INSL5 is a high affinity specific agonist for GPCR142 (GPR100).

Insulin-like peptide 5 (INSL5) is a peptide that belongs to the relaxin/insulin family, and its receptor has not been identified. In this report, we demonstrate that INSL5 is a specific agonist for GPCR142. Human INSL5 displaces the binding of (125)I-relaxin-3 to GPCR142 with a high affinity (K(i) = 1.5 nM). In a saturation binding assay, (125)I-INSL5 binds GPCR142 with a K(d) value of 2.5 nM. In functional guanosine (gamma-thio)-triphosphate binding and cAMP accumulation assays, INSL5 potently activates GPCR142 with EC(50) values of 1.3 and 1.2 nM, respectively. In addition, INSL5 stimulates Ca(2+) mobilization in HEK293 cells expressing GPCR142 and G alpha(16). Overall, INSL5 behaves as an agonist for GPCR142 similar to relaxin-3. However, unlike relaxin-3, which is also a potent agonist for GPCR135 and LGR7, INSL5 does not activate either GPCR135 or LGR7. INSL5 inhibits (125)I-relaxin-3 binding to GPCR135 with a low potency (K(i) = 500 nM). A functional assay shows that INSL5 (1 microm) is a weak antagonist for GPCR135. In addition, INSL5 (up to 1 microm) shows no affinity or activity at LGR7 or LGR8 either in a binding assay or a bio-functional assay. Previously, we have demonstrated that GPCR142 mRNA is expressed in peripheral tissues, particularly in the colon. Here we show that INSL5 mRNA is expressed in many peripheral tissues, similar to GPCR142. The high affinity interaction between INSL5 and GPCR142 coupled with their co-evolution and partially overlapping tissue expression patterns strongly suggest that INSL5 is an endogenous ligand for GPCR142.

The relaxin/insulin family peptides include insulin (1), IGF1 1 (2), IGF2 (3), relaxin (4,5), INSL3 (6), INSL4 (7), INSL5 (8), INSL6 (9), and relaxin-3/INSL7 (10). Except for IGF1 and IGF2, which are single chain peptides, each member of the family consists of two peptide subunits (an A-chain and a B-chain) that are linked by three disulfide bonds (4 -14). Insulin, IGF1, and IGF2 are known to be involved in the regulation of glucose metabolism (15) and signal through tyrosine kinase/ growth factor receptors, which are single transmembrane re-ceptors (16,17). Relaxin plays multifunctional roles including uterus relaxation, reproductive tissue growth, and collagen remodeling in females (18). In addition, relaxin has been reported to play important roles in nonreproductive functions including wound healing, cardiac protection, and allergic responses (19). The receptor for relaxin has been identified recently as a leucine-rich repeat containing the G-protein-coupled receptor (LGR) LGR7 (20). Although relaxin also activates LGR8 in vitro (20), recent studies show that LGR8 is likely the endogenous receptor for INSL3 and is involved in testis descent (21,22). To date, the receptors for INSL4, INSL5, and INSL6 have not been identified. Relaxin-3 (also known as INSL7), the most recently identified member of the family, was reported to be an additional ligand for LGR7 (23). We recently identified relaxin-3 as a ligand for two orphan G-protein-coupled receptors GPCR135 (14) and GPCR142 (24). The predominant brain expression for both relaxin-3 (10,14,25,26) and GPCR135 (14,26), coupled with their high affinity interaction, strongly suggests that relaxin-3 is the endogenous ligand for GPCR135. The tissue expression pattern of GPCR142 (also known as GPR100), which is primarily in peripheral tissues (24), is drastically different from that of relaxin-3, suggesting that GPCR142 may have an endogenous ligand other than relaxin-3. Furthermore, despite the high conservation of relaxin-3 in different species, GPCR142 is less conserved in the mouse and is a pseudogene in the rat (26), suggesting that GPCR142 may have a diminished role in rodents and may function as a receptor for a different ligand (other than relaxin-3) in other mammals. Sequence analysis among insulin/relaxin family members indicates that INSL5 shares high homology to relaxin-3 ( Fig. 1A), suggesting that it may be an additional ligand for GPCR135, GPCR142, LGR7, or LGR8. In this report, we demonstrate that INSL5 is an agonist for GPCR142 but not for GPCR135, LGR7, or LGR8.

Analysis of INSL5 mRNA Expression in Different Human
Tissues by Quantitative PCR Two primers for human INSL5 (forward primer, 5Ј GTG GGC TAG AAT ACA TAC GG 3Ј; reverse primer, 5Ј TCT TCA GTG GGC ATC TGT CC 3Ј), designed according to the published INSL5 cDNA sequence (8), and two primers for human GPCR142 (forward primer, 5Ј TCC TGG TGG CTT CCT TCT TC 3Ј; reverse primer, 5Ј CAG GAG ACA GTA CAG CAC AG 3Ј), designed according to published GPCR142 cDNA sequence (GenBank TM accession number AY394502), were used to PCR-amplify the human INSL5 and GPCR142 cDNA, respectively. Human cDNA pools made from RNAs extracted from different human tissues (BD Biosciences) were used as PCR templates. Human ␤-actin primers (forward primer, 5Ј CAC TCT TCC AGC CTT CCT TC 3Ј; reverse primer, 5Ј CGA TCC ACA CGG AGT ACT TG 3Ј) were used to amplify the actin cDNA as the internal controls. TaqDNA polymerase with TaqStart antibody (BD Biosciences) was used as the polymerase for PCR. SyBr Green (Molecular Probe) was added to the PCR to monitor * 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. the DNA synthesis. The PCRs were performed in a SmartCycler (Cepheid, Sunnyvale, CA) at conditions of 95°C for 5 s for denaturing, 62°C for 7 s for annealing, and 72°C for 10 s for extension. The PCRs were performed at the above conditions for 40 cycles. The signals were collected after each extension step. Purified human GPCR142, INSL5, and actin PCR products at different concentrations were used in PCRs as the standards for quantitation. The relative abundance of GPCR142 and INSL5 mRNA in different tissues were normalized to the actin expression level for each tissue.

Identification of the Cleavage Sites in INSL5 Pre-propeptide
Identification of the Signal Peptide Cleavage Site of INSL5-The human INSL5 cDNA was PCR-amplified from human colon cDNA (BD Biosciences) using two primers (forward primer, 5Ј ACT AGA GAA TTC  GCC ACC ATG AAG GGC TCC ATT TTC ACT CTG TTT TTA TTC TCT  G 3Ј; reverse primer, 5Ј GTC ATC AAG CTT TCA CTT GTC ATC GTC  ATC CTT GTA GTC GCA AAG AGC ACT CAA ATC AGT CAT 3Ј). The PCRs were performed at 94°C for 30 s, 65°C for 30 s, 72°C for 1 min and 30 s for 40 cycles. The PCR product was cloned into pCMVsport1 (Invitrogen) between EcoRI and HindIII sites. A FLAG tag coding region was inserted before the 3Ј end stop codon. The DNA construct was sequenced to confirm the sequence identity and then transfected into African green monkey kidney COS-7 cells (ATCC number CRL-1651, ATCC, Manassas, VA) by using Lipofectamine (Invitrogen) as the transfection reagent. Three days after transfection, the C-terminally FLAG-tagged INSL5 peptide was first affinity-purified from the conditioned medium by using an anti-FLAG affinity column and then further purified by HPLC by using a C-18 column (4.6 ϫ 250 mm, Vydac, Hesperia, CA). The N-terminal sequences of the purified proteins were analyzed by Edman degradation (Molecular Structure Facility, University of California, Davis).
Identification of the Cleavage Site between the C-chain and A-chain of INSL5-The N-terminal sequencing results for the C-terminal FLAGtagged INSL5, which reveals the signal peptide cleavage site, were used to guide the construction of the N-terminal FLAG-tagged INSL5 expression vector. The INSL5 coding region, excluding the signal peptide coding region, was PCR-amplified by using the above C-terminal FLAGtagged INSL5 expression vector as the template and two primers (forward primer, 5Ј ATA TAG GAA TTC GAC GAC GAC GAC AAG AAG GAG TCT GTG AGA CTC TGT GGG C 3Ј; reverse primer, 5Ј ATG ATA AAG CTT TTA GCA AAG AGC ACT CAA ATC AGT CAT 3Ј). The PCRs were performed at 94°C for 30 s, 65°C for 30 s, 72°C for 1 min 30 s for 20 cycles. The PCR products were then cloned into the modified pCMV-sport1 vector (14) between EcoRI and HindIII sites. The resulting DNA construct was sequenced to confirm the sequence identity and co-transfected into COS-7 cells with a furin expression vector (14) for transient expression. The N-terminally tagged INSL5 peptide secreted into the cell culture medium was affinity-purified with an anti-FLAG column as described for purification of recombinant relaxin-3 (14). The N-terminal FLAG tag was removed by digesting the peptides with enterokinase (Novagen, Madison, WI) at 37°C overnight. The untagged INSL5 peptide was further purified by reverse phase HPLC. The HPLC fractions were lyophilized, reconstituted in water, and tested for ligand activity by using SK-N-MC/␤-gal cells stably expressing human GPCR142 (24). The peptides in active fractions were analyzed by mass spectrometry (MS) as described previously (14) for molecular mass determination, which was then used to deduce the cleavage sites of the B-chain/C-chain and the C-chain/A-chain.

Expression and Purification of Human Recombinant INSL5
In INSL5, at the junction of the B-chain and the A-chain, there is a natural RWRR sequence, which is a preferred furin cleavage site (27). The MS results indicated that the cleavage site between the C-chain and the A-chain occurs after a single Arg 114 of the INSL5 pre-propeptide. Our results from expression of the N-terminally FLAG-tagged INSL5 indicated that the cleavage after Arg 114 is not efficient even when co-expressed with furin. To produce the wild type INSL5 mature peptide at higher efficiency, we engineered the INSL5 expression construct similarly to that of relaxin-3 (14) by introducing an artificial preferred furin site at the junction of the C-chain and the A-chain. The human INSL5 propeptide coding region was PCR-amplified through two steps to introduce a mutation that encodes the artificial furin site (RWRR) between the C-chain and the A-chain. The first step was performed using cloned human INSL5 cDNA as the template and two primers (forward primer, 5Ј ATA TAG GAA TTC GAC GAC GAC GAC AAG AAG GAG TCT GTG AGA CTC TGT GGG C 3Ј; reverse primer, 5Ј GTG CAA CAC AAA GTT TGT AAA TCT TGT CTT CGC CAA CGT GAA TGC TTC TTT GAC TTC CAA AGC TC 3Ј). The PCR was performed at conditions of 94°C for 30 s, 65°C for 30 s, and 72°C for 1 min for 20 cycles. The resulting PCR products from the first step were used as the template for the second step PCR using two primers (forward primer, 5Ј ATA TAG GAA TTC GAC GAC GAC GAC AAG AAG GAG TCT GTG AGA CTC TGT GGG C 3Ј; reverse primer, 5Ј ACT AGA AAG CTT TTA GCA AAG AGC ACT CAA ATC AGT CAT GGA ACA GCC ATC AGT GCA ACA CAA AGT TTG TAA ATC TTG TC 3Ј). The PCR was performed at conditions of 94°C for 30 s, 65°C for 30 s, and 72°C for 1 min for 20 cycles. The product of the second step PCR was then digested with EcoRI and HindIII and cloned into the pCMV-signal FLAG. The resulting DNA construct was sequenced to confirm the sequence identity and then co-expressed with furin in COS-7 cells. The N-terminally FLAG-tagged INSL5 was again first purified by an affinity column, cleaved to removed the N-terminal tag, and then further purified by HPLC as described above.

Radioligand Binding Assays
To characterize INSL5 as the possible ligand for GPCR135 and GPCR142, cell membranes from COS-7 cells transiently expressing human GPCR135 or GPCR142 were used in radioligand binding assays by using 125 I-relaxin-3 (2,200 Ci/mmol) as the radioligand as described (14,24). Briefly, unlabeled INSL5 or other peptides at various concentrations were added to the membranes from COS-7 cells expressing either GPCR135 or GPCR142 in the presence of 100 pM 125 I-relaxin-3. The binding assays were incubated at room temperature for 1 h. The binding mixtures were then filtered through GFC filters pre-saturated with 0.3% polyethylenimine (Sigma) and washed with cold binding buffer. The bound radioligand was then counted in a microscintillation counter (TopCount/NTX, Packard Instrument Co.) with Microscint-40.
To characterize INSL5 as a possible ligand for LGR7 and LGR8, COS-7 cells transiently expressing human LGR7 and LGR8 were used in the whole cell binding assay. Briefly, COS-7 cells transiently expressing LGR7 or LGR8 were detached from culture dishes with phosphatebuffered saline plus 10 mM EDTA and seeded in 96-well opaque culture plates. Radioligand ( 125 I-relaxin-3 for LGR7 and 125 I-INSL3 for LGR8) was added to each well at a final concentration of 100 pM in binding buffer (DMEM plus 50 mM HEPES and 1% bovine serum albumin). Unlabeled INSL5 or other peptides at various concentrations were added to the binding mixture as the competitors. The binding assays were carried out at 4°C for 5 h. The unbound radioligand was removed by aspirating the binding buffer. Cells in the plates were washed with cold phosphate-buffered saline. Fifty microliters of Microscint-40 was then added to each well of the 96-well plate. The plates were then counted in a microscintillation counter (TopCount/NTX, Packard In-strument Co.) to measure the bound radioligand. The results were analyzed by PRISM 3.0 software (GraphPad Software, San Diego). The IC 50 values are the concentration of unlabeled ligand that inhibits 50% of the total specific binding.
To determine the K d values of INSL5 to GPCR135, GPCR142, LGR7, and LGR8, INSL5 was labeled using chloramine T (Sigma) in the presence of Na 125 I (PerkinElmer Life Sciences). The HPLC-purified 125 I-INSL5 (4,400 Ci/mmol) at different concentrations was incubated with COS-7 cells (in 24-well tissue culture plates) transiently expressing either GPCR135, GPCR142, LGR7, or LGR8 in binding buffer (DMEM plus 50 mM HEPES, pH 7.2, and 1% bovine serum albumin in a final volume of 100 l). The binding reaction was incubated at room temperature for 1 h. The binding buffer was aspirated, and unbound ligands were removed by washing the cells in the 24-well plates with cold phosphate-buffered saline. The bound ligand was dissolved in 1% SDS plus 1 N NaOH and counted in a gamma counter (Cobra, Quantum, Packard Instrument Co.). The nonspecific binding was assessed by performing the binding assay in the presence of 1 M human relaxin-3. The results were analyzed using PRISM 3.0 software (GraphPad Software).

GTP␥S Binding Assays
The GTP␥S binding assays were performed as described (24

Ca 2ϩ Mobilization Assays
HEK293 cells (ATCC number CRL-1573, ATCC, Manassas, VA) were co-transfected with human GPCR142 and G␣ 16 expression constructs by using Lipofectamine as the transfection reagent (Invitrogen). Two days after transfection, the cells were detached with phosphate-buffered saline plus 10 mM EDTA and washed with DMEM/F12 (Invitrogen) without phenol red. Cells were then seeded in 96-well black wall poly-D-lysine-coated plates with 4 M Ca 2ϩ dye Fluo-3 (TEFLABS, Austin TX) in DMEM/F12 plus 2 mM Probenecid (Sigma). Different concentrations of INSL5 or relaxin-3 were added to cells to stimulate intracellular Ca 2ϩ mobilization, which was monitored by a fluorescence imaging plate reader instrument (Molecular Devices, Sunnyvale, CA). The results were analyzed using PRISM 3.0 software. The EC 50 value is the ligand concentration that stimulates 50% of maximum Ca 2ϩ release.

␤-Galactosidase Assays for cAMP Accumulation
SK-N-MC/␤-gal cells harbor a ␤-galactosidase gene under the control of a cAMP-response element promoter (28). In these cells, elevation in cAMP concentration leads to increased ␤-galactosidase gene expression, whose activity can be measured as described previously (28). SK-N-MC/ ␤-gal cells were used to establish cell lines stably expressing human GPCR135, GPCR142, LGR7, and LGR8. INSL5 or relaxin-3 at various concentrations was added to cells expressing GPCR135 or GPCR142 to inhibit forskolin (5 M, Sigma)-stimulated ␤-galactosidase gene expression. For characterization of LGR7 and LGR8, INSL5, porcine relaxin, or INSL3, at various concentrations, was added to LGR7-or LGR8expressing cells to induce the ␤-galactosidase expression. The intracellular ␤-galactosidase activity was measured by a colorimetric method using Chlorophenol Red-␤-D-galactopyranoside (Roche Diagnostics) as the substrate and reading the absorbance at a wavelength of 570 nm.

Identification of the Cleavage Sites for INSL5 Maturation and Recombinant Expression of INSL5-Members
of the insulin/relaxin family, such as insulin, relaxin, relaxin-3, and INSL3, have a two-subunit (B-chain and A-chain) structure and are processed from their pre-propeptides through the removal of the signal peptides and C-peptides. INSL5 is a member of insulin/relaxin family, and the cleavage sites for the INSL5 pre-propeptide have not been determined. The junction of B-chain and C-chain has an RWRR sequence (Fig. 1A), which is a preferred furin cleavage site and is conserved in relaxin-3. The cleavage sites in relaxin-3 have been proven through investigation of the endogenous peptide as well as the recombinant peptide expressed in mammalian cells (14). Therefore, we are confident that this cleavage site will be used in the INSL5 maturation process. However, the signal peptide cleavage site and the cleavage site between the C-chain and A-chain have remained unclear. A C-terminally FLAG-tagged INSL5 expression vector was constructed for expression of INSL5 in mammalian cells to determine its cleavage sites for INSL5 maturation, which includes the signal peptidase cleavage site, the cleavage site between the B-chain and C-chain, as well as the site between the C-chain and the A-chain. The C-terminal FLAG-tagged INSL5 was transiently expressed in COS-7 cells. The protein secreted into the medium was purified by an anti-FLAG affinity column and then further purified by a reverse phase HPLC. The N terminus of the purified INSL5 peptide was then analyzed by Edman degradation. The N-terminal sequence analysis results revealed a dominant major sequence, KESVRL-, matching the B-chain sequence of INSL5, indicating that the signal peptidase cleaves between Ser 22 and Lys 23 of the INSL5 pre-propeptide. The N-terminal sequencing results also revealed a minor sequence of HXXXIPQ (X stands for any amino acid), which matches the predicted N terminus of the C-chain after cleavage, confirming that the RWRR sequence between the B-chain and the C-chain is utilized for INSL5 maturation. No apparent sequence from the sequencing results matches the possible N terminus of the A-chain, indicating that the cleavage between the C-chain and the A-chain either did not occur or was incomplete. It was also possible that the N terminus of the A-chain, if cleaved, may have been modified and blocked from Edman degradation. To further investigate the cleavage site between the C-chain and the Achain, the INSL5 pro-peptide coding region (excluding the signal peptide coding region) was cloned into a pCMV-sport1 signal-FLAG expression vector (14) for expression of a secreted N-terminally FLAG-tagged INSL5 (with the tag at the N terminus of the B-chain) in COS-7 cells with co-expression of furin. The N-terminally tagged INSL5 was purified with an affinity column. SDS-PAGE analysis indicated that the purified FLAG-tagged peptides are heterogeneous in molecular mass ranging from 7 to 15 kDa (data not shown), suggesting the C-chain removal was incomplete. The peptide was cleaved with enterokinase to remove the FLAG tag and then further purified by HPLC. Peptides in HPLC fractions were analyzed by Edman degradation for N-terminal sequencing as well as by MS for molecular weight determination. Peptides from one major HPLC peak demonstrated an apparent molecular mass of 5043 daltons, which is much smaller than that of the propep- tide (12,837 daltons) and indicates the C-chain has been removed. Edman degradation of the peptide in the fraction showed a single sequence of KESVRL-, matching the N terminus of the B-chain. Again, no sequencing results match the possible A-chain sequence, suggesting that the N terminus of the A-chain is blocked. The molecular mass of the peptide (5043 daltons) as measured by MS is 17 daltons smaller than the predicted molecular mass (5060 daltons) of the INSL5 mature peptide if the cleavage between the C-chain and the A-chain occurs between Arg 114 and Gln 115 . Because newly exposed Gln at the N terminus is often modified into pyro-Glu (ϽGlu), which results in reduction of the molecular mass of 17 daltons by losing an NH 3 group, our results strongly indicate that the cleavage between the C-chain and the A-chain does occur between Arg 114 and Gln 115 . Based on the above result, we redesigned the INSL5 expression construct to express N-terminal FLAG-tagged INSL5 with a modification of the DNA to encode an additional preferred furin cleavage site (RWRR) (27) just before Gln 115 . The N-terminally FLAG-tagged INSL5 was transiently co-expressed with human furin in COS-7 cells and affinity-purified by an anti-FLAG column. The affinity-purified peptides appeared uniform in molecular weight in the SDS-PAGE analysis (data not shown), suggesting that the C-chain removal is complete. The N-terminal FLAG tag was removed, and the untagged wild type INSL5 was further purified by reverse phase HPLC. The HPLC-purified peptide was analyzed by mass spectroscopy and showed that it has a molecular mass of 5043 daltons, which confirms that the cleavages between B/C-chain as well as the C/A-chain were processed as we predicted. In addition, the N-terminal Gln of the A-chain is transformed into pyro-Glu (Fig. 1B). Ligand for GPCR135, GPCR142, LGR7, and LGR8 in Radioligand Binding Assays-We characterized INSL5 peptide as a possible ligand for GPCR135, GPCR142, LGR7, and LGR8 in radioligand binding assays. 125 I-Relaxin-3 was used as the tracer for binding assays involving GPCR135, GPCR142, and LGR7. 125 I-INSL3 was used as the tracer to characterize LGR8. Relaxin-3 potently displaced 125 I-relaxin-3 binding to GPCR135 with an IC 50 value of 0.5 Ϯ 0.1 nM (Fig. 2A) (Fig. 3A) show that 125 I-INSL5 binds GPCR142 at high affinity with a K d value of 2.5 Ϯ 0.3 nM. The saturation binding also show that 125 I-INSL5 binds GPCR135 at a much lower affinity (Fig. 3B). The K d value of 125 I-INSL5 to GPCR135 was not assessed because the saturation was not reached at the highest concentration tested (10 nM). In parallel experiments, mock-transfected cells or cells expressing LGR7 or LGR8 were used in saturation binding assays by using 125 I-INSL5 as the tracer, and no specific binding was observed for those cells (data not shown).

Pharmacological Characterization of INSL5 as a Possible
INSL5 Activates GPCR142 but Not GPCR135, LGR7, or LGR8 -INSL5 was investigated for functional activity at GPCR142 and GPCR135 by using GTP␥S binding assays, cAMP accumulation assays, and Ca 2ϩ mobilization assays. The results showed that in the GTP␥S binding assay, INSL5, up to 10 M, did not stimulate detectable GTP␥S incorporation in GPCR135-expressing cells, whereas relaxin-3 potently stimulated GPCR135 with an EC 50 value of 0.4 Ϯ 0.1 nM (Fig. 4A). In a parallel experiment, INSL5, similar to relaxin-3, potently stimulates GTP␥S binding in GPCR142 membranes with an EC 50 value of 1.3 Ϯ 0.1 nM (Fig. 4B).
Both GPCR135 and GPCR142 are linked to cAMP inhibition (14,24). To evaluate whether INSL5 activates GPCR135 or GPCR142 in cAMP inhibition assays, SK-N-MC/␤-gal cells harboring a ␤-galactosidase gene under the control of a cAMPresponse element promoter were used as the host cells (28) to express GPGR142 and GPCR135. In SK-N-MC/␤-gal cells, an increase in cAMP concentration is correlated with increased levels of ␤-galactosidase expression. Forskolin stimulates intracellular cAMP accumulation and induces ␤-galactosidase expression in SK-N-MC/␤-gal cells. Activation of GPCR135 or GPCR142 leads to inhibition of forskolin-induced ␤-galactosidase expression. INSL5 did not inhibit forskolin-stimulated ␤-galactosidase expression in GPCR135-expressing cells (Fig.  4C) but potently inhibited forskolin-stimulated ␤-galactosidase expression in GPCR142-expressing cells in a dose-dependent manner with an EC 50 value of 1.2 Ϯ 0.1 nM (Fig. 4D). In the same experiment, relaxin-3 potently inhibited forskolin-stimulated ␤-galactosidase expression in both GPCR135-and GPCR142-expressing cells with an EC 50 value of 0.3 Ϯ 0.05 and 0.9 Ϯ 0.1 nM, respectively. In the Ca 2ϩ mobilization assays, while relaxin-3 stimulated Ca 2ϩ mobilization in HEK293 cells co-expressing G␣ 16 and either GPCR135 (Fig. 4E) or GPCR142 (Fig. 5F), INSL5 stimulated Ca 2ϩ response in GPCR142-expressing cells (Fig. 4F) but not in GPCR135-expressing cells (Fig. 4E). The agonist activity of INSL5 was also tested in SK-N-MC/␤-gal cells expressing LGR7 or LGR8. Whereas relaxin-3 or INSL3, as positive controls, potently stimulated ␤-galactosidase activity in LGR7-or LGR8-expressing cells, respectively, INSL5 did not demonstrate any detectable activity for cells expressing either LGR7 or LGR8 (data not shown).
INSL5 Is a Weak Antagonist for GPCR135-Although INSL5 binds GPCR135 with low potency, it shows no agonist activity for GPCR135 in the bio-functional assays. Therefore, INSL5 was tested for antagonist activity at GPCR135. To assess the antagonism of INSL5 for GPCR135, relaxin-3-stimulated inhibition of cAMP accumulation was measured in the presence of INSL5. The results demonstrated that INSL5 (1 M) produces a rightward shift in the relaxin-3 dose-response curve (Fig. 5). However, no significant antagonist activity of INSL5 was observed when lower concentrations (i.e. 1, 10, and 100 nM) of INSL5 were used (data not shown).

Molecular Characterization of INSL5 mRNA Expression in Different Human
Tissues-INSL5 mRNA has been reported to be expressed in the human uterus and colon but not in the brain by Northern blot (8). Because GPCR142 mRNA has been detected in the human brain (24), the human INSL5 mRNA expression in human brain alongside 17 other human tissues was investigated using quantitative reverse transcription-PCR, which is more sensitive than a Northern blot. The results indicate that INSL5 mRNA is detected in human brain, fetal brain, kidney, prostate, ovary, thymus, bone marrow, placenta, spleen, and heart in addition to colon (Fig. 6). Among those tissues expressing INSL5 mRNA, all tissues express detectable GPCR142 mRNA. The colon is the tissue that expresses both high levels of INSL5 and GPCR142, which is consistent with previous reports (8,24). In fetal brain and pituitary, high levels of INSL5 mRNA expression were detected; however, the GPCR142 mRNA expression was barely detectable. DISCUSSION An amino acid sequence comparison of the B-chain sequence from different members of the insulin/relaxin family indicated LGR8 (29 -32). Insulin and IGFs do not have those positively charged amino acids at the two positions, and they signal through single trans-membrane/tyrosine kinase receptors, namely the insulin receptor and IGF receptor (16,17). Hsu (33) predicted  that relaxin, INSL3, relaxin-3, INSL4, INSL5, and INSL6 should all be ligands for LGR7 and LGR8 or related receptors, which we think is a very rational prediction. Relaxin, INSL3, and relaxin-3 have been found to either activate LGR7 or LGR8, which are hormone receptor GPCRs. More recently, relaxin-3 has been identified as a ligand for GPCR135 and GPCR142, which are typical peptide-like GPCRs. Although GPCR135 and GPCR142 are not in the exact same sub-family as LGR7 and LGR8, they are still in the type I GPCR family. The identification of INSL5 as an additional ligand for GPCR142 adds further support for Hsu's hypothesis. Although the receptors for INSL4 and INSL6 remain unknown, we predict that they may belong to the GPCR family as well.
INSL5 is the closest member to relaxin-3 in the insulin/ relaxin family. Because relaxin-3 is a ligand for GPCR135, GPCR142, and LGR7, one might expect that INSL5 is an additional ligand for those receptors as well. Previously, we reported that conditioned medium from COS-7 cells expressing INSL5 cDNA did not activate GPCR135, LGR7, or LGR8 (14). In this report, we show that purified INSL5 recombinant peptide activates GPCR142, but not GPCR135, LGR7, or LGR8. However, in the functional assay, INSL5 antagonizes GPCR135 at relatively high concentrations (1 M). Because both relaxin-3 and GPCR135 are predominantly expressed in the brain, it is likely that relaxin-3 is the endogenous ligand for GPCR135. The fact that both INSL5 and GPCR142 are expressed in the brain as well as in the peripheral tissues, particularly in the colon, coupled with their high affinity interaction, strongly suggests that INSL5 is an endogenous ligand for GPCR142. We reported previously that GPCR142 is a pseudo- gene in the rat (24,27). We searched the rat genome for rat INSL5 gene and found a genomic sequence (GenBank TM accession number NW_047717.1) that is highly homologous to the mouse INSL5 cDNA sequence. However, that rat genomic sequence only contains the putative coding regions for the signal peptide and the A-chain, whereas the coding region for the B-chain is missing. Therefore, the rat INSL5 gene will not encode a functional peptide and we conclude that rat INSL5 gene is a pseudogene as well. Most interestingly, the recently completed dog genome contains a sequence highly homologous to the human INSL5 peptide (GenBank TM accession number NW_ AAEX01024390.1). However, this sequence does not encode an open reading frame, with two stop codons present in the C peptide sequence and one upstream of a homologous B-chain sequence. Furthermore, there was no recognizable signal peptide or methionine start codon upstream of this B-chain sequence, and it was missing the critical Cys-29 in the B-chain. Hence the dog INSL5 is nonfunctional. Searches of the dog genome failed to find any trace of a dog GPCR142 ortholog, although a dog GPCR135 ortholog was found. Further phylogenetic analysis shows that INSL5 and GPCR142 emerged after duplications of relaxin-3 and GPCR135, respectively. Given that interacting partners typically co-evolve (34), and both INSL5 and GPCR142 have similar expression profiles, these findings strongly suggest that GPCR142 is the INSL5 receptor. However, our results do not exclude that GPCR142 could still be an endogenous receptor for relaxin-3. Because relaxin-3, INSL5, GPCR135, and GPCR142 are all expressed in brain and testis (8,10,14,24), it is possible that in these tissues both relaxin-3 and INSL5 are ligands for GPCR142, whereas relaxin-3 is the ligand for GPCR135. Detailed anatomical mapping of those ligands and receptors in the brain and testis is required to elucidate the natural relations of these ligand/receptor pairs. In vivo studies using specific ligand(s) for either GPCR135 or GPCR142 will help answer those questions. Additionally, ligand/receptor knockout studies will also provide valuable information to sort out this complicated ligand receptor pairing system. Further analysis of the expression profile of INSL5 indicates that there are tissues (i.e. fetal brain and pituitary) expressing high levels of INSL5 but very low or no levels of GPCR142 (Fig. 6), which opens the possibility that there are additional receptor(s) for INSL5.
When we made the recombinant relaxin-3 (14), we knew the cleavage sites of the relaxin-3 pre-propeptides because we purified the endogenous relaxin-3 from the pig and identified the termini for the A-chain and B-chain. When we were trying to produce the recombinant INSL5, we were confident about the cleavage site between the B-chain and the C-chain because there is an RWRR sequence at the junction, which is a preferred furin site (27) and is conserved between relaxin-3 and INSL5. However, the cleavage site for the signal peptidase and the junction site for the C-chain and the A-chain remained unclear. Although those two sites have been predicted (8), it is necessary to prove the actual sites in native conditions. By expression and analysis of the C-and the N-terminally tagged INSL5 fusion peptides, we were able to confirm the cleavage sites used for INSL5 pre-propeptide maturation. It is worth noting that when we were trying to remove the N-terminal FLAG tag using enterokinase, the FLAG-INSL5 is partially resistant to enterokinase cleavage. Longer times (24 h) and higher concentrations of enterokinase (1 unit of enterokinase for every 10 g of INSL5 peptide) were required to remove the FLAG tag from the FLAG-INSL5 fusion peptide, which we did not experience when we produced the recombinant FLAG-relaxin-3 and other relaxin-related peptides in the same way.
Although recent data suggest that the INSL3/LGR8 and relaxin/LGR7 systems are independent in rodents, the high expression of relaxin receptors in the brain (25,(35)(36)(37), and actions of relaxin peptides on the brain, suggests that the relaxin-3 biology may be more complex. Updated K d , K i , and EC 50 values of relaxin-2, relaxin-3, INSL3, and INSL5 to  GPCR135, GPCR142, LGR7, and LGR8, respectively, are presented to compare the affinities of different ligands to different receptors (Table I). The receptors for INSL4 and INSL6 are still to be identified but are predicted to be GPCRs, and their recent emergence from a relaxin, as opposed to relaxin-3 ancestor, suggests they should interact with LGR-like receptors (33).
In summary, we have identified INSL5 as an additional ligand for GPCR142. Among GPCR135, GPCR142, LGR7, and LGR8, INSL5 only activates GPCR142. The partially overlapping tissue expression profiles of INSL5 and GPCR142, combined with their high affinity interaction, strongly suggest that INSL5 is the endogenous ligand for GPCR142. Our finding has added new infusion to the fast growing relaxin-related research field and will help us to better understand the complicated signaling system for relaxin and relaxin-related peptides.