Is myeloid-derived growth factor a ligand of the sphingosine-1-phosphate receptor 2?

Secretory myeloid-derived growth factor (MYDGF) exerts beneficial effects on organ repair, probably via a plasma membrane receptor; however, the identity of the expected receptor has remained elusive. In a recent study, MYDGF was reported as an agonist of the sphingosine-1-phosphate receptor 2 (S1PR2), an A-class G protein-coupled receptor that mediates the functions of the signaling lipid, sphingosine-1-phosphate (S1P). In the present study, we conducted living cell-based functional assays to test whether S1PR2 is a receptor for MYDGF. In the NanoLuc Binary Technology (NanoBiT)-based ß-arrestin recruitment assay and the cAMP-response element (CRE)-controlled NanoLuc reporter assay, S1P could efficiently activate human S1PR2 overexpressed in human embryonic kidney (HEK) 293T cells; however, recombinant human MYDGF, overexpressed either from Escherichia coli or HEK293 cells, had no detectable effect. Thus, the results demonstrated that human MYDGF is not a ligand of human S1PR2. Considering the high conservation of MYDGF and S1PR2 in evolution, MYDGF is also probably not a ligand of S1PR2 in other vertebrates.

The ghrelin receptor GHSR has two efficient agonists in the lobe-finned fish Latimeria chalumnae.

The peptide hormone ghrelin (an agonist) and LEAP2 (an antagonist) play important functions in energy metabolism via their receptor GHSR, an A-class G protein-coupled receptor. Ghrelin, LEAP2, and GHSR are widely present from fishes to mammals. However, our recent study suggested that fish GHSRs have different binding properties to ghrelin: a GHSR from the lobe-finned fish Latimeria chalumnae (coelacanth) is efficiently activated by ghrelin, but GHSRs from the ray-finned fish Danio rerio (zebrafish) and Larimichthys crocea (large yellow croaker) have lost binding to ghrelin. Do fish GHSRs use another peptide as their agonist? In the present study we tested to two fish motilins from D. rerio and L. chalumnae because motilin is distantly related to ghrelin. In ligand binding and activation assays, the fish GHSRs from D. rerio and L. crocea displayed no detectable or very low binding to all tested motilins; however, the fish GHSR from L. chalumnae bound to its motilin with high affinity and was efficiently activated by it. Therefore, it seemed that motilin is not a ligand for GHSR in the ray-finned fish D. rerio and L. crocea, but is an efficient agonist for GHSR in the lobe-finned fish L. chalumnae, one of the closest fish relatives of tetrapods. The results of present study suggested that GHSR might have two efficient agonists, ghrelin and motilin, in ancient fishes; however, this feature might be only preserved in some extant fishes with ancient evolutionary origins.

Nanomolar range of FAM237B can activate receptor GPR83.

Our recent study confirmed that the mature neuropeptide FAM237A, also known as neurosecretory protein GL (NPGL), is an efficient agonist for GPR83. The paralog FAM237B was previously reported as a weak agonist for GPR83. In the present study, we prepared mature human FAM237B via an intein-fusion approach and demonstrated that it could cause a significant activation effect at the nanomolar range (1‒10 nM) in a NanoBiT-based ß-arrestin recruitment assay. Thus, FAM237B appears to be another endogenous agonist for GPR83 and future in vivo studies will be required to confirm this.

Unusual orthologs shed new light on the binding mechanism of ghrelin to its receptor GHSR1a.

The gastric peptide ghrelin has important functions in energy metabolism and cellular homeostasis by activating growth hormone secretagogue receptor type 1a (GHSR1a). The N-terminal residues of ghrelin orthologs from all vertebrates are quite conserved; however, in orthologs from Cavia porcellus and Phyllostomus discolor, Ser2 and Leu5 are replaced by a smaller Ala and a positively charged Arg, respectively. In the present study, we first demonstrated that the hydrophobic Leu5 is essential for the function of human ghrelin, because Ala replacement caused an approximately 100-fold decrease in activity. However, replacement of Leu5 by an Arg residue caused much less disruption; further replacement of Ser2 by Ala almost restored full activity, although the [S2A] mutation itself showed slight detriments, implying that the positively charged Arg5 in the [S2A,L5R] mutant might form alternative interactions with certain receptor residues to compensate for the loss of the essential Leu5. To identify the responsible receptor residues, we screened GHSR1a mutants in which all conserved negatively charged residues in the extracellular regions and all aromatic residues in the ligand-binding pocket were mutated separately. According to the decrease in selectivity of the mutant receptors towards [S2A,L5R]ghrelin, we deduced that the positively charged Arg5 of the ghrelin mutant primarily interacts with the essential aromatic Phe286 at the extracellular end of the sixth transmembrane domain of GHSR1a by forming cation-π and π-π interactions. The present study provided new insights into the binding mechanism of ghrelin with its receptor, and thus would facilitate the design of novel ligands for GHSR1a.

LEAP2 has antagonized the ghrelin receptor GHSR1a since its emergence in ancient fish.

Recent studies have demonstrated that liver-expressed antimicrobial peptide 2 (LEAP2) antagonizes the ghrelin receptor GHSR1a in mammals. However, its antagonistic function in lower vertebrates has not yet been tested. LEAP2 orthologs have been identified from a variety of fish species; however, previous studies all focused on their antimicrobial activity. To test whether LEAP2 functions as a GHSR1a antagonist in the lowest vertebrates, we studied the antagonism of a fish LEAP2 from Latimeria chalumnae, an extant coelacanth that is one of the closest living fish relatives of tetrapods. Using binding assays, we demonstrated that the coelacanth LEAP2 and ghrelin bound to the coelacanth GHSR1a with IC50 values in the nanomolar range. Using activation assays, we demonstrated that the coelacanth ghrelin activated the coelacanth GHSR1a with an EC50 value in the nanomolar range, and this activation effect was efficiently antagonized by a nanomolar range of the coelacanth LEAP2. In addition, we also showed that the human LEAP2 and ghrelin were as effective as their coelacanth orthologs towards the coelacanth GHSR1a; however, the coelacanth peptides had moderately lower activity towards the human GHSR1a. Thus, LEAP2 serves as an endogenous antagonist of the ghrelin receptor GHSR1a in coelacanth and the ghrelin-LEAP2-GHSR1a system has evolved slowly since its emergence in ancient fish.

Identifying key residues and key interactions for the binding of LEAP2 to receptor GHSR1a.

Liver-expressed antimicrobial peptide 2 (LEAP2) was recently identified as a competitive antagonist for the G protein-coupled receptor GHSR1a, the cognate receptor for the gastric peptide ghrelin. LEAP2 plays important functions in energy metabolism by tuning the ghrelin-GHSR1a system. However, the molecular mechanism by which LEAP2 binds to GHSR1a is largely unknown. In the present study, we first conducted alanine-scanning mutagenesis on the N-terminal fragment of human LEAP2 and demonstrated that the positively charged Arg6 and the aromatic Phe4 are essential for LEAP2 binding to GHSR1a. To identify the receptor residues interacting with the essential Arg6 and Phe4 of LEAP2, we conducted extensive site-directed mutagenesis on GHSR1a. After all conserved negatively charged residues in the extracellular regions of human GHSR1a were mutated, only mutation of Asp99 caused much more detriments to GHSR1a binding to LEAP2 than binding to ghrelin, suggesting that the absolutely conserved Asp99 of GHSR1a probably interacts with the essential Arg6 of LEAP2. After five conserved Phe residues in the predicted ligand-binding pocket of human GHSR1a were mutated, three of them were identified as important for GHSR1a binding to LEAP2. According to a structural model of GHSR1a, we deduced that the adjacent Phe279 and Phe312 might interact with the essential Phe4 of LEAP2, while Phe119 might interact with the aromatic Trp5 of LEAP2. The present study provided new insights into the interaction of LEAP2 with its receptor, and would facilitate the design of novel ligands for GHSR1a in future studies.

Functionality of an absolutely conserved glycine residue in the chimeric relaxin family peptide R3/I5.

The insulin superfamily is a group of homologous proteins that are further divided into the insulin family and relaxin family according to their distinct receptors. All insulin superfamily members contain three absolutely conserved disulfide linkages and a nonchiral Gly residue immediately following the first B-chain cysteine. The functionality of this conserved Gly residue in the insulin family has been studied by replacing it with natural L-amino acids or the corresponding unnatural D-amino acids. However, such analysis has not been conducted on relaxin family members. In the present study, we conducted chiral mutagenesis on the conserved B11Gly of the chimeric relaxin family peptide R3/I5, which is an efficient agonist for receptor RXFP3 and RXFP4. Similar to the effects on insulin family foldability, L-Ala or L-Ser substitution completely abolished the in vitro refolding of a recombinant R3/I5 precursor; whereas, D-Ala or D-Ser substitution had no detrimental effect on refolding of a semi-synthetic R3/I5 precursor, suggesting that the conserved Gly residue controls the foldability of relaxin family members. In contrast to the effect on insulin family activity, D-Ala or D-Ser replacement had no detrimental effect on the binding and activation potencies of the mature R3/I5 towards both RXFP3 and RXFP4, suggesting that the conserved Gly residue is irrelevant to the relaxin family's activity. The present study revealed functionality of the conserved B-chain Gly residue for a relaxin family peptide for the first time, providing an overview of its contribution to foldability and activity of the insulin superfamily.

Cholesterol modulates the binding properties of human relaxin family peptide receptor 3 with its ligands.

Relaxin family peptide receptor 3 (RXFP3) is implicated in the regulation of food intake and stress response upon activation by its cognate agonist relaxin-3. As an A-class G protein-coupled receptor, RXFP3 is an integral plasma membrane protein with seven transmembrane domains, yet influence of the membrane lipids on its function remains unknown. In the present study, we disclosed that cholesterol, an essential membrane lipid for mammalian cells, modulated the binding properties of human RXFP3 with its ligands. We first demonstrated that depletion of cholesterol from host human embryonic kidney (HEK) 293T cells by methyl-ß-cyclodextrin altered ligand-binding properties of the overexpressed human RXFP3, such as increasing its binding potency with some antagonists and decreasing its binding affinity with a NanoLuc-conjugated R3/I5 tracer. Thereafter, we demonstrated that two B-chain residues, B5Tyr and B12Arg, were primarily responsible for the increased binding potency of these antagonists with human RXFP3 under the cholesterol depletion condition. Our results suggest that cell membrane cholesterol interacts with human RXFP3 and modulates its ligand-binding properties, providing new insights into the influence of membrane lipids on RXFP3 function.

Development of a novel ligand binding assay for relaxin family peptide receptor 3 and 4 using NanoLuc complementation.

Relaxin family peptides perform a variety of biological functions by binding and activating relaxin family peptide receptor 1-4 (RXFP1-4), four A-class G protein-coupled receptors. In the present work, we developed a novel ligand binding assay for RXFP3 and RXFP4 based on NanoLuc complementation technology (NanoBiT). A synthetic ligation version of the low-affinity small complementation tag (SmBiT) was efficiently ligated to the A-chain N terminus of recombinant chimeric agonist R3/I5 using recombinant circular sortase A. After the ligation product R3/I5-SmBiT was mixed with human RXFP3 or RXFP4 genetically fused with a secretory large NanoLuc fragment (sLgBiT) at the N terminus, NanoLuc complementation was induced by high-affinity ligand-receptor binding. Binding kinetics and affinities of R3/I5-SmBiT with sLgBiT-fused RXFP3 and RXFP4 were conveniently measured according to the complementation-induced bioluminescence. Using R3/I5-SmBiT and the sLgBiT-fused receptor as a complementation pair, binding potencies of various ligands with RXFP3 and RXFP4 were quantitatively measured without the cumbersome washing step. The novel NanoBiT-based ligand binding assay is convenient for use and suitable for automation, thus will facilitate interaction studies of RXFP3 and RXFP4 with ligands in future. This assay can also be applied to some other plasma membrane receptors for pharmacological characterization of ligands in future studies.

Interaction mechanism of insulin-like peptide 5 with relaxin family peptide receptor 4.

Insulin-like peptide 5 (INSL5) is a gut peptide hormone belonging to the insulin/relaxin superfamily. It is implicated in the regulation of food intake and glucose homeostasis by activating relaxin family peptide receptor 4 (RXFP4). Previous studies have suggested that the B-chain is important for INSL5 activity against RXFP4. However, functionalities of the B-chain residues have not yet been systematically studied. In the present work, we conducted alanine-scanning mutagenesis of the B-chain residues of human INSL5 to obtain an overview of their contributions. Binding and activation assays of these INSL5 mutants with human RXFP4 identified two essential exposed B-chain C-terminal residues (B23Arg and B24Trp) and one important exposed central B-chain residue (B16Ile). These three determinant residues together with the C-terminal carboxylate moiety probably constitute a central receptor-binding patch that forms critical hydrophobic and electrostatic interactions with RXFP4 during INSL5 binding. Some other exposed residues, including B10Glu, B12Ile, B13Arg, B17Tyr, B21Ser, and B22Ser, made minor contributions to INSL5 function. These auxiliary residues are scattered around the edge of the central receptor-binding patch, and thus form a peripheral receptor-binding patch on the surface of INSL5. Our present work provides new insights into the interaction mechanism of INSL5 with its receptor RXFP4.

A novel BRET-based binding assay for interaction studies of relaxin family peptide receptor 3 with its ligands.

Relaxin family peptide receptor 3 (RXFP3) is an A-class G protein-coupled receptor that is implicated in the regulation of food intake and stress response upon activation by its cognate agonist relaxin-3. To study its interaction with various ligands, we developed a novel bioluminescence resonance energy transfer (BRET)-based binding assay using the brightest NanoLuc as an energy donor and a newly developed cyan-excitable orange fluorescent protein (CyOFP) as an energy acceptor. An engineered CyOFP without intrinsic cysteine residues but with an introduced cysteine at the C-terminus was overexpressed in Escherichia coli and chemically conjugated to the A-chain N-terminus of an easily labeled chimeric R3/I5 peptide via an intermolecular disulfide linkage. After the CyOFP-conjugated R3/I5 bound to a shortened human RXFP3 (removal of 33 N-terminal residues) fused with the NanoLuc reporter at the N-terminus, high BRET signals were detected. Saturation binding and real-time binding assays demonstrated that this BRET pair retained high binding affinity with fast association/dissociation. Using this BRET pair, binding potencies of various ligands with RXFP3 were conveniently measured through competition binding assays. Thus, the novel BRET-based binding assay facilitates interaction studies of RXFP3 with various ligands. The engineered CyOFP without intrinsic cysteine residues may also be applied to other BRET-based binding assays in future studies.

Rapid preparation of bioluminescent tracers for relaxin family peptides using sortase-catalysed ligation.

Relaxin family is a group of peptide hormones with a variety of biological functions by activating G protein-coupled receptors RXFP1-4. We recently developed bioluminescent tracers for their receptor-binding assays by chemical conjugation with the ultrasensitive NanoLuc reporter. To simplify preparation of the bioluminescent tracers, in the present study, we established a sortase-catalysed ligation approach using the chimeric R3/I5 as a model. Following catalysis by recombinant sortase A, a NanoLuc reporter carrying the LPETG sortase recognition motif at the C-terminus was efficiently ligated to an R3/I5 peptide carrying four successive Gly residues at the A-chain N-terminus, via the formation of a peptide bond between the C-terminal LPET sequence of NanoLuc and the A-chain N-terminal Gly residue of R3/I5. Saturation binding assays demonstrated that the NanoLuc-ligated R3/I5 retained high binding affinity to RXFP3 and RXFP4, with the calculated dissociation constants (K d) of 4.34 ± 0.33 nM (n = 3) and 5.66 ± 0.54 nM (n = 3), respectively. Using the NanoLuc-ligated R3/I5 as a tracer in competition binding assays, binding potencies of various ligands towards RXFP3 and RXFP4 were conveniently quantified. This work provides a simple method for rapid preparation of bioluminescent tracers for relaxin family peptides and other protein/peptide hormones for ligand-receptor interaction studies.

A negatively charged transmembrane aspartate residue controls activation of the relaxin-3 receptor RXFP3.

Relaxin-3 is an insulin/relaxin superfamily neuropeptide involved in the regulation of food intake and stress response via activation of its cognate receptor RXFP3, an A-class G protein-coupled receptor (GPCR). In recent studies, a highly conserved ExxxD motif essential for binding of relaxin-3 has been identified at extracellular end of the second transmembrane domain (TMD2) of RXFP3. For most of the A-class GPCRs, a highly conserved negatively charged Asp residue (Asp(2.50) using Ballesteros-Weinstein numbering and Asp128 in human RXFP3) is present at the middle of TMD2. To elucidate function of the conserved transmembrane Asp128, in the present work we replaced it with other residues and the resultant RXFP3 mutants all retained quite high ligand-binding potency, but their activation and agonist-induced internalization were abolished or drastically decreased. Thus, the negatively charged transmembrane Asp128 controlled transduction of agonist-binding information from the extracellular region to the intracellular region through maintaining RXFP3 in a metastable state for efficient conformational change induced by binding of an agonist.

Identification of hydrophobic interactions between relaxin-3 and its receptor RXFP3: implication for a conformational change in the B-chain C-terminus during receptor binding.

Relaxin-3 is an insulin/relaxin superfamily neuropeptide implicated in the regulation of food intake and stress response via activation of the G protein-coupled receptor RXFP3. Their electrostatic interactions have been recently identified, and involves three positively charged B-chain residues (B12Arg, B16Arg, and B26Arg) of relaxin-3 and two negatively charged residues (Glu141 and Asp145) in a highly conserved ExxxD motif at the extracellular end of the second transmembrane domain of RXFP3. To investigate their hydrophobic interactions, in the present work we deleted the highly conserved B-chain C-terminal B27Trp residue of relaxin-3, and mutated four highly conserved aromatic residues (Phe137, Trp138, Phe146, and Trp148) around the ExxxD motif of RXFP3. The resultant [∆B27W]relaxin-3 exhibited approximately tenfold lower binding potency and ~1000-fold lower activation potency towards wild-type RXFP3, confirming its importance for relaxin-3 function. Although the RXFP3 mutants could be normally trafficked to cell membrane, they had quite different activities. [F137A]RXFP3 could normally distinguish wild-type relaxin-3 and [∆B27W]relaxin-3 in binding and activation assays, whereas [W138A]RXFP3 lost most of this capability, suggesting that the Trp138 residue of RXFP3 forms hydrophobic interactions with the B27Trp residue of relaxin-3. The hydrophobic Trp138 residue and the formerly identified negatively charged Glu141 and Asp145 residues in the highly conserved WxxExxxD motif may thus form a functional surface that is important for interaction with relaxin-3. We hypothesize that the relaxin-3 B-chain C-terminus changes from the original folding-back conformation to an extended conformation during binding with RXFP3, to allow its B27Trp and B26Arg residues to interact with the Trp138 and Glu141 residues of RXFP3, respectively.

Application of the novel bioluminescent ligand-receptor binding assay to relaxin-RXFP1 system for interaction studies.

Relaxin is a prototype of the relaxin family peptide hormones and plays important biological functions by binding and activating the G protein-coupled receptor RXFP1. To study their interactions, in the present work, we applied the newly developed bioluminescent ligand-receptor binding assay to the relaxin-RXFP1 system. First, a fully active easily labeled relaxin, in which three Lys residues of human relaxin-2 were replaced by Arg, was prepared through overexpression of a single-chain precursor in Pichia pastoris and in vitro enzymatic maturation. Thereafter, the B-chain N-terminus of the easily labeled relaxin was chemically cross-linked with a C-terminal cysteine residue of an engineered NanoLuc through a disulfide linkage. Receptor-binding assays demonstrated that the NanoLuc-conjugated relaxin retained high binding affinity with the receptor RXFP1 (K d = 1.11 ± 0.08 nM, n = 3) and was able to sensitively monitor binding of a variety of ligands with RXFP1. Using the novel bioluminescent binding assay, we demonstrated that three highly conserved B-chain Arg residues of relaxin-3 had distinct contributions to binding of the receptor RXFP1. In summary, our present work provides a novel bioluminescent ligand-receptor binding assay for the relaxin-RXFP1 system to facilitate their interaction studies, such as characterization of relaxin analogues or screening novel agonists or antagonists of RXFP1.

Novel bioluminescent receptor-binding assays for peptide hormones: using ghrelin as a model.

Peptide hormones perform important biological functions by binding specific cell membrane receptors. For hormone-receptor interaction studies, receptor-binding assays are widely used. However, conventional receptor-binding assays rely on radioactive tracers that have drawbacks. In recent studies, we established novel non-radioactive receptor-binding assays for some recombinant protein hormones based on the ultrasensitive bioluminescence of a newly developed nanoluciferase (NanoLuc) reporter. In the present work, we extended the novel bioluminescent receptor-binding assay to peptide hormones that have small size and can be conveniently prepared by chemical synthesis. Human ghrelin, a 28-amino acid peptide hormone carrying a special O-fatty acid modification, was used as a model. To prepare a bioluminescent ghrelin tracer, a chemically synthesized ghrelin analog with a unique cysteine residue at the C-terminus was site-specifically conjugated with an engineered NanoLuc with a unique exposed cysteine residue at the C-terminus via a reversible disulfide linkage. The NanoLuc-conjugated ghrelin retained high binding affinity with the ghrelin receptor GHSR1a (K d = 1.14 ± 0.13 nM, n = 3) and was able to sensitively monitor the receptor-binding of various GHSR1a ligands. The novel bioluminescent receptor-binding assay will facilitate the interaction studies of ghrelin with its receptor. We also proposed general procedures for convenient conjugation of other peptide hormones with NanoLuc for novel bioluminescent receptor-binding assays.

Secretory overexpression and isotopic labeling of the chimeric relaxin family peptide R3/I5 in Pichia pastoris.

Relaxin family peptides are a group of peptide hormones with divergent biological functions. Mature relaxin family peptides are typically composed of two polypeptide chains with three disulfide linkages, rendering their preparation a challenging task. In the present study, we established an efficient approach for preparation of the chimeric relaxin family peptide R3/I5 through secretory overexpression in Pichia pastoris and in vitro enzymatic maturation. A designed single-chain R3/I5 precursor containing the B-chain of human relaxin-3 and the A-chain of human INSL5 was overexpressed in PichiaPink strain 1 by high-density fermentation in a two-liter fermenter, and approximately 200 mg of purified precursor was obtained from one liter of the fermentation supernatant. We also developed an economical approach for preparation of the uniformly (15)N-labeled R3/I5 precursor by culturing in shaking flasks, and approximately 15 mg of purified (15)N-labeled precursor was obtained from one liter of the culture supernatant. After purification by cation ion-exchange chromatography and reverse-phase high performance liquid chromatography, the R3/I5 precursor was converted to the mature two-chain form by sequential treatment with endoproteinase Lys-C and carboxypeptidase B. The mature R3/I5 peptide had an α-helix-dominated conformation and retained full receptor-binding and receptor activation activities. Thus, Pichia overexpression was an efficient approach for sample preparation and isotopic labeling of the chimeric R3/I5 peptide. This approach could also be extended to the preparation of other relaxin family peptides in future studies.

Role of ZIC1 methylation in hepatocellular carcinoma and its clinical significance.

Hepatocellular carcinoma (HCC) is one of the most aggressive malignancies in humans, and its prognosis is generally poor even after surgery. The zinc finger of the cerebellum (ZIC1) gene is a novel tumor suppressor gene that plays a crucial role in vertebrate development. Altered expression of ZIC1 is observed in various types of human cancers. The aims of the present study were to investigate the methylation status of ZIC1 in HCC and evaluate its clinical implication. The methylation status of ZIC1 was analyzed in 132 pairs of HCC and corresponding noncancerous tissues by methylation-specific polymerase chain reaction (PCR) (MSP). The expression of ZIC1 messenger RNA (mRNA) in HCC tissues was examined by real-time PCR. Methylation frequency of ZIC1 in HCC was significantly higher than that in the corresponding noncancerous tissues (P < 0.001), and it was correlated with tumor size (P = 0.022), histological differentiation (P = 0.033), and tumor stage (P = 0.009). The downregulation of the ZIC1 mRNA expression in HCC was correlated with the ZIC1 methylation (P < 0.001). The patients with methylated ZIC1 had a poorer overall survival than those without methylated ZIC1 (P < 0.001). Taken together, our results suggested that the hypermethylation may lead to promoter silencing of ZIC1 mRNA and associated with poor survival in HCC. Overall, aberrant methylation is an important mechanism for ZIC1 inactivation in HCC, and ZIC1 methylation may be a promising biomarker for the diagnosis and prognosis of HCC.

Identification of important residues of insulin-like peptide 5 and its receptor RXFP4 for ligand-receptor interactions.

Insulin-like peptide 5 (INSL5) is an insulin/relaxin superfamily peptide involved in the regulation of glucose homeostasis by activating its receptor RXFP4, which can also be activated by relaxin-3 in vitro. To determine the interaction mechanism of INSL5 with its receptor RXFP4, we studied their electrostatic interactions using a charge-exchange mutagenesis approach. First, we identified three negatively charged extracellular residues (Glu100, Asp104 and Glu182) in human RXFP4 that were important for receptor activation by wild-type INSL5. Second, we demonstrated that two positively charged B-chain Arg residues (B13Arg and B23Arg) in human INSL5 were involved in receptor binding and activation. Third, we proposed probable electrostatic interactions between INSL5 and RXFP4: the B-chain central B13Arg of INSL5 interacts with both Asp104 and Glu182 of RXFP4, meanwhile the B-chain C-terminal B23Arg of INSL5 interacts with both Glu100 and Asp104 of RXFP4. The present electrostatic interactions between INSL5 and RXFP4 were similar to our previously identified interactions between relaxin-3 and RXFP4, but had subtle differences that might be caused by the different B-chain C-terminal conformations of relaxin-3 and INSL5 because a dipeptide exchange at the B-chain C-terminus significantly decreased the activity of INSL5 and relaxin-3 to receptor RXFP4.

The highly conserved negatively charged Glu141 and Asp145 of the G-protein-coupled receptor RXFP3 interact with the highly conserved positively charged arginine residues of relaxin-3.

Relaxin-3 is a newly identified insulin/relaxin superfamily peptide that plays a putative role in the regulation of food intake and stress response by activating its cognate G-protein-coupled receptor RXFP3. Relaxin-3 has three highly conserved arginine residues, B12Arg, B16Arg and B26Arg. We speculated that these positively charged arginines may interact with certain negatively charged residues of RXFP3. To test this hypothesis, we first replaced the negatively charged residues in the extracellular domain of RXFP3 with arginine, respectively. Receptor activation assays showed that arginine replacement of Glu141 or Asp145, especially Glu141, significantly decreased the sensitivity of RXFP3 to wild-type relaxin-3. In contrast, arginine replacement of other negatively charged extracellular residues had little effect. Thus, we deduced that Glu141 and Asp145, locating at the extracellular end of the second transmembrane domain, played a critical role in the interaction of RXFP3 with relaxin-3. To identify the ligand residues interacting with the negatively charged EXXXD motif of RXFP3, we replaced the three conserved arginines of relaxin-3 with negatively charged glutamate or aspartate, respectively. The mutant relaxin-3s retained the native structure, but their binding and activation potencies towards wild-type RXFP3 were decreased significantly. The compensatory effects of the mutant relaxin-3s towards mutant RXFP3s suggested two probable interaction pairs during ligand-receptor interaction: Glu141 of RXFP3 interacted with B26Arg of relaxin-3, meanwhile Asp145 of RXFP3 interacted with both B12Arg and B16Arg of relaxin-3. Based on these results, we proposed a relaxin-3/RXFP3 interaction model that shed new light on the interaction mechanism of the relaxin family peptides with their receptors.