Desensitization and Internalization of Human and Xenopus ...

0013-7227/00/$03.00/0 Endocrinology Copyright © 2000 by The Endocrine Society

Vol. 141, No. 12 Printed in U.S.A.

Desensitization and Internalization of Human and Xenopus Gonadotropin-Releasing Hormone Receptors Expressed in ␣T4 Pituitary Cells Using Recombinant Adenovirus* JAMES N. HISLOP, MICHAEL T. MADZIVA, HELEN M. EVEREST, TOM HARDING, JAMES B. UNEY, GARY B. WILLARS, ROBERT P. MILLAR, BRIGITTE E. TROSKIE, JAMES S. DAVIDSON, AND CRAIG A. MCARDLE University Research Centre for Neuroendocrinology (J.N.H., M.T.M., H.M.E., T.H., J.B.U., C.A.M.), University of Bristol, Bristol, United Kingdom; Department of Cell Physiology and Pharmacology (G.B.W.), University of Leicester, Leicester LE1 9HN, United Kingdom; Medical Research Council Human Reproductive Science Unit (R.P.M.), Centre for Reproductive Biology, Edinburgh EH3 9ET, Scotland, United Kingdom; and Medical Research Council Research Unit for Molecular Reproductive Endocrinology (B.E.T., R.P.M., J.S.D.), Department of Medical Biochemistry, University of Cape Town, Observatory 7925, South Africa ABSTRACT Nonmammalian vertebrates express at least two forms of GnRH and distinct forms of GnRH receptor (GnRH-R) have coevolved with their ligands. Mammalian and nonmammalian GnRH-R have key structural differences (notably the lack of C-terminal tails in mammalian GnRH-R) and comparative studies are beginning to reveal their functional relevance. However, cellular context and receptor density influence G protein-coupled receptor function and may be important variables in such work using heterologous expression systems. Here we report a comparative study using ␣T4 cells (gonadotrope progenitors that lack endogenous GnRH-R) transfected with a mammalian (human) or nonmammalian (Xenopus laevis type I) GnRH-R. Because conventional transfection strategies proved inefficient, recombinant adenovirus expressing these receptors were constructed, enabling controlled and efficient GnRH-R expression. When

expressed in ␣T4 cells at physiological density, these GnRH-Rs retain the pharmacology of their endogenous counterparts (as judged by ligand specificity in radioligand binding and inositol phosphate accumulation assays) but do not activate adenylyl cyclase and are not constitutively active. Moreover, the Xenopus GnRH-R rapidly desensitizes and internalizes in these cells, whereas the human GnRH-R does not, and the internalization rates are not dependent upon receptor number. These data extend studies in COS, HEK, and GH3 cells showing that other GnRH-R with C-terminal tails desensitize and internalize rapidly, whereas tail-less mammalian GnRH-R do not. Retention of these distinctions at physiological receptor density in gonadotrope lineage cells, supports the argument that the evolution of nondesensitizing mammalian GnRH-Rs is functionally relevant and related to the development of mammalian reproductive strategies. (Endocrinology 141: 4564 – 4575, 2000)

T

to facilitate ␤-arrestin binding are most often located within the C-terminal tails of GPCRs, this region plays a crucial role in desensitization and internalization of many receptors (2, 5). GnRH is a hypothalamic decapeptide that acts via GPCRs on gonadotropes to stimulate the secretion of LH and FSH. Activation of GnRH-R causes a Gq/11 mediated stimulation of phospholipase C (PLC), which hydrolyzes membrane phosphoinositides yielding inositol phosphates (IPs), including Ins (1,4,5)P3, which mobilizes Ca2⫹ from intracellular stores (6 –9). GnRH also increases Ca2⫹ entry into gonadotropes, predominantly via voltage-operated Ca2⫹ channels, and the increase in cytosolic Ca2⫹ caused by GnRH is primarily responsible for the increase in exocytotic hormone release (6 –10). Sustained stimulation of gonadotropes with GnRH and reduces GnRH-stimulated gonadotropin secretion, and this homologous desensitization underlies the suppression of the reproductive system, which is exploited in the major clinical applications of GnRH analogues (6, 11). Most vertebrates investigated express at least two forms of GnRH. Typically, the highly conserved chicken GnRH-II (cGnRH-II) is found with one or more additional form of the

HE RHODOPSIN family of receptors act via heterotrimeric G proteins to regulate effector proteins including phospholipase C (PLC) and adenylyl cyclase. Activation of such receptors is typically followed by their desensitization and internalization, and these processes involve rapid agonist-induced receptor phosphorylation. This phosphorylation, which may be mediated by specific G protein receptor kinases (GRKs), permits association with ␤-arrestin, which impairs G protein binding and activation (1, 2). ␤-arrestin can also act as an adapter, targeting desensitized GPCRs for internalization via clathrin-coated vesicles (CCVs) from which the receptors are either resensitized and recycled to the cell surface or are targeted to lysosomes for proteolytic degradation (3, 4). Because the amino acids phosphorylated Received May 17, 2000. Address all correspondence and requests for reprints to: Dr. Craig A. McArdle, University of Bristol, Department of Medicine, Bristol Royal Infirmary, Marlborough Street, Bristol BS2 8HW, United Kingdom. Email: [email protected]. * The work was supported in part by the Wellcome Trust (054949), the Medical Research Council (G78/6046), the South African Research Council, and National Research Foundation.

4564

GnRH-R DESENSITIZATION AND INTERNALIZATION IN ␣T4 CELLS

peptide, and these distinct forms may play different physiological roles. Amphibians, like mammals, express mammalian GnRH, which is thought to regulate reproductive function, and cGnRH-II, which may function as a neuromodulator. The multiple forms of GnRH have apparently evolved in parallel with distinct forms of the GnRH-R, which have different pharmacological characteristics. Thus the cloned GnRH-R of nonmammalian vertebrates (catfish, goldfish, chicken, and Xenopus laevis) have high selectivity for cGnRH-II over GnRH, whereas the opposite is true for the cloned mammalian GnRH-R (9, 12). Mammalian and nonmammalian GnRH show considerable similarity in a number of key areas (12), including conserved disulphide bridges (between extracellular loops I and II) as well as conservation of residues involved in G protein coupling and of residues in the TM domains implicated in ligand binding (but not specificity). There are also, however, clear structural differences between mammalian and nonmammalian GnRH-R (12). For example, the interacting asparagine and aspartate acid of TM II and TM VII (Asn87 and Asp318) in hGnRH-R), which are necessary for mammalian GnRH-R function are not conserved in nonmammalian GnRH-R (where aspartic acid is found in both positions) and two cysteins in the amino terminus and extracellular loop II (Cys14 and Cys184) in the hGnRH-R), which are thought to form a disulphide bridge, are not conserved in nonmammalian GnRH-R. However, the most striking structural difference is that all cloned mammalian GnRH-Rs lack C-terminal tails, whereas all cloned nonmammalian GnRH-Rs possess C-terminal tails with multiple potential sites for phosphorylation (9, 12–17). The mouse GnRH-R of ␣T3–1 cells (a gonadotrope-derived cell line) does not show rapid homologous desensitization, and this has been attributed to the lack of necessary C-terminal tail and its phosphorylation sites (18 –25). In accord with this, it has been shown that tailed GnRH-Rs (e.g. catfish and chicken) desensitize and/or internalize more rapidly than nontailed GnRH-Rs (e.g. rat and human) when expressed in COS or HEK cells (15, 23, 25). Although it was originally thought that the fundamental pharmacological characteristics (e.g. ligand specificity, effector coupling, and desensitization) would be constant for any given GPCR, it is now clear that these features can vary according to cell type and receptor density. Indeed, cell-tocell differences in the stoichiometry of receptors to effectors and accessory proteins can cause receptor signaling and regulation to vary dramatically from cell to cell. For example, in addition to activating PLC, 5-HT2C receptors can stimulate adenylyl cyclase at low density, but inhibit it at high density (26). Moreover, rapid desensitization of thromboxane A2 receptors was pronounced at low receptor density and absent at high density (27) and desensitization of TRH receptors was more pronounced in two pituitary cell lines than in COS, Hela, or HEK cells (28). GnRH-R signaling is also dependent upon receptor density and cellular context. It appears, for example, that GnRH-R expressed in Sf9 insect cells, COS cells, and GH3 cells can activate Gs (29 –33), whereas there is no compelling evidence for such activation in gonadotropes or in the gonadotrope-derived ␣T3–1 cell line (34 –36). In addition, reducing GnRH-R number in GH3 cells has been found to increase GnRH-stimulated cAMP accumulation

4565

while reducing GnRH-stimulated IP accumulation (37). Moreover, catfish GnRH-R have been shown to desensitize more rapidly in COS cells than in HEK cells (23). Because pituitary GnRH-R number can vary dramatically (e.g. through the oestrous cycle) and GnRH-R are also expressed in several extra-pituitary sites (e.g. prostate, testes, ovary, and breast), such observations raise the intriguing possibility that the pharmacology and/or signaling of GnRH-R varies between different cell types and/or under different physiological conditions. They also raise the question of whether differences in receptor density, signaling, or cellular context have contributed to the reported functional differences between mammalian and nonmammalian GnRH-R. To address these issues, we have developed recombinant adenovirus (Ad) expressing mammalian and nonmammalian GnRH-R and have used these for a functional comparison of the human and type I Xenopus laevis GnRH-Rs (hGnRH-R and XGnRH-R) expressed in ␣T4 cells, a gonadotrope/thyrotrope progenitor derived cell line which expresses the gonadotropin ␣-subunit, but not GnRH or TRH receptors (38). Materials and Methods Materials and cell culture GnRH and chicken GnRH-II were purchased from Peninsula Laboratories, Inc. Europe Ltd. (Merseyside, UK) or from Sigma (Poole, UK). Buserelin and [125I]Buserelin (2000 Ci/mmol) were provided by Prof. Sandow (Aventis Pharma GmbH, Frankfurt, Germany). [125I]cGnRH-II (approximately 3400 Ci/mmol, determined by self-displacement) was prepared using chloramine-T and purified by G25 Sephadex column chromatography. Culture media, sera, and plasticware were from Life Technologies, Inc. (Paisley, UK) or Falcon (Becton Dickinson and Co., Oxford, UK). Lipofectin, Lipofectamine, and the Plus Reagent were from Life Technologies, Inc. and Fugene 6 was from Roche Molecular Biochemicals (Lewes, UK). Fura 2/AM was from Molecular Probes, Inc. (Eugene, OR). [2-3H]Inositol (14 –16 Ci/mmol) was from Amersham International PLC (Little Chalfont, UK). All other reagents were from standard commercial suppliers. ␣T3–1 and ␣T4 cells were cultured in serum-supplemented DMEM as described (35, 38, 39). For experiments they were harvested by trypsinization and then incubated for 1–3 days in flasks or culture plates as described in the figure legends. For Ca2⫹ imaging, cells were cultured in 12-well plates (2 ml/well) containing untreated round glass cover slips.

Generation of recombinant adenovirus Recombinant, E1 deleted Ad, were produced according to standard techniques (40 – 43). Ad encoding the enhanced green fluorescent protein (EGFP), were generated for assessment of transfection efficiency. The reporter gene was excised from the plasmid pEGFP-N1 (CLONTECH Laboratories, Inc., Basingstoke, UK) and cloned into the Ad transfer vector pXCXCMV under control of the human cytomegalovirus (CMV) 1E promoter enhancer fragment (663 bp of pcDNA 1 from Invitrogen, Nu Leek, The Netherlands) followed by a GH poly A tail. The recombinant virus (Ad-EGFP) was then generated by homologous recombination with pJM17 (Microbix Systems Inc., Toronto, Canada) in HEK-293 cells, grown to high titer and then purified by CsCl density gradient centrifugation. The transfer vector (pXCXCMV) and a control Ad construct which contains no insert in the E1 region (Ad0) was kindly provided by A. Byrnes (University of Oxford, Oxford, UK). Ad-EGFP titer (determined using a plaque assay) was 0.75 ⫻ 1010 plaque forming U/ml. To generate Ad encoding GnRH-R, DNA encoding human and type I Xenopus GnRH-Rs was excised from pcDNA1/Amp plasmids (12, 14) using XbaI and BamHI (human) or XhoI and BamHI (Xenopus). The inserts were purified and ligated into an identically digested pXCXCMV which, after transformation and growth in Escherichia coli was purified on a CsCl gradient. Homologous recombination was then achieved by CaPO4 transfection of HEK-293 cells with 6 ␮g of pXCXCMV-GnRH-R


4566

construct and 6 ␮g pJM17 per 60 mm culture dish of 70% confluent cells. The cells were overlaid in medium with 0.5% agarose and individual recombinant Ad plaques were then picked and amplified. Restriction analysis (HindIII digests) of the Ad genome was then used to confirm homologous recombination. PCR was also performed amplifying viral DNA with primers (Cruachem, Glasgow, UK) flanking the multiple cloning sites of pXCXCMV: (forward, dACAACAGATGGCTGGCAAC; reverse, dAAATGGGCGGTAGGCGTG). The PCR mixture was run on an agarose gel to verify the appropriate size of the amplified bands and these were also excised and sequenced. After sequence confirmation, Ad stocks were bulked up by infection of 8⫻ T175 flasks of HEK-293 cells followed by extraction and caesium chloride gradient purification, filter sterilization and aliquoted storage at ⫺80 C. Viral titer was determined using a standard plaque assay and is reported as multiplicity of infection (m.o.i.) where an m.o.i. of 1 is defined as 1 plaque forming unit per plated ␣T4 cell. Because ␣T4 cells will have proliferated between plating and infection, the values given over-estimate the actual m.o.i. at the time of infection.

Flow cytometry

Endo • 2000 Vol. 141 • No. 12

Dynamic video imaging of cytosolic Ca2⫹ Video imaging of fura 2 loaded ␣T4 cells was performed as described (39). Cells were loaded for 30 min in 1 ml PSS containing 2 ␮m fura 2. They were then washed several times in PSS and the cover slips were loaded into a holder which was fitted into a heating chamber at 37 C. Image capture was typically performed within 10 –25 min of loading in approximately 500 ␮l of PSS or in Ca2⫹-free PSS (containing 250 ␮m EGTA instead of CaCl2) using MagiCal hardware, Tardis software and a Nikon Diaphot microscope (39). The cells were excited alternately at 340 and 380 nm and emitted light was collected at 510 nm, averaging the data from 8 or 16 video frames, and subtracting background values before ratioing. The ratio of fluorescence at 340 and 380 nm, was calculated on a pixel-by-pixel basis and used to determine the Ca2⫹ concentration assuming a dissociation constant of 225 nm for fura-2 and Ca2⫹ at 37 C. Calibration was performed as described (39), and where spike and plateau Ca2⫹ values are reported these were defined as the maximum response within 10 sec of stimulation and the response after 1 min, respectively.

Radioligand binding

Flow cytometry was used to compare the efficiencies of different transfection procedures by transfecting ␣T4 cells with pEGFP-N1 plasmid (encoding EGFP) or with recombinant Ad encoding EGFP (AdEGFP). Analyses were performed on a FACScalibur flow cytometer using CellQuest software (both from Becton Dickinson and Co., Oxford, UK) for data acquisition. Cells were seeded into 24 well plates (100,000 cells per well), transfected 24 h later and then collected for flow cytometry after a further 18 h of culture. Standard procedures were used for CaPO4 transfection and manufacturers instruction were followed for transfections using 1 ␮g of EGFP DNA and Fugene 6 (3 ␮l for 6 h), Lipofectin (2.5 ␮l for 6 h) or Lipofectamine with the Plus Reagent (4 ␮l Plus Reagent and 1 ␮l Lipofectamine for 3 h). For the Ad-mediated transfections, cells were incubated in the presence of the Ad-EGFP at an m.o.i. of 1–1000 as indicated. The cells were then trypsinized, suspended in serum supplemented DMEM, and centrifuged (4 min, 300 ⫻ g, 4 C). The cells were washed in 4 ml FACS buffer (PBS, 5% FCS, 0.1% sodium azide), centrifuged and resuspended in FACS buffer. The cells were then analyzed at 100 –300 cells/sec and a minimum of 1 ⫻ 104 gated single cell events was acquired. The laser excitation wavelength was 488 nm and the photo-multiplier band pass filter was Fl1, 505 nm. Plots of event number against fluorescence intensity were used for calculation of mean fluorescence intensity, the proportion of cells that were positive (e.g. fluorescence intensity above a threshold set to exclude ⬎90% of untransfected cells), and the mean fluorescence intensity within the positive cell population. 3

3

Accumulation of [ H]inositol phosphates ([ H]IPx ) and Ins (1,4,5)P3 quantification [3H]IPx accumulation was used as a measure of PLC activity as described (19) using cells labeled by preincubation with [3H]inositol and stimulated in the presence of LiCl. Cells were cultured in 24-well plates in 1 ml of media and 2 ␮Ci [2-3H]inositol (14 –16 Ci/mmol) was added to each well for the final 16 h of incubation. After two washes in physiological salt solution (PSS: 127 mm NaCl, 1.8 mm CaCl2, 5 mm KCl, 2 mm MgCl2, 0.5 mm NaH2PO4, 5 mm NaHCO3, 10 mm glucose, 0.1% BSA, and 10 mm HEPES, pH 7.4) each well was stimulated for the period indicated in the figures with 200 –250 ␮l of PSS containing 10 mm LiCl and the indicated concentration of GnRH, Buserelin, or cGnRH-II. The stimulation was terminated by adding 1 ml of water at 95 C. The cells were lyzed by freezing and thawing, and [3H]IPx was separated from free [3H]inositol using anion exchange chromatography in formate form Dowex-1 columns (19). d-Ins (1,4,5)P3 mass was determined using a RRA as described (19). Briefly, ␣T4 cells cultured and infected with Ad in 24 well plates were washed and incubated for 30 min in 1 ml of Krebs/ HEPES buffer at 37 C. The medium was aspirated and replaced with 150 ␮l Krebs/HEPES with GnRH or cGnRH-II (both at 10⫺7 m). Incubations were performed in duplicate and were terminated at 5–300 sec by addition of 150 ␮l ice-cold 1 m TCA. The control (0 sec) time-point was obtained by adding the stimulus after the TCA. The d-Ins (1,4,5)P3 was then extracted as described (19) using duplicate aliquots for each sample and standards (0.1 nm–3 ␮m).

GnRH-R expression was assessed using whole cell binding assays in which approximately 50,000 cells were incubated in suspension for 30 min at 21 C in 100 ␮l of PSS containing 1 mg/ml bacitracin with approximately 10⫺10 m radiolabel and 0 or 10⫺11–10⫺5 m of the unlabeled competitor peptide (36). Free and bound peptide were then separated by centrifugation through oil (36). For human GnRH-R, the radiolabel was [125I]Buserelin. For Xenopus GnRH-R the radiolabel was [125I]cGnRH II. Receptor internalization was quantified in a modified whole cell binding assay in which approximately 50,000 cells were grown in 24-well plates, were washed in PSS and then incubated at 37 C in 200 ␮l PSS containing approximately 10⫺10 m radiolabel and 0 (total binding) or 10⫺6 m (nonspecific binding) of Buserelin or cGnRH-II. After the required incubation period (2– 60 min) the cells were rapidly rinsed in ice-cold PSS and then incubated for 2 min either in PSS or in 150 mm NaCl with 50 mm acetic acid (pH 3– 4). The cells were then washed again in PSS and solubilized in 0.5 ml of 0.2 m NaOH with 1% SDS. Radiolabel in the solubilized cells was determined by ␥-counting and specific cell-associated radioactivity was determine by subtraction of nonspecific from the total. Total specific binding is defined as the specific binding in cells receiving no acid wash, whereas acid-resistant (internalized) specific binding is defined as that seen in the acid washed cells. For one series of experiments, an internalization index was calculated by expressing acid-resistant specific binding as a percentage of total cell-associated specific binding.

Statistical analysis and data presentation The figures show data from a single representative experiment or the mean ⫾ sem of data pooled from “n” independent experiments (raw data or data normalized as described in the figure legends). Data are typically reported in the text as mean ⫾ sem and statistical analysis was by Student’s t test, accepting P ⬍ 0.05 as statistically significant. EC50 values were estimated by nonlinear regression using GraphPad Software, Inc. Prism (GraphPad Software, Inc., San Diego, CA). For Ca2⫹ measurements, image analysis was used to quantify the mean ionized Ca2⫹ in all of the cells in each field of view (which typically contained 10 –50 cells) as well as in individual cells. The figures show the mean (with or without sem) of data pooled from the indicated number of fields of view or in individual cells, as indicated.

Results

In preliminary experiments we attempted to transfect ␣T4 cells with hGnRH-R by CaPO4 precipitation but transfection efficiency was extremely poor. Because others have shown efficient Ad-mediated transfection of pituitary cells (28), we exploited EGFP expressing vectors to compare transfection efficiencies using Ad and 4 other transfection strategies. Flow cytometry (used to measure mean fluorescence intensity in the entire population and reported in arbitrary units) revealed that mean EGFP levels (expressed as arbitrary fluo-


rescence units) were much greater in cells transfected using the recombinant Ad (470,000) than in control cells (100) or in cells transfected with pEGFP-N1 using CaPO4 (150), Fugene 6 (1900), Lipofectin (30,000) or Lipofectamine with the Plus Reagent (50,000). This distinction was largely due to the fact that ⬎95% of cells expressed EGFP after infection with AdEGFP, whereas ⬍15% expressed EGFP after transfection using the other strategies (not shown). We therefore developed recombinant Ad encoding human and Xenopus GnRH-R (Ad hGnRH-R and Ad XGnRH-R). Insertion of receptor complementary DNA (cDNA) into the adenoviral DNA was confirmed by HindIII restriction digests, by PCR and by sequencing of the PCR products, as described in Materials and Methods. In each case, inserts of appropriate length and sequence were verified (not shown). We next explored the relationship between viral titer and receptor expression by infecting ␣T4 cells with Ad hGnRH-R and constructing competition binding curves using approximately 0.25 nm [125I]Buserelin and varied amounts of unlabeled Buserelin. No specific binding of [125I]Buserelin was seen in the untransfected ␣T4 cells, but infection with Ad hGnRH-R at increasing titer (from m.o.i. values of 3–300) increased [125I]Buserelin binding and this was inhibited in a concentration-dependent manner by Buserelin (Fig. 1A, main panel). Fitting these data to a single site, competition model revealed no dependence of Kd values on viral titer, so Bmax values were estimated by re-fitting the data with the Kd fixed at the mean value of 2.0 nm (2.0 ⫾ 1.2, n ⫽ 3). As shown (Fig. 1A, inset), increasing Ad titer from 3 to 100, increased receptor number from approximately 300 to 30,000 sites/cell. Similar studies performed with [125I]cGnRH-II and cGnRH-II in ␣T4 cells infected with Ad XGnRH-R (Fig. 1B, main panel) revealed high affinity binding sites and again, no relationship was observed between Ad titer and Kd. The data were therefore fitted through the mean Kd value of 3.0 nm (3.0 ⫾ 1.1, n ⫽ 4), revealing that increasing Ad titer from 3 to 100 increased receptor number from approximately 1,000 to 220,000 sites/cell (Fig. 1B, inset). To establish whether these binding sites were indeed functional GnRH-R, GnRH-stimulated [3H]IPx accumulation was measured in cells infected with Ad hGnRH-R at varied titer. As shown (Fig. 2, upper panel) GnRH did not stimulate [3H]IPx accumulation in untransfected ␣T4 cells but clearly did so after infection, and this effect was dependent upon both GnRH dose and viral titer. Similar results were obtained in cells infected with Ad XGnRH-R and then stimulated with cGnRH-II (Fig. 3, lower panel), although the higher levels of receptor expression achieved with the Ad XGnRH-R (compare insets in Fig. 1) were not associated with greater [3H]IPx responses (compare upper and lower panels in Fig. 2). Indeed, a plot of receptor number against maximal [3H]IPx response (data derived from those in Figs. 1 and 2) revealed halfmaximal [3H]IPx accumulation with at least 30,000 XGnRH-R per cell, compared with only 6,000 hGnRH-R per cell (Fig. 3). In parallel experiments, effects of 10⫺7 m GnRH and 10 ␮m forskolin on cAMP accumulation were measured by stimulation of cells infected with Ad hGnRH-R, for 30 min in medium with 1.0 mm IBMX. Infection with Ad hGnRH-R did not measurably alter basal cAMP values. Moreover, although forskolin clearly stimulated cAMP accumulation (11.1 ⫾ 1.1

4567

FIG. 1. Titer-dependence of receptor expression in ␣T4 cells infected with Ad hGnRH-R or Ad XGnRH-R. ␣T4 cells cultured in 60-mm Petri dishes were infected with hGnRH-Ad (upper figure) or XGnRH-R (lower figure) at an m.o.i. of 10 (triangles), 30 (open circles) or 100 (filled circles) then cultured for 2 days before being scraped from the culture vessels and used for suspension binding assays using approximately 0.25 nM [125I]Buserelin and the indicated concentration of unlabelled Buserelin (upper figure) or approximately 0.125 nM [125I]cGnRH-II and the indicated concentration of cGnRH-II (lower figure). Pooled Kd values were 2.0 ⫾ 1.2 nM (n ⫽ 3) for Buserelin binding to hGnRH-R and 3.0 ⫾ 1.1 (n ⫽ 4) for cGnRH-II binding to XGnRH-R and the values shown are means ⫾ SEM (n ⫽ 3– 4) normalized as a percentage of the binding seen without competitor in cells infected at m.o.i. values of 100. The insets show the numbers of receptors per cell calculated from Bmax values derived by fitting curves through the pooled Kd values and normalization according to cell number determined in parallel experiments (mean ⫾ SEM, n ⫽ 3).

fold over basal, n ⫽ 5), GnRH failed to do so in untransfected cells, or in cells infected with Ad hGnRH-R at m.o.i. values of 10 –300 (not shown). Similarly, cGnRH-II (10⫺7 m) failed to increase cAMP accumulation in cells infected with Ad XGnRH-R at m.o.i. values of 3–300 (not shown). Because activation of PLC-coupled GnRH-R increases [Ca2⫹]i in ␣T3–1 cells (39), we investigated this possibility in ␣T4 cells. After infection with Ad hGnRH-R, GnRH caused a robust increase in [Ca2⫹]i that began to decline after approximately 20 sec and returned rapidly to control values on transfer to Ca2⫹-free medium (Fig. 4). A similar effect was seen when

4568


FIG. 2. Titer-dependence of [3H]IPx accumulation in ␣T4 cells infected with Ad hGnRH-R or Ad XGnRH-R. ␣T4 cells cultured in 24-well plates were infected with Ad hGnRH-R or Ad XGnRH-R at m.o.i. values of 0 (open circles), 10 (filled circles), 30 (open triangles) or 100 (filled triangles) then cultured for 2 days. [3H]inositol was added to the medium for the final 16 h of culture, after which the cells were washed and stimulated for 30 min with the indicated concentration of GnRH (upper panel hGnRH-R) or cGnRH-II (lower panel, XGnRH-R), in the presence of 10 mM LiCl. Data shown are the means (n ⫽ 3) from three experiments, each having duplicate or triplicate determinations. SEM, which have been omitted for clarity, were mostly ⬍5%. For data normalization the [3H] eluted in the IP fraction was expressed as a % of that in the free [3H]inositol and IP fractions (e.g. fraction 3 as a % of fraction 1 plus fraction 3). This provides an internal control for cell number and labeling efficiency and thereby facilitates pooling of data from repeated experiments (51).

XGnRH-R infected cells were stimulated with cGnRH-II (10⫺7 m) although in this case a more pronounced biphasic (spike-plateau) response was observed. Similar responses were observed in parallel experiments using ␣T3–1 cells (not shown, see Ref. 39). The hGnRH-R data shown in Fig. 4 are from cells infected with Ad at an m.o.i. of 100, but were taken from a series of experiments in which m.o.i. values were varied between 0.3 and 1000. In these experiments, GnRH failed to increase [Ca2⫹]i in control cells but caused a titer-dependent increase in [Ca2⫹]i after infection with Ad hGnRH-R (Fig. 5, upper panel). Because these data were obtained by [Ca2⫹]i imaging, Ca2⫹ traces for individual cells were available, so the proportion of cells responding to GnRH could be estimated. As shown (Fig. 5, upper panel), The proportion of GnRH-responsive cells also increased with increasing viral titer but, interestingly, over 50% of the cells were judged as GnRH responsive after infection with an m.o.i. of 3, whereas an m.o.i. of 30 was required to achieve a 50% maximal increase in [Ca2⫹]i. Assuming that Ca2⫹ responses are only observed in GnRH-R expressing cells, this implies that the vast majority of cells express GnRH-R at low viral titer and that increasing titer then increases responses by increasing the amount of

Endo • 2000 Vol. 141 • No. 12

FIG. 3. Relationship between receptor number and maximal GnRHstimulated [3H]IPx accumulation. Maximal [3H]IPx accumulation (e.g. that seen with 10⫺8–10⫺6 M GnRH or cGnRH-II) was determined in cells which had been infected with Ad hGnRH-R (open circles) or Ad XGnRH-R (filled circles) at m.o.i. values of 3–300 (data taken from Fig. 2) and plotted against the corresponding receptor number, determined by saturation analysis (data from Fig. 1). Curve fitting (assuming a sigmoid relationship between [3H]IPx response and log10 receptor number) revealed half-maximal responses at 6,000 and 100,000 sites per cell for the hGnRH-R and XGnRH-R, respectively.

FIG. 4. Comparison of GnRH effects on [Ca2⫹]i in ␣T4 cells infected with Ad hGnRH-R or Ad XGnRH-R. ␣T4 cells were infected with Ad hGnRH-R (m.o.i. 1000) or Ad XGnRH-R (m.o.i. 100) and then cultured for one day before being loaded with fura 2 and used for Ca2⫹ imaging. During imaging Ad hGnRH-R infected cells were stimulated with 10⫺7 M GnRH (filled circles) and Ad XGnRH-R infected cells were stimulated with cGnRH-II (filled triangles) in normal medium before being transferred to Ca2⫹-free medium (still with peptide). Control cells (open circles) were infected with Ad and then stimulated with medium alone before transfer to Ca2⫹-free medium without peptide. The control data for hGnRH-R and XGnRH-R expressing cells were indistinguishable and have been pooled for clarity. The data shown are means ⫾ SEM (n ⫽ 3– 6) where the average response of the entire field of view (containing 10 –50 cells) was calculated for each experiment.

receptor per cell, rather than the proportion of cells expressing the receptor. To investigate the relationship between viral titer and protein expression more directly, flow cytometry was used to define both the proportion of cells expressing EGFP and the intensity of EGFP fluorescence in the EGFP


FIG. 5. Titer-dependence of [Ca2⫹]i responses to GnRH and EGFP expression in ␣T4 cells. Upper panel, ␣T4 cells were infected with Ad hGnRH-R at the indicated m.o.i. and then cultured for 1 day before being loaded with fura 2 and used for [Ca2⫹]i imaging precisely as described under Fig. 4. The maximum concentration after stimulation with 10⫺7 M GnRH (average response in each field of view containing 10 –50 cells), and the proportion of cells responding to GnRH in each field of view, were both determined. The figure shows means ⫾ SEM (n ⫽ 2– 8) for the maximal [Ca2⫹]i increases (filled circles) and for the proportion of cells responding (open circles). Curve fitting (assuming a sigmoid relationship between response and log10 of Ad titer) revealed half-maximal responses at m.o.i. values of 30 and 2.5 for the mean [Ca2⫹]i response, proportion of cells responding, respectively. Lower panel, ␣T4 cells were infected with Ad-EGFP at the indicated m.o.i. then cultured for 24 h before flow cytometric analysis which was used to calculate the proportion of positively stained cells (e.g. those having fluorescence values at least 3-fold higher than basal, open circles) and the mean fluorescence in the positively stained cell population (filled circles, arbitrary units). The data are means of duplicate observations in a representative experiment.

positive cells, after infection with Ad-EGFP at varied titer (Fig. 5, lower panel). As shown, ⬎95% of cells were positively stained, and fluorescence intensity in these cells was increased to ⬎100 fold above basal, after infection with AdEGFP at an m.o.i. ⬎10. Importantly, increasing viral titer from m.o.i. values of 10 to 1000 increased mean fluorescence by 40-fold without appreciably altering the proportion of cells fluorescing, demonstrating (as implied by the Ca2⫹ imaging data) that increasing viral titer above an m.o.i. of 10, causes an increase in protein expression per cell rather than an increase in proportion of cell expressing the protein. A further intriguing observation was that the [Ca2⫹]i responses to GnRH were qualitatively dependent upon viral titer. As

4569

shown, oscillations in basal [Ca2⫹]i were seen in many control cells (Fig. 6, left-hand traces), which were unaltered by GnRH but were rapidly extinguished on transfer to Ca2⫹-free medium. In cells infected with hGnRH-R at low m.o.i. (1 or 3) most cells failed to respond to GnRH but where responses were observed, GnRH increased the amplitude and/or frequency of the [Ca2⫹]i oscillations and these were, again, inhibited by transfer to Ca2⫹-free medium (Fig. 6, central traces). In contrast, addition of GnRH to cells infected with Ad hGnRH-R at high titer (10 –100), caused spike-plateau type responses instead of oscillations (Fig. 6, right-hand traces). When competition binding assays were performed with [125I]Buserelin and unlabeled ligands (GnRH, Buserelin and cGnRH-II) in ␣T4 cells infected with Ad hGnRH-R (m.o.i. 100), all 3 peptides competed for [125I]Buserelin with the rank order of potency (Buserelin ⬎ GnRH⬎⬎cGnRH-II) expected for a mammalian GnRH-R Fig. 7, upper left panel). Similarly, all three peptides competed for binding of [125I]cGnRH-II in cells infected with XGnRH-R (m.o.i. 100, Fig. 7, upper right panel), but in this case the rank order of potency was cGnRHII⬎⬎Buserelin⬎⬎GnRH. Identical rank-orders of potency were seen when ␣T4 cell infected with Ad hGnRH-R or Ad XGnRH-R were used for construction of dose-response curves for [3H]IPx accumulation (Fig. 7, lower left and right panels, respectively). To test for rapid desensitization of GnRH-R, ␣T4 cells were infected with Ad XGnRH-R or with Ad hGnRH-R (both at a m.o.i. values of 100 –300) and then used in time-course experiments in which accumulation of [3H]IPx and Ins (1,4,5)P3 mass were measured. As shown (Fig. 8, upper panel), GnRH caused a rapid accumulation of [3H]IPx in hGnRH-R expressing cells, with no obvious reduction in accumulation rate over time. In cells expressing XGnRH-R, cGnRH-II also caused a rapid increase in [3H]IPx accumulation, but in this case, the initial rate of accumulation was not maintained beyond 2 min Thus, although the initial rates of [3H]IPx accumulation (0 –2 min) were indistinguishable, the final rate of XGnRH-R mediated accumulation (2–15 min) was only 30% of that with the hGnRH-R, and only 40% of the initial XGnRH-R mediated rate (Fig. 8, upper panel). The timecourses of receptor-mediated elevation of Ins (1,4,5)P3 mass were also markedly different for the two receptors (Fig. 8, lower panel). In Ad XGnRH-R infected cells, cGnRH-II caused a rapid and pronounced increase in Ins (1,4,5)P3, but this response was transient with a maximum at 5 sec and a marked reduction thereafter. In contrast, the response to GnRH in Ad hGnRH-R infected cells was less pronounced and was not transient. GnRH increased Ins (1,4,5)P3 mass to a maximal level at 60 sec, and this was maintained for 5 min. To quantify internalization of GnRH-R, ␣T4 cells were infected with Ad hGnRH-R or with Ad XGnRH-R (both at a m.o.i. values of 100 –300) and then used in time-course experiments in which total specific binding and acid-resistant specific binding were measured. The time-courses for specific association of [125I]Buserelin with hGnRH-R, and of [125I]cGnRH-II with XGnRH-R, were indistinguishable (not shown). However, the acid-resistant fraction of the specific binding was greater with the XGnRH-R at all time-points (Fig. 9), such that after 60 min, 75% of specific binding to cells

4570


Endo • 2000 Vol. 141 • No. 12

FIG. 6. Titer-dependence of Ca2⫹ responses of individual cells to GnRH in ␣T4 cells infected with Ad hGnRH-R. The figure shows responses for individual cells from the experiments described in Fig. 4. These have been selected to show the characteristic [Ca2⫹]i responses seen in uninfected cells (left-hand traces), in cells infected with Ad hGnRH-R at low titer (middle traces) or at high titer (right-hand traces). Traces A, B, C, D, and E are off-set on the vertical axis by 5,000; 4,000; 3,000; 2,000; and 1,000 nM (respectively) for clarity.

expressing XGnRH was internalized (acid resistant), whereas only 25% of binding to hGnRH-R was internalized. In the final experiments, the possible influence of receptor number of GnRH-R internalization was assessed by calculation of an internalization index (specific acid-resistant binding as a percentage of total specific binding) for binding to ␣T4 cells infected with Ad hGnRH-R and Ad XGnRH-R at m.o.i. values of 12.5–200. As shown (Fig. 10, upper panel), the internalization index for XGnRH-R was greater than that for the GnRH-R at all m.o.i. values tested. Discussion

Most vertebrates express at least two forms of GnRH (9) and distinct classes of GnRH-Rs have apparently evolved in parallel with their ligands (44, 45). The cloned mammalian GnRH-Rs have high sequence homology and are selective for GnRH but have lower affinity for cGnRH-II, whereas the nonmammalian GnRH-Rs have higher selectivity for cGnRH-II (9, 12, 15–17). Comparative studies in heterologous expression systems (15, 23, 25) have revealed that nonmammalian GnRH-R (chicken and catfish) desensitize and internalize more rapidly than their mammalian counterparts (human and rat), leading to the suggestion (15) that the lack of receptor desensitization is advantageous in mammals for generation of the preovulatory gonadotropin surge driven by increasing exposure of gonadotropes to GnRH (15). However, this argument rests on the untested assumption that the functional characteristics of mammalian and nonmammalian GnRH-R observed in COS and HEK-293 cells, are maintained in gonadotropes and at physiological receptor density. Given the increasing evidence that the signaling and/or desensitization of GPCRs (including GnRH-R) can depend upon receptor density and cellular context (26 –33), we sought to test this assumption by comparative studies in gonadotrope progenitor ␣T4 cells. In the first experiments, we compared the efficiency of transfection using CaPO4 precipitation and liposome based

strategies to infection with Ad-EGFP and found Ad mediated transfection to be far more effective (higher levels of EGFP expressed and larger proportion of cells transfected). We therefore prepared recombinant Ad expressing human GnRH-R and found that infection with these causes expression of GnRH-R in ␣T4 cells, which have the pharmacological characteristics of their endogenous counterparts. Thus, the affinity of [125I]Buserelin binding to Ad hGnRH-R infected ␣T4 cells (Kd approx. 2 nm), the potency of GnRH-stimulated [3H]IPx accumulation (EC50 approx. 1 nm), the rank orders of potency for the 3 peptides in [3H]IPx accumulation assays and competitive binding assays (Buserelin ⬎ GnRH⬎⬎cGnRH-II) are all similar to data obtained with mammalian GnRH-Rs in pituitary cells and with hGnRH-R expressed in other heterologous systems (9, 30, 46, 47). Moreover, Ca2⫹ imaging studies revealed that these receptors, like endogenous mammalian GnRH-R in other systems (6 –9, 19, 39) cause a biphasic (spike-plateau) increase in [Ca2⫹]i. In each case the spike response is mediated by Ca2⫹ mobilization from intracellular stores [presumably Ins (1,4,5)P3 mediated], as judged by retention of a spike response in Ca2⫹free medium (not shown), whereas the plateau phase is dependent upon Ca2⫹ entry across the plasma membrane, as demonstrated by loss of the response on transfer to Ca2⫹-free medium. Although XGnRH-R have recently been expressed in COS cells, receptor levels were too low for radioligand binding studies (12). This problem was obviated using Ad XGnRH-R infected ␣T4 cells that expressed receptors with high affinity for [125I]cGnRH-II (Kd ⬃3 nm) and specificity for cGnRH-II (cGnRH-II⬎⬎Buserelin ⬎ GnRH). These characteristics are similar to those reported for the endogenous GnRH-R of amphibian pituitary extracts (48). Moreover, the potency of cGnRH-II stimulated [3H]IPx accumulation (EC50 approx. 2 nm) and the rank order of potency for the three peptides in [3H]IPx accumulation assays (cGnRHII⬎⬎Buserelin ⬎ GnRH) parallel the binding and functional data obtained with other nonmammalian GnRH-Rs and the


4571

FIG. 7. Effects of GnRH, cGnRH-II and Buserelin on [3H]IPx accumulation and radioligand binding in ␣T4 cells infected with Ad hGnRH-R or Ad XGnRH-R. Upper panels, ␣T4 cells were infected with Ad hGnRH-R or Ad XGnRH-R (each at m.o.i. values of 100 –300), then cultured for 1–2 days before being scraped from the culture vessels and used for suspension binding assays using approximately 0.25 nM [125I]Buserelin (hGnRH-R) or 0.125 nM [125I]cGnRH-II (XGnRH-R) and the indicated concentration of unlabeled GnRH, open circles), Buserelin (open triangles) or cGnRH-II (filled circles). Lower panels, ␣T4 cells cultured in 24-well plates were infected with hGnRH Ad or Ad XGnRH-R (each at an m.o.i. of 100 –300) and then cultured for 1–2 days. [3H]inositol was added to the medium for the final 16 h of culture, after which the cells were washed and stimulated for 30 min with the indicated concentration of peptide (symbols as above) in the presence of 10 mM LiCl. The data shown in both panels are mean ⫾ SEM (n ⫽ 2– 4) from repeated experiments (each having duplicate or triplicate observations). Binding data were normalized as a % of that seen without competitor and [3H]IPx responses were normalized as a % of the maximum response seen in cells expressing hGnRH-R.

XGnRH-R (12, 48, 49). Similarly, the biphasic effect of cGnRH-II on [Ca2⫹]i in XGnRH-R infected ␣T4 cells, parallels that seen on activation of nonmammalian GnRH-R in gonadotropes (50). Interestingly, when the relationship between viral titer and EGFP expression (Ad-EGFP) was defined, increasing Ad titer from an m.o.i. of 3 to 1000 caused a major increase in EGFP fluorescence, without appreciably altering the proportion of positively stained cells (80 –100%). This implies that increasing m.o.i. above 10 causes an increase in protein per cell, rather than in the proportion of cells expressing the protein. This apparently holds true for the GnRH-R because in the Ca2⫹ imaging experiments, increasing Ad hGnRH-R from an m.o.i. of 10 to 1000 did not increase the proportion of cells responding to GnRH, but instead, increased the amplitude of the response. Assuming that only GnRH-R expressing cells show Ca2⫹ responses to GnRH, the vast majority of cells must express GnRH-R after infection at an m.o.i. of 10, so that the 20- fold increase in receptor number caused by increasing Ad hGnRH-R from 10 to 300 (Fig. 1) reflects an increase in receptors per cell rather than an increase in the

proportion of cells expressing GnRH-R. Thus, recombinant Ad provides a simple means of controlling the number of GnRH-R in ␣T4 cells, highlighting an important distinction to conventional transient transfection strategies, where a relatively low (and usually unknown) proportion of cells are transfected so that it is unclear whether increasing DNA increases the proportion of cells expressing the protein or the amount of protein per positive cell. Using this strategy we are able to vary receptor number from approximately 1,500 to over 30,000 sites per cell for the hGnRH-R. This caused a corresponding increase in maximal GnRH-stimulated [3H]IPx accumulation, although receptor affinity (Kd for [125I]Buserelin ⬃2 nm) and potency of GnRH (EC50 approx. 0.2 nm) were not measurably influenced by viral titer. No evidence was obtained for stimulation of cAMP accumulation or for constitutive receptor signaling (e.g. elevation of [3H]IPx accumulation or cAMP levels in the absence of stimulus) at any Ad hGnRH-R titer. We also found that stimulation of hGnRH-R in cells infected at low titer tended to cause asynchronous oscillatory Ca2⫹ responses, whereas stimulation of cells infected at high titer caused more sus-

4572


FIG. 8. Time-courses of [3H]IPx accumulation and Ins(1,4,5)P3 elevation in Ad XGnRH-R and Ad hGnRH-R infected cells. Upper panel, ␣T4 cells cultured in 24-well plates were infected with hGnRH Ad (open symbols) or with Ad XGnRH-R (filled symbols), each at an m.o.i. of 100 –300, and then cultured for 1–2 days. [3H]inositol was added to the medium for the final 16 h of culture, after which the cells were washed and stimulated for the indicated time in medium containing 10 mM LiCl and 10⫺7 M GnRH (hGnRH-R) or cGnRH-II (XGnRH-R). The data are the means ⫾ SEM (n ⫽ 4 –11). [3H]IPx data were first normalized as above (Fig. 2 legend) and then pooled after calculation of the fold increase over basal values. Linear regression analysis revealed comparable initial rates of [3H]IPx accumulation (0.18- and 0.19-fold basal/min during 0 –2 min) for XGnRH-R and hGnRH-R respectively, whereas the final rate of [3HIPx accumulation in XGnRH-R expressing cells was only 30% of that in hGnRH-R expressing cells (0.08- and 0.28-fold basal/min, respectively, during 2–15 min). Lower panel, Ad hGnRH-R (open circles) or Ad XGnRH-R (filled circles) infected cells were stimulated with 10⫺7 M GnRH or cGnRH-II (respectively) for the indicated time. The incubations were terminated by addition of TCA and Ins(1,4,5)P3 was then extracted and measured as described in Materials and Methods. The figure shows the means ⫾ SEM (n ⫽ 4) pooled from four separate experiments, each with duplicate observations.

tained spike-plateau responses (Fig. 7). In rat gonadotropes, low concentrations of GnRH cause oscillatory Ca2⫹ response and increasing GnRH concentration increases spikefrequency until sustained responses are achieved (8, 51). Thus, it appears that this same relationship is seen when the number of active receptors is controlled by altering ligand concentration (in the face of a constant receptor number) or by altering receptor number (in the face of constant ligand concentration). These data were, however, unexpected because oscillatory Ca2⫹ responses are not seen in ␣T3–1 cells (39, 51) implying that immortalization of this cell line has somehow prevented them from displaying this fundamental characteristic of gonadotropes. Clearly, this is not the case for ␣T4 cells, which may therefore prove to be valuable models for exploration of the mechanisms and relevance of oscillatory Ca2⫹ signals in gonadotropes.

Endo • 2000 Vol. 141 • No. 12

FIG. 9. Internalization of XGnRH-R is more rapid than internalization of hGnRH-R in ␣T4 cells. ␣T4 cells cultured in 24-well plates were infected with Ad hGnRH-R or Ad XGnRH-R (each at m.o.i. values of 100 –300), then cultured for 1–3 days before being washed and incubated for the indicated period in medium containing approximately 0.25 nM [125I]Buserelin (hGnRH-R, open circles) or 0.125 nM [125I]cGnRH-II (XGnRH-R, filled circles) with 0 or 10⫺6 M Buserelin or cGnRH-II (nonspecific binding). The binding was terminated by transfer to ice-cold PSS and washing before solubilization of the cells in NaOH. Alternatively, the cells were treated exactly as described above except that, after binding, the cells were incubated in acid medium to remove radioligand from cell surface receptors. The data shown are means ⫾ SEM (n ⫽ 3) of acid-resistant specific binding, each having duplicate or triplicate observations, which were expressed as a percentage of total specific binding (acid labile plus acid resistant) at each time point.

FIG. 10. Internalization of XGnRH-R is more rapid than internalization of hGnRH at varied receptor density. ␣T4 cells cultured in 24 well plates were infected with hGnRH-Ad (open circles) or Ad XGnRH-R (filled circles) at the indicated m.o.i. They were then cultured and used for radioligand binding experiments exactly as described above (Fig. 9 legend), except that all incubations were for 5 min. The data shown are means ⫾ SEM (n ⫽ 5–9) of acid-resistant specific binding expressed as a percentage of total specific binding pooled from repeated experiments, each having duplicate or triplicate observations.

GnRH-R expression in gonadotropes is hormonally regulated. The dynamic range for rat GnRH-R is approximately 20,000 –75,000 sites/gonadotrope through the oestrous cycle (49, 52), and sheep GnRH-R can be manipulated through a range of 500 –20,000 sites/cell by in vitro exposure to gonadal steroids (53). The levels of hGnRH-R expression reported here (1,500 –30,000 sites/cell) therefore compare favorably to the physiological range, as well as to the levels in stable GnRH-R expressing cells lines [approximately 10,000 sites/ cell in GGH3–1⬘ cells (32, 33) and 65,000 sites/cell in ␣T3–1 cells (36)]. Higher expression levels (up to 220,000 sites/cell)


were obtained with Ad XGnRH-R, although data are not available to relate this to physiological XGnRH-R receptor density. The higher expression of XGnRH-R could reflect differences in protein synthesis, elaboration to the cell surface or stability. A similar distinction has been observed for rat and catfish GnRH-R expressed transiently in GH3 cells where levels were 5- to 8-fold greater with the tailed GnRH-R and expression of the rat GnRH-R was increased by addition of the catfish GnRH-R tail (54). The lack of a second glycosylation site toward the amino terminus of the human GnRH-R may also contribute to the relatively low expression level in ␣T4 cells, as it does in COS cells (55). As with the hGnRH-R, no evidence was obtained for constitutive signaling or for stimulation of cAMP accumulation with the XGnRH-R in ␣T4 cells. The latter observation was unexpected because the closely related catfish GnRH-R was cloned by virtue of its ability to activate a cAMP response element binding protein reporter in HEK-293 cells (17) and because mammalian GnRH-R mediate activation of Gs in GH3 and COS cells (31–33). It therefore appears that the specific coupling of GnRH-R to Gq seen in ␣T3–1 cells, also exists in ␣T4 cells and is retained for both hGnRH-R and XGnRH-R. When [3H]IPx responses were compared in cells infected with Ad hGnRH-R and Ad XGnRH-R at varied titer (Fig. 4), increasing receptor number increased ligand-stimulated [3H]IPx accumulation but [3H]IPx responses were lower in XGnRH-R expressing cells, despite the fact that XGnRH-R number was greater than that for hGnRH-R. Indeed, a plot of maximal [3H]IPx response against estimated receptor number revealed that a 50% maximal [3H]IPx response required activation of at least 30,000 XGnRH-R as opposed to only 6,000 hGnRH-R. Because these data are obtained using a 30-min stimulation, this difference could be due to more pronounced desensitization and/or internalization of the XGnRH-R, occurring during the stimulation. The hallmarks of rapid homologous desensitization of PLC-activating GPCRs are a failure to maintain initial rates of ligand stimulated [3H]IPx accumulation (against a LiCl block of IP metabolism) and a transient elevation of Ins (1,4,5)P3 mass caused by concomitant Ins (1,4,5)P3 metabolism and a reduction in the rate of Ins (1,4,5)P3 generation due to receptor desensitization (20). These characteristics were observed in XGnRH-R infected cells but not in hGnRH-R infected cells. In these experiments, the initial rates of XGnRH-R and hGnRH-R mediated [3H]IPx accumulation were indistinguishable and this rate was sustained by the hGnRH-R, but not by the XGnRH-R. Moreover, the XGnRH-R mediated a transient (spike-plateau) increase in Ins (1,4,5)P3 mass, whereas the hGnRH-R mediated only a monophasic increase to a maintained plateau. A similar distinction was observed when receptor internalization was assessed. Only 25% of specific binding to the hGnRH-R was acid resistant (internalized) after incubation for 60 min at 37 C, whereas over 75% of the binding to the XGnRH-R was acid resistant under the same conditions. These receptor-specific [3H]IPx and Ins (1,4,5)P3 profiles are unlikely to reflect differences in receptor number because we have shown (22) that the monophasic Ins (1,4,5)P3 profile seen on activation of mouse GnRH-R in ␣T3–1 cells, is maintained at a wide range of receptor den-

4573

sities (receptor density manipulated by partial irreversible blockade). Moreover, the differences in receptor internalization rates were not due to differences in receptor number because when Ad titer was varied (m.o.i. 12.5–200) ⬍10% of hGnRH-R binding was internalized, whereas ⬎50% of XGnRH-R binding was internalized, irrespective of Ad titer. Using the data inset in Fig. 3, A and B, we estimate that hGnRH-R number was varied from 2,000 to 30,000 sites/cell in this experiment, whereas XGnRH-R was varied from 2,000 to 220,000 sites/cell. Thus, the greater rate of XGnRH-R internalization was seen at a range of receptor densities encompassing that for the hGnRH-R, and cannot therefore reflect a difference in receptor number. One of the most remarkable aspects of this receptor family is that all cloned nonmammalian GnRH-Rs have C-terminal tails, whereas none of the cloned mammalian GnRH-Rs have such tails (9, 12, 15–17). Although GnRH-stimulated gonadotropin secretion clearly desensitizes in mammals, the lack of C-terminal tails in mammalian GnRH-R raised the question of whether receptor desensitization is involved. The endogenous murine GnRH-R of ␣T3–1 cells were found not to undergo rapid homologous desensitization, and this was shown to be a characteristic of the receptor rather than cell type or receptor density (18 –22). The implied causal relationship between the unique structural and functional features of these receptors (mammalian GnRH-R are the only GPCRs known to completely lack C-terminal tails and the only PLC-activating GPCRs known not to rapidly desensitize) is supported by the demonstration that activation of heterologously expressed catfish GnRH-R (or rat GnRH-R with added TRH receptor tails) causes ␤-arrestin translocation along with receptor phosphorylation, internalization, and desensitization, whereas all of these processes were absent or slower with the rat GnRH-R (23, 25). Similarly, the chicken GnRH-R internalized about 4 times more rapidly than the human GnRH-R, and this internalization rate is reduced by successive C-terminal truncations (15). Thus, although we have not tested this directly, the absence and presence of C-terminal most likely underlies the differences in desensitization and internalization rates of the hGnRH-R and XGnRH-R reported herein. In summary, development of recombinant Ad encoding GnRH-R has enabled the first comparative studies of different GnRH-R in gonadotrope lineage cells. In these cells, variation of Ad titer (above an m.o.i. of 10) provides an efficient means of controlling receptor number (per cell) within the physiological range. Human and Xenopus GnRH-R expressed in this way had pharmacological characteristics comparable to their endogenous counterparts (in terms of ligand recognition profiles and activation of PLC/Ca2⫹ signaling). No evidence was obtained for coupling to adenylyl cyclase or for constitutive receptor signaling at any receptor density but Ad hGnRH-R titer (and hence hGnRH-R number) had both quantitative and qualitative effects on the [Ca2⫹]i responses to GnRH. Interestingly, measurement of [3H]IPx accumulation in 30-min incubations revealed that sustained activation of PLC by XGnRH-R is much less efficient than activation by hGnRH-R, possibly because of differences in desensitization and/or internalization occurring during the stimulation. Indeed, internalization of the XGnRH-R was

4574


found to be more rapid than for the hGnRH-R, and the XGnRH-R rapidly desensitized, whereas the hGnRH-R did not. The retention of such differences at physiological receptor density in gonadotrope lineage cells supports the argument that the evolution of nondesensitizing GnRH-R is related to the development of mammalian reproductive strategies. Acknowledgments We are grateful to Dr. P. Mellon (UCSD, Department of Reproductive Medicine, La Jolla, CA) for providing the ␣T3–1 and ␣T4 cells, to Prof. Sandow (Aventis Pharma GmbH, Frankfurt, Germany) for providing the Buserelin and [125I]Buserelin, and to Dr. A. Byrne (University of Oxford, Oxford, UK) for providing vectors.

22.

23.

24. 25.

26.

References 1. Dohlman HG, Thorner J, Caron MG, Lefkowitz RJ 1991 Model systems for the study of seven transmembrane segment receptors. Annu Rev Biochem 60:653– 688 2. Lefkowitz RJ, Hausdorff WP, Caron MG 1990 Role of phosphorylation in desensitization of the ␤-adrenoceptor. Trends Pharmacol Sci 11:190 –194 3. Ferguson SSG, Downey WE, Colapietro A-M, Barak LS, Menard L, Caron MG 1997 Role of ␤-arrestin in mediating agonist-promoted G-protein-coupled receptor internalization. Science 271:363–366 4. Krueger KM, Daaka Y, Pitcher JA, Lefkowitz RB 1997 The role of sequestration in G-protein coupled receptor resensitization. J Biol Chem 272:5– 8 5. Lattion A-L, Diviani D, Cottechia S 1994 Truncation of the receptor carboxyl terminus impairs agonist-dependent phosphorylation and desensitization of the ␣1B-adrenergic receptor. J Biol Chem 269:22887–22893 6. Conn PM 1995 Gonadotropin-releasing hormone action. In: Adashi EY, Rock JA, Rosenwaks Z (eds) Reproductive Endocrinology, Surgery and Technology. Lippincott-Raven, Philadelphia, vol 1:163–179 7. Hille B, Tse A, Tse FW, Bosma MM 1995 Signalling mechanisms during the response of pituitary gonadotropes to GnRH. Recent Prog Horm Res 50:5–95 8. Stojilkovic SS, Catt KJ 1995 Expression and signal-transduction pathways of gonadotropin-releasing hormone receptors. Recent Prog Horm Res 50:161–205 9. Sealfon SS, Weinstein H, Millar RP 1997 Molecular mechanisms of ligand interaction with the gonadotropin-releasing hormone receptor. Endocr Rev 18:180 –205 10. McArdle CA, Poch A 1992 Dependence of gonadotropin-releasing hormone release upon intracellular Ca2⫹ pools is revealed by desensitization and by thapsigargin blockade. Endocrinology 130:3567–3574 11. Barbieri RL 1992 Clinical applications of GnRH and its analogues. Trends Endocrinol Metab 3:30 –34 12. Troskie BE, Hapgood JP, Millar RP, Illing N 2000 Complementary deoxyribonucleic acid cloning, gene expression, and ligand selectivity of a novel gonadotropin-releasing hormone receptor expressed in the pituitary and midbrain of Xenopus laevis. Endocrinology 141:1764 –1771 13. Tsutsumi M, Zhou W, Millar RP, Mellon PL, Roberts JL, Flanagan CA, Dong K, Gillo B, Sealfon SC 1992 Cloning and functional expression of a mouse gonadotropin-releasing hormone receptor. Mol Endocrinol 6:1163–1169 14. Chi L, Zhou W, Prikhozhan A, Flanagan C, Davidson JS, Golembo M, Illing N, Millar RP, Sealfon SC 1993 Cloning and characterization of the human GnRH receptor. Mol Cell Endocrinol 91:R1–R6 15. Pawson A. Katz A, Sun Y-M, Lopes J, Illing N, Millar RP, Davidson JS 1998 Contrasting internalization kinetics of human and chicken gonadotropinreleasing hormone receptors mediated by C-terminal tail. J Endocrinol 196:R9 –R12 16. Illing N, Troskie BE, Nahorniak CS, Hapgood JP, Peter RE, Millar RP 1999 Two gonadotropin-releasing hormone receptor subtypes with distinct ligand selectivity and differential distribution in brain and pituitary on the goldfish (Carassius auratus). Proc Natl Acad Sci USA 96:2256 –2531 17. Tensen C, Okuzawa K, Blomenrohr M, Rebers F, Leurrs R, Bogerd J, Schulz R, Goos H 1997 Distinct efficacies for two endogenous ligands on a single cognate gonadoliberin receptor. Eur J Biochem 243:134 –140 18. Davidson JS, Wakefield IK, Millar RP 1994 Absence of rapid desensitization of the mouse gonadotropein-releasing hormone receptor. Biochem J 300:299 –302 19. McArdle CA, Willars GB, Fowkes RC, Nahorski SR, Davidson JS, ForrestOwen W 1996 Desensitization of gonadotropin-releasing hormone action in ␣T3–1 cells due to uncoupling of inositol 1,4,5-trisphosphate generation and Ca2⫹ mobilization. J Biol Chem 271:23711–23717 20. McArdle CA, Davidson JS, Willars GB 1999 Desensitization at, and distal to, G-protein coupled receptors: the tail of the GnRH receptor. Mol Cell Endocrinol 151:129 –136 21. Willars GB, McArdle CA, Nahorski SR 1998 Acute desensitization of phos-

27. 28.

29. 30. 31. 32.

33.

34. 35.

36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

Endo • 2000 Vol. 141 • No. 12

pholipase C-coupled muscarinic M3 receptor, but not gonadotropin-releasing hormone receptor co-expressed in ␣T3–1 cells: implications for mechanisms of rapid desensitization. Biochem J 333:301–308 Forrest-Owen W, Willars GB, Nahorski SR, Assefa D, Davidson JS, Hislop J, McArdle CA 1998 The lack of gonadotrophin-releasing hormone receptor desensitization in ␣T3–1 cells is not due to GnRH receptor reserve or phosphatidylinositol 4,5 bis-phosphate pool size. Mol Cell Endocrinol 147:161–173 Heding A, Vrecl M, Bogerd J, McGregor A, Sellar R, Taylor PL, Eidne KA 1998 Gonadotropin-releasing hormone receptors with intracellular carboxylterminal tails undergo acute desensitization of total inositol phosphate production and exhibit accelerated internalization kinetics. J Biol Chem 273:11472–11477 McArdle CA, Forrest-Owen W, Willars G, Davidson J, Poch A, Kratzmeier M 1995 Desensitization of gonadotropin-releasing hormone action in the gonadotrope-derived ␣T3–1 cell line. Endocrinology 136:4864 – 4871 Heding A, Vrecl M, Hanyaloglu AC, Sellar R, Taylor PL, Eidne KA 2000 The rat gonadotropin-releasing hormone receptor internalizes via a ␤-arrestinindependent, but dynamin-dependent, pathway: addition of carboxyl-terminal tail confers ␤-arrestin dependency. Endocrinology 141:299 –306 Lucaites VL, Nelson DL, Wainscott DB, Baez M 1996 Receptor subtype and density determine the coupling repertoire of the 5-HT2 receptor family. Life Sci 59:1081–1095 Spurney RF 1998 Effect of receptor number on desensitization of the mouse thromboxane receptor. Biochem Pharmacol 55:1271–1281 Falck-Pedersen E, Heinflink M, Alvira M, Nussenzweig DR, Gershengorn MC 1994 Expression of thyrotropin-releasing hormone receptors by adenovirus-mediated gene transfer reveals that thyrotropin releasing hormone desensitization is cell-specific. Mol Pharmacol 45:684 – 689 Delahaye R, Manna PR, Berault A, Berreur-Bonnenfant J, Berreur P, Counis R 1997 Rat GnRH receptor expressed in insect cells induces activation of adenylyl cyclase. Mol Cell Endocrinol 135:119 –127 Kaiser UB, Conn PM, Chin WW 1997 Studies of gonadotropin-releasing hormone (GnRH) action using GnRH receptor-expressing pituitary cell lines. Endocr Rev 18:46 –70 Arora KK, Krsmanovic LZ, Mores N, O’Farrell H, Catt KJ 1998 Mediation of cyclic AMP signaling by the first intracellular loop of the gonadotropinreleasing hormone receptor. J Biol Chem 273:25581–25586 Kuphal D, Janovick JA, Kaiser UB, Chin WW, Conn PM 1994 Stable transfection of GH3 cells with rat gonadotropin-releasing hormone receptor complementary deoxyribonucleic acid results in expression of a receptor coupled to cyclic adenosine 3⬘,5⬘- monophosphate-dependent prolactin release via a G-protein. Endocrinology 135:315–320 Ulloa-Aguirre A, Stanislaus D, Arora V, Vaananen J, Brothers S, Janovick JA, Conn PM 1998 The third intracellular loop of the rat gonadotropin-releasing hormone receptor couples the receptor to Gs and Gq/11-mediated signal transduction pathways: evidence from loop fragment transfection in GGH3 cells. Endocrinology 139:2472–2478 Clayton RN, Shakespear RA, Marshall JC 1978 LH-RH binding to purified pituitary plasma membranes: absence of adenylate cyclase activation. Mol Cell Endocrinol 11:63–78 Horn F, Bilezikjian LM, Perrin MH, Bosma MM, Windle JJ, Huber KS, Blount AL, Hille B, Vale W, Mellon PL 1991 Intracellular responses to gonadotropin-releasing hormone in a clonal cell line of the gonadotrope lineage. Mol Endocrinol 5:347–355 McArdle CA, Schomerus E, Groener I, Poch A 1992 Estradiol regulates gonadotropin-releasing hormone receptor number, growth and inositol phosphate production in ␣T3–1 cells. Mol Cell Endocrinol 87:95–103 Pinter JH, Janovick JA, Conn PM, Efficiency of differential receptor-effector coupling is GnRHR concentration-dependent. Program of the 80th Annual Meeting of The Endocrine Society, New Orleans, LA, 1998, P3– 425 Windle JJ, Weiner RI, Mellon PL 1990 Cell lines of the pituitary gonadotrope lineage derived by targeted oncogenesis in transgenic mice. Mol Endocrinol 4:597– 603 McArdle CA, Bunting R, Mason WT 1992 Dynamic video imaging of cytosolic Ca2⫹ in the ␣T3–1, gonadotrope-derived cell line. Mol Cell Neurosci 3:124 –132 McGrory WJ, Bautista DS, Graham FL 1988 A simple technique for the rescue of early region I mutations into infectious human adenovirus type 5. Virology 163:614 – 617 Graham FL, Prevec L 1995 Methods for construction of adenovirus vectors. Mol Biotechnol 3:207–220 Geddes BJ, Harding TC, Lightman SL, Uney JB 1997 Long-term gene therapy in the CNS: reversal of hypothalamic diabetes insipidus in the Brattleboro rat by using an adenovirus expressing arginine vasopressin. Nat Med 3:1402–1404 Harding TC, Geddes BJ, Murphy D, Knight D, Uney JB 1998 Switching transgene expression in the brain using an adenoviral tetracycline-regulatable system. Nat Biotechnol 16:553–555 Troskie B, King JA, Millar RP, Peng YY, Kim J, Figueras H, Illing N 1997 Chicken GnRH II-like peptides and a GnRH receptor selective for chicken GnRH II in amphibian sympathetic ganglia. Neuroendocrinology 65:396 – 402 Troskie B, Illing N, Rumbak E, Sun YM, Hapgood J, Sealfon S, Conklin D, Millar RP 1998 Identification of three putative GnRH receptor subtypes in vertebrates. Gen Comp Endocrinol 112:296 –302

GnRH-R DESENSITIZATION AND INTERNALIZATION IN ␣T4 CELLS 46. Davidson JS, Wakefield IK, King JA, Mulligan GP, Millar RP 1988 Dual pathways of calcium entry in spike and plateau phases of luteinizing hormone release from chicken pituitary cells: sequential activation of receptor-operated and voltage-sensitive calcium channels by gonadotropin-releasing hormone. Mol Endocrinol 2:382–390 47. McArdle CA, Gorospe WC, Huckle WR, Conn PM 1987 Homologous downregulation of gonadotropin-releasing hormone receptors and desensitization of gonadotropes: lack of dependence on protein kinase C. Mol Endocrinol 1:420 – 429 48. Fasano S, de Leeuw R, Pieranoni R, Chieffi G, van Oordt PG 1990 Characterization of gonadotropin-releasing hormone (GnRH) binding sites in the pituitary and testis of the frog, Rana esculenta. Biochem Biophys Res Commun 168:923–932 49. Clayton RN, Solano AR, Garcia-Vela A, Dufau ML, Catt KJ 1980 Regulation of pituitary receptors for gonadotropin-releasing hormone during the rat estrous cycle. Endocrinology 107:699 –706 50. Davidson JS, Wakefield IK, King JA, Mulligan GP, Millar RP 1988 Dual pathways of calcium entry in spike and plateau phases of luteinizing hormone release from chicken pituitary cells: sequential activation of receptor-operated

51. 52.

53.

54.

55.

4575

and voltage-sensitive calcium channels by gonadotropin-releasing hormone. Mol Endocrinol 2:382–390 Stojilkovic SS, Catt KJ 1992 Calcium oscillations in anterior pituitary cells. Endocr Rev 13:256 –280 Loumaye E, Catt KJ 1983 Agonist-induced regulation of pituitary receptors for gonadotropin-releasing hormone. Dissociation of receptor recruitment from hormone release in cultured gonadotrophs. J Biol Chem 258:12002–12009 Laws SC, Webster JC, Miller WL 1990 Estradiol alters the effectiveness of gonadotropin-releasing hormone (GnRH) in ovine pituitary cultures: GnRH receptors versus responsiveness to GnRH. Endocrinology 127:381–386 Lin X, Janovick JA, Brother S, Blomenrohr M, Bogerd J, Conn PM 1998 Addition of catfish gonadotropin-releasing hormone (GnRH) receptor intracellular carboxyl-terminal tail to rat GnRH receptor alters receptor expression and regulation. Mol Endocrinol 12:161–171 Davidson J, Flanagan C, Myburgh D, Elario R, Forrest-Owen W, McArdle CA 1996 Incorporation of an additional glycosylation site enhances expression of functional human gonadotropin-releasing hormone receptor. Endocrine 4:207–212