Fundulus Heteroclitus): Diverences in Populations from Polluted and Reference Environments

General and Comparative Endocrinology 150 (2007) 174–188 www.elsevier.com/locate/ygcen Communications in Genomics and Proteomics Cloning of three estrogen receptors (ER) from killiWsh (Fundulus heteroclitus): DiVerences in populations from polluted and reference environments

Sarah R. Greytak, Gloria V. Callard ¤

Department of Biology, Boston University, Boston, MA 02215, USA

Received 13 March 2006; revised 10 June 2006; accepted 26 July 2006 Available online 8 September 2006

Abstract

Epidemiological, wildlife, and laboratory studies support the hypothesis that chemicals released into the environment through anthro- pogenic activities are responsible for abnormalities of reproduction and development. Although the New Bedford Harbor (NBH) killiWsh population has survived and reproduced successfully for >50 yr (»20 generations), Wsh have high body burdens of the major NBH contaminants (polychlorinated biphenyls); elevated levels of P450 aromatase B and vitellogenin mRNA (markers of estrogen eVect); and evidence of endocrine disruption. To investigate possible adaptive changes in the estrogen response system of NBH killiWsh, we cloned the estrogen receptors (ER) from killiWsh populations resident in NBH and a relatively unpolluted reference site (Scorton Creek MA, SC). ER, -a, and -b cDNAs encoding full-length polypeptides of 620, 543, and 672 amino acids, respectively, were identiWed. Each ER subtype had multiple splice variants, single nucleotide polymorphisms (SNPs), and a characteristic tissue distribution and developmental proWle. As measured by real-time quantitative RT-PCR analysis, the overall tissue distribution of each ER was similar in NBH and SC Wsh, implying that the regulatory pathways which maintain tissue-speciWc expression are largely unchanged by long term pollutant exposure. Nonetheless, a striking diVerence was seen in the quantity of mRNA of the estrogen-inducible gene ER, which was signiWcantly lower in brain, liver and ovaries of reproductively active NBH as compared to SC females. Paradoxically, despite the “estrogenic” NBH environment, ER mRNA levels did not diVer in reproductively inactive NBH and SC females, or in males at the two sites at any time of year. We interpret results in NBH Wsh as due in part to a deWcit of circulating estrogen, and in part to pollutant-mediated hyporesponsive- ness of the ER gene. In marked contrast to adult Wsh, ER was elevated »5-fold in NBH as compared to SC embryos/larvae, perhaps indicative of estrogenic chemicals in yolk. We conclude that contaminants in the NBH environment impact molecular components of the estrogen signaling pathways in resident killiWsh populations. Whether these changes are transient or heritable requires further study. © 2006 Elsevier Inc. All rights reserved.

Keywords: Estrogen receptor; ER; PCBs; KilliWsh; New Bedford Harbor Superfund Site; Molecular markers; Endocrine disruption; Reproduction

1. Introduction in the environment, have the potential to disrupt critical hormone-regulated processes of reproduction and develop- In addition to their toxic and teratogenic eVects, which ment, even at low concentrations and transient exposures generally result from high dose exposures, halogenated aro- (reviewed in Phillips and Harrison, 1999; Sumpter and matic hydrocarbons (HAH) including polychlorinated dib- Johnson, 2005). Although the “endocrine disruptor hypoth- enzodioxins (PCDDs), dibenzofurans (PCDFs), and esis” is still a matter of debate (Safe, 2005), strong theoreti- polychlorinated biphenyls (PCBs), pesticides, plasticizers, cal support comes from laboratory studies, which show drugs, heavy metals, and other chemicals that accumulate that these chemicals can mimic or block hormone-receptor interactions or otherwise aVect signal transduction path- * Corresponding author. Fax: +1 617 353 2923. ways (Crisp et al., 1998; Phillips and Harrison, 1999). Addi- E-mail address: [email protected] (G.V. Callard). tional evidence is based on epidemiological and wildlife

0016-6480/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ygcen.2006.07.017 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 175 studies which report an increased incidence, or geographi- (Zhu and Conney, 1998) or production (Baba et al., 2005) cal clusters, of reproductive and developmental abnormali- of endogenous estrogens. In addition to dioxin-like PCBs, a ties in the natural environment (Crisp et al., 1998; Phillips subset of PCB congeners and in vivo metabolites can aVect and Harrison, 1999; Mills and Chichester, 2005). Nonethe- reproductive processes by virtue of their ability to bind to less, neither laboratory nor observational approaches per se ER directly and exert agonist or antagonist eVects, depend- reveal the biological consequences of chronic, multigenera- ing on their aYnity and concentration relative to endoge- tional exposures to complex chemical mixtures, and only nous ligand (Connor et al., 1997; Lind et al., 1999). rarely have physiological or genetic mechanisms that per- Our knowledge of the biological activities of NBH pollu- mit or promote survival and reproductive success of tants is far from complete; however, results of our earlier exposed populations been investigated. study show that the NBH environment is “estrogenic”, as To address these issues, killiWsh (Fundulus heteroclitus) measured by increased mRNA levels of hepatic vitellogenin at the New Bedford Harbor MA (NBH) Superfund site, (vtg) and brain P450 aromatase B (P450aromB) in NBH when compared to killiWsh at a relatively unpolluted site males and reproductively inactive females (Greytak et al., (Scorton Creek MA, SC), are an extremely valuable 2005). In the same study, we obtained evidence of endocrine resource. NBH is one of the most extensively PCB-contam- disruption. In reproductively active NBH females, gonado- inated sites in the United States (Weaver, 1983). The history somatic index, hepatosomatic index, plasma estrogen, and of pollution in NBH indicates that the resident killiWsh expression of hepatic vtg and brain P450aromB were sig- population has been exposed to complex mixtures of PCBs, niWcantly lower than in NBH females (Greytak et al., 2005). heavy metals, and other pollutants for >50 yr (»20 genera- The data are consistent with an impact of NBH contami- tions). Despite high body burdens of contaminants (Black nants at one or more levels of the hypothalamic–pituitary– et al., 1998), NBH killiWsh continue to survive and repro- gonadal (HPG) axis. Nonetheless, the seasonal onset and duce (Black et al., 1998). Indeed, killiWsh are highly abun- regression of reproductive condition occur at essentially the dant in NBH (D. Nacci, personal communication). Survival same time in SC and NBH Wsh (Greytak et al., 2005); males of the NBH killiWsh population has been ascribed to and females display a semilunar spawning cycle at both acquired resistance to the toxic eVects of dioxin and dioxin- sites (Black et al., 1998); and viable embryos can be like PCBs (Nacci et al., 1999; Bello et al., 2001). Toxicity of obtained after natural spawning or in vitro fertilization dioxin-like chemicals (DLC) is mediated mainly by arylhy- (Black et al., 1998; Powell et al., 2000). drocarbon receptors (AHR) (Schmidt and BradWeld, 1996). By analogy to acquired resistance to AHR-mediated Although the mechanism of resistance is unclear, compari- toxicity, we postulated that changes in ER-mediated signal- son of NBH and SC killiWsh shows diVerences in tissue-spe- ing pathways of NBH killiWsh could attenuate or neutralize ciWc expression and induction of AHR1 and CYP1A1 eVects of excess environmental estrogen, and thereby con- mRNAs (Powell et al., 2000; Bello et al., 2001), and site-spe- tribute to the reproductive success of the NBH population. ciWc single nucleotide polymorphisms (SNPs) in AHR1 In mammals, the diversity and complexity of estrogen sig- (Hahn et al., 2004). Some but not all of these traits are heri- naling systems is accomplished by two distinct ER genes ( table, indicating that NBH killiWsh reXect current as well as and ) and multiple splice variants (Hall et al., 2001). By historical population exposures to PCBs and other contam- contrast, teleost Wsh have one ER and two ER-subtypes inants. (Hawkins et al., 2000; Menuet et al., 2002; Sabo-Attwood AHR are expressed in many reproductively relevant tis- et al., 2004), termed ERa and ERb (Hawkins and sues (Hahn, 1998) and, in addition to their role in toxicity, Thomas, 2004). Each Wsh ER has one or more splice vari- activation of AHR can impact reproduction by altering the ants (Tchoudakova et al., 1999; Pakdel et al., 2000; Menuet estrogen signaling pathway (reviewed in Safe et al., 2000). et al., 2001) and a characteristic ligand binding proWle, tis- Estrogen receptors (ER) are ligand-activated transcription sue distribution, and regulation (Hawkins and Thomas, factors, which bind to estrogen response elements (ERE) in 2004; Menuet et al., 2004). Interestingly, the teleostean ER regulatory regions of target genes. Like all members of the gene itself is highly estrogen inducible, at least in liver, steroid/thyroid hormone /retinoic acid receptor superfam- indicative of a physiologically relevant autoregulatory ily, ER have a prototypical nuclear receptor structure with mechanism (Urushitani et al., 2003; Menuet et al., 2004). In conserved functional domains designated A–F (AB, activa- theory, therefore, chronic multigenerational exposure to tion function1; C, DNA binding domain; D, hinge region; environmental estrogen could alter the estrogen response E, ligand binding domain; F, activation function 2) (Krust system by changing the amount, ratio, ligand binding or et al., 1986). In mammals, AHR ligands can interfere with regulatory characteristics of the diVerent ER forms. To estrogen signaling by altering ER mRNA or protein levels investigate possible adaptive mechanisms in estrogen sig- (reviewed in Tian et al., 1998; Safe et al., 2000); interacting naling pathways, we cloned and characterized all three ER with DNA to competitively inhibit ER binding to EREs in both NBH and SC killiWsh, and used real-time PCR to (reviewed in Safe et al., 2000); interacting with unbound ER compare expression in tissues of males and females during protein to potentiate transcription through EREs or atten- seasonal periods of reproductive activity and inactivity, and uating the function of bound ER (Ohtake et al., 2003). Fur- in developmentally staged embryos and larvae at the two ther, activation of the AHR pathway can alter the clearance sites. During the course of our study, a killiWsh ER cDNA 176 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188

Table 1 Oligonucleotide primers Primer Target Sequence nts E1 ER; ATHCARGGDCAYAAYGRCTAYATSTG F 672–698 ERa F 1023–1049 E2 ER; AANCCWGGVABYTYTTTBGCCCAGBYRATCAT R 1101–1133 ERa R 1430–1462 A1 ER ACCGCCGATACTCTGTTCCCGTCAA F 1001–1026 A2 ER CGACCCTACACCGAGGTCACCATGA F 1035–1060 A3 ER GTGACCTCGGTGTAGGGTCGGCTAAG R 1029–1055 A4 ER CCCGAGGATGATTCATGTATAAAG F 2–25 A5 ER TAAACTGCATGTGCTCATTCATAA R 1900–1924 A6 ER GCATGCTCAAGCTCAAACC F 1339–1358 A7 ER GTGCCGGTGCAGAAAGA R 1410–1427 BA1 ERa TCATGAAGGACATAAAGAGGCCGCTGA F 1386–1412 BA2 ERa TGTCGCTCACAAACCTGGCGGATAA F 1430–1455 BA3 ERa CACGTCCAGAGTGAGAGAACTCAAGC F 1692–1718 BA4 ERa AGAACTCAAGCTGACGAGGGAGGAAT F 1707–1733 BA5 ERa TAGTGCATTGATTGGTTGCTGGACA R 1087–1111 BA6 ERa CCCCTGACCGGAGTAGCTGAAAATG R 720–745 BA7 ERa ATGGTCGGGCTGTAGAAAGGGATGG R 698–723 BA8 ERa GATGTCGTAATTGTCGGGTTTC F 461–483 BA9 ERa GAAAGGCATCAGTCCTCCTGT R 2220–2241 BA10 ERa GTGGTATTCGCAAGGAGCGT F 1184–1204 BA11 ERa GGTGTGGTTAACACCTTGGGTC R 1268–1290 BB1 ERb GTCTGCATCCCCTCTCCGTA F 324–344 BB2 ERb GCTTTTACGTCGGTTCTTGTCTAT R 847–870 BB3 ERb TCATGTGAAGGTTGCAAGGCTTTTT F 762–787 BB4 ERb TACATTTGCCCCGCCACAAATCAAT F 816–841 BB5 ERb CGTAGTGGTACCCAGAGGCGTAGTC R 729–754 BB6 ERb TCATTGTGTCCCTGGATACTCCTTTT R 789–815 BB7 ERb TGCAGTAGAGCTGCAACATCAT F 122–144 BB8 ERb TTAATCTTCTTCCAAGCTGGTTTC R 2175–2199 BB9 ERb TGCACCCAACCTCTTGTCTACA F 522–544 BB10 ERb TTCTTCGGATTTCTTCCCCAG R 630–651 N1 Actin GGCCAACAGGGAGAAGATGACCCAGAT ND N2 Actin GGATTCCGCAGGACTCCATTCCGA ND The position of targeted sequences and orientation is based on the newly isolated killiWsh ER, a, and b cDNAs (GenBank Nos. AY571785, AY570922, and AY570923, respectively). ND, not determined. was isolated and characterized by another laboratory (Uru- designed using Primer3 (ER) or Primer Express (ERa, -b) (Applied shitani et al., 2003). Biosystems, Foster City CA). Real-time PCR primers were chosen to target sequences that were found to be identical in SC and NBH Wsh. Primer speciWcity was validated using authentic ER-containing plasmids and tis- 2. Materials and methods sue samples. Oligonucleotides N1-2 used for real time PCR ampliWcation of killiWsh actin were previously reported (Garcia-Reyero et al., 2004; 2.1. Collection sites Greytak et al., 2005).

NBH, is on Buzzard’s Bay approximately 55 miles south of Boston and 2.3. Fish and tissue sampling has been described previously (Nacci et al., 1999; Bello et al., 2001; Grey- tak et al., 2005). SC is located on Cape Cod Bay in Sandwich, MA, approx- KilliWsh in this study were those described in our earlier report, which imately 30 miles northeast of NBH and in the same climatic zone. details their physiological and reproductive status as determined by size, Although we cannot rule out diVerences in microclimate at the two sites, age, gonadosomatic index (GSI), hepatosomatic index (HSI), plasma this collection site has previously been used as a relatively unpolluted ref- androgen, estrogen, and vitellogenin, expressed levels of P450 aromatase erence location for killiWsh research (Bello et al., 2001; Greytak et al., B, P450 aromatase A, and vitellogenin mRNAs, and gonadal, and hepatic 2005). histology. (Greytak et al., 2005). Further, in our previous study using the above markers we conWrmed that reproductive seasonality was identical in 2.2. Oligonucleotides these two populations (Greytak et al., 2005). BrieXy, adult male and female SC Wsh (12, 6 each) were collected monthly between May and October Oligonucleotides (Invitrogen Corp., Carlsbad, CA) used as PCR prim- using minnow traps. Within 12 h of SC collections, Wsh of the same size ers are listed in Table 1. Oligonucleotides E1 and E2 are degenerate prim- range (6.5–8.6 cm standard length) were obtained from NBH in collabora- ers based on conserved nucleotides in vertebrate ER (Tchoudakova et al., tion with the EPA Atlantic Ecology Division, Narragansett RI. Our previ- 1999). Gene-speciWc primers A1–A5 (ER), BA1–BA9 (ERa), and BB1– ous study showed that Wsh of this size were »2–3 yr of age at both sites BB8 (ERb) were based on our initial sequence data and designed using (Greytak et al., 2005). Fish were anesthetized with 0.06% ethyl 3-amino- the Primer3 program (Rozen and Skaletsky, 2000). Oligonucleotides A6– benzoate methanesulfonate salt (MS-222, Sigma–Aldrich, St. Louis, MO) A7, BA10–BA11, and BB9–BB10 were real time quantitative PCR primers and killed by decapitation. For RNA analysis, brain, eye, gill, gonad, S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 177 spleen, liver, heart, gut, and muscle were immediately frozen on dry ice and was analyzed in triplicate. AmpliWcation was performed under default stored at ¡70 °C. conditions with the addition of a denaturing step to conWrm the presence of a single amplicon. Data were analyzed using the Q-gene package and 2.4. Egg and embryo sampling normalized to actin after correcting for diVerences in ampliWcation eYciency as recommended in the Q-gene package (Simon, 2003). Gravid females were collected the day before predicted natural spawning (at the new or full-moon) and immediately stripped to obtain unfertil- 2.8. Sequence comparisons and phylogenetic analysis ized eggs which were quick frozen. To obtain embryos for the developmental series, additional male and female Wsh were kept for 2 Consensus nucleotide sequences of newly isolated killiWsh ER, -a, weeks in Xow-through seawater tanks at the National Health and Envi- and -b and three variant ER forms ERx, -ax, and -bx were entered ronmental Research Laboratory (Atlantic Ecology Division, US EPA) as into Genbank and given Accession Nos. AY571785, AY570922, described previously (Nacci et al., 1999). The tanks were maintained at AY570923, DQ413179, DQ413180, and DQ413181, respectively. The ambient temperature and light cycles. The Wsh were fed three times daily a translation start site and open reading frame were determined using diet consisting of Tetramin® Wsh Xakes and Biodry® 1000 pellet food (Bio- ATGpr (Salamov et al., 1998). The three killiWsh ER polypeptide products, Warrington, Oregon) supplemented with krill and Artemia sp. sequences together with previously reported vertebrate ER (for GenBank Naturally spawned fertilized eggs were collected in baskets and then main- accession numbers, see legend to Fig. 2), were aligned using ClustalX tained in petri dishes in 35 ppt saltwater at room temperature on a rotary 1.81(Chenna et al., 2003) and a Gonnet scoring matrix. Using all charac- shaker until hatch (»10–12 dpf under these conditions). Post-hatch larvae ters and default parameters, neighbor-joining and maximum likelihood were transferred to glass beakers and fed Artemia sp. daily. Embryos and trees were constructed in Phylip 3.6 (Felsenstein, 1993). To conWrm larvae were killed by quick freezing on dry ice. topology, bootstrapping was performed and values were recorded as a percentage of times out of 1000 that a node was recovered. 2.5. RNA isolation and reverse transcription 2.9. Statistics RNA was extracted and reverse transcribed as described previously (Greytak et al., 2005). BrieXy, RNA was extracted from frozen tissues The number of independent samples analyzed for each gene, tissue-type, using Trizol (Sigma–Aldrich), treated with DNAseI (Roche Diagnostics stage of development, and collection site is indicated in results and or Wgure Indianapolis, IN), and size-separated on 1% agarose gels to assess quality. legends. Statistical analyses were performed using the SigmaStat 2.0 package Yield was determined spectrophotometrically. For use in cloning reactions (Jandell ScientiWc). Data sets were Wrst tested for normality. Data were ana- and real-time PCR, cDNA was synthesized from 3 g total RNA using lyzed by ANOVA with all pairwise comparisons performed by the Tukey Superscript II (Invitrogen) and Oligo (dT) (Promega Corp. Madison, WI). Test or by one-way ANOVA on RANKS as indicated in results and legends.

2.6. Cloning strategy 3. Results A stepwise PCR cloning strategy similar to that described for the killi- Wsh aromatases (Greytak et al., 2005) was employed to isolate ER cDNAs 3.1. Isolation of killiWsh ER cDNA of SC Wsh. Initial fragments of ER and ERa were ampliWed from liver and ovarian cDNA using degenerate primers E1 and E2 and previously The ER cDNA was cloned from the ovary and liver of described conditions (Tchoudakova et al., 1999). These fragments were SC killiWsh in three overlapping segments. An initial fragment extended using the Generacer kit (Invitrogen) in conjunction with primers W A1–A3 () and BA1–BA7 (a) and liver cDNA. PCR products were puri- of 461bp was ampli ed using RT-PCR with primers E1/E2. Wed from a 1% agarose gel using GeneCleanII kit (Bio101 Inc), subcloned This fragment was extended through the 5Ј and 3Ј UTR in into pGEM-T Easy vector (Promega), and sequenced (Macrogen Inc., RACE reactions with gene-speciWc primers A1–A3, yielding Seoul Korea). An initial fragment of ERb was obtained by amplifying products of 1055 and 906bp, respectively. Alignments showed liver RNA with primers BB1 and BB2 with the following conditions: 94 °C only 4 nucleotide diVerences (nt 369, 917, 1392, and 1817) /5 min; 35 cycles of 94 °C/ 30 s, 57 °C/ 30 s, 72 °C /1 min, and 72 °C/10 min. The remainder of the ERb cDNA was obtained from liver RNA using when our ER sequence was compared to the ER cDNA the Generacer kit in conjunction with primers BB3–BB6. reported while our study was in progress (Urushitani et al., In order to verify sequence data and identify possible polymorphisms 2003). The assembled fragments resulted in a 1941 bp cDNA within the SC population, six end-to-end PCR reactions each for ER, -a, with an 1860 bp open reading frame (ORF) that encoded a and - b were carried out using primers A4/5, BA8/9, and BB7/8 and liver full-length ER peptide with two potential translational start cDNA. The reaction products were separately cloned and sequenced. To diVerentiate PCR error from polymorphisms, a second independent PCR sites resulting in protein sequences of 620 aa or 573 aa. The cloning step was performed for each cDNA in which a nucleotide change size of the corresponding 5Ј UTRs were 15 and 156bp, was identiWed. To assess possible site-dependent diVerences in the coding respectively. Of the 6 SC and 7 NBH Wsh from which regions of the ER cDNAs of SC and NBH Wsh, liver RNAs from seven sequence data was obtained, 8 single nucleotide polymor- (ER) or four (ERa and ERb) NBH Wsh were individually reverse tran- W W phisms (SNP’s) were identi ed, only one of which was site- scribed and ampli ed. Products were then cloned and sequenced as W described above. speci c (H494Q) (Table 2). Of the 8 SNP’s, 5 resulted in no change in protein sequence. Two of the substitutions that 2.7. Real-time PCR analysis aVected protein sequence (nt 507 and 851) were non-conservative, resulting in the following amino acid changes: M164T, Real-time PCR was performed using an ABI PRISM 7900 (Applied and I278T. An additional substitution (nt 987) resulted in a Biosystems) instrument. Reverse transcribed RNA, prepared as described conservative change in sequence L324F. In addition, 2 of 7 above, was diluted Wvefold. Primers were A6/7, BA10/11, and BB9/10 (ER, -a, and -b); and N1/2 (actin). The Wnal reaction contained 50% individuals from NBH were found to have a SNP (nt 1497) Sybr Green Master Mix (Applied Biosystems), 1 M of each primer resulting in H494Q within the E domain (Fig. 1). This substi- (0.5 M for actin), and an aliquot (3%) of the RT reaction. Each sample tution was not found in any of the 6 SC Wsh screened. In addi- 178 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188

Table 2 Position and frequency of identiWed single nucleotide polymorphisms (SNPs), splice variants, insertions or deletions of each ER subtype identiWed in SC and NBH killiWsh ER subtype nt Substitution AA substitution Domain Population identiWed Sequence changes that alter coding 507T > C M164T AB SC + NBH 851 T > C I278T D SC + NBH 987C > T L324F E SC + NBH 1497T > G H494Q E NBH (2 of 7) 1425–1755 Deletion partial EF domains EF SC + NBH a 752T > C F84S AB SC + NBH 866A > G K122R AB SC + NBH 1373C > T T291R E SC + NBH 1728 S408 3Ј alternate truncated product E SC + NBH b 232-235 Del E19 AB SC + NBH 605–608 Insert S143 AB SC + NBH 991 G > A T273A D NBH (2 of 4) 1545 S456 3Ј alternate truncated product E SC + NBH Sequence changes that do not alter coding 231C > T A72 AB SC + NBH 368T > C F117 AB SC + NBH 1107C > T I364 E SC + NBH 1164C > T L383 E SC + NBH 1167G > C L384 E SC + NBH a 802G > C S99 AB NBH (2 of 4) 808T > C A101 AB NBH(1 of 4) 813C > T A102 AB SC + NBH 1165G > A G222 D NBH(1 of 4) b166G>A5ЈUTR NA NBH (1 of 4) 368C > T H64 AB NBH (1 of 4) 501T > C N107 AB NBH (1 of 4) 576G > A E133 AB NBH (2 of 4) 655A > C A159 AB SC + NBH 674C > G S166 AB SC + NBH 687G > T S169 AB NBH (3 of 4) 726C > T C183 C SC + NBH 792A > G K205 C SC + NBH 959A > G R261 D NBH (2 of 4) 987A > T R269 D NBH(1 of 4) 1002G > A R275 D NBH (2 of 4) 1056A > G E293 D SC + NBH 1379G > A K401 E SC + NBH 1389C > T S405 E SC + NBH 1404G > A K408 E SC + NBH 1431A > G V418 E SC + NBH 1665C > G S495 E NBH (2 of 4) 1898C > A A574 F SC + NBH 1898C > A A574 F SC + NBH 2055T > A I626 F NBH(1 of 4) cDNAs encoding full-length ER, -a, and -b peptides were cloned and sequenced from SC (n D 6) and NBH (n D 6, ER; n D 4, ERa and -b) killiWsh, as described in text. tion to the major product of 1941bp, a minor product of fragment of approximately 439 bp was ampliWed using RT- »1600 bp was produced in all of the end to end PCR reac- PCR with primers E1/E2. This fragment was extended tions of both SC and NBH Wsh. The 1600bp product was to include both UTRs with RACE resulting in 5Ј and 3Ј sequenced and found to contain a deletion (nt 1425 to 1755) products of 1754 and 1087, respectively. The assembled which aVected the E (Fig. 1) and F domains. This splice vari- sequence resulted in a coding region of 1659 nt with a 501 nt ant has been termed ERx to correspond to variants isolated 5Ј UTR and a 3Ј UTR of 983 nt. Translation of the cDNA previously in human (Ogawa et al., 1998). resulted in a 553 aa sequence. Among the 6 SC Wsh from which we obtained sequence data, 4 single nucleotide poly- 3.2. Isolation of killiWsh ERa cDNA morphisms were identiWed (Table 2). One of the substitutions resulted in no change in the coding sequence (nt 813), The ERa cDNA was cloned in Wve overlapping frag- another (nt 866) resulted in a conservative change (K122R). ments from the ovary and liver of SC killiWsh. An initial Two additional substitutions (nt 752 and 1373) resulted in S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 179

Fig. 1. Alignment of the LBD domain of ER from humans and killiWsh showing normal and variant forms and SNPs within this region. Deduced amino acid sequences of the LBD of killiWsh ER, a, and b (GenBank Nos. AY571785, AY570922, and AY570923, respectively) and three truncated transcripts: ERx, ERax, and ERbx (GenBank Nos. DQ413179, DQ413180, and DQ413181, respectively) isolated in this study were aligned with human ER and ER (GenBank Nos. NP_000116 and NP_001428, respectively), and human ERcx (GenBank No. BAA31966). Helices are designated by boxes and amino acid residues identical to the human ER are replaced by a dash. Residues important for estradiol binding (¤) and those that form hydrogen bonds with estradiol (᭹) are based on the human ER (Brzozowski et al., 1997; Pike et al., 1999). Single nucleotide polymorphisms (ER L324F, ER H494Q, and ERa T291R) identiWed in this study are circled. Only ER H494Q is speciWc to NBH (Note helix 10 in LBD). Locations of the splice sites that result in the truncated ERx (black), ERax (white), ERbx (white), and human ERcx (gray) are designated with arrows. non-conservative changes (F84S and T291R) which aVect both sites. Similarly to ERa, an alternate 3Ј truncation the AB and E domains, respectively. Three SNPs that resulted in a coding sequence of 456 aa that lacked part of resulted in no change in coding sequence were identiWed the E and F domains and had an alternate 3Ј UTR. This solely in the NBH population (nts 802, 808, and 1165) form has been termed ERbx. (Table 2). An additional 3Ј truncated cDNA, which trans- lated into a protein of 408 aa was identiWed in Wsh from 3.4. Sequence comparisons and phylogenetics both populations (Fig. 1). This 3Ј truncation resulted in a partial deletion of the E and F domains and the use of an In order to further characterize the killiWsh ER, a phylo- alternate 3Ј UTR. This truncated form has been termed genetic tree was constructed using 17 previously identiWed ERax. vertebrate ER sequences for comparison (Fig. 2). Consis- tent with earlier reports, the vertebrate ER forms were 3.3. Isolation of killiWsh ERb cDNA clearly divided into and clades. In teleosts the clade was further divided into a and b subtypes. The tree The ERb cDNA was obtained from the liver of SC shows that killiWsh ER, -a, and -b have the highest killiWsh in three overlapping fragments. An initial fragment degree of similarity to the corresponding ER forms in of 500 bp was obtained by PCR with primers BB1 and BB2. largemouth bass. Of interest, in the shark only one ER form This fragment was then extended by RACE reactions with has been found and is more closely related to the ER primers BB3–BB6, resulting in 5Ј and 3Ј products of 754 forms in tetrapods than to the teleost ER. and 1436 nt, respectively. The assembled cDNA was In agreement with the tree, sequence alignments 2252 bp with an ORF of 2019 bp. The isolated 5Ј and 3Ј showed that killiWsh ER, -a, and -b have 74, 80, and UTRs were 177 and 55 nt, respectively. The translation of 79% identity, respectively, when compared to their coun- the nucleotide sequence resulted in a peptide of 673 aa. Of terparts in largemouth bass, but only 58, 62, and 54% the 6 SC Wsh from which ERb sequence data was identity, respectively, when compared to the three obtained, 11 SNPs were identiWed but none resulted in a zebraWsh ER. By contrast, the two killiWsh ER subtypes change in coding sequence (Table 2). An additional 10 had only 37% overall identity to the killiWsh ER and SNPs that did not alter coding sequence were found in only 48% identity to each other, indicative of a greater NBH killiWsh (Table 2). A single nucleotide substitution at degree of conservation of orthologs than paralogs. When nt 991 (T273A) in the hinge region was identiWed in 2 of 4 diVerent functional domains within either the or Wsh from NBH. In addition to the SNPs, a 3 bp codon dele- clades were compared, the highest degree of identity was tion in the AF-1 domain (nt 232–235, aa E19) was identiWed found in the DBD (>80%) and LBD/AF2 domains (>43– in 4 Wsh, including those from both sites. A 3 bp codon 74%). By contrast, the AF1 and hinge regions showed insertion in the AF-1 domain (nt 605–608, aa S143) was only very low levels of identity (10–40%) across the three also identiWed in Wve individuals representing Wsh from killiWsh ER. 180 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188

Fig. 2. Phylogenetic tree of vertebrate ER. Deduced amino acid sequences of killiWsh ER, -a, and -b (GenBank Nos. AY571785, AY570922, and AY570923, respectively) were aligned with ER of representative vertebrates, and analyzed to produce a consensus neighbor-joining (shown) and maximum likelihood (not shown) tree with lamprey as the designated outgroup. GenBank accession numbers are as follows: ER, NP_990514 (chicken), NP_000116 (human), NP_036821 (rat), AAG16713 (croaker), AAG44622 (bass), AAK16740 (zebraWsh); ER, NP_990125 (chicken), NP_001428 (human), NP_036886 (rat), AAK57823 (dogWsh); ERa, AAG16712 (croaker ), AAO39211 (bass ), AAK16742 (zebraWsh), ERb AAG16711 (croaker ), AAO39210 (bass ), AAK16741 (zebraWsh B), and AAK20929 (lamprey). Numbers at nodes indicate percentage of times a clade was recovered from 1000 replicates using neighbor joining and maximum likelihood (in brackets) methods. Branch lengths are proportional to the amino acid substitutions per amino acid position.

3.5. Tissue distribution times higher than ERb in spleen and heart. Additionally, although ER, -a, and -b mRNA were most abundant Real-time PCR was applied to determine the tissue dis- in the liver and gonads, sex diVerences in the mRNA lev- tribution of ER, -a, and -b mRNAs in reproductively els were gene- and tissue-speciWc. Thus, in liver, ER active SC and NBH killiWsh. With the exception of the mRNA was signiWcantly higher in females than in males, gonads and liver, which were pooled by sex, tissues in this but ERa and -b were the same in both sexes. While in experiment were pooled from an equivalent number of testis ERa was the predominant form, all three ER males and females (see M&M and legend to Fig. 3). As mRNAs were present at similar levels in the ovary. Inter- shown in SC Wsh (Fig. 3), all three ER mRNAs were pres- estingly, although the maximum mRNA level was similar ent in brain, eye, gill, ovary, testis, spleen, liver, heart, gut, for each of the three forms, the range from highest to low- and muscle, but steady-state mRNA levels were gene- and est abundance tissues was much narrower for ERa (10- tissue-speciWc: ER female liver > male liver D ovary > fold, liver vs. spleen) than for ER (1000-fold, liver vs gill) brain D testis > spleen); ERa (testis D ovary >liver> or ERb (100-fold, liver vs. spleen). Whereas the overall brain > gut > muscle D eye); ERb (ovary > liver > testis tissue distribution for each ER form in Wsh from NBH > brain > gut) (Fig. 3). Also, the ratio of the three mRNA (data not shown) was the same as in SC (Fig. 3), quantita- forms was tissue-speciWc. To illustrate, ERa was the pre- tive diVerences between the two populations were dominant isoform in eye, gill, and muscle (ratio a: or observed with certain genes, tissues, and times of year (see a:b D >50:1), whereas ER and ERa were both »50- below). S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 181

Fig. 3. Tissue distribution of ER (A), -a (B), and -b (C) mRNA in male and female SC killiWsh. Tissues were collected from SC Wsh during the period of reproductive activity (May). With the exception of liver and gonads, like-tissues were pooled from two male and two female Wsh for each RNA extract. Liver, testes and ovaries were pooled from two males (M) or two females (F) per RNA extract. Equal amounts of input RNA were analyzed by real-time PCR. Each bar represents normalized mRNA levels § SEM of four independent sample pools from SC Wsh. The same experiment was performed in May using NBH Wsh; overall tissue distribution did not diVer from SC Wsh (not shown).

3.6. ER mRNA levels in liver, gonads, and brain of son. ER levels did not diVer in SC and NBH males with reproductively active and inactive male and female killiWsh time of year. from SC and NBH Similarly, when levels of ER mRNA were analyzed in the ovary of female killiWsh by two-way ANOVA, signiW- Real-time PCR analysis of ER was applied to deter- cant site (P D <0.001; F D 34.659) and seasonal (P D <0.001; mine how mRNA levels in reproductively relevant tissues F D 82.040) eVects were observed (Fig. 4B). Ovarian ER were aVected by sex, reproductive status, and polluted vs. mRNA was lower in reproductively active NBH females unpolluted environments. As shown in Fig. 4A, when rel- than in reproductively active SC controls; however, a ative mRNA levels of ER in liver were analyzed by decrease in ovarian ER mRNA was observed in both pop- three-way ANOVA, signiWcant sex (P D <0.001; ulations during the period of reproductive inactivity. In F D 70.708), seasonal (P D <0.001; F D 35.087), and site contrast, when levels of ER mRNA in testis were analyzed (P D 0.008; F D 8.384) diVerences were obtained. The liver in male killiWsh by two-way ANOVA, no signiWcant site of reproductively active SC females had >3-fold higher (P D 0.738; F D 0.116) or seasonal (P D 0.835; F D 0.0450) ER levels than reproductively inactive SC females, or eVects were observed (Fig. 4B). reproductively active or inactive SC males. Compared to As shown in Fig. 4C, when levels of ER mRNA in the SC females, NBH females had signiWcantly lower hepatic brains of killiWsh from both populations were analyzed by ER mRNA levels, but only during the reproductive sea- three-way ANOVA, signiWcant site (P D <0.001; F D 22.877) 182 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188

A 3.7. ERa and -b mRNA in liver, gonads, and brain of reproductively active and inactive male and female killiWsh from SC and NBH

To further assess the eVects of sex, site, and season on ER levels in killiWsh, real-time PCR was applied to measure ERa and ERb mRNAs. When levels of ERa mRNA were analyzed in the liver of reproductively active and inactive male and female Wsh from both sites by three-way ANOVA, signiWcant site (P D 0.036; F D 4.935), but not seasonal (P D 0.078; F D 3.395) or sex (P D 0.832; F D 0.0461) diVerences were obtained. Similarly, results of two-way W B ANOVA of ER a mRNA levels showed no signi cant eVect of site or season in ovary (P D 0.134; F D 2.531; P D 0.556; F D 0.363, respectively) or testis (P D 0.934; F D 0.0075; P D 0.939; F D 0.00614, respectively). Further, in brain no signiWcant eVects of sex (P D 0.064; F D 3.809), season (P D 0.409; F D 0.709) or site (P D 0.157; F D 2.146) were identiWed. Three-way ANOVA of ERb mRNA levels revealed no signiWcant eVects of sex, season or site in liver (P D 0.482; F D 0.510; P D 0.703; F D 0.149; P D 0.925; F D 0.00893, respectively); ovary (P D 0.250; F D 1.474; P D 0.143; F D 2.491, respectively); testis (P D 0.773; C F D 0.0853,;P D 0.876; F D 0.0250, respectively) or brain (P D 0.943; F D 0.0052; P D 0.432; F D 0.641; P D 0.789; F D 0.0731, respectively). Because no sex or seasonal diVer- ences were obtained by three-way ANOVA of ERa or -b, data were combined and diVerences between sites retested by one-way ANOVA on RANKS (Fig. 5). No diVerences in mRNA levels of either ERa or -b were found in ovary, brain, or testis of the two populations. However, in liver, ERa, but not ERb, was signiWcantly reduced in the NBH when compared to SC killiWsh (P <0.01).

3.8. Developmental patterns Fig. 4. ER abundance in (A) liver, (B) gonad, and (C) brain of reproduc- W tively active and inactive male and female killi sh from SC and NBH, as To determine stage-related patterns of ER expression, measured by real-time PCR. Males and females were collected from SC W and NBH during seasonal periods of reproductive activity (June and July) mRNA levels in developing embryos/larvae were quanti ed and inactivity (August and September). Extracted RNA was pooled based by reverse transcription real-time PCR at intervals between on reproductive status. The pools were comprised of like tissues from two 1 and 18 dpf (Fig. 6). Two-way ANOVA separately for each (ovaries), four (liver, brain), or one (testis) Wsh per site and status. Each gene showed that stage of development signiWcantly bar represents normalized ER mRNA levels § SEM of four (ovary, liver, aVected steady-state levels of ER (P D <0.001; F D 8.084), and brain) or 5–7 (testis) samples. Data were analyzed by three-way ANOVA for site, sex, and season (see Section 3). Within each panel, bars ER a (P D <0.001; F D 28.315); and ER b (P D <0.001; with diVerent letters diVered signiWcantly (P < 0.05) as determined by the F D 24.6720), but only ER mRNA was site-related Tukey test. Note diVerences in the scale of the y-axis in each panel. (P D <0.001; F D 104.689). In SC Wsh, all three ER mRNAs increased progressively between 1 and 18 dpf, but only ER and ERb displayed peaks associated with hatching at 10– but not seasonal (P D 0.125; F D 2.533) or sex (P D 0.211; 12 dpf. In NBH W sh, ER mRNA was »5-fold higher than F D 1.657) diVerences were observed. Similar to observa- in SC Wsh at all time points after 1 dpf, and the peaks at 6 tions in liver and ovary, brain ER mRNA levels were sig- and 12 dpf were correspondingly exaggerated. niWcantly higher in reproductively active SC females than in reproductively inactive SC females or in SC males, 3.9. Unfertilized eggs regardless of reproductive condition. Nonetheless, both males and females from the NBH population had signiW- To obtain insight into the regulation of ER, -a, and -b cantly reduced brain ER levels when compared to corre- mRNAs in the earliest stages of embryogenesis, unfertilized sponding reference Wsh, but only during the period of eggs obtained from SC, and NBH Wsh were analyzed by seasonal reproductive activity. reverse transcription real-time PCR. All three mRNAs were S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 183

A B

C D

Fig. 5. ERa and ERb mRNA abundance in (A) liver; (B) ovary, (C) brain, and (D) testis of killiWsh of SC and NBH, as measured by real-time PCR. Liver, ovaries, and brain were collected, pooled and processed as described in the legend to Fig. 4, and aliquots taken for ERa and ERb analysis. When resultant data for each tissue were analyzed by three-way ANOVA, no signiWcant sex or seasonal diVerences were observed; therefore, data from reproductively active and inactive males and females were combined by tissue-type and time of year (see Section 3). Each bar represents normalized mRNA levels § SEM of sixteen (liver and brain), eight (ovary), or ten-fourteen (testis) independent biological samples. Within each panel, bars with diVerent letters diVered signiWcantly (P < 0.05) as determined on combined data by two-way ANOVA for site and tissue. Note diVerences in the scale of the y-axis in each panel. detected in unfertilized eggs of Wsh at both sites, indicating a gene duplication event prior to the divergence of teleost maternal synthesis. When data were analyzed by two-way Wsh from the tetrapod lineage, and that the two -subclades ANOVA, relative ER abundance was gene speciWc resulted from a teleost-speciWc genome duplication event (PD<0.001; FD811.782) but not site-related (PD0.339; after the divergence of the sturgeons and gars (Hoegg et al., FD0.992) (Table 3). Interestingly, levels of ERa and ERb 2004). Compared to other Wshes, killiWsh ER are most were, respectively, 10,000- and 100-fold lower in unfertilized closely related to their respective orthologs in largemouth eggs than in whole ovary, but mean ER levels were enriched bass, reXecting the close taxonomic relationship of the two in eggs. Further, although ER mRNA was signiWcantly species. lower in ovaries of NBH than of SC killiWsh, this diVerence Like ER previously isolated from Wsh (Pakdel et al., was not signiWcant in unfertilized eggs (Table 3). 2000; Menuet et al., 2004; Sabo-Attwood et al., 2004) and humans (reviewed in Herynk and Fuqua, 2004), the three 4. Discussion ER cloned in this study have multiple transcripts. In rainbow trout, two distinct ER transcripts arise from alter- To investigate possible adaptive changes in estrogen sig- nate promoter usage and splicing (Pakdel et al., 2000). naling pathways, we compared the molecular characteris- The “long form” has two possible translational start sites tics and expressed levels of ER, -a, and b in killiWsh and encodes a protein with a complete A domain, while from polluted (NBH) and nearby reference (SC) sites. the “short form” has only the downstream start site and Using a multistep PCR cloning approach and RNA gives rise to an ER with a truncated A domain (Pakdel extracts from NBH and SC Wsh, we identiWed three ER et al., 2000). Short and long ER cDNAs have also been genes in killiWsh. Based on sequence comparisons, espe- identiWed in channel catWsh (Patino et al., 2000). In sea cially within conserved functional domains, the three bass, the possibility of two transcripts was discussed but cloned ER were identiWed as ER, -a, and -b, to conform only the long forms were actually cloned (Halm et al., to the current naming of ER in Atlantic croaker and 2004). The killiWsh ER cDNA cloned in this study has zebraWsh (Lassiter et al., 2002; Hawkins and Thomas, two possible translational start sites, and in that respect, 2004). Similar to previously identiWed teleostean ER, killi- resembles the rainbow trout “long form”, but we did not Wsh ER segregate into two main clades ( and ) with a fur- Wnd evidence for a second 5Ј UTR nor were any of our ther division into two -subclades (a and b). Our identiWed transcripts lacking an A domain. Our killiWsh phylogenetic analysis is consistent with the prevailing view, ER cDNA is in agreement with the sequence described namely, that the vertebrate ER and - clades arose due to earlier (Urushitani et al., 2003). In contrast to ER, 184 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188

Fig. 6. Relative abundance of ER (A), -a (B), and -b (C) mRNA during development. Fertilized eggs were obtained by natural spawning of SC and NBH Wsh, maintained in running seawater, and collected at timed intervals in development between 1 and 18 dpf. As deWned previously (Oppenheimer, 1937; Armstrong and Child, 1965), developmental stages and major landmarks are indicated. For each RNA extract, 25 embryos/larvae were pooled and analyzed by real-time PCR. Each data point shows normalized mRNA levels § SEM of three independent biological samples. Note diVerences in the scale of the y-axes of (B) as compared to (A and C). For each ER gene, data were analyzed by two-way ANOVA for stage of development and parental collection site. ER showed signiWcant eVects of developmental stage and site, whereas ERa and ERb showed signiWcant eVects of stage but not site. Refer to Table 3 for levels in unfertilized eggs. Stage-related diVerences in mRNA abundance between each developmental time point and the previous time point are indicated (asterisks, P <0.05). killiWsh ERa, and -b transcripts have a single transla- same 3Ј UTR as the transcript encoding the complete ER tion start site, but all three ER have variants that diVer at protein. Similarly, each of the ER’s have transcripts that the 3Ј-end. In the case of ER, a variant cDNA predicts a predict partial deletions of E and F domains, but these partial deletion aVecting the E/F domains, but has the deletion variants have alternative 3Ј UTRs and

Table 3 Comparison of ER, ERa, and ERb levels in whole ovary of reproductively active females, unfertilized eggs, and 1 dpf embryos of SC and NBH Wsh Mean normalized expression § SEM Ratio: ER:a:b SC NBH SC NBH £ 102 £ 104 £ 102 £ 102 a £ 104 b £ 102 Ovary 11.06 § 0.54 1123.89 § 123.21 7.76 § 0.34 4.17 § 0.74c 873.21 § 105.16 7.15 § 0.43 3:3:2 1:2:2 Eggs 29.26 § 9.56 2.10 § 0.23a 2.28 § 0.01a 18.6 § 4.61 2.22 § 1.34a 1.97 § 0.18a 14000:1:1100 8000:1:800 Embryos 0.14 § 0.03a 0.34 § 0.01a,b 0.11 § 0.03a,b 0.49 § 0.38a,b 0.50 § 0.01a 0.11 § 0.00a,b 400:1:300 1000:1:225 Values for ovary (n D 4 pools of 2) and embryos (n D 3 pools of 25) are those shown in Figs. 5 and 6, respectively. Values for unfertilized eggs represent 3 pools of 25 eggs each. SigniWcance of diVerences was determined by two-way ANOVA followed by Tukey test for pairwise comparisons. aP <0.05 when compared to corresponding ovary; bP < 0.05 when compared to corresponding eggs. cP < 0.05 when site related diVerences in mRNA levels were found solely in ovary and not in eggs or embryos. S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 185 presumably arise from usage of alternate splice sites. with multiple transcripts. Also, the tissue-distribution of We have termed the truncated killiWsh polypeptides ER forms diVers somewhat in diVerent teleostean species ERx, -ax, and -bx to conform to terminology used for (Filby and Tyler, 2005). a previously isolated human ERcx variant, which is The overall tissue distribution or each ER subtype is known to act as a dominant negative on ER function similar for reproductively active Wsh from SC and NBH, (Ogawa et al., 1998). Further, the ER deletion variant implying that mechanisms regulating tissue-speciWc expres- resembles one of the ER7 variants found in normal sion are largely unaVected by long term pollutant exposure. and cancerous human tissues (Herynk and Fuqua, 2004). Nonetheless, there is a notable diVerence in the quantity of Although ERx, -ax, and -bx cDNAs were cloned from ER mRNA in brain, liver, and ovaries of reproductively both populations, we have not determined if their relative active SC and NBH females. When compared to reproduc- abundance is site-speciWc. If so, it would be important to tively inactive SC females, or seasonally active or inactive investigate if they are translatable and have functional SC males, ER mRNA levels in liver and ovaries of repro- importance. ductively active SC females are elevated 3–4-fold, and a SNPs identiWed throughout the coding region of AHR1 smaller but signiWcant elevation is seen in the brain. In in killiWsh have been linked to dioxin-resistance although marked contrast to SC females, Wsh collected from NBH the mechanism is unknown (Hahn et al., 2004). In cloning display little (ovary) or no (liver, brain) seasonal increase in each of the ER from 4 to 7 individuals from each collection ER expression. We interpret the normally occurring site, we sought to determine if there are obvious site-speciWc increase in ER expression as the estrogen inducible-com- SNPs, for example, in conserved functional domains, which ponent. In rainbow trout and zebraWsh, ER transcription might predict altered estrogen signaling. Multiple SNPs are is autoregulated by a positive feedback loop driven by found in each of the three ER, but only two that aVect cod- estrogen and ERE-like half-sites in its promoter (Menuet ing sequence are NBH-speciWc (one each in ER and et al., 2004). The presence of ERE-like binding sites, ERb). Moreover, neither of these two SNPs occurs within together with SP1 and AP1 sites that are known to mediate a conserved region of a functional domain. It is worth not- estrogen actions on mammalian promoters, are conserved ing, however, that non-coding SNPs in the human ER across known teleostean ER genes (Emmith et al., 2005). gene have been linked to altered estrogen responsiveness as In the natural environment, hepatic ER expression indicated by increases in bone density after hormone increases and decreases concomitant with seasonal Xuctua- replacement therapy (Yahata et al., 2005). The authors of tions in gonadal size, plasma estrogen, and vitellogenin syn- this paper hypothesize that an eVect might occur through thesis in females but not in males (Sabo-Attwood et al., altering splice acceptor or donor sites, as has been seen with 2004), indicating that it is a physiologically important regu- SNPs in the hypoxanthine phosphoribosyltransferase gene latory mechanism. Moreover, in vivo treatment with estro- (O’Neill et al., 1998). Systematic analysis of coding and gen induces ER mRNA in several Wsh species, including non-coding sequences of the ER genes in a much larger killiWsh (Urushitani et al., 2003; Menuet et al., 2004). One number of individuals from the NBH and SC killiWsh pop- explanation for attenuation of seasonal increases in ER ulations would be required to determine if polymorphisms expression in reproductively active NBH Wsh, is that circu- identiWed in our study are signiWcantly related to environ- lating estrogen is deWcient. In our earlier study, we reported mental conditions. that reproductively active NBH as compared to SC To examine the possibility that chronic exposure to envi- females, have a low GSI, and low plasma estrogen (Greytak ronmental pollutants impacts mechanisms regulating tis- et al., 2005). Additionally, HSI is low and molecular mark- sue-speciWc expression of one or more of the three ER ers of estrogen eVect in liver (vtg) are indicative of estrogen genes, we Wrst applied real-time PCR to quantify mRNA deWciency. Another possibility is that low ER expression levels in reproductively active SC Wsh. Of the ten tissues in NBH females is due to dysregulation of the estrogen- examined (brain, eye, gill, ovary, testis, spleen, liver, heart, inducible component by dioxin- or estrogen-like NBH con- gut, and muscle), all expressed ER, -a, and -b to some taminants. Studies in rodents or cell lines show that ER extent, but the overall distribution pattern and range of levels decrease after TCDD exposure, due either to a expression from high- to low-expressing tissues were gene- decrease in ER transcription (Tian et al., 1998); post- speciWc. In general, all three ER were expressed at highest translational eVects (Harris et al., 1990; Wang et al., 1993); levels in reproductively relevant tissues. Additionally, the or interference with estrogen actions on estrogen responsive relative abundance of the three mRNA forms in a given tis- genes (see Section 1). Still other studies have reported estro- sue (ratio, ER:ERa:ERb) was form speciWc. Our results gen-mediated decreases in ER mRNA (Saceda et al., 1988; are similar to those reported previously in largemouth bass Read et al., 1989) or protein (Read et al., 1989; Alarid et al., (Sabo-Attwood et al., 2004), seabass (Halm et al., 2004) and 1999). Although our previous report concluded that the to our results with zebraWsh (unpubl. data), but diVer from NBH environment is “estrogenic”, as measured by values previously reported for zebraWsh in which all three increases in estrogen marker genes in males and reproduc- ER mRNAs were found to be lower in ovary than in testis tively inactive female killiWsh (Greytak et al., 2005), we can- (Menuet et al., 2002). The diVerence could be due to primer not rule out the possibility that NBH contaminants are placement, which is an important consideration for genes estrogenic when endogenous estrogen is low (seasonal 186 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 inactivity) and antiestrogenic when endogenous estrogen is Whether these eVects are transient, due to chemically high (seasonal activity). A Wnal, more intriguing possibility induced changes in physiological or biochemical processes, is that the ER gene of NBH Wsh has acquired resistance to or heritable, resulting from multigenerational selection estrogen induction. The most compelling evidence to sup- pressures, are questions that require further study. Of great port this hypothesis is the absence of any diVerences in ER interest is the possibility that resistance in ER-mediated expression when SC and NBH males and females are com- signal transduction pathways is similar or related to AHR- pared during the period of seasonal inactivity. Our earlier mediated resistance described in the same killiWsh popula- study shows that two other estrogen responsive genes (vtg tion and, if so, whether observed changes somehow attenu- and P450aromB) are robustly induced in the same NBH ate endocrine disrupting eVects of environmental killiWsh (Greytak et al., 2005). Detailed dose-response stud- pollutants. ies are required to discriminate among these possibilities, which are not mutually exclusive. Acknowledgments Because interpretation of ER expression in adult Wsh is V complicated by site-related di erences in reproductive con- This work was supported by grants from the National dition, we applied real-time PCR to compare relative abun- Institutes of Health (NIEHS P42 ES07381), the U.S. Envi- dance of the ER , - a, and - b mRNAs in unfertilized eggs ronmental Protection Agency (RD831301), and the and developing embryos/larvae derived from naturally National Sea Grant College Program (R/P-68 22227323). W spawned NBH and SC sh. In contrast to published stud- SRG was the recipient of a National Institutes of Health ies, which report that ER mRNA is not detectable in predoctoral traineeship (NICHD 2T32 HD 073897). The W unfertilized zebra sh eggs (Bardet et al., 2002; Tingaud- authors are grateful to Diane Nacci and Denise Champlin, W Sequeira et al., 2004), we nd that ER , ER a, and - b Atlantic Ecology Division, US EPA, Narragansett RI, for W mRNAs well within detection limits in killi sh eggs with invaluable advice and assistance in collecting NBH Wsh and our real-time PCR method, and the same is true for eggs; and to Mark E. Hahn, Woods Hole Oceanographic W V zebra sh eggs (unpubl. data). As stated above, this di er- Institute, for inspiration and generously sharing his exper- V ence may be due to a di erence in primer placement or tise and resources. assay conditions. Comparison of ER, -a, and -b mRNA levels in unfertilized eggs, on the one hand, and whole ovary or 1 dpf embryos, on the other, suggests that (a) ER References mRNA, in contrast to the ER forms, is selectively Alarid, E.T., Bakopoulos, N., Solodin, N., 1999. Proteasome-mediated pro- expressed by preovulatory oocytes; and (b) all three ER teolysis of estrogen receptor: a novel component in autologous down- mRNAs are maternally transferred but rapidly degraded regulation. Mol. Endocrinol. 13, 1522–1534. post-fertilization. Interestingly, despite lower levels of ER Armstrong, P.B., Child, J.S., 1965. Stages in the normal development of in ovaries of NBH as compared to SC Wsh, we found no Fundulus heteroclitus. Biol. Bull. 128, 143–168. site-related diVerences in levels of ER, -a, or -b mRNA Baba, T., Mimura, J., Nakamura, N., Harada, N., Yamamoto, M., Moroh- V ashi, K.-i., Fujii-Kuriyama, Y., 2005. Intrinsic function of the aryl in eggs or early embryos. This suggests that e ects of NBH hydrocarbon (Dioxin) receptor as a key factor in female reproduction. pollutants on ovarian ER expression are selective for the Mol. Cell. Biol. 25, 10040–10051. non-germinal compartment. During early development, the Bardet, P.L., Horard, B., Robinson-Rechavi, M., Laudet, V., Vanacker, three ER genes are diVerentially expressed. Whereas the J.M., 2002. Characterization of oestrogen receptors in zebraWsh (Danio abundance of ER and ERb mRNA increases slowly until rerio). J. Mol. Endocrinol. 28, 153–163. W Bello, S.M., Franks, D.G., Stegeman, J.J., Hahn, M.E., 2001. Acquired a peak at hatch (12 dpf), a signi cant incremental increase resistance to Ah receptor agonists in a population of Atlantic killiWsh in ERa is not seen until 18 dpf. It will be interesting to (Fundulus heteroclitus) inhabiting a marine superfund site: in vivo and relate these patterns to morphogenetic events by in situ in vitro studies on the inducibility of xenobiotic metabolizing enzymes. hybridization. Toxicol. Sci. 60, 77–91. As with adults, signiWcant site-related diVerences in ER Black, D.E., Gutjahr-Gobell, R., Pruell, R.J., Bergen, B., Mills, L., McEl- roy, A.E., 1998. Reproduction and Polychlorinated biphenyls in Fundu- expression during early development were seen with ER , lus heteroclitus (Linnaeus) from New Bedford Harbor, Massachusetts, but not -a or -b. Indeed, at each time point after the Wrst USA. Environ. Toxicol. Chem. 17, 1405–1414. 24 h of development, ER mRNA levels were Wvefold Brzozowski, A.M., Pike, A.C., Dauter, Z., Hubbard, R.E., Bonn, T., Eng- higher in NBH as compared to SC embryos/larvae, imply- strom, O., Ohman, L., Greene, G.L., Gustafsson, J.A., Carlquist, M., ing exceptionally high rates of embryonic synthesis or 1997. Molecular basis of agonism and antagonism in the oestrogen V receptor. Nature 389, 753–758. unusual mRNA stability. These e ects could be due to Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T.J., Higgins, estrogen-like pollutants sequestered in yolk. Paradoxically, D.G., Thompson, J.D., 2003. Multiple sequence alignment with the high ER expression predicts hypersensitivity to estrogen clustal series of programs. Nucleic Acids Res. 31, 3497–3500. rather than the hyposensitivity that could be adaptive in an Connor, K., Ramamoorthy, K., Moore, M., Mustain, M., Chen, I., Safe, S., estrogenic environment. Zacharewski, T., Gillesby, B., Joyeux, A., Balaguer, P., 1997. Hydroxyl- ated polychlorinated biphenyls (PCBs) as estrogens and antiestrogens: We conclude estrogenic or other contaminants in the structure–activity relationships. Toxicol. Appl. Pharmacol. 145, 111–123. NBH environment impact molecular components of the Crisp, T.M., Clegg, E.D., Cooper, R.L., Wood, W.P., Anderson, D.G., Bae- estrogen signaling pathway in resident killiWsh populations. tcke, K.P., HoVmann, J.L., Morrow, M.S., Rodier, D.J., SchaeVer, J.E., S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188 187

Touart, L.W., Zeeman, M.G., Patel, Y.M., 1998. Environmental endo- Menuet, A., Le Page, Y., Torres, O., Kern, L., Kah, O., Pakdel, F., 2004. crine disruption: an eVects assessment and analysis. Environ. Health Analysis of the estrogen regulation of the zebraWsh estrogen receptor Perspect. 106 (Suppl 1), 11–56. (ER) reveals distinct eVects of ERalpha, ERbeta1 and ERbeta2. J. Mol. Emmith, N., Novillo, A., Callard, G., 2005. Multispecies Comparison of Endocrinol. 32, 975–986. Estrogen Receptors in Teleosts. International Congress of Compara- Menuet, A., Pellegrini, E., Anglade, I., Blaise, O., Laudet, V., Kah, O., Pak- tive Endocrinology, Boston, MA. del, F., 2002. Molecular characterization of three estrogen receptor Felsenstein, J., 1993. PHYLIP PHYLIP (Phylogeny Inference Package) forms in zebraWsh: binding characteristics, transactivation properties, version 3.6. Department of Genetics, University of Washington, Seat- and tissue distributions. Biol. Reprod. 66, 1881–1892. tle, Distributed by the author. Mills, L.J., Chichester, C., 2005. Review of evidence: are endocrine-disrupt- Filby, A.L., Tyler, C.R., 2005. Molecular characterization of estrogen ing chemicals in the aquatic environment impacting Wsh populations? receptors 1, 2a, and 2b and their tissue and ontogenic expression Sci. Total Environ. 343, 1–34. proWles in fathead minnow (Pimephales promelas). Biol. Reprod. 73, Nacci, D., Coiro, L., Champlain, D., Ayaraman, S., McKinney, R., Glea- 648–662. son, W.R.M.T., Specker, J.J., Cooper, K., 1999. Adaptations of wild Garcia-Reyero, N., Raldua, D., Quiros, L., Llaveria, G., Cerda, J., Barcelo, populations of the estuarine Wsh Fundulus heteroclitus to persistent D., Grimalt, J.O., Pina, B., 2004. Use of vitellogenin mRNA as a environmental contaminants. Mar. Biol. 134, 9–17. biomarker for endocrine disruption in feral and cultured Wsh. Anal. Ogawa, S., Inoue, S., Watanabe, T., Orimo, A., Hosoi, T., Ouchi, Y., Mura- Bioanal. Chem. 378, 670–675. matsu, M., 1998. Molecular cloning and characterization of human Greytak, S.R., Champlin, D., Callard, G.V., 2005. Isolation and character- estrogen receptor betacx: a potential inhibitor of estrogen action in ization of two cytochrome P450 aromatase forms in killiWsh (Fundulus human. Nucleic Acids Res. 26, 3505–3512. heteroclitus): diVerential expression in Wsh from polluted and unpol- Ohtake, F., Takeyama, K., Matsumoto, T., Kitagawa, H., Yamamoto, Y., luted environments. Aquat. Toxicol. 71, 371–389. Nohara, K., Tohyama, C., Krust, A., Mimura, J., Chambon, P., Yana- Hahn, M.E., 1998. The aryl hydrocarbon receptor: a comparative perspec- gisawa, J., Fujii-Kuriyama, Y., Kato, S., 2003. Modulation of oestrogen tive. Comp. Biochem. Physiol. C Pharmacol. Toxicol. Endocrinol. 121, receptor signalling by association with the activated dioxin receptor. 23–53. Nature 423, 545–550. Hahn, M.E., Karchner, S.I., Franks, D.G., Merson, R.R., 2004. Aryl hydro- O’Neill, J.P., Rogan, P.K., Cariello, N., Nicklas, J.A., 1998. Mutations that carbon receptor polymorphisms and dioxin resistance in Atlantic killi- alter RNA splicing of the human HPRT gene: a review of the spec- Wsh (Fundulus heteroclitus). Pharmacogenetics 14, 131–143. trum. Mutat. Res. 411, 179–214. Hall, J.M., Couse, J.F., Korach, K.S., 2001. The multifaceted mechanisms Oppenheimer, J.M., 1937. The normal stages of Fundulus heteroclitus. of estradiol and estrogen receptor signaling. J. Biol. Chem. 276, 36869– Anat. Rec. 68, 1–15. 36872. Pakdel, F., Metivier, R., Flouriot, G., Valotaire, Y., 2000. Two estrogen Halm, S., Martinez-Rodriguez, G., Rodriguez, L., Prat, F., Mylonas, receptor (ER) isoforms with diVerent estrogen dependencies are gener- C.C., Carrillo, M., Zanuy, S., 2004. Cloning, characterisation, and ated from the trout ER gene. Endocrinology 141, 571–580. expression of three oestrogen receptors ER, ER1 and ER2 in the Patino, R., Xia, Z., Gale, W.L., Wu, C., Maule, A.G., Chang, X., 2000. European sea bass, Dicentrarchus labrax. Mol. Cell. Endocrinol. 223, Novel transcripts of the estrogen receptor gene in channel catWsh. 63–75. Gen. Comp. Endocrinol. 120, 314–325. Harris, M., Zacharewski, T., Safe, S., 1990. EVects of 2,3,7,8-tetrachlo- Phillips, B., Harrison, P., 1999. Overview of the Endocrine Disruptor Issue. rodibenzo-p-dioxin and related compounds on the occupied nuclear Royal society of chemistry, Cambridge, UK. estrogen receptor in MCF-7 human breast cancer cells. Cancer Res. 50, Pike, A.C., Brzozowski, A.M., Hubbard, R.E., Bonn, T., Thorsell, A.G., 3579–3584. Engstrom, O., Ljunggren, J., Gustafsson, J.A., Carlquist, M., 1999. Hawkins, M.B., Thomas, P., 2004. The unusual binding properties of the Structure of the ligand-binding domain of oestrogen receptor in the third distinct teleost estrogen receptor subtype ERbetaa are accompa- presence of a partial agonist and a full antagonist. EMBO J. 18, 4608– nied by highly conserved amino acid changes in the ligand binding 4618. domain. Endocrinology 145, 2968–2977. Powell, W.H., Bright, R., Bello, S.M., Hahn, M.E., 2000. Developmental Hawkins, M.B., Thornton, J.W., Crews, D., Skipper, J.K., Dotte, A., and tissue-speciWc expression of AHR1, AHR2, and ARNT2 in dioxin- Thomas, P., 2000. IdentiWcation of a third distinct estrogen receptor sensitive and -resistant populations of the marine Wsh Fundulus hetero- and reclassiWcation of estrogen receptors in teleosts. Proc. Natl. Acad. clitus. Toxicol. Sci. 57, 229–239. Sci. USA 97, 10751–10756. Read, L.D., Greene, G.L., Katzenellenbogen, B.S., 1989. Regulation of Herynk, M.H., Fuqua, S.A., 2004. Estrogen receptor mutations in human estrogen receptor messenger ribonucleic acid and protein levels in disease. Endocr. Rev. 25, 869–898. human breast cancer cell lines by sex steroid hormones, their antago- Hoegg, S., Brinkmann, H., Taylor, J.S., Meyer, A., 2004. Phylogenetic nists, and growth factors. Mol. Endocrinol. 3, 295–304. timing of the Wsh-speciWc genome duplication correlates with the Rozen, S., Skaletsky, H., 2000. Primer3 on the WWW for general users and diversiWcation of teleost Wsh. J. Mol. Evol. 59, 190–203. for biologist programmers. Methods Mol. Biol. 132, 365–386. Krust, A., Green, S., Argos, P., Kumar, V., Walter, P., Bornert, J.M., Cham- Sabo-Attwood, T., Kroll, K.J., Denslow, N.D., 2004. DiVerential expres- bon, P., 1986. The chicken oestrogen receptor sequence: homology with sion of largemouth bass (Micropterus salmoides) estrogen receptor v-erbA and the human oestrogen and glucocorticoid receptors. EMBO isotypes , , and by estradiol. Mol. Cell. Endocrinol. 218, 107– J. 5, 891–897. 118. Lassiter, C.S., Kelley, B., Linney, E., 2002. Genomic structure and embry- Saceda, M., Lippman, M.E., Chambon, P., Lindsey, R.L., Ponglikit- onic expression of estrogen receptor beta a (ERa) in zebraWsh (Danio mongkol, M., Puente, M., Martin, M.B., 1988. Regulation of the rerio). Gene 299, 141–151. estrogen receptor in MCF-7 cells by estradiol. Mol. Endocrinol. 2, Lind, P.M., Eriksen, E.F., Sahlin, L., Edlund, M., Orberg, J., 1999. EVects 1157–1162. of the antiestrogenic environmental pollutant 3,3Ј,4,4Ј, 5- pentachlo- Safe, S., 2005. Clinical correlates of environmental endocrine disruptors. robiphenyl (PCB #126) in rat bone and uterus: diverging eVects in Trends Endocrinol. Metab. 16, 139–144. ovariectomized and intact animals. Toxicol. Appl. Pharmacol. 154, Safe, S., Wormke, M., Samudio, I., 2000. Mechanisms of inhibitory aryl 236–244. hydrocarbon receptor-estrogen receptor crosstalk in human breast Menuet, A., Anglade, I., Flouriot, G., Pakdel, F., Kah, O., 2001. Tissue-spe- cancer cells. J. Mammary Gland Biol. Neoplasia 5, 295–306. ciWc expression of two structurally diVerent estrogen receptor alpha Salamov, A.A., Nishikawa, T., Swindells, M.B., 1998. Assessing protein isoforms along the female reproductive axis of an oviparous species, coding region integrity in cDNA sequencing projects. Bioinformatics the rainbow trout. Biol. Reprod. 65, 1548–1557. 14, 384–390. 188 S.R. Greytak, G.V. Callard / General and Comparative Endocrinology 150 (2007) 174–188

Schmidt, J.V., BradWeld, C.A., 1996. Ah receptor signaling pathways. Urushitani, H., Nakai, M., Inanaga, H., Shimohigashi, Y., Shimizu, A., Annu. Rev. Cell Dev. Biol. 12, 55–89. Katsu, Y., Iguchi, T., 2003. Cloning and characterization of estrogen Simon, P., 2003. Q-Gene: processing quantitative real-time RT-PCR data. receptor alpha in mummichog, Fundulus heteroclitus. Mol. Cell. Endo- Bioinformatics 19, 1439–1440. crinol. 203, 41–50. Sumpter, J.P., Johnson, A.C., 2005. Lessons from endocrine disruption and Wang, X., Porter, W., Krishnan, V., Narasimhan, T.R., Safe, S., 1993. their application to other issues concerning trace organics in the Mechanism of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)-mediated aquatic environment. Environ. Sci. Technol. 39, 4321–4332. decrease of the nuclear estrogen receptor in MCF-7 human breast can- Tchoudakova, A., Pathak, S., Callard, G.V., 1999. Molecular cloning of an cer cells. Mol. Cell. Endocrinol. 96, 159–166. estrogen receptor subtype from the goldWsh, Carassius auratus. Gen. Weaver, G., 1983. PCB contamination in and around New Bedford, Mass. Comp. Endocrinol. 113, 388–400. Environ. Sci. Technol. 18, 22A–27A. Tian, Y., Ke, S., Thomas, T., Meeker, R.J., Gallo, M.A., 1998. Transcriptional Yahata, T., Quan, J., Tamura, N., Nagata, H., Kurabayashi, T., Tanaka, suppression of estrogen receptor gene expression by 2,3,7,8-tetrachlo- K., 2005. Association between single nucleotide polymorphisms of rodibenzo-p-dioxin (TCDD). J. Steroid. Biochem. Mol. Biol. 67, 17–24. estrogen receptor gene and eYcacy of HRT on bone mineral den- Tingaud-Sequeira, A., Andre, M., Forgue, J., Barthe, C., Babin, P.J., 2004. sity in post-menopausal Japanese women. Hum. Reprod. 20, 1860– Expression patterns of three estrogen receptor genes during zebraWsh 1866. (Danio rerio) development: evidence for high expression in neuromasts. Zhu, B.T., Conney, A.H., 1998. Functional role of estrogen metabolism in Gene Expr. Patterns 4, 561–568. target cells: review and perspectives. Carcinogenesis 19, 1–27.