Does a Nonclassical Signaling Mechanism Underlie an Increase of Estradiol-Mediated Gonadotropin-Releasing Hormone Receptor Binding in Ovine Pituitary Cells?

BIOLOGY OF REPRODUCTION 85, 770–778 (2011)Published online before print 6 July 2011.DOI 10.1095/biolreprod.111.091926

Does a Nonclassical Signaling Mechanism Underlie an Increase of Estradiol-MediatedGonadotropin-Releasing Hormone Receptor Binding in Ovine Pituitary Cells?1

Tracy L. Davis,3 Jennifer D. Whitesell,4 Jeremy D. Cantlon, Colin M. Clay, Terry M. Nett2

Animal Reproduction and Biotechnology Laboratory, Department of Biomedical Sciences, Colorado State University,Fort Collins, Colorado

ABSTRACT

Estradiol-17beta (E2) is the major regulator of GnRH receptor(GnRHR) gene expression and number during the periovulatoryperiod; however, the mechanisms underlying E2 regulation ofthe GNRHR gene remain undefined. Herein, we find that E2conjugated to BSA (E2-BSA) mimics the stimulatory effect of E2on GnRH binding in primary cultures of ovine pituitary cells. Thetime course for maximal GnRH analog binding was similar forboth E2 and E2-BSA. The ability of E2 and E2-BSA to increaseGnRH analog binding was blocked by the estrogen receptor (ER)antagonist ICI 182,780. Also, increased GnRH analog binding inresponse to E2 and the selective ESR1 agonist propylpyrazoletriol was blocked by expression of a dominant-negative form ofESR1 (L540Q). Thus, membrane-associated ESR1 is the likelycandidate for mediating E2 activation of the GNRHR gene. AscAMP response element binding protein (CREB) is an establishedtarget for E2 activation in gonadotrophs, we next explored apotential role for this protein as an intracellular mediator of theE2 signal. Consistent with this possibility, adenoviral-mediatedexpression of a dominant-negative form of CREB (A-CREB)completely abolished the ability of E2 to increase GnRH analogbinding in primary cultures of ovine pituitary cells. Finally, thepresence of membrane-associated E2 binding sites on ovinepituitary cells was demonstrated using a fluorescein isothiocy-anate conjugate of E2-BSA. We suggest that E2 regulation ofGnRHR number during the preovulatory period reflects amembrane site of action and may proceed through a nonclas-sical signaling mechanism, specifically a CREB-dependentpathway.

GnRH receptors, membrane estrogen receptors, ovine, pituitary

INTRODUCTION

The most dynamic time during the estrous cycle withrespect to hypothalamic-pituitary-ovarian interaction is the

preovulatory period when serum concentrations of progester-one decline as a result of prostaglandin-F

2a (PGF2a)-mediated

luteolysis and serum concentrations of estradiol-17b (E2)increase with the development of the preovulatory follicle. Thispattern of ovarian hormone secretion leads to importantchanges at the hypothalamus and anterior pituitary. Duringthe preovulatory period, there is an E2-induced increase inGnRH receptor (GnRHR) numbers in the anterior pituitary [1,2] followed by an increase in release of GnRH [3]. Thus, bothheightened pituitary responsiveness to GnRH and increasedGnRH secretion underlie the generation of the ovulatory LHsurge. These effects appear to be conserved across mostmammalian species, including sheep [1, 3], cattle [2, 4],monkeys [5], mice [6], and rats [7, 8].

The stimulatory effect of E2 on the number of GnRHR isevident in the absence of hypothalamic input and in theprimary cultures of ovine [9], rat [10, 11], and murine [6]pituitary cells. Thus, this effect of E2 is mediated directly at theanterior pituitary gland. To determine whether the largenumber of GnRHR during the preovulatory period is the resultof increased expression of the GnRHR gene, severalinvestigators have measured GNRHR mRNA after inductionof luteolysis in sheep [12–14]. Concentrations of GNRHRmRNA are increased as early as 12 h after luteolysis andremain high at 24 h [12]. In the early preovulatory period,increased GNRHR gene expression preceded an increase in thenumber of GnRHR [12], and the maximal number of GnRHRwere observed later in the preovulatory period near the onset ofthe LH surge [1, 14], which occurs approximately 56 h after theadministration of PGF

2a [1]. These temporal relationshipssupport the hypothesis that increased amounts of GNRHRmRNA lead to greater numbers of GnRHR that in turnmaximizes pituitary sensitivity to GnRH in preparation for theLH surge. Thus, the density of GnRHR on gonadotrophsdetermines their ability to respond to GnRH [15].

To date, virtually nothing is known about the mechanismsunderlying the regulation of GnRHR expression by E2. This istroubling because E2 is considered the primary regulator ofgene expression and subsequent number of GnRHR ongonadotrophs. Based on the ability of membrane-impermeableE2 conjugates to elicit an acute decrease in LH secretionindependent of a hypothalamic site of action [16, 17], wespeculated that E2 activation of the GNRHR gene may beinitiated at a plasma membrane site of action and increase thenumber of GnRHR, as determined by GnRH binding, in anonclassical signaling manner. Herein, we show that E2conjugated to BSA (E2-BSA) mimics the effects of free E2on enhancing GnRHR binding in ovine pituitary cells.Furthermore, we suggest that ESR1 (also known as ERa) isthe likely candidate for mediating E2 input to the GNRHR genevia a mechanism that is dependent on a member of the cAMPresponse element binding (CREB) family of transcriptionfactors.

1Supported by the National Research Initiative Competitive Grant no.2005-35203-15376 from USDA Cooperative State Research, Education,and Extension Service; a grant from the Colorado State UniversityExperiment Station; NIH Training Grant T32HD07031-29 (T.L.D.); andNIH HD0655943. T.M.N. has previously consulted for BoehringerIngelheim Pharmaceuticals, Inc., and Mylan Laboratories, Inc., and hasreceived lecture fees from the Society of Toxicology. He holds an equityposition in Gonex, Inc.2Correspondence: FAX: 970 491 3557; e-mail: [email protected] address: Department of Animal and Veterinary Science,University of Idaho, Moscow, ID 83844.4Current address: Neuroscience Program, University of Colorado,Anschutz Medical Campus, Aurora, CO 80045.

Received: 22 February 2011.First decision: 21 March 2011.Accepted: 7 June 2011.� 2011 by the Society for the Study of Reproduction, Inc.eISSN: 1529-7268 http://www.biolreprod.orgISSN: 0006-3363

770

Dow

nloaded from w

ww

.biolreprod.org.

MATERIALS AND METHODS

Materials

Collagenase Type II, hylauronidase Type V, deoxyribonuclease I, fluo-rescently labeled E2-BSA, E2, and Sephadex G-25 were purchased from Sigma(St. Louis, MO). Estradiol conjugated to BSA was purchased from Steraloids Inc.(batch R288; Newport, RI). Fetal bovine serum (FBS) was purchased fromGemini Bio-Products (West Sacramento, CA). Penicillin/streptomycin andDulbecco modified Eagle medium (DMEM) were purchased from Mediatech,Inc. (Herndon, VA). Matrigel was purchased from BD Biosciences (San Jose,CA). Glass bottom culture dishes were purchased from MatTek Corp. (Ashland,MA). The estrogen receptor (ER) antagonist (ICI 182,780) and propylpyrazoletriol (PPT) were purchased from Tocris Bioscience (Ellisville, MO). Theantibody to the estrogen receptor (clone H222) was purchased from Lab VisionCorp. (Fremont, CA). Concanavalin-A conjugated to Alexa 594 was obtainedfrom Molecular Probes (Invitrogen, Carlsbad, CA). All the remaining chemicalswere purchased from Sigma or Fisher Scientific (Fairlawn, NJ).

Preparation of the E2 Conjugate and Fluorescently LabeledE2 Conjugate

The E2 conjugated to BSA (E2-BSA; 1,3,5(10)-estratrien-3, 17b-diol-6-one6-carboxymethyl oxime:BSA) contained six molecules of E2 linked to eachmolecule of BSA. To remove free E2 from E2 conjugated to BSA, 10 mg ofE2-BSA were resuspended in 1 ml PBS. Five ml of diethyl ether were added,and the contents of the tube were vortexed for 1 min and then frozen in a dry-ice methanol bath. The organic phase was decanted and discarded. Extractionof the E2-BSA stock was repeated an additional six times. Based on previousextractions, approximately 20 pg/ml of free E2 remained within the E2-BSAconjugate after seven extractions (Arreguin-Arevalo and Nett, unpublished). Asdescribed previously [17], 6-keto-17b-estradiol-6-carboxymethyl oxime (E2-6-CMO; Steraloids Inc.) was conjugated to a 15-amino acid sequence (N-terminus-SGGEVVVDQPMERLY-C-terminus) on the amino group of theserine reside. The conjugation reaction was added to a Sephadex LH20 columnto separate E2-6-CMO from the conjugated form (E2-PEP). The presence of theconjugate and absence of free E2-6-CMO in the first peak eluted from theSephadex LH20 column was confirmed using a Waters QTOF-microelectrospray mass spectrometer with the sample introduced by direct infusion(Macromolecular Resources, Colorado State University, CO). The massspectrometry confirmed that there was one molecule of E2 linked to eachpeptide molecule and that the conjugate was devoid of free E2. Fluorescentlylabeled E2 conjugate (E2-BSA-FITC) was prepared as previously described toremove potentially contaminating free fluorescein isothiocyanate (FITC). Toseparate FITC from the E2-BSA-FITC, 1 mg of E2-BSA-FITC diluted in PBSwas added to a Sephadex G-25 column (0.6 3 11 cm) and eluted with 0.05 MPBS-gelatin [18]. One-ml fractions were collected, and the fraction containingthe dark yellow band (fraction 3) was dialyzed at 48C against 1 L of PBS for 24h to remove any remaining free FITC.

Dissociation and Culture of Ovine Pituitary Cells

All the procedures involving animals were approved by Colorado StateUniversity Animal Care and Use Committee and were in compliance withNational Institutes of Health (NIH) guidelines. Sexually mature ewes that hadbeen ovariectomized for at least 30 days prior to pituitary collection were used.Anterior pituitary glands were collected from ewes following anesthesia withsodium pentobarbital and exsanguination. Pituitaries were immediately placedin dissociation medium (137 mM NaCl, 25 mM HEPES, 10 mM glucose, 5mM KCl; pH 7.4) and transported to the laboratory on ice. Pituitaries weredissociated as previously described [19] with minor modifications. Briefly,pituitary tissue was sectioned into 0.5 mm slices using a Stadie-Riggs tissueslicer. Pituitary slices were washed with dissociation medium and transferred toa flask containing dissociation medium with collagenase Type II (3850 units),hylauronidase Type V (100 units), and deoxyribonuclease I (100 ng). Pituitarytissue was incubated at 378C in a Dubnoff metabolic shaker for 90 min. Cellswere collected by centrifugation and washed with dissociation medium withoutenzymes; following the final wash, the cells resuspended in phenol red-freeDMEM medium containing 10% charcoal-stripped FBS. For all the cultureexperiments, 2 3 106 cells were plated in 6-well tissue culture dishes at 378C inan atmosphere of 5% CO

2for at least 36 h prior to addition of the treatments.

Effect of E2 on Percent Change of GnRH Receptors

An initial experiment was conducted to determine if a short duration ofexposure to E2, after which the E2 was removed from the culture and the cells

were incubated for an additional 12 h, would increase the number of GnRHanalog binding to GnRHR in ovine pituitary cells compared to vehicle-treatedcells. Anterior pituitary glands were collected from ovariectomized ewes duringthe month of May and prepared as previously described. Pituitary cells weretreated for 15, 30, or 60 min at 378C with either vehicle (0.01% ethanol), E2 (40nM), or E2-BSA (40 nM). The concentration of E2-BSA was based on themolecular weight of BSA (66 kDa). After 15-, 30-, and 60-min treatments, themedium was removed, cells were washed twice with PBS, and E2-free mediumwas added to the cells. Cells were then cultured until 12 h posttreatment, whichhas been shown to be the time of maximal increase in number of GnRHR inovine pituitary cells [9]. An additional treatment was included in whichpituitary cells were cultured with E2 (40 nM) or E2-BSA (40 nM) for the entire12 h incubation. Relative number of GnRHR was determined by radioreceptorassay [20]. Data are presented as GnRH analog binding as a percent of controlfor three pituitaries.

Because the previous data were obtained using a single, rather high,concentration of E2 and E2-BSA, an experiment was conducted using variousconcentrations (0.1 pM to 10 nM) of E2 or E2-BSA. Vehicle (0.01% ethanol)or the various concentrations of E2 or E2-BSA were added to the cells andincubated for 12 h. The relative number of GnRHR was determined byradioreceptor assay in triplicate for each treatment. Data are presented as GnRHanalog binding as a percent of control and replicated three times using anteriorpituitary glands collected from ovariectomized ewes during the months ofJanuary and February. Because of the controversy of using E2-BSA conjugateand the question of removal of all the free E2, an additional study wasconducted following the experiments described above using pituitary cellstreated with various concentrations (0.1 pM to 10 nM) of E2-PEP for 12 h. Therelative number of GnRHR was determined by radioreceptor assay. Data arepresented as GnRH analog binding as a percent of control and replicated fourtimes using anterior pituitary glands collected from ovariectomized ewes duringthe anestrous period (month of July).

To determine if the time course for increasing the number of GnRHR ingonadotrophs was similar for E2 and E2-BSA, pituitary cells were culturedwith vehicle (0.01% ethanol), 0.1 nM E2, or 10 nM E2-BSA for 3, 6, 9, or 12 hbefore the end of the incubation. Relative number of GnRHR was determinedin duplicate by radioreceptor assay and reported as GnRH analog binding as apercent of control. The experiment was replicated four times using anteriorpituitary glands collected from ovariectomized ewes during the months ofJanuary and February.

To determine if the effects of the estrogens could be inhibited, cells werecultured in the presence or absence of an ER antagonist, ICI 182,780. Pituitarycells were cultured for 12 h with 0.1 nM E2 or 10 nM E2-BSA with or without10 lM ICI 182,780, and then the relative number of GnRHR was determined intriplicate for each treatment. Data were replicated four times using anteriorpituitary glands collected from ovariectomized ewes during the months ofJanuary and February and are presented as percent change of GnRH analogbinding.

Effect of a Dominant-Negative ESR1 Mutant on E2and an ESR1 Agonist-Stimulated Increase in GnRH Receptors

To determine if the increase in GnRH analog binding was mediated throughESR1, a dominant-negative mutant for ESR1 (L540Q) and PPT, a selectiveESR1 agonist, were used. The dominant-negative ESR1 mutant was kindlyprovided by Dr. J. Larry Jameson (Northwestern University, Feinberg Schoolof Medicine, Chicago, IL). This form of ESR1 has a mutation within theCOOH-terminal activation domain and exhibits similar binding affinity for E2as the wild-type receptor [21]. Further, it has been suggested that this mutantESR1 functions as a dominant-negative protein by its inability to recruit thesteroid receptor coactivator proteins [22]. An adenovirus was constructed usinga bacterial recombination of the L540Q dominant-negative mutant inpAdTrackCMV with pAd-Easy in BJ5183 Escherichia coli as described byHe et al. [23]. Pituitary cells were infected for 24 h with the adenoviral vectorcontaining the L540Q dominant-negative mutant form of ESR1. Twenty-fourhours postinfection, pituitary cells were treated with either 4 nM E2 or 0.82 nMPPT for 12 h, and GnRH analog binding was determined by radioreceptorassay. The experiment was replicated three times.

Effect of a Dominant-Negative Inhibitor of CREBon E2-Stimulated Increase in GnRH Receptors

To determine if the CREB pathway was involved in the estrogen-stimulatedincrease in numbers of GnRH receptors, a dominant-negative inhibitor ofCREB (A-CREB) was used. The dominant-negative inhibitor, A-CREB, waskindly provided by Charles Vinson (Laboratory of Biochemistry, NationalCancer Institute, National Institutes of Health, Bethesda, MD). The dominant-

INCREASE IN NUMBER OF GnRHR VIA MEMBRANE ER 771

Dow

nloaded from w

ww

.biolreprod.org.

negative form of A-CREB prevents the basic region of wild-type CREB frombinding to DNA [24]. The coding sequence of A-CREB including the N-terminal FLAG epitope was amplified from pRc/CMV, and a Kpn1 site wasengineered immediately before the start codon. The PCR product was digestedby Kpn1 and HindIII and inserted into pAdTrackCMV at these sites.Adenovirus was constructed using bacterial recombination of A-CREB inpAdTrackCMV with pAd-Easy in BJ5183 E. coli using standard procedures[23]. Cells were infected for 24 h with adenoviral vectors containing A-CREBor GFP, which was used as a control. Cells were treated with 0.1 nM E2 for 12h, and GnRH analog binding was determined by radioreceptor assay. Theexperiment was replicated five times.

Cellular Localization of Estrogen Receptors

Dissociated pituitary cells were plated on Matrigel-coated 35 mm glassbottom culture dishes. After overnight incubation, the pituitary cells werewashed twice with PBS and incubated at room temperature for 2 min with 10nM E2-BSA-FITC in PBS. Cells were washed three times with PBS after E2-BSA-FITC labeling and incubated 10 min with 2% paraformaldehyde. Cellswere washed following fixation, and the plasma membranes were labeled withconcanavalin-A conjugated to Alexa 594. To inhibit binding of E2-BSA-FITCto the plasma membrane, cells were pretreated with E2 (10 lM), an antibody

directed against the ligand-binding domain of the ER (H222 clone; 1 lg), or anER antagonist (ICI 182,780; 10 lM) at room temperature for 10 min. With thepretreatments remaining in the PBS, 10 nM E2-BSA-FITC was added andincubated at room temperature for 2 min. Confocal images were acquired usinga 488–594-nm argon ion laser.

Radioreceptor Assay

After treatment, cells were washed with PBS and gently scraped from theculture dish wells, pelleted by centrifugation (750 3 g), and resuspended in200 ll GnRH receptor assay buffer (10 mM Tris-HCl, 1 mM CaCl

2, 0.1%

BSA, pH 7.4). One million cells were incubated with 0.2 nM [125I]-D-Ala6-GnRH-ethylamide at 48C for 4 h. Nonspecific binding was determined foreach treatment well by incubating 1 3 106 cells with 0.2 lM unlabeled D-Ala6-GnRH-ethylamide. After incubation, 3 ml ice-cold GnRH receptorbuffer were added, and the cells centrifuged for 15 min at 750 3 g.Supernatants were decanted, and radioactivity in the cell pellet wasquantified. Specific binding was calculated by subtracting nonspecificbinding from the total binding.

Statistical Analyses

The relative changes in number of GnRHR caused by treatment wasreported as a change of percentage of the specific binding of [125I]D-Ala6-GnRH-ethylamide in E2-treated cells compared to percentage of the specificbinding of [125I]D-Ala6-GnRH-ethylamide in untreated control cells (control¼100% binding). Differences in GnRH analog binding were determined byANOVA using the general linear model of SAS. Differences in GnRH analogbinding between treated and untreated cells were separated using the least-significant-difference test. Data are presented as mean 6 SEM.

RESULTS

E2 and E2 Conjugates Increase GnRH Analog Bindingin Ovine Pituitary Cells

Estradiol at 0.1 pM to 10 nM increased the percent of GnRHanalog binding above the control (Fig. 1A). Although a greaterconcentration was required (10 nM), treatment with E2-BSAalso led to a significant increase in the percent of GnRH analogbinding and to the same total increase as that induced by E2.Estradiol conjugated to the 15-amino acid peptide (E2-PEP)increased the percent of GnRH analog binding above thecontrol in a dose response similar to that of E2 (Fig. 1B). Anincrease was observed with E2-PEP treatment at 1 pM to 10nM, albeit at a lesser extent than when the E2 and E2-BSAstudy was conducted, presumably due to the sheep being inanestrus with fewer pituitary estrogen receptors [25] anddecreased responsiveness (Davis and Nett, unpublishedresults).

A Similar Time Course of Increased GnRH Analog BindingIs Evident for Pituitary Cells Treated with Either E2or E2-BSA

If E2 and E2-BSA are working through a similar mechanismto elicit enhanced GnRH analog binding in pituitary cells, thena similar time course of response should be evident. Consistentwith this, 0.1 nM E2 or 10 nM E2-BSA increased GnRHanalog binding at 6, 9, or 12 h but not at either 0 or 3 h (Fig.2A). Also, we sought to determine whether a short-termexposure (15 min) to E2 (40 nM) followed by removal of E2and a 12 h incubation is equivalent to exposure of pituitarycells with E2 for an entire 12 h. Thus, primary cultures of ovinepituitary cells were cultured with E2 for 0, 15, 30, 60 min or 12h. Consistent with a relatively rapid signaling event, a brief 15min exposure to E2 followed by removal of the E2 for theremaining 12 h incubation period was sufficient to increaseGnRH analog binding to the same extent as a 12 h exposure(Fig. 2B).

FIG. 1. Effect of dose of E2 and E2-BSA (A) and E2-PEP (B) on relativenumber of GnRH receptors as indicated by GnRH analog binding inpituitary cells. Treatments were added 12 h before the end of theincubation. Binding is represented as a percent of the control (100%).Values are the mean 6 SEM for three replicates (A) and four replicates (B).*P , 0.05.

772 DAVIS ET AL.

Dow

nloaded from w

ww

.biolreprod.org.

E2 and E2-BSA Increase in GnRH Analog Binding

Is Sensitive to Blockage by ICI 182,780

Next we sought to determine if the ability of E2 and E2-BSA to increase GnRH analog binding is mediated by one ofthe classical forms of ER (ESR1 or ESR2, also known asERa and ERb, respectively). Pituitary cells were culturedwith either 0.1 nM E2 or 10 nM E2-BSA in the presence ofthe pure ER antagonist ICI 182,780. The inclusion of ICI182,780 effectively blocked the increase in GnRH analogbinding in response to either E2 or E2-BSA (Fig. 3). Thus, itis likely that the ability of E2 to enhance pituitary number ofGnRHR is mediated via one or both of the classical ERisoforms rather than via novel ER proteins such as GPR30[26, 27].

Adenoviral-Mediated Overexpression of Dominant-Negative ESR1 Blocks Both E2- and PPT-Mediated Increasein GnRH Analog Binding

Based on the previous data, it appears that the ability of E2to enhance GnRH analog binding is sensitive to antagonism ofone of the classical isoforms of ER. Given that E2-mediatednegative feedback on secretion of LH is dependent on ESR1 asdetermined by the use of ESR1 knockout mice [28–32], wereasoned that the likely ER subtype mediating the increase inGnRH analog binding is ESR1. To test this possibility, weexamined the impact of a dominant-negative form of ESR1(L540Q) on the ability of both E2 and ESR1 agonist (PPT) toincrease GnRH analog binding. The dominant-negative L540Qmutant effectively blocked both E2- and PPT-enhanced GnRHanalog binding in primary cultures of ovine pituitary cells (Fig.4).

Adenoviral-Mediated Overexpression of A-CREB BlocksE2-Induced GnRH Analog Binding

Cumulatively, our data support the notion that E2 regulationof the GnRHR number may reflect a membrane-initiatedsignaling event or a nonclassical signaling mechanismmediated via ESR1. The ability of membrane forms of ESR1to integrate signaling through both AP-1 [33, 34] and CREB iswell established [35, 36]. Furthermore, recent work hasestablished that E2 leads to rapid phosphorylation of CREBin ovine gonadotrophs [37]. To determine if this pathwaycontributes to E2 regulation of GnRHR number, we applied anexperimental paradigm based on overexpression of a dominant-negative form of CREB termed A-CREB. Consistent with apotential role for CREB family members in E2 activation of theGnRHR gene, the ability of 0.1 nM E2 to enhance GnRHanalog binding was blocked by adenoviral-mediated overex-pression of A-CREB but not GFP (included as a control forinfection) (Fig. 5).

Membrane Fluorescence Is Evident on Ovine Pituitary CellsIncubated with FITC Conjugated to E2-BSA

To determine if plasma membrane-associated E2 bindingsites are detectable, ovine pituitary cells were incubated withE2-BSA-FITC (Fig. 6). Importantly, membrane fluorescencewas attenuated when cells were preincubated with either free

FIG. 2. Effect of duration of E2 treatment on relative number of GnRHreceptors as indicated by GnRH analog binding. A) E2 (0.1 nM) or E2-BSA(10 nM) were added for 0, 3, 6, 9, or 12 h prior to the end of theincubation. Pituitary cells for all the points were cultured for 12 h. Valuesare the mean 6 SEM for four replicates. B) Pituitary cells were incubatedfor 15, 30, or 60 min with 40 nM E2, washed with PBS, and incubated inmedium without steroid for the remaining 12 h. As a positive control forGnRH analog binding, cells were incubated in the presence of E2 (40 nM)for 12 h. Binding is representative as a percent of the control (100%).Values are the mean 6 SEM for three replicates. *P , 0.05.

FIG. 3. Effects of an ER antagonist on E2 stimulated an increase in GnRHreceptors. Pituitary cells were cultured with E2 (0.1 nM) or E2-BSA (10nM) for 12 h in the presence or absence of ICI 182,780 (10 lM). Valuesare the mean 6 SEM for four replicates. *P , 0.001.

INCREASE IN NUMBER OF GnRHR VIA MEMBRANE ER 773

Dow

nloaded from w

ww

.biolreprod.org.

E2, antibody directed against the ligand binding domain ofESR1 (clone H222), or ICI 182,780 (Fig. 6).

DISCUSSION

The historical paradigm of E2 action posits that the hormonediffuses across both the plasma and nuclear membrane where itinteracts with high-affinity ER. While there is little doubt thatthis model appropriately describes E2 regulation of many of itstarget genes, it has become very apparent that this hormone iscapable of eliciting changes in gene expression that areinitiated by interacting with nonnuclear receptors located at theplasma membrane [38] or even intracellular membranecompartments such as the smooth endoplasmic reticulum[39]. This is especially evident in ER knockout and knock-inmice that contain a mutated ER that cannot bind to ERE(estrogen response element); therefore, E2-stimulated geneexpression and subsequent protein translation is via anonclassical pathway, and it has been shown that ERE-independent E2 genomic effects can occur independently of ERdirectly binding to DNA [40]. This is likely the mechanism thatoccurs with the GNRHR gene promoter because an ERE hasnot been identified in the GNRHR gene promoter from anyspecies [41–45], and it is exquisitely clear that E2 has profoundeffects on the regulation of this gene. Herein we have exploredthe possibility that this effect of E2 may reflect a mechanismmediated via an ER on the plasma membrane through thenonclassical signaling pathway. Our interest in this possibilitywas stimulated by previous work examining the impact ofmembrane-impermeable E2 conjugates on LH secretion insheep [16]. In that study, E2, E2-PEP, and E2-BSA led to arapid suppression of LH secretion that persisted for approxi-mately 6 h. The rapidity of this response (within 30 min) is notconsistent with an exclusively genomic mechanism. Followingthe transient suppression of LH secretion, a preovulatory-likesurge of LH was evident at approximately 12 h post-E2infusion. In contrast, the E2-PEP and E2-BSA conjugates didnot induce a preovulatory-like surge of LH; however, the meanLH measured during the surge window was significantlygreater than that measured during the preinfusion period,suggesting that sensitivity of the pituitary gland may have

increased at that time even through there was no increase inGnRH secretion [46].

The E2-induced preovulatory LH surge is due to at least twocomponents: first, an increase in GnRHR numbers in thepituitary [1] and second, a large and temporally delayed,prolonged release of GnRH from the hypothalamus [3]. Theabsence of a full LH surge in animals treated with the BSA-conjugated E2 [16] likely reflects the loss of the secondcomponent as a result of either an inability of the E2 conjugatesto cross the blood-brain barrier or the lack of E2 membranesignaling to GnRH neurons [46]. In contrast, if E2 activation ofmembrane-signaling events are involved in the synthesis ofGnRHR, it would explain the discrete increase in LH secretionobserved during the expected preovulatory-like surge of LH insheep treated with membrane-impermeable E2-BSA [16].Consistent with this notion, we find that E2-BSA stimulatesan increase in GnRH analog binding in primary cultures ofovine pituitary cells in a time course similar to that measuredfor free E2. Also, because gonadotrophs are the only ovinepituitary cell type thought to express E2 receptors [47],detection of displaceable membrane binding of a fluorescentE2 conjugate supports the hypothesis that membrane-associat-ed binding sites for E2 are present on the same cells thatexpress GnRHR in the ovine anterior pituitary.

If a membrane site of action underlies the ability of E2 toenhance the number of GnRHR, then at issue is the identity ofthe ER. Since the first characterization of rapid, nongenomicsignaling induced by E2, much effort has been devoted toisolating and characterizing the protein that mediates theseeffects. It is clear that this issue is complex and likely involvesboth the classical ER (ESR1 and ESR2) as well as novelproteins such as GPR30 that are not members of thesuperfamily of nuclear receptors. Certainly much evidencehas been generated that supports the concept that ER aretransported to and localized in the plasma membrane, which isdependent on posttranslational modification and associationwith other proteins [48–53]. Thus, the ability of E2 to signalthrough a membrane site of action does not require invokingthe presence of novel ER isoforms. Consistent with this, wefind that the ability of both E2 and E2-BSA to elicit increased

FIG. 4. Expression of a dominant-negative ESR1 mutant inhibits E2- andESR1 agonist (PPT)-mediated increase in GnRH receptors. Pituitary cellswere infected for 24 h with adenovirus encoding a dominant-negativemutant form of ESR1 (L540Q). At 24 h postinfection, cells were treatedwith 4 nM E2 or 0.82 nM PPT for 12 h. Relative numbers of GnRHreceptors are indicated as GnRH analog binding as a percent of thecontrol (100%). Values are the mean 6 SEM for three replicates. *P ,0.05.

FIG. 5. Expression of a dominant-negative CREB blocks E2-mediatedincrease in GnRH receptors. Pituitary cells were infected for 24 h withadenovirus encoding green fluorescent protein (GFP, control for cytotoxiceffects) or dominant-negative CREB. At 24 h postinfection, cells weretreated with 0.1 nM E2 for 12 h. Cells that were not infected and treatedwith vehicle for 12 h served as a binding control (100%). Values are themean 6 SEM for five replicates. *P , 0.05.

774 DAVIS ET AL.

Dow

nloaded from w

ww

.biolreprod.org.

GnRH analog binding is completely blocked by inclusion ofthe ER antagonist ICI 182,780. Because this compound has notbeen shown to block E2 signaling through GPR30, it seemslikely that the operative receptor is either ESR1 or ESR2.Several additional lines of evidence seem to best support ESR1rather than ESR2 as the key receptor. First, the selective ESR1agonist PPT mimics the effects of E2 on GnRH analog binding.Second, overexpression of a dominant-negative ESR1 (L540Q)effectively blocks both E2 and PPT increase in GnRH analogbinding. Third, we have previously shown that responsivenessof the gonadotrophs to E2 is primarily mediated via ESR1when cells or animals are treated with PPT [54].

The first evidence for nongenomic signaling by steroidhormones are based on the ability of these molecules to elicitbiochemical and cellular events that are simply too rapid toreflect a requirement for biosynthesis of new mRNA andprotein. The ability of E2, E2-PEP, and E2-BSA to rapidlysuppress GnRH-induced LH secretion [16, 17] clearly fits withthis line of evidence. In contrast, the most distal mechanismunderlying the effects of E2 on GnRHR expression is, almostcertainly, transcriptional. In support of this notion, it is wellestablished that E2 elicits coordinate changes in GnRHRmRNA and GnRHR protein [9, 14, 55–59]. While these datado not eliminate the possibility of a posttranscriptionalmechanism such as mRNA stabilization, the ability ofactinomycin D to block E2 up-regulation of GnRHR numberin ovine pituitary cells does not support this theory [56]. Thus,the most likely mechanism is that E2 activates GNRHR geneexpression. Consistent with this notion, we have established E2responsiveness of the ovine GNRHR gene promoter in multiplelines of transgenic mice [60]. Furthermore, as has beendemonstrated in sheep, this effect of E2 is evident in theabsence of GnRH input [9, 56, 61]. Thus, E2 is exerting adirect pituitary effect to increase GNRHR promoter activity in

both sheep and in the transgenic mouse model. Based on thesedata, one might logically predict the presence of high-affinitybinding sites for ER within the ovine GNRHR gene promoter.This, however, is not the case. A canonical ERE is not presentin the ovine GNRHR gene [43], and E2 responsiveness is notdetectable using standard in vitro analyses of promoterregulation [60]. Thus, it appears that E2 leads to an increasein GNRHR gene expression through a noncanonical pathway.At issue then is the identity of this pathway.

It is interesting to note that the presence of a cAMP responseelement (CRE) is one of the most conserved structural featuresin the proximal promoter of the GNRHR gene from multiplespecies, including sheep and humans. In the ovine promoter,the CRE is capable of binding CREB and mediating enhancedtranscriptional activity in response to cAMP and forskolin [62].Although the ability of E2 to module transcriptional activity oftarget genes through cAMP, protein kinase A, and CREBpathways has been well described in neurons [63–69], recentlya 30-min exposure to E2 was shown to increase levels ofphosphorylated CREB in gonadotrophs in the ovine pituitarygland [37, 47]. Based on these date, we were intrigued with thepossibility that E2 may influence transcriptional activity of theGnRHR gene and subsequent protein synthesis through aCREB/ATF-dependent mechanism. Consistent with this pos-sibility, adenoviral-mediated overexpression of a dominant-negative form of CREB prevented E2-induced increase ofGnRH analog binding in sheep pituitary cells. Importantly, thisresult was not due to a nonspecific effect of infection becausean equivalent dose of adenovirus expressing GFP did notattenuate the increase in GnRH binding in response to E2.These data are the first to clearly implicate a candidateintracellular pathway that underlies E2 regulation of theGNRHR gene and subsequent protein synthesis. We stressthe term candidate in this context because of the finding that an

FIG. 6. Fluorescently labeled E2-BSA (E2-BSA-FITC) staining associated with theplasma membrane of ovine pituitary cells.Pituitary cells were incubated with E2-BSA-FITC alone (A) or in the presence of 1000-fold excess of E2 (B), an antibody specificfor the ligand-binding domain of the ER(H222 clone) (C), or an ER-specific antag-onist (ICI 182,780) (D). Confocal images ofE2-BSA-FITC are shown in the left panel.The plasma membrane was identified byconcanavalin A conjugated to Alexa 594(middle panel). Colocalization of E2-BSA-FITC and concanavalin A appears yellow inthe merged view (right panel).

INCREASE IN NUMBER OF GnRHR VIA MEMBRANE ER 775

Dow

nloaded from w

ww

.biolreprod.org.

A-CREB molecule in which the basic amino acids comprisingthe DNA-binding domain have been replaced with acidicamino acids nevertheless retains the capacity to dimerize but isincapable of DNA binding [24]. As such, A-CREB exerts itsdominant-negative effects by effectively sequestering wild-type partners in a nonproductive dimer. What is important hereis that CREB does not simply homodimerize but is capable ofheterodimerization with other CREB/ATF family members aswell as members of the AP-1 family of transcription factors[70]. Thus, at present we cannot conclude that CREB is theoperative player in mediating E2 input to the GNRHR gene.Ultimately, this issue will likely have to be resolved by acombination of DNA-protein binding and chromatin immuno-precipitation assays. Unfortunately, given the diversity ofendocrine cells types in the anterior pituitary gland, this willnot be a trivial undertaking. Finally, as mentioned above, theovine GNRHR gene promoter contains a functional CRE.Certainly we speculate that this site may be important inmediating E2 input; however, because transgenic micerepresent the only reliable model in which E2 responsivenessof the GNRHR gene promoter is recapitulated, it will be neitherquick nor easy to directly test this hypothesis.

Over the past decade. it has become increasingly clear thatE2 signaling involves both membrane- and nuclear-mediatedmechanisms. Based on the present data, the ability of E2 toregulate GNRHR gene expression may be added to theexpanding list of biological responses to E2 in which theinitial signal is propagated at the plasma membrane, probablythrough a nonclassical signaling mechanism. This does notseem to be an unreasonable hypothesis considering the numberof hormones that signal via G-protein-coupled receptors at theplasma membrane of cells to alter gene expression andsubsequent protein synthesis in cells. Estrogen receptorsassociated with the plasma membrane have been shown toactivate G-proteins [71–74], indicating that this steroidhormone receptor may be considered a nontraditional G-protein-coupled receptor. Furthermore, it has been shown thatGnRH itself acting through the GnRHR, a G-protein-coupledreceptor, regulates the GNRHR gene [75]. There is nowrelatively extensive evidence supporting a membrane-initiated,nonclassical signal underlying the ability of E2 to rapidly andtransiently suppress GnRH-induced LH secretion. As such, it istempting to speculate that the E2 signals pathways that lead toacute suppression of LH secretion, and, subsequently,increased GNRHR gene expression may result from a commonmembrane site of action. These two actions of E2 may beinvolved with 1) decreasing secretion of LH prior to theovulatory surge to permit a buildup of LH stores in the anteriorpituitary and 2) increasing sensitivity of the anterior pituitary toGnRH, both of which might contribute to maximizing secretionof LH during the ovulatory surge.

The use of E2 conjugated to BSA to study the effects of E2at the membrane has been quite controversial. The linkage ofthe large BSA molecule to the C-6 position of E2 has beenshown to have differential effects in cells relative to free E2[76, 77]; however, in the present study, it was shown that E2-BSA increased GnRHR numbers in a manner similar to freeE2. Other concerns have been raised that the bulkiness of BSAcould alter the ability of binding of E2-BSA to the ER and alterthe biological outcome [46, 76]. That a greater concentration ofE2-BSA was required to increase the numbers of GnRHR isnot surprising because binding of E2-BSA to the soluble E2receptor is only about 7% that of E2 (Arreguin-Arevalo andNett, unpublished results). Further, if the ER is tethered in theplasma membrane, it is likely to be much less able to interactwith E2-BSA than soluble receptor. Thus, the need for a greater

concentration of E2-BSA to induce the same increase inGnRHR as E2 would be expected. It should also be noted that10 nM E2-BSA was used in the present study, and this is thecommonly used concentration of free E2 for in vitro studies,which is much greater than circulating concentrations of E2 invivo. Based on previous experiments conducted by Arreguin-Arevalo and Nett (unpublished results), there is approximately20 pg/ml of free E2 remaining in the E2-BSA conjugatefollowing the extraction procedure. Based on this concentra-tion, approximately 3.0 fM of free E2 was added to cellsreceiving E2-BSA. This concentration is well below thatrequired for a physiological response based on our dose-response experiments. Two other major concerns of usingconjugated E2-BSA include release of free E2 or proteolyticdegradation of the BSA, which would allow smaller aminoacids attached to E2 to gain access to nuclear receptors [78].Our laboratory has conducted studies to refute these claims.Administration of E2-BSA to ewes causes a rapid nongenomiceffect as observed by a decrease in secretion of LH; however,the preovulatory surge release of LH that is considered to be agenomic effect is not observed [16, 46]. The results from thesestudies provide evidence that the diethyl ether extractionmethod removes free E2 from the conjugated form and that E2-BSA is stable in vivo for a number of hours and does notappear to be degraded by proteolysis because genomic effectswere not observed. Finally, a classical study of administrationof E2 and E2-BSA to rodents to measure the effect on uterineweight was conducted (Clay et al., unpublished results) inwhich 0.1 lg E2 or 250 lg E2-BSA was administered to theanimals. Only 0.1 lg E2 increased uterine weight, and therewas no difference in uterine weight between controls and E2-BSA treated animals, again providing further evidence that ourmethod of removing free E2 from the conjugated form iseffective and that E2-BSA is stable in vivo such that free E2 isnot available for classical genomic effects.

We should emphasize that while our data support a role formembrane-associated ESR1 in GNRHR gene regulation, we donot reject a role for the nuclear receptor. What is intriguing inthis regard is that the ESR1 molecule containing the L540Qmutation clearly blocks the E2-induced increase in GnRHanalog binding. However, there is no evidence we are aware ofthat this particular mutation in L540Q affects the ability of E2to signal through a membrane site of action. Rather, thepresence of this single amino acid conversion appears to alterthe ability of ER to recruit coactivators to an E2 responsivepromoter [22]. As such, enhanced transcription of the GNRHRgene to E2 may be initiated by a membrane form of ESR1 butrequires participation of nuclear ER for the ultimate transcrip-tional response. The latter almost certainly does not involvedirect binding of ER to the GNRHR gene but rather may reflectinteractions of ER with primary DNA-binding proteins (e.g.,AP-1 and CREB) and the formation of a ternary complex thatthen acts to recruit key coactivators. If correct, then E2signaling to the GNRHR gene may well reflect an importantphysiological model of the concept of parallel pathways asoriginally proposed by Vasudevan and Pfaff [35].

REFERENCES

1. Crowder ME, Nett TM. Pituitary content of gonadotropins and receptorsfor gonadotropin-releasing hormone (GnRH) and hypothalamic content ofGnRH during the periovulatory period of the ewe. Endocrinology 1984;114:234–239.

2. Schoenemann HM, Humphrey WD, Crowder ME, Nett TM, Reeves JJ.Pituitary luteinizing hormone-releasing hormone receptors in ovariecto-mized cows after challenge with ovarian steroids. Biol Reprod 1985;32:574–583.

3. Moenter SM, Caraty A, Locatelli A, Karsch FJ. Pattern of gonadotropin-

776 DAVIS ET AL.

Dow

nloaded from w

ww

.biolreprod.org.

releasing hormone (GnRH) secretion leading up to ovulation in the ewe:existence of a preovulatory GnRH surge. Endocrinology 1991; 129:1175–1182.

4. Kesner JS, Convey EM, Anderson CR. Evidence that estradiol induces thepreovulatory LH surge in cattle by increasing pituitary sensitivity toLHRH and then increasing LHRH release. Endocrinology 1981;108:1386–1391.

5. Adams TE, Normal RL, Spies HG. Gonadotropin-releasing hormonereceptor binding and pituitary responsiveness in estradiol-primed mon-keys. Science 1981; 213:1388–1390.

6. Naik SI, Young LS, Charlton HM, Clayton RN. Evidence for a pituitarysite of gonadal steroid stimulation of GnRH receptors in female mice. JReprod Fertil 1985; 74:615–624.

7. Menon M, Peegel H, Katta V. Estradiol potentiation of gonadotropin-releasing hormone responsiveness in the anterior pituitary is mediated byan increase in gonadotropin-releasing hormone receptors. Am J ObstetGynecol 1985; 151:534–540.

8. Yasin M, Dalkin AC, Haisenleder DJ, Kerrigan JR, Marshall JC.Gonadotropin-releasing hormone (GnRH) pulse pattern regulates GnRHreceptor gene expression: augmentation by estradiol. Endocrinology 1995;136:1559–1564.

9. Gregg DW, Nett TM. Direct effects of estradiol-17 beta on the number ofgonadotropin-releasing hormone receptors in the ovine pituitary. BiolReprod 1989; 40:288–293.

10. Drouin J, Lagace L, Labrie F. Estradiol-induced increase of the LHresponsive to LH releasing hormone (LHRH) in rat anterior pituitary cellsin culture. Endocrinology 1976; 99:1477–1481.

11. Drouin J, Labrie F. Interactions between 17 beta-estradiol and progester-one in the control of luteinizing hormone and follicle-stimulating hormonerelease in rat anterior pituitary cells in culture. Endocrinology 1981;108:52–57.

12. Turzillo AM, Campion CE, Clay CM, Nett TM. Regulation ofgonadotropin-releasing hormone (GnRH) receptor messenger ribonucleicacid and GnRH receptors during the early preovulatory period in the ewe.Endocrinology 1994; 135:1353–1358.

13. Brooks J, Taylor PL, Saunders PT, Eidne KA, Struthers WJ, McNeilly AS.Cloning and sequencing of the sheep pituitary gonadotropin-releasinghormone receptor and changes in expression of its mRNA during theestrous cycle. Mol Cell Endocrinol 1993; 94:R23–R27.

14. Hamernik DL, Clay CM, Turzillo A, Van Kirk EA, Moss GE. Estradiolincreases amounts of messenger ribonucleic acid for gonadotropin-releasing hormone receptors in sheep. Biol Reprod 1995; 53:179–185.

15. Wise ME, Nieman D, Stewart J, Nett TM. Effect of number of receptorsfor gonadotropin-releasing hormone on the release of luteinizing hormone.Biol Reprod 1984; 31:1007–1013.

16. Arreguin-Arevalo JA, Nett TM. A nongenomic action of estradiol as themechanism underlying the acute suppression of secretion of luteinizinghormone in ovariectomized ewes. Biol Reprod 2006; 74:202–208.

17. Arrequin-Arevalo JA, Nett TM. A nongenomic action of 17b-estradiol asthe mechanism underlying the acute suppression of secretion of luteinizinghormone. Biol Reprod 2005; 73:115–122.

18. Niswender GD, Roess DA, Sawyer HR, Silvia WJ, Barisas BG.Differences in the lateral mobility of receptors for luteinizing hormone(LH) in the luteal cell plasma membrane when occupied by ovine LHversus human chorionic gonadotropin. Endocrinology 1985; 116:164–169.

19. Adams TE, Wagner TO, Sawyer HR, Nett TM. GnRH interaction withanterior pituitary. II. Cyclic AMP as an intracellular mediator in the GnRHactivated gonadotroph. Biol Reprod 1979; 21:735–747.

20. Nett TM, Crowder ME, Moss GE, Duello TM. GnRH-receptor interaction.V. Down-regulation of pituitary receptors for GnRH in ovariectomizedewes by infusion of homologous hormone. Biol Reprod 1981; 24:1145–1155.

21. Ince BA, Zhuange Y, Wrenn CK, Shapiro DJ, Katzenellenbogen BS.Powerful dominant negative mutants of the human estrogen receptor. JBiol Chem 1993; 268:14026–14032.

22. Acevedo ML, Lee KC, Stender JD, Katzenellenbogen BS, Kraus WL.Selective recognition of distinct classes of coactivators by a ligand-inducible activation domain. Mol Cell 2004; 13:725–738.

23. He TC, Zhou S, Da Costa LT, Yu J, Kinzler KW, Vogelstein B. Asimplified system for generating recombinant adenoviruses. Proc NatlAcad Sci U S A 1998; 95:2509–2514.

24. Ahn S, Olive M, Aggarwal S, Krylov D, Ginty DD, Vinson C. Adominant-negative inhibitor of CREB reveals that it is a general mediatorof stimulus-dependent transcription of c-fos. Mol Cell Biol 1998; 18:967–977.

25. Glass JD, Amann RP, Nett TM. Effects of season and sex on the

distribution of cytosolic estrogen receptors within the brain and theanterior pituitary gland of sheep. Biol Reprod 1984; 30:894–902.

26. Thomas P, Pang Y, Filardo EJ, Dong J. Identity of an estrogen membranereceptor coupled to a G protein in human breast cancer cells.Endocrinology 2005; 145:624–632.

27. Revankar CM, Cimino DF, Sklar LA, Arterburn JB, Prossnitz ER. Atransmembrane intracelular estrogen receptor mediates rapid cell signaling.Science 2005; 307:1625–1630.

28. Lindzey J, Jayes FL, Yates MM, Couse JF, Korach KS. The bi-modaleffects of estradiol on gonadotropin synthesis and secretion in female miceare dependent on estrogen receptor-a. J Endocrinol 2006; 191:309–317.

29. Couse JF, Yates MM, Walker VR, Korach KS. Characterization of thehypothalamic-pituitary-gonadal axis in estrogen receptor (ER) null micereveals hypergonadism and endocrine sex reversal in females lacking ERabut not ERb. Mol Endocrinol 2003; 17:1039–1053.

30. Couse JF, Korach KS. Estrogen receptor null mice: what have we learnedand where will they lead us? Endocr Rev 1999; 20:358–417.

31. Wersinger SR, Haisenleder DJ, Lubahn DB, Rissman EF. Steroidfeedback on gonadotropin release and pituitary gonadotropin subunitmRNA in mice lacking a functional estrogen receptor alpha. Endocrine1999; 11:137–143.

32. Lindzey J, Wetsel WC, Couse JF, Stoker T, Cooper R, Korach KS. Effectsof castration and chronic steroid treatments on hypothalamic gonadotro-pin-releasing hormone content and pituitary gonadotropins in male wild-type and estrogen receptor-a knockout mice. Endocrinology 1998;139:4092–4101.

33. Jakacka M, Ito M, Weiss J, Chien PY, Gehm BD, Jameson JL. Estrogenreceptor binding to DNA is not required for its activity through thenonclassical AP1 pathway. J Biol Chem 2001; 276:13615–13621.

34. Jakacka M, Ito M, Martinson F, Ishikawa T, Lee EJ, Jameson JL. Anestrogen receptor (ER)alpha deoxyribonucleic acid-binding domainknock-in mutation provides evidence for nonclassical ER pathwaysignaling in vivo. Mol Endocrinol 2002; 16:2188–2201.

35. Vasudevan N, Pfaff DW. Membrane-initiated actions of estrogens inneuroendocrinology: emerging principles. Endocr Rev 2007; 28:1–19.

36. Kushner PJ, Agard DA, Greene GL, Scanlan TS, Shiau AK, Uht RM,Webb P. Estrogen receptor pathways to AP-1. J Steroid Biochem Mol Biol2000; 74:311–317.

37. Iqbal J, Latchoumanin O, Clarke IJ. Rapid in vivo effects of estradiol-17bin ovine pituitary gonadotropes are displayed by phosphorylation ofextracellularly regulated kinase, serine/threonine kinase, and 30, 50-cyclicadenosine 50-monophosphate-responsive element-binding protein. Endo-crinology 2007; 148:5794–5802.

38. Bjornstrom L, Sjoberg M. Mechanisms of estrogen receptor signaling:convergence of genomic and nongenomic actions on target genes. MolEndocrinol 2005; 19:833–842.

39. Revankar CM, Cimino DF, Sklar LA, Arterburn JB, Prossnitz ER. Atransmembrane intracellular estrogen receptor mediates rapid cellsignaling. Science 2005; 307:1625–1630.

40. McDevitt MA, Glidewell-Kenney C, Jimenea MA, Ahearn PC, Weiss J,Lameson JL, Levine JE. New insights into the classical and non-classicalactions of estrogen: evidence from estrogen receptor knock-out and knock-in mice. Mol Cell Endocrinol 2008; 290:24–30.

41. Kakar SS, Musgrove LC, Devor DC, Sellers JC, Neill JD. Cloning,sequencing, and expression of human gonadotropin releasing hormone(GnRH) receptor. Biochem Biophys Res Commun 1992; 189:289–295.

42. Albarracin CT, Kaiser UB, Chin WW. Isolation and characterization of the50-flanking region of the mouse gonadotropin-releasing hormone receptorgene. Endocrinology 1994; 135:2300–2306.

43. Campion CE, Turzillo AM, Clay CM. The gene encoding the ovinegonadotropin- releasing hormone (GnRH) receptor: cloning and initialcharacterization. Gene 1996; 170:277–280.

44. Fan NC, Peng C, Krisinger J, Leung PC. The human gonadotropin-releasing hormone receptor gene: complete structure including multiplepromoters, transcription initiation site, and polyadenylation signals. MolCell Endocrinol 1995; 107:R1–R8.

45. Reinhart J, Xiao S, Arora KK, Catt KJ. Structural organization andcharacterization of the promoter region of the rat gonadotropin-releasinghormone receptor gene. Mol Cell Endocrinol 1997; 130:1–12.

46. Arreguin-Arevalo A, Ashley RL, Wagenmaker ER, Oakley AE, KarschFJ, Nett TM. Membrane-initiated actions of estradiol (E2) in the regulationof LH secretion in ovariectomized (OVX) ewes. Reprod Biol Endocrinol2010; 8:40.

47. Clarke IJ, Tobin VA, Pompolo S, Pereira A. Effects of changinggonadotropin-releasing hormone pulse frequency and estrogen treatmenton levels of estradiol receptor-a and induction of Fos and phosphorylatedcyclic adenosine monophosphate response element binding protein in

INCREASE IN NUMBER OF GnRHR VIA MEMBRANE ER 777

Dow

nloaded from w

ww

.biolreprod.org.

pituitary gonadotropes: studies in hypothalamo-pituitary disconnectedewes. Endocrinology 2005; 146:1128–1137.

48. Li L, Haynes MP, Bender JR. Plasma membrane localization and functionof the estrogen receptor a variant (ER46) in human endothelial cells. ProcNatl Acad Sci U S A 2003; 100:4807–4812.

49. Razandi M, Oh P, Pedram A, Schnitzer J, Levin ER. ERs associate withand regulate the production of caveolin: implications for signaling andcellular actions. Mol Endocrinol 2002; 16:100–115.

50. Razandi M, Alton G, Pedram A, Ghonshani S, Webb P, Levin ER.Identification of a structural determinant necessary for the localization andfunction of estrogen receptor a at the plasma membrane. Mol Cell Biol2003; 23:1633–1646.

51. Acconcia F, Ascenzi P, Fabozzi G, Visca P, Marino M. S-Palmitoylationmodulates human estrogen receptor-a functions. Biochem Biophys ResCommun 2004; 316:878–883.

52. Acconcia F, Ascenzi P, Bocedi A, Spisni E, Tomasi V, Trentalance A,Visca P, Marino M. Palmitoylation-dependent estrogen receptor (a)membrane localization regulation by 17b-estradiol. Mol Biol Cell 2005;16:231–237.

53. Pedram A, Razandi M, Sainson RC, Kim JK, Hughes CC, Levin ER. Aconserved mechanism for steroid receptor translocation to the plasmamembrane. J Biol Chem 2007; 282:22278–22288.

54. Arrequin-Arevalo JA, Davis TL, Nett TM. Differential modulation ofgonadotropin secretion by selective estrogen receptor 1 and estrogenreceptor 2 agonists in ovariectomized ewes. Biol Reprod 2007; 77:320–328.

55. Ghosh BR, Wu JC, Strahl BD, Childs GV, Miller WL. Inhibin andestradiol alter gonadotropes differentially in ovine pituitary cultures:changing gonadotrope numbers and calcium responses to gonadotropin-releasing hormone. Endocrinology 1996; 137:5144–5154.

56. Turzillo AM, DiGregorio GB, Nett TM. Messenger ribonucleic acid forgonadotropin-releasing hormone receptor and numbers of gonadotropin-releasing hormone receptors in ovariectomized ewes after hypothalamic-pituitary disconnection and treatment with estradiol. J Anim Sci 1995;73:1784–1788.

57. Gregg DW, Allen MC, Nett TM. Estradiol-induced increase in number ofgonadotropin-releasing hormone receptors in cultured ovine pituitary cells.Biol Reprod 1990; 43:1032–1036.

58. Laws SC, Webster JC, Miller WL. Estradiol alters the effectiveness ofgonadotropin-releasing hormone (GnRH) in ovine pituitary cultures:GnRH receptors versus responsiveness to GnRH. Endocrinology 1990;127:381–386.

59. Sealfon SC, Laws SC, Wu JC, Gillo B, Miller WL. Hormonal regulationof gonadotropin-releasing hormone receptors and messenger RNA activityin ovine pituitary culture. Mol Endocrinol 1990; 4:1980–1987.

60. Duval DL, Farris AR, Quirk CC, Nett TM, Hamernik DL, Clay CM.Responsiveness of the ovine gonadotropin-releasing hormone receptor geneto estradiol and gonadotropin-releasing hormone is not detectable in vitrobut is revealed in transgenic mice. Endocrinology 2000; 141:1001–1010.

61. Clarke IJ, Cummins JT, Crowder ME, Nett TM. Pituitary receptors forgonadotropin-releasing hormone in relation to changes in pituitary andplasma gonadotropins in ovariectomized hypothalamo-pituitary-discon-nected ewes. II. A marked rise in receptor number during the acutefeedback effects of estradiol. Biol Reprod 1988; 39:349–354.

62. Holtzen ML, Quirk CC, Duval DL, Clay CM. Regulation of the ovineGnRH receptor (oGnRHR) gene by cAMP and steroidogenic factor-1 (SF-

1). In: Program for the 32nd Annual Meeting of the Society for the Studyof Reproduction, July 31–August 3, 1999; 60 Special Issue: Abstract 586.

63. Vasudevan N, Pfaff DW. Non-genomic actions of estrogens and theirinteraction with genomic actions in the brain. Front Neuroendocrinol2008; 29:238–257.

64. Szego EM, Barabas K, Balog J, Szilagyi N, Korach KS, Juhasz G,Abraham IM. Estrogen induces estrogen receptor alpha-dependent cAMPresponse element-binding protein phosphorylation via mitogen activatedprotein kinase pathway in basal forebrain cholinergic neurons in vivo. JNeurosci 2006; 26:4104–4110.

65. Boulware MI, Weick JP, Becklund BR, Kuo SP, Groth RD, MermelsteinPG. Estradiol activates group I and II metabotropic glutamate receptorsignaling, leading to opposing influences on cAMP response element-binding protein. J Neurosci 2005; 25:5066–5078.

66. Lee SJ, Campomanes CR, Sikat PT, Greenfield AT, Allen PB, McEwenBS. Estrogen induces phosphorylation of cyclic AMP response elementbinding (pCREB) in primary hippocampal cells in a time-dependentmanner. Neuroscience 2004; 124:549–560.

67. Wade CB, Dorsa DM. Estrogen activation of cyclic adenosine 50-monophosphate response element-mediated transcription requires theextracellularly regulated kinase/mitogen-activated protein kinase pathway.Endocrinology 2003; 144:832–838.

68. Zhou Y, Watters JJ, Dorsa DM. Estrogen rapidly induces thephosphorylation of the cAMP response element binding protein in ratbrain. Endocrinology 1996; 137:163–166.

69. Abraham IM, Han SK, Todman MG, Korach KS, Herbison AE. Estrogenreceptor beta mediates rapid estrogen actions on gonadotropin-releasinghormone neurons in vivo. J Neurosci 2003; 23:5771–5777.

70. Vinson C, Myakishev M, Acharva A, Mir AA, Moll JR, Bonovich M.Classification of human B-ZIP proteins based on dimerization properties.Mol Cell Biol 2002; 22:6321–6335.

71. Razandi M, Pedram A, Greene GL, Levin ER. Cell membrane and nuclearestrogen receptors (ERs) originate from a single transcript: studies ofERalpha and ERbeta expressed in Chinese hamster ovary cells. MolEndocrinol 1999; 13:307–319.

72. Razandi M, Pedram A, Merchenthaler I, Greene GL, Levin ER. Plasmamembrane estrogen receptors exist and function as dimers. MolEndocrinol 2004; 18:2854–2865.

73. Hammes SR, Levin ER. Extranuclear steroid receptors: nature and actions.Endocr Rev 2007; 28:726–741.

74. Levin ER, Pietras RJ. Estrogen receptors outside the nucleus in breastcancer. Breast Cancer Res Treat 2008; 108:351–361.

75. Rispoli LA, Nett TM. Pituitary gonadotropin-releasing hormone (GnRH)receptor: structure, distribution and regulation of expression. Anim ReprodSci 2005; 88:54–74.

76. Stevis PE, Deecher DC, Suhadolnik L, Mallis LM, Frail DE. Differentialeffects of estradiol and estradiol-BSA conjugates. Endocrinology 1999;140:5455–5458.

77. Temple JL, Wray S. Bovine serum albumin-estrogen compoundsdifferentially alter gonadotropin-releasing hormone-1 neuronal activity.Endocrinology 2005; 146:558–563.

78. Harrington WR, Kim SH, Funk CC, Madak-Erdogan Z, Schiff R,Katzenellenbogen JA, Katzenellenbogen BS. Estrogen dendrimer conju-gates that preferentially activate extranuclear, nongenomic versus genomicpathways of estrogen action. Mol Endocrinol 2006; 20:491–502.

778 DAVIS ET AL.

Dow

nloaded from w

ww

.biolreprod.org.