Sexually Dimorphic Transcription of Estrogen Receptors in Cod Gonads Throughout a Reproductive Cycle

K NAGASAWA and others Sex-biased expression of 52:3 357–371 Research cod esr genes

Sexually dimorphic transcription of estrogen receptors in cod gonads throughout a reproductive cycle

Kazue Nagasawa1,*, Christopher Presslauer1, Lech Kirtiklis1,2, Igor Babiak1 and Jorge M O Fernandes1

1Faculty of Biosciences and Aquaculture, University of Nordland, 8049 Bodø, Norway Correspondence 2Department of Zoology, Faculty of Biology and Biotechnology, University of Warmia and Mazury in Olsztyn, should be addressed 10-718 Olsztyn, Poland to J M O Fernandes *(Present address: Laboratory of Aquacultural Biology, Graduate School of Agricultural Science, Tohoku University, Email Miyagi 981-8555, Japan) [email protected]

Abstract

The role of sex steroid regulation in gonadal maturation is a very complex process that is far Key Words from being fully understood. Hence, we have investigated seasonal changes in gonadal " androgen receptor (ar) expression of estrogen receptors (ERs) in Atlantic cod (Gadus morhua L.), a batch spawner, " Atlantic cod throughout the annual reproductive cycle. Three nuclear ER partial cDNA sequences (esr1, " cytochrome P450 aromatase esr2a,andesr2b) were cloned and all esr transcripts were detected mainly in liver and gonads (cyp19a1a) of ﬁsh of both sexes. In situ hybridization of esrs along with germ cell (vasa) and gonadal " estrogen receptor (esr) somatic cell markers (gonadal soma-derived factor (gsdf), 3b-hydroxysteroid dehydrogenase " sexual maturation (3bhsd), and anti-Mu¨ llerian hormone (amh) for testicular, or gsdf for ovarian somatic cells) showed that all three esrs were preferentially localized within interstitial ﬁbroblasts composed

Journal of Molecular Endocrinology of immature and mature Leydig cells in testis, whereas they were differentially expressed in both follicular cells and oocytes in ovary. Quantitative real-time PCR analysis revealed a sexually dimorphic expression pattern of the three esr paralogs in testis and ovary. A signiﬁcant increase in esr2a expression was identiﬁed in testis and of esr2b in ovary, whereas esr1 transcripts were elevated in both testis and ovary in February and March before the spawning period. The localization and sexually dimorphic expression of esr genes in gonads indicate a direct function of estrogen via ERs in gonadal somatic cell growth and differentiation for Leydig cell in testis and follicular cells in ovary throughout the annual Journal of Molecular reproductive cycle in Atlantic cod. Endocrinology (2014) 52, 357–371

Introduction

Estrogens, mainly 17b-estradiol (E2),actdirectlyto stimulate vitellogenin and choriogenin production during promote sex differentiation and gonadal maturation oocyte growth in vitellogenic stages for all vertebrates (Korach 1994). Estrogen action has been classically studied (Kazeto et al. 2011). Two distinct subtypes of shown to act in liver via speciﬁc nuclear estrogen receptors nuclear ERs (ERa and ERb) have been cloned in a number (ERs), which act as ligand-dependent DNA-binding trans- of organisms from humans (Keaveney et al. 1991)to cription factors (Dahlman-Wright et al. 2006) in order to nematodes (Mimoto et al. 2007). The role of ERs in gonadal

http://jme.endocrinology-journals.org Ñ 2014 Society for Endocrinology Published by Bioscientiﬁca Ltd. DOI: 10.1530/JME-13-0187 Printed in Great Britain Downloaded from Bioscientifica.com at 09/24/2021 06:26:41AM via free access Research K NAGASAWA and others Sex-biased expression of 52:3 358 cod esr genes

development has been elucidated from the phenotypic Expression of esr is first detected in teleosts at the embryonic alterations of knockout (KO) mice. KO mice lacking Esr1 stage, indicating an essential role of estrogen signaling showed complete infertility in both males and females in larval development (Lassiter et al.2002). Despite the (Lubahn et al. 1993), whereas both male and female Esr2- paramount importance of estrogen signaling mediated by KO animals developed normally, albeit with reduced ERs in fish gonadal development, information about the ovulation efficiency (Krege et al. 1998). precise localization of ERs is still limited to a few fish species: Interestingly, there have been a considerable number ERa in the interstitial fibroblasts (the Leydig cell precursors) of reports that show predominant ER transcripts in gonads in testis of rainbow trout Oncorhynchus mykiss (Bouma & rather than other organs, indicating the possible direct Nagler 2001), ERa in thecal, granulosa, and interstitial cells action of estrogen in gonadal cells in vertebrates. It is in ovary and in Sertoli cells in testis of eelpout Zoarces known that the concentration of estrogen in testis and rete viviparus (Andreassen et al.2003), and ERa in granulosa cells testis fluid far exceeds that in male serum in various species in ovary and ERb2 in cells in undifferentiated gonad in (Hess 2000), indicating a central role for estrogen may medaka Oryzias latipes (Chakraborty et al.2011a). Therefore, function in testicular and epididymal cells (O’Donnell et al. information about localization patterns and expression 2001). To obtain an overview of the localization of ER profiles for ER subtypes in teleost gonadal cells is required subtypes within gonads, O’Donnell et al. (2001) surmised in order to understand their potential role in regulating gene the likely localization of ERs in adult rodent testis from transcription and gonadal development process in teleosts. numerous items of conflicting data: Leydig cells express Atlantic cod (Gadus morhua) is a cold-water marine both ERa and ERb, whereas Sertoli cells and testicular germ fish species that has an annual reproductive cycle cells (e.g. spermatogonia, pachytene spermatocytes, and (Kjesbu 1989). Importantly, precocious sexual maturation round spermatids) predominately expressed ERb.The (also known as early puberty) often occurs in both sexes direct effects of estrogen on gonad development have in Atlantic cod when they are 2 years old under in been observed following administration of diethylstilbes- aquaculture conditions. This is an important bottleneck trol (DES) to rats. DES exposure revealed that estrogen plays for the cod farming industry, as it is associated with a an inhibitory role in Leydig cell proliferation (Abney & significant reduction in somatic growth (Karlsen et al. Myers 1991) and Sertoli cell maturation (Sharpe et al. 1998). 2006). Vitellogenesis generally starts in the autumn and In addition, there are several lines of evidence showing that subsequent reproduction occurs from February to April estrogen has a stimulatory effect on germ cells. In along the Norwegian coastline. During the 50–60-day particular, a report in which neonatal rats were treated spawning period, females can release up to 17–19 batches Journal of Molecular Endocrinology with estrogen displayed increased numbers of undiffer- of eggs at intervals of several days (Kjesbu 1994). The entiated and differentiating type A spermatogonia (Kula timing of sexual maturation and the spawning period are 1988). Within rodent ovaries, both ERa and ERb are strongly associated with a peak of the plasma E levels in expressed in granulosa cells in follicles (Drummond et al. 2 1999). Administration of estrogen to hypophysectomized 2- and 3-year-old Atlantic cod under aquaculture con- rats resulted in a stimulatory effect on granulosa cell growth ditions (Norberg et al. 2004). However, molecular and differentiation (Drummond & Findlay 1999). In mechanisms underlying the direct effects of E2 signaling addition, androgen receptor (Ar)KOmicethatlack on gonadal maturation and batch spawning have been functional aromatase (Britt et al. 2001) – an alternative poorly investigated, even if the final maturation process model of ER depletion – exhibit abnormal folliculogenesis by maturation-inducing steroids has been studied ex vivo with seminiferous tube-like structures due to apoptotic (Tveiten et al. 2010). To deepen our understanding of granulosa cells (Drummond & Findlay 1999, Drummond the role played by estrogen regulation in gonadal et al. 2002). Moreover, an in vitro culture study of mouse maturation, we kept 2-year-old Atlantic cod in land- ovarian follicles reported that ERb, but not ERa, plays a based tanks and examined their annual reproductive direct role in folliculogenesis to facilitate follicle matu- cycle throughout 1 year. In this report, we focus on the

ration including E2 production and ovulation (Emmen ERs in gonads: ERa (esr1) and two paralogs of ERb (esr2a et al. 2005). and esr2b). We also analyzed the relative mRNA levels of In ﬁsh, the major target organs of estrogens are gonads, ar, which mainly acts in spermatogenesis (Shi et al. 2012), liver, pituitary, and brain (Filby & Tyler 2005). However, and cytochrome P450 aromatase (cyp19a1a) catalyzing the direct effect of estrogen on gonadal development and the biosynthesis of estrogens (Johnsen et al. 2013)to the role of ERs are poorly understood (Shi et al.2011). determine their association with esr expression in gonads.

Materials and methods sampled from individual fish at each sampling time point, fixed with 4% paraformaldehyde (PFA)/PBS at 4 8C for 24 h, Fish husbandry and sample collection or snap-frozen and stored at K80 8C respectively. PFA-fixed Atlantic cod were hatched and reared at Mørkvedbukta gonads were then washed, embedded in paraffin, sectioned Research Station (University of Nordland, Norway). In at 4.5-mm thickness, and stained with hematoxylin–eosin. 2009, when the fish were 2-year-old adults (initial weight All procedures used in this study were in accordance with the of 1.1G0.2 kg, meanGS.D., nZ18), they were transferred to guidelines set by the National Animal Research Authority three land-based flow-through tanks (40 m3) at a density (Forsøksdyrutvalget, Norway) and approved by the Faculty of w100 individuals/tank from August 2009 to May 2010. of Biosciences and Aquaculture (University of Nordland, Seawater (7.4G0.4 8C) was pumped from 200 m depth. Norway) Ethics Committee. A commercial diet (Amber Neptun, Skretting AS, Stavanger, Norway) was provided daily at 5% (w/w) body weight of the RNA extraction and cDNA synthesis fish by automatic belt feeders. Eight fish were taken from Total RNA was extracted from various tissues using QIAzol each tank (nZ24), killed by exposure to 0.5 g/l tricaine lysis reagent (Qiagen) as detailed elsewhere (Campos et al. methanesulfonate (Sigma–Aldrich), total length, total body 2010) and quantified by spectrophotometry with a weight (TW), and gonad weight (GW) were measured, and NanoDrop ND-1000 instrument (Thermo Scientific, the gonadosomatic index (GSI, 100!GW/TW (%)) was Saven & Werner AS, Kristiansand, Norway). RNA integrity calculated at each sampling time point (August, November was assessed by electrophoresis on a 1% (w/v) agarose in 2009, and February, March, and May in 2010; Table 1). gel. Total RNA (1 mg) was transcribed to cDNA using a Differences in all numerical data for each sampling point Quantitect reverse transcriptase kit (Qiagen) as reported within fish of the same sex were determined using a Tukey’s previously (Campos et al. 2010). multiple comparison test using the GraphPad Prism Soft- ware, when the one-way ANOVA was significant. Normality cDNA cloning of ER (esr) genes in Atlantic cod and equal variance conditions were met, and the significance level was set at P!0.05. In order to minimize blood In order to identify genomic sequences that con- contamination, the fish were exsanguinated after terminal tain putative coding regions for esr genes in Atlantic anesthesia by cutting their gill arches before sampling of cod, TBLASTN searches were performed in Ensembl other tissues. Excised tissues were then rinsed three times (http://www.ensembl.org/) against the cod genome

Journal of Molecular Endocrinology with PBS. For semi-quantitative RT-PCR, brain, gill, heart, assembly (release 69, October 2012; Flicek et al. 2013). head kidney, kidney, liver, spleen, stomach, midgut, testis The following protein sequences were used as queries: (GSIZ4.7%), ovary (GSIZ1.5%), skeletal muscle, skin, and i) gilt-head sea bream ERa (AAD31032.2), ERb1 blood were collected, snap-frozen in liquid nitrogen, and (AAD31033.1), and ERb2 (CAE30470.1) and ii) rainbow stored at K80 8C for subsequent RNA extraction. For trout ERa1 (NP_001117821.1), ERa2 (NP_001118030.1), histological observations and quantitative real-time PCR ERb1 (NP_001118225.1), and ERb2 (NP_001118042.1). (qPCR) analysis, testes (nZ5) and ovaries (nZ5) were After extracting the corresponding genomic sequences

Table 1 Total length (TL), total body weight (TW), gonad weight (GW), and gonadosomatic index (GSI, 100!GW/TW) of Atlantic cod sampled throughout a sexual maturation cycle

Sex Sampling point TL (cm) TW (g) GW (g) GSI (%) n

Male August 2009 44.3G0.7a 1050.0G75.8a 1.0G0.5a 0.1G0.0a 9 November 2009 50.7G1.0b 1566.7G92.5a,b 37.9G6.9a 2.5G0.4a 9 February 2010 50.9G1.4b 1746.2G158.0b 218.1G27.9b 12.4G1.0b 8 March 2010 53.6G1.7b 2346.0G211.6c 198.0G25.9b 11.2G1.2b 9 May 2010 51.4G1.3b 1452.3G121.2b 12.1G1.4a 0.9G0.1a 7 Female August 2009 45.2G1.0a 1143.3G65.7a 15.9G1.8a 1.4G0.1a 9 November 2009 51.5G0.9b 1563.6G72.5a 38.9G3.2a 2.5G0.2a 14 February 2010 52.6G1.0b 2098.2G134.1b 231.4G21.1b 11.2G1.0b 14 March 2010 53.8G1.0b 1855.0G119.9a,b 405.1G51.7c 18.3G1.7c 16 May 2010 53.6G0.9b 1726.6G83.1a,b 234.5G58.6b 12.9G2.8b 14

Values with different lower-case letters display signiﬁcant differences within the same sex (P!0.05).

Table 2 Primers used for cDNA cloning, semi-quantitative RT-PCR, quantitative real-time PCR (qPCR), and in situ hybridization (ISH). Gene names, GenBank accession numbers, PCR efﬁciencies (E, %), coefﬁcient of determination values (R2), and amplicon sizes (bp) are indicated

Gene GenBank Forward primer (50/30) Reverse primer (50/30) Purpose ER2 Size (bp)

esr1 JX178935 TGCAGTCCCTGGG- TGACCTCGGTG- Cloning/ISH – – 822 CAGTGGGTCCACCA TACGGCCGGTTCATCT – – AGGCAGCTGGAGAACAG- TGGGCGTCCAGCATCTCCAG- Cloning/ISH – – 643 GACGTCGCC CAGCAGGT – – AAAGGAGGTATGCGCAAGG TGACCTCGGTG- RT-PCR/qPCR 83.0 0.999 307 TACGGCCGGTTCATCT esr2a JX178936 GGACACTGACATGGACT- TGCAGCTGGA- Cloning – – 655 GAAGGAGTA GAACTGGTTGCTCCCGCT (50RACE)/ISH – – CCTTACACCGACCTGGGCCAT- GCATGATGTGGGCGTCCAG- Cloning/ISH – – 1438 GACT CATCTC – – TTGGTATGGCAACCTCCTCCC- GGCCGTAGAAGGG- Cloning – – 213 CAA CACGGCGCCAA – – CCTTACACCGACCTGGGCCAT- ACGACACCA- RT-PCR/qPCR 85.4 0.999 304 GACT CACTCTCCTCGCTGTCCT esr2b JK993476 CAGATCGCTGCTGTCGTCCAA- CCGTAGTGGTACCCGGACGCG- Cloning/RT-PCR/ 92.3 0.977 183 TAAGA TAGTCCT ISH/qPCR ar FJ268742 CGATATGTTCCCAGGAATGAG GGTGGTTCTGTTTACCTGCTG QPCR 80.4 0.982 152 cyp19a1 DQ402370 ACAACAACAAGTACGGCAG- GTAGAGGAGCTGCTGAGGAT- QPCR 86.0 0.991 76 CAT GAG actb EX739174 TGACCCTGAAGTACCCCATC TCTTCTCCCTGTTGGCTTTG QPCR 83.0 0.996 162 arp EX741373 TGATCCTCCACGACGATGAG CAGGGCCTTGGCGAAGA RT-PCR/qPCR 91.6 0.994 113 eef1a EX721840 CACTGAGGTGAAGTCCGTTG GGGGTCGTTCTTGCTGTCT QPCR 88.9 0.995 142 vasa HM451456 CTGCGTGCGTCCAGTGGTGTT- TCTTTCTGGAGTCCGTGGAGG- ISH – – 920 GTA CAA gsdf KC204828 ATGA- ACAGCCACACTGGCGGGT- ISH – – 600 CAGTCCTGCTGGGCTCTTC- CAGCACCG CAT 3bhsd KC204829 ATGTCTCTGAGTGGG- TGCCTTCTGGTAGCCGAAGCT- ISH – – 786 GACGTGTGTCT GAACG amh JN802292 TCCTACACCATCCCGGTGTCTT- TCAGCGGCAGCCA- ISH – – 849 CAGG CATTCTTTGGCA Journal of Molecular Endocrinology (esr1, GeneScaffold_4271; esr2a, GeneScaffold_1546; and corresponding orthologs in various species (Supplementary esr2b, GeneScaffold_26), putative cDNA sequences were Table 1, see section on supplementary data given at the obtained with the Augustus gene prediction software end of this article) using MUSCLE (www.drive5.com). (http://augustus.gobics.de/) and used to design specific Poorly aligned or divergent regions were trimmed using primer sets (Table 2). PCRs were performed as described Gblocks 0.91b (molevol.cmima.csic.es). The resulting by Nagasawa et al. (2012). For 50 RACE of esr2a, the gene- multiple sequence alignment was used for Bayesian specific primers listed in Table 2 and the GeneRacer Kit (MrBayes v3.1.2, mrbayes.csit.fsu.edu) phylogenetic recon- with SuperScript III RT (Invitrogen) were used according struction as detailed elsewhere (Fernandes et al.2010). to the manufacturer’s instructions. Specific amplicons A graphical representation of the phylogenetic tree was were separated by electrophoresis, purified, cloned, and obtained using the PhyloWidget (www.phylowidget.org). sequenced as detailed elsewhere (Nagasawa et al. 2012).

Semi-quantitative RT-PCR Bioinformatic analysis Semi-quantitative RT-PCR was performed with cDNAs Identity of cod esr cDNA sequences was confirmed by from the various tissues mentioned above (male, GSIZ TBLASTN searches against the non-redundant database at 4.7% and female, GSI Z1.5%) using gene-specific primer NCBI BLAST (www.ncbi.nlm.nih.gov). For phylogenetic sets (Table 2). Thermocycling parameters were 94 8C for analysis, Atlantic cod cDNA sequences were extended by 3 min, followed by 35 cycles (25 cycles for acidic ribosomal in silico cloning using the Ensembl genome browser and protein (arp)) of 30 s at 94 8C, 30 s at 60 8C, and 30 s their deduced amino acid (AA) sequences were aligned with at 72 8C, with a final elongation step of 72 8C for 3 min.

PCR products were separated by electrophoresis on a 1% Quantitative real-time PCR (w/v) agarose gel and visualized with the Kodak gel Twenty-fold diluted samples of cDNA synthesized as documentation system v.4.0.5. Expression patterns were described above were used in duplicate, as were negative confirmed using two biological replicates of each sex. (minus reverse transcriptase) and no template (water) controls. Gene-specific qPCR primers for the three esr In situ hybridization paralogs were designed using the GenScript Real-time PCR Spatial expression patterns of esr paralogs were determined Software (www.genscript.com) to span at least one intron/ by in situ hybridization (ISH), along with those of known exon border whenever possible (Fernandes et al. 2008). germ cell and gonadal somatic cell markers reported in The qPCR was run at 95 8C for 15 min, followed by 45 other teleosts. The gene vasa was used for detection of cycles of 15 s at 94 8C, 20 s at 60 8C, and 20 s at 72 8C using spermatogonia/spermatocytesintestesoroogonia/ LightCycler 480 SYBR Green I Master chemistry (Roche) oocytes in ovaries because of its well-conserved spatio- on a LightCycler 480 (Roche), as described previously temporal expression pattern among teleosts, including (Campos et al. 2010). Specificity of the reactions was Atlantic cod (Presslauer et al. 2012). The gonadal soma- evaluated by melting curve analysis and further by Sanger derived factor (gsdf) was used to identify Sertoli cells in sequencing of amplicons. Amplification efficiencies were testes or follicular cells in ovaries, as it is a proven marker calculated by standard curves from a fivefold dilution of these cells in rainbow trout (Sawatari et al. 2007), series (1:1, 1:5, 1:25, 1:125, and 1:625) of pooled cDNA as

medaka (Shibata et al. 2010), and zebrafish (Gautier et al. detailed by Fernandes et al. (2008). Cycle threshold (CT) 2011). In addition to gsdf, anti-Mu¨llerian hormone (amh) values were determined using the LightCycler 480 Soft- has been used as a Sertoli cell marker in zebrafish ware with a level of fluorescence intensity arbitrarily set (Rodriguez-Mari et al. 2005). Leydig cells were detected at 1. Three endogenous reference genes, b-actin (actb), arp, using 3b-hydroxysteroid dehydrogenase (3bhsd)asa and eukaryotic elongation factor 1a (eef1a), were validated marker, as reported for rainbow trout (Kobayashi et al. as reported by Fernandes et al. (2008) and Nagasawa et al. 1998), swamp eel (Lo Nostro et al. 2004), and Nile tilapia (2012). Their geNorm stability values (M) were 0.812, 0.779, (Ruksana et al. 2010). The gsdf and 3bhsd genes were and 0.618 for actb, arp,andeef1a respectively. The most C cloned from Atlantic cod in this study. Digoxigenin- stable gene combination was found to be arp eef1a with an Z labeled sense and anti-sense RNA probes for all genes average pairwise variation V2/3 0.258. Therefore, raw qPCR examined (esr1, esr2a, esr2b, vasa, gsdf, amh, and 3bhsd) data of target genes were corrected using the geometric average of arp and eef1a quantities as normalization factors.

Journal of Molecular Endocrinology were individually in vitro transcribed using T3 or T7 RNA polymerase (Roche) from corresponding DNA fragments Differences in relative expression levels of three genes esrs, amplified using the primers listed in Table 2. Tissue ar,andcyp19a1a among different sampling time points samples were obtained from 2-year-old Atlantic cod reared or between sexes were examined by Mann–Whitney U tests at Mørkvedbukta Research Station. Testes (GSIZ0.4%) and using the GraphPad Prism Software (San Diego, CA, USA), ovaries (GSIZ1.3%) were fixed with Bouin’s solution as the data were not normally distributed. The significance ! (Sigma–Aldrich) and 4% PFA/PBS respectively at 4 8C for level was set at P 0.05. 12–24 h, washed out with nuclease-free water, dehydrated with an ascending ethanol series, and then embedded in paraffin. Cross sections of testes and ovaries samples were Results prepared at 4-mm thickness, attached onto polylysine- Gonadal maturation status throughout an annual coated glass slides (Polysine, VWR international, Leuven, reproductive cycle in 2-year-old Atlantic cod Belgium), and used for ISH analysis as reported by Fernandes et al. (2006) with slight modifications. Sense Atlantic cod kept in land-based tanks throughout a year and anti-sense probes for all genes (esr1, esr2a, esr2b, vasa, showed a clear reproductive cycle with a significant gsdf, amh, and 3bhsd) were individually hybridized at 70 8C increase in GW (Table 1) and differentiation of germ for 24 h and then washed out with RNase A treatment cells in both testes and ovaries (Fig. 1). Throughout a (Sigma–Aldrich) as reported previously (Nagasawa et al. reproductive cycle, testis and ovary weights reached a peak 2009). Sections were mounted on gelatin embedding in February (GSIZ12.4G1.0%, nZ8) and March (GSIZ medium (Sigma–Aldrich) and observed under a BX-51 18.3G1.7%, nZ16) respectively (Table 1). Spermiation microscope (Olympus). was seen in some male individuals in February and March,

Testes Ovaries

A F

nu f August n

GSI=0.1% GSI=1.2%

B G

SC SC nu ch

November ca SC SC

GSI=2.7% GSI=2.3%

C H

SZ ch SC February yg

GSI=8.5% GSI=9.4% f

D I SZ

yg Journal of Molecular Endocrinology March f SZ SC ch GSI=11.4% GSI=15.9%

E J gmc f

ct SZ May SZ SZ ch he gmc

GSI=0.7% GSI=7.3%

Figure 1 Histology of representative testes and ovaries at different maturation status the cytoplasm and a large central nucleus. (I) Vitellogenic oocyte with using hematoxylin–eosin staining in 2-year-old Atlantic cod throughout an enlarged yolk granules, eccentric nucleus, and chorion. (J) Hydrated egg in annual reproductive cycle. (A) Spermatogonia and cysts of primary follicle sac with post-ovulatory follicle. Arrowheads, spermatogonia; arrows, spermatocytes. (B) Cysts of spermatocytes. (C and D) Cysts of spermatocytes primary spermatocytes; ca, cortical alveoli; ch, chorion; ct, connective tissue; and lobules containing spermatozoa. (E) Gonadal myoid cells with remaining f, follicle layer; gmc, gonadal myoid cells; he, hydrated egg; n, nucleus; nu, spermatozoa. (F) Perinuclear stage oocytes with a large circular nucleus and nucleolus; SC, spermatocytes; SZ, spermatozoa; and yg, yolk granule. The scale peripheral nucleoli. (G) Oocyte with cortical alveoli, chorion, and nucleus with bars represent 100 mm. A full colour version of this ﬁgure is available at http:// detached nucleoli. (H) Vitellogenic oocytes with yolk granules ﬁlling most of dx.doi.org/10.1530/JME-13-0187.

whereas ovulation was seen in a few female individuals in 1.0 O. niloticus ERα 0.76 H. burtoni ERα 1.0 A. melanopus ERα May. The testis weight was 218-fold larger in February 1.0 F. heteroclitus ERα O. latipes ERα 1.0 α compared with August, whereas the ovary weight 1.0 G. aculeatus ER α 0.88 S. schlegelii ERα ER (teleost) E. coioides ERα increased by 25.5-fold from August to March. The GSIs 1.0 α 0.98 1.0 A. schlegelii ER S. auratus ERα for both sexes decreased significantly from March to May. 1.0 0.98 D. labrax ERα G. morhua ERα Histological observations of gonads revealed the matu- 1.0 O. mykiss ERα1 1.0 O. mykiss ERα2 D. rerio ERα ration and differentiation status of germ cells: in testes, α 1.0 S. denticulatus ER 1 α α 1.0 α ER 1/ER 2 (teleost) 1.0 S. denticulatus ER 2 spermatogonia and several cysts of primary spermatocytes V. barbatulus ERα 1.0 α 1.0 0.89 C. auratus ER 1 were observed in immature testis in August (Fig. 1A). C. auratus ERα2 α 1.0 M. musculus ER ERα (tetrapods) In November, testes were mainly filled with the cysts of H. sapiens ERα 1.0 O. hiloticus ERβ1 0.75 H. burtoni ERβ1 spermatocytes (Fig. 1B), and lobules containing sperma- β 0.94 1.0 F. heteroclitus ER 1 O. latipes ERβ1 tozoa were subsequently observed in February and March A. melanopus ERβ1 1.0 0.73 G. aculeatus ERβ1 (Fig. 1C and D). After spermiation, a number of gonadal 1.0 E. coioides ERβ1 ERβ1 (teleost) S. schlegelii ERβ1 1.0 β 1.0 1.0 A. schlegelii ER 1 myoid cells were preferentially observed in testis with 1.0 S. auratus ERβ1 1.0 D. labrax ERβ1 remaining spermatozoa in lobules (Fig. 1E). In ovaries, 1.0 G. morhua ERβ1 1.0 O. mykiss ERβ1 perinuclear stage oocytes were mainly observed in August 1.0 S. denticulatus ERβ1 β 1.0 C. auratus ER 1 0.82 V. barbatulus ERβ2 (Fig. 1F) and then some oocytes with cortical alveoli were D. rerio ERβ2 0.91 S. denticulatus ERβ2 observed in November (Fig. 1G). Vitellogenic oocytes with 1.0 V. barbatulus ERβ1 1.0 C. auratus ERβ2 yolk granules accumulating in their cytoplasm were found 1.0 D. rerio ERβ1 1.0 O. niloticus ERβ2 1.0 H. burtoni ERβ2 mainly in February (Fig. 1H). Enlarged yolk granules and F. heteroclitus ERβ2 0.96 O. latipes ERβ2 0.92 0.56 eccentric nuclei were observed in vitellogenic oocytes in 0.67 A. melanopus ERβ2 E. coioides ERβ2 S. schlegelii ERβ2 ERβ2 (teleost) March (Fig. 1I). Post-ovulatory follicles and remaining 0.97 0.85 1.0 1.0 D. labrax ERβ2 β hydrated eggs in follicle sacs were found in May (Fig. 1J). 1.0 0.89 A. schlegelii ER 2 1.0 S. auratus ERβ2 1.0 G. aculeatus ERβ2 G. morhua ERβ2 O. mykiss ERβ2 1.0 M. musculus ERβ ERβ (tetrapods) Three esr paralogs in Atlantic cod H. sapiens ERβ 1.0 M. musculus ESRRA H. sapiens ESRRA ESRRA Partial cDNA sequences of three esr genes in Atlantic cod D. rerio ESRRA were obtained using a combination of in silico and experimental cloning: esr1 (JX178935), esr2a (JX178936), Figure 2 Phylogenetic inference of ER subtypes found in vertebrates. The numbers Journal of Molecular Endocrinology and esr2b (JK993476). Partial ERa AA sequences (479 AA) at the nodes indicate posterior probability obtained from Bayesian overlapped on 83% of the region of full-length orange- analysis. GenBank accession numbers of all ERs are given in spotted grouper ERa with 81% similarity and 75% Supplementary Table 1. identity, and five conserved essential domains of ERs were identified: A/B (98 AA (partial)), C (84 AA), D (62 AA), and Spinibarbus denticulatus) was observed. G. morhua ERa E (240 AA), and F (64 AA). For ERb1, the 532 AA partial was grouped in the ERa clade. The two clades of teleost sequence covered 94% of the region of the full-length ERb1 and ERb2 displayed similar topologies. In both the Japanese flounder ERb with 83% similarity and 74% teleost ERb1 and ERb2 clades, G. morhua ERb1 and ERb2 identity and contained the essential domains A/B (157 were most closely related to O. mykiss ERb1 and ERb2 AA), C (82 AA), D (47 AA), and E (246 AA (partial)). In respectively. Unexpectedly, some Cyprinidae species addition, the 61 AA ERb2 partial sequence corresponded to (e.g. S. denticulatus, Varicorhinus barbatulus, C. auratus, 9% of full-length gilt-head sea bream ERb2 and comprised and Danio rerio) showed an isolated subcluster distinct A/B (24 AA in (partial)) and C (37 AA in (partial)) domains. from teleost ERb1 and ERb2. Bayesian analysis using in silico-extended cod sequences containing the E-domain produced a consensus phyloge- Tissue distribution of esr transcripts netic tree that displayed a clear separation of ERa,ERb1, and ERb2 subtypes into three main clades, whereas all the The genes esr1, esr2a,andesr2b were differentially homologs of estrogen-related receptor a (ESRRA) were expressed across various tissues in both adult male and grouped outside of all ER clades (Fig. 2). In the teleost ERa female Atlantic cod (Fig. 3). The esr1 transcripts were clade, an isolated subcluster containing ERa1 and ERa2 mainly found in liver and gonads in both sexes, but from a few teleost species (e.g. Carassius auratus, O. mykiss, prominent in testes and in livers of females. The esr2a

that of esr1 (Fig. 4H and I). These two esrs were preferentially observed in the testicular interstitial ﬁbroblasts, whereas

Brain Gill HeartHead KidneykidneyLiver SpleenStomachMidgutGonadMuscleSkin BloodNC weak signals for esr2b were detected in some interstitial Male fibroblasts (Fig. 4J). No hybridization signals were observed esr1 Female with sense probes (Supplementary Figure 1,seesectionon supplementary data given at the end of this article). Male esr2a Female Localization of esr transcripts in ovary Male esr2b Female Sequential cross sections of ovary were prepared and used for histological analyses (Fig. 5A and B). Several develop- Male arp mental stages of oocytes could be visualized by hemat- Female oxylin staining (Fig. 5C). ISH with a vasa probe strongly stained pre- and early-vitellogenic oocytes rather than mid- Figure 3 vitellogenic oocytes (Fig. 5D), while a gsdf probe stained Tissue distribution of three esr paralogs in 2-year-old Atlantic cod adult tissues was analyzed by semi-quantitative RT-PCR. The three esr genes were the follicular cell layer surrounding oocytes at all develop- differentially expressed among the tissues tested and showed a sexually mental stages (Fig. 5E). Signals for esr1 were predominantly Z Z dimorphic expression pattern (male, GSI 4.7% and female, GSI 1.5%). observed in follicular cells, (Fig. 5F), whereas, esr2a signals The arp gene was used as an endogenous reference gene, as it showed stable expression among all tissues in Atlantic cod. were mainly detected in the follicular cell layer and the cytoplasm of pre- and early-vitellogenic oocytes (Fig. 5G). transcripts were widely distributed in several tissues with In both follicular cells and the cytoplasm of all oocytes, varied expression levels. Nevertheless, the esr2a transcripts esr2b signals were detected (Fig. 5H). No hybridization exhibited a sex-dependent distribution in liver and signals were observed in any of the cells with negative gonads, similar to esr1. Expression of esr2a was found in control probes (Supplementary Figure 2, see section on supplementary data given at the end of this article). skin of males and blood of females, whereas the esr2a transcripts were seen in the pronephros and excretory kidney in males but not much in females. The esr2b Sexually dimorphic expression of ar, cyp19a1a, and transcripts had a broad distribution in female tissues, esr genes during a reproductive cycle while their expression was found in fewer male tissues, Journal of Molecular Endocrinology The levels of esr1 mRNA were upregulated by 12.6- and including testis and liver. Notably, expression of esr genes 7.0-fold from August to March in both testes and ovaries in liver was more prominent in females than in males and respectively (Fig. 6A). In March, esr1 mRNA levels in all esr transcripts were consistently present at higher levels testes were 5.0-fold higher than those in ovaries (Fig. 6A). in testis than in ovary, except for esr2b. In brain, low A significant increase in esr2a mRNA level in testes was transcript levels of esr1 in females and esr2a in males were particularly evident in February and March (P!0.05; observed but expression of esr2b was not detected. Fig. 6B), but no significant alterations in esr2a in ovaries were observed throughout the reproductive cycle. The Localization of esr transcripts in testis level of esr2a mRNA in testes was 13.3- and 12.0-fold higher than that in ovaries in February and March The central region of the testes was sampled for histological respectively (Fig. 6B). The level of esr2b mRNA in ovaries analyses (Fig. 4A) and sequential cross sections of the was elevated up to 24.9-fold from August to March and testicular lobe were prepared (Fig. 4B). Hematoxylin staining then rapidly decreased by May, showing a temporal showed the presence of spermatogonia and spermatocytes pattern similar to esr1 in testis (Fig. 6C). In addition, ar (Fig. 4C). ISH with a vasa probe specifically stained germ cells mRNA levels in testes were elevated up to 6.5-fold from (Fig. 4D), while staining with gsdf or amh probes affected August to March and then decreased in May, while ar Sertoli cells surrounding germ cell cysts (Fig. 4E and F). A expression in ovaries had a modest peak in November 3bhsd probe detected Leydig cells in the interstitial tissue (Fig. 6D). Remarkably, cyp19a1a transcript levels in ovaries between lobule segments (Fig. 4G). There was no visible increased by 37.5-fold from August to March, whereas difference in localization between esr1 and esr2a in testis, cyp19a1a was barely detected in testes throughout the even though the staining intensity of esr2a was higher than reproductive cycle (Fig. 6E).

A B

a p Efferent duct

CGHematoxylin 3βhsd

DHvasa esr1

EIgsdf esr2a

FJamh esr2b Journal of Molecular Endocrinology

Figure 4 Localization of transcripts of three esrs, vasa, gsdf, 3bhsd,andamh in staining (C) and ISH staining with sense or anti-sense probes for vasa (D), gsdf Atlantic cod testes. (A) Excised testis from 2-year-old Atlantic cod (E), amh (F), 3bhsd (G), esr1 (H), esr2a (I), and esr2b (J). (C, D, E, F,G, H, I and J) (GSIZ0.4%). The dashed line indicates the region sampled for ISH analysis. a, High-magnification fields taken from the areas adjacent to insets. White anterior and p, posterior. (B) Schematic representations of a testicular lobe arrowheads indicate spermatogonia. SC, spermatocytes. Arrows indicate connected to efferent duct (left panel in (B)) and its cross-sectional image the position of weak signals that were detected in interstitial fibroblasts. (right panel in (B)). The dashed line indicates the section location. The area The bars represent 5 cm (A), 100 mm (C, D, E, F, G, H, I and J), and 200 mm enclosed by the black box in (B) indicates the position of low-magnification (insets in (C, D, E, F, G, H, I and J)). A full colour version of this figure is fields (insets in (C, D, E, F, G, H, I and J)). (C, D, E, F, G, H, I and J) Hematoxylin available at http://dx.doi.org/10.1530/JME-13-0187

were hitherto unknown in the teleost superorder Discussion Paracanthopterygii. The sexually dimorphic expression Knowledge about the direct functions of estrogen in of esr transcriptsingonadswithpeaksbeforethe gonadal cells, and particularly the role of ER subtypes, spawning period throughout a reproductive cycle is still limited to a few mammal species. In this study, represents the ﬁrst observation, to our knowledge, in a we characterized three esr paralogs in Atlantic cod that batch spawner.

a p

CFHematoxylin esr1

DGvasa esr2a

EHgsdf esr2b

Figure 5 Journal of Molecular Endocrinology Localization of transcripts of three esrs, vasa, and gsdf in Atlantic cod ovary. staining (C) and ISH staining with anti-sense probes for vasa (D), gsdf (E), (A) Excised ovary from 2-year-old Atlantic cod (GSIZ1.3%). The region esr1 (F), esr2a (G), and esr2b (H). (C, D, E, F, G and H) High-magnification sampled for ISH analysis is indicated by the dashed line. a, anterior and p, fields taken from the areas adjacent to insets. The bars represent 5 cm (A), posterior. (B) Schematic representation of cross section of ovary. The area 100 mm (C, D, E, F, G and H), and 200 mm (insets in (C, D, E, F, G and H)). enclosed by the black box in (B) indicates the position of low-magnification A full colour version of this figure is available at http://dx.doi.org/10.1530/ fields (insets in (C, D, E, F, G and H)). (C, D, E, F, G and H) Hematoxylin JME-13-0187

In teleost fish, two subtypes of ERb (ERb1 and ERb2– localization of esr2a and esr2b indicates that these paralogs also named ERg) have been identified in distant taxa. arose from the 3R teleost-specific whole-genome dupli- In contrast, most fish species have only ERa, but two ERa cation event that occurred 350 million years ago (Meyer & isoforms (ERa1 and ERa2) have recently been reported in Van de Peer 2005). salmonids (rainbow trout; Nagler et al. (2007))and Teleost esrs exhibit a wide variety of tissue distribution

cyprinids (goldﬁsh C. auratus and phoenix barb patterns but with common E2 target tissues associated with S. denticulatus; Zhu et al. (2008)). Atlantic cod has a single reproduction such as liver and gonads (Filby & Tyler 2005, ERa subtype, supporting the hypothesis that the two ERa Chen et al.2011). In Atlantic cod, all three esr transcripts subtypes (ERa1 and ERa2) reported in goldﬁsh, rainbow were present in liver and gonads with a sexual dimorphic trout, and phoenix barb are derived from the recent expression pattern. Prominent expression of esrs in livers of tetraploidization event that is thought to have occurred females compared with livers of males could be caused by

in the salmonid and cyprinid lineages. Within the ERb the higher endogenous E2 level in female serum (Norberg subtype, esr2a and esr2b were in silico mapped on different et al.2004). In gonads, esr1 and esr2a were predominantly scaffolds in the Atlantic cod genome. The chromosomal expressed in testes compared with ovaries indicating an

ABCDE* 2.0 esr12.0 esr2a* 2.0 esr2b1.0 ar2.0 cyp19a1a * * * B,C d A Testis B,C A c 0.8 B 1.5 1.5 1.5 1.5 * Ovary a,b,c,d c B 0.6 * 1.0 1.0 1.0 1.0 * c * * A 0.4 * A A,B A,B A,B * 0.5 0.5 A,B A 0.5 0.5 c c 0.2 b * Relative expression Relative expression Relative expression Relative expression b,c B Relative expression A A,B B b a,c a A b A b a a a a B a,c c a a,b a a A a A A A A 0.0 0.0 0.0 0.0 0.0

May May May May May August March August March August March August March August March February February February February February November November November November November Time Time Time Time Time

Figure 6 Quantification of three esrs (A, B and C), ar (D), and cyp19a1a (E) in gonads differences within the same sex throughout the reproductive cycle (capital of Atlantic cod throughout the reproductive cycle. Black and white bars letters, testes and lower-case letters, ovaries; nZ25). Asterisks (*) indicate show the relative mRNA expression in testes and ovaries respectively. significant differences between testes and ovaries at a particular sampling Transcript levels were quantified by real-time PCR. Error bars show S.E.M.at point at P!0.05 (nZ5). each sampling time point. Different superscript letters indicate significant

important role of both ERs in spermatogenesis, as proposed (Wu et al. 2001), indicating a diversity of E2 signaling for other teleost species (Pinto et al.2006, Zhu et al.2008). requirements of spermatogenic cells among teleosts. There In rodents, several studies have demonstrated that are only two other reports describing the localization of the spatial distribution of ER subtypes in testis varies esrs in teleost ovarian cells so far: esr1 in eelpout significantly among species, as reviewed by O’Donnell (Andreassen et al. 2003) and esr1, esr2a, and esr2b in et al. (2001), but there is a consensus that both ERa and medaka (Chakraborty et al. 2011b). In Atlantic cod, ERb in rodent ovary are expressed in granulosa cells of transcripts of the three esr genes were differentially follicles (Drummond & Findlay 1999). In teleosts, there is localized in ovaries even though they were all observed very limited knowledge about the localization of ER mRNA in follicular cells surrounding mid-vitellogenic oocytes. or protein in adult gonads and their expression pattern Some degree of conservation in esr expression patterns is controversial (Bouma & Nagler 2001, Wu et al. 2001, may be noted between cod and the above two species. Andreassen et al. 2003, Chakraborty et al. 2011b). This First, esr1 is preferentially weakly expressed in follicular study delivers significant insights into this matter by cells of pre- to mid-vitellogenic oocytes. In contrast, esr2a identifying the cells expressing esr paralogs in fish gonads is predominantly expressed in the cytoplasm of pre- and for both sexes. In cod testis, all three esr transcripts were early-vitellogenic oocytes rather than their follicular cells.

Journal of Molecular Endocrinology preferentially found in testicular interstitial fibroblasts, Furthermore, esr2b is mainly expressed in both the similar to the localization of rainbow trout ERa protein in follicular layer and cytoplasm of oocytes. These obser- testis (Bouma & Nagler 2001) but not in the Sertoli cells vations of ovarian cells highlight the potential importance surrounding spermatogonia. It has been reported that of all three esrs in follicular cell growth and differentiation rainbow trout ERa protein is expressed in interstitial during cod folliculogenesis, as suggested for rodents fibroblasts containing Leydig cell precursors and is (Drummond & Findlay 1999, Drummond et al. 1999). involved in differentiation of precursors into mature Further studies are required to specify the function of ER Leydig cells (Bouma & Nagler 2001). In Atlantic cod testis, in fish oocytes. the number of the esr1/esr2a-expressing interstitial fibro- It is noteworthy that Atlantic cod testes have a unique blasts was larger than that of the 3bhsd-expressing structure in each individual testicular lobe: undifferentiated interstitial fibroblasts (Fig. 3G, H and I), indicating that spermatogonia and the most advanced germ cells are esr1/esr2a-expressing cells in testicular interstitial fibro- distributed in the periphery of the lobe and in the vicinity blasts could be both precursor and mature Leydig cells. of the collecting duct respectively (Almeida et al.2008).

Therefore, the main target cells of E2 via ERs in Atlantic Interestingly, this gradient distribution of different germ cell cod testicular cells could be Leydig cells, and ERs types was also visualized in our study using the ISH detection (particularly in esr1 and esr2a) may be involved in growth system with a gsdf probe (upper right panel in Fig. 4E). Similar and maturation of Leydig cells to maintain their steroido- to rainbow trout (Sawatari et al.2007), Atlantic cod gsdf was genesis. Nevertheless, there are reports of results conflict- highly expressed in the Sertoli cells surrounding undiffer- ing with our observations, which show predominant entiated and/or early spermatogonia rather than the Sertoli localization of esr in fish mitotic or meiotic germ cells cells surrounding cysts of advanced germ cells such as in eelpout (Andreassen et al. 2003) and channel catfish spermatocytes and spermatozoa. In contrast, amh transcripts

were distributed in Sertoli cells with similar signal intensities spermatogonial clones. In our study, we observed a regardless of germ cell developmental progression. gradual increase in ar expression in testis toward the To the best of our knowledge, this is the first paper spawning season and significantly higher transcript levels examining seasonal changes in all three esr transcripts in in February and March along with the increase in GSI in both testes and ovaries of a batch spawner throughout an testis, indicating that the number of Sertoli cells is annual reproductive cycle. At several time points, the increasing at this point. Taken together with the con- three esr paralogs were differentially expressed in male vs comitant increase in plasma androgen levels (Almeida female gonads. The results of GSI changes and histological et al. 2009), our data indicate the potential importance of observations of gonads enabled us to assume that their androgen signaling in Sertoli cells for maintaining main spawning period may be April and revealed a spermatogenesis and spermiogenesis in cod. In contrast, clear annual reproductive cycle in 2-year-old Atlantic we found only a modest peak of ar expression in ovaries in cod under the experimental conditions used, with a November, implying proliferation of the follicle cells and similar sexual maturation process to that reported by the epithelial cells of the ovigerous lamellae, as reported in Karlsen et al. (2006). ovary of Japanese eel Anguilla japonica (Tosaka et al. 2010). Gonadal esr1 mRNA expression increased significantly It should be noted that the molecular function of AR in in both testes and ovaries as gonadal maturation pro- teleost ovaries still remains to be investigated. gressed and decreased after the spawning period, resem- Aromatase is a key enzyme for converting androgens bling the results of esr1 expression in orange-spotted to estrogens during sexual development. It is encoded by a grouper ovaries at different developmental stages (Chen single gene in mammals, with exception of the domestic et al. 2011). A similar increase in mRNA level was also pig and other suiformes, which have multiple cyp19 genes found for esr2a in testes and for esr2b in ovaries. These (Corbin et al. 2009). In contrast, teleosts have two cyp19 sexually dimorphic expression patterns of esr paralogs genes: cyp19a1a is preferentially expressed in ovaries, along with the observed significant increase in their whereas cyp19a1b is found predominantly in brain

relative mRNA levels indicate that E2 signaling via ERs (Tchoudakova & Callard 1998). Our current knowledge may be essential for Leydig cells in testes and follicular of seasonal changes of cyp19a1a expression in ﬁsh ovaries cells in ovaries during gonadal maturation period. As throughout a reproductive cycle is still very limited esr1/esr2a genes are preferentially expressed in Leydig cells (Sampath Kumar et al. 2000, Rocha et al. 2009, Johnsen in cod testes and esr1 is speciﬁcally expressed in the et al. 2013). In Atlantic cod, cyp19a1a transcripts are follicular layer in cod ovaries, it is plausible that this constantly present at higher levels in ovaries than in

Journal of Molecular Endocrinology increase in their mRNA levels may be associated with testes, with a signiﬁcant increase in expression in ovaries Leydig cell proliferation before the spawning period and but not in testes towards the spawning season. Impor- proliferation of follicular cells during oocyte growth and tantly, these data indicate that cyp19a1a has potential as a maturation respectively. Recently, an in vitro study molecular indicator of ovarian maturation status in batch- uncovered the novel pathway of the direct function of spawning teleosts. It should be noted that the upregula- estrogen in the marsupial mammal gonad: the function of tion pattern of mRNA levels for cyp19a1a and esr1/esr2b in

estrogen mediated by ER blocks the nuclear entry of SOX9 ovaries coincided with the increase in plasma E2 levels in into gonadal somatic cells, preventing the testicular gene female cod reported by Norberg et al. (2004), indicating activation pathway and allowing upregulation of key that estrogen biosynthesis and its signaling via ERs are female genes in ovarian development (Pask et al. 2010). particularly important during the period before spawning. Therefore, further studies may examine the suppression of In conclusion, we identified three ER subtypes in Atlantic the sox9 transcription pathway by activated ER in gonadal cod, representing the superorder Paracanthopterygii, and somatic cells in fish. found that esr1, esr2a, and esr2b transcripts were detected AR plays a key role in androgen action and it is mainly in liver and gonads of fish of both sexes. All three essential for the maintenance of spermatogenesis in testis esr paralogs were predominantly localized within inter- (Ikeuchi et al. 2001). Almeida et al. (2009) reported that stitial fibroblasts in testes, whereas they were differentially Atlantic cod plasma 11-ketotestosterone levels increased expressed in both follicular cells and oocytes in ovaries. in February and March along with the advancement of Throughout an annual reproductive cycle, the mRNA level spermiogenesis and spermiation, in agreement with an of three esr genes, as well as ar and cyp19a1a, displayed sex- increase in GSI. Moreover, they showed that ar mRNA was dependent upregulation in gonads with a peak during the prominently localized in Sertoli cells regardless of growing period before spawning, indicating their importance in

preparation for this event, perhaps through activation of of the reproductive cycle. Biology of Reproduction 65 60–65. ER-responsive genes and proliferation of Leydig/Sertoli (doi:10.1095/biolreprod65.1.60) Britt KL, Drummond AE, Dyson M, Wreford NG, Jones ME, Simpson ER & and follicular cells. These findings provide new insights Findlay JK 2001 The ovarian phenotype of the aromatase knockout into the role of ERs in gonadal development during the (ArKO) mouse. Journal of Steroid Biochemistry and Molecular Biology 79 181–185. (doi:10.1016/S0960-0760(01)00158-3) reproductive cycle of batch-spawning teleosts. Campos C, Valente LMP, Borges P, Bizuayehu T & Fernandes JMO 2010 Dietary lipid levels have a remarkable impact on the expression of growth-related genes in Senegalese sole (Solea senegalensis Kaup). Journal of Experimental Biology 213 200–209. (doi:10.1242/jeb.033126) Supplementary data Chakraborty T, Shibata Y, Zhou LY, Katsu Y, Iguchi T & Nagahama Y 2011a This is linked to the online version of the paper at http://dx.doi.org/10.1530/ Differential expression of three estrogen receptor subtype mRNAs in JME-13-0187. gonads and liver from embryos to adults of the medaka, Oryzias latipes. Molecular and Cellular Endocrinology 333 47–54. (doi:10.1016/j.mce. 2010.12.002) Chakraborty T, Katsu Y, Zhou LY, Miyagawa S, Nagahama Y & Iguchi T Declaration of interest 2011b Estrogen receptors in medaka (Oryzias latipes) and estrogenic The authors declare that there is no conflict of interest that could be environmental contaminants: an in vitro–in vivo correlation. Journal of perceived as prejudicing the impartiality of the research reported. Steroid Biochemistry and Molecular Biology 123 115–121. (doi:10.1016/ j.jsbmb.2010.11.015) Chen H, Zhang Y, Li S, Lin M, Shi Y, Sang Q, Liu M, Zhang H, Lu D, Meng Z et al. 2011 Molecular cloning, characterization and expression Funding profiles of three estrogen receptors in protogynous hermaphroditic This work was funded by research grants from the Research Council of orange-spotted grouper (Epinephelus coioides). General and Comparative Norway to J M O F and I B (references 190350/S40 and 182653/V10 Endocrinology 172 371–381. (doi:10.1016/j.ygcen.2011.03.025) respectively), with additional support from the aid of the Norwegian Corbin CJ, Berger T, Ford JJ, Roselli CE, Sienkiewicz W, Trainor BC, Roser JF, Mechanism Europejskiego Obszaru Gospodarczego (EOG) and the Vidal JD, Harada N & Conley AJ 2009 Porcine hypothalamic aromatase Norwegian Financial Mechanism under the Scholarship and Training Fund. cytochrome P450: isoform characterization, sex-dependent activity, regional expression, and regulation by enzyme inhibition in neonatal boars. Biology of Reproduction 81 388–395. (doi:10.1095/ biolreprod.109.076331) Author contribution statement Dahlman-Wright K, Cavailles V, Fuqua SA, Jordan VC, Katzenellenbogen JA, I B and J M O F conceived and designed the experiments. K N, C P, and L K Korach KS, Maggi A, Muramatsu M, Parker MG & Gustafsson JA 2006 performed the experiments. K N and C P analyzed the data. I B and J M O F International Union of Pharmacology. LXIV. Estrogen receptors. contributed reagents/materials/analysis tools. K N, C P, L K, I B, and J M O F Pharmacological Reviews 58 773–781. (doi:10.1124/pr.58.4.8) wrote the paper. Drummond AE & Findlay JK 1999 The role of estrogen in folliculogenesis. Molecular and Cellular Endocrinology 151 57–64. (doi:10.1016/ S0303-7207(99)00038-6) Drummond AE, Baillie AJ & Findlay JK 1999 Ovarian estrogen receptor a Journal of Molecular Endocrinology Acknowledgements and b mRNA expression: impact of development and estrogen. We are grateful to Heidi Hovland Ludviksen, Dalia Dahle, and Hilde Ribe Molecular and Cellular Endocrinology 149 153–161. (doi:10.1016/ (University of Nordland, Norway) for providing Atlantic cod specimens and S0303-7207(98)00247-0) invaluable technical assistance. Drummond AE, Britt KL, Dyson M, Jones ME, Kerr JB, O’Donnell L, Simpson ER & Findlay JK 2002 Ovarian steroid receptors and their role in ovarian function. Molecular and Cellular Endocrinology 191 27–33. (doi:10.1016/S0303-7207(02)00052-7) References Emmen JM, Couse JF, Elmore SA, Yates MM, Kissling GE & Korach KS 2005 In vitro growth and ovulation of follicles from ovaries of estrogen Abney TO & Myers RB 1991 17b-Estradiol inhibition of Leydig cell receptor (ER)a and ERb null mice indicate a role for ERb in follicular regeneration in the ethane dimethylsulfonate-treated mature rat. maturation. Endocrinology 146 2817–2826. (doi:10.1210/en.2004-1108) Journal of Andrology 12 295–304. (doi:10.1002/j.1939-4640.1991. Fernandes JMO, Mackenzie MG, Wright PA, Steele SL, Suzuki Y, tb01604.x) Kinghorn JR & Johnston IA 2006 Myogenin in model pufferfish Almeida FF, Kristoffersen C, Taranger GL & Schulz RW 2008 Spermato- species: comparative genomic analysis and thermal plasticity of genesis in Atlantic cod (Gadus morhua): a novel model of cystic germ expression during early development. Comparative Biochemistry cell development. Biology of Reproduction 78 27–34. (doi:10.1095/ and Physiology. Part D, Genomics & Proteomics 1 35–45. (doi:10.1016/ biolreprod.107.063669) j.cbd.2005.09.003) Almeida FF, Taranger GL, Norberg B, Karlsen O, Bogerd J & Schulz RW 2009 Fernandes JMO, Mommens M, Hagen O, Babiak I & Solberg C 2008 Photoperiod-modulated testis maturation in Atlantic cod (Gadus Selection of suitable reference genes for real-time PCR studies of morhua, L.). Biology of Reproduction 80 631–640. (doi:10.1095/biol- Atlantic halibut development. Comparative Biochemistry and Physiology. reprod.108.071340) Part B, Biochemistry & Molecular Biology 150 23–32. (doi:10.1016/j.cbpb. Andreassen TK, Skjoedt K, Anglade I, Kah O & Korsgaard B 2003 Molecular 2008.01.003) cloning, characterisation, and tissue distribution of oestrogen receptor Fernandes JMO, Ruangsri J & Kiron V 2010 Atlantic cod piscidin and its a in eelpout (Zoarces viviparus). General and Comparative Endocrinology diversification through positive selection. PLoS ONE 5 e9501. 132 356–368. (doi:10.1016/S0016-6480(03)00101-1) (doi:10.1371/journal.pone.0009501) Bouma J & Nagler JJ 2001 Estrogen receptor-a protein localization in the Filby AL & Tyler CR 2005 Molecular characterization of estrogen receptors testis of the rainbow trout (Oncorhynchus mykiss) during different stages 1, 2a, and 2b and their tissue and ontogenic expression profiles in

fathead minnow (Pimephales promelas). Biology of Reproduction 73 development after insertional disruption of the mouse estrogen 648–662. (doi:10.1095/biolreprod.105.039701) receptor gene. PNAS 90 11162–11166. (doi:10.1073/pnas.90.23.11162) Flicek P, Ahmed I, Amode MR, Barrell D, Beal K, Brent S, Carvalho-Silva D, Meyer A & Van de Peer Y 2005 From 2R to 3R: evidence for a fish-specific Clapham P, Coates G, Fairley S et al. 2013 Ensembl 2013. Nucleic Acids genome duplication (FSGD). BioEssays: News and Reviews in Molecular, Research 41 D48–D55. (doi:10.1093/nar/gks1236) Cellular and Developmental Biology 27 937–945. (doi:10.1002/bies. Gautier A, Sohm F, Joly JS, Le Gac F & Lareyre JJ 2011 The proximal 20293) promoter region of the zebrafish gsdf gene is sufficient to mimic the Mimoto A, Fujii M, Usami M, Shimamura M, Hirabayashi N, Kaneko T, spatio-temporal expression pattern of the endogenous gene in Sertoli Sasagawa N & Ishiura S 2007 Identification of an estrogenic hormone and granulosa cells. Biology of Reproduction 85 1240–1251. (doi:10.1095/ receptor in Caenorhabditis elegans. Biochemical and Biophysical Research biolreprod.111.091892) Communications 364 883–888. (doi:10.1016/j.bbrc.2007.10.089) Hess RA 2000 Oestrogen in fluid transport in efferent ducts of the male Nagasawa K, Takeuchi Y, Miwa M, Higuchi K, Morita T, Mitsuboshi T, reproductive tract. Reviews of Reproduction 5 84–92. (doi:10.1530/ror.0. Miyaki K, Kadomura K & Yoshizaki G 2009 cDNA cloning and 0050084) expression analysis of a vasa-like gene in Pacific bluefin tuna Thunnus Ikeuchi T, Todo T, Kobayashi T & Nagahama Y 2001 Two subtypes of orientalis. Fisheries Science 75 71–79. (doi:10.1007/s12562-008-0021-9) androgen and progestogen receptors in fish testes. Comparative Nagasawa K, Giannetto A & Fernandes JMO 2012 Photoperiod influences Biochemistry and Physiology. Part B, Biochemistry & Molecular Biology 129 growth and mll (mixed-lineage leukaemia) expression in Atlantic cod. 449–455. (doi:10.1016/S1096-4959(01)00375-X) PLoS ONE 7 e36908. (doi:10.1371/journal.pone.0036908) Johnsen H, Tveiten H, Torgersen JS & Andersen O 2013 Divergent and Nagler JJ, Cavileer T, Sullivan J, Cyr DG & Rexroad C III 2007 The complete sex-dimorphic expression of the paralogs of the Sox9–Amh–Cyp19a1 nuclear estrogen receptor family in the rainbow trout: discovery of the regulatory cascade in developing and adult Atlantic cod (Gadus novel ERa2 and both ERb isoforms. Gene 392 164–173. (doi:10.1016/ morhua L.). Molecular Reproduction and Development 80 358–370. (doi:10. j.gene.2006.12.030) 1002/mrd.22170) Norberg B, Brown CL, Halldorsson A, Stensland K & Bjornsson BT 2004 Karlsen O, Norberg B, Kjesbu OS & Taranger GL 2006 Effects of photoperiod Photoperiod regulates the timing of sexual maturation, spawning, sex steroid and thyroid hormone profiles in the Atlantic cod (Gadus and exercise on growth, liver size, and age at puberty in farmed Atlantic morhua). Aquaculture 229 451–467. (doi:10.1016/S0044- cod (Gadus morhua L.). ICES Journal of Marine Science 63 355–364. 8486(03)00393-4) (doi:10.1016/j.icesjms.2005.10.013) O’Donnell L, Robertson KM, Jones ME & Simpson ER 2001 Estrogen and Kazeto Y, Tosaka R, Matsubara H, Ijiri S & Adachi S 2011 Ovarian spermatogenesis. Endocrine Reviews 22289–318.(doi:10.1210/er.22.3.289) steroidogenesis and the role of sex steroid hormones on ovarian growth Pask AJ, Calatayud NE, Shaw G, Wood WM & Renfree MB 2010 Oestrogen and maturation of the Japanese eel. Journal of Steroid Biochemistry and blocks the nuclear entry of SOX9 in the developing gonad of a Molecular Biology 127 149–154. (doi:10.1016/j.jsbmb.2011.03.013) marsupial mammal. BMC Biology 8 113. (doi:10.1186/1741-7007-8-113) Keaveney M, Klug J, Dawson MT, Nestor PV, Neilan JG, Forde RC & Gannon F Pinto PI, Passos AL, Martins Deborah MPRS & Canario AV 2006 1991 Evidence for a previously unidentified upstream exon in the human Characterization of estrogen receptor bb in sea bream (Sparus auratus): oestrogen receptor gene. Journal of Molecular Endocrinology 6 111–115. phylogeny, ligand-binding, and comparative analysis of expression. (doi:10.1677/jme.0.0060111) General and Comparative Endocrinology 145 197–207. (doi:10.1016/ Kjesbu OS 1989 The spawning activity of cod, Gadus morhua L. Journal of j.ygcen.2005.08.010) Fish Biology 34 195–206. (doi:10.1111/j.1095-8649.1989.tb03302.x) Presslauer C, Nagasawa K, Fernandes JMO & Babiak I 2012 Expression of Kjesbu OS 1994 Time of start of spawning in Atlantic cod (Gadus morhua) vasa and nanos3 during primordial germ cell formation and migration females in relation to vitellogenic oocyte diameter, temperature, fish in Atlantic cod (Gadus morhua L.). Theriogenology 78 1262–1277. Journal of Molecular Endocrinology length and condition. Journal of Fish Biology 45 719–735. (doi:10.1111/ (doi:10.1016/j.theriogenology.2012.05.022) j.1095-8649.1994.tb00939.x) Rocha A, Zanuy S, Carrillo M & Gomez A 2009 Seasonal changes in gonadal Kobayashi T, Nakamura M, Kajiura-Kobayashi H, Young G & Nagahama Y expression of gonadotropin receptors, steroidogenic acute regulatory 1998 Immunolocalization of steroidogenic enzymes (P450scc, protein and steroidogenic enzymes in the European sea bass. General b P450c17, P450arom, and 3 -HSD) in immature and mature testes of and Comparative Endocrinology 162 265–275. (doi:10.1016/j.ygcen. rainbow trout (Oncorhynchus mykiss). Cell and Tissue Research 292 2009.03.023) 573–577. (doi:10.1007/s004410051086) Rodriguez-Mari A, Yan YL, Bremiller RA, Wilson C, Canestro C & Korach KS 1994 Insights from the study of animals lacking functional Postlethwait JH 2005 Characterization and expression pattern of estrogen receptor. Science 266 1524–1527. (doi:10.1126/science. zebrafish anti-Mu¨llerian hormone (amh) relative to sox9a, sox9b, and 7985022) cyp19a1a, during gonad development. Gene Expression Patterns 5 Krege JH, Hodgin JB, Couse JF, Enmark E, Warner M, Mahler JF, Sar M, 655–667. (doi:10.1016/j.modgep.2005.02.008) Korach KS, Gustafsson JA & Smithies O 1998 Generation and Ruksana S, Pandit NP & Nakamura M 2010 Efficacy of exemestane, a new reproductive phenotypes of mice lacking estrogen receptor b. PNAS 95 generation of aromatase inhibitor, on sex differentiation in a 15677–15682. (doi:10.1073/pnas.95.26.15677) gonochoristic fish. Comparative Biochemistry and Physiology. Toxicology & Kula K 1988 Induction of precocious maturation of spermatogenesis in Pharmacology 152 69–74. (doi:10.1016/j.cbpc.2010.02.014) infant rats by human menopausal gonadotropin and inhibition by Sampath Kumar R, Ijiri S & Trant JM 2000 Changes in the expression of simultaneous administration of gonadotropins and testosterone. genes encoding steroidogenic enzymes in the channel catfish (Ictalurus Endocrinology 122 34–39. (doi:10.1210/endo-122-1-34) punctatus) ovary throughout a reproductive cycle. Biology of Reproduc- Lassiter CS, Kelley B & Linney E 2002 Genomic structure and embryonic tion 63 1676–1682. (doi:10.1095/biolreprod63.6.1676) expression of estrogen receptor beta a (ERba) in zebrafish (Danio rerio). Sawatari E, Shikina S, Takeuchi T & Yoshizaki G 2007 A novel transforming Gene 299 141–151. (doi:10.1016/S0378-1119(02)01050-8) growth factor-b superfamily member expressed in gonadal somatic cells Lo Nostro FL, Antoneli FN, Quagio-Grassiotto I & Guerrero GA 2004 enhances primordial germ cell and spermatogonial proliferation in Testicular interstitial cells, and steroidogenic detection in the proto- rainbow trout (Oncorhynchus mykiss). Developmental Biology 301 gynous fish, Synbranchus marmoratus (Teleostei, Synbranchidae). 266–275. (doi:10.1016/j.ydbio.2006.10.001) Tissue & Cell 36 221–231. (doi:10.1016/j.tice.2004.03.001) Sharpe RM, Atanassova N, McKinnell C, Parte P, Turner KJ, Fisher JS, Kerr JB, Lubahn DB, Moyer JS, Golding TS, Couse JF, Korach KS & Smithies O 1993 Groome NP, Macpherson S, Millar MR et al. 1998 Abnormalities in Alteration of reproductive function but not prenatal sexual functional development of the Sertoli cells in rats treated neonatally

with diethylstilbestrol: a possible role for estrogens in Sertoli cell Tchoudakova A & Callard GV 1998 Identification of multiple CYP19 genes development. Biology of Reproduction 59 1084–1094. (doi:10.1095/ encoding different cytochrome P450 aromatase isozymes in brain and biolreprod59.5.1084) ovary. Endocrinology 139 2179–2189. (doi:10.1210/endo.139.4.5899) Shi D, Wen HS, He F, Li JF, Yang YP, Chen CF, Zhang JR, Chen XY, Jin GX, Tosaka R, Todo T, Kazeto Y, Mark Lokman P, Ijiri S, Adachi S & Yamauchi K Shi B et al. 2011 The physiology functions of estrogen receptor a (ERa) 2010 Expression of androgen receptor mRNA in the ovary of Japanese in reproduction cycle of ovoviviparous black rockfish, Sebastes eel, Anguilla japonica, during artificially induced ovarian development. schlegeli Hilgendorf. Steroids 76 1597–1608. (doi:10.1016/j.steroids. General and Comparative Endocrinology 168 424–430. (doi:10.1016/ 2011.09.014) j.ygcen.2010.05.005) Shi Y, Liu X, Zhang H, Zhang Y, Lu D & Lin H 2012 Molecular identification Tveiten H, Frantzen M, Scott AM & Scott AP 2010 Synthesis of b of an androgen receptor and its changes in mRNA levels during 17,20 ,21-trihydroxypregn-4-en-3-one by ovaries of reproductively Gadus morhua Journal of Fish Biology 17a-methyltestosterone-induced sex reversal in the orange-spotted mature Atlantic cod . 77 33–53. (doi:10.1111/j.1095-8649.2010.02655.x) grouper Epinephelus coioides. Comparative Biochemistry and Physiology. Wu C, Patino R, Davis KB & Chang X 2001 Localization of estrogen receptor Part B, Biochemistry & Molecular Biology 163 43–50. (doi:10.1016/j.cbpb. a and b RNA in germinal and nongerminal epithelia of the channel 2012.04.011) catfish testis. General and Comparative Endocrinology 124 12–20. Shibata Y, Paul-Prasanth B, Suzuki A, Usami T, Nakamoto M, Matsuda M & (doi:10.1006/gcen.2001.7668) Nagahama Y 2010 Expression of gonadal soma derived factor (GSDF) is Zhu P, Zhang Y, Zhuo Q, Lu D, Huang J, Liu X & Lin H 2008 Discovery spatially and temporally correlated with early testicular differentiation of four estrogen receptors and their expression profiles during testis in medaka. Gene Expression Patterns 10 283–289. (doi:10.1016/j.gep. recrudescence in male Spinibarbus denticulatus. General and Comparative 2010.06.005) Endocrinology 156 265–276. (doi:10.1016/j.ygcen.2008.01.017)

Received in ﬁnal form 12 March 2014 Accepted 17 March 2014 Accepted Preprint published online 18 March 2014 Journal of Molecular Endocrinology