestrial anurans. Therefore, it could be important to elucidate the physiological role of AQP5, as well as AQPa2, in various species for an understanding of the molecular mechanisms, diversity, and evolution of anuran osmoregulation.
This work was supported by Grants-in Aid for Scientific Research (20570055; 23370028; 26440165) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, by a SUNBOR grant from Suntory Institute for Bioorganic Research, and by a fund from the Cooperative Program (no. 113, 2011) of the Atmosphere and Ocean Research Institute, The University of Tokyo.
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: Y.S., T.S., R.O., M.S., and S.T. conception and design of research; Y.S., T.S., and H.M. performed experiments; Y.S., T.S., N.T., R.O., M.S., and S.T. analyzed data; Y.S., T.S., and M.S. interpreted results of experiments; Y.S., T.S., and N.T. prepared figures; Y.S. and M.S. drafted manuscript; Y.S. and M.S. edited and revised manuscript; Y.S. and M.S. approved final version of manuscript.
We express our sincere gratitude to Dr. Susumu Hyodo and Ms. Sanae Hasegawa, Atmosphere and Ocean Research Institute, The University of Tokyo, Japan, and Mr. Junya Hara, Shizuoka University, Japan, for their technical assistance in this study. We are also grateful to Dr. Bridget I Baker for critical reading of the manuscript.
- Copyright © 2014 the American Physiological Society