Environ. Sci. Technol. 2007, 41, 6305-6310
Estrogen-Induced Alterations in amh and dmrt1 Expression Signal for Disruption in Male Sexual Development in the Zebrafish RU ¨ DIGER W. SCHULZ,† JAN BOGERD,† RUNE MALE,‡ JONATHAN BALL,§ M A R T I N A F E N S K E , §,⊥ LISBETH C. OLSEN,‡ AND C H A R L E S R . T Y L E R * ,§ Laboratory for Endocrinology, Department of Biology, Faculty of Science, University of Utrecht, Padualaan 8, NL-3584 CH Utrecht, The Netherlands, Department of Molecular Biology, University of Bergen, HiB, Thormoehlensgt. 55, N-5020 Bergen, Norway, and Environmental and Molecular Fish Biology Group, School of Biosciences, Hatherly Laboratories, University of Exeter, Prince of Wales Road, Exeter, Devon, EX4 4PS, UK
Dmrt1 and amh are genes involved in vertebrate sex differentiation. In this study, we cloned dmrt1 and amh cDNAs in zebrafish (Danio rerio) and investigated the effects of exposure to 17R-ethinylestradiol (EE2), during early life on their patterns of expression and impact on the subsequent gonadal phenotype. Expression of both amh and dmrt1 in embryos was detected as early as at 1 day post fertilization (dpf) and enhanced expression of amh from 25 dpf was associated with the period of early gonadal differentiation. Sex-dependent differences in enhanced green fluorescent protein transgene expression driven by the promoter of the germ cell-specific vas gene were exploited to show that at 28dpf and 56dpf both amh and dmrt1 mRNA were overexpressed in males compared with females. Exposure during early life to environmentally relevant concentrations of EE2 had a suppressive effect on the expression of both amh and dmrt1 mRNAs and this was associated with a cessation/retardation in male gonadal sex development. Our findings indicate that estrogen-induced suppression in expression of dmrt1 and amh during early life correlate with subsequent disruptive effects on the sexual phenotype in males.
Introduction
Materials and Methods
In mammals, the Y-linked sex-determining gene, Sry, governs the development of the male phenotype (1) by activating the expression of Sox9 (2), and this factor (with other transcriptional regulators) triggers male-specific differentiation of the reproductive system (3, 4). A suite of genes expressed downstream of Sry in mammals is conserved in the differentiating testis in different vertebrate * Corresponding author phone: 00 44 1392 264450; e-mail
[email protected]. † University of Utrecht. ‡ University of Bergen. § University of Exeter. ⊥ Current address: Syngenta, Jealott’s Hill International Research Centre, Berkshire, UK. 10.1021/es070785+ CCC: $37.00 Published on Web 07/26/2007
groups (e.g., refs 5, 6) although the sequence of molecular events leading to male sex differentiation can differ (5, 7). Male-specific differentiation in mammals includes upregulation of the expression of anti-Mu ¨ llerian hormone (AMH, a member of the transforming growth factor β family, produced by Sertoli cells) (8) during embryogenesis and even though teleost fish do not have a Mu ¨ llerian duct, an amh orthologue has been identified in various fish species (e.g., refs 9-12), including in zebrafish (13, 14). Interestingly, amh shows a male-biased pattern of expression during sex differentiation in most of these fishes (except for medaka, Oryzias latipes; 12), suggesting that it has a function during testis differentiation. Dmrt1 is also expressed at high levels during testicular differentiation and is either not expressed or down regulated during ovarian differentiation in vertebrates, including in fish (4, 15). The significance of dmrt1 for sex differentiation, has been re-enforced by the identification of dmy, also referred to as dmrt1b(Y), as a gene duplicate of the autosomal dmrt1 on the Y chromosome in the medaka and the first sex-determining gene in a nonmammalian vertebrate, expressed in preSertoli and Sertoli cells (16, 17). Outside the genus Oryzias, however, dmy/dmrt1b(Y) has not been identified (18), suggesting that dmy/dmrt1b(Y) has evolved only recently as a sex-determining gene. Fish show a plasticity of germ cell differentiation along male or female developmental pathways and environmental factors, e.g., temperature (19) and exposure to hormones and their environmental mimics, can play decisive roles in the sex differentiation process. Disruptions in gonadal development including the simultaneous presence of both male and female germ cells in the same gonad (20) are widely reported and these responses can be induced in fish through controlled exposures to steroidal estrogens including the pharmaceutical 17R-ethinylestradiol (EE2, e.g., for zebrafish, refs 21, 22). In this study, we investigated the effects of exposure to EE2 during the period of sex differentiation on amh and dmrt1 gene expression in juvenile zebrafish and assessed the subsequent impacts on the gonadal sex phenotype. EE2 was chosen as the test chemical because it is an extremely potent estrogen and is present in wastewater treatment work (WwTW) effluent discharges at concentrations known to cause feminized responses in fish, including in zebrafish (21, 22). The zebrafish is especially interesting for the study of endocrine-disrupting chemicals on sexual development because there are no sex chromosomes and no evidence for a simple (monogenic) sex determination mechanism; morphologically, testis differentiation takes place via a juvenile protogynous stage (23).
2007 American Chemical Society
Fish and their Maintenance. Zebrafish of the AB strain were used for the cloning experiments, and zebrafish of the wild type Kalkutta (WIK) strain were used for the estrogen exposure study; both strains were obtained from the Max-Plank Institute in Tu ¨ bingen. Viable embryos were collected from the breeding tank 2 h post fertilization (hpf). Transgenic zebrafish expressing enhanced green fluorescent protein (eGFP) under the control of the germ cellspecific vas promoter have been described previously (24). Animal culture and experimentation using zebrafish was consistent with the respective national British, Dutch, and Norwegian regulations, and the protocols used were approved by the local (Exeter, Utrecht, Bergen) Ethics and Animal Care and Use Committees. VOL. 41, NO. 17, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6305
Molecular Cloning of amh and dmrt1. The procedures for the cloning and sequence analysis of zebrafish amh and dmrt1 cDNAs and the results obtained are given in the Supporting Information. Expression Analysis of amh and dmrt1 mRNA. Northern blot analysis served to determine the size of the amh transcript(s) from pooled, adult testis and ovaries, as detailed in the Supporting Information. To determine the temporal patterns of expression of amh and dmrt1 during early life, RT-PCR analysis was performed on embryos through to fish at 1-35 days post hatch. RNA was isolated from pooled samples of whole-body extracts from embryos and juveniles and from dissected adult testes and ovaries using the Trizol reagent (Invitrogen). Reverse transcription reactions were primed with decamer primer (Ambion, Austin, TX; http://www.ambion.com/). Single stranded cDNA was synthesized from total RNA (1 µg) using Superscript II reverse transcriptase (Invitrogen) according to the manufacturer’s protocol. Primers and the PCR conditions are detailed in the Supporting Information. Real-time, quantitative (rtq) PCR was employed to quantify estrogen effects on amh and dmrt1 gene expression (see below) and to quantify expression of these genes in zebrafish that have been sorted into males and females based on the gender-related difference in eGFP expression under the control of the vas promoter (25). Primers, probes (Tab. S1), and other experimental conditions are detailed in the Supporting Information. Whole mount in situ hybridization was employed to investigate the gonadal tissue localization of amh and the approach and results are described in the Supporting Information. Ethinylestradiol (EE2) Exposure. Three hundred zebrafish embryos (∼2 hpf) were divided in two identical flow-through 18 L tanks and exposed immediately to one of three nominal concentrations (0.05, 0.5, and 5 ng/L) of EE2 in ethanol (SigmaAldrich, St. Louis, MO; http://www.sigmaaldrich.com/) for a period of up to 56 days. The tank dosing flow rate was 100 mL per 24 h, in combination with 10 L of clean water, and EE2 dosing bottles were renewed every 7 days. Duplicate dilution water and solvent control tanks were run under the same conditions without the addition of EE2. Water flow rates and EE2 dosing rates were monitored daily during the course of the exposure. Fish were sampled at 28 and 56 dpf. Whole bodies of zebrafish were fixed in Bouin’s solution (0.9% picric acid, 9% formaldehyde, 5% acetic acid), embedded in wax, and sectioned to 5 µm. Sections were stained with haematoxylin and eosin, before capturing digital images. The stages of sexual development were characterized in females and males by the advancement of the germ cells. Vitellogenin (VTG) was quantified in whole body homogenates of fish at 28 dpf and 56 dpf from all treatment groups using a competitive homologous ELISA for zebrafish vtg, adapted from a previously published protocol (26; see the Supporting Information for details about the assay procedure). The results are expressed as nanograms of VTG per gram fish body weight. Statistical Analysis. Expression of amh and dmrt1 mRNAs in samples collected from transgenic zebrafish at 28 and 56 dpf and sorted according to their eGFP expression level were compared by Student’s t test using the GraphPad Prism4 software package (GraphPad Software, San Diego, CA; http:// www.graphpad.com/). For the EE2 exposure study morphometric measurements were compared using a one way analysis of variance (ANOVA) and F test. Analysis of whole body VTG employed ANOVA, followed by Fisher’s partial least-squares difference test. Significant differences in gene expression (amh and dmrt1 mRNAs) between the EE2 exposure groups and controls were tested by ANOVA. Individual treatment groups were com6306
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 17, 2007
pared to the controls employing Dunnett’s multiple comparison test. Some individual treatment groups were additionally tested against the control using Student’s t test. Data were log transformed to provide for normal distribution and equal variance, prior to the analysis, using SigmaStat version 2.03 (Systat Software Inc, Point Richmond, CA; http:// www.systat.com/).
Results cDNA Cloning and Transcript Size. BLAST analysis of the isolated zebrafish amh cDNA (accession number AY 721604) predicted amino acid sequence revealed the highest similarity with other known AMH proteins (Figure S1). It also had strong similarities with TGFβ, BMP, and other sequences containing a conserved C-terminal cysteine knot-like domain, classified as a TGFβ domain and found in bilateral organisms. AMH sequences grouped together and were vertebrate-specific. Outside the C-terminal cysteine knot-like domain, there was little conservation. The zebrafish amh mRNA covers 3234 nt plus a poly A tail. Northern hybridization using a randomly labeled probe representing the zebrafish amh ORF revealed one amh transcript of 3.7 kb in the testis (Figure S3), while no apparent signal could be detected in the ovary. The zebrafish dmrt1 cDNA (accession number AF 439562) encoded a 267 amino acid protein. Several cDNA variants of dmrt1 have now been reported in the database, but they all appear to be derived from a single gene (27). BLAST analysis showed that the characteristic DM domain was highly conserved, and that there was little similarity between different subfamilies outside this domain, and phylogenetic analysis showed that dmrt1 is restricted to vertebrates (Figure S2). Information on dmrt1 transcript sizes has been published recently (27). Temporal and Spatial amh and dmrt1 Expression. amh mRNA was detectable in embryos at 1 dpf, and low-level expression was observed throughout 12-20 dpf (Figure 1A). There was an apparent decrease in amh expression at 20dpf, but the level of β actin, which served as a control for cDNA synthesis and loading control was also lower in this sample. There was a clear up-regulation of amh expression during early gonadal differentiation (from day 25 pf). Amh expression was detected in both adult testes and ovaries and was localized to Sertoli cells in the testis, which concurs with previous reports in zebrafish (13, 14). However, not all Sertoli cells expressed amh mRNA at levels readily detectable by in situ hybridization (Figure S4). Using RT-PCR (Figure 1A), we detected low-level expression of dmrt1 mRNA during the early stages of ontogenesis. In contrast to amh mRNA, an up-regulation was not apparent for dmrt1 mRNA during the period of early gonadal sex differentiation (starting at 25 dpf), supporting and extending previous findings of Guo and colleagues (27) on early embryonal (1 dpf) and adult expression. Ethinylestradiol Exposure. There were no statistical differences between the dilution water controls and solvent controls for growth (weight and length), condition factor (weight/length3), VTG induction or gene expression (with the exception of a marginal difference for amh at 28 dph), and so all EE2 exposures are compared against the dilution and solvent controls combined. The length and condition factor of fish in the different EE2-exposed groups did not differ significantly from controls (data not shown). However, there were significant but small effects of the EE2 treatment on fish weight at 56 dpf (Figure S5), with a stimulatory effect at the lowest concentration (p e 0.01; 12% increase), and an inhibitory effect at the highest concentration of EE2 (p e 0.01; 23% decrease), compared with controls. EE2 exposure, up to and including the intermediate concentration (nominal 0.5 ng EE2/L) did not induce a
FIGURE 1. (A) Expression of amh (upper panel), dmrt1 (middle panel), and an actin fragment (lower panel) analyzed by RT-PCR of RNA isolated from fish throughout days 1-35, adult testes (T) and adult ovaries (O). (B) Expression of amh and dmrt1 mRNA in male and female zebrafish as separated for low (male) and high (female) eGFP expression at 28 and 56 dpf. Significant differences between sexes were assessed by two-tailed Student t test (*, p < 0.05; **, p < 0.005; n ) 4-6 per group). vitellogenin response at either 28 dpf or 56 dpf. There was, however, a significant induction of VTG in zebrafish exposed to the highest EE2 concentration (nominal 5 ng EE2/L) at both 28 and 56 dpf compared with controls (p e 0.01; 640fold and 24 000-fold induction, respectively; Figure S6). In controls at 28 dpf all fish contained either a small group of primary oocytes, or undifferentiated, premeiotic germ cells only (Figure 2A, B). There were no differences between the stages of development across the different EE2 treatment groups and the controls (Figure S7). At 56 dpf, the control population contained sexually differentiated females (Figure 2C), and fish undergoing gonadal transformation (viz. degrading primary oocytes). Control populations also contained sexually differentiated males, characterized by gonads without primary oocytes but with spermatocytes (Figure 2D; Figure S7). At 56 dpf, stages of gonadal development in the two lower concentrations of EE2 did not differ from the controls. The stages of ovarian development in these treatment groups ranged from those containing primary oocytes only to ovaries containing oocytes at the cortical alveolus stage (Figure 2C). In the 0.5 ng EE2/L treatment group 33% (n ) 8 out of 24) of the differentiated males had a feminized reproductive duct (an ovarian cavity; Figure 2E). In the 5 ng EE2/L treatment group (at 56 dpf) testis development was strongly inhibited and the male gonad had not developed beyond that seen at 28 dpf (Figure 2F). No sexually differentiated males were found in this EE2 treatment group at 56 dpf (n ) 12) and only one “male” fish was undergoing (partial) transformation (Figure S7). Furthermore, the gonads were smaller compared with contemporaries in the other EE2 treatments and controls. Expression of both genes, amh and dmrt1, was highly variable at 28 dpf and 56 dpf in the controls (and for some of the low-dose EE2 treatment groups) (Figure 3), and there appeared to be two subpopulations of fish in these groups as defined by bimodal expression patterns of both amh and dmrt1. To test if this was gender related, we analyzed individual fish sorted into presumptive males and females
based on expressing eGFP under the control of the vas promoter. At 28 dpf, expression of amh and dmrt1 in presumptive males was 14-fold and 2-fold higher, respectively, than in females (Figure 1B), confirming sex-dependent differences in the levels of the expression of these transcripts. The male-biased overexpression was more pronounced at 56 dpf (Figure 1B), mounting up to 52- and 22-fold higher expression levels in males than in females for amh and dmrt1, respectively. The high variability in expression of amh and dmrt1 (likely gender-related) in our exposure study, associated with an inability to sex the same fish (RNA preparations for amh and dmrt1 mRNA quantification were derived from whole body extracts, and animals expressing eGFP under the control of the vas promoter were not available for the EE2 exposure study) limited our ability to distinguish any subtle effects of EE2 on gene expression. Nevertheless, concentrations as low as 0.5 ng EE2/L were found to affect the expression of both genes (Figure 3). At 28 dpf, the intermediate concentration augmented the level of dmrt1 mRNA, with a suggestion for a similar effect on amh mRNA, but this was not statistically significant. In contrast, at the highest EE2 exposure concentration both mRNAs showed a significant decrease in expression compared with the control, together with a diminished variation in expression between individuals (especially for amh mRNA). At 56 dpf, exposure to the lowest concentration, EE2 induced an elevation in the expression dmrt1 mRNA. At this later life stage, exposure to both the intermediate and highest concentrations of EE2 induced significant down-regulations and a concomitant diminished variation in expression between individuals within the treatment groups for both amh and dmrt1 mRNAs (Figure 3).
Discussion We cloned amh and dmrt1 in zebrafish to provide us with the means to investigate their roles in gender assignment and assess the effects of altered expression during early life VOL. 41, NO. 17, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6307
FIGURE 2. Stages of gonadal development in control and EE2exposed zebrafish at 28 dpf and 56 dpf. At 28 dpf, all fish contained either undifferentiated, premeiotic germ cells only (A) or a small group of primary oocytes (B), and there were no apparent effects of estrogen exposure. At 56 dpf, the control group comprised of sexually differentiated females with ovaries containing oocytes ranging in development from primary oocytes to cortical alveolus stage (C), transforming males (not shown), and sexually differentiated males (D). At 56 dpf, some males exposed to 0.5 ng EE2/L had an ovarian cavity and showed premeiotic germ cells only (E). At 56 dpf, development of the “male” gonad in the 5.0 ng EE2/L exposure group (F) had not progressed beyond that seen in fish at 28 dpf and there were no sexually differentiated males in this treatment group. Gonads in the 5.0 ng EE2/L treatment group were also smaller than contemporaries in the other EE2 treatment and the controls. Ca, primary oocyte in cortical alveoli stage; Gc, germ cell; Oc, ovarian cavity; Po, primary oocyte in perinucleolar stage; Lv, liver; S1, spermatogonia; S2 , spermatocytes; S3, spermatids. (induced by oestrogen exposure) on the subsequent gonadal phenotype. During the course of this work, papers appeared describing the cloning of zebrafish amh (13, 14) and dmrt1 (27), and a discussion of the structural features of the protein and nucleotide sequences can be found in the Supporting Information. Northern hybridization (Figure 1A) demonstrated a single transcript only for amh in the testis. In situ hybridizations (Figure S4) localized amh mRNA to Sertoli cells in the adult testis, consistent with other teleost species (10, 12), and higher vertebrates (28). A discrepancy in the Sertoli cell numbers identified morphologically and by in situ hybridization using amh mRNA detection may reflect that only Sertoli cells in contact with germ cells at early stages of spermatogenesis express detectable levels of amh mRNA (using in situ techniques) as has been shown to occur in mammals (28). The increase in expression of amh from day 25 dpf, is in 6308
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 17, 2007
accordance with the period of gonadal transformation and testicular differentiation in zebrafish (29). At 28 dpf, both dmrt1 and amh expression were higher in males than in females (as determined using the vas eGFP transgenic fish), similar to the male-biased expression patterns described in rainbow trout for dmrt1 (15) and amh (11). However, the rainbow trout shows a stable, genetic sex determination mechanism, and there are large (at least 50-fold) differences in the levels of gonadal expression for these transcripts between males and females during sex differentiation and before gonads are morphologically distinct. In zebrafish, even though there are no sex chromosomes and the testis differentiation pathway is via a juvenile protogynous stage, sex related differences in expression of amh also differed considerably, and they were an order of magnitude higher in the males compared with those in females at 28 dpf (before the gonadal sex is morphologically distinguishable for males). The difference in the level of expression of dmrt1 between the sexes at 28 dpf was less pronounced compared with that for amh, but was nevertheless clearly distinguishable. This may be related to the fact that dmrt1 expression also occurs in oocytes up until the beginning of vitellogenesis (27). At 56 dpf, the male-biased overexpression of both genes appeared to have further increased, even though amh mRNA is also known to occur in granulosa cells of maturing follicles (13). amh and dmrt1 mRNA were also detected in zebrafish embryos at life stages before sexual differentiation and notably even at 1dpf (Figure 1A), but their functional role(s) at these life periods are not known. Exposure to EE2 during early life induced well characterized phenotypic effects. Vitellogenin was induced at 5 ng EE2/L (Figure S6) and an ovarian cavity was present in onethird of the sexually differentiated males exposed to 0.5 ng EE2/L (at 56 dpf, Figure 2E). These findings concur with previous studies where zebrafish (21, 30, 31) and other fish species have been exposed to estrogens during early life (32, 33). Exposure to 5 ng EE2/L during early life caused an inhibition of male zebrafish gonad development (Figure 2F). Previous studies in zebrafish (22, 31) have demonstrated that the effects of EE2 on male gonad differentiation strongly depends on the concentration, timing and duration of the exposure. Effects during early life are reversible when exposure is transient and to 3 ng EE2/L, while continuous exposure to 5ng EE2/L from embryos to sexual maturity induced reproductive failure (21). The mechanisms by which these suppressive effects of EE2 on testis development in zebrafish occur have not been established. The findings in this paper, however, have shown that both amh and dmrt1 are estrogen-sensitive in zebrafish (Figure 3), a species with no sex chromosomes and no evidence for a simple (monogenic) sex determination mechanism, and that suppressive effects of EE2 on both amh and dmrt1 expression (at 5ng EE2/L at 28 dpf, and both 0.5 and 5 ng EE2/L at 56 dpf) were associated with the retardation in male gonad development. In drawing functional associations between changes in the levels of expression of the target genes and the phenotypes seen, we have assumed that the changes in the Amh and Dmrt-1 proteins are directly proportional to the changes in the transcripts. The molecular responses occurred at a lower concentration of EE2 (10-fold lower) than that for effects seen on either VTG induction or retardation of gonadal development determined via histology, and these genes may thus serve as good biomarkers of estrogenic disruption of reproductive function in males. We measured amh and dmrt1 mRNA levels in trunk extracts, but this is considered to reflect gonadal mRNA since expression of these genes has been shown to be gonadspecific in different fish species (6, 12, 14, 34). In genetically all-male rainbow trout, dmrt1 similarly has been shown to
FIGURE 3. Expression of amh and dmrt1 mRNA in control and EE2-exposed zebrafish at 28 and 56 dpf. Data are presented as box and whisker plots: lower and upper boundary of box correspond to 25th and 75th percentiles, lower and upper whiskers to 10th and 90th percentiles, 5th and 95th percentiles are given as black circles, and the median is given as line within the box. Asterisks (*) indicate significant differences of the exposure groups from the control, tested by ANOVA (p )