Deletions of portions of the extracellular loops of the lutropin/choriogonadotropin receptor decrease the binding affinity for ovine luteinizing hormone, but not human choriogonadotropin, by preventing the formation of mature cell surface receptor.

The rat lutropin/choriogonadotropin receptor (rLHR) is a G protein-coupled receptor which binds either human choriogonadotropin (hCG) or lutropin (luteinizing hormone, LH) and, therefore, plays a central role in reproductive physiology. In addition to the seven transmembrane helices, three extracellular loops, three intracellular loops, and a cytoplasmic tail characteristic of all G protein-coupled receptors, the rLHR also contains a relatively large N-terminal extracellular domain. Since high affinity hormone binding occurs to this N-terminal extracellular domain and since G proteins are activated by intracellular regions of the receptor, it has been hypothesized that upon hormone binding a portion of the hormone or the receptor's extracellular domain might interact with the receptor's extracellular loops and/or transmembrane helices, thus evoking an intracellular conformational change. To explore this possibility, we prepared and characterized several mutants of the rLHR in which portions of the extracellular loops were deleted. Ultimately, it was not possible to examine the signal transduction properties of the mutants because all but one mutant were retained intracellularly. Although the intracellularly retained mutants must be somewhat misfolded, all were found to bind hCG with high affinity if the cells were first solubilized in detergent. However, the binding of oLH to the detergent solubilized mutants was altered. Thus, whereas the wild-type rLHR bound oLH with two apparent affinities, the solubilized deletion mutants bound oLH with only one apparent affinity. Although these data could be interpreted to suggest that an ovine LH (oLH) binding site on the extracellular loops of the rLHR was deleted, data shown argue against this hypothesis. Rather, the results presented suggest that the two apparent affinities of the wild-type rLHR for oLH represent the binding affinities of two populations of rLHR where the mature, cell surface form binds oLH with a higher affinity than the immature, intracellular form. Furthermore, we show that mutations of the rLHR which cause intracellular retention of the receptor result in a decrease from two to one apparent binding sites for oLH due to the absence of the high affinity oLH binding component contributed by the mature cell surface receptor. Therefore, whereas hCG cannot discriminate between the mature cell surface wild-type receptor and an intracellularly retained rLHR mutant, oLH can make this discrimination, thus suggesting a conformational difference between the two forms of the receptor.

The rat lutropin/choriogonadotropin receptor (rLHR) is a G protein-coupled receptor which binds either human choriogonadotropin (hCG) or lutropin (luteinizing hormone, LH) and, therefore, plays a central role in reproductive physiology. In addition to the seven transmembrane helices, three extracellular loops, three intracellular loops, and a cytoplasmic tail characteristic of all G protein-coupled receptors, the rLHR also contains a relatively large N-terminal extracellular domain. Since high affinity hormone binding occurs to this Nterminal extracellular domain and since G proteins are activated by intracellular regions of the receptor, it has been hypothesized that upon hormone binding a portion of the hormone or the receptor's extracellular domain might interact with the receptor's extracellular loops and/or transmembrane helices, thus evoking an intracellular conformational change. To explore this possibility, we prepared and characterized several mutants of the rLHR in which portions of the extracellular loops were deleted. Ultimately, it was not possible to examine the signal transduction properties of the mutants because all but one mutant were retained intracellularly. Although the intracellularly retained mutants must be somewhat misfolded, all were found to bind hCG with high affinity if the cells were first solubilized in detergent. However, the binding of oLH to the detergent solubilized mutants was altered. Thus, whereas the wild-type rLHR bound oLH with two apparent affinities, the solubilized deletion mutants bound oLH with only one apparent affinity. Although these data could be interpreted to suggest that an ovine LH (oLH) binding site on the extracellular loops of the rLHR was deleted, data shown argue against this hypothesis. Rather, the results presented suggest that the two apparent affinities of the wild-type rLHR for oLH represent the binding affinities of two populations of rLHR where the mature, cell surface form binds oLH with a higher affinity than the immature, intracellular form. Furthermore, we show that mutations of the rLHR which cause intracellular retention of the receptor result in a decrease from two to one apparent binding sites for oLH due to the absence of the high affinity oLH binding component contributed by the mature cell surface receptor. Therefore, whereas hCG cannot discriminate between the mature cell surface wild-type receptor and an intracellularly retained rLHR mutant, oLH can make this discrimination, thus suggesting a conformational difference between the two forms of the receptor.
The rat lutropin/choriogonadotropin receptor (rLHR) 1 is a 674-amino acid single polypeptide chain belonging to the superfamily of G protein-coupled receptors (1). Similar to other members of this superfamily such as the adrenergic (2) and tachykinin receptors (3), the rLHR is thought to consist of seven transmembrane regions connected by alternating intracellular and extracellular loops with an extracellular N-terminal domain and an intracellular C-terminal tail (1,4). The rLHR is also a member of the glycoprotein hormone receptor family, a subfamily of the G protein-coupled receptors that binds large 28 -38-kDa ligands (5). This subfamily consists of the LHR, follitropin receptor (FSHR), and the thyrotropin receptor. Unlike the adrenergic receptors, which have a short N-terminal extension and have been shown to bind their ligands within their transmembrane helices (6,7), the glycoprotein receptors have the unusual feature of a large, approximately 340-amino acid, N-terminal domain which has been demonstrated to bind hormones with high affinity (8 -10).
In addition to the high affinity binding of hormone to the large N-terminal domain of the rLHR, there is some evidence that the extracellular loops of the rLHR may play a role in low affinity interactions with the hormone. Using short synthetic peptides that represent extracellular portions of the rLHR, it was shown that micromolar concentrations of a peptide corresponding the third extracellular loop were able to inhibit the binding of hCG to the rLHR, suggesting that hCG may bind with low affinity to the third extracellular loop (11). Also, rLHR truncation mutants expressing the C-terminal portion of the receptor and either a 10-or 49-amino acid N-terminal extension were able to bind hCG with low affinity and stimulate cAMP production at high concentrations of hCG (12). Both of these studies suggest a potential low affinity interaction of hCG with the extracellular loops of the rLHR.
To further examine whether the extracellular loops of the rLHR play a role in the binding of glycoprotein hormones, we created a series of rLHR extracellular loop deletion mutants, the hypothesis being that if hormone interacted with a loop or a portion thereof, deletion of a portion of that loop would disturb hormone binding. The data presented demonstrate that, while mutations within all three extracellular loops of the rLHR do not affect hCG binding to the rLHR, they reduce the affinity of the rLHR for oLH from two apparent affinities to one apparent affinity. However, we further show that the reduction in oLH binding to one apparent affinity is not due to the deletion of a binding site, but instead due to the absence of the mature population of the rLHR.

MATERIALS AND METHODS
Supplies-Highly purified hCG (CR-125 and CR-127) and oLH (NIADDK-oLH-26) were kindly provided by the National Hormone and Pituitary Agency of the NIDDK (Baltimore, MD). hCG was iodinated as described previously (13). Crude hCG, Nonidet P-40, fibronectin, and most of the other reagents were obtained from Sigma. Tissue culture plastic ware and reagents were obtained from Corning (Corning, NY) and Life Technologies, Inc., respectively. Modular incubator chambers were obtained from Billups-Rothenberg, Inc. (Del Mar, CA). Poly(vinylidene fluoride) membranes and electrophoresis reagents were purchased from Bio-Rad. Enhanced chemiluminescence blotting reagents were obtained from Amersham Corp.
Preparation of rLHR and Mutants-For preparation of the wild-type rLHR, the full coding sequence of the rLHR cDNA (1) was subcloned into pcDNA I/neo (InVitrogen, San Diego, CA). The rLHR extracellular loop deletion mutants were created by polymerase chain reaction using the overlap extension method (14,15) and were subcloned into pcDNA I/neo. The entire region that was synthesized by a polymerase chain reaction was sequenced to verify the removal of the correct base pairs and the integrity of the remaining sequence. rLHR-t616 was constructed and initially characterized as described previously (16).
Cells and Transfection-Human embryonic kidney 293 cells (ATCC CRL 1573; American Type Culture Collection, Rockville, MD) were maintained in a high glucose Dulbecco's modified Eagle's medium supplemented with 50 g/ml gentamicin, 10 mM HEPES, and 10% newborn calf serum (growth medium) and kept in a sterile, humidified, 5% CO 2 environment. Using the calcium phosphate precipitation method (17), cells were transiently transfected with pcDNA I/neo containing either the wild-type rLHR cDNA or one of the extracellular loop deletion mutant cDNAs, or with pCIS containing rLHR-t616 cDNA. Sixteen to twenty hours after transfection, cells were washed twice with warm Waymouth's MB752/1 medium containing 1 mg/ml BSA and 50 g/ml gentamicin and then replaced with growth medium. Twenty-four hours later, cells were used for experiments as described below. Unless otherwise specified, all experiments were performed on transiently transfected 293 cells.
For the creation of stable cell lines, cells were transfected as described above. Sixteen to twenty hours following transfection, cells were washed as described above and then plated with growth medium containing 700 g/ml geneticin. For nonclonal stables such as rLHR-wt and rLHR-E3M, surviving cells were pooled, maintained in growth medium containing 700 g/ml geneticin, and utilized for experiments. Clonal stable cell lines rLHR-wt12 and rLHR-wt16 were kindly donated by Mario Ascoli (University of Iowa).
Western Blotting-High molecular weight standards, soluble extracts from 293 cells expressing the wild-type rLHR or one of the rLHR mutants and solubilized extracts from untransfected 293 cells were applied to a 7.5% polyacrylamide gel. Solubilized extracts (0.25% Nonidet P-40, final concentration) were prepared as described for soluble binding assays. Extracts were loaded at a concentration of 256 g of protein/lane (57.2 g of protein/mm of gel). The Western blots were performed as described previously (18).
For Western blots performed on temperature shifted cells, solubilized extracts (0.5% Nonidet P-40, final concentration) were prepared from 293 cells stably transfected with rLHR-wt16 and rLHR-E3M preincubated at 37 or 26°C and from 293 cells stably expressing the empty pcDNA I/neo vector as described for soluble binding assays. High molecular weight standards and extracts were loaded onto a 7% polyacrylamide gel, where each lane received an equal amount of hCG binding activity (2.6 ng of hCG bound per lane for extracts containing rLHR-wt and 5.1 ng of hCG bound per lane for extracts containing rLHR-E3M).
Determination of the Percentage of Wild-type and Extracellular Loop Mutant rLHRs Expressed on the Cell Surface-Because both intracellular and surface receptors are solubilized by detergent, 125 I-hCG binding to detergent-solubilized extracts is an indication of the total amount of rLHR present. The percentage of receptor expressed at the cell surface was calculated by dividing the amount of 125 I-hCG bound to intact cells by the amount bound to detergent solubilized extracts of these same cells.
To measure binding to intact cells, dishes were placed on ice for 10 min and subsequently washed twice with cold Waymouth's MB752/1 medium lacking bicarbonate, but supplemented with 1 mg/ml BSA and 50 g/ml gentamicin (assay medium). Dishes were incubated overnight at 4°C with a saturating concentration (100 ng/ml, final concentration) of 125 I-hCG in the presence or absence of an excess of unlabeled hCG (50 IU/ml). After 24 h, the reaction was terminated by scraping the cells into Hanks' balanced salt solution supplemented with 1 mg/ml BSA and 50 g/ml gentamicin. Cells were pelleted by centrifugation (1500 ϫ g, 10 min), washed with the same buffer, and counted in a gamma counter. All determinations were performed in duplicate.
For the soluble binding assays, cells were placed on ice for 15 min and washed with cold buffer A (150 mM NaCl and 20 mM HEPES, pH 7.4). The cells were then scraped into cold buffer A containing protease inhibitors (19). Cells were collected by centrifugation at 4°C (1,500 ϫ g, 5 min), and the pellet was resuspended in cold buffer A containing 1% Nonidet P-40 and 20% glycerol. The suspension was incubated for 15 min on ice and subsequently centrifuged (12,000 ϫ g, 15 min, 4°C). The supernatant, which contains the solubilized receptor, was diluted 10fold with cold buffer A containing 20% glycerol and protease inhibitors to bring the detergent concentration to 0.1% Nonidet P-40. A saturating concentration of 125 I-hCG (100 ng/ml, final concentration) was added to aliquots of the detergent extracts in the presence or absence of excess unlabeled hCG (50 IU/ml). The binding assay was terminated by filtration through polyethyleneimine-treated filters (20), and the filters were subsequently counted in a gamma counter. All determinations were performed in duplicate.
Determination of Equilibrium Binding Constants-To determine the equilibrium binding constants for hCG and oLH, competition experiments were performed using detergent solubilized cell extracts in buffer A containing 20% glycerol, 0.1% Nonidet P-40, and protease inhibitors prepared as described above. Aliquots of extracts were incubated overnight at 4°C with a subsaturating concentration (2 ng/ml, final concentration) of 125 I-hCG and increasing concentrations of unlabeled hCG (0 -4.3 g/ml) or unlabeled oLH (0.03-105 g/ml). Assays were terminated by filtration as described above. All determinations were performed in duplicate. The data from each experiment were analyzed for both one site and two site models of ligand binding using the computer program LIGAND (21), and the best statistical fit for the data was determined.
For experiments with GTP␥S, detergent-solubilized extracts (0.1% Nonidet P-40, final concentration) of a nonclonal rLHR-wt expressing cell line were prepared as described above and incubated with a trace concentration of 125 I-hCG and increasing concentrations of unlabeled hCG or oLH as described above in the absence or presence of 100 -200 M GTP␥S. Reactions were incubated at 4°C and terminated 24 h later by filtration as described above. Filters were counted on a gamma counter and determinations were performed in duplicate.
For the determination of hCG and oLH binding affinities to intact cells expressing rLHR-wt, competition experiments were performed. Cells were washed twice with cold assay medium, and then incubated overnight at 4°C with a trace concentration of 125 I-hCG and increasing concentrations of unlabeled hCG or oLH. After 24 h, cells were scraped and collected in Hanks' balanced salt solution supplemented with 1 mg/ml BSA and 50 g/ml gentamicin (wash medium). Cells were centrifuged (1500 ϫ g, 10 min, 4°C), washed with wash medium, vortexed, centrifuged again, and counted in a gamma counter. All determinations were performed in duplicate.
Temperature Shift Experiments-For temperature shift experiments, cells stably expressing either rLHR-wt12 or rLHR-E3M were propagated in a sterile, humidified, 5% CO 2 environment at 37°C. Cells were plated on 60-mm dishes coated with fibronectin and grown for 24 h under the same conditions as for propagation. After 24 h, dishes were transferred to modular incubator chambers filled with a humidified 95% O 2 , 5% CO 2 mixture and then placed in either a 26 or 37°C incubator. Forty-eight hours later, cells were harvested. Total 125 I-hCG binding to intact cells and detergent solubilized extracts thereof were performed to determine cell surface expression as described above. Competition binding assays utilizing detergent solubilized extracts rLHR and Binding Affinity for oLH were performed to determine equilibrium binding constants as described above.

RESULTS
To examine whether the extracellular loops of the rLHR are involved in the binding of gonadotropin hormones, a series of deletion mutants were created in which small portions, three to six amino acids at a time, were removed from regions of the extracellular loops. The deduced amino acid sequence of the rLHR and the specific amino acids deleted in each mutant are shown in Fig. 1. The E in the designations of the mutants is to indicate that these are deletions in the extracellular loops. The number defines whether it is the first, second, or third loop relative to the N terminus, and N, M, and C refer to the N-terminal, middle, and C-terminal portions of each extracellular loop, respectively. For example, rLHR-E1M corresponds to the full-length rLHR in which the middle portion of the first extracellular loop of the rLHR was deleted.
Initially, it was important to determine if the rLHR extracellular loop deletion mutants were expressed. As shown in Fig. 2, Western blotting of solubilized extracts from transiently transfected 293 cells revealed that all nine mutants were stably expressed. The wild-type rLHR appears as a mature 85-kDa protein and an immature 68-kDa precursor protein (18,22). In addition to the specific rLHR bands at 85 and 68 kDa, a prominent band appears between the 85-and 68-kDa bands in the wild-type rLHR and in all the extracellular loop deletion mutants (Fig. 2). However, this band represents a nonspecific interaction of the antibody with an unidentified protein since it was also detected in the untransfected 293 cells (Fig. 2). Further examination of the Western blot of the rLHR extracellular loop deletion mutants reveal that E2N migrates as two proteins with the same molecular masses as the wild-type rLHR. All the other mutants, however, appear only as the smaller 68-kDa protein (Fig. 2).
The following experiments were performed to determine if the receptor mutants could bind hCG and to determine if they were properly localized to the cell surface or not. To address these questions, intact 293 cells transiently expressing the extracellular loop deletion mutants and solubilized extracts thereof were tested for their ability to bind a saturating concentration of 125 I-hCG. The results of these experiments are shown summarized in Table I. Intact cells expressing wildtype, E2N, and E3M receptors bound hCG, but the intact cells expressing the other mutants failed to bind hormone. However, solubilized extracts from cells expressing the wild-type or any of the mutant rLHRs were able to bind hCG. The absolute amounts of hormone bound depended on the efficiency of transfection and varied from experiment to experiment. However, by comparing the amount of 125 I-hCG bound to intact cells with the total binding detected in solubilized extracts within a given experiment, the percentage of rLHR expressed at the cell surface can be calculated. As seen in Table I, approximately 75% of the rLHR-wt and rLHR-E2N are expressed on the cell surface. In contrast, only 20% of rLHR-E3M and 1.5% or less of all the other extracellular loop deletion mutants are expressed on the cell surface. The lack of cell surface expression of most of these extracellular loop mutants does not exclude the possibility that these receptors may bind hormone with high affinity as it has been previously demonstrated that many rLHR mutants that fail to reach the cell surface can still bind hCG with high affinity (5,(23)(24)(25).
The next set of experiments tested the crucial question concerning the potential effects of the deletion of small portions of the rLHR extracellular loops on the affinity of the receptor for hormone. Previously, it was shown that hCG binds to the rLHR with picomolar affinity (1). It was also demonstrated that the extracellular domain alone binds hCG with an affinity comparable to that of the full-length rLHR (8). Because of these prior results, the possibility exists that even if hCG interacted with the extracellular loops with low affinity, a decrease in the affinity for hCG in the solubilized extracts of cells expressing the extracellular loop deletion mutants might not be detected because of the high affinity interaction of hCG with the Nterminal 340 amino acids of the rLHR. Consequently, we chose to measure the affinity of the receptor mutants for oLH, as well as for hCG, because oLH has a much lower affinity for the rLHR (25).
To measure the binding affinities of the wild-type rLHR and deletion mutants for hCG and oLH, competition assays were performed using solubilized extracts from transiently transfected 293 cells using unlabeled hCG or oLH to displace 125 I-hCG. Consistent with previous reports, computer analyses of the data utilizing the LIGAND program (21) indicated that a single class of high affinity (360 pM) hCG binding sites was present in extracts from cells expressing the wild-type rLHR (Table II and Fig. 3) (23,25). The results for the binding of hCG to all the mutants are compiled in Table II. Significantly, the deletion of any portion of any of the three extracellular loops had no effect on the affinity of the receptor mutants for hCG. When competition assays were performed with unlabeled oLH as the displacing ligand, analyses of data from extracts of cells expressing rLHR-wt revealed two apparent classes of oLH binding sites, one of 7.1 nM and a lower affinity site of 64 nM (Table III). Fig. 4, Panels A and B, shows the binding of oLH to the wild-type rLHR, where the curvilinear nature of the Scatchard plot is readily apparent. oLH binding affinities for the wild-type rLHR and all rLHR extracellular loop deletion mutants are summarized in Table III. Solubilized extracts of cells expressing rLHR-E2N bound oLH with two apparent affinities similar to the wild-type rLHR. In contrast to the rLHR-wt and rLHR-E2N, however, deletion of any of the other portions of any of the extracellular loops resulted in oLH binding with only a single apparent binding affinity of 33-46 nM (see Table III). As an example, Fig. 4, Panels C and D, illustrates the binding of oLH to rLHR-E1M. A comparison of Panels B and D of Fig.   4 clearly shows a shift in oLH binding from two apparent affinities for rLHR-wt to a single affinity for rLHR-E1M.
Thus far, the shift from two apparent affinities to one apparent affinity observed for all the extracellular loop deletion mutants, with the exception of rLHR-E2N, could be explained by the deletion of a binding site for oLH on the extracellular loops. However, reevaluation of the Western blot in Fig. 2, cell surface expression data in Table I, and the oLH binding affinity data in Table III shows that rLHR-E2N is the only extracellular deletion mutant that is expressed as both the mature 85-kDa and the immature 68-kDa forms of the receptor, that is expressed at the cell surface at levels comparable to rLHR-wt, and that binds oLH with two apparent affinities. These observations suggest an alternative hypothesis to explain the shift from two apparent affinities to one affinity observed for all the extracellular loop deletion mutants other than rLHR-E2N. Thus, the introduction of mutations into the extracellular loops of the rLHR may result in the improper folding and retention of the mutant receptors intracellularly. This retention may result in a lack of processing of the mutants to the mature 85-kDa cell surface receptor which may be reflected by the absence of high affinity oLH binding. To examine this further, the following experiments were performed.
If the loss of high affinity oLH binding to the intracellularly retained rLHR mutants is due to the absence of the fully processed cell surface rLHR, one would predict that expression of the extracellular loop deletion mutants at the cell surface should restore high affinity oLH binding. Incubation of some mislocalized proteins such as the ⌬F508 CFTR mutant (26,27), some vesicular stomatitis virus G proteins with mutations in their glycosylation sites (28,29), and some rLHR mutants 2 at reduced temperatures restored proper localization of these mutant proteins to the cell surface. Therefore, 125 I-hCG binding assays were performed on nonclonal 293 cells stably transfected with the different intracellularly trapped rLHR extracellular loop deletion mutants after a 48-h preincubation at 26 versus 37°C. With the exception of rLHR-E3M, little or no cell surface binding was detected for all of the intracellularly trapped extracellular loop deletion mutants incubated at either temperature (data not shown), indicating that these mutants continued to be mislocalized intracellularly even when incubated at reduced temperature. However, rLHR-E3M, which was expressed at Ͻ20% at the cell surface at 37°C, was expressed at the cell surface with levels similar to the wild-type  rLHR and Binding Affinity for oLH receptor when preincubated at 26°C (Table IV). As shown in Fig. 5, Western blots of detergent solubilized extracts from rLHR-E3M expressing cells preincubated at 37°C exhibited only the immature 68-kDa form of the receptor. However, extracts from cells preincubated at 26°C clearly showed both the 85-and 68-kDa foms of the rLHR, similar to wild-type rLHR. Equilibrium binding assays were subsequently performed with detergent-solubilized extracts prepared from a clonal cell line stably expressing the wild-type rLHR or from a nonclonal 293 cell line stably transfected with rLHR-E3M preincubated at 37 versus 26°C. Similar to experiments with extracts from transiently transfected cells, one class of high affinity binding sites was determined for the binding of hCG to extracts from rLHR-wt and rLHR-E3M expressing cells preincubated at either 37 or 26°C (data not shown). A model of two apparent affinities was determined for the binding of oLH to rLHR-wt regardless of whether the cells were preincubated at 37 or 26°C (Table IV). As shown in Table IV and Fig. 6, Panels A and B, a model of one apparent affinity for oLH binding to rLHR-E3M was determined for cells preincubated at 37°C, similar to that observed for equilibrium binding assays performed with extracts from cells transiently expressing rLHR-E3M. Importantly, as shown in Table IV and Fig. 6, Panels C and D, detergent solubilized extracts from rLHR-E3M-expressing cells preincubated at 26°C now bound oLH with two apparent affinities similar to rLHR-wt. A comparison of the Scatchard plots of oLH binding to rLHR-E3M from cells preincubated at 37 versus 26°C (Fig. 6, Panel B versus Panel D) clearly demonstrates the shift from one apparent affinity for rLHR-E3M expressing cells preincubated at 37°C to two apparent affinities for those incubated at 26°C. This correlation between the appearance of cell surface hCG binding and the mature 85-kDa receptor when rLHR-E3M expressing cells are preincubated at 26°C with the shift from one apparent oLH binding affinity at 37°C to two apparent affinities at 26°C suggests that the detection of only one apparent oLH binding affinity for most of the extracellular loop deletion mutants is due to the absence of the mature population of rLHR.
To further examine the hypothesis that the absence of high affinity oLH binding for the intracellularly retained mutants is due to the absence of the mature, cell surface rLHR, a search was undertaken to identify other rLHR mutants that were mislocalized intracellularly similar to the extracellular loop deletion mutants, but that did not involve the alteration of a putative hormone binding site. One such mutant, rLHR-t616, is a truncation mutant of the cytoplasmic tail of the rLHR after amino acid 616. One would predict that a truncation of the cytoplasmic tail should not alter a hormone binding site. 125 I-hCG binding assays to intact cells transiently expressing rLHR-t616 and detergent-solubilized extracts thereof revealed that Ͻ2% of rLHR-t616 was expressed at the cell surface, suggesting the retention of this mutant intracellularly (Table V). Similar to rLHR-wt and the extracellular loop deletion mutants, a model of a single class of high affinity (337 pM) binding sites was determined for the binding of hCG to extracts from rLHR-t616-expressing cells. However, extracts of rLHR-t616-expressing cells bound oLH with only one apparent affinity similar to that observed for most of the intracellularly retained extracellular loop deletion mutants (Table V and Fig. 7). Therefore, these experiments with rLHR-t616 also suggest that the absence of high affinity oLH binding is due to the absence of the mature, cell surface rLHR.
Thus far, the data suggest that a mature, cell surface rLHR is required for high affinity oLH binding. The question remains, though, whether the two apparent affinities for oLH represent two binding sites for oLH present on the mature, cell FIG. 3. Binding of hCG to the wild-type rLHR. Detergent-solubilized extracts from 293 cells transiently expressing the wild-type rLHR were incubated with a trace concentration of 125 I-hCG and increasing concentrations of unlabeled hCG as described under "Materials and Methods." The results are representative of 11 independent experiments. Each point represents the mean of duplicate determinations. The computer program LIGAND (21) was used to analyze the data. The program determined that a single class of binding sites was the best statistical fit through the data points (p Ͻ 0.05). surface rLHR, or whether the two apparent affinities represent the binding affinities for two populations of rLHR, where the 85-kDa cell surface population of rLHR binds oLH with a high affinity, and the 68-kDa intracellular population of rLHR binds oLH with a lower affinity. To attempt to discern if the mature, cell surface rLHR binds oLH with two apparent binding affinities or with single high binding affinity, equilibrium binding assays were performed on intact rLHR-wt-expressing cells with the rationale that only the mature, fully processed rLHR should be at the cell surface and be able to bind hormone. As shown in Table VI, intact cells expressing rLHR-wt bound oLH with two apparent affinities. However, a comparison of the distribution of high and low affinity sites observed with intact cells expressing rLHR-wt as compared to solubilized extracts thereof revealed distinct differences in the percentage of total receptor with a high or low affinity (Table VI). As measured in intact cells, approximately 90% of oLH binding to rLHR-wt was of high affinity while Ͻ10% was of low affinity. In contrast, when assayed in detergent solubilized extracts, only 60% binding to rLHR-wt was of high affinity, while 40% was of low affinity (Table VI). These data suggest that the cell surface rLHR is predominantly in the high affinity form and that the cell surface low affinity binding component is most likely contributed to by the presence of a small percentage of incompletely folded rLHR on the cell surface. Taken altogether, these data clearly indicate that the shift from two apparent oLH binding affinities for rLHR-wt to one apparent affinity with most of the extracellular loop deletion FIG. 4. Binding of oLH to rLHR-wt and rLHR-E1M. Detergent solubilized extracts from 293 cells transiently transfected with the cDNA for rLHR-wt or rLHR-E1M were incubated with a trace concentration of 125 I-hCG and increasing concentrations of unlabeled oLH as described under "Materials and Methods." The computer program LIGAND (21) was used to analyze the data to determine the best fit through the data points. A and B are a representative displacement curve and Scatchard plot, respectively, for rLHR-wt from 11 independent experiments. For the binding of oLH to rLHR-wt, a model of two classes of binding sites was found to be the best statistical fit for the data points (p Ͻ 0.05). C and D are a representative displacement curve and Scatchard plot, respectively, for rLHR-E1M from four independent experiments. For the binding of oLH to rLHR-E1M, a model of a single class of binding sites was determined to be the best statistical fit (p Ͻ 0.05)

TABLE IV
Affinities of temperature shifted rLHR-wt and rLHR-E3M for oLH Data shown are the mean Ϯ range or standard error of the mean for the indicated number of experiments. 125 I-hCG bindings to intact cells and detergent solubilized extracts thereof were performed as described under "Materials and Methods." The percentage of receptor expressed on the cell surface was calculated by dividing the amount of 125 I-hCG binding to intact cells by the amount bound to solubilized extracts. Equilibrium binding assays where a trace concentration of 125 I-hCG was incubated with increasing concentrations of unlabeled oLH were performed with detergent-solubilized extracts from a clonal stable 293 cell line expressing rLHR-wt or with extracts from a nonclonal stable 293 cell line expressing rLHR-E3M as described under "Materials and Methods." mutants is not due to the deletion of an oLH binding site on the rLHR. Instead, deletion of portions of the extracellular loops and even part of the cytoplasmic tail most likely perturb the normal folding of the rLHR resulting in the retention of the mutant receptors intracellularly and the absence of the mature cell surface receptor. These data also suggest that the observation of two binding affinities for the binding of oLH to rLHR-wt is not due to the presence of two binding sites on the mature rLHR for oLH, but is instead most likely due to the presence of two populations of rLHR with different binding affinities for oLH where the mature, 85-kDa cell surface form of the receptor binds oLH with a higher affinity than the immature, 68-kDa intracellular form of the receptor.

DISCUSSION
The present study was undertaken to identify potential low affinity binding sites for hormone on the extracellular loops of the rLHR. There is much evidence for the involvement of the extracellular loops of other G protein-coupled receptors in ligand binding. For example, experiments in which portions of the neurokinin-1 receptor were substituted with the corresponding sequences of the neurokinin-2 and -3 receptors revealed that the first and second extracellular loops of the neurokinin-1 receptor, in addition to a portion of the N-terminal tail and transmembrane 3, bind substance P (30, 31) In addition, Glu-301 of the third extracellular loop of the mammalian gonadotropin-releasing hormone receptor has recently been shown to be involved in an electrostatic interaction with Arg-8 of gonadotropin-releasing hormone (32). The second and third extracellular loops of the interleukin-8 type A receptor have also been demonstrated to be part of a ligand binding site (33,34).
Although it has been shown that hormone binds with high affinity to the N-terminal extracellular domain of the rLHR (8,(35)(36)(37)(38), low affinity interactions of hormone with the extracellular loops and/or transmembrane helices of the rLHR have FIG. 5. Western blot of extracts from rLHR-wt and rLHR-E3Mexpressing cells incubated at 37 or 26°C. Detergent-solubilized extracts prepared from rLHR-wt or rLHR-E3M expressing cells incubated at 37 or 26°C were run under nonreducing conditions on 7% SDS-gels and transferred to poly(vinylidene fluoride) membranes. The membranes were probed as described under "Materials and Methods" with the polyclonal antibody Bugs prepared to the purified rLHR. Only the relevant portion of the gel is shown.
FIG. 6. Binding of oLH to extracts from rLHR-E3M-expressing cells incubated at 37°C or 26°C. Nonclonal stably transfected 293 cells expressing rLHR-E3M were incubated for 48 h at either 37 or 26°C as described under "Materials and Methods." Detergent-solubilized extracts were prepared from these cells and incubated with a trace concentration of 125 I-hCG and increasing concentrations of unlabeled oLH. A representative experiment from two such experiments is shown and all points represent the mean of duplicate determinations. The computer program LIGAND (21) was utilized to determine the best fit through the data points. A and B are a representative displacement curve and Scatchard plot, respectively, from rLHR-E3M expressing cells grown at 37°C. A single class of binding sites was determined to be the best statistical fit for rLHR-E3M incubated at 37°C (p Ͻ 0.05). C and D are from the same cell line incubated at 26°C. A model of two classes of binding sites was determined to be the best statistical fit for rLHR-E3M incubated at 26°C (p Ͻ 0.05). also been suggested (11,12,18,25). For example, a low affinity interaction of hCG with the third extracellular loop of the rLHR was suggested by the finding that micromolar concen-trations of a synthetic peptide corresponding to the third extracellular loop of the rLHR was able to inhibit hCG binding to the full-length rLHR (11). In addition, it was reported that a rLHR truncation mutant expressing the C-terminal portion of the receptor and either a 10-or 49-amino acid N-terminal extension was able to bind hCG with micromolar affinity and to stimulate low levels of cAMP production (12). However, without data to demonstrate that this truncation mutant was properly located to the plasma membrane, it is difficult to distinguish such a low binding affinity for hCG and low cAMP stimulation from nonspecific effects. Studies by Ascoli and coworkers found that a substitution of Asp-383 for Asn in the second transmembrane helix of the rLHR produced a receptor that had a decreased affinity for oLH in the absence of sodium (25), suggesting a binding site for oLH on the C-terminal half of the rLHR. Lastly, our laboratory has shown that, unlike the full-length rLHR which binds oLH with two apparent affinities, the rLHR extracellular domain alone binds oLH with a single affinity (18).
In the present studies, we examined the effects of deletions of portions of the extracellular loops of the rLHR on hCG and oLH binding affinities. Western blotting revealed that all of the mutants were expressed as stable proteins. However, unlike rLHR-wt, which is expressed as both the mature 85-kDa and immature 68-kDa forms, all the mutants, with the exception of rLHR-E2N, were expressed as only the 68-kDa precursor form of the rLHR. In addition, all the extracellular loop deletion mutants, with the exception of rLHR-E2N, were severely or completely retained intracellularly. Binding assays to detergent solubilized extracts of cells showed that although the affinities for hCG were unaffected by deletions of any portion of any of the extracellular loops, deletions of portions of all the extracellular loops with the exception or rLHR-E2N resulted in altered oLH binding. A shift from two apparent binding affinities of oLH for rLHR-wt to a single binding affinity for oLH was observed for all the mutants with the exception of rLHR-E2N. The results of experiments presented herein, therefore, at first suggested that the extracellular loops of the rLHR were contributing to the binding of oLH. However, further experiments confirmed the correlation between the presence of two binding affinities for oLH and the expression of the cell surface 85-kDa form of the receptor. Taken altogether, these data suggest that the shift from two apparent oLH binding affinities for rLHR-wt to a single binding affinity for the most of the mutants is not due to the loss of an oLH binding site on the rLHR, but is due to the absence of the fully folded, cell surface form of the receptor. We hypothesize that the mutations are causing a misfolding of the receptor that causes them to be retained intracellularly, thus preventing the further folding and processing of the mutant receptor into the mature cell surface form. Our data also show that the observation of two oLH binding affinities measured for rLHR-wt does not represent two binding sites on an individual receptor. Rather, the two oLH binding affinities for rLHR-wt represent the binding affinities of oLH for two populations of rLHR. The 85-kDa mature cell surface form of the rLHR binds oLH with a higher affinity, while the 68-kDa intracellular precursor form binds oLH with a lower affinity.
For certain other G protein-coupled receptors, in particular the adrenergic receptors, the affinity of agonist for the receptor can be affected by the receptor's association with a G protein (39). This is manifest by the ability of exogenously added GTP to decrease the binding affinity of agonist for the receptor (39). The possibility that perhaps the two apparent binding affinities of the rLHR for oLH were arising from solubilization of one pool of rLHR associated with G proteins and one pool not FIG. 7. Binding of oLH to the truncation mutant rLHR-t616. Detergent-solubilized extracts from 293 cells transiently expressing rLHR-t616 were incubated with a trace concentration of 125 I-hCG and increasing concentrations of unlabeled oLH as described under "Materials and Methods." A representative experiment from three independent experiments is shown. Each point represents the mean of duplicate determinations. The computer program LIGAND (21) was utilized to determine the best statistical fit for the data points. A model of a single class of binding sites was determined to be the best statistical fit for the data (p Ͻ 0.05). rLHR and Binding Affinity for oLH associated with G proteins was, therefore, also considered. It is generally not possible to observe a change in hCG binding affinity to membranes containing the rLHR upon addition of GTP (40 -44). This is thought to be due to the high affinity interactions of hCG with the extracellular domain of the rLHR (8), a situation quite unlike the binding of catecholamines to the transmembrane helices of adrenergic receptors (45). Thus, whereas it can be envisioned how an intracellular conformational change in an adrenergic receptor accompanying G protein activation could lead to an increased dissociation of catecholamine from their binding site, a similar conformational change in the rLHR would not be as likely to result in a dissociation of hCG from the extracellular binding domain. The possibility remained, however, that the binding affinity of oLH, which is much lower than that of hCG for the rLHR, might be affected by G protein activation, especially if there were additional contact sites of oLH on the carboxyl half of the receptor. If so, then this might also account for the two populations of oLH binding affinities observed in detergent extracts of cells expressing the wild-type rLHR. To test this possibility, we examined the effects of 200-300 M GTP␥S on oLH binding to detergent soluble extracts of cells expressing the wild-type rLHR. In three separate experiments, however, the same results were observed regardless of the presence or absence of GTP␥S (data not shown). Therefore, it is highly unlikely that the association or lack of association of G proteins with the rLHR contribute to alterations in oLH binding to this receptor. As such, the higher affinity binding of oLH to the mature cell surface receptor as compared to the intracellular precursor form of the receptor does not appear to be due to differences in the associations of these receptor pools with G proteins. Previously, our laboratory reported that rLHR-t338, a rLHR truncation mutant at amino acid 338 that expresses only the N-terminal extracellular domain of the rLHR, was retained intracellularly and bound oLH with a single apparent affinity (18). Based upon these observations, we proposed that the loss of high affinity oLH binding was due to the loss of an oLH binding site on the C-terminal half of the rLHR (18). However, the evidence presented herein suggests that the absence of high affinity oLH binding to rLHR-t338 is most likely due to the misfolding of this mutant resulting in its retention intracellularly and lack of folding and processing to a mature form.
The data presented herein clearly illustrate important differences in the binding of hCG and oLH to the rLHR. Unlike oLH, hCG is not capable of distinguishing between the immature and mature forms of the rLHR. Consistent with reports on other intracellularly retained rLHR mutants (5,23,24,46), hCG binds to the intracellularly trapped extracellular loop deletion mutants with the same high affinity as it binds to the cell surface form of the rLHR as long as a hormone binding site has not been disturbed. Therefore, although each of these mutations is causing a degree of misfolding resulting in intracellular retention of the mutant protein, it cannot be a gross perturbation of structure since hCG can still bind with normal high affinity. Because hCG binds with equally high affinities to the mature and immature forms of the rLHR, one might conclude that the immature form of the rLHR was folded in the same conformation as the mature cell surface form. However, our observations that oLH binds with lower affinity to the immature receptor suggest otherwise and argue that although the immature receptor is folded completely enough to bind hCG, it must not be in exactly the same conformation as the mature cell surface receptor.
An interesting comparison can be made between mutants of the rLHR and those of the rFSHR, a receptor which shares a high degree of amino acid identity with the rLHR (47). In this respect, our laboratory has recently demonstrated that mutations of the rFSHR which cause intracellular retention of the mutant rFSHR in the endoplasmic reticulum result in a lack of detectable FSH binding even if a hormone binding site has not been altered (48). For example, rFSHR-t637 is a truncation mutation of the cytoplasmic tail of the rFSHR at amino acid 637 that, similar to rLHR-t616, should theoretically not perturb a hormone binding site. Similar to rLHR-t616, rFSHR-t637 is retained intracellularly, as evidenced by its sensitivity to endoglycosidase H. Detergent-solubilized extracts of cells expressing rFSH-t637, however, have no detectable FSH binding activity. Thus, whereas intracellularly retained mutants of the rLHR show comparable high affinity hCG binding and reduced oLH binding affinity as compared to the cell surface mature rLHR, intracellularly retained mutants of the rFSHR lack any detectable binding activity. Assuming that mutants of the rLHR and rFSHR are retained in the same intracellular compartment (i.e. the endoplasmic reticulum), these observations suggest that the newly synthesized rLHR in the endoplasmic reticulum is already folded into a conformation that can bind hCG with high affinity. It must still be in a conformation distinct from the cell surface mature receptor, though, since it does not yet have the same high affinity oLH binding as the mature rLHR. In contrast, newly synthesized rFSHR in the endoplasmic reticululum does not appear to be folded appropriately enough to allow for any hormone binding. Taken altogether, the data on intracellularly retained mutants of the rLHR and rFSHR suggest that the rLHR folds more readily than the highly related rFSHR.
Clearly more studies will be required in order to better understand the complex folding and maturation of the gonadotropin receptors, as well as other G protein-coupled receptors. Toward this end, the differential binding of oLH to mature versus immature rLHR should prove useful in elucidating the folding of the rLHR.

TABLE VI
Distribution of high and low affinity oLH binding sites Data shown are the mean Ϯ standard error of the mean for the indicated number of experiments. Equilibrium binding assays where a trace concentration of 125 I-hCG was incubated with increasing concentrations of unlabeled oLH were performed on intact cells or detergent-solubilized extracts (as indicated) from transiently transfected 293 cells as described under "Materials and Methods." Affinities and the percentage of total receptor with a given affinity were calculated using the computer program LIGAND (21).