INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

UREA, a major end product of nitrogen catabolism, is used in a variety of metabolic and physiological processes in animals (47). For example, it is used 1) simply as an excretory waste product resulting from the detoxification of ammonia; 2) in nitrogen recycling, during hibernation or fasting, with the help of gut microbes that break down urea and use its nitrogen to make amino acids; and 3) as an osmolyte. As an osmolyte, it acts to raise osmolarity throughout the body as seen in the elasmobranchs or locally as seen in the renal medulla of mammals, allowing retention of water inside the body and prevention of dehydration. In contrast to most terrestrial vertebrates that are ureotelic, most teleost fishes are ammonotelic via the gill. For these reasons, urea transporters (UTs) have been studied mainly in the mammalian kidneys, where urea accumulation in the inner medulla is a key component of the urine-concentrating process, and also in the elasmobranchs, which retain a large amount of urea (~400 mM) to raise the osmolarity of their body fluids close to that of seawater through the specialized apparatus in the kidney, the peritubular sheath (14).

A UT was first cloned from the rabbit renal medulla by expression cloning (49). Subsequent homology screening revealed the presence of family members that derive from the two genes, UT-A and UT-B, tandemly arranged on the same chromosome, in the case of mouse and human, on chromosome 18 (5, 22, 23). Five splice variants of the mammalian UT-A gene have been identified, characterized, and termed UT-A1, UT-A2, UT-A3, UT-A4, and UT-A5 (6, 8, 33, 35, 49). The UT-A isoforms are predominantly expressed in the kidney except the testis isoform UT-A5. The UT-B isoform originally identified in the erythropoietic tissue (24) is relatively ubiquitous (27, 40). In addition to the mammalian UTs, an elasmobranch UT has recently been cloned from the kidney of the spiny dogfish Squalus acanthias (36). This dogfish shark UT is highly expressed in the kidney and moderately in the brain. In the kidney, it is considered to be involved in the reabsorption of urea by the tubules and maintenance of its high plasma and tissue levels by minimizing its loss into the urine. More recently, a urea transporter has been cloned from the gill of the gulf toadfish Opsanus beta (42), a teleost fish that is unique in its ureotelic nature because most teleosts are ammonotelic. The gulf toadfish is also unique in its amphibious nature; it can breathe air for long periods. Under such conditions, facultative ureogenesis appears to be essential to avoid ammonia intoxication because the gill or the site of ammonia excretion is no longer bathed in water. In fact, the gulf toadfish has been shown to contain a complete set of enzymes involved in the hepatic ornithine-urea cycle, including the rate-limiting enzyme carbamoyl phosphate synthase I (CPSase I). The gulf toadfish UT has been demonstrated to be confined to the gill by Northern blot analysis, providing direct molecular evidence for the concept, previously suggested by divided chamber experiments (37) and pharmacokinetic analyses using [¹⁴C]urea (19), that the gill is the major site of urea excretion. However, the detailed site, namely the type of gill cells expressing the UT, has not been determined.

Ammonotelic teleost fishes lack the hepatic ornithine-urea cycle, but they can synthesize substantial amounts of urea through CPSase III, which is present in various tissues, including the muscle, and prefers glutamine as an NH₃ donor (note that the CPSase I mentioned above uses ammonia as a physiological substrate; both CPSase isozymes are mitochondrial enzymes). This route of urea synthesis was studied in detail, by Wright et al. (46) and Korte et al. (11), in the rainbow trout Oncorhynchus mykiss, an ammonotelic teleost fish, and was shown to be developmentally regulated. They showed that the CPSase III activity and message are highly expressed in an embryonic stage and decline sharply in adulthood, but their levels remain significant in the muscle. Although exceptional, very high levels of CPSase III have been demonstrated in the muscle of the Lake Magadi tilapia Oreochromis alcalicus grahami, the only fish in the alkaline lake (pH 10) (28). These facts and previous physiological observations that 10-30% of the total waste nitrogen is excreted as urea even in ammonotelic teleost fishes (20, 45) point to the importance of identification, characterization, and cellular localization of the urea transporter involved in the facilitated movement of urea in the ammonotelic teleosts as well as those of the mammals and ureotelic teleosts. Here we report cloning, molecular characterization, salinity-dependent regulation of expression, and immunohistochemical localization of a UT from the eel Anguilla japonica, a euryhaline ammonotelic teleost fish.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials. Japanese eels (Anguilla japonica) were purchased from a dealer and adapted to a freshwater tank for 2 wk and then transferred to a seawater tank for time-course analysis and acclimated there for more than 2 wk before use. Restriction enzymes, T4 polynucleotide kinase, PerfectHyb hybridization buffer solution, and Klenow DNA polymerase were obtained from Toyobo (Tokyo, Japan); the Expand Long Template PCR system was from Roche Molecular Biochemicals (Mannheim, Germany); pBluescript II SK, Escherichia coli strain XL1-Blue MRF', Taq DNA polymerase, ZAP II, and the Gigapack III Gold in vitro packaging kit were from Stratagene (San Diego, CA); the mRNA purification kit and Ready-To-Go DNA labeling kit were from Amersham Pharmacia Biotech (Uppsala, Sweden); E. coli strain Top 10, pZEro-2, and pRSET-B were from Invitrogen (San Diego, CA); the DNA ligation kit version 2 was from Takara (Kyoto, Japan); the SequiTherm Long-read-LC sequencing kit was from Epicentre Technologies (Madison, WI); [-³²P]dCTP and Hybond-N⁺ nylon membranes were from Amersham Pharmacia Biotech (Little Chalfont, Bucks, UK); Protran BA85 nitrocellulose filters were from Schleicher and Schuell (Daddel, Germany); Immobilon-P membrane was from Millipore (Bedford, MA).

RNA isolation and construction of subtracted cDNA library. Eels were adapted to either seawater or freshwater for 2 wk as described above. Total RNA was isolated from the seawater- and freshwater-adapted eel gills by the guanidinium thiocyanate/CsCl centrifugation method (1), and poly(A)⁺ RNA was affinity purified using an oligo(dT)-cellulose mRNA purification kit (Amersham Pharmacia Biotech). Two types of subtracted cDNA libraries from seawater and freshwater eel gills were constructed using the PCR-Select cDNA Subtraction kit (Clontech) according to the user manual (http://www.clontech.com; PT1117-1, PR85431). In brief, double-stranded cDNA was synthesized from 2 µg of poly(A)⁺ RNA of seawater (tester) and freshwater (driver) eel gills. Tester and driver cDNAs were digested with RsaI. One-third of the tester cDNA was then ligated with adaptor 1, and another one-third was ligated with adaptor 2R (contained in the kit). First and second hybridizations were performed following the instruction manual. First, only the residual differentially expressed single-stranded cDNAs were specifically amplified by PCR with Ex Taq (Takara) using a primer corresponding to the common sequence of the 5'-end of the adaptor (PCR primer 1). They were then amplified again with nested-PCR primer 1 and nested-PCR primer 2R corresponding to the 3'-side of each adaptor according to the instruction manual. Next, these PCR products were directly inserted into the NotI-cleaved pBluescript II SK vector (Stratagene) using the NotI site on adaptor 1 and the XmaIII site on adaptor 2R of the fragments for site-specific cloning, and subtracted cDNA libraries (seawater and freshwater) were generated by selection using ampicillin-containing LB plates. About 300 individual clones were sequenced from each library, and the expression levels of the corresponding mRNAs were determined by Northern blot analysis using RNA preparations from freshwater and seawater eel gills to confirm their differential expression.

Northern blot analysis. Total RNA was isolated from various tissues of seawater and freshwater eels by the acid guanidinium thiocyanate-phenol-chloroform method. For Northern analysis, 20 µg/lane of total RNA was electrophoresed on formaldehyde-agarose (1%) denaturing gels in 1× MOPS running buffer (20 mM MOPS, pH 7.0, 8 mM acetate, 1 mM EDTA) and then transferred onto Hybond-N⁺ nylon membranes (Amersham Pharmacia Biotech) by the vacuum-blotting method using a MilliBlot-V system (American Bionetics) and 10× SSC (SSC: 0.15 M NaCl, 15 mM sodium citrate, pH 7.0) as the transfer buffer. After transfer, membranes were baked for 2 h at 80°C and prehybridized for 2.5 h at 65°C in PerfectHyb hybridization solution (Toyobo). The cDNA clones isolated from the above-described subtraction libraries were used as probes. The probes were labeled with [-³²P]dCTP (3,000 Ci/mmol) using a Ready-To-Go DNA labeling kit, and the unincorporated nucleotides were removed by passage through a Sephadex G-50 column (Amersham Pharmacia Biotech). The membranes were hybridized separately with each ³²P-labeled probe in the same buffer at 68°C for 16 h. The blots were subsequently washed with increasingly stringent conditions (final wash: 0.1× SSC and 0.1% SDS for 30 min at 60°C). Membranes were exposed to imaging plates (Fuji Film, Tokyo, Japan) in a cassette overnight. The results were analyzed using a Fujix BAS2000 Bio-image analyzer (Fuji Film).

cDNA library screening. The eel gill cDNA library in ZAP II (Stratagene) was prepared as described (9). Approximately 3 × 10⁵ plaque-forming units of the library were plated on XL1-Blue MRF' E. coli (Stratagene) to a density of approximately 3 × 10⁴ plaques per 10 × 14-cm agar plate and replicated on Protran BA85 nitrocellulose filters (Schleicher and Schuell). Filters were prehybridized for 2 h at 42°C in a solution containing 20% formamide, 5× Denhardt's solution (0.1% each of Ficoll, polyvinylpyrrolidone, and bovine serum albumin), 6× SSPE (SSPE: 0.15 M NaCl, 1 mM EDTA, and 10 mM NaH₂PO₄, pH 7.4), and 0.1% SDS. The probe was labeled with [-³²P]dCTP as described in Northern blot analysis. Hybridization was performed for 16 h at 42°C in prehybridization solution by adding radiolabeled cDNA probe at ~10⁶ cpm/ml. To identify positive clones, the filters were rinsed twice with 2× SSC, 0.1% SDS for 15 min at room temperature and subsequently washed with increasingly stringent buffers (final wash: 0.1× SSC containing 0.1% SDS for 30 min at 55°C). Finally, the filters were exposed to Kodak X-Omat AR film for 48 h at 80°C with an intensifying screen. Positive plaques were isolated by three rounds of screening under identical conditions as the primary screening and transformed into pBluescript II SK by in vivo excision using ExAssist helper phage and the E. coli XL1-Blue MRF'.

DNA sequencing and analysis. Nucleotide sequences of both strands were determined by the dideoxy chain termination method with an automated sequencer (LI-COR model 4000L) using a SequiTherm Long-read Cycle Sequencing kit-LC (Epicentre Technologies). The DNA sequences were compiled and analyzed using the Genetyx MAC computer program (Software Development) and also compared with the GenBank/EMBL/DDBJ database using the advanced BLAST program of National Center for Biotechnology Information (NCBI). Multiple protein sequence alignments were carried out using the program ClustalW, and final adjustments were made manually. A phylogenetic tree was constructed by the neighbor-joining method (29, 30) using the MEGA (Molecular Evolutionary Genetics Analysis) software, version 2, developed by S. Kumar, K. Tamura, I. Jakobsen, and M. Nei (Pennsylvania State University) (http://www.megasoftware.net). The confidence of the neighbor-joining tree was tested by bootstrap analyses, running 1,000 replicates. Primary sequence motifs were identified using the PROSITE database of ExPASy. The transmembrane-spanning domains in the amino acid sequences were predicted using the program PSORT (21).

Determination of 5'-terminal sequence of cDNA by 5'-RACE. To obtain a full-length cDNA, the 5'-ends of eUT were amplified using a 5'/3'-RACE (rapid amplification of cDNA ends) kit (Roche Molecular Biochemicals) according to the instruction manual. First-strand cDNA was synthesized from 2 µg of total RNA of seawater eel gill using avian myeloblastosis virus reverse transcriptase and the gene-specific primer SP1 (5'-CTTGGCTGAGAACACAGCCAT-3'). To attach a known sequence to the 3'-end of the first-strand cDNA, a homopolymeric tail was appended using terminal transferase and dATP. The tailed cDNA was directly amplified by PCR using the oligo(dT)-anchor primer and the gene-specific nested primer SP2 (5'-GGATTGTTGACGAACATCAC-3'). To reduce the background, the first PCR products were amplified again using the PCR anchor primer and a second nested gene-specific primer, SP3 (5'-TCCTTTCATCCATTGCCCAA-3'). Next, the PCR products were purified by 1.5% agarose gel electrophoresis, and an ~250-bp band was isolated and subcloned into pZEro-2 vector using a Zero Background/Kan cloning kit (Invitrogen) and sequenced.

Identification of an in-frame stop codon upstream of the initiator methionine by PCR amplification of the genomic DNA sequence. Because no in-frame stop codon was identified in the 5'-RACE products, PCR amplification of genomic DNA was performed using two sets of primers complementary to the cDNA sequence. Eel genomic DNA (2 µg) was digested with EcoRV and HincII at 37°C for 3 h and self-ligated using a DNA ligation kit (Takara) at 16°C overnight to produce circular DNA; one of the two restriction enzymes should have a cleavage site in the cDNA sequence, and the primers should be designed using the cDNA sequence upstream of the restriction site. The first PCR amplification was performed using antisense primer N1 (5'-CCTCTGCAAGATGCGAATCA-3') and sense primer C1 (5'-ACTGCAAACACTGATGGA-3'). The PCR reaction was carried out in an MJ Research thermal cycler for 30 cycles of denaturation (94°C, 20 s), annealing (60°C, 30 s), and extension (72°C, 45 s) with an additional 7-min primer extension after the final cycle using Ex Taq polymerase (Takara). To reduce background and increase target fragments, a second PCR amplification was performed using another set of nested primers, N2 (5'-GGCATAAATGTGTCCAGA-3') and C2 (5'-AAGCCCTGTTGCAGAGAG-3'), under identical PCR conditions. The PCR reactions produced a band of ~950 bp, which was purified, subcloned, and sequenced as described above.

Antibody production. A 208-bp fragment encoding part of the intracellular NH₂-terminal region of eUT (amino acid residues 11-79) was subcloned into bacterial expression vector pRSET B (Invitrogen), and the construct was transformed into E. coli XL1-Blue. When the turbidity of the bacterial culture measured at 600 nm reached 0.6, isopropyl--D-thiogalactopyranoside was added to a final concentration of 1 mM, and incubation was continued for an additional 5 h at 37°C. The cells were harvested from a 1-liter culture by centrifugation and resuspended in 20 ml of lysis buffer (50 mM Na₂PO₄, 300 mM NaCl), disrupted by sonication, and centrifuged at 10,000 g for 20 min. The pellet was solubilized by resuspension and mixing in 8 M urea, 50 mM Na₂PO₄, 300 mM NaCl, pH 7.6, for 2 h at 4°C. The mixture was then centrifuged at 10,000 g for 40 min to remove any insoluble materials. Urea-solubilized recombinant fusion protein (His₆-eUT-N) was purified in the denatured state using a Ni²⁺-NTA agarose column (Qiagen) and dialyzed against saline at 4°C. Polyclonal antibodies to His₆-eUT-N were prepared in Japanese white rabbits by injecting about 250-300 µg of purified recombinant protein emulsified in complete Freund's adjuvant (1:1) subcutaneously at multiple sites. The rabbits were injected five times at 2-wk intervals and bled 7 days after the fifth immunization.

Affinity purification of anti-eUT antibody. The antibody was purified on an N-hydroxysuccinimide (NHS)-activated HiTrap affinity column (Amersham Pharmacia Biotech, Uppsala, Sweden) according to the manufacturer's instructions. His₆-eUT-N fusion protein (400 µg in 500 µl of saline) was coupled to 1 ml of NHS-activated Sepharose, and then 1 ml of anti-eUT antiserum (diluted 1:1 in PBS) was applied to the column and incubated overnight at 4°C. The bound antibody was eluted with 4 M MgCl₂ and dialyzed against 1 liter of PBS overnight at 4°C.

Western blot analysis. Freshwater and seawater eel gills were homogenized in ice-cold Tris-buffered saline (TBS: 10 mM Tris · HCl and 150 mM NaCl, pH 7.4) containing 2 mM phenylmethylsulfonyl fluoride, 20 µg/ml leupeptin, 10 µg/ml aprotinin, and 10 µg/ml pepstatin. The homogenates were centrifuged at 5,000 g for 20 min, and the pellets were resuspended in the same buffer. These procedures were repeated three times at 4°C. Finally the pellets were solubilized in TBS containing 1% Triton X-100. Total protein concentrations in the samples were measured using a Pierce bicinchoninic acid protein assay kit (Pierce, Rockford, IL) and adjusted to ~1 µg/µl with the solubilization solution. The membrane proteins were separated by SDS-PAGE using a 10% polyacrylamide gel and electroblotted onto a PVDF membrane. After blocking in 10 mM Tris · HCl, pH 8.0, containing 150 mM NaCl, 0.05% Tween 20, and 5% nonfat milk for 1 h at room temperature, the membrane was incubated with anti-eUT antiserum, anti-Na⁺-K⁺-ATPase antiserum, preimmune serum, or preabsorbed antiserum at 1:1,000 dilution overnight at 4°C. Preabsorption was carried out by incubating 10 µl of the antiserum with 400 µg of affinity-purified fusion protein (His₆-eUT-N) in 500 µl of PBS overnight at 4°C. The immune complexes on the membrane were then reacted with alkaline phosphatase-conjugated goat anti-rabbit IgG at 1:3,000 dilution for 1 h at room temperature. The bound secondary antibody was visualized using 5-bromo-4-chloro-3-indolyl phosphate-nitroblue tetrazolium chloride as chromogenic substrates.

Immunohistochemistry. Seawater eel gills were fixed in 0.1 M PBS, pH 7.4, containing 4% (wt/vol) paraformaldehyde, successively incubated in 1× Hanks' balanced salt solution (HBSS: 130 mM NaCl, 5 mM KCl, 0.5 mM Na₂PO_4, 100 mM HEPES, pH 7.4) containing 10% sucrose overnight and 1× HBSS containing 20% sucrose for 30 min at 4°C, and finally frozen in Tissue Tek optimum cutting temperature compound. Frozen sections (6 µm) were cut in a cryostat at 20°C and mounted on Vectabond-coated glass slides and dried in air for 45 min. The sections were washed twice with PBS, fixed with 2.5% (vol/vol) H₂O₂ in PBS for 1 h at room temperature, and again washed with PBS. They were subsequently incubated for 2 h at room temperature with 2.5% (vol/vol) normal goat serum, and then for 48 h at 4°C with affinity-purified anti-eUT antibody (1:2,000), preimmune serum (1:1,000), or anti-Na⁺-K⁺-ATPase -subunit antiserum (1:30,000) unless stated otherwise. After washing with PBS, the sections were incubated with peroxidase- and goat anti-rabbit IgG-conjugated dextran polymer (EnVision+, Dako Japan, Kyoto, Japan) for 2 h at room temperature. The bound antibody was visualized using 3,3'-diaminobenzidine tetrahydrochloride containing nickel chloride and 0.02% H₂O₂ in 50 mM Tris · HCl, pH 7.4, according to standard double-labeling techniques.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Identification of differentially expressed genes from subtracted cDNA libraries. In an attempt to isolate genes that are specifically expressed in freshwater- or seawater-adapted euryhaline fishes, we have generated two types of subtracted cDNA libraries from poly(A)⁺ RNA preparations of freshwater and seawater eel gills, as described in METHODS. The subtracted cDNA libraries were expected to be enriched in sequences specifically expressed either in freshwater or seawater. We picked up ~300 individual clones from each library, sequenced their inserts, and classified them by a database search into several categories including channels, transporters, receptors, transcription factors, structural proteins, housekeeping enzymes, and proteins with no match to previously studied ones. The subtraction is usually not complete, and therefore false positives should be eliminated. Furthermore, for the analysis of the functional role of the freshwater- or seawater-specific clones, identification of the cells that express the corresponding proteins is necessary, which means that their expression levels should be relatively high. To fulfill these criteria, we performed Northern analysis using total RNA preparations from freshwater and seawater eel gills and obtained seven candidate clones that exhibited differential expression between freshwater and seawater and relatively dense bands on Northern analysis. Among them, there was a 337-bp clone encoding a part of a UT and exhibiting strong, although not exclusive, expression in seawater eel gill (Fig. 1). As mentioned above, although the gill has been established as the major site of urea secretion, the cells involved have not been determined. We therefore decided to characterize the UT clone first.

View larger version (47K): [in this window] [in a new window]   Fig. 1.   Differential expression of eel urea transporter (eUT) mRNA in the gills of seawater (SW)- and freshwater (FW)-adapted eels. Total RNA preparations (20 µg/lane) isolated from seawater- and freshwater-adapted eel gills were analyzed by Northern blot analysis using a 32P-labeled eUT probe under high-stringency conditions. A single transcript was detected at 1.9 kb. Hybridization to an eel -actin probe (corresponding to nt 206-343 in rat sequence) demonstrated equal loading of each lane.

Cloning and sequence analysis of full-length cDNA encoding eUT. On Northern blot analysis, the 337-bp eUT probe hybridized to a 1.9-kb transcript (Fig. 1). To isolate eUT cDNA clones of this size, we constructed a seawater eel gill cDNA library in the ZAP II vector and obtained four positive clones by screening 3 × 10⁵ recombinant plaques. The longest clone was of 1,828 bp and contained the apparent 3'-end of the sequence, including a polyadenylation signal (AATAAA) and a poly(A) tail. It was, however, not clear whether the clone contains the 5'-noncoding region because it lacks an in-frame stop codon upstream of the first methionine. We therefore performed 5'-RACE to determine the 5'-end of eUT mRNA. The 5'-RACE yielded 18 bp of additional sequence, extending the total length of the sequence from 1,828 to 1,846 bp (Fig. 2A), but no in-frame stop codon was present. To establish the translational start site (ATG), we next isolated a genomic DNA clone containing the corresponding region and sequenced it (Fig. 2B). The genomic clone contained an in-frame stop codon 24 bp upstream of the initiator methionine, firmly establishing the open reading frame of eUT cDNA (Fig. 2A). The nucleotide sequence around the initiator methionine codon ATG (ATATTCATGC), however, does not conform well to the consensus sequence for translation initiation in higher eukaryotes [GCC(A/G)CCATGG] (12, 13).

View larger version (74K): [in this window] [in a new window]   Fig. 2.   Nucleotide and deduced amino acid sequences of eUT cDNA. A: the nucleotide sequence of the longest clone including that of 5'-rapid amplification of cDNA ends (5'-RACE) product and its amino acid sequence are shown. Numbers at left refer to the first amino acids on the lines, and numbers at right refer to the last nucleotides on the lines. *, Stop codon. Ten putative transmembrane (TM)-spanning regions (TM 1-TM 10) are indicated with solid underlines, and LP-box and TYPE domain are marked with dotted underlines. Potential phosphorylation sites for protein kinase C (PKC) are indicated by circles, and putative polyadenylation signal is boxed. Conserved potential N-linked glycosylation site is shaded. The DDBJ/EMBL/GenBank accession no. is AB049726. B: sequence of 0.8-kb segment upstream of the transcription start site. Negative numbers at left refer to the nucleotide positions upstream of the 5'-end of eUT cDNA. TATA box, GC box, and CAAT box sequences are in boxes. The 5'-end sequence of eUT cDNA is shaded. *, The first in-frame stop codon upstream of the initiator Met. Note the absence of the splice acceptor site (AG) between the in-frame stop codon and the first Met, eliminating the possibility that the in-frame stop codon is in an intron. The DDBJ/EMBL/GenBank accession no. is AB049727.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Structural model of eUT protein. Putative transmembrane-spanning regions were predicted using the computer program PSORT (21). Potential membrane-spanning regions are numbered 1-10. Potential PKC phosphorylation sites are indicate by solid triangles. A potential N-glycosylation (Asn-239) site is present on the 3rd relatively long extracellular loop.

View larger version (151K): [in this window] [in a new window]   Fig. 4.   Amino acid sequence alignment of eUT and the UT family members of other species. The abbreviations, references, and accession numbers are as follows: tUT, toadfish gill UT (42), AF165893; mtUT, Magadi tilapia gill UT (41), AF278537; shUT, shark kidney UT (36), AF257331; fUT, frog UT (2), Y12784.1; rUT-A2, rat kidney UT of the A2 type (35), Q62668; rUT-A3, rat kidney UT of the A3 type (8), AF041788; rUT-B, rat UT of the B type (3), X98399.1. Black boxes denote identical amino acid residues; gray shading denotes residues identical to that of eUT; white areas denote nonidentity. *, A potential N-glycosylation site.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Genomic organization of the mammalian UT genes to clarify the nomenclature currently used (A) and a phylogenetic tree of the UTs of various species so far identified and sequenced (B). UT-A isoforms of mammalian urea transporters arise, by alternative usage of multiple transcription initiation sites and by alternative splicing of pre-mRNA (15), from a single gene (UT-A1) that is considered to be generated by gene duplication of an ancestral UT gene into UT-A2 and UT-A3.

Sequence of the promoter region of the eUT gene. Figure 2B shows a partial 872-bp sequence of the eUT gene covering the possible transcription initiation site. The gene contains a consensus TATA box, several CAAT boxes, and a number of putative transcription factor-binding sites, including the AP-3, Sp1, and GATA-1 sites.

Tissue distribution and time course of induction. To determine the multiple tissue distribution of eUT mRNA and to compare its expression levels in seawater and freshwater eels, we performed Northern blot analysis using total RNA preparations from various tissues of seawater and freshwater eels, including the gill, kidney, head kidney, heart, liver, posterior intestine, anterior intestine, stomach, and muscle. A strong signal of ~1.9 kb was detected only in the gill but not in other tissues examined (Fig. 6). Comparison between the seawater and freshwater eel gill samples indicated that there was a significant difference in the eUT mRNA expression during adaptation to seawater. Figure 7 shows the time course of the induction of eUT mRNA expression after transfer of eels from freshwater to seawater. The adaptive alteration occurred relatively slowly over a time course of hours to days (Fig. 7).

View larger version (75K): [in this window] [in a new window]   Fig. 6.   Tissue distribution of eUT using high-stringency Northern analysis. Eels were adapted to seawater or freshwater for 2 wk, and total RNAs were isolated from the indicated tissues of eels as described under METHODS. A: autoradiogram of a Northern filter containing RNA preparations from seawater eels. B: Northern analysis of freshwater eel samples. Hybridization with an eel -actin probe was used to estimate relative amounts of mRNA loaded to each lane. Amounts of total RNA were roughly adjusted by measuring their absorbance at 260 nm; weak signals in the liver lanes are due to relatively low abundance of the actin message compared with rRNA. -Actin transcripts were present as ~2.0-kb (ubiquitous) and ~1.8-kb (specific for skeletal muscle) species.

View larger version (68K): [in this window] [in a new window]   Fig. 7.   Time course of eUT expression in gills under the conditions of hypertonic stress. Freshwater eels were transferred to seawater, and their total RNA was isolated from the gills of 3 eels separately at the time indicated and subjected to Northern blot analysis. The set of 3 samples gave similar results; 2 of them are shown to show low variability among individuals. Hybridization to an eel -actin probe demonstrated equal loading of each lane.

Production of antiserum and immunochemical characterization of eUT. A rabbit polyclonal antiserum was generated against a bacterially expressed His-tagged protein containing amino acid residues 11-79 of the cytoplasmic NH₂ terminus of eUT. The specificity of the antiserum was characterized by Western blot analysis. A single band of 54 kDa was observed when Triton extracts of membrane preparations of the eel gill were subjected to Western blotting (Fig. 8). The size of the band corresponds well to the calculated molecular mass of eUT if one assumes that native eUT is glycosylated at Asn-293. Seawater samples gave a denser band than freshwater samples, indicating a higher expression level of eUT in seawater eels. This confirms the result of Northern analysis at the protein level. The staining was almost abolished by preabsorption of the antiserum with the antigen (Fig. 8).

View larger version (48K): [in this window] [in a new window]   Fig. 8.   Western blot analysis of eUT to determine the size of the antigen and specificity of antiserum. Membrane preparations from branchial cell homogenates were treated with Triton X-100 to solubilize membrane proteins, and the Triton extracts were electrophoresed under reducing conditions, transferred to polyvinyl difluoride (PVDF) membranes, and stained with either anti-eUT antiserum, absorbed antiserum, or preimmune serum. FW, sample from freshwater eel gill; SW, sample from seawater eel gill.

Production of antiserum against eel Na+-K⁺-ATPase. To identify the chloride cells in the eel gill, we raised an antiserum to one of its marker proteins, the -subunit of Na⁺-K⁺-ATPase (amino acid residues 469-773). The antiserum obtained was of high titer and monospecific; it stained a single band of ~100 kDa corresponding to the -subunit of Na⁺-K⁺-ATPase on Western blot analysis (Fig. 9) and could be used at a dilution of 1:30,000 for immunohistochemical staining as shown below.

View larger version (29K): [in this window] [in a new window]   Fig. 9.   Specificity of antiserum raised against the -subunit of eel Na+-K+-ATPase. PVDF membranes prepared as in Fig. 8 from Triton extracts of seawater eel gill were stained with anti-Na+-K+-ATPase antiserum.

Cellular localization of eUT in the gill. The localization of eUT in the gill was examined by immunohistochemistry using antiserum characterized above. The anti-eUT antiserum specifically stained chloride cells that are located near the basal region of the secondary lamella (Fig. 10A). The chloride cell localization was confirmed by staining consecutive sections with an antiserum against Na⁺-K⁺-ATPase, a well-established marker protein of the chloride cell (34) (Fig. 10B). Because chloride cells have a complex tubular system formed by extensive invaginations of the basolateral plasma membrane into the cytoplasmic space, basolateral membrane proteins usually give a diffuse, intracellular pattern of staining as typically seen in the cases of Na⁺-K⁺-ATPase (34, 44), Na⁺-K⁺-2Cl cotransporter (43), and a K⁺ channel (39). A similar staining pattern was obtained (Fig. 10A), indicating that eUT is present in the basolateral membrane of the chloride cells. No specific staining was observed with preimmune serum or absorbed antiserum (Fig. 10C).

View larger version (61K): [in this window] [in a new window]   Fig. 10.   Immunohistochemical localization of eUT in seawater-adapted eel gill. Serial sections of the gill were stained with affinity-purified antibody against recombinant eel UT (A), antiserum against recombinant eel Na+-K+-ATPase -subunit (B), and preimmune serum (C). Clear staining was obtained with anti-eUT antibody but not with preimmune serum. The Na+-K+-ATPase antiserum was used to identify chloride cells (CC). SL, secondary lamella. Scale bars, 25 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

While attempting to identify differentially expressed genes in freshwater and seawater eel gills using subtracted cDNA libraries, we isolated a clone encoding a UT and relatively highly expressed in seawater eel gill. Sequence analysis indicated that the UT consists of 486 amino acid residues; it lacks the ALE sequence that is diagnostic of UT-B (3). Although eel UT is similar to those of other species in the core region, including the 10 potential transmembrane segments, it completely differs from others in its much longer NH₂- and COOH-terminal regions (Fig. 4), including those of Lake Magadi tilapia gill (41), gulf toadfish gill (42), shark kidney (36), amphibian urinary bladder (2), and mammalian kidneys (31, 32). Because the eel sequence represents the first example in the ammonotelic teleosts, it remains to be seen whether the structural difference is a property unique to the eel gill UT or a feature common to those of ammonotelic teleost fishes.

Of the nine tissues studied, only the gill exhibited a strong signal of UT on Northern blot analysis (Fig. 6). Furthermore, immunohistochemistry demonstrated that the infoldings of the basolateral membranes or the tubular system of chloride cells is the site of UT localization. This location of the UT is consistent with the previous observation that organic substances such as methylene blue and [¹⁴C]urea were concentrated and excreted by chloride cells of seawater eels (18, 19). Its expression level, especially in seawater, was much higher than that expected for an ammonotelic teleost fish, which excretes most of the nitrogenous waste as ammonia; the eel gill UT message levels could easily be determined by Northern blot analysis using total RNA preparations without enrichment of poly(A)⁺ RNA, suggesting the previously unrecognized importance of UTs even in ammonotelic teleost fishes. It may be involved in facilitated transport of urea derived from the direct cleavage of dietary arginine and from the uricolytic pathway of purine catabolism. It may also constitute a component of a fail-safe system to avoid ammonia intoxication, namely, it may serve as a carrier of urea synthesized de novo through CPSase III when pathways of ammonia excretion are precluded by environmental parameters. For example, in seawater, euryhaline fishes such as the eel excrete only small amounts of concentrated urine compared with the production of very large amounts of dilute urine in freshwater; in such seawater conditions, ammonia excretion from the kidney becomes virtually inoperative. Chloride cell localization of eUT suggests the possible dependency of its transport activity on Na⁺ because the basolateral membrane of the chloride cell has been demonstrated to be heavily laden with Na⁺-K⁺- ATPase (34).

Although the gill has long been known as the site of urea excretion since the classic work of Smith (37), it is only recently that the process has been suggested to involve a transport protein. For example, Wright et al. suggested, in their earlier work (48), a mechanism of passive simple diffusion, but have recently demonstrated the presence of a phloretin-sensitive UT in the gill (7, 25). As mentioned above, more direct evidence for the presence of a branchial UT has been provided by Walsh et al. (42), who isolated cDNA encoding a phloretin-sensitive UT from the gill of the ureotelic gulf toadfish and showed its exclusive expression in the gill. More recently, Walsh et al. (41) have cloned a UT from the gill of the Lake Magadi tilapia (Alcolapia grahami), an unusual fish excreting all its nitrogenous waste as urea because of its highly alkaline habitat, which is unfavorable for the passive diffusion of ammonia. Our identification of a UT in the ammonotelic eel gill, if combined with these two examples of unusual species of teleosts that synthesize significant amounts of urea to cope with environmental restrictions, may help establish the generality of the carrier-mediated mechanism of urea transport even in the gill and make it necessary to modify the traditional view of the simple diffusional model of urea transport in the gill. In some elasmobranchs, including the freshwater sawfish Pristis microdon Latham (38) and skates (26), a good correlation was found between the salinity and the rate of branchial urea secretion. In this context, the enhanced expression of the UT observed in seawater eel gills is intriguing, and it would be interesting to determine 1) whether gill urea excretion increases when eels are adapted to seawater and 2) whether the message levels of UTs undergo salinity-dependent changes in the gills of the above-mentioned elasmobranchs when their cDNA probes become available.

Sequence analysis of the 5'-flanking region of the eUT gene indicated the presence of canonical TATA and CAAT boxes in the appropriate location. These elements are common to promoters of a variety of highly regulated eukaryotic genes, including cell-type specifically, developmentally, nutritionally, and hormonally regulated genes. The eUT promoter region also contains several putative transcription regulatory elements, including the binding sites for Sp1 and GATA-1. Further characterization of the promoter sequence will help to identify the transcription factors that orchestrate the expression of eUT and determine its chloride cell specificity.

In summary, we have isolated a cDNA clone encoding the UT isoform of the UT family from an eel gill cDNA library and demonstrated chloride cell specificity and salinity-dependent regulation of its expression. High-level expression of the transporter in the chloride cells indicates that the UT plays an unexpectedly important physiological role even in ammonotelic teleost fishes.

Perspectives