LGR8 Signal Activation by the Relaxin-like Factor*

The relaxin-like factor (RLF) is thought to be responsible for the intra-abdominal migration of the testis during mammalian development. Our latest studies of RLF and LGR8 have revealed that the N-terminal region of the A chain is not required for receptor binding but is indispensable for cyclic AMP generation. RLF derivatives with six residues deleted from the N terminus of the A chain are active, whereas further truncation, up to the first A chain cysteine (A-10), yields tightly binding ligands devoid of signaling activity. These derivatives are specific competitive inhibitors (RLFi) of RLF. Although receptor binding is dependent upon B chain residues, the N-terminal region of the A chain is a generic trigger of the trans-membrane signaling activity.

Ades-(1-7) RLF is a specific competitive inhibitor of RLF function. In this paper we report that the receptor-binding region of RLF is physically separate from the trans-membrane signal initiation site in the N-terminal region of the A chain.

EXPERIMENTAL PROCEDURES
Materials-Fmoc amino acid derivatives were purchased from either Advanced ChemTech (Louisville, KY) or Novabiochem. Reagents and solvents for peptide synthesis were obtained from Advanced ChemTech. The LGR8 cloned into the pcDNA3.1.zeo plasmid was a gift from Dr. Hsueh, Department of Obstetrics and Gynecology, Stanford University School of Medicine. The human embryo kidney cell line 293T/17 was obtained from the American Type Culture Collection (ATCC CRL-11286) and used for LGR8 expression.
Peptide Synthesis-All RLF chains were synthesized by Fmoc chemistry, using an automated peptide synthesizer (Applied Biosystems, Foster City, CA, model 433A). The C-terminal amino acids were linked to Wang resin. Side chain-protecting groups of all trifunctional amino acids were trifluoroacetic acid-labile, except for methionine (sulfoxide) in position B-5, acetamidomethyl protected cysteines in positions A-11 and B-11, and tertiary butyl protected cysteine in position A-24. O-Benzotriazol-N,N,NЈ,NЈ-tetramethyluronium-tetrafluoroborate (HBTU) (12) was used for carboxyl activation. Each polypeptide chain was deprotected and removed from the solid support by treatment with a freshly prepared mixture of trifluoroacetic acid/water/ethanedithiol/ thioanisole/phenol (100/5/2.5/5/7.5) (v/v/v/v/w) for 2 h at room temperature (13). The resin was filtered off, the peptide was precipitated with ice-cold diethylether, the pellet was collected by centrifugation, washed two times with diethylether, air-dried, resuspended in water, and lyophilized. Each RLF chain was purified by reversed-phase HPLC on a Rainin C 18 -column (41.4 mm ϫ 250 mm) using 0.1% trifluoroacetic acid in water (solvent A) and 0.1% trifluoroacetic acid in 83% acetonitrile (solvent B) at a flow rate of 40 ml/min. The peptide was eluted using a 30-min linear gradient and detected by UV light absorbance of the effluent. Fractions of 10 ml were collected, and 10 l of each fraction analyzed for homogeneity by HPLC on a Bakerbond column (see below). Fractions containing the pure peptides were pooled and lyophilized.
The distribution of the cysteine-protecting groups allowed for the stepwise synthesis of the three disulfide bonds. In the A chain, cysteines A-10 and A-15, liberated during the trifluoroacetic acid treatment, were oxidized by titration with iodine in 50% acetic acid (peptide concentration ϭ 10 mg/ml). Excess iodine was immediately reduced with 1 M ascorbic acid in water. The reaction mixture was diluted with water to Ͻ20% (in acetic acid) and purified by preparative HPLC. The fractions containing the pure peptide were pooled and lyophilized. To convert cysteine (tert-butyl) (A-24) to the 2-pyridylsulfenyl derivative the protected A chain (17 mol) and 2,2Ј-dipyridyl disulfide (Aldrich) (240 mol) were dissolved in 900 l of trifluoroacetic acid and chilled on ice. Thioanisole (100 l) and 1 ml trifluoromethanesulfonic acid in trifluoroacetic acid (1:4 v/v) were added, and the reaction continued for 30 min at 0°C (10,14). The A chain was precipitated with chilled diethylether, and the pellet was collected by centrifugation, washed 3ϫ with 10 ml of diethylether, and air-dried. The RLF A chain derivative was dissolved in 1 M acetic acid and desalted on Sephadex G25 sf in 1 M acetic acid and lyophilized.
To produce the first interchain disulfide bond, equimolar amounts (4 mol) of the thiol-activated A chain and the monothiol B chain were reacted in 3 ml of 8 M guanidinium chloride, 0.1 M acetic acid (adjusted to pH 4.5 with NaOH) for 24 h at 37°C. The product was isolated by size separation on Sephadex G50 sf (2.5 cm ϫ 55 cm) in 1 M acetic acid followed by separation on HPLC on a Jupiter C 18 -column (10 mm ϫ 250 mm, Phenomenex). The last disulfide bond was formed by oxidative removal of * This work was supported by Grant 1-R01-HD40406 from the National Institutes of Health. 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.  1 The abbreviations used are: RLF, relaxin-like factor; CD, circular dichroism; HPLC, high performance liquid chromatography; MALDI-MS, matrix-assisted laser desorption/ionization mass spectrometry; Fmoc, N-(9-fluorenyl)methoxycarbonyl; hRLF, human RLF. the acetamidomethyl groups in positions A-11/B-11 using iodine in 95% acetic acid as solvent (10,14). After 60 min at room temperature in the dark the reaction mixture was slowly added to a vigorously stirring solution of 0.1 M ascorbic acid. The peptide was isolated by size separation on Sephadex G25 sf in 1 M acetic acid and lyophilized. Lastly, the methionine sulfoxide (B-5) (1.25 mol) was reduced with 800 l of 50 mM ammonium iodide in 90% trifluoroacetic acid for 15 min on ice. The reaction was quenched by addition of 10 mM ascorbic acid in water (5 ml), and the peptide was isolated by gel filtration on Sephadex G25 sf in 1 M acetic acid, followed by HPLC purification on a Jupiter C 18 -column. The solvents were removed by lyophilization. Final yields and mass spectroscopy data are presented in Table I.
High Performance Liquid Chromatography-Peptides were analyzed in two HPLC systems. For both systems solvent A consisted of 0.1% trifluoroacetic acid in water, and solvent B consisted of 0.1% trifluoroacetic acid in acetonitrile/water 4/1 v/v.
In system 1, ϳ10 -20 g of peptide was loaded onto a Bakerbond wide-pore C 18 -column (4.1 mm ϫ 250 mm). The peptide was eluted with a 30-min linear gradient from 20 to 60% B at a flow rate of 1 ml/min. The UV light absorbance of the effluent was recorded at 220 nm.
In system 2 an Aquapore 300 column (C 8 , 2.1 mm ϫ 30 mm) was used in combination with an ABI HPLC model 130A. Approximately 1 g of peptide was loaded and eluted with a 60-min linear gradient from 20 to 45% B at a flow rate of 100 l/min. The UV light absorbance of the effluent was recorded at 230 nm.
RLF analogs (10 g) were dissolved in 20 l of water and reduced with dithiothreitol (20 l, 50 mM) in 0.2 M Tris/HCl (pH 8.6) containing 6 M guanidinium chloride for 30 min at 37°C. Glacial acetic acid (5 l) was added, and the chains were separated by analytical HPLC in system 2.
RLF analogs (5 g) were dissolved in 10 l of 50 mM phosphate buffer, pH 7.5, and digested with Staphylococcal protease V8 (enzyme to substrate ratio (w/w), 1:10) for 16 h at 25°C (15) followed by trypsin (enzyme to substrate ratio (w/w), 1:20) for 30 min at 37°C. The reaction was quenched by the addition of 50 l of 0.1% trifluoroacetic acid in water and was subjected to analytical HPLC on Aquapore 300 using a 40-min linear gradient from 0 to 40% B.
Amino Acid Composition-Peptides were hydrolyzed in vapor phase 6 M HCl containing 1% phenol. Amino acids were modified with phenylisothiocyanate and separated by HPLC, using the Waters Pico . Tag system.
Matrix-assisted laser desorption/ionization (MALDI) mass spectrometry was performed at the mass spectrometry facility at the Medical University of South Carolina, Charleston. The RLF analog (1 g), dissolved in 1 l of 0.1% trifluoroacetic acid or 1 l of the enzymatic digest, was mixed with 3 l of ␣-cyano-4-hydroxycinnamic acid (50 mM in acetonitrile/water 4/1 v/v), and 1 l was spotted on a sample plate. Mass spectra were acquired on an Applied Biosystems (ABI) Voyager DR-STE instrument in the linear mode.
Circular Dichroism (CD) Spectroscopy-RLF derivatives (ϳ0.5 mg/ml) were dissolved in water, and the protein content was determined accurately by UV spectroscopy, using calculated absorbance coefficients. An aliquot of 25 g of RLF analog was diluted to 300 l with a Tris/HCl buffer (pH 7.5, 25 mM final concentration) for spectroscopic analysis. A Jasco J710 spectropolarimeter was used to acquire far-ultraviolet CD spectra at a resolution of 0.2 nm and a bandwidth of 2 nm. A cell of 0.1-cm path length was used, and 10 spectra were averaged. Molar ellipticity was calculated according to the literature (16).
Receptor-binding Assay-Transfected 293T/17 cells were grown to 80% confluence and then dislodged with 0.5 M EDTA (pH 7.5, 1 ml/10 ml of conditioned medium) for 10 min at 37°C. Cells were collected by centrifugation at 2000 ϫ g for 10 min, suspended two times in 1 ml of ice-cold binding buffer (20 mM Hepes, pH 7.5, 1% bovine serum albumin, 0.1 mg/ml lysine, 1.5 mM CaCl 2 , 50 mM NaCl, 0.01% NaN 3 ), and recentrifuged at 4°C for 10 min at 2000 ϫ g. The pellet was reconstituted to 25 ϫ 10 6 cells/ml of binding buffer. For binding assays 0.025 pmol of A 125 I-labeled Tyr-9 hRLF (17) in 20 l of binding buffer was placed into a 1.5-ml Eppendorf vial together with 40 l of binding buffer with or without human RLF or human RLF derivative and 40 l (10 6 cells) of the cell suspension and incubated at room temperature for 60 min. Thereafter the cells were diluted with 1 ml of ice-cold binding buffer and collected by centrifugation (10 min at 2000 ϫ g at 4°C); the pellet was washed with 1 ml of the same buffer and centrifuged for 10 min at 5000 ϫ g at 4°C. The tips of the vials containing the pellets were placed in counting tubes and transferred to a ␥ counter for analysis. The total binding was determined in the absence and nonspecific binding in the presence of 32 pmol of unlabeled RLF. In a regular assay, total binding was ϳ30,000 cpm, and nonspecific binding amounted to less than 5% of the total binding. Data points were collected in duplicates, and the data of at least 3 independent binding assays were pooled and presented as the mean (Ϯ S.E.).
Cyclic AMP Assay-Stably transfected LGR8-293T/17 cells were seeded in 24-well tissue culture plates at a density of 250,000 cells/well and grown in 500 l of Dulbecco's modified Eagle's medium, supplemented with fetal bovine serum (10%) and antibiotics. Twenty-four hours later the conditioned medium was replaced by 100 l of 3-isobutyl-1-methylxanthine (5 mM) in Dulbecco's modified Eagle's medium containing 1% bovine serum albumin. The cells were incubated for 1 h at 37°C in 5% CO 2 before 100 l of hormone dilution, prepared in the same buffer, was added. The cultures were grown at 37°C in 5% CO 2 in air for 16 h. Thereafter the plates were moved into a Ϫ80°C freezer and kept for 6 h at least. Prior to the assay, 20 l of 0.5 M EDTA was added to each well, the plate was incubated for 15 min at 37°C, and the content was transferred into 1.5-ml bullet tubes and heated to Ͼ90°C for 10 min in a water bath. Cell debris was removed by centrifugation at 14,000 rpm for 10 min at 4°C. The supernatant (100 l) was reacted with acetic anhydride, and the 2Ј-O-acyl-cAMP concentration was determined by radioimmunoassay using [ 125 I]succinyl-cAMP-tyrosyl methyl ester (18) (20,000 cpm ϭ 5 fmol) and anti-cAMP antiserum (Chemicon, Temecula, CA). The total response was determined in the presence of 80 pmol (2.5 g/ml) of human RLF, and the nonspecific response was determined in the absence of RLF. Assays of hRLF derivatives were compared with human RLF standards run in parallel. Assays were performed in duplicate, and 2-4 independent assays were averaged (Ϯ S.E.).

RESULTS
In the search for additional receptor-binding residues of human RLF (Fig. 1) we have synthesized RLF derivatives with truncations at the N termini. The yields of the corresponding derivatives and the MALDI-MS data are presented in Table I. All derivatives were homogeneous in two different HPLC systems. Upon reduction, two polypeptides were generated, and the HPLC elution profiles were compared with a human RLF standard treated in parallel. As expected, altered retention times were observed only for the chain carrying the modification. Enzymatic digests and peptide mapping confirmed the parallel arrangement of the peptide chains. Treatment with Staphylococcus aureus V8 protease, followed by trypsin, liberated the N-terminal A-(9 -19)-B- (9 -16)  19)-B-(9 -16) disulfide peptide. Peptides were identified by MALDI-MS and HPLC. The fragmentation pattern is in agreement with the parallel arrangement of the two chains (see supplemental material) as observed in natural and synthetic bovine RLF (9).
Truncation at the N terminus of the A chain up to the first cysteine in position A-10 resulted in derivatives with full receptor binding properties ( Fig. 2A). CD spectra (Fig. 2B) showed signal intensities similar to full-length RLF at the absolute minimum but showed a slight reduction at 222 nm. In addition, a small blue-shift of the absolute minimum and the crossover point were observed amounting to ϳ0.5 nm for the Ades-(1-6) and Ades-(1-7) RLF derivatives and ϳ1 nm for Ades-(1-8) and Ades-(1-9) RLF derivatives. Measurements of the cAMP-response yielded a complete dose-response curve for Ades-(1-6) RLF albeit with a reduced activity of ϳ10% relative to the full-length RLF. In contrast, the activity of derivatives truncated by 7, 8, and 9 residues from the N terminus of the A chain showed significantly reduced cAMP production; none of them reached full response even at the highest concentration. This effect is more pronounced for Ades-(1-8) and Ades-(1-9) RLF than for Ades-(1-7) RLF (Fig. 2C).
RLF shortened at the N terminus of the B chain by 5 residues retains full receptor binding affinity (Fig. 3A). Its CD spectrum indicates a slight increase of signal intensity and a very small red-shift (Ͼ0.5 nm) when compared with the CD spectrum of the full-length RLF (Fig. 3B). Bdes-(1-5) RLF retains the wildtype trans-membrane signal activity (Fig. 3C). The additional deletion of 6 residues from the N terminus of the A chain still does not affect receptor binding (Fig. 3A), or the secondary structure, as evidenced by the CD spectrum (Fig. 3B). However, the dose-dependent cAMP production is reduced to 10% of the RLF response (Fig. 3C).
Derivatives in which either arginine in position A-8 or tyrosine in position A-9 are replaced by alanine have 100% binding avidity, an intact secondary structure, and 100% cAMP response (Fig. 4). Ades-(1-9) RLF supplemented with the missing A chain nonapeptide in equimolar amounts could not restore adenylate cyclase activity (data not shown).
To evaluate the inhibitory effect we measured the RLF-dependent cAMP response in the presence of a constant concentration (2 ng/ml) of Ades-(1-8) RLF. In parallel, the inverse experiment was conducted by increasing the dose of the inhibitor Ades- (1)(2)(3)(4)(5)(6)(7)(8) in the presence of a uniform concentration (2 ng/ml) of hRLF. The two dose-response curves crossed at a concentration of 1.5 ng/ml, indicating that Ades-(1-8) behaves as a competitive antagonist. Inhibitory efficacy is a function of the direct cAMP attenuation and binding avidity. For the Ades-(1-8) RLF derivative the ED 50 is attained at a 1:1 ratio of RLF to inhibitor. DISCUSSION Results of previous studies suggest that tryptophan B-27 (10) and arginine B-16 2 contribute to the binding energy of RLF/ receptor interaction. However, the reduction in affinity upon 2 E. E. Bü llesbach and C. Schwabe, unpublished results.  2. RLF truncated at the N terminus of the A chain. A, competitive binding of RLF derivatives to LGR8 using A-125 I-labeled Tyr-9 hRLF (17) as tracer. Each data point was measured in duplicate, three independent experiments were performed, and the mean (ϮS.E.) of the pooled data is presented. B, far-ultraviolet CD spectra of the RLF derivatives were acquired. The protein concentration of each derivative was 83 g/ml in 25 mM Tris/HCl buffer at pH 7.5. Ten spectra were averaged. C, cAMP assays on LGR8 stably transfected 293T/17 cells. Each data point was collected in duplicate, and the data of two assays were pooled, and the mean (ϮS.E.) were reported. replacement of either amino acid by alanine is only ϳ50-fold, suggesting that additional binding sites exist that contribute to the relatively high energy of the RLF/LGR8 interaction (19). While searching for these binding regions by measuring both receptor-binding and cAMP production, we discovered a signal initiation region of RLF apart from the receptor-binding contacts.
The first target of our investigation was the N-terminal region of the A chain. The data suggest that binding to the LGR8 receptor expressed in 293T cells is not significantly perturbed by the removal of up to 9 residues from the N-terminal end of the RLF A chain. Small changes of the CD spectra imply changes in secondary structure, in particular, the blue-shift I-labeled Tyr-9 hRLF (17) as a tracer. Each data point was measured in duplicate, three independent experiments were performed, and the mean (ϮS.E.) of the pooled data is presented. B, far-ultraviolet CD spectra of the RLF derivatives were acquired. The protein concentration of each derivative was 83 g/ml in 25 mM Tris/HCl buffer at pH 7.5. Ten spectra were accumulated. C, cAMP assays on LGR8 stably transfected 293T/17 cells. Each data point was done in duplicate, data of two assays were pooled, and the mean (ϮS.E.) was reported.
FIG. 4. Characteristics of AR8A (AArg8Ala) and AY9A (ATyr9Ala) RLF derivatives. A, competitive binding of RLF derivatives to LGR8 using A 125 I-labeled Tyr-9 hRLF as the tracer. Each data point was measured in duplicate, three independent experiments were performed, and the mean (ϮS.E.) of the pooled data is presented. B, far-ultraviolet CD spectra of the RLF derivatives were acquired. The protein concentration of each derivative was 83 g/ml in 25 mM Tris/ HCl buffer at pH 7.5. Ten spectra were accumulated. C, cAMP assays on LGR8 stably transfected 293T/17 cells. Each data point was collected in duplicate, and the data of two assays were pooled, and the mean (ϮS.E.) is reported. and the reduction of the 222-nm minimum suggest presence of the less stable 3 10 -helix (20) next to the ␣-helix (16). Increased changes are observed for derivatives truncated up to 8 residues. It is unclear how many of these changes can be contributed to the secondary structure of the removed N-terminal peptide. Residue 9 is tyrosine, which may make a structureindependent contribution to the spectrum.
RLF B chain truncation experiments further enhance the impression that the molecule can be pared down substantially without affecting the core of the binding region. Extensive deletion of amino acids from both ends of the B chain does not interfere with receptor binding. Even in combination with N-terminal shortening of the A chain by 6 residues, the B chain N-terminal deletion of residues 1-5 and elimination of C-terminal residues beyond residue 27 from the B chain does not attenuate receptor binding. It appears that RLF can be shortened to 42 residues without significant reduction in receptor affinity (Fig. 3A) and without disturbance of the CD spectra. Note that removal of the Nterminal pentapeptide from the B chain compensates for the differences in the CD spectrum observed after truncation of the A chain alone (Fig. 3B). This would speak against the idea that water exclusion from the hydrophobic core is important for either binding or signaling activity. The B chain truncation would admit water to the same region that is uncovered if the A chain N-terminal tail is removed so that one can assume that removal of both the N-and C-terminal fragments would certainly admit water to the core region.
Although deleting the N-terminal end of the A chain has no effect on receptor binding, it reduces cAMP production by RLF as a function of chain length. Shortening the B chain N terminus by 5 residues and converting the new C-terminal tryptophan B-27 to an amide have no effect on cAMP production. However, in combination with deletion of 6 amino acids from the A chain N terminus, the B chain deletions caused suppression of cAMP production similar to the effect of the A chain hexapeptide deletion alone.
The cAMP production begins to decrease notably when 6 residues are removed from the N terminus of the A chain and goes to very low levels with the deletion of residues 7, 8, and 9. It appears that at least 5 amino acids have to precede cysteine (A-10) in the A chain to elicit the full cAMP response. The sequences of RLF from various species show His/Arg or Arg/His in positions 8 and 9, respectively, with human Arg/ Tyr providing the only known exception (Ref. 21 and references therein). To investigate the specificity requirements of the signaling initiation effect we synthesized RLF with ala-nine in position 8 and a second one with alanine in position 9; both derivatives were fully active. The fact that alanine is sufficient to restore signal initiation activity leaves one with the novel idea that, in this case, primary structure per se may be sufficient to illicit the most important communication function in an organism. The role of the A chain tail is merely to provide bulk to maintain a certain conformation of the receptor activation complex. Furthermore, it seems that RLF truncated beyond position A-6, while retaining some background activity, cannot stimulate full activity by a large excess of the ligand. This observation suggests that the A chain-truncated molecule is a specific competitive RLF inhibitor.
A potent inhibitor of RLF activity in vivo would provide an important research tool for investigating the ramifications and mechanisms of cryptorchidism. The idea was tested by measuring suppression of cAMP production by a uniform amount of truncated RLF added to each point of a human RLF doseresponse curve as well as by adding a uniform amount of active RLF to a RLF inhibitor curve. The result, depicted in Fig. 5, confirms that the truncated molecule is a competitive inhibitor of cAMP production by RLF and that approximately equimolar concentrations of the inhibitor are needed to suppress the cAMP production by 50%.
These experiments have led to the most surprising observation that a segment of a primary structure (as opposed to specific amino acids) is responsible for signal initiation of RLF as it binds to the LGR8 receptor. It invites the suggestion that receptor side chains must be in place to make and break contact, in the mode of a vintage telegraph. FIG. 5. Inhibition assay of human RLF shortened at the N terminus of the A chain by 8 residues. Either 2 ng/ml of Ades-(1-8) RLF was added together with variable amounts of human RLF or 2 ng/ml human RLF was added together with variable amounts of Ades-(1-8) RLF. Each assay was performed in duplicate, and two assays were pooled. The mean (ϮS.E.) is presented in the graph.