Using zebrafish embryos and larvae, we examined the temporal patterns of cortisol and expression of genes involved in corticosteroid synthesis and signaling. Embryonic cortisol levels decreased ∼70% from 1.5 h postfertilization (hpf) to hatch (∼42 hpf) and then increased 27-fold by 146 hpf. The mRNA abundances of steroidogenic acute regulatory protein, 11β-hydroxylase and 11β-hydroxysteroid dehydrogenase type 2, increased severalfold after hatch and preceded the rise of cortisol levels. In contrast to other teleosts that possess two glucocorticoid receptors (GRs) and one mineralocorticoid receptor (MR), only one GR and MR were identified in zebrafish, which were cloned and sequenced. GR mRNA abundance decreased from 1.5 to 25 hpf, rebounded, and then was stable from 49 to 146 hpf. MR transcripts increased continuously from 1.5 hpf and were 52-fold higher by 97 hpf. An acute cortisol response to a stressor was not detected until 97 hpf, whereas melanocortin type 2 receptor mRNA increased between 25 and 49 hpf. Collectively, the patterns of cortisol and the expression of cortisol biosynthetic genes and melanocortin type 2 receptor suggest that the corticoid stress axis in zebrafish is fully developed only after hatch. The temporal differences in GR, MR, and 11β-hydroxysteroid dehydrogenase type 2 gene expression lead us to propose a key role for MR signaling by maternal cortisol during embryogenesis, whereas cortisol secretion after hatch may be regulating GR expression and signaling in zebrafish.
- glucocorticoid receptor
- mineralocorticoid receptor
- steroidogenic acute regulatory protein
- 11β-hydroxysteroid dehydrogenase type 2
- melanocortin type 2 receptor
- genome duplication
- Danio rerio
cortisol is the main corticosteroid hormone in teleosts secreted in response to stressor exposure and plays a key role in stress adaptation (28, 45). The stressor-induced elevation of plasma cortisol is mediated via activation of melanocortin type 2 receptor (MC2R) by ACTH binding (12). Although several other hormones have been implicated in cortisol secretion, ACTH is the major secretagogue (28, 45). The molecular regulation of the steroidogenic pathway in nonmammalian vertebrates is unclear; however, the steroidogenic acute regulatory protein (StAR), which shuttles cholesterol from the outer to the inner mitochondrial membrane, is thought to be a key rate-limiting step in steroid synthesis (35). Indeed, mRNA abundances of StAR and 11β-hydroxylase, the final step in cortisol synthesis, increase in response to acute stressor exposure in trout (1).
Corticosteroid signaling is mediated by the glucocorticoid receptor (GR) and mineralocorticoid receptor (MR), which are ligand-activated transcription factors. GRs and MRs possess the four functional domains common to nuclear hormone receptors: A/B, C (DNA-binding), D (hinge region), and E (ligand-binding) domains. The C and E domains are highly conserved across vertebrates (8). In teleosts, GR and its ligand cortisol are best known for their roles in the stress response (10, 45) and, specifically, modulate aspects of intermediary metabolism, growth, behavior, and immune function (10, 28). The roles of MR and its ligand are less clear. In mammals, the mineralocorticoid system controls salt and water balance via MR and its ligand aldosterone (30). Although cortisol is a high-affinity ligand for MR, this steroid is inactivated in MR-specific tissues by the enzyme 11β-hydroxysteroid dehydrogenase type 2 (11βHSD2), allowing aldosterone binding to this receptor (18). Although teleosts lack the capacity to synthesize aldosterone (24), a recent study raises the possibility that 11-deoxycorticosterone (DOC) may be the MR ligand in rainbow trout (39).
Although most vertebrates possess one GR, the teleosts that have been examined to date [e.g., rainbow trout (Oncorhynchus mykiss) (8) and a cichlid (Haplochromis burtoni) (20)] and two species of puffer fish [Tetraodon nigroviridis and Takifugu rubripes (36)] have two GRs. Two GRs are thought to have arisen from the whole genome duplication that occurred in ray-finned fish ∼350 million years ago (27, 43). After a genome duplication event, the majority of the new paralogs undergo nonfunctionalization, where one of the two duplicated genes is lost (46). However, in some instances, both genes are retained; it is estimated that ≥20% of the genes in zebrafish exist as paralogs (7). In these cases, the duplicated genes can acquire new functions, or the original functions of the single precursor gene are divided between the two genes (7).
The development of the corticoid system during the early life stages of vertebrates is not well understood. Specifically, little is known about the timing of the activation of cortisol synthesis or the molecular mechanisms involved. We tested the hypothesis that the molecular events underlying the emergence of the corticosteroid signaling pathway are tightly linked to the steroid production capacity in teleosts. Therefore, we quantified cortisol levels and the temporal expression of genes required for corticosteroid synthesis (StAR, 11β-hydroxylase, and MC2R), cortisol metabolism or inactivation (11βHSD2), and corticosteroid signaling (GR and MR). We cloned and sequenced a full-length GR and MR in zebrafish and demonstrate for the first time that this species has only one isoform of the GR, in contrast to other teleosts examined to date. Also, we subjected zebrafish embryos and larvae to a handling stressor to determine the timing of the activation of the corticosteroid stress axis.
MATERIALS AND METHODS
Adult zebrafish (Danio rerio; 0.6–0.9 g) were purchased from a commercial supplier and maintained in a recirculating system (Aquatic Habitats, Apopka, FL) at 28°C in well water (hardness 400 mg CaCO3/l, pH 8.0) in a 12:12-h light-dark photoperiod. Adults were fed two or three times daily a mixture of tropical fish flakes and bloodworms. All experiments with live animals were approved by the University of Waterloo Animal Care Committee and conformed to the guidelines of the Canadian Council on Animal Care.
Breeding and embryo maintenance.
Fertilized eggs were acquired by breeding 12 adult zebrafish (6 male and 6 female) in 30-liter aquaria at 28°C equipped with aeration and a charcoal filter. Before the end of the light period, an egg-capturing tray, which consisted of a plastic container 5 cm high × 25 cm wide × 35 cm long, was placed in each tank. The lid was replaced with plastic mesh, which allowed the eggs to fall through while preventing fish from consuming the eggs. Plastic plants were glued to the mesh. In the morning, 1 h after the commencement of the light period, the trays were removed from the tanks and the eggs were transferred to 120-ml beakers containing 60 ml of water at 28.5°C (30 eggs/beaker). At 36 h postfertilization (hpf) and 96 hpf, 95% of the water was replaced in each beaker. At 97 hpf, 40 larvae were transferred to 250-ml beakers with 100 ml of water and fed tropical fish flake food that was ground to a powder with a mortar and pestle. Larvae were fed four times per day, and water was changed daily until fish had reached 146 hpf (time of last sample).
Embryos and larvae were sampled at 1.5, 8, 25, 49, 73, 97, and 146 hpf (n = 5 independent samples with pooled individuals; see below) for cortisol and temporal gene expression measurements. Pools of 25 fish were used for cortisol measurements, whereas gene transcripts were quantified in pools of 100 embryos at 1.5 h, down to 15 larvae at 146 h. Fish were collected, immediately frozen on dry ice, and subsequently stored at −80°C. In all experiments, each n represents eggs collected from a different batch of fish.
At 25, 49, 73, and 97 hpf, pools of 25 zebrafish (n = 3 independent samples per treatment per time point) were frozen on dry ice or subjected to a stressor regimen consisting of swirling fish in a 20-ml glass vial containing 5 ml of water for 30 s. After the stress, vials were placed in a 28.5°C incubator for 5 min, and then the fish were frozen on dry ice and, subsequently, stored at −80°C for cortisol analysis. The sampling time was based on previous work that showed consistent elevation of cortisol levels 5 min after a physical stressor exposure in carp larvae (38).
Cortisol was extracted by partial thawing of samples (pools of 25 fish) on ice and homogenization in 800 μl of ice-cold phosphate-buffered saline (pH 7.4) for 30 s with a rotor-stator homogenizer. Cortisol was extracted from 750 μl of homogenate three times with 5 ml of diethyl ether. The ether was evaporated by placement of tubes in a 45°C water bath for 1 h. The tubes were allowed to air dry for an additional 3 h before reconstitution (see below).
Cortisol was quantified using a colorimetric 96-well enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI). Evaporated samples were reconstituted in 750 μl of enzyme immunoassay buffer and kept at 4°C for 12–24 h, with occasional vortexing before use directly in the assay. Samples were tested in duplicate. Cortisol was normalized to the protein content of the homogenates, which was determined with a bicinchoninic acid protein assay (Pierce, Rockford, IL), with bovine serum albumin as the standard. The detection limit was 12 pg cortisol/ml. Cortisol extraction efficiency was ∼75%, which was similar to that reported previously (32).
Cloning and sequencing of GR and MR.
In the zebrafish genomic database, we identified only one GR, which was homologous to rainbow trout GR-2 and Haplochromis burtoni GR-1 (which is misnamed; see below). Attempts to identify the trout GR-1 and H. burtoni GR-2 homolog in zebrafish were subsequently performed using RT-PCR with primers designed in two areas of the receptor (supplemental data for this article are available online at the American Journal of Physiology-Regulatory, Integrative, and Comparative Physiology website). First, trout GR-1 (12 bp) and H. burtoni GR-2b (27 bp) possess inserts in their DNA-binding domains that are not found in the duplicate GR (8, 20). Therefore, primers were constructed to overlap different portions of this insert. Second, sequence alignments (ClustalW) of the trout and H. burtoni GR revealed areas in the A/B domain that differed between GR-1 and GR-2 in both species. Primers were then designed in these areas. All the primers were paired with others that were constructed in the highly conserved C and E domains. Primers were tested in multiple combinations and with cDNA from at least two adult tissues (ovary, gastrointestinal tract, liver, and gill) and one early life stage (28 or 52 hpf). However, RT-PCR primarily failed to amplify a product. If a product was amplified, in every instance the resulting sequence was identical to the GR that had already been identified.
The adult zebrafish ovary and gill were used to amplify and sequence cDNA transcripts for GR and MR, respectively. Fish were euthanized with an overdose of MS-222 (0.25 g/l) followed by spinal severance. Tissues were excised, immediately frozen in 1.5-ml Eppendorf tubes on dry ice, and stored at −80°C. Total RNA was extracted with QIAzol lysis reagent (phenol and guanidine thiocyanate) and purified with an RNeasy Mini Kit (Qiagen, Mississauga, ON, Canada). Total RNA was treated with DNase (Qiagen) to remove genomic DNA and quantified at 260/280 nm using a Nanodrop spectrophotometer (Wilmington, DE).
Rapid amplification of cDNA ends (RACE) PCR was performed with a SMART RACE cDNA amplification kit (Clontech, Mountain View, CA). Primers were designed with Primer3 software (Table 1). RACE PCR products were sequenced at the York University Core Molecular Biology and DNA Sequencing Facility (Toronto, ON, Canada).
GRs and MRs of zebrafish were compared with those of other species through alignments of the deduced amino acid sequences with use of the ClustalW algorithm. A phylogenetic tree was also constructed with PHYLIP version 3.66 using an alignment (ClustalW) of the coding domain sequences of GRs and MRs, with the Homo sapiens androgen receptor as the outgroup. Trees were constructed using the maximum-likelihood method and bootstrap algorithms, with the human androgen receptor as the outgroup. The solidity of the nodes of the tree was determined by the bootstrap algorithm with 1,000 simulations.
Quantitative real-time PCR.
The mRNA abundance of GR, MR, StAR, 11β-hydroxylase, 11βHSD2, MC2R, and β-actin was examined in embryos and larvae using quantitative PCR. RNA was extracted as described above for RACE PCR. First-strand cDNA was synthesized using a commercial kit (MBI Fermentas, Burlington, ON, Canada), where 1 μg of total RNA was reverse transcribed in 20 μl using Moloney murine leukemia virus reverse transcriptase (40 U), oligo(dT)18 primers (0.5 μg), dNTPs (1 mM each), and an RNase inhibitor (20 U) in a total volume of 20 μl.
A relative standard curve was constructed for each gene with use of plasmid vectors with inserted target sequences. The standard curve and samples were quantified according to previous studies (1). Briefly, 1 μl of cDNA or diluted plasmid was used as the template in 25-μl reactions. Each sample was run in triplicate with Platinum Quantitative PCR SuperMix-uracil DNA glycosylase (Invitrogen, Carlsbad, CA). Every 25-μl reaction contained 1.5 U of Platinum Taq DNA polymerase, 20 mM Tris·HCl (pH 8.4), 50 mM KCl, 3 mM MgCl2, 200 μM dGTP, 200 μM dATP, 200 μM dCTP, 400 μM dUTP, and 1 U of uracil DNA glycosylase. The reaction also contained 0.2 μM forward and reverse primers (Table 1), fluorescein calibration dye (1:2,000 dilution; Bio-Rad), and SYBR Green I nucleic acid gel stain (1:100,000 dilution; Roche, Laval, PQ, Canada). The PCR was as follows: amplification at 95°C for 3 min and 40 cycles at 95°C for 30 s for denaturation, annealing at 60°C for 30 s, and extension at 72°C for 30 s. The transcript abundance was obtained from their respective standard curves and normalized to β-actin expression. The threshold cycle (Ct) for β-actin was similar across all time points and treatments, similar to that reported by Sawyer et al. (34), and, therefore, was used for normalization.
Temporal patterns of cortisol and gene expression were initially screened for normality and homogeneity of variance before a one-way ANOVA. Then Tukey's honestly significant difference test for multiple comparisons was used to determine significant differences among groups (SPSS). Data that did not meet the assumptions of normality and homogeneity of variance were logarithmically transformed, and, if logarithmic transformation did not normalize the data, nonparametric tests (Kruskal-Wallis H-test followed by Mann-Whitney U test) were performed. Cortisol levels in control and stressed conditions were compared using a Student's t-test. Differences were considered significant if P < 0.05.
Eggs sampled at 1.5 hpf contained 0.24 pg cortisol/μg protein, or 4.0 pg/egg, which represents the maternal deposition of cortisol into the yolk. By 25 hpf, cortisol had decreased 70% and remained at this level (0.07 pg cortisol/μg protein or 1.1 pg/larva) until immediately after hatch (i.e., 49 hpf). Then cortisol levels increased and were 27-fold higher (1.87 pg cortisol/μg protein or 21.5 pg/larva) by 146 hpf (Fig. 1).
The transcript abundance of StAR (Fig. 2A) and 11β-hydroxylase (Fig. 2B) was higher after than before hatch. Specifically, StAR mRNA abundance among the prehatch times was low and not different over the first 25 hpf. The mRNA abundance increased severalfold immediately after hatch and reached maximal levels at 73 hpf; then the level dropped and was significantly lower at 146 than 73 hpf (Fig. 2A). 11β-Hydroxylase showed a transient elevation before hatch (i.e., 8 hpf) and then dropped by 25 hpf. The levels increased severalfold by 49 hpf, and this level was maintained until 146 hpf (Fig. 2B).
11βHSD2 gene expression.
The mRNA abundance of 11βHSD2 was low before hatch but increased after hatch (73 hpf) to reach a maximum at 97 hpf. Beyond 97 hpf, transcript levels dropped and were significantly lower after feeding (146 hpf) that at 97 hpf (Fig. 3).
Cortisol response to stress and MC2R expression.
Exposure to a physical stressor did not elevate whole animal cortisol above control levels at 25, 49, or 73 hpf (Fig. 4A). However, cortisol levels increased 63% after stressor exposure compared with the control levels at 97 hpf (Fig. 4A). The MC2R transcript levels were elevated after hatch compared with before hatch (Fig. 4B). Specifically, the mRNA abundance at 49 and 97 hpf, but not at 73 hpf, was significantly higher than at 25 hpf (Fig. 4B).
GR and MR cDNA sequences and analysis.
Only one GR was identified in the zebrafish genomic database (ENSDARG 00000025032). BLASTn searches of the database with use of the conserved C and E domains of GR or the GR-1 and GR-2 sequences of other fish species consistently resulted in hits for a single GR in zebrafish. For example, a BLASTn search with the zebrafish GR C domain generated hits on chromosome 14 (GR), chromosome 1 (MR), chromosome 24 (progesterone receptor), and chromosome 5 (androgen receptor). A BLASTn search with the E domain produced hits on chromosome 14 (GR), chromosome 9 (unknown), and chromosome 5 (androgen receptor). In contrast, we were able to locate two GRs in the genomic databases of medaka [Oryzias latipes: ENSORLT00000001940 (GR-1) and ENSORLT00000007571 (GR-2)] and stickleback [Gasterosteus aculeatus: ENSGACT00000027452 (GR-1) and ENSGACT 00000024121 (GR-2)].
RACE PCR with primers designed for GR amplified a product with a predicted coding domain of 2,241 bp or 746 amino acids (EF567112; see supplemental data) based on alignments with GR from other species and analysis using Open Reading Frame Finder software (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). The C and E domains of zebrafish GR are highly homologous to those of all vertebrate GRs (Table 2).
RACE PCR for MR amplified a product with a predicted coding domain of 2,913 bp or 970 amino acids (EF567113; see supplemental data) based on the criteria described for GR. The zebrafish MR has high homology in the C and E domains to MRs from all other species (Table 2). The putative MR gene is localized to chromosome 1 (ENSDARG00000037025). The two MR transcripts in the zebrafish genomic database (ENS DART00000053820 and ENSDART00000053818) appear to be splice variants; they are found at the same chromosomal location and, except for a nine nucleotide insertion (TGC AGA AAG) between the two zinc fingers in the C domain, similar to rainbow trout MRa and MRb (39), are identical. The zebrafish MR that was amplified and sequenced from the gill did not possess the nine-nucleotide insert.
The zebrafish GR clustered with other teleostean GRs and, specifically, within the clade of GR-2 (Fig. 5). The GRs of H. burtoni appear to have been incorrectly numbered (Fig. 5), as also noted by Stolte et al. (37). Of the species used in the analysis, the zebrafish GR is most closely related to the fathead minnow (Pimephales promelas) GR (Fig. 5). All domains, even the highly variable A/B domain, were highly homologous between these two species (Table 2).
GR and MR expression.
During embryogenesis, GR transcripts dropped (77%) from 1.5 to 25 hpf (Fig. 6A). However, posthatch GR mRNA abundance rebounded to the level at 1.5 hpf and was significantly higher than at 25 hpf, and this level was maintained until 146 hpf (Fig. 6A). MR transcripts continuously increased 52-fold between 1.5 and 97 hpf and remained at this level at 146 hpf (Fig. 6B). There appeared to be a greater number of GR than MR transcripts in embryos and larvae. For example, at 1.5 hpf, the Ct was ∼29 for MR and 22 for GR. In addition, at 146 hpf (when MR mRNA abundance was high), the Ct was ∼23 for MR and 21 for GR.
In contrast to other teleosts examined to date showing two isoforms of GR and one MR, we have identified only a single GR and MR in zebrafish. The expression patterns of these two receptors were distinct during embryogenesis. Although MR expression showed a continuous elevation during development, GR mRNA abundance was transient and followed closely the cortisol profiles seen in the embryos. Specifically, the maternal cortisol deposit is depleted during embryogenesis; then steroid biosynthesis is activated at around the time of hatch. Collectively, the cortisol profiles, along with the expression of steroidogenic genes and MC2R, suggest that the corticoid stress axis in zebrafish is fully developed only after hatch, toward the onset of exogenous feeding. We hypothesize a role for maternal cortisol in mediating MR signaling during embryogenesis, whereas GR expression and signaling may be activated after hatch in zebrafish.
Zebrafish corticoid receptors.
Previous detailed examinations of a number of fish revealed two GRs in each species, including rainbow trout (8), a cichlid (20), European sea bass [GR-1 (42) and GR-2 (44)], T. rubripes, and T. nigroviridis (36), and we have also located two GRs in the O. latipes and G. aculeatus genomic databases (see results). Two GRs are thought to have been the result of the whole genome duplication that occurred in ray-finned fish ∼350 million years ago (9, 27, 43). Indeed, the zebrafish genome shows substantial evidence of this duplication, such as two copies of numerous genes when only one is found in mammals. The discovery of seven Hox gene clusters in zebrafish, compared with four in mammals, was also one of the first pieces of evidence suggesting that genome duplication had occurred in teleosts (3). Therefore, zebrafish certainly possessed two GRs at some point in their history but appear to have lost one GR gene. However, it is possible that a second GR does exist in zebrafish but has eluded our in silico and RT-PCR-based searches. For the majority of duplicated genes, one copy is lost via nonfunctionalization mutations soon after duplication. However, it is estimated that ≥20% of zebrafish genes remain in duplicated form, whereas the rest have returned to single gene systems (31, 46). In addition, different species of fish have retained different sets of paralogs: although zebrafish and T. nigroviridis have similar numbers of duplicate genes, only 48% of those in zebrafish are also present in duplicate in T. nigroviridis, and the remaining 52% have returned to a single gene (46).
Phylogenetic analysis placed the GR of zebrafish in the GR-2 clade of teleost corticoid receptors. The rainbow trout GR-2 has a much greater transactivational sensitivity to cortisol (∼60-fold) and dexamethasone (∼10-fold) than GR-1 (8). In H. burtoni, GR-1 and GR-2 have similar transactivational sensitivities but different maximum activations (20). It will be interesting to compare the binding, activation, and functional characteristics of single vs. double GR systems in teleosts once data are acquired. No studies have examined the functional differences of GR-1 and GR-2.
Only one GR has been reported in brown trout (AY863149) and Japanese flounder (42), but on the basis of phylogenetic analysis, we predict that these fish would yield a second GR on further examination. However, only one GR is presently reported for the fathead minnow (17), a close relative of zebrafish. The fathead minnow GR sequence is very similar to that of the zebrafish, even in the A/B and D domains, which are not well conserved across vertebrates (Table 2). Therefore, the fathead minnow may also possess a single GR system. On the other hand, the common carp (Cyprinus carpio) has two GRs (37), indicating that a single GR is not a characteristic of the entire Cyprinidae family. However, this example is complicated by the recent genome duplication in carp, ∼12 million years ago (13). Therefore, thorough searches for a second GR in the fathead minnow, other cyprinids, and increasingly distant teleosts are required to determine the extent of a single GR. With this knowledge, along with functional studies of the one- and two-GR systems in teleosts, we may begin to understand why some fish have lost a GR gene.
Ontogeny of corticosteroidogenesis.
The temporal changes in cortisol levels during the early life stages of teleosts are similar across a number of species. The initial maternal deposit of cortisol in the yolk is utilized during embryogenesis and reaches its lowest concentration around the time of hatch; then the larva begins to synthesize cortisol de novo [Japanese flounder (14), tilapia (Oreochromis mossambicus) (23), rainbow trout (5), Asian sea bass (Lates calcarifer) (33), common carp (C. carpio) (38), sea bream (Sparus aurata) (40), and salmon (Salmo salar) (29)]. Cortisol levels in zebrafish also show the same pattern. In addition, the expression of StAR, 11β-hydroxylase, and MC2R is upregulated immediately before the rise in larval cortisol levels, pointing to the activation of the steroidogenic pathway at around the time of hatch. This in agreement with another study in zebrafish showing the expression of cytochrome P-450 11a1 (P450scc) in the primordial interrenal tissue at 36 hpf (21). Although genes encoding key proteins involved in steroidogenesis are detectable during embryogenesis, the upregulation of the steroid biosynthetic pathway at around the time of hatch leads us to propose that ACTH production and/or ACTH receptor (MC2R) synthesis may be a key signal for activation of corticosteroidogenesis during development.
In zebrafish, the ability to synthesize cortisol after hatch does not immediately give rise to the stress-induced stimulation of cortisol production. This lack of response may not be related to MC2R availability, as we observed an upregulation (5-fold) of MC2R transcripts in zebrafish at around the time of hatch, and this preceded the de novo cortisol synthesis. However, despite higher MC2R gene expression and basal cortisol production capacity at around the time of hatch, the delayed activation of the cortisol stress axis (just before exogenous feeding) suggests a role for the brain-pituitary axis as a key determinant in activating the stress axis during development. In support of this argument, a cortisol response similar to that of zebrafish was observed in rainbow trout larvae. Trout could synthesize cortisol at 1 wk after hatch but did not experience stress-induced increases in cortisol until 2 wk after hatch (5). However, interrenal tissue from embryonic trout (i.e., before hatch) increased cortisol synthesis when stimulated with ACTH in vitro (6). Similarly, MC2R transcripts in zebrafish increased between 25 and 49 hpf, 2 days before we detected a cortisol increase after stress. Collectively, these studies suggest that the final step in the onset of the cortisol stress response may involve activation of sensory inputs in the brain, leading to higher ACTH release.
Many of the conclusions in this study are based on measurements of mRNA abundance, with the assumption that they correspond directly to functional proteins. Although this may not always be the case, the genes involved in cortisol synthesis (StAR, 11β-hydroxylase, and MC2R) are upregulated along with cortisol levels, suggesting that, in some instances, mRNA abundance correlates well with gene function.
Role of GR and MR during development.
Most studies related to corticosteroid receptors in fish focused on cortisol stimulation of GR signaling (2, 28). The recent discoveries of MRs in fish (20, 39), along with functional studies, suggest a role for this receptor in ion regulation (19). The finding that DOC may be the ligand for MR in fish is very interesting (39), but physiological studies to confirm this finding are warranted. Little is known about GR and MR expression or function in fish development. Here we show, for the first time, that both of these receptors are present during embryogenesis but exhibit distinct expression patterns.
Specifically, the reduction in GR mRNA abundance after fertilization correlates with the decrease in maternal cortisol content. Inasmuch as zygotic transcription in zebrafish begins at 3–5 hpf (25), it is likely that the GR transcripts at 1.5 hpf in the present study are maternally derived and appear unstable. GR signaling may therefore not be critical during embryogenesis. In agreement with this idea, a recent study identified a mutant zebrafish line that does not develop corticotropic pituitary cells (15). This results in decreased larval cortisol levels; however, there is no mention of abnormal embryonic development (15). Also, mice lacking a functional GR survive until birth but die shortly thereafter due to impaired lung development (11). Collectively, these results highlight a temporal lag in GR expression, suggesting a more important role for this receptor after hatch in zebrafish.
In addition to cortisol binding to GR, this hormone is also a high-affinity ligand for MR. However, the enzyme 11βHSD2 metabolizes cortisol to cortisone, thereby eliminating receptor binding (9). The coexpression of 11βHSD2 with MR allows ligands other than cortisol, such as aldosterone in tetrapods, to bind to MR (18). The expression of 11βHSD2 in zebrafish increased just before the rise of larval cortisol. This suggests that, in tissues that express both 11βHSD2 and MR, a ligand other than cortisol, perhaps DOC (39), may be responsible for MR signaling after hatch. However, the situation may be different during embryogenesis (1.5–25 hpf). Because there were very few 11βHSD2 transcripts during this time, cortisol would have access to MR, as well as GR. We hypothesize that MR activation by maternal cortisol may be a key signal during embryogenesis, since transcripts for this receptor increased before hatch, while GR transcripts decreased over the same time period.
Perspectives and Significance
The identification of only one GR in zebrafish sparks a number of physiological and evolutionary questions. For example, from a functional standpoint how could zebrafish (or its ancestor) afford to lose one GR? Also, what exactly are the individual functions of the two GRs in other fish? What were the evolutionary or environmental pressures that favored a return to a single-GR system, and what is the extent of a single GR among fish? This finding also has broad implications. First, a single-GR gene model allows this species to be a much simpler and powerful tool for gene silencing (e.g., morpholino and small interfering RNA) studies to characterize the roles of GR and MR in fish development and physiology. Second, zebrafish are a popular model species used to study aspects of development, disease, drug discovery, and toxicology in mammals (16, 22, 26). A single GR and MR give zebrafish a corticoid-signaling pathway similar to that of mammals, lending strong support for its suitability as a model organism in studies involving the corticosteroid axis (4). Our examination of the corticosteroid axis during zebrafish development has shown distinct expression patterns for GR and MR and, along with the temporal expression of 11βHSD2, lead us to propose a key role for MR signaling by maternal cortisol during embryogenesis. In contrast, GR and cortisol upregulation after hatch may be a key signal for metabolic adjustments to exogenous feeding, as well as stressor challenges. Future studies using in situ hybridization may help elucidate the spatial and tissue-specific distribution of GR, MR, and 11βHSD2, whereas morpholino experiments may reveal the specific functions of the receptors during the early life stages of fish.
This study was funded by the Natural Sciences and Engineering Research Council of Canada Discovery Grant to M. M. Vijayan and an E. B. Eastburn Fellowship to D. Alsop.
We thank Andrew Doxey and Brendan Knight (University of Waterloo) for helpful discussions on genome/phylogenetic analyses and breeding zebrafish, respectively.
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