Photoaffinity Labeling of the Lutropin Receptor with Synthetic Peptide for Carboxyl Terminus of the Human Choriogonadotropin (cid:97) Subunit*

Human choriogonadotropin (hCG) consists of an (cid:97) subunit and a (cid:98) subunit. The existing evidence from various studies using truncation, substitution, synthetic hormone peptides, and hCG crystals suggests that the C-terminal region of the (cid:97) subunit contacts the luteinizing hormone/choriogonoadotropin (LH/CG) receptor and is involved in receptor activation. Despite a deluge of the speculation and the important role of the (cid:97) C-terminal region, direct evidence for its interaction with the receptor has been elusive. Because of the significant biological activity, it is imperative to prove the interac- tion of the (cid:97) C-terminal region. For this purpose, decamer peptides corresponding to the (cid:97) subunit se-quence from His 83 to Ser 92 ( (cid:97) 83–92 ) were derivatized with the N -hydroxysuccinimide ester of 4-azidobenzoylgly- cine (ABG) and radioiodinated. The resulting ABG- 125 I-(cid:97) 83–92 was capable of binding and activating the LH/CG receptor. Furthermore, UV-sensitive ABG- 125 I- (cid:97) 83–92 exclusively photoaffinity-labeled an (cid:59) 86-kDa molecule. This labeled molecule was shown to be the LH/CG recep- tor by various methods including immunoprecipitation by anti-LH/CG receptor antiserum. In addition, evi- dence is presented that the amino group of (cid:97) Lys

hCG 1 is a placental hormone and is involved in maintenance of the corpus luteum during pregnancy in human females. It is a member of the glycoprotein hormone family which includes LH, FSH, TSH and equine choriogonadotropin. These hormones consist of a common ␣ subunit and a distinct ␤ subunit which are noncovalently associated (1).
Photoaffinity labeling studies demonstrated that both ␣ and ␤ subunits affinity-labeled the LH/CG receptor (2). The truncation or substitution of hCG␣ C-terminal amino acid residues reduces the receptor-binding affinity and abolishes cAMP induction (3)(4)(5)(6). This is consistent with the observation that a dodecamer peptide corresponding to the hCG␣ C-terminal region, ␣ 81-92 , inhibited 125 I-hCG binding to the receptor (7) and that a decamer peptide, ␣ 83-92 , was capable of binding to cells possessing the LH/CG receptor and inducing cAMP synthesis (6). Also, the crystal structure of HF treated hCG suggests the ␣C-terminal region as part of the potential receptor binding site (8,9). Recently, a study using a set of reciprocal mutants of hCG and the LH/CG receptor, hCG ␣Lys 91 3 Asp and LH/CG-R Asp 397 3 Lys , indicated that the two amino acids were complementary in receptor activation (10). These results predict that the C-terminal region of hCG␣ interacts with the receptor and is crucial for receptor-activation to induce cAMP synthesis. Despite the continuing increase in speculation, direct proof has been elusive for the crucial interaction of the C-terminal region of hCG␣ with the receptor. For this reason, it is necessary to demonstrate the direct interaction of the C-terminal region of hCG␣ with the receptor. In this communication, we present the first unequivocal evidence for this interaction. Furthermore, evidence is presented that ␣Lys 91 of hCG is near to a carboxyl group of the receptor and that this ion pair is susceptible to cross-linking with EDC to form an amide bond.

EXPERIMENTAL PROCEDURES
Materials-hCG CR127 was supplied by the National Hormone and Pituitary Program. Peptides were synthesized by Multiple Peptide System (San Diego, CA) and purified as described previously (7). The N-hydroxysuccinimide ester of 4-azidobenzoylglycine (NHS-ABG) was synthesized, ␣ 83-92 was derivatized with NHS-ABG, and it was radioiodinated as described previously (11). The resulting ABG-125 I-␣ 83-92 was fractionated on a Sephadex G-10 column. It is important to carefully monoiodinate the peptide and preserve its biological activity during the harsh iodination reaction steps.
Binding, cAMP Induction, and Photoaffinity Labeling-Human embryonic kidney 293 cells were transfected with the LH/CG receptor construct as described previously (10). Cells transiently expressing the LH/CG receptor (293 ϩ ) were incubated with 125 I-␣ 83-92 , ABG-125 I-␣ 83-92 or 125 I-NAc/CONH 2 -␣ 83-92 in the presence of increasing concentrations of ␣ 83-92 . After washing cells three times until nonspecific binding was insignificant, the radioactivity of bound 125 I-␣ 83-92 , ABG-125 I-␣ 83-92 or 125 I-NAc/CONH 2 -␣ 83-92 , was determined. Heat denatured fetal calf serum and bovine serum albumin, 15 and 3%, respectively, were used to reduce nonspecific binding. For cAMP assay, 293 ϩ cells were incubated with increasing concentrations of ␣ 83-92 , ABG-␣ 83-92 or NAc/CONH 2 -␣ 83-92 and intracellular cAMP was determined as described previously (10). For photoaffinity labeling, 293 ϩ cells were incubated with ABG-125 I-␣ 83-92 and washed three times. The cells were irradiated with an ACME-LITE model 228A xenon flash lamp (12). UV light from the flashes are capable of activating UV sensitive groups such as the arylazide of ABG. Each flash lasts ϳ0.2 ms and minimizes random colli-* This work was supported by National Institutes of Health Grant HD-18702. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § sional crosslinks and heating of cells (12). The irradiated cells were solubilized in 3% sodium dodecylsulfate and 100 mM dithiothreitol. The samples were electrophoresed on polyacrylamide gels which were dried on filter paper and exposed to x-ray film.
Affinity Cross-linking of 125 I-NAc/CONH 2 -␣ 83-92--Transiently transfected 293 ϩ cells were incubated with 125 I-NAc/CONH 2 -␣ 83-92 , and washed to remove nonspecifically bound and unbound peptide as described above. The cells were incubated with 1 mM EDC at room temperature for 20 min for cross-linking, washed, and solubilized for gel electrophoresis. To examine the reaction specificity of EDC, the cells were pretreated with the N-hydroxysuccinimde ester of caproic acid (13) for 30 min or 100 mM of sodium acetate and treated with EDC. Gels were dried and exposed to x-ray film.
Immunoprecipitation-LH/CG receptors on 293 ϩ cells were photoaffinity labeled with ABG-125 I-␣ 83-92 as described above. They were washed three times on ice with ice cold 150 mM NaCl, 20 mM HEPES, pH 7.4. Cells were scraped in ice cold 150 mM NaCl, 20 mM HEPES (pH 7.4), containing protease inhibitors 30 mM phenylmethylsulfonyl fluoride, 1 mM leupeptin, and 1 mM EDTA. After pelleting cells by centrifugation at 1300 ϫ g for 10 min, pellets were resuspended in 0.5 ml of the buffer containing 1% Triton X-100, 30% glycerol and the protease inhibitors, incubated on ice for 45 min, and diluted with 1.5 ml of the buffer containing 0.1% Triton X-100, 30% glycerol and the protease inhibitors. The mixture was centrifuged at 100,000 ϫ g for 30 min and the supernatant was incubated with normal rabbit sera overnight at 4°C. The incubation mixture was further incubated with goat antirabbit antibody-Sepharose for 2 h and centrifuged 300 ϫ g for 3 min. This precleared supernatant was incubated with rabbit anti-LH/CG receptor antiserum and immunoprecipitated with goat anti-rabbit antibody-Sepharose. The Sepharose was washed several times and boiled in sodium dodecyl sulfate under the reducing condition for 5 min and centrifuged at 300 ϫ g for 3 min. The solubilized supernatant was electrophoresed and the gel was processed for autoradiography. In a control sample, the precleared pellet of goat antirabbit antibody-Sepharose was boiled in sodium dodecyl sulfate for 5 min and centrifuged 300 ϫ g for 3 min. The solubilized supernatant was electrophoresed on polyacrylamide gel. RESULTS  In addition to this 293 ϩ cell dependence, the formation of the ϳ86-kDa band requires photolysis as it is dependent on the number of xenon flashes and 125 I-ABG-␣ 83-92 bound to 293 ϩ cells failed to produce the ϳ86-kDa band when the sample was not photolyzed. These results suggest that the radioactive ϳ86-kDa band was generated by the photoactivation of 125 I-ABG-␣ 83-92 and photoaffinity labeling of a ϳ86-kDa molecule. The band was not produced by nonspecific association of 125 I-ABG-␣ 83-92 with a ϳ86-kDa molecule. Furthermore, the data indicate that 125 I-ABG-␣ 83-92 exclusively labeled a ϳ86-kDa molecule on 293 cells expressing the LH/CG receptor. As expected, the size of the ϳ86-kDa band material corresponds to the mature 85-86-kDa LH/CG receptor (13,14). Fig. 3 demonstrates the specificity of the photoaffinity labeling. 125 I-ABG-␣ 83-92 did not bind to mock transfected (293 Ϫ ) cells lacking the LH/CG receptor and did not form the ϳ86-kDa band even after UV photolysis. This result indicates the requirement of the LH/CG receptor for 125 I-ABG-␣ 83-92 binding and labeling the ϳ86-kDa band. Unlabeled hCG or ␣ 83-92 , but not FSH or TSH, blocked 125 I-ABG-␣ 83-92 binding to 293 ϩ cells, another indication of the specificity of 125 I-ABG-␣ 83-92 binding to the LH/CG receptor. Furthermore, ␣ 26 -46 or ␤ 38 -57 failed to block 125 I-ABG-␣ 83-92 binding to 293 ϩ cells. Since ␣ 26 -46 (7) and ␤ 38 -57 (15) inhibit 125 I-hCG binding to the LH/CG receptor, this result suggests that there is an exclusive and specific binding site for 125 I-ABG-␣ 83-92 in the LH/CG receptor.
Taken together these results clearly indicate 125 I-ABG-␣ 83-92 binding to the LH/CG receptor. Therefore, they suggest that the ϳ86-kDa band represents photoaffinity labeled LH/CG receptors. To obtain more direct evidence for this conclusion, the photoaffinity labeled band material was immunoprecipitated using rabbit anti-LH/CG receptor antiserum (Fig. 4). When 293 ϩ cells were incubated with 125 I-ABG-␣ 83-92 , photolyzed, solubilized in Triton X-100, and immunoprecipitated with anti-LH/CG receptor antiserum, the ϳ86-kDa-band was precipitated (Fig. 4). However, it was not precipitated when normal rabbit sera was used, when 293 ϩ cells were processed in the I-ABG-␣ 83-92 has two amino groups which could have been derivatized with ABG. They are the amino group of Lys 91 and the N-terminal amine. Our data cannot specify whether or not the amino group of Lys 91 was involved in the photoaffinity labeling of the receptor. Furthermore, it is difficult to predict the chemical group(s) of the receptor which was labeled as the phenyl nitrene reacts nonspecifically with a variety of functionalities (16). Since this information is necessary to define the cross-linked point of the peptide and the receptor, we have synthesized 125 I-NAc/CONH 2 -␣ 83-92 of which the N-terminal amine of ␣ 83-92 was acetylated and the C-terminal carboxylate was amidated. This modified peptide was used for affinity labeling the receptor. Transiently transfected 293 ϩ cells were incubated with 125 I-NAc/CONH 2 -␣ 83-92 and washed to remove unbound peptide. The cells complexed with 125 I-NAc/CONH 2 -␣ 83-92 were treated with EDC and solubilized for gel electrophoresis. The sample treated with EDC showed the ϳ86-kDa band of the receptor which was labeled with 125 I-NAc/CONH 2 -␣ 83-92 (Fig. 5). FSH or TSH did not prevent binding and crosslinking of 125 I-NAc/CONH 2 -␣ 83-92 . On the other hand, hCG or unlabeled ␣ 83-92 blocked binding of 125 I-NAc/CONH 2 -␣ 83-92 to the cells. Also 125 I-NAc/CONH 2 -␣ 83-92 did not bind to 293 Ϫ (data not shown). These results indicate the specificity of 125 I-NAc/CONH 2 -␣ 83-92 binding to the LH/CG receptor. Carbodiimides including EDC react with carboxyl groups to produce Oacylisoureas which in turn react with amino groups to produce amide bonds (16). Therefore, the primary reactive group of 125 I-NAc/CONH 2 -␣ 83-92 with EDC is the amino group of Lys 91 . To verify this reaction specificity the cells complexed with 125 I-NAc/CONH 2 -␣ 83-92 were treated with the N-hydroxysuccinimde ester of caproic acid which monofunctionally reacts with amino group (13). As expected, it prevented cross-linking of 125 I-NAc/CONH 2 -␣ 83-92 to the receptor as did acetate. These data indicate that the cross-linking reaction of EDC involves both amino and carboxyl groups, consistent with the reaction specificity of carbodiimides (16). Carbodiimides may also react with phenolic groups of Tyr and sulfhydryls of Cys. These reactions are, however, monofunctional and therefore, do not produce cross-linked products (17). Our data indicate that the amino group of 125 I-NAc/CONH 2 -␣ 83-92 is cross-linked to a carboxyl group to form an amide. They suggest that the amino group and side chain of hCG ␣Lys 91 is in the proximity of a carboxyl group of the receptor in the hormone receptor complex. EDC induced a zero length cross-link between these two counter ions suggests the existence of a salt bridge between them. This conclusion is consistent with the complementarity of hCG ␣Lys 91 and LH/CG receptor's Asp 397 (10).
The binding and labeling site of 125 I-ABG-␣ 83-92 appears to be specific since it was blocked by ␣ 83-92 , but not by two other peptides, ␣ 26 -46 and ␤ 38 -57 which are known to inhibit 125 I-hCG binding to the receptor. This is consistent with the x-ray crystal structure of HF treated hCG (8,9). In the crystal, the ␣C terminus shows a structure up to ␣Tyr 89 . The last three residues of the ␣C terminus, His 90 -Lys 91 -Ser 92 does not show a structure, despite their presence in an open ϳ5 nm solvent channel. This result suggests that ␣His 90 -Lys 91 -Ser 92 are flexible. It is interesting to speculate the ␣C-terminal positions up to ␣Tyr 89 , in relationship with ␣ 26 -46 and ␤ 38 -57 . ␣ 26 -46 is part of a loop in the hCG crystal and ␣Met 47 , the residue closest to ␣Tyr 89 and the ␣C-terminal peptide backbone, is 16.5 Å away from ␣Tyr 89 . Therefore, it is not expected for ␣ 26 -46 to compete with 125 I-ABG-␣ 83-92 for receptor binding, consistent with the data in Fig. 3. On the other hand, ␤ 38 -57 is closer to ␣Tyr 89 , ␤Gln 54 being 8.6 Å away from ␣Tyr 89 . Therefore, it may be possible for ␤ 38 -57 to weakly interfere with, although may not completely block, binding of 125 I-ABG-␣ 83-92 to the receptor. In fact, the intensity of the photoaffinity labeled ϳ86-kDa-band is somewhat reduced in the presence of ␤ 38 -57 in Fig. 3.
In this study we demonstrate that the ␣ 83-92 specifically interacts with the LH/CG receptor. In this complex of ␣ 83-92  5. Affinity cross-linking. Cells, 293 ϩ , were incubated with 125 I-NAc/CONH 2 -␣ 83-92 in the presence or absence of 1000 ϫ concentrations of FSH or TSH. After washing cells were treated with EDC, solubilized and electrophoresed. The gel was exposed to x-ray film. As controls, cells complexed with 125 I-NAc/CONH 2 -␣ 83-92 were treated with EDC in the presence of sodium acetate, a carboxyl specific competitor for EDC induced cross-link, or the N-hydroxysuccinimide of caproic acid, a monofunctional reagent specific to amino groups. The cells did not bind 125 I-NAc/CONH 2 -␣ 83-92 in the presence of hCG or unlabeled ␣ 83-92 . Also 293 Ϫ cells did not bind 125 I-NAc/CONH 2 -␣ 83-92 . Since these sample lanes were blank, the data are not shown. and the receptor, the amino group of hCG ␣Lys 91 is near to a carboxyl group of the receptor to form an ion pair. This interaction between the peptide and the receptor is sufficient to activate the receptor to induce cAMP synthesis as does the hormone itself.