Claudins are a large family of integral transmembrane tight junction (TJ) proteins involved in regulating the permeability of the paracellular pathway. In these studies, we clone and describe the tissue distribution of four claudin-3 genes (designated Tncldn3a, Tncldn3b, Tncldn3c, and Tncldn3d) from the euryhaline spotted green puffer fish Tetraodon nigroviridis and examine the response of Tetraodon and Tncldn3 mRNAs to salinity variation (freshwater, FW; seawater, SW; and hypersaline seawater, HSW). In Tetraodon, genes encoding for claudin-3 TJ proteins are widely expressed, suggesting that claudin-3 proteins participate in regulating paracellular permeability across various epithelia within fishes. Of particular note is the widespread distribution of Tncldn3 genes in tissues that regulate hydromineral balance (gills, skin, kidney, and intestine). Renal and intestinal tissues express all four Tncldn3 genes, while the gills and skin specifically express Tncldn3a and Tncldn3c. In response to salinity variation, Tetraodon exhibits characteristics typical of a euryhaline fish species: moderate changes in blood osmolality and muscle moisture content; alterations in gill, kidney, and intestinal Na+-K+-ATPase activity; and unaltered Na+-K+-ATPase activity in the integument. In conjunction with these changes, Tncldn3 mRNA expression exhibits marked and significant salinity-dependent alterations that are both tissue and gene specific. Overall, our data suggest that a decreased abundance of claudin-3 in Tetraodon occurs in “leakier” epithelia and that claudin-3 TJ proteins will likely play an important role in the maintenance of hydromineral balance across osmoregulatory epithelia of euryhaline fishes.
- paracellular permeability
varying components of the paracellular and transcellular transport pathways collectively govern a myriad of physiological properties across vertebrate epithelia. The transcellular pathway uses membrane-bound apical and basolateral transporters/channels to move solutes through epithelial cells, while the paracellular pathway regulates overall epithelial “tightness,” as well as solute permeability, charge, and (or) size selectivity between epithelial cells. Properties of the paracellular pathway are dictated largely by the tight junction (TJ) complex, which is located between epithelial cells at the apical-most region of the lateral cell membrane (13). Adjacent epithelial cells form protein strands of TJs that adhere to one another, creating a solute-selective semipermeable “gate” or barrier. In addition, TJs polarize epithelial cells by encircling the apical-lateral region of a cell, forming a “fence” that separates membrane-bound apical and basolateral proteins. The TJ complex is composed of integral transmembrane TJ proteins, as well as peripheral scaffolding TJ proteins and to date, upward of 40 proteins have been described in association with the mammalian TJ complex alone (13, 40).
Claudin (Cldn) proteins are a large family of integral transmembrane TJ proteins, the first members of which were described in 1998 (12). Cldns are now generally accepted to be major determinants of normal paracellular function, as well as key elements in the pathophysiology of epithelial dysfunction (for a review, see Ref. 40). In mammals there are ∼24 Cldn proteins that are expressed in a tissue-specific manner in association with particular epithelial characteristics (21, 40). Overexpression of individual Cldn proteins in cultured epithelial monolayers variously increases or decreases transepithelial resistance (TER; an indication of epithelial “tightness”) and alters epithelial permeability to small tracer molecules and ions such as Na+ and Cl− (40). Cldns have also been identified in other vertebrates, and as a group, appear to have generally expanded along the chordate lineage (20). However, genome duplication events in fishes (2, 6, 39), as well as tandem gene duplication, presumably in response to the unique physiological challenges associated with diversification in an aquatic habitat, have seen the Cldn family greatly expand in this group (24).
Hydromineral balance in teleost fishes is regulated across epithelia of the gill, skin, gastrointestinal (GI) tract, and renal system (for a review, see Ref. 27). Briefly, skin and gill epithelia are directly exposed to surrounding water and are responsible for the separation of the blood space from the aquatic environment. The skin is generally seen as a relatively impermeable barrier to salt and water movement in fishes; however, the gill epithelium intricately controls the movement of solutes between the two fluids. In freshwater (FW), the gill actively acquires salts from surrounding water, while in seawater (SW), the gill actively secretes salt from the blood to the water. Excess systemic salt in SW teleosts mostly comes from the tendency of these fishes to imbibe surrounding salt water. In a SW environment, water is needed to replace that lost by osmosis across the gill epithelium. This water is acquired across the GI tract; however, GI water absorption cannot occur without an associated obligatory salt load. It is this salt load that is secreted primarily across the gill. FW fish generally do not imbibe water as they must cope with tissue hydration rather than dehydration. Therefore, FW fish use the kidney to produce relatively large volumes of dilute urine. This is in contrast to the SW kidney, which produces an isotonic urine that is rich in the salts not secreted across the gill.
All of the aforementioned processes take place using a suite of membrane-bound transcellular transport proteins, as well as the semipermeable paracellular pathway. Mechanisms of ion transport across the transcellular pathway are well characterized in fishes (10, 26). Furthermore, ultrastructural studies on the TJ complex (e.g., 37), as well as electrophysiological studies using ionoregulatory epithelia (18, 25, 26) strongly support TJ heterogeneity as a critical component of the overall physiological strategies that allow fishes to dominate the aquatic environment. Yet despite this, far less is known about the paracellular pathway and its components in this vertebrate group, and almost nothing is known about TJ proteins in fishes and how these proteins respond to the physiological challenges associated with an aquatic lifestyle (5, 20, 24).
With this background in mind, we set out to further understand the TJ complex in fishes by examining the response of Cldn3 to environmental change in a euryhaline model organism, the spotted green puffer fish (Tetraodon nigroviridis). Cldn3 is an important and widely distributed integral transmembrane TJ protein in vertebrates; therefore, we hypothesized that Cldn3 would be distributed widely in T. nigroviridis (Tetraodon). Furthermore, on the basis of recent observations of gene distribution in a closely related model organism, Fugu rubripes (24), we hypothesized that Cldn3 in Tetraodon would be strongly associated with tissues involved in the maintenance of hydromineral balance in fishes. To address this, we built upon studies that used F. rubripes, in which an expanded claudin (cldn) gene superfamily was annotated relative to the mammalian Cldn superfamily (24) and cloned an expanded cldn3 family in Tetraodon using its recently sequenced genome (17). We then hypothesized that cldn3 gene expression would respond to environmental change in a tissue-specific manner and addressed this by acclimating Tetraodon to hypoosmotic and hyperosmotic surroundings (freshwater, FW; seawater, SW; hypersaline seawater, HSW). Tissues were collected and examined for endpoints of hydromineral status, as well as alterations in cldn3 gene expression. To the best of our knowledge, these studies are among the first to comprehensively address the response of TJ proteins to environmental change in fishes, a vertebrate group that is often faced with unique physiological challenges.
MATERIALS AND METHODS
Spotted green puffer fish (T. nigroviridis) were purchased from a local supplier and held in 75-liter opaque aquaria with recirculating brackish water composed of 5 g/l synthetic sea salt (Instant Ocean, Aquarium Systems) dissolved in dechlorinated tap water. The salt content of brackish water was confirmed to be 5‰ by measuring salinity with a hand-held refractometer (SR-6 VitalSine Refractometer). Water was aerated, and temperature was held constant at 24 ± 1°C using an immersion heater. Fish received a constant photoperiod cycle of 12:12-h light-dark cycle and were fed ad libitum once daily with BioPure blood worms (Hikari Sales). After a settling period of 2 wk, fish used for salinity acclimation experiments were divided into three groups and held in 60-liter opaque glass aquaria with recirculating brackish water, as described previously. These fish were gradually acclimated to one of three experimental salinities: freshwater (FW, ∼0–1‰), seawater (SW, 33‰), and hypersaline seawater (HSW, 45‰). Fish held in SW and HSW were gradually acclimated to new conditions by adding Instant Ocean synthetic sea salt at a rate resulting in daily salinity increments of 1–2‰. Fish were acclimated to FW by diluting brackish water with dechlorinated tap water (daily reductions ∼0.5–1‰). Once the desired salinity of each group was obtained, the fish were allowed to acclimate to this new environment for 2 wk. During the course of the experiment, fish were fed as previously outlined. Salinity and water temperature were monitored daily.
Blood and tissue sampling.
Fish were net captured and anesthetized in 0.5 g/l tricaine methanesulfonate (MS-222; Syndel Laboratories). In each case, fish were anesthetized in water of appropriate salinity, and tissues for all experiments were collected from animals that were unfed 24 h before sampling. For gene identification and tissue expression profiles, appropriate tissues were carefully dissected from fish taken from brackish water stock tanks. Tissues were immersed in TRIzol Reagent (Invitrogen) and either processed immediately (for RNA extraction procedures, see below) or quick-frozen in liquid nitrogen and stored at −85°C for later analysis. For experiments designed to examine the response of TJ gene expression in response to alterations in environmental salt concentrations, blood was collected from anesthetized fish prior to tissue removal. Briefly, the skin of anesthetized fish was quickly rinsed with distilled water and blotted dry, and the dorsal skin was partly removed to expose the underlying musculature. Blood was then collected into microhematocrit capillary tubes (Fischer Scientific) after severing the spine. Blood was allowed to clot at room temperature for 30 min and centrifuged (using a Haematokrit 20 centrifuge; Hettich Zentrifugen) at 10,000 rpm for 5 min. Serum was separated from packed blood cells and stored at −30°C for later analysis of osmolality. A full flank of white muscle was removed to analyze muscle moisture content. Gill (complete basket), kidney (one full lobe), intestine (midway between anterior- and posterior-most region of the GI tract), and skin (dorsal only) samples were removed, quick-frozen in liquid nitrogen, and stored at −85°C until further analysis. All experiments were carried out in accordance with the principles published in the Canadian Council on Animal Care's guide to the care and use of experimental animals and an experimental protocol approved by York University Animal Care Committee.
Claudin-3 gene identification in tetraodon.
Using the National Center for Biotechnology Information (NCBI), we retrieved F. rubripes (Fugu) claudin-3 gene (fu-cldn 3) sequences. To identify homologous claudin-3 loci in Tetraodon nigroviridis (Tetraodon), fu-cldn3 sequences were used to search the Tetraodon genome using Genoscope's BLAT program (http://www.genoscope.cns.fr/externe/tetranew/). Specifically, the sequence with the highest BLAT score (and thus highest degree of sequence identity) was chosen and was submitted to a BLAST search against the nonredundant (nr) nucleotide database (1) to cross-reference and confirm sequence identity with fu-cldn 3 orthologs. Once a putative Tetraodon claudin gene was confirmed by BLASTn, the sequence was aligned with both the relevant fu-cldn and fu-cldn coding sequence (CDS) using GeneDoc (32). Subsequently, a Tetraodon CDS with high homology to a specific fu-cldn CDS was used as a template for primer design.
Nomenclature and phylogenetic analysis.
Identified Tetraodon claudin genes (Tncldn) were designated a name/number based on the nomenclature of the Fugu claudin gene family (24), where fu-cldn3 identity was assigned according to a human ortholog. Phylogenetic analysis was conducted using all identified Tetraodon claudin-3 (Tncldn3) genes to confirm gene orthologs and to infer the correct order of different possible gene duplication events (i.e., tandem gene duplication and/or whole-genome duplication). The deduced amino acid sequences for each Tncldn3 gene, all four previously reported fu-cldn 3 amino acid sequences (fu-cldn 3a, GenBank accession no. AY554377; fu-cldn 3b, GenBank accession no. AY554378; fu-cldn 3c, GenBank accession no. AY554367; fu-cldn 3d, GenBank accession no. AY554368) and fu-cldn 33a (outgroup, GenBank accession no. AY55466) were aligned with ClustalX version 1.8 (16). Using PHYLIP 3.6b's PROTDIST (Jones-Taylor-Thornton model) and CONSENSE programs (11), phylogeny was inferred, and the tree was visualized using Treeview X (http://darwin.zoology.gla.ac.uk/∼rpage/treeviewx/index.html).
Gene expression profiles.
RT-PCR was used to examine gene expression profiles for each identified Tncldn3 gene. To extract total RNA from tissues taken from brackish water fish, samples were first homogenized in TRIzol Reagent (Invitrogen) using a PRO250 homogenizer (ProScientific) and then extracted as per the manufacturer's instructions. The following macroscopically dissected organs of Tetraodon were used: brain, eye, gill, heart, intestine, liver, spleen, kidney, muscle, skin, ovary, and testis. Initially, all RNA samples were treated with DNase I (Amplification Grade; Invitrogen) following the manufacturer's instructions. Next, first-strand cDNA was generated using Superscript III Reverse Transcriptase along with Oligo(dT)12–18 primers (Invitrogen). Primers used for each Tncldn3 and β-actin (acting as an internal control) can be seen in Table 1. β-actin primers were designed based on a previously published study (38). PCR amplification of Tncldn3s and β-actin was performed under the following conditions: 4 min denaturation at 95°C for one cycle followed by 40 cycles of denaturation at 95°C (30 s), annealing at 51–55°C (45 s), and extension at 72°C (30 s), respectively, and a final single extension at 72°C for 5 min (0.2 mM dNTP, 0.2 μM forward and reverse primers, 1 × Taq DNA polymerase buffer, 1.5 mM MgCl2, and 1 U Taq DNA polymerase). Tncldn3 and β-actin PCR amplicons were resolved by agarose gel electrophoresis (1%; stained with ethidium bromide) for ∼90 min at 105 V. Images used for expression profiles were captured using a Multiimage Light Cabinet (AlphaImager HP model; AlphaInnotech). Sequence identity of PCR products was examined by sequence analysis (Molecular Core, York University, Canada).
Quantitative real-time PCR analysis.
Total RNA was extracted from tissues of interest (i.e., gills, intestine, kidney, and skin) using TRIzol reagent and DNase treated; then, cDNA was synthesized using Superscript III as outlined previously. Quantitative real-time PCR (qRT-PCR) was carried out using a Chromo4 Detection System (CFB-3240, Bio-Rad Laboratories) and SYBR Green I Supermix (Bio-Rad Laboratories). A standard curve was constructed for each gene of interest to 1) optimize the template cDNA concentration and 2) verify that the threshold cycle (Ct) fell into an acceptable range. Primers that were used for the expression profiles were also used for qRT-PCR analysis (see Table 1). However, for qRT-PCR reactions, conditions were changed slightly: 4 min at 95°C for one cycle, then 30 s at 95°C, 45 s at 51–55°C and 30 s at 72°C for 40 cycles. A melting curve analysis was carried out after each run to confirm the efficiency of the qRT-PCR conditions and to verify that no primer-dimers or other nonspecific products were synthesized in the reaction tubes. β-actin was used as an internal control for qRT-PCR analysis of Tncldn3 genes, and all samples were run in duplicate.
Serum osmolality, muscle moisture analysis, and Na+-K+-ATPase activity.
Serum osmolality was determined using an osmometer (5500 Vapor Pressure Osmometer; Mandel Scientific). Muscle moisture content was determined gravimetrically after drying muscle to a constant weight at 60°C for 3 days. Na+-K+-ATPase activity was determined in gill, intestine, kidney, and skin tissues using methods previously outlined (28).
All data are expressed as mean values ± SE (n) where n equals the number of fish in a treatment group. To examine for significant differences between groups, data were subjected to a one-way ANOVA. When statistical differences were found by one-way ANOVA, a Holm-Sidak method was used to delineate differences between groups. A fiducial limit of P ≤ 0.05 was used throughout. All statistical analyses were conducted using SigmaStat 3.1 (Systat Software).
Using a combination of bioinformatic tools [i.e., NCBI database, Tetraodon genome database (Genoscope), BLAT and BLASTn], four claudin-3 genes were identified in the Tetraodon genome. These were designated Tetraodon nigroviridis claudins-3a, -3b, -3c, and -3d, and abbreviated Tncldn3a, Tncldn3b, Tncldn3c, and Tncldn3d, respectively. All Tncldn3 genes comprise one exon and range in size from 1,122 to 4,181 bp. Tncldn3a and Tncldn3b are located adjacent to one another on chromosome 16 and exhibit 89.1 and 88.9% sequence identity to Fugu claudins-3a and -3b, respectively. In contrast, Tncldn3c and Tncldn3d are located adjacent to one another on chromosome 7. When Tncldn3c and Tncldn3d were compared with Fugu cldn3c and cldn3d, sequence identity was found to be 88.1 and 85.2%, respectively. Full details of the Tncldn3 genes are outlined in Table 2. Partial regions of identified Tncldn3 genes amplified by RT-PCR yielded PCR amplicons of expected band size (see Table 1). When sequenced and aligned, these PCR products were confirmed to be identical to those initially identified using GeneDoc (data not shown).
Phylogenetic analysis of Tetraodon cldn3 genes.
Our analysis of the Tetraodon genome, relative to Fugu, revealed only a single highly similar ortholog for each cldn3 gene. To assess phylogenetic relatedness between and among the Tncldn3 genes and their respective orthologs in Fugu, we generated a phylogenetic tree using PHYLIP (Fig. 1). Cladistic analysis was carried out using the deduced amino acid sequences for the Tncldn3 genes. We were able to infer the following: 1) each Tncldn3 gene has a respective ortholog in Fugu (e.g., Tncldn3a and fu-cldn 3a are grouped together) and 2) cldn3 genes are grouped in pairs; i.e., cldn3a and cldn3c are grouped together and closely related, whereas cldn3b and cldn3d are more related.
Tncldn3 expression profiles.
Using RT-PCR and agarose gel electrophoresis, we examined tissue-specific expression patterns of Tncldn3 genes. We show that these genes are variously distributed in tissues examined, with Tncldn3a and Tncldn3c exhibiting widespread distribution while Tncldn3b was found to be moderately widespread, and Tncldn3d exhibited a more restricted distribution pattern (Fig. 2). More specifically, moderate to strong expression of Tncldn3a was found in tissues of the brain, gill, heart, intestine, liver, kidney, and gonads, whereas weaker expression of Tncldn3a was detected in the skin. Very weak to undetectable levels of Tncldn3a were found in eye, spleen, and muscle tissue. Strong Tncldn3c expression was detected in the gill and skin, while moderate expression was detected in the intestine and testis (Fig. 2). In contrast, weak expression levels seemed to occur in brain, eye, heart, kidney, muscle, and ovary tissue, with barely detectable expression occurring in the liver and spleen. Strong expression of Tncldn3b was detected in heart, intestine, liver, and gonads, while moderate to weak expression was found in brain, eye, and kidney tissue. No Tncldn3b expression was detected in gill, spleen, muscle, or skin. Tncldn3d was only detected in the intestine and kidney, where expression was marked (Fig. 2). As an internal control, β-actin levels were highly expressed in all tissues (Fig. 2).
For a quantitative measure of Tncldn3 gene expression in tissues involved in the regulation of hydromineral balance, qRT-PCR analysis of gill, intestine, kidney, and skin were conducted for Tncldn3 genes present in each respective organ. For tissue exposed directly to the external environment, that is, gill and skin, Tncldn3c is highly expressed relative to Tncldn3a (Fig. 3). In contrast, kidney and intestine tissues exhibit high levels of Tncldn3d relative to any other Tncldn3 gene (Fig. 3).
We also confirmed the absence of select Tncldn3 genes in Tetraodon under varying salinity conditions. That is, where a tissue of interest did not express a Tncldn3 in our expression profiles (i.e., Tncldn3b and Tncldn3d were not detected in the gill or skin), 3 or 4 random samples from FW, SW, and HSW fish were examined using RT-PCR and qRT-PCR to confirm that no environmentally induced transcript appearance could be detected. In all cases, no Tncldn3b or Tncldn3d was found in gill or skin tissue.
Serum osmolality, muscle moisture content and Na+-K+-ATPase activity in response to salinity variation.
Serum osmolality significantly increased in a stepwise manner with increasing salinity. Fish held in SW exhibited a marginal but significant increase in osmolality relative to FW fish, while the HSW group exhibited a significantly greater serum osmolality than both FW and SW groups (Fig. 4A). In contrast to osmolality, muscle moisture content was seen to decrease as salinity increased (Fig. 4B). In particular, the HSW treatment group showed a significant decrease compared with the other groups.
Na+-K+-ATPase activity responded to salinity variation in all tissues examined with the exception of the skin (Fig. 5). The gills of Tetraodon exhibited a stepwise increase in Na+-K+-ATPase activity in response to increasing salinity, with fish in SW and HSW displaying an ∼1.5- and 2.5-fold increase in ionomotive enzyme activity, respectively (Fig. 5A). A similar stepwise increase in enzyme activity was observed in the intestine of Tetraodon (Fig. 5B). In contrast, renal Na+-K+-ATPase responded in an opposite manner, exhibiting a stepwise decrease in activity in response to increments in the water salt levels (Fig. 5B).
qRT-PCR analysis of tncldn3 gene expression in response to salinity variation.
Acclimation of Tetraodon to varying salinity resulted in tissue-specific alterations in Tncldn3 expression. For example, Tncldn3a responded to salinity variation in the gills, intestine, and kidney by decreasing in a salinity-dependent manner (Figs. 6 and 7). In contrast, Tncldn3a expression in the skin increased in response to increased salinity (Fig. 6B). For clarity, the response of Tncldn3 mRNA expression to salinity change will be considered tissue by tissue.
Only Tncldn3a and Tncldn3c are expressed in the gills of Tetraodon. Both TJ genes exhibited a stepwise expression decrease in response to increased salinity (Fig. 6A). In contrast, the same TJ genes in the skin of Tetraodon were found to increase in response to increased salinity (Fig. 6B).
The expression of Tncldn3 in response to elevated salinity varied from gene to gene in intestinal tissue (Fig. 7A). More specifically, Tncldn3c did not alter at all as a result of acclimation of Tetraodon to SW or HSW. Tncldn3d appeared to exhibit a marginal decline in expression levels in SW and HSW; however, this reduction was not significant (P = 0.053). Significant reductions in both Tncldn3a and Tncldn3b were apparent in fish acclimated to SW and HSW (Fig. 7A). In the former (i.e., Tncldn3a), reductions in mRNA expression were stepwise, in accordance with a stepwise increase in salinity. In contrast, Tncldn3b significantly reduced in SW and exhibited no further decline in HSW (Fig. 7A).
In kidney tissue, Tncldn3a, Tncldn3b, and Tncldn3c exhibited significant changes in expression, whereas Tncldn3d showed no significant variation (Fig. 7B). Tncldn3a and Tncldn3b expression decreased significantly in a salinity-dependent manner, whereas Tncldn3c only reduced (albeit very significantly), in HSW fish (Fig. 7B).
Spotted green puffer fish (T. nigroviridis) possess four genes putatively encoding for claudin-3 (Cldn3) TJ proteins (Tncldn3a, Tncldn3b, Tncldn3c, and Tncldn3d). Tncldn3 genes are differently expressed in various tissues where TJs play an important role in maintaining the functional integrity of epithelia and endothelia (see Fig. 2). Furthermore, Tncldn3 genes are often highly expressed in tissues involved in the maintenance of hydromineral balance in fishes (gill, kidney, intestine, and skin). The intestine and kidney are the only tissues that express all four Tncldn3 genes, and epithelia that are exposed directly to the external environment (i.e., gill and skin) specifically express Tncldn3a and high levels of Tncldn3c. Taken together, these observations suggest extensive Cldn3 TJ protein distribution in Tetraodon and the potential involvement of Cldn3 TJ proteins in numerous physiological processes in fishes. Moreover, the specific expression patterns of Tncldn3 genes in osmoregulatory tissues and the response of these genes to salinity change, as reported in our study, provide convincing evidence that one role for Cldn3 TJ proteins in Tetraodon (and potentially other fishes) will include participation in the control of water and solute movement across osmoregulatory epithelia.
Identification and phylogenetic analysis of Tncldn3 genes.
The possession of four cldn3 genes by Tetraodon is consistent with the presence of four cldn3 genes in a closely related counterpart F. rubripes (24). It has been previously suggested that the Fugu claudin superfamily of genes underwent extensive duplications caused by tandem gene duplication and/or a whole genome duplication (WGD) event, proposed to have occurred in the fish lineage (2, 24). The most likely order of these events in Tetraodon appears to mirror those described by Loh et al. (24) for Fugu based on the following observations. First, each Tncldn3 gene (i.e., Tncldn3a-d) has only a single ortholog in Fugu (Fig. 1). Second, Tn/fu-cldn3a and cldn3c genes are grouped together while Tn/fu-cldn3b and cldn3d are more related (Fig. 1); hence, we can infer that a single tandem gene duplication initially occurred followed by WGD, as has been postulated (24).
Tissue-specific expression patterns of Tncldn3 genes.
Tncldn3 genes are widely expressed (Fig. 2), consistent with the relatively wide expression of cldn3 genes in other vertebrates (14, 19, 22, 35, 41, 42), as well as in Fugu (24). When comparing fish species (i.e., Tetraodon and Fugu), there are several instances in which the distribution of Tncldn3 genes seems to be broader in scope than those of Fugu. For example, Tncldn3a was found in all tissues except for the eye and spleen, and Tncldn3c was found in all tissues except for very low to negligible expression in the liver and spleen. In contrast, Loh et al. (24) found fu-cldn 3a in the intestine, kidney, and heart only, and fu-cldn 3c was restricted to the skin, gill, brain, eye, and intestine. Furthermore, the generally broad distribution of Tncldn3a, Tncldn3b, and Tncldn3c is in accord with the broad distribution of Cldn3 in other vertebrate tissues (14, 19, 22, 35, 41, 42), but the restricted presence of Tncldn3d to the intestine and kidney appears to be unique, at least for now, to Tetraodon and Fugu (this study, 24). A particularly notable difference between the two puffer fish species was the presence of several Tncldn3 genes in Tetraodon gonads (Fig. 2) compared with the absence of all fu-cldn 3 genes in Fugu gonadal tissue (24). In this regard, the occurrence of cldn3 genes in Tetraodon gonads is consistent with observations made in other vertebrates where Cldn3 has been reported in both testis (30) and ovarian epithelial cells (3).
When considering osmoregulatory tissues only, considerable overlap occurs between the expression of cldn3 genes in Tetraodon and Fugu. For example, there is a conspicuous presence of all four Tetraodon (this study) and Fugu (24) cldn3 genes in the intestine, suggesting that Cldn3 has an important role to play in the gastrointestinal tract. A single cldn3 gene has also recently been reported from the intestine of another fish species, Dicentrarchus labrax (5) that has been suggested to be involved in osmoregulation. In contrast, Fugu only strongly expresses one cldn3 gene in tissues exposed directly to surrounding water (i.e., fu-cldn 3c in the gill and skin), whereas Tetraodon prominently expresses two cldn3 genes in gill and skin tissue (Tncldn3a and Tncldn3c). Furthermore, Tncldn3a and Tncldn3c respond markedly to alterations in the salt content of surrounding water. These observed differences would be interesting to relate to the haloplasticity of the different species since Tetraodon is a strongly euryhaline organism capable of surviving in both FW and SW, whereas the FW tolerance of F. rubripes is limited (23).
The response of Tetraodon and Tncldn3 genes to altered environmental salinity.
Tetraodon responded to salinity variation in a manner typical of a euryhaline fish that naturally inhabits rivers and estuaries (36). Serum osmolality increased marginally in SW and to a greater extent in HSW, but otherwise remained relatively stable, whereas muscle moisture content only exhibited significant alterations (i.e., a decline) under hypersaline conditions. The maintenance of a stable internal environment in euryhaline fishes acclimating from one environment to another occurs, in part, due to significant modulation/reorganization of the epithelial transcellular pathway, that is, modulation of ionomotive enzyme activity, as well as membrane-bound apical and basolateral transport proteins and channels (for a review, see Refs. 10 and 26). In the current study, modulation of the transcellular pathway is reflected by changes in the primary ionomotive enzyme of fish osmoregulatory organs, Na+-K+-ATPase (Fig. 5). The second component that facilitates acclimation of euryhaline fishes to a changing environment is the reorganization of the paracellular pathway. To date, this latter phenomenon has been acknowledged by observed alterations in epithelial ultrastructure, electrophysiological properties, and ion efflux rates. In this study, we address a molecular basis for some of the observed changes.
Tetraodon gill Na+-K+-ATPase significantly increased in SW and further increased in HSW. In the euryhaline fish gill, higher Na+-K+-ATPase activity in SW drives active salt secretion by generating a transmembrane Na+ gradient that facilitates Cl− transport into gill mitochondria-rich cells (MRCs) via a Na+-K+-2Cl− cotransporter. Cl− then exits the MRC through apical Cl− channels (CFTR-type anion channels) (26). In contrast to Cl−, Na+ is not removed from the internal fluids of SW fish through a transcellular route, but instead, Na+ follows an electrochemical gradient along a paracellular pathway between the SW gill MRC and accessory cell (AC) (for a review, see Refs. 10 and 26). The paracellular pathway between the MRC and AC is constructed as a shallow “leaky” TJ (37). These “leaky” TJs are structurally very different from deep TJs, which are found between pavement cells (PVC) in the fish gill epithelium, as well as between MRCs and PVCs and the ACs and PVCs (37). Functionally, the “leaky” TJs of the SW fish gill are thought to create a leakier epithelium relative to the tighter properties of the FW fish gill. Although epithelial “tightness” has not been measured directly across the fish gill, because of limitations imposed by architectural complexity, transepithelial resistance measurements across surrogate models of the FW and SW fish gill support this contention, being reported in the range of 200–300 Ωcm2 vs. 600–10,000 Ωcm2 for SW and FW, respectively (44). Given these observations, it is interesting to consider the response of gill Tncldn3a and Tncldn3c to increased environmental salinity. Both genes exhibit a significant, stepwise, reduction in expression as salinity increases (Fig. 6A). This would suggest that Cldn3 abundance in the gill decreases in response to increasing environmental salt concentrations. Because Cldn3 is linked to regions of higher resistance in vertebrate epithelia (13) and Cldn3 overexpression has been demonstrated to reduce the paracellular permeability of vertebrate epithelia (8), it would seem reasonable to hypothesize that a reduction in Cldn3 expression in the gills would be adaptive in fish moving from FW to SW, possibly by contributing to the “leaky” SW gill phenotype.
In contrast to the gill epithelium, the skin of Tetraodon did not exhibit any alterations in Na+-K+-ATPase activity in response to salinity change, suggesting a passive barrier role for the epidermis in these fish. This is supported by histological studies of epidermal structure in two closely related species within the Tetraodon genus (T. steindachneri and T. fluviatilis), in which no definitive cytological evidence for an active role for the skin in osmoregulation has been described (15, 31). Nevertheless, the relatively uniform cells of the most superficial epidermal layer of Tetraodon skin are connected by TJs, and below this layer, cell-to-cell contact is maintained only by desmosomes (15). Therefore, as a barrier to ion and water loss or loading (salinity dependent), TJs of the superficial layer of epithelial cells in Tetraodon skin are likely to play an important role in the overall maintenance of hydromineral balance in these fish. Substantial increments in the expression of Tncldn3a and Tncldn3c in the skin of Tetraodon would seem to support this line of reasoning (Fig. 6B), and on the basis of our current knowledge of Cldn3 and its role in regulating paracellular permeability in vertebrate epithelia, it is possible that the skin of Tetraodon “tightens” in response to elevated environmental salt concentrations.
Cldn3 is broadly expressed along the gastrointestinal tract of mammals (35) and has been reported in the intestine of aves (14). In the rat, Cldn3 is found in the duodenum, jejunum, ileum, and colon, but not in the stomach (35). Cldn3 function in the vertebrate intestine has yet to be clearly defined. Recent studies have reported that treatment of Caco-2 epithelia with a mycotoxin (ochratoxin A), caused a marked decrease (∼87%) in Cldn3 expression, which was coupled to a substantial reduction in TER and concomitant increase in paracellular permeability (29). These observations are in general accord with studies relating pathological dysfunction in the gastrointestinal tract to increased intestinal permeability, a phenomenon normally associated with reduced TJ protein expression and changes in TJ protein distribution (7). In light of these observations, reductions in Tncldn3a and Tncldn3b (and to some extent Tncldn3d) expression may contribute to an increase in permeability across SW- and HSW-acclimated Tetraodon intestinal epithelia. However, it should be noted that an increase in intestinal permeability in SW and HSW Tetraodon would be a normal adaptive physiological phenomenon, which would be consistent with reported reductions in TER across the intestinal epithelium of numerous euryhaline fishes upon acclimation to SW (for a review, see Ref. 25). Supporting this are the observed changes in intestinal Na+-K+-ATPase activity, in which an increase in the activity of SW and HSW fish is a typical euryhaline response that energizes the absorption of Na+ and Cl−, which, in turn, is tightly linked to the absorption of water.
Studies using the murine kidney reported segment-specific expression patterns for Cldn TJ proteins along the nephron (19). A recent report also indicates that region-specific expression patterns of TJ proteins occur along the fish nephron (H. Chasiotis and S. P. Kelly, unpublished observations). Segment-specific Cldn expression likely reflects the varying physiological properties of the vertebrate nephron (4). For example, Cldn8 is found primarily in the distal regions of the nephron (where paracellular permeability is low), and overexpression of this protein in a renal cell line has been reported to reduce the paracellular permeability of monovalent and divalent cations (4, 46). In the mouse nephron, Cldn3 is found in the tighter middle and distal segments, including the thin and thick ascending limb of Henle, the distal tubule, and the collecting duct, but not in the leakier proximal regions (19). This associates Cldn3 with regions of reduced paracellular permeability in the nephron; however, to the best of our knowledge, functional studies regarding the physiological role of Cldn3 in the nephron have yet to be conducted. Observations of the FW fish nephron have drawn comparisons between the distal segment and the mammalian thick ascending limb of Henle. In the goldfish nephron, the distal tubule expresses abundant Na+-K+-ATPase and occludin protein, supporting the belief that this region likely acts as a “tight” diluting segment in FW fishes (9, 34). However, marine fish either appear not to possess a distal nephron segment [but instead two or three proximal segments, (27)] or have a distal segment that exhibits varying levels of structural and functional degeneration (33). If Cldn3 TJ proteins are associated with the distal regions of the teleost fish nephron, similar to Cldn3 association with the more distal regions of the mammalian nephron, then the above observations could partly explain a reduction in Tncldn3a, Tncldn3b, and Tncldn3c expression in Tetraodon acclimated to SW and HSW. This would also explain the observed reductions in kidney Na+-K+-ATPase activity (Fig. 5B), since a degeneration in distal segment abundance and/or function would presumably impact total protein activity from renal tissue. Another contributing factor could be a reduction in collecting tubule abundance. This possibility is supported by recent observations that collecting tubule abundance is markedly reduced in euryhaline fish acclimated to SW relative to those in FW (43).
A final comment should be made regarding the response of Tncldn3d to salinity change. That is, Tncldn3d exhibits the highest constitutive expression of all the Tncldn3s in both the intestine and kidney (see Fig. 3). However, in response to salinity change, there is no significant alteration in kidney Tncldn3d mRNA expression, and although a decrease in Tncldn3d mRNA expression is observed in the intestine, this alteration is marginal (i.e., P = 0.053). At this stage, it is difficult to speculate as to why Tncldn3d did not respond to salinity; however, it will be interesting to explore intratissue expression and any shifts of Tncldn3 (e.g., Tncldn3d) expression in future studies. In addition, if we were to assume that high mRNA levels reflect increased protein abundance, it is also possible that with respect to the role of claudin-3 proteins in these tissues, Tncldn3d maintains baseline epithelial permeability and that physiological alterations in response to environmental change may be modulated entirely by “lesser expressed” claudin-3 isoforms. This area will require further study.
Perspectives and Significance
As research continues to consolidate the fundamental role of TJ proteins in the regulation of epithelial barrier properties within mammalian systems, opportunities to examine the often unique properties of other vertebrate epithelia have yet to be fully realized. Cldn proteins have been described in various chordates and may have appeared with chordates themselves (see Refs. 20 and 24). Furthermore, members of the Cldn superfamily of integral transmembrane TJ proteins are now generally accepted to be key players, not only in the formation of TJs but also major determinants in varying the magnitude of the paracellular pathway, as well as its permselectivity. In epithelia of aquatic organisms, including those directly exposed to an aqueous environment that often fluctuates greatly in composition, paracellular permeability and permselectivity are critical for the maintenance of functional epithelial integrity and ultimately internal homeostasis. In these studies, our observations of cldn3 genes in Tetraodon, in particular, their expression profiles and their tissue and gene-specific responses to salinity variation, clearly underscore the potential importance of these TJ proteins in regulating the paracellular pathway in fishes. This provides a strong impetus for further study.
This work was supported by a National Sciences and Engineering Research Council Discovery Grants and a Canada Foundation for Innovation New Opportunities Fund to S. P. Kelly and S. I. Wright.
We thank Dr. Eric Clelland for helpful comments.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society