703

BIOLOGY OF REPRODUCTION 73, 703–712 (2005)Published online before print 8 June 2005.DOI 10.1095/biolreprod.105.042796

Expression of Estrogen Receptor ESR1 and Its 46-kDa Variantin the Gubernaculum Testis1

Christophe Staub,2,5 Michel Rauch,3,5 Francois Ferriere,3,6 Melanie Trepos,5 Isabelle Dorval-Coiffec,5

Philippa T. Saunders,7 Gilda Cobellis,4,5 Gilles Flouriot,6 Christian Saligaut,6 and Bernard Jegou5

INSERM,5 U625, GERHM, IFR 140, Campus de Beaulieu, Univ Rennes I, Rennes, Bretagne F-35042, FranceUMR-CNRS 6026,6 IFR 140, Campus de Beaulieu, Univ Rennes I, Rennes, Bretagne F-35042, FranceMRC Human Reproductive Sciences Unit,7 Centre for Reproductive Biology, Edinburgh EH16 4SB, United Kingdom

ABSTRACT

Testicular descent corresponds to migration of the testis fromthe abdominal cavity to the scrotum and is essential for properfunctioning of the testis. Recent advances in the characteriza-tion of estrogen receptor (ESR) subtypes and isoforms in varioustissues prompted us to study ESRs within the gubernaculum tes-tis, a structure involved in testicular descent. In the rat guber-naculum, we searched for ESR alpha (Esr1) and beta (Esr2) andfor the androgen receptor (Ar), androgens being known to reg-ulate testicular descent. Reverse transcription-polymerase chainreaction (RT-PCR) revealed that Esr1, Esr2, and Ar mRNAs wereall expressed in the gubernaculum. Using PEETA (Primer exten-sion, Electrophoresis, Elution, Tailing, and Amplification), we es-tablished that all Esr1 leader exons, previously identified in oth-er organs, such as the uterus and pituitary, were transcribed inthe gubernaculum, with the major form being O/B. The RNAprotection assays, RT-PCR, and Western blot experiments re-vealed that isoform-specific mRNA transcripts generated by al-ternative splicing of the C-leader sequence on coding exons 1and 2 of the Esr1 gene gave the 46- and 66-kDa ESR1 proteins.The ESR1 and AR proteins were found to colocalize in the pa-renchymal cells of the gubernaculum early in development,whereas AR also was strongly expressed in the muscular cells,both during fetal and postnatal life. The ESR2 protein was weaklyexpressed, principally in the muscular cells, but only once tes-ticular descent had occurred. The levels of the 46-kDa ESR1variant (ER46) exceeded those of the 66-kDa ESR1 form (ER66)at periods when the gubernaculum developed. Conversely, the66-kDa form appears to predominate clearly when the guber-naculum growth was low or completed. The possible role of es-trogens on the modulation of the androgen-dependent growthof the gubernaculum and, more widely, on testicular descent isdiscussed.

androgen receptor, developmental biology, estradiol receptor,male reproductive tract

1Supported by the INSERM, Ministere de la Recherche et Ministere de laSante, and the EDEN European program, contract number QLK4-CT-2002-00603.2Correspondence: Bernard Jegou, INSERM, U625, GERHM, Campus deBeaulieu, Univ Rennes I, Rennes, Bretagne F-35042 France.FAX: 33 2 23 23 50 55; e-mail: bernard.jegou@rennes.inserm.fr3These authors contributed equally to this work.4Current address: Dipartimento di Medicina Sperimentale, II Universita diNapoli, Via Costantinopoli, Naples 16-80138, Italy.

Received: 15 April 2005.First decision: 12 May 2005.Accepted: 3 June 2005.Q 2005 by the Society for the Study of Reproduction, Inc.ISSN: 0006-3363. http://www.biolreprod.org

INTRODUCTION

During the last decade, concern has been rising over in-creases in the number of abnormalities associated with thehuman urogenital tract [1–3]. One of the abnormalities be-coming more frequent is malpositioning of the testes, a con-dition called cryptorchidism [4, 5]. Cryptochidism leads tohypofertility and sterility and is a major risk factor for tes-ticular cancer, the incidence of which also is increasingthroughout the world [5, 6]. At present, the cause of crypt-orchidism is unclear. A better understanding of the mech-anisms controlling the process of gonad descent—a processdefined by Scorer [7] as ‘‘the movement of the testis fromthe abdominal cavity to the bottom of a fully developedand fully relaxed scrotum’’—is needed before the cause ofthis condition can be elucidated.

During the initial phase of fetal development, the testesare in the renal position, held by the cranial suspensoryligament (CSL) at the upper pole and the caudal genitalligament (CGL) at the lower pole. The CGL is comprisedof parenchymal cells, which form a cord making up part ofthe gubernaculum organ. This organ also consists of a bulbof mesenchymal cells, which differentiate into the cremas-ter muscle; and the cremaster muscle is involved in the finalpositioning of the testes within the scrotum. The initialphase of transabdominal testicular descent, which is con-trolled by insulin-like factor 3 (INSL3) and androgens, re-sults in the testis being positioned in the lower abdomenduring embryo growth [8–10]. The INSL3 controls guber-naculum swelling via its receptor, LGR8 (leucine-rich re-peat containing G protein-coupled receptor 8 also knownas GREAT or relaxin receptor 2) [11, 12], a process result-ing in thickening of the CGL because of increases in water,glycosaminoglycan, and hyaluronic acid content [13]. An-drogens mediate the beginning of the regression phase ofthe CSL [14]. During the second phase of transabdominaldescent, androgens trigger the end of the CSL regressionphase and stimulate growth of the gubernaculum bulb and,thus, differentiation of the muscular part of the gubernac-ulum. The traction resulting from this growth facilitatesmovement in the inguinoscrotal phase of the testicular de-scent [9, 15–17].

Several lines of evidence indicate that estradiol, dieth-ylstilbestrol (DES), and estrogenic compounds in generalalso are involved in mediating testicular descent. For ex-ample, administration of estradiol [18, 19] or DES [20, 21]to gestating female rats inhibits testicular descent, leadingto cryptorchidism in the male offspring. Both gubernacu-lum swelling and transabdominal testicular migration areblocked by this prenatal exposure to estrogens [19]. Theexposure of pregnant women to DES leads to an increase

704 STAUB ET AL.

TABLE 1. Oligonucleotides used in the RT-PCR experiments.

mRNA Primers 59-39 Start location Product size (bp) NCBI accession no.

Esr1 AGTGAAGCCTCAATGATGGGCAAAGATCTCCACCATGCCT

13001580

281 NM012689

Esr2 CTACTGAACGCGGTGACAGACGTGTGAGCATTCAGCATCT

16541878

255 U57439

Ar GTGTCGTCTCCGGAAATGTTGGAATCAGGCTGGTTGTTGT

29483213

266 NM012502

Cyp19a1 ACGTGGAGACCTGACAAAGGCATGACCAAGTCCACGACAG

9641240

277 NM017085

Srd5a1 GAGATATTCAGCTGAGACCCTTAGTATGTGGGCAGCTTGG

21642350

187 J0535

Actb GACTACCTCATGAAGATCCTTTGCTGATCCACATCTGCTG

6401138

512 NM031144

in the incidence of cryptorchidism in male children [22].There also may be a relationship between increased levelsof placental estradiol and the occurrence of cryptorchidismin the exposed babies [23, 24]. Furthermore, transgenicmale mice, which overexpress Cyp19a1 (human aromatase)and, thus, exhibit elevated serum estradiol concentrations,display several reproductive tract abnormalities, includingcryptorchidism [25]. Finally, evidence recently has accu-mulated that indicates some xenoestrogens (the so-calledendocrine disruptors) are involved in the occurrence of dif-ferent types of abnormalities of the urogenital tract, includ-ing cryptorchidism [3, 26, 27]. The effects of estrogens ontesticular descent generally are believed to be indirect, re-sulting from inhibition of LH production and, consequently,testosterone secretion [28] and/or from inhibition of fetalINSL3 production in the Leydig cells [29–31]. However,the finding that Esr1 knockout male mice have retractedgonads [32] suggests that estrogens play a direct role in thescrotal positioning of the testis.

The present study searched for the presence of the es-trogen receptors ESR1 and ESR2 within the rat gubernac-ulum testis as an indication of the possible direct implica-tion of estrogens in the process that leads to final position-ing of the testis. We demonstrate that whereas ESR2 proteinwas not detected in the fetal gubernaculum, high expressionlevels of ESR1 were found in the parenchymal cells of thegubernaculum testis, colocalized with the androgen receptor(AR), from 16 days postcoitum (dpc) to 22 days postpartum(dpp). Two isoforms of ESR1, namely ER66 and ER46,were characterized. We hypothesize that changes in theirexpression during gubernacular development and their pos-sible competitive interaction could represent a mechanismby which estrogens modulate the action of androgens dur-ing the different phases of testicular descent.

MATERIALS AND METHODS

Animals and Sample Collection

Timed pregnant female Sprague-Dawley rats were purchased fromElevage Janvier (Le Genest Saint Isle, Laval, France). They were anes-thetized by an i.p. injection of 40 mg/kg of sodium pentobarbital (Sanofi-Synthelabo, Libourne, France). Gubernaculi were microdissected from ratfetuses aged 16, 17, 18, 20, and 21 days dpc and 0.5, 3, 4, 8, 12, and 22days dpp under a binocular microscope (Olympus B061). After dissection,gubernaculi were immediately frozen in liquid nitrogen and either used toobtain mRNA and protein extracts, fixed in Bouin fluid, and processedinto paraffin wax or cryopreserved in Tissue-Teck O.C.T. (Sakura FinetekEurope BV, The Netherlands) for immunohistochemical analysis. All ex-periments were conducted in accordance with the Guide for the Care andUse of Laboratory Animals (Institute for Laboratory Animal Research[ILAR] of the National Academy of Sciences, Bethesda, MD).

RNA Extraction and Reverse Transcription-PolymeraseChain Reaction Analysis

Total RNA was extracted from each gubernaculum using the methoddescribed by Chomczynski and Sacchi [33]. Sequences encoding Esr1(estrogen receptor a), Esr2 (estrogen receptor b), Ar (AR), Cyp19a1 (aro-matase), Srd5a1 (5a-reductase isoform I), and Actb (b-actin) mRNA wereamplified by reverse transcription-polymerase chain reaction (RT-PCR) us-ing the primers listed in Table 1.

Complementary DNA was prepared from 4 mg of RNA using 40 ngof random hexanucleotides (Boehringer Mannheim, Germany) and 200 Uof Moloney murine leukemia virus-reverse transcriptase (M-MLV-RT; Pro-mega, Madison, WI) according to the manufacturer’s instructions. ThePCR was carried out on 40 ng of cDNA in a final volume of 25 mlcontaining 0.6 U of Taq polymerase (Qiagen, Courtaboeuf, France) and0.2 mM of each appropriate 59 and 39 sequence-specific primers (Table 1).Actin amplification was performed as a control for RNA quality, quantityestimation, and RT efficiency. The samples were denatured at 948C for 5min. Amplification was carried out using 35 or 40 standard PCR cycleswith a 588C annealing temperature using a thermal cycler. Aliquots of 15ml of each PCR sample were subjected to electrophoresis on 1.8% agarosegels. Product size was determined by comparison to markers (100-basepair [bp] DNA ladder; New England Biolabs). Bands were visualized afterstaining with 0.5 mg/ml of ethidium bromide by observation under ultra-violet illumination (Multimage Light Cabinet; Alpha Innotech Corpora-tion, San Leandro, CA). The RT-PCR products were identified by sequenc-ing using the BigDye system (Applied Biosystems, Foster City, CA) andan ABI 310 sequencer (PerkinElmer, Wellesley, MA).

Long Esr1 cDNA AmplificationsComplementary DNA was synthesized from 1 mg of total RNA by RT-

PCR using 50 U of expand reverse transcriptase and the conditions rec-ommended by the supplier (Boehringer Mannheim) with the oligonucle-otide primer R1 from the 39-untranslated region (UTR; exon 8) of rEsr1mRNAs (Table 2; see Fig. 4A). An aliquot (2 ml) of the RT product wasthen used as a template for two rounds of 35 cycles of PCR amplification.The first round was done in the presence of the C1 and R2 primer pair;the second round was done using the nested primers, C2 and R3 (Table2; see Fig. 4A). The 59-primers (C1 and C2) used for the C-rEsr1 cDNAamplification were designed using the published sequence for exon C(GenBank accession no. X98237) [34]. The 39-primer R2 and the nestedprimer R3 were designed from the 39 region common to all rEsr1 cDNAs,located immediately upstream from the region used to design the RT prim-ers (see Fig. 4A). Both rounds of amplification were performed using theExpand Long Template PCR System (Boehringer Mannheim) and the con-ditions recommended by the manufacturer. Ten-microliter aliquots fromeach reaction were analyzed on a 1% agarose gel stained with ethidiumbromide.

Modified Primer Extension and S1 Nuclease Mappingof Esr1

The strongly labeled, single-stranded DNA probe needed for S1 nu-clease mapping and the long labeled primer used for primer extension wereobtained using a specific primer and T7 DNA polymerase in the presenceof [a-32P]dCTP (3000 Ci/mmol) [35, 36]. The probe (0/B) or the longprimer (S) were then allowed to hybridize to the appropriate RNA samples

705ESTROGEN RECEPTORS IN THE GUBERNACULUM TESTIS

TABLE 2. Oligonucleotides used in primer extension, S1 nuclease mapping, and long cDNA amplification of Esr1.

Name Primers 59-39 Location Descriptiona

R1b

C1c

R2c

C2c

R3c

TTCTGCTAAAGTCATACAGGACATAGTCTTGCTGTGTGTAGGGCGGATTCGCAGAACCTTGTGGCTGAGGAGGTTCCTGGAAGGGGAGTTCTCAGATGGTGTT

Exon 8Exon CExon 8Exon CExon 8

RT primerPCR (F)PCR (R)Nested PCR (F)Nested PCR (R)

F1d

R4d

F8e

R3e

OdGf,g

R5g

ATGACCATGACCCTTCACACTTCTGGCGTCGATTGTCAGACTGTCTCAGCTCTTGACGGGAGTTCTCAGATGGTGTTGGCCACGCGTCGACTAGTACGGGIIGGGIIGGGIIGCCGAGGTACAGATTGGCTT

Exon 1Exon 2Exon O/BExon 8Poly(C) tail

Exon 1

Primer S PCR (F)Primer S PCR (F)O/B probe PCR (F)O/B probe PCR (R)Seq PCR (F)

Seq PCR (R)a F, forward primer; R, reverse primer.b Specific primer for reverse transcription (RT) of Esrl mRNAs.c Primers for PCR and nested PCR for long Esr1 cDNA amplification.d PCR primers used to generate the long primer S.e PCR primers used to generate the O/B probe.f OdG, Oligo (dG) adapter primer.g PCR primers used to sequence the 59-Esr1 extension products.

FIG. 1. Detection of Esr1, Esr2, Ar,Cyp19a1, Srd5a1, and Actb mRNA by RT-PCR analysis at various stages of guberna-cular development. Positive controls wereas follows: oviducts (Ovi) from 23-dpp ratsfor Esr1, ovaries (Ova) from 23-dpp rats forEsr2 and Cyp19a1, and epididymis (Epi)from 15-dpp rats for Ar and Srd5a1. Nega-tive controls were examined by PCR per-formed with mRNA (2) and water (0) in-stead of cDNA. Numbers in parenthesesare the expected product sizes (see Table1); the letters a and b correspond to 35and 40 amplification cycles, respectively.The RT-PCR products were sequenced us-ing the BigDye system (Applied Biosys-tems, Foster City, CA) and an ABI 310 se-quencer (PerkinElmer, Wellesley, MA).

and subjected to an S1 nuclease digestion or to reverse transcription, re-spectively.

The S long primer was generated by PCR using 1 ng of a vectorcontaining the rEsr1 open reading frame and two specific primers: a bio-tinylated 59-primer F1, and a downstream primer R4 located in exon 2(Table 2; see Fig. 2A).

The 0/B probe was generated by PCR using the rEsr1 cDNA vectorand two specific primers: F8 specific to the exon 0/B in the mRNA of therEsr1 isoform, and primer R3 located in exon 8 (Table 2; see Fig. 3). TheRT-PCR product was subcloned into the TA cloning vector pCR3.1 (In-vitrogen, Carlsbad, CA) downstream of the T7 promoter. Finally, the 0/Bprobe was prepared by PCR using the biotinylated T7 and R3 primer pair.

All biotinylated PCR products were isolated on the basis of their bind-ing to streptavidin-coated magnetic beads (Dynal Biotech, Oslo, Norway).This step was carried out as recommended by the manufacturer, and thenonbiotinylated DNA strands were removed using 0.1 M NaOH.

Probe 0/B and S primer were then extended by allowing the R4 primerto anneal to the corresponding biotinylated, single-stranded template. Afterelution of the single-stranded DNA probe and primer by alkaline treatmentand magnetic separation, 2 3 105 cpm of the probe or primer were co-precipitated with 30 mg of total RNA and then dissolved in 20 ml ofhybridization buffer (80% formamide, 40 mM PIPES [pH 6.4], 400 mM

NaCl, and 1 mM EDTA [pH 8]). The probe and primer were then dena-tured by incubation at 658C for 10 min. Hybridization reactions were car-ried out overnight at 558C.

The S1 digestion and reverse transcription were carried out as previ-ously described [37]. Products were separated by electrophoresis on de-naturing polyacrylamide/urea gels.

The most abundant extension products were cut out and eluted fromthe gel using 500 ml of 0.5 M ammonium acetate, 10 mM magnesiumacetate, 1 mM EDTA (pH 8), and 0.1% SDS and then extracted using theQiaex II gel extraction kit (Qiagen). A poly(C) tail was then added to eachof the extension products to allow subcloning into the pCR3.1 plasmidvector (Invitrogen) for sequencing of the product from a PCR using anoligo(dG) adapter primer and a specific primer R5 located in exon 1 (Table2; see Fig. 2A). All clones were sequenced.

ESR1 Western Immunoblot AnalysisGubernaculum lysates were prepared using ice-cold lysis buffer (50

mM Tris [pH 6.8], 5 M urea, and 2% SDS; Bio-Rad Laboratories, Her-cules, CA) and 1 M dithiothreitol (Quantum-Appligen, Illkirch, France)and then centrifuged at 105 000 3 g for 1 h at 48C. The protein concen-tration in the supernatants was determined using a red-dot-blot protein-

706 STAUB ET AL.

FIG. 2. Expression of 59-UTR mRNA iso-forms of the rat estrogen receptor-a (rEsr1)in the liver, uterus, pituitary gland, and gu-bernaculum. A) Genomic organization ofthe 59 region of the rat Esr1 gene. The po-sition of all upstream exons described sofar is shown. The ATG start codon, the ac-ceptor splice site in exon 1, and the posi-tion of the F1, R4, and R5 primers are in-dicated. Broken lines symbolize splicingsites. The corresponding transcripts areshown as well. B) Total RNA from yeast(negative control), liver (Liv), uterus (Ut),pituitary (Pit), and gubernaculum (Gub)was allowed to hybridize to primer S andsubjected to RT. The extension productswere then separated on a sequencing gel.The extension products corresponding totranscripts I, II, III, IV, and V are indicated.The distribution pattern and intensities ofexpression of the Esr1 leader exons aresimilar between the gubernaculum and thepituitary.

assay technique based on ponceau S staining [38]. Equal amounts of pro-tein from each extract were subjected to electrophoresis on 8% (w/v) SDS-PAGE under reducing conditions. The proteins were then electrotrans-ferred (over 2 h) onto nitrocellulose membranes (Immobilon-PSQ;Millipore, Bedford, MA). Membranes were incubated overnight at 48Cwith Tris-buffered saline (TBS [pH 7.6]; 20 mM Tris, 1.4% HCl, and 130mM NaCl) containing 3% bovine serum albumin, fraction V (BSA; Eu-robio, Les Ulis, France) to block nonspecific binding. They were thenincubated for 2 h at room temperature in TBS containing 0.1% BSA, 0.2%Tween 20 (Sigma, St Louis, MO), and either the NCL-ER-6F11/2 (finaldilution, 1:100; Novocastra, Newcastle upon Tyne, UK) or the MC20 (fi-nal dilution, 1:1000; Santa Cruz Biotechnology, Santa Cruz, CA) anti-ESR1 antibodies or an anti-b-tubulin antibody as a standardization control(final dilution, 1:2000; Pharmingen, BD Biosciences, Franklin Lakes, NJ).After washing with TBS containing BSA (0.1%) and Tween-20 (0.2%),the PSQ membranes were incubated for 1 h at room temperature in TBS-0.1% BSA-0.2% Tween-20 with horseradish peroxidase (HRP)-conjugatedanti-rabbit (NA9340V) or anti-mouse (NA931V; final dilution, 1:10 000;Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) antibod-ies. After washing in TBS-0.1% BSA-0.2% Tween-20, the signal was de-tected using enhanced chemiluminescence plus Western immunoblot de-tection system (Enhanced Chemiluminescence System; Amersham Bio-sciences, Little Chalfont, Buckinghamshire, UK) and Kodak Biomax MRscientific imaging film (Kodak, New Haven, CT). The films were scannedwith an Epson Perfection 1200 photo scanning system (Epson France S.A.,Levallois Perret, France). The intensity of the bands were then measuredusing the Alpha Ease FC Stand Alone software (Alpha Innotech, SanLeandro, CA).

Immunohistochemistry

Cryosections (thickness, 7 mm) were cut using a cryostat (Cryo-StarHM 560 M; Microm, Walldorf, Germany) at 2208C and then collected onSuperFrost Plus microscope slides (CML, Nemours, France) and air-dried.Sections were then fixed using 3.7% formaldehyde in PBS for 10 min,quickly immersed in cold methanol for 2 min, and stored in PBS/glycerol(v/v) containing 250 mM sucrose and 7 mM MgCl2 at 2208C until furtherprocessing. After washing in PBS, sections were incubated for 20 min atroom temperature in PBS containing 1% BSA to prevent nonspecific pro-tein binding. Sections were washed with PBS containing 0.1% BSA. Serialsections were incubated overnight at 48C in PBS containing 0.1% BSAand one of the following primary antibodies: the NCL-ER-6F11/2 anti-ESR1 antibody (final dilution, 1:10; Novocastra), the MC20 anti-ESR1antibody (final dilution, 1:300; Santa Cruz Biotechnology), or the M22Aanti-AR antibody (final dilution, 1:200). This M22A antibody was ob-tained from a mouse monoclonal hybridoma and was selected by immu-noprecipitation of cytosolic tritiated R1881 calf AR complex. It binds toAR on Western blots and has been used on rat prostate as positive control(not shown). Incubation with mouse immunoglobulin (IgG1) (X0931;DAKO, Glostrup, Denmark) was used as a negative control. The sectionswere then washed with 0.05% Tween 20 in PBS, incubated first withbiotin-labeled anti-rabbit (E0432) or anti-mouse (E0464) secondary anti-bodies (final dilution, 1:500; DAKO) for 60 min at room temperature, thenwith streptavidin-HRP (P0397; final dilution, 1:500, DAKO) for 30 minat room temperature, and finally with diaminobenzidine (DAB) for 10 min(Sigma Fast 3,39 DAB tablet sets; Sigma). After mounting in Eukitt (Kin-dler, Freiburg, Germany), the sections were analyzed using a bright-field

707ESTROGEN RECEPTORS IN THE GUBERNACULUM TESTIS

FIG. 3. S1 nuclease protection assays. The principles of the (0/B-1) rEsr1 mRNA detection method are shown on the left. The size and position on thegel of the single-stranded 0/B probe and protected fragments after S1 digestion of the probe/rEsr1 mRNA hybrids also is shown, as are the methioninestart codon (AUG) and the position of F8 and R3 primers. Open boxes represent the exons (1–8) encoding rEsr1, and the gray box represents the unique0/B sequence. Experiments were carried out on total RNA extracts from liver (Liv), uterus (Ut), pituitary (Pit), gubernaculum (Gub), and yeast (negativecontrol). Extracts were allowed to hybridize to the labeled 0/B probe and then treated with S1 nuclease. The undigested (protected) hybrids were thenseparated on a sequencing gel (right). The undigested probe is shown as well. The leader exon 0/B is spliced on coding exon 1 in the uterus, pituitary,and gubernaculum. Other leader sequences also are spliced at the coding exon 1 site in these tissues. The 0/B leader sequence was not found in theliver, indicating that Esr1 mRNA is produced from a different leader sequence in this tissue.

FIG. 4. Long template PCR. A) Genomic organization of the rat Esr1 gene showing the position of the oligonucleotides C1, C2, R1, R2, and R3 used.B) Presence of rER66 and rER46 mRNA in the gubernaculum (Gub), MMQ cells (MMQ; positive control), pituitary gland (Pit), uterus (Ut), and liver(Liv). The tRNA was used as a negative control (Control). Both rER66 and rER46 mRNA transcripts were detected in the 20.5-dpc gubernaculum. C)Schematic representation of the structures of the full-length and the amino-terminal truncated rESR1 proteins. The A/B domain is responsible for AF1transactivation and is missing from the rER46 protein. The 6F11 antibody, which is directed against this amino-terminal domain of the protein, did notrecognize the truncated isoform. The exon C domain constitutes the DNA-binding domain (DBD) of the protein. The D domain contains the nuclearlocalization sequence (NLS), and the E domain constitutes the ligand-binding domain (LBD). The MC20 antibody is directed against the carboxy-terminalF domain and, thus, recognizes both ESR1 isoforms.

Olympus AX60TF microscope with monochromatic objectives (Olympus,France) and photographed using a digital macro camera (Kigamo).

For ESR2 immunohistochemistry, we used a sheep polyclonal antibodythat previously demonstrated cell-specific patterns of expression in themale urogenital tract of adult mice [39]. Briefly, sections (thickness, 5mm) were mounted on charged slides (Superfrost; BDH Laboratory Sup-plies, Poole, UK) and subjected to heat-induced antigen retrieval in 10mM citrate (pH 6; 5 min at full pressure and then allowed to stand for 20min). Endogenous peroxidases were blocked by incubating slides in asolution of 3% (v/v) hydrogen peroxide (in methanol) for 30 min on a

rocker platform. Sections were washed in water and in TBS (0.05 M Tris[pH 7.4] and 0.85% NaCl) for 5 min. Tissue was then blocked in normalrabbit serum (NRS)/TBS/BSA (Diagnostics Scotland, Carluke, Lanarkshi-re, UK) diluted 1:4 in TBS plus 5% BSA containing four drops per 1 mlof avidin for 30 min at room temperature (blocking kit SP-2001; VectorAvidin/Biotin, Peterborough, UK). Two 5-min TBS washes were followedby blocking in biotin (four drops/1 ml of TBS) for 20 min. The sectionswere then washed twice in TBS, then incubated for 16 h at 48C with sheepanti-estrogen receptor b diluted 1:500 in TBS. Sections were washed twicefor 5 min each time in TBS and incubated with rabbit anti-sheep IgG

708 STAUB ET AL.

FIG. 5. A) ESR1 detection by Westernblot analysis. Similar amounts of homoge-nized lysates from gubernaculi at 16, 17,18, 19, 20, and 21 dpc as well as at 0, 3,4, 8, 12, and 22 dpp were separated by8% SDS-PAGE and immunoblotted witheither ESR1 MC20 antibody or tubulin an-tibody. The oviduct (ovi) from 23-dpp ratswas used as positive control. Arrowheadsindicate the positions of the ESR1 66- and46-kDa isoforms. B) Representation of thep66:p46 ratio calculated using the relativeintensity of the bands containing the 66-and 46-kDa proteins during gubernaculardevelopment. Results are presented as themean ratio 6 SEM calculated from threeindependent experiments. Three differentphases (I–III) were characterized on thebasis of the fluctuations in the p66:p46 ra-tio. Means between phase I and phase IIare different (P 5 0.003) as well as meansbetween phase II and phase III (P 50.041). In contrast, no significant differ-ence were observed between phase I andphase III (P 5 0.526).

(Vector, Peterborough, UK) diluted 1:500 in NRS/TBS/BSA for 1 h. Sec-tions were washed in TBS (5 min each wash) incubated in avidin-biotincomplex-HRP complex (DAKO) for 30 min, then washed in TBS (twowashes for 5 min each wash), and bound antibodies were visualized byincubation with 3,39-diaminobenzidine tetrahydrochloride (liquid DAB;catalog no. K3468; DAKO). Sections were counterstained with hematox-ylin. Images were captured using an Olympus Provis microscope (Olym-pus Optical Co., London, UK) equipped with a Kodak DCS330 camera(Eastman Kodak, Rochester, NY), stored on a Macintosh PowerPC com-puter (Apple, Cupertino, CA), and assembled using Photoshop 7 (Adobe,Mountain View, CA).

StatisticsDifferences among the p66:p46 ratios between the three characterized

phases were analyzed by the nonparametric Mann-Whitney test. Analyseswere made using the Statview 5.0 software (SAS Institute, Inc., Cary, NC).

RESULTS

ESR1 Is Expressed Within the Gubernaculum

The expression of Esr1, Esr2, Ar, Cyp19a1, and Srd5a1in the gubernaculum was investigated by RT-PCR analysisthroughout development (Fig. 1). Esr1, Ar, and Srd5a1were found at all investigated stages when 35 PCR cycleswere performed. Esr2 was only detected in the gubernaculidissected from 18- and 20-dpc rats when 35 PCR cycleswere carried out but was detected in the gubernaculi dis-sected from 18-dpc to 22-dpp rats after 40 PCR cycles.Cyp19a1 was never detected at any stage or in any condi-tion.

Bearing in mind the complex genomic organization ofthe Esr1 gene, in which multiple promoters have been iden-tified [40], we first studied what was the major leader exonin the gubernaculum. The most abundant extension prod-ucts were then sequenced, and analysis revealed the pres-

ence of several rEsr1 mRNA isoforms with different 59-UTRs. rEsr1 mRNA was abundant in the gubernaculum,and all previously described leader exons were present (Fig.2). The pattern of distribution and cumulative intensity ofthe rEsr1 mRNA in the gubernaculum were similar to thosefound in the pituitary and uterus. In contrast, only a fewleader sequences were found in the liver (Fig. 2). The 0/Bwas the leader exon preferentially transcribed in both thepituitary and the gubernaculum.

To confirm the relative predominance of 0/B rEsr1mRNA isoform over the other isoforms expressed in thegubernaculum, we carried out an S1 nuclease-mapping ex-periment using a single-stranded DNA probe that was spe-cific for 0/B rEsr1 mRNAs (0/B probe). The S1 nucleaseprotection assay experiments confirmed our earlier findingthat high levels of rEsr1 mRNA were produced in the gu-bernaculum and that, as in the uterus, the 0/B exon leadersequence predominated in the gubernaculum (O/B mRNAsrepresent .50% of total protected fragments) (Fig. 3). Inaddition to the large fragment corresponding to isoformswith the 0/B exon leader sequence, we also isolated a short-er protected fragment, indicating that other 59-UTR iso-forms also were produced in the tissues tested (Fig. 3). The0/B leader exon was not found in the liver, indicating thatEsr1 mRNA is produced from a different leader sequencein this organ.

Two Isoforms of ESR1 Are Present Within theGubernaculum

Because the Esr1 gene is able to generate by alternativesplicing at least two different isoforms of the ESR1 pro-tein, RT-PCR analysis were performed to analyze theoverall pattern of rEsr1 expression in the gubernaculum.

709ESTROGEN RECEPTORS IN THE GUBERNACULUM TESTIS

FIG. 6. Immunolocalization of ESR1 in 21.5-dpc (A, B, and D) and 22-dpp (E and G) gubernaculi using the 6F11 antibody (A and E) and theMC20 antibody (B) of AR at 21.5 dpc (C) and 22 dpp (F) and of ESR2 in22-dpp (H and I) gubernaculi. Mouse IgG was used as a control (D, G,and I). The ESR1 immunostaining was strong both in fetal and postnatalgubernaculum and was restricted to the nuclei of the parenchymal cells(arrows in A, B, and E). In the fetal gubernaculum, AR immunostainingwas strong and was found both in the nuclei of the parenchymal cells(arrows in C) and in the nuclei of the mesenchymal cells (arrowheads inC). Postnatally, AR labeling was restricted to the muscular cells (arrow-heads in F) and in the nuclei of vascular cells (open arrow in F). The ESR2immunostaining was consistently weak, only observed postnatally, andessentially localized in the nuclei of muscular cells (arrowheads in H)with only a few labeled parenchymal cells (not shown) and in vascularcells (open arrow in H). Bar 5 50 mm. Insets in A and C represent a 1.5-fold magnification.

Full-length rEsr1 cDNA (rER66) was detected in all thetissues and cells tested (Fig. 4B). In addition, a shorteramplicon also was detected in the gubernaculum, MMQcells, and the liver. This amplicon was approximately 600bp shorter than the full-length cDNA. Sequencing of thecorresponding cDNA sequences revealed that this differ-ence resulted from the deletion of exon 1 (538 bp) and thesplicing of the exon C leader sequence on coding exon 2(rER46).

Western blots of ESR1 were performed with the MC20and 6F11 antibodies directed against the carboxy- and ami-no-terminal, respectively, of the ESR1 protein. The MC20antibody recognized the two ESR1 isoforms, whereas the6F11 antibody recognized only the full-length protein (Fig.4C). When the Western blots were performed with theMC20 antibody, strong signals were observed for the 66-kDa isoform in gubernaculum aged from 17 to 20 dpc andfor the 46-kDa isoform in gubernaculum aged from 19 dpcto 3 dpp (Fig. 5A). The 66-kDa protein was consistentlybarely detectable between 4 and 12 dpp but was abundantagain by 22 dpp. The same pattern of expression was ob-served for this full-length ESR1 using the 6F11 antibody(not shown). In contrast, the 46-kDa protein was either veryweakly or not detected after 8 dpp. In all the Western blotexperiments, a 55-kDa nonspecific band was detected withboth ESR1 antibodies. This unspecific staining already hasbeen reported in the literature [41].

On the basis of fluctuations in the p66:p46 ratio calcu-lated using the relative intensity of the bands detected usingthe MC20 antibody, we have defined three main phases ofESR1 production (phases I–III) (Fig. 5B). During the firstphase (16–21 dpc), the 66-kDa form predominates over the46-kDa form. During the second phase (birth to 8 dpp), thep66:p46 ratio markedly decreases (P 5 0.003) to close to1, with the two ESR1 forms being detected in near-com-parable amounts. During the third phase (12–22 dpp), pro-duction of the 66-kDa form increases again, whereas ex-pression of the 46-kDa form is very low. As a result, thep66:p46 ratio in phase III is higher than that in phase II (P5 0.041), but it is not different from that in phase I (P 50.526).

Localization of Both Isoforms of ESR1 and ESR2 and ofAR in Gubernaculum

Immunostaining of ESR1 was observed in the nuclei ofparenchymal cells at 21.5 dpc using both the 6F11 antibodyand the MC20 antibody (Fig. 6, A and B, respectively),suggesting the absence of a differential location of the twoESR1 isoforms. In contrast, ESR2 immunostaining wasnever detected in the fetal gubernaculum (data not shown).The AR was detected in the nuclei of the parenchymal cellsand in the nuclei of the mesenchymal cells of the guber-naculum at this stage of development (Fig. 6C). The ESR1was still detectable in the parenchymal cells in the outerring of the gubernaculum of 22-dpp rats and also in exten-sions of parenchymal cells within the muscular tissue ofthis structure (Fig. 6E). In contrast to the strong staining ofESR1 observed in the parenchymal cells, a much weakerstaining of ESR2 was detected only postnatally and essen-tially in the muscular cells (Fig. 6H), with only few paren-chymal and vascular cells being labeled. After birth, ARimmunostaining was restricted to muscular areas and invascular cells (Fig. 6F).

DISCUSSION

The work of Donaldson et al. [32] revealed that the tes-tes are positioned near the bladder neck rather than withinthe scrotum in most homozygous Esr1 knockout mice [42,43]. This finding led us to investigate whether the guber-naculum expresses estrogen receptors and, therefore, likelywould be a direct target for estrogens during the process oftesticular descent and development in the rat.

Our RT-PCR analysis showed that Esr1 and Esr2

710 STAUB ET AL.

mRNAs were present in the gubernaculum at all stages ofdevelopment studied. The PEETA (Primer extension,Electrophoresis, Elution, Tailing, and Amplification) tech-nique revealed that Esr1 was expressed within the guber-naculum and that the expression level and the pattern ofdistribution of the various Esr1 mRNAs were similar tothose found in other structures. The presence of all pre-viously identified leader exons—the major form being O/B—in the gubernaculum, like in other organs, such as thepituitary, uterus, and liver, suggests that several isoformsof this receptor are present in these organs [44, 45]. Thisidea is consistent with the results obtained from the S1nuclease mapping, which showed that some of the Esr1mRNAs in the gubernaculum were generated by splicingof the 0/B leader sequence on coding exon 1 whereas oth-ers were generated using a different leader sequence. Boththe RT-PCR experiments using primers designed from theC leader sequence and coding exon 8 and the Western blotexperiments confirmed these findings. They revealed thepresence of mRNA transcripts corresponding to 46- and66-kDa ESR1 proteins. Thus, we conclude that two majorESR1 isoforms are produced in the gubernaculum and thatthese proteins are derived from isoform-specific mRNAtranscripts generated by alternative splicing of the C lead-er sequence on coding exon 1 or 2.

The present study establishes that the ESR1 proteinswere abundant in the parenchymal cells of the fetal andpostnatal gubernaculum testis. It also demonstrated majordifferences between the expression of ESR1 and that ofESR2. In fact, both Esr2 mRNA and protein were ex-pressed at lower levels than Esr1. Furthermore, ESR2 pro-tein could only be detected postnatally and was localizedprincipally in the muscular cells. Although the role playedby ESR2 in the gubernaculum muscular cells remains to beelucidated, it probably is not involved in the process oftesticular descent, because the ESR2 protein is absent inthe fetal gubernaculum. This idea is consistent with the factthat no abnormality in testicular positioning was noted inEsr2 knockout male mice [46]. In contrast, our resultsshowing that Esr1 is expressed in the gubernaculum at allages studied is consistent with the results showing malpo-sitioning of the testes occurs in Esr1 knockout mice, causedby a reduction in the volumetric capacity of the cremastersac (the inner part of the scrotum), leading to overdevel-opment of the gubernaculum [32, 47].

As expected, and considering the well-known involve-ment of androgens in testicular descent, Ar mRNA also wasdetected in the gubernaculum at all stages of development.We confirm that during fetal life, AR protein is present inthe mesenchymal core [48, 49], and we show that it also ispresent in the parenchymal cell ring of the fetal gubernac-ulum. After birth, AR localization is restricted to the mus-cular area, as described previously [49].

Considering the colocalization of ESR1 and AR in theparenchymal cells of the gubernaculum, it is tempting tohypothesize that estrogens and androgens cointervene indevelopment of the gubernaculum. Androgens are knownto stimulate gubernacular development [9], whereas estro-gens would be expected to inhibit it, which would be con-sistent with the abundant literature indicating that estradiol,DES, and xenoestrogens are all involved in the occurrenceof cryptorchidism [18–27]. We analyzed the promoting se-quence of the Ar gene in the rat and found two potentialestrogen-responsible elements (ERE mismatched motives),providing a possible molecular pathway via which estro-gens could influence androgen action. A role for the estro-

gen receptor in mediating androgen-driven cell growth anddifferentiation already has been proposed for AR in theprostate, where estrogens are suspected to limit androgen-induced prostatic growth [50]. Interestingly, the amino-ter-minal truncated product (ER46) reportedly acts as an in-hibitor of ER66 by competing for DNA binding and isthought to play a role in the genomic regulation of breastcancer proliferation [40, 51].

If we draw a parallel between the latter results and thedata presented herein, we hypothesize that when the 66-kDa form clearly exceeds that of the 46-kDa form (i.e.,during fetal life), the action of estrogens could preventgrowth of the gubernaculum. When the p66:p46 ratio isdecreasing, competitive inhibition of ER66 by ER46 couldreduce estrogen inhibition on androgen action and allowthe initiation of the growth of the organ, which actuallyoccurs during the corresponding time frame [47]. Betweenbirth and 8 dpp, when the 46-kDa form is expressed at leastas much as the 66-kDa form, inhibition of ESR1 couldoccur and be required at a time when the gubernaculumdevelops exponentially [17]. The strong resurgence of ex-pression of the 66-kDa form coincides with the end of gub-ernacular development [52]. Therefore, it is tempting to hy-pothesize that the fluctuations in the expression levels ofESR1 and the changes in the ratio of its two isoforms inthe parenchymal cells of the gubernaculum are part of thepossible regulatory mechanism by which estrogens couldmodulate androgen action during testicular descent. How-ever, the reality of this interplay between sex hormonesawaits further experiments demonstrating direct experimen-tal effects of estrogens on gubernaculum growth.

The lack of Cyp19a1 expression and the presence ofSrd5a1 in the gubernaculum are consistent with results ob-tained in a previous study [16]. These data indicate that theestrogens acting on the gubernaculum most likely originatefrom another source within the fetus, and even more likelyfrom the mother, because testicular descent and final posi-tioning of gonads within the scrotum are normal inCyp19a1 knockout male mice [53]. In contrast, our dataconfirm that conversion of testosterone to 5a-reduced me-tabolites can occur locally [16].

To summarize, we demonstrate here that two isoformsof ESR1, ER66 and its truncated variant ER46, are highlyexpressed in the gubernaculum testis before testicular trans-abdominal descent occurs. These proteins are translatedfrom isoform-specific mRNA transcripts generated by al-ternative splicing. The differences in both timing and levelsof expression of these two competitive ESR1 proteins mayhelp us to understand, through future studies, how estrogensmight regulate the growth of the gubernaculum during tes-ticular descent. We hypothesize that by modulating the an-drogen-dependent growth of gubernaculum, estrogens (to-gether with INSL3 and androgens) could be involved incoordinating events that control the positioning of the testeswithin the scrotum. This is important in the context of thesearch for the cause of cryptorchidism and the xenoestro-gen-induced abnormalities of the male urogenital tract, andit is consistent with a recent publication indicating thatESR1 is overexpressed in the cremasteric muscle of crypt-orchid boys [54].

ACKNOWLEDGMENTS

The authors wish to thank Julie Lassurguere-Labbe and SheilaMacPherson for their help at various stages of the experiments and AlainLe Nedic for taking care of the animals.

711ESTROGEN RECEPTORS IN THE GUBERNACULUM TESTIS

REFERENCES

1. Toppari J, Larsen JC, Christiansen P, Giwercman A, Grandjean P,Guillette LJ Jr, Jegou B, Jensen TK, Jouannet P, Keiding N, LeffersH, McLachlan JA, Meyer O, Muller J, Rajpert-De Meyts E, ScheikeT, Sharpe R, Sumpter J, Skakkebaek NE. Male reproductive healthand environmental xenoestrogens. Environ Health Perspect 1996; 104:741–803.

2. Jegou B, Auger J, Multigner L, Pineau C, Thonneau P, Spira A, Jouan-net P. The saga of the sperm count decrease in humans and wild andfarm animals. In: Press CR (ed.), The Male Gamete: From Science toClinical Applications. Vienna, IL: Claude Gagnon; 1999:445–454.

3. Sharpe RM, Irvine DS. How strong is the evidence of a link betweenenvironmental chemicals and adverse effects on human reproductivehealth? BMJ 2004; 328:447–451.

4. Paulozzi LJ. International trends in rates of hypospadias and crypt-orchidism. Environ Health Perspect 1999; 107:297–302.

5. Boisen KA, Kaleva M, Main KM, Virtanen HE, Haavisto AM,Schmidt IM, Chellakooty M, Damgaard IN, Mau C, Reunanen M,Skakkebaek NE, Toppari J. Difference in prevalence of congenitalcryptorchidism in infants between two Nordic countries. Lancet 2004;363:1264–1269.

6. Adami HO, Bergstrom R, Mohner M, Zatonski W, Storm H, EkbomA, Tretli S, Teppo L, Ziegler H, Rahu M. Testicular cancer in ninenorthern European countries. Int J Cancer 1994; 59:33–38.

7. Scorer CG. The descent of the testis. Arch Dis Child 1964; 39:605–609.

8. Adham IM, Emmen JM, Engel W. The role of the testicular factorINSL3 in establishing the gonadal position. Mol Cell Endocrinol2000; 160:11–16.

9. Emmen JM, McLuskey A, Adham IM, Engel W, Grootegoed JA,Brinkmann AO. Hormonal control of gubernaculum development dur-ing testis descent: gubernaculum outgrowth in vitro requires both in-sulin-like factor and androgen. Endocrinology 2000; 141:4720–4727.

10. Ivell R, Bathgate RA. Reproductive biology of the relaxin-like factor(RLF/INSL3). Biol Reprod 2002; 67:699–705.

11. Gorlov IP, Kamat A, Bogatcheva NV, Jones E, Lamb DJ, Truong A,Bishop CE, McElreavey K, Agoulnik AI. Mutations of the GREATgene cause cryptorchidism. Hum Mol Genet 2002; 11:2309–2318.

12. Ferlin A, Simonato M, Bartoloni L, Rizzo G, Bettella A, Dottorini T,Dallapiccola B, Foresta C. The INSL3-LGR8/GREAT ligand-receptorpair in human cryptorchidism. J Clin Endocrinol Metab 2003; 88:4273–4279.

13. Heyns CF, Human HJ, Werely CJ, De Klerk DP. The glycosamino-glycans of the gubernaculum during testicular descent in the fetus. JUrol 1990; 143:612–617.

14. Emmen JM, McLuskey A, Grootegoed JA, Brinkmann AO. Androgenaction during male sex differentiation includes suppression of cranialsuspensory ligament development. Hum Reprod 1998; 13:1272–1280.

15. George FW, Peterson KG. Partial characterization of the androgenreceptor of the newborn rat gubernaculum. Biol Reprod 1988; 39:536–539.

16. George FW. Developmental pattern of 5a-reductase activity in the ratgubernaculum. Endocrinology 1989; 124:727–732.

17. Hutson JM, Hasthorpe S, Heyns CF. Anatomical and functional as-pects of testicular descent and cryptorchidism. Endocr Rev 1997; 18:259–280.

18. Grocock CA, Charlton HM, Pike MC. Role of the fetal pituitary incryptorchidism induced by exogenous maternal estrogen during preg-nancy in mice. J Reprod Fertil 1988; 83:295–300.

19. Shono T, Hutson JM, Watts L, Goh DW, Momose Y, MiddlesworthB, Zhou B, Ramm-Anderson S. Scanning electron microscopy showsinhibited gubernacular development in relation to undescended testesin estrogen-treated mice. Int J Androl 1996; 19:263–270.

20. McLachlan JA, Newbold RR, Bullock B. Reproductive tract lesionsin male mice exposed prenatally to diethylstilbestrol. Science 1975;190:991–992.

21. Perez-Martinez C, Garcia-Iglesias MJ, Ferreras-Estrada MC, Bravo-Moral AM, Espinosa-Alvarez J, Escudero-Diez A. Effects of in uteroexposure to zeranol or diethylstilbestrol on morphological develop-ment of the fetal testis in mice. J Comp Pathol 1996; 114:407–418.

22. Newbold RR, McLachlan JA. Transplacental hormonal carcinogene-sis: diethylstilbestrol as an example. Prog Clin Biol Res 1996; 394:131–147.

23. Husmann DA. Excess fetal estrogen as an etiological cause of humancryptorchidism. J Urol 2000; 164:1696.

24. Hadziselimovic F, Geneto R, Emmons LR. Elevated placental estra-

diol: a possible etiological factor of human cryptorchidism. J Urol2000; 164:1694–1695.

25. Li X, Makela S, Streng T, Santti R, Poutanen M. Phenotype charac-teristics of transgenic male mice expressing human aromatase underubiquitin C promoter. J Steroid Biochem Mol Biol 2003; 86:469–476.

26. Toppari J. Environmental endocrine disrupters and disorders of sexualdifferentiation. Semin Reprod Med 2002; 20:305–312.

27. Sharpe RM. The ‘estrogen hypothesis’—where do we stand now? IntJ Androl 2003; 26:2–15.

28. Rajfer J, Walsh PC. Hormonal regulation of testicular descent: exper-imental and clinical observations. J Urol 1977; 118:985–990.

29. Imajima T, Shono T, Zakaria O, Suita S. Prenatal phthalate causescryptorchidism postnatally by inducing transabdominal ascent of thetestis in fetal rats. J Pediatr Surg 1997; 32:18–21.

30. Emmen JM, McLuskey A, Adham IM, Engel W, Verhoef-Post M,Themmen AP, Grootegoed JA, Brinkmann AO. Involvement of insu-lin-like factor 3 (Insl3) in diethylstilbestrol-induced cryptorchidism.Endocrinology 2000; 141:846–849.

31. Nef S, Shipman T, Parada LF. A molecular basis for estrogen-inducedcryptorchidism. Dev Biol 2000; 224:354–361.

32. Donaldson KM, Tong SY, Washburn T, Lubahn DB, Eddy EM, Hut-son JM, Korach KS. Morphometric study of the gubernaculum in maleestrogen receptor mutant mice. J Androl 1996; 17:91–95.

33. Chomczynski P, Sacchi N. Single-step method of RNA isolation byacid guanidinium thiocyanate-phenol-chloroform extraction. AnalBiochem 1987; 162:156–159.

34. Freyschuss B, Grandien K. The 59 flank of the rat estrogen receptorgene: structural characterization and evidence for tissue- and species-specific promoter utilization. J Mol Endocrinol 1996; 17:197–206.

35. Flouriot G, Nestor P, Kenealy MR, Pope C, Gannon F. An S1 nucleasemapping method for detection of low abundance transcripts. AnalBiochem 1996; 237:159–161.

36. Flouriot G, Pope C, Kenealy MR, Gannon F. Improved efficiency forprimer extension by using a long, highly labeled primer generatedfrom immobilized single-stranded DNA templates. Nucleic Acids Res1997; 25:1658–1659.

37. Ausubel FM, Brent R, Kingstom RE, Moore DD, Seidman JG, SmithJA, Struhl K. Current Protocols in Molecular Biology. New York:John Wiley & Sons; 1998.

38. Morcol T, Subramanian A. A red-dot-blot protein assay technique inthe low nanogram range. Anal Biochem 1999; 270:75–82.

39. Zhou Q, Nie R, Prins GS, Saunders PT, Katzenellenbogen BS, HessRA. Localization of androgen and estrogen receptors in adult malemouse reproductive tract. J Androl 2002; 23:870–881.

40. Flouriot G, Brand H, Denger S, Metivier R, Kos M, Reid G, Sonntag-Buck V, Gannon F. Identification of a new isoform of the humanestrogen receptor-a (hER-a) that is encoded by distinct transcripts andthat is able to repress hER-a activation function 1. EMBO J 2000;19:4688–4700.

41. Saji S, Jensen EV, Nilsson S, Rylander T, Warner M, Gustafsson JA.Estrogen receptors a and b in the rodent mammary gland. Proc NatlAcad Sci U S A 2000; 97:337–342.

42. Lubahn DB, Moyer JS, Golding TS, Couse JF, Korach KS, SmithiesO. Alteration of reproductive function but not prenatal sexual devel-opment after insertional disruption of the mouse estrogen receptorgene. Proc Natl Acad Sci U S A 1993; 90:11162–11166.

43. Korach KS. Insights from the study of animals lacking functionalestrogen receptor. Science 1994; 266:1524–1527.

44. Pendaries C, Darblade B, Rochaix P, Krust A, Chambon P, KorachKS, Bayard F, Arnal JF. The AF-1 activation-function of ERa may bedispensable to mediate the effect of estradiol on endothelial NO pro-duction in mice. Proc Natl Acad Sci U S A 2002; 99:2205–2210.

45. Okuda Y, Hirata S, Watanabe N, Shoda T, Kato J, Hoshi K. Novelsplicing events of untranslated first exons in human estrogen receptora (ERa) gene. Endocr J 2003; 50:97–104.

46. Krege JH, Hodgin JB, Couse JF, Enmark E, Warner M, Mahler JF,Sar M, Korach KS, Gustafsson JA, Smithies O. Generation and re-productive phenotypes of mice lacking estrogen receptor beta. ProcNatl Acad Sci U S A 1998; 95:15677–15682.

47. Bartlett JE, Washburn T, Eddy EM, Korach KS, Temelcos C, HutsonJM. Early development of the gubernaculum and cremaster sac inestrogen receptor knockout mice. Urol Res 2001; 29:163–167.

48. Cooke PS, Young PF, Cunha GR. A new model system for studyingandrogen-induced growth and morphogenesis in vitro: the bulboure-thral gland. Endocrinology 1987; 121:2161–2170.

49. Husmann DA, McPhaul MJ. Localization of the androgen receptor inthe developing rat gubernaculum. Endocrinology 1991; 128:383–387.

712 STAUB ET AL.

50. Weihua Z, Makela S, Andersson LC, Salmi S, Saji S, Webster JI,Jensen EV, Nilsson S, Warner M, Gustafsson JA. A role for estrogenreceptor beta in the regulation of growth of the ventral prostate. ProcNatl Acad Sci U S A 2001; 98:6330–6335.

51. Denger S, Reid G, Kos M, Flouriot G, Parsch D, Brand H, KorachKS, Sonntag-Buck V, Gannon F. ERa gene expression in human pri-mary osteoblasts: evidence for the expression of two receptor proteins.Mol Endocrinol 2001; 15:2064–2077.

52. Hrabovszky Z, Di Pilla N, Yap T, Farmer PJ, Hutson JM, Carlin JB.

Role of the gubernacular bulb in cremaster muscle development ofthe rat. Anat Rec 2002; 267:159–165.

53. Robertson KM, Simpson ER, Lacham-Kaplan O, Jones ME. Charac-terization of the fertility of male aromatase knockout mice. J Androl2001; 22:825–830.

54. Przewratil P, Paduch DA, Kobos J, Niedzielski J. Expression of estro-gen receptor alpha and progesterone receptor in children with unde-scended testicle previously treated with human chorionic gonadotro-pin. J Urol 2004; 172:1112–1116.