Sauvagine Cross-links to the Second Extracellular Loop of the Corticotropin-releasing Factor Type 1 Receptor*

Contact sites between the corticotropin-releasing factor receptor type 1 (CRFR1), the sauvagine (SVG) radioligands [Tyr0,Gln1]SVG (125I-YQS) and [Tyr0,Gln1, Leu17]SVG (125I-YQLS) were examined. 125I-YQLS or 125I-YQS was cross-linked to CRFR1 using the chemical cross-linker, disuccinimidyl suberate (DSS), which cross-links the ε amino groups of lysine residues that have a molecular distance of 11.4 Å. DSS specifically and efficiently cross-linked 125I-YQLS and 125I-YQS to CRFR1. CRFR1 contains 5 putative extracellular lysine residues (Lys110, Lys111, Lys113, Lys257, and Lys262) that can cross-link to the 4 lysine residues (Lys16, Lys22, Lys25, and Lys27) of the radioligands. Identification of the CNBr-cleaved fragments of CRFR1 cross-linked to 125I-YQLS or 125I-YQS established that the second extracellular loop of CRFR1 cross-links to Lys16 of YQS. Additionally, site-directed mutagenesis (changing Lys to Arg in CRFR1 individually and in combination) revealed that Lys257 in the second extracellular loop of CRFR1 is an important cross-linking site. In conclusion, it was shown that in SVG-bound CRFR1, Lys257 of CRFR1 lies in close proximity (11.4 Å) to Lys16 of SVG.

dation-resistant SVG analog, [Tyr 0 ,Gln 1 ,Leu 17 ]SVG (YQLS), which binds to CRFR1 and CRFR2 with high affinity and which can be cross-linked to CRF receptors with high efficiency (16), to map cross-linked residues. The results showed that Lys 16 17 ]SVG (YQLS) were synthesized in the Massachusetts General Hospital Biopolymer Facility. They were HPLC-purified and analyzed by mass spectroscopy, amino-terminal sequencing, and acid hydrolysis. SVG was obtained from Bachem (King of Prussia, PA) and disuccinimidyl suberate (DSS) was purchased from Pierce. Tissue culture media were from the Massachusetts General Hospital Media Facilities (Boston, MA). Centricon 30 centrifugal filters were purchased from Millipore (Bedford, MA). The molecular mass markers were purchased from Amersham Biosciences.
Cell Culture-COS-7 cells were transiently transfected with the murine CRFR1 and various CRFR1 Lys to Arg mutants at 90% confluency using the DEAE-dextran method (15). The cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 200 units/ml penicillin, and 20 g/ml streptomycin sulfate. The cells were cultured at 37°C in a humidified atmosphere in 95% air, 5% CO 2 .
Measurement of Receptor Expression on the Cell Surface-Nine of the ten c-Myc epitope tag sequences (QKLISEEDL) was introduced within the amino-terminal domain of mCRFR1 between Glu 31 and Ser 32 (17). Stably transfected LLCPK-1 cells (90 -95% confluent) and transiently transfected COS-7 cells in 24-well plates (72 h after transfection) were rinsed with phosphate-buffered saline (PBS) containing 5% heat-inactivated fetal bovine serum and incubated with the 9E10 monoclonal antibody (17) at 1:1000 dilution. The cells were incubated for 2 h at room temperature, rinsed with PBS, and incubated for 2 more hours with 125 I-labeled sheep anti-mouse immunoglobulin G (200,000 cpm/ well) diluted in PBS, 5% fetal bovine serum. The supernatant was removed, and then the cells were washed and lysed with 1 N NaOH. The cell lysates were collected and counted in a Micromedic gamma counter.
Radioligand Binding to Intact COS-7 Cells-Binding assays were performed as described previously (16,17). Intact COS-7 cells transiently transfected with mCRFR1 or mCRFR1 mutants were plated into 24-well plates, and a binding assay was performed when cells reached 90 -95% confluency. The cells were rinsed with a Tris-based binding buffer (50 mM Tris-HCl, pH 7.6, 100 mM NaCl, 2 mM CaCl 2 , 5 mM KCl, 5% heat-inactivated horse serum, 0.5% heat-inactivated fetal bovine serum). 125 I-YQLS (100,000 cpm/well), prepared and purified as previously described (14), was added in the presence of increasing concentrations of the competing peptide for 2-4 h at room temperature. The cells were then rinsed three times with binding buffer and lysed with 1 N NaOH (0.25 ml, 3ϫ). The cell lysates were collected, and the radio-activity was counted in an automated gamma counter.
Chemical Cross-linking of 125 I-YQLS or 125 I-YQS to CRFR1 and CRFR1 Mutants-Cells expressing wild type or mutant receptors were plated in 24-well plates and allowed to reach 90 -95% confluency. The cells were then rinsed with PBS. 125 I-YQLS or 125 I-YQS (1,000,000 cpm/well) was added in HEPES binding buffer (25 mM HEPES, pH 7.6, 125 mM NaCl, 5 mM KCl, 5% heat-inactivated horse serum, 0.5% heatinactivated fetal bovine serum) for 2-4 h at room temperature. The buffer was removed, the cells were rinsed with PBS to remove bound tracer, and DSS (0.5 mM) was added to the cells in PBS (pH 8.2) for 20 -30 min. The cells were then rinsed with PBS, lysed with SDS sample buffer, and analyzed on a 5-20% SDS-polyacrylamide gel. After electrophoresis the gels were dried and autoradiographed using x-ray film or a phosphorimager screen.
cAMP Stimulation Assay-COS-7 cells were plated in 24-well plates and were allowed to reach 90 -95% confluency. The cells were chilled on ice and rinsed with ice-cold PBS. The cells were then challenged with YQLS at different concentrations in Dulbecco's modified Eagle's medium containing 2 mM 3-isobutyl-1-methylxanthene (IBMX), 1 mg/ml bovine serum albumin, and 35 mM HEPES, pH 7.4 at 37°C for 15 min. The medium was then removed, and the cells were rapidly frozen on dry ice. Intracellular cAMP was extracted by thawing the cells in 1 ml of 50 mM HCl, and cAMP was then measured using radioimmunoassay (18).
Cyanogen Bromide (CNBr) Peptide Cleavage-After autoradiography, the radioactive protein bands were cut, and the proteins were electroeluted in 1ϫ SDS running buffer (25 mM Tris base, 192 mM of glycine, and 0.1% SDS). The eluate was concentrated using a Centricon 30 unit (Millipore). CNBr (100 mM) was added to the sample and dissolved in 70% formic acid at room temperature for 24 h. CNBr and formic acid were removed by lyophilization (3ϫ). The proteins were analyzed on a 5-20% gradient SDS gel or on Tricine/SDS-PAGE (12% acrylamide) according to the method of Schä gger and von Jagow (19) and autoradiographed using x-ray film or a phosphorimager screen.

DSS Cross-links 125 I-YQLS but Not 125 I-CRF to CRFR1
-125 I-CRF and 125 I-YQLS bind to CRFR1 with total specific binding of about 6 and 13%, respectively. As reported previously (16), addition of DSS caused covalent cross-linking of 125 I-YQLS to mCRFR1. No cross-linking occurred with 125 I-CRF. In contrast, 125 I-YQLS specifically cross-linked to cells expressing CRFR1 but not to cells transfected with the pcDNA1 plasmid (Fig. 1A). The cross-linked receptor was resolved as a broad band of ϳ80 kDa, which is consistent with the predicted molecular mass of the glycosylated CRFR1 (Fig.  1A). Addition of increasing concentrations of unlabeled SVG buffer decreased the 125 I-YQLS-CRFR1 cross-linked band in a concentration-dependent manner (Fig. 1B). These data indicate that 125 I-YQLS but not 125 I-CRF cross-links to CRFR1 efficiently and specifically.
Binding and Cross-linking of 125 I-YQLS to CRFR1 Bearing Lysine to Arginine Mutations-Because DSS cross-links free amino groups lying at a molecular distance of 11.4 Å; the efficient cross-linking of 125 I-YQLS to mCRFR1 provides an opportunity to determine which lysine residues in the ligand and the receptor are in close proximity to each other. Therefore we mutated the putative extracellular Lys to Arg residues in mCRFR1 and generated the following mutants: K257R, K262R, K257R/K262R, K110R/K111R, and K110R/K111R/ K113R. The mutant receptor plasmids were transiently transfected into COS-7 cells, and cell surface expression was determined using a double antibody binding assay to the epitope tag. Expression of K257R, K262R, K257R/K262R, K110R/K111R, and K110R/K111R/K113R was 90.1 Ϯ 0.4, 81.0 Ϯ 1.9, 72.8 Ϯ 0.6, 90.3 Ϯ 0.5, and 82.6 Ϯ 1.4% of that of wild type (Table I). Specific binding of 125 I-YQLS to K262R was similar to that of CRFR1 (Table I). In contrast, specific binding of 125 I-YQLS to K110R/K111R and K110R/K111R/K113R was 245.9 Ϯ 5.3 and 157.0 Ϯ 2.9% of that of CRFR1 (Table I); whereas its specific binding to K257R and K257R/K262R was 56.1 Ϯ 0.5 and 40.6 Ϯ 2.6% of that of CRFR1, respectively (Table I). All mutants exhibited comparable apparent K i values ( Fig. 2A and Table I).
YQLS increased cAMP accumulation in cells transfected with K110R/K111R, K110R/K111R/K113R, and K262R with an R max that was 79.9 Ϯ 17.3, 47.1 Ϯ 1.3, and 54.4 Ϯ 4.2% of that observed in cells transfected with the wild type CRF receptor ( Fig. 2B and Table I). Additionally, YQLS increased cAMP accumulation in cells transfected with K110R/K111R, K110R/ K111R/K113R, and K262R with an EC 50 that was similar to that observed in cells transfected with the wild type CRF receptor ( Fig. 2B and Table I). In contrast, the EC 50 of YQLS in cells transfected with K257R and K257R/K262R was greater than that observed in cells transfected with the wild type CRF receptor ( Fig. 2B and Table I), and their R max was 31.6 Ϯ 4.7 and 45.2 Ϯ 5.8% of the wild type ( Fig. 2B and Table I). The binding and the cAMP data suggest that the lysine residue at position 257 is an important residue for SVG interaction.
The different CRFR1 mutants were then cross-linked to 125 I-YQLS using DSS; the cells were lysed, and the labeled receptors were analyzed on a gradient SDS-PAGE followed by autoradiography. The cross-linked mutant receptors appeared as diffuse bands that were similar in size to that of the wild type receptor (Fig. 1C). Cross-linking efficiency to the mutant receptors was evaluated by loading equal amounts of radioactivity of cell lysates on SDS-PAGE. The intensity of the cross-linked mutated receptors, with the exception of that of K257R, was similar to or more than that of the wild type (Fig. 1C). This indicates that none of the Lys residues at positions 110, 111, 113, and 262 participates in the cross-linking reaction. In contrast, Lys 257 in the second extracellular loop (EC2) appears to be an important cross-linking site.
Identification of the CRFR1 Cross-linking Domain-To confirm that the EC2 region is involved in the cross-linking reaction, CNBr-cleavage of wild type and mutant receptors was used. CRFR1 contains several methionine residues in its backbone; among them Met 230 and Met 276 flank the predicted cross- linking site, Lys 257 . CNBr cleavage is predicted to generate a 46-residue fragment from CRFR1 cross-linked to the 41-residue radioligand, 125 I-YQLS. The predicted size of the crosslinked fragment (87 residues) is ϳ9 kDa (the CNBr-cleaved fragments are schematically represented in Fig. 3C). To resolve the small CNBr-cleaved fragments, equal amounts of radioactivity of the cross-linked receptor were loaded on high resolution Tricine/SDS-PAGE (19). As predicted, CNBr cleavage of 125 I-YQLS-cross-linked CRFR1 produced a major ϳ 9-kDa band and a less intense 11-kDa band (Fig. 3A, CRFR1). The intensity of the 11-kDa band was variable from experiment to experiment and represents incomplete CNBr cleavage. The intensity of the ϳ 9-kDa band decreased dramatically in K257R as compared with the wild type (Fig. 3A, R257), confirming that Lys 257 is an important cross-linking site. Additionally, the ϳ9-kDa fragment in K262R had a similar intensity to that of the wild type (Fig. 3A, R262), indicating that Lys 262 is not an important cross-linking site. However the ϳ9-kDa band disappeared completely in K257R/K262R indicating that Lys 262 participates in the cross-linking reaction only when Lys 257 is mutated (Fig. 3A, R257/R262). These data identify the EC2 region as a cross-linking site.
To positively identify the 9-kDa band a M230L mutant was constructed by changing the methionine residue at position 230 to leucine. The M230L mutant had an expression level and specific binding that were 85.3 Ϯ 5.2 and 167.4 Ϯ 0.9% of the wild type values, respectively (Table I). The apparent binding  affinity was also similar to that of the wild type receptor (Table  I). Additionally, YQLS increased cAMP accumulation in cells transfected with M230L with an EC 50 that was similar to that observed in cells transfected with the wild type CRF receptor (Table I). M230L was found to cross-link to 125 I-YQLS efficiently (Fig. 1C). The Leu 230 mutation increased the size of the cross-linked CNBr-cleaved fragment from ϳ9 to ϳ11 kDa. The size of the CNBr fragment in M230L is identical to the less intense ϳ11-kDa band seen in the wild type, consistent with an increase in the size of the fragment by 25 residues (CNBr cleavage occurs at Met 205 instead of Met 230 ; Fig. 3C, M 230 ).
Identification of the Sauvagine Cross-linking Residue-To identify the Lys residue within sauvagine that cross-links to the second extracellular domain of CRFR1, 125 I-YQS was used as the radioligand. 125 I-YQS contains a native methionine residue at position 17 that could be cleaved by CNBr (Fig. 3D). The 125 I-YQS radioligand is radiolabeled on Tyr 0 (at position zero). CNBr cleavage of 125 I-YQS is predicted to result in a 17-amino acid fragment bearing the 125 I label on Tyr 0 and containing one lysine residue (Fig. 3D, K 16 ). If Lys 16 is cross-linked to CRFR1, a labeled receptor fragment of 63 residues is predicted (17 residues from YQS and 46 residues from CRFR1). In contrast, if cross-linking occurs through the other lysine residues of SVG (Lys 22 , Lys 25 , Lys 27 ) then the receptor fragment is not labeled with 125 I. Analysis of the CNBr-cleaved, 125 I-YQS-cross-linked CRFR1 with high resolution Tricine-PAGE revealed a labeled band with a size of ϳ6 kDa (Fig. 3B). This band is smaller than the ϳ9-kDa band produced from CRFR1 cross-linked to 125 I-YQLS (Fig. 3A). The difference in size between the ϳ6and ϳ9-kDa bands reflects CNBr cleavage of SVG at Met 17 ; this identifies Lys 16 , the only Lys residue in the cleaved fragment, as the cross-linking residue.

DISCUSSION
Photoaffinity cross-linking of radiolabeled [Tyr 0 ]ovine CRF to rat CRFR1 was previously reported; however, the crosslinked residues were not characterized (20). We also reported that the oxidation-resistant sauvagine radioligand, 125 I-YQLS, which exhibits high affinity binding to the CRF receptors, cross-links efficiently to mCRFR1 through 1 (or more) of the 4 lysine residues of the ligand (positions 16, 22, 25, and 26) to 1 (or more) of the lysine residues of the receptor clustered at the juxtamembrane region (positions 110, 112, and 113) and/or the second extracellular loop (positions 257 and 262) (16). The data presented in this study indicate that Lys 16 of the ligand preferentially cross-links to Lys 257 on the EC2 of mCRFR1. When Lys 257 is mutated to Arg, the other Lys residue in EC2, Lys 262 , substitutes for Lys 257 , indicating that ligand binding to CRFR1 is not rigid and that residue 16 of CRFR1-bound YQS is within 11.4 Å from Lys 257 more often than from Lys 262 of CRFR1.
It is interesting to note that the single mutation of Lys 257 to Arg and the double mutation of Lys 257 and Lys 262 to Arg result in mutant receptors with decreased binding activity and decreased cAMP accumulation efficiency. This suggests that Lys 257 may be important for ligand-receptor interaction. In contrast, none of the other mutants has a decreased binding activity. The decreased binding of Lys 257 is consistent with data from human CRFR1 and CRFR2 hybrids showing important regions for SVG interaction in the EC2 (Asp 254 ) and at the junction of the EC2 with the fifth transmembrane domain (12,13,16).
The nature of bimolecular interactions of polypeptide ligands with their cognate receptors from group B of G-protein-coupled receptors has been recently investigated. Polypeptide ligands, such as CRF and PTH, exhibit strong helical tertiary structures (21,22). Important binding determinants, both in the amino terminus as well as the carboxyl terminus of the ligands, have been characterized. In general, cross-linking studies identified residues within the carboxyl terminus of the ligands that lie within close proximity to the amino terminus of their respective receptors and residues within the amino terminus of the ligands that lie within close proximity to the extracellular loops and the transmembrane domains of their respective receptors. However, differences were uncovered in some receptors. For instance, residues 1 and 13 of PTH-(1-34) cross-linked to Met 425 in the third extracellular domain and Arg 186 in the juxtamembrane region that is amino-terminal to transmembrane 1 of the PTH receptor type 1, respectively (23,24). Surprisingly, residues 6, 22, and 26 of secretin cross-linked to the most distal part of the amino terminus of the secretin receptor at positions Val 4 , Leu 17 , and Leu 36 , respectively (25), and residue 23 of PTH cross-linked to a region between residues 23 and 40 of the PTH receptor type 1 (26). Our data that show that residue 16 of SVG and residue 257 of CRFR1 are in close proximity to each other indicate that the residues in close proximity in ligand-bound CRFR1 are different from those found in the other members of this family of G-protein-coupled receptors. These results should facilitate modeling of ligandreceptor interaction in CRF receptors.