Identification of two amino acid residues on the extracellular domain of the lutropin/choriogonadotropin receptor important in signaling.

The lutropin/choriogonadotropin receptor (LH/CG-R) is a G protein-coupled receptor with a relatively large extracellular domain. The cDNAs of LH/CG-R wild type and 15 point and double mutants, which encoded residues of opposite charge to that of wild type, were transiently transfected into COS-7 cells. Human choriogonadotropin (hCG) binding was determined, as was hCG-mediated cAMP production. Most of the replacements resulted in no substantive effect on the binding affinity of hCG to LH/CG-R or on hCG-stimulated cAMP production, although the mutants expressed at a lower level than LH/CG-R wild type. The most interesting observation was noted with two point mutants of LH/CG-R, Glu332-->Lys and Asp333-->Lys, which bound hCG but failed to give increased cAMP production. Several of the mutant forms of LH/CG-R that expressed at low levels were further analyzed by soluble binding assays and Western blots. There was no evidence of any significant degree of intracellular trapping of hCG-binding mutant receptors. The expected major (93 kDa) and minor (78 kDa) forms were found for LH/CG-R wild type and several of the mutants. The Lys235-->Asp and Asp333-->Lys mutants exhibited primarily the lower M(r) form, indicating that receptor processing was impaired or that the mutant higher M(r) form was more rapidly degraded than LH/CG-R wild type. These results demonstrate that Glu332 and Asp333, which are located near the first transmembrane helix, are important in receptor activation, while other conserved ionizable residues of LH/CG-R appear important in cell surface expression or stability but not in binding or signaling.

Considerable data have been obtained to delineate specific amino acid residues and regions on the hCG ␣ and ␤ subunits that participate in receptor binding and signal transduction. The recent availability of a crystal structure of partially deglycosylated hCG (Lapthorn et al., 1994;Wu et al., 1994) adds a new dimension to the analysis of glycoprotein hormone structure-function relationships. In contrast, relatively little is known about the conformation of the receptor and specific amino acid residues that may be involved in hormone binding and signaling, although mutagenesis has been used to map certain of these regions of LH/CG-R (Xie et al., 1990;Ji, 1991, 1993;Segaloff and Ascoli, 1993;Quintana et al., 1993;Wang et al., 1993aWang et al., , 1993b. A perusal of the available amino acid sequences of LH/CG-R from several species shows the presence of a number of conserved ionizable amino acid residues, many of which are also conserved in the follitropin receptor and the thyrotropin receptor. Such conserved residues may be important in signaling or as hormone contact sites, e.g. to oppositely charged conserved side chains on LH, CG, follitropin, and thyrotropin. The present study addresses 11 of the ionizable groups on the extracellular portion of rat LH/CG-R: Arg 21 and Arg 31 (encoded by exon 1), Lys 55 (encoded by exon 2), Glu 65 and Glu 68 (encoded by exon 3), Arg 114 and Lys 121 (encoded by exon 5), Glu 221 and Lys 235 (encoded by exon 9), and Glu 332 and Asp 333 (encoded by exon 11). Site-directed mutagenesis was used to prepare single and double mutants where each ionizable side chain was replaced by one of opposite charge. The data suggest that two point mutants of LH/CG-R, Lys 332 and Lys 333 , bind hCG with nearly the same affinity as LH/CG-R wild type but exhibit diminished signaling.

Supplies-The
Transformer TM site-directed mutagenesis kit was obtained from Clontech Laboratories, Inc, the Sequanase version 2.0 kit was a product of U. S. Biochemical Corp., the plasmid kit was from Qiagen Inc. (Chatsworth, CA), and the Magic TM Minipreps DNA Purification System was from Promega. Bovine serum albumin, isobutylmethylxanthine, DEAE-dextran, chloroquine, dextran sulfate, dimethyl sulfoxide, and Nonidet P-40 were purchased from Sigma. hCG (CR-121) was provided by National Institutes of Health and by Dr. Steven Birken (Columbia University, NY); [ 125 I]hCG (100 -150 Ci/g) was from ICN (Horsham, PA). The cAMP radioimmunoassay kit was from DuPont NEN. Dulbecco's modified Eagle's medium and fetal bovine serum were obtained from Life Technologies, Inc.; NuSerum was from Collaborative Biomedical Products (Bedford, MA).
Mutagenesis of the LH/CG-R cDNA-The cDNA for the rat LH/CG receptor was kindly provided by Dr. William Moyle (Piscataway, NJ) in the expression vector pSVL. The desired 21-base deoxyoligonucleotides were synthesized by Dr. Rudolf Werner (University of Miami, FL) on an Applied Biosystems model 380B DNA synthesizer and by the Molecular Genetics Core Facility of the University of Georgia. In vitro mutagenesis was performed using the Transformer TM site-directed mutagenesis kit as described (Deng and Nickoloff, 1992), and mutated plasmid clones were identified by double-stranded sequencing.
Expression of LH/CG-R-COS-7 cells (African green monkey kidney cells), obtained from ATCC CRL, were grown at 37°C in humidified air containing 5% CO 2 in the following medium: Dulbecco's modified Eagle's medium (90%), fetal bovine serum (10%), and 100 units/ml each of penicillin and streptomycin. Cells were transiently transfected at 60 -80% confluency by the DEAE-dextran method. Transfection efficiencies were estimated by cotransfecting a pSV-␤-galactosidase plasmid with the pSVL vectors containing the LH/CG-R wild type and mutant cDNAs. 48 h after transfection, the cells were stained in situ for ␤-galactosidase. This was accomplished by first washing the cells with PBS five times, then fixing them in a 0.05% (v/v) glutaraldehyde solution in PBS for 10 -15 min. After three additional washes with PBS, the cells were incubated in a solution containing 1 mg/ml 5-bromo-4-chloro-3-indoyl ␤-D-galactoside 1 mM MgCl 2 , 20 mM potassium ferricyanide/ ferrocyanide in PBS for at least 1 h at 37°C. The cells transfected with the pSV-␤-galactosidase control plasmid were readily visualized under the microscope and counted. From the total number of cells in the field, transfection efficiencies were obtained.
[ 125 I]hCG Binding to COS-7 Cells-At 48 h post-transfection, the cells were assayed for [ 125 I]hCG binding, along with untransfected cells as control. Cells (1 ϫ 10 6 cells/well) were washed twice with serum-free Waymouth's medium containing 1 mg of bovine serum albumin/ml, then 1 ml of this medium was added to each well. For competitive binding assays, [ 125 I]hCG (10 5 cpm) and different concentrations of unlabeled hCG were added to each well, while for direct binding assays, cells were incubated with increasing concentrations of [ 125 I]hCG. Nonspecific binding was determined by inclusion of a 1000-fold excess of unlabeled hormone. The cells were incubated overnight with gentle shaking at room temperature. On the following day, cells were washed twice with PBS, then trypsinized, collected, and counted in a ␥ counter. Specific binding of [ 125 I]hCG in the absence of unlabeled hormone gave 4,000 -5,000 cpm, and nonspecific binding was about 5%. The results were analyzed by In-Plot (Graphpad Software, San Diego, CA), by Ligand Program (Munson and Rodbard, 1980), and by graphical analysis, the agreement between the various methods being good. Unless indicated otherwise, the data from a typical experiment are shown as mean Ϯ S.E.
[ 125 I]hCG Binding to Detergent-solubilized COS-7 Cell Extracts-This procedure is based on that described by others (Xie et al., 1990;Thomas and Segaloff, 1994). Control cells and transfected cells (48 h) were placed on ice and washed twice with cold 0.15 M NaCl, 20 mM Hepes, pH 7.4. They were then scraped into a small volume of the same buffer containing 1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, and 5 mM N-ethylmaleimide. The cells were centrifuged (1500 ϫ g, 15 min, 4°C) and lysed by vortexing in a solution containing the same buffer with 1% Nonidet P-40 and 20% glycerol. After incubation on ice for 15 min, the cell lysate was centrifuged (12,000 ϫ g, 15 min, 4°C), and the supernatant was diluted to 0.1% Nonidet P-40 and incubated overnight with [ 125 I]hCG (2-3 ϫ 10 5 cpm) in the presence and absence of a 1000-fold excess of hCG. Then, the binding solution was added to Whatman GF/B filters that had been soaked in 0.3% polyethylenimine (pH 9.1) for at least 1 h (Roche et al., 1985). After filtration and washing five times, the filters were counted in a ␥ counter. The specific binding of [ 125 I]hCG to solubilized LH/CG-R was between 5 and 15 ϫ 10 3 cpm, and nonspecific binding was about 25% of the total binding.
cAMP Assay-After transfection (48 h), the COS-7 cells were washed twice with Dulbecco's modified Eagle's medium containing 1 mg of bovine serum albumin/ml and incubated in 0.5 ml of the medium containing isobutylmethylxanthine (0.8 mM) at 37°C for 15 min. Increasing concentrations of hCG were then added, and the incubation was continued for 30 min at 37°C. After removing the medium, the cells were lysed in 100% ethanol at Ϫ20°C overnight. The extract was collected, dried under flowing nitrogen gas, and resuspended in the buffer of the cAMP assay kit. cAMP concentrations were determined with the [ 125 I]cAMP radioimmunoassay kit. Unless indicated otherwise, the results of a typical experiment are given as mean Ϯ S.E.
An anti-LH/CG-R rabbit antiserum (Rosemblit et al., 1988), kindly provided by Dr. Deborah Segaloff (University of Iowa,), was diluted 4-fold with cold PBS (pH 8.0) and absorbed to a slurry of protein A-Sepharose in a small column at 4°C. The column was then washed extensively with PBS, and the IgG fraction was eluted by addition of a 50 mM glycine buffer, pH 3.0, containing 0.1 M NaCl, and collected in tubes of 1 ml each containing 75 l of 0.2 M Tris-HCl and 23.3 l of 0.85 M Hepes buffer. The most concentrated fractions were pooled and stored at Ϫ70°C. Western blots were based on the technique of Hipkin et al. (1992) with slight modifications. Prestained M r standards (ovalbumin, bovine serum albumin, ␤-galactosidase, and myosin from Bio-Rad) and equal amounts of cell lysates in sample buffer, with no boiling and in the absence of reducing agent, were applied to 7% SDS-polyacrylamide gels. An equivalent amount of lysate from untransfected COS-7 cells was used as a negative control. The proteins were electrophoretically transferred to polyvinylidine difluoride membranes, which were then washed twice (5 min each) with 20 mM Tris-HCl buffer, pH 7.5, containing 0.5 M NaCl, and blocked for 2 h at room temperature with PBS containing 5% instant nonfat dry milk, 0.2% Tween 20, and 10% glycerol. Next, the filters were incubated overnight at room temperature with the same blocking solution containing 3 g/ml rabbit anti-LH/CG-R IgG. The blot was then washed five times (5 min each) with blocking solution, followed by incubation for 1 h at room temperature with a 1:5000-fold dilution of a horseradish peroxidase-labeled donkey anti-rabbit IgG whole antibody fraction in the blocking solution. The second antibody was replaced, and the filters were washed twice (5 min each) with 20 mM Tris-HCl, pH 7.5, 0.5 M NaCl; this was followed by two washes (5 min first wash, 10 min second wash) with the same buffer containing 1% Nonidet P-40. The filters were then washed for 5 min with the 20 mM Tris-HCl, 0.5 M NaCl (pH 7.5) solution, for 10 min with the same buffer containing 1% Nonidet P-40, and finally for 5 min with the buffer solution only. The blots were treated for 1 min with Amersham ECL developer solutions 1 and 2 (combined in equal amounts), then wrapped in Saran wrap and exposed to autoradiography film for about 30 s, followed by immediate development. The apparent M r and density of the various bands were obtained after scanning densitometry (PDI, Huntington Station, NY).

RESULTS
The DNA sequences of the rat LH/CG-R wild type and mutants were determined for the mutated and adjacent regions in all clones by dideoxynucleotide sequencing. Following this confirmation of the desired wild type and mutant clones, transfection efficiencies were monitored by transient cotransfection of pSVL-LH/CG-R wild type and mutants into COS-7 cells with pSVL-␤-galactosidase. The cotransfected cells were stained in situ, and transfection efficiencies of 5-10% were estimated. hCG binding to transfected cells was determined via competition of [ 125 I]hCG with hCG and direct binding of [ 125 I]hCG. The results from multiple studies using competition and saturation binding with LH/CG-R wild type yielded an apparent K d of 0.17 Ϯ 0.02 nM (n ϭ 11) and an average of about 1 Ϯ 0.1 ϫ 10 4 receptors/cell, uncorrected for transfection efficiencies (Table I). We also found that untransfected COS-7 cells, as expected (Loosfelt et al., 1989), do not contain LH/CG-R (data not shown).
Figs. 1-3 show the results of competition and saturation binding of the LH/CG-R point mutants, and the K d values and receptor numbers are given in Table I. Generally, the K d values are within severalfold of that obtained for wild type LH/CG-R, indicating that no major alteration in hormone affinity accompanied the mutations. Several of the receptor point mutants, however, had K d values slightly lower than that of wild type, e.g. Lys 55 3 Arg, Lys 55 3 Asp, and Lys 121 3 Asp. In contrast, the number of receptors/cell is, with the exception of the Arg 31 replacements and the Lys 55 3 Arg replacement, considerably less than that obtained for the wild type receptor. Most of the other LH/CG-R mutants yield on the order of 1-4 ϫ 10 3 receptors/cell, although receptors with the Lys 121 3 Asp and Asp 333 3 Lys replacements contained Ͻ1 ϫ 10 3 receptors/cell, again uncorrected for transfection efficiencies.
Two double mutants of LH/CG-R, E65K,E68K and E332K,D333K, were also characterized by direct and competitive binding (Fig. 4). The K d of the former double mutant was somewhat lower than that of LH/CG-R wild type, and the average number of receptors/cell was 625 Ϯ 106 (n ϭ 3) ( Fig. 4 and Table I). No specific binding was detected with LH/CG-R(Lys 332 , Lys 333 ) under the conditions used. Soluble binding assays were performed on several of the point LH/CG-R mutants that exhibited a relatively low number of surface receptors (Table II). Compared to wild type LH/ CG-R, the Lys 121 3 Asp and Lys 235 3 Asp mutants yielded only very low levels of specific binding. In contrast, the Glu 332 3 Lys and the Asp 333 3 Lys mutants gave specific binding intermediate to that of control and wild type LH/CG-R.
To assess the relative amounts of mutant receptors in cells with a low number of surface receptors, e.g. Ͻ25% LH/CG-R wild type, Western blots were obtained. As reported by others (Segaloff and Ascoli, 1993), purified rat LH/CG-R gives an apparent molecular mass of about 93 kDa. Transiently transfected COS-7 cells express a major band of about 93 kDa (Fig.  5), which is presumed to correspond to the 85-kDa form obtained by Hipkin et al. (1992) using stably transfected human

Parameters of hCG binding and stimulation of cAMP production in COS-7 cells expressing LH/CG-R wild type and mutants
Cells were transiently transfected with pSVL containing the cDNA to rat LH/CG-R wild type or mutant. After 48 h, competitive binding studies with hCG and [ 125 I]hCG were performed to determine IC 50 values, K d values, and average receptor numbers/cell; direct binding studies were also conducted with [ 125 I]hCG in several cases to determine K d values and average receptor numbers/cell. The production of cAMP in response to hCG was determined in transfected cells, and the ED 50 is given. The results are presented as mean Ϯ S.E. when n Ͼ 2; for n ϭ 2, the results are given as mean Ϯ range, i.e. upper and lower values.
The average value for wild type receptors (9.8 Ϯ 1.3 ϫ 10 3 receptors/cell) was normalized to 100%, and the numbers for mutant receptors are presented as a percentage of that for wild type. The numbers are not corrected for transfection efficiencies (5-10%) and thus reflect approximate relative values.
b Represents the highest concentration of hCG used. c No detectable receptors under the conditions used.

FIG. 1. hCG binding to COS-7 cells transfected with pSVL vectors containing the wild type and mutant LH/CG-R cDNAs.
Competition binding of [ 125 I]hCG and hCG to seven point mutant forms of LH/CG-R with replacements of Arg 21 , Arg 31 , Lys 55 , Glu 65 , and Glu 68 is shown. In each case wild type (Wt) receptor was included for comparison.

FIG. 2. hCG binding to COS-7 cells transfected with pSVL vectors containing the wild type and mutant LH/CG-R cDNAs.
Competition binding of [ 125 I]hCG and hCG to six point mutant forms of LH/CG-R with replacements of Arg 114 , Lys 121 , Glu 221 , Lys 235 , Glu 332 , and Asp 333 is shown. Wild type (Wt) receptor was also included for comparison. embryonic kidney 293 cells. In addition, a minor band of apparent molecular mass of 78 kDa is also detected from LH/ CG-R wild type, probably corresponding to the 68-kDa band reported by Hipkin et al. (1992), as are apparent higher and lower M r minor bands, which may be nonspecific. Interestingly, the LH/CG-R point mutants, Lys 55 3 Asp and Glu 332 3 Lys, as well as the double mutant, E65K,E68K, yield both the 93-and 78-kDa forms in about equal amounts, while the point mutants, Lys 121 3 Asp, Lys 235 3 Asp, and Asp 333 3 Lys, exhibit predominantly the lower M r form ( Fig. 5 and Table II). Equal amounts of cell lysate protein were added from cells transfected with wild type and mutant LH/CG-R cDNAs to facilitate comparison of the relative total amount of immunoreactive receptor in each case. This amount of protein, which was necessary to visualize the relatively low levels of mutant receptors, resulted in overloading the gel with wild type LH/CG-R and obscuring the resolution of the major and minor bands. These were resolved, however, by densitometric scanning of the blots and by adding reduced amounts of cell lysate protein (data not shown). Comparable studies on cell lysates that were not purified by lectin chromatography gave similar results (data not shown), indicating that no other non-glycosylated forms of LH/CG-R are present.
The functionality of each of the mutant receptors was evaluated by cAMP production in response to added hCG. Figs. 6 and 7 show typical results for those receptor point mutants that exhibited ED 50 values and maximal cAMP levels similar to LH/CG-R wild type. Interestingly, the LH/CG-R(Asp 121 ) mutant, which gave a low level of expression, exhibited an ED 50 and maximal cAMP response like that of wild type receptor. In contrast, the LH/CG-R point mutants Lys 332 and Lys 333 were non-responsive to added hCG (Fig. 8). Likewise, the double mutant, LH/CG-R(Lys 332 , Lys 333 ), for which no receptors could be detected, exhibited no hCG-responsive cAMP production. The ED 50 values for cAMP production by hCG in the transiently transfected COS-7 cells are given in Table I. DISCUSSION The results presented herein enable us to conclude that replacement of either Glu 332 or Asp 333 with Lys in LH/CG-R yields mutant receptors that bind hCG as well as LH/CG-R wild type but exhibit diminished signaling. Thus, this region of the extracellular domain of the receptor, which is Ͻ10 amino acid residues from the first transmembrane helix, appears critical in gonadotropin-mediated receptor activation. While receptor numbers are reduced with these particular mutants relative to LH/CG-R wild type, other single and double receptor mutants with comparable receptor densities, e.g. Asp 21 , Asp 55 , Asp 121 , Lys 65 , and Lys 68 gave cAMP responses like that of LH/CG-R wild type. Thus, the reduced hormone-mediated signaling noted with Lys 332 and Lys 333 is believed to be an intrinsic property of the mutant receptors and not a result of diminished receptor number.
The two M r forms of LH/CG-R expressed in COS-7 cells, i.e. 93 and 78 kDa as obtained herein, were first reported by Ascoli and colleagues (Hipkin et al., 1992) who expressed the receptor in stably transfected human embryonic kidney 293 cells, al-  a The values were normalized to 100% for LH/CG-R wild type, and the specific binding to mutant receptors is given as a percentage of that of wild type receptor.
b Western blots were densitometrically scanned, and the total intensity of the 93-and 78-kDa bands were normalized to 100%. The percentage of each of the forms is given. though the molecular mass values they determined were somewhat different than the values we obtained, e.g. 85 and 68 kDa, respectively. They reported data showing that the lower M r form was a precursor to the higher M r species; in addition, they suggested that the lower M r form bound hormone with reduced affinity relative to the major 85-kDa form. Since a variety of M r values has been reported for LH/CG-R purified from various tissues/species and expressed in cell lines (cf. Segaloff and Ascoli, 1993), the glycosylation patterns may differ depending upon the source of the receptor. In any case, we find that the lower M r form of rat LH/CG-R expressed in COS-7 cells is predominant for most of the mutants that lead to low receptor densities. Since there is no evidence of significant intracellular accumulation of these mutant forms, this observation suggests that the mature mutant receptors have a short half-life relative to LH/CG-R wild type, that processing proceeds less efficiently, or that the mRNA is degraded more rapidly or translated less efficiently.
Single replacements of LH/CG-R at several conserved sites with oppositely charged amino acid residues have no significant effect on the apparent K d relative to LH/CG-R wild type. That these particular replacements in LH/CG-R result in little or no change in the apparent K d of hCG binding does not necessarily mean that they are not contact sites. For example, Clackson and Wells (1995) replaced each of the 33 side chains on the extracellular domain of human growth hormone recep- Results are also shown for wild type (Wt) LH/CG-R and for untransfected (control) cells. The cAMP response to LH/CG-R(Asp 235 ) was lower than that of wild type receptor in this experiment, but this finding was not reproducible.
tor, known to be involved in single ligand binding (de Vos et al., 1992), and found that fewer than half of the mutant receptors exhibited any significant loss in binding affinity. The greatest contributions to hormone-receptor binding were attributed to six hydrophobic amino acid residues on the receptor and, to a lesser extent, to five ionizable side chains. These particular 11 contact sites form a contiguous region in which the hydrophobic groups are surrounded by the charged groups. Thus, about two-thirds of the receptor amino acid residues that are rendered solvent-inaccessible upon hormone binding contribute little to binding affinity. Whether a similar relationship will emerge for LH/CG-R remains to be determined.
With the exception of the replacements of Glu 332 and Asp 333 with Lys, the charge reversals we made on LH/CG-R did not alter receptor functionality as judged by cAMP production in response to added hCG. There was, however, a dramatic effect of most of these mutations on receptor density. A reduction in receptor number could result from inefficient translation of the mRNA, instability of the mutant receptor, and/or inefficient assembly into the plasma membrane, as has been noted for several point and deletion mutants of LH/CG-R (Segaloff and Ascoli, 1993).
It is interesting and somewhat surprising that cAMP responsiveness appears independent of receptor density over a fairly wide concentration range. Excluding replacements at positions 332 and 333, LH/CG-R wild type and the various mutants are present at average receptor numbers/cell ranging over 2 orders of magnitude; yet, the maximal amount of cAMP produced at high concentrations of hCG is about the same. These results suggest that the amount of G s proteins or adenylate cyclase is limiting with regard to hCG-LH/CG-R-mediated cAMP production.
In summary, Glu 332 and Asp 333 , which are located Ͻ10 amino acid residues from the beginning of the first transmembrane helix and are invariant in the known LH/CG, follitropin, and thyrotropin receptors, appear to participate in transmembrane signaling. Thus, these particular amino acid residues may have fundamental roles in signaling for all of the glycoprotein hormone receptors.