The RFamide peptide family is well represented in invertebrates, but so far, only four RFamide prepropeptide precursors have been described in mammals. These include neuropeptide FF (NPFF)/neuropeptide AF (Perry et al., 1997; Vilim et al., 1999), prolactin-releasing peptide (PrRP) (Hinuma et al., 1998), RFamide-related peptides (RFRPs) (Hinuma et al., 2000), and KiSS-1 (Ohtaki et al., 2001).

NPFF, the first RFamide peptide identified in mammals, acts as a modulator of morphine analgesia, tolerance, and dependence, and influences many functions including pain mechanisms (Panula et al., 1996). Recently, two human receptors for NPFF have been identified: hNPFF1 and hNPFF2 (Bonini et al., 2000; Elshourbagy et al., 2000; Hinuma et al., 2000). These receptor subtypes have different tissue localizations in human and rat (Bonini et al., 2000). The recently cloned NPFF2 receptor (named HLWAR77 by Elshourbagy et al., 2000) has been described as G_i/o-coupled when stably expressed in HEK 293 cells (Elshourbagy et al., 2000) or CHO cells (Kotani et al., 2001). Among the receptors with the highest amino acid sequence homology to NPFF1 and NPFF2 are members of the orexin, human neuropeptide Y (NPY), and cholecystokinin family, which have been implicated in feeding (Bonini et al., 2000). BIBP3226, an anorexigenic Y1 receptor ligand, has been shown to bind to the NPFF1 receptor, thus further suggesting a potential role of the NPFF1 receptor in regulation of feeding (Bonini et al., 2000).

PrRP, the second RFamide peptide, was identified as the endogenous ligand for the orphan G protein-coupled receptor (GPCR), hGR3 (Hinuma et al., 1998). The full physiological role of PrRP remains to be elucidated, but it has been envisioned to play a broader role in brain function than originally suggested, and the ability to stimulate prolactin release may not represent its primary biological function (Samson et al., 2000). The hGR3 receptor (named GPR10 by Marchese et al., 1995; and in this paper referred to as the hPrRP receptor) and its rat counterpart, UHR-1, have been shown to couple to at least G_q and G_i but not to G_s in CHO cells (Hinuma et al., 1998, 1999).

Mammalian genes encoding a third class of RFamide peptides, namely hRFRP1 and hRFRP3 and their receptor OT7T022, have recently been reported (Hinuma et al., 2000). OT7T022 is activated by NPFF and corresponds to the NPFF1 receptor characterized by Bonini et al. (2000). RFRPs (more specifically hRFRP1) have been shown to regulate prolactin secretion (Hinuma et al., 2000), but additional physiological functions cannot be ruled out.

KiSS-1 is a metastasis suppressor gene that encodes the fourth class of RFamide peptides (Ohtaki et al., 2001). The gene product of KiSS-1, also known as “metastin,” was found to be the endogenous ligand of an orphan GPCR, hOT7T175 (Ohtaki et al., 2001).

The cellular mechanisms by which NPFF, PrRP, and RFRP exert their functions in vivo are poorly understood. To clarify the cellular and pharmacological actions activated by these RFamide peptides, well characterized, functional in vitro assays on recombinant cell lines and/or cell lines expressing these receptors endogenously need to be developed. In the present report, we have used [³⁵S]guanosine-5′-O-(3-thio)triphosphate ([³⁵S]GTPγS) binding induced by RFamide peptides on the hNPFF2 and hPrRP receptors stably transfected in CHO cells (CHO-hNPFF2 and CHO-hPrRP-R, respectively) to determine agonist efficacies and rank orders of potency. The agonist potencies of RFamide peptides were compared with their affinities obtained in competition binding assays with cell membrane preparations. Receptor autoradiography was conducted to visualize and quantitate the NPFF binding sites in rat cervical spinal cord sections to investigate the binding properties of RFamide peptides at the NPFF receptor in a native environment.

Materials and Methods

Materials. The following natural peptides and analogs were purchased from Bachem (Bubendorf, Switzerland): (1DMe)Y8Fa [DYL(NMe)FQPQRF-NH₂], hPrRP20 (TPDINPAWYASRGIRPVGRF-NH₂), hPrRP31 (SRTHRHSMEIRTPDINPAWYASRGIRPVGRF-NH₂), human neuropeptide Y (hNPY) (YPSKPDNPGEDAPAEDMARYYSALRHYINLITRQRY-NH₂), hNPY18-36 (ARYYSALRHYINLITRQRY-NH2), NPFF (FLFQPQRF-NH₂), and BIBP3226. HPFRP-1 (MPHSFANLPLRF-NH₂), and hPRFP-3 (VPNLPQRF-NH₂) were purchased from Phoenix Pharmaceuticals Inc. (Belmont, CA). NPFF-OH (FLFQPQRF-OH), rat neuropeptide SF (rNPSF) (SLAAPQRF-NH₂), FMRFamide (FMRF-NH₂), hPrRP24–31, hRFRP(6G)-3 (VPNLPGRF-amide), and bPrRP20 (TPDINPAWYAGRGIRPV-GRFamide) were custom synthesized by Peptide Technologies Corp. (Gaithersburg, MD). Substance P was purchased from Peninsula Laboratories (Belmont, CA). Iodinated (1DMe)Y8Fa was custom ordered from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK). The specific activity of ¹²⁵I-(1DMe)Y8Fa was 2000 Ci/mmol, as indicated by the product specification sheet. [³⁵S]GTPγS was purchased from PerkinElmer Life Sciences (Boston, MA). The specific activity of [³⁵S]GTPγS was 1250 Ci/mmol.

¹²⁵I-(1DMe)Y8Fa Binding Assays on Rat Spinal Cord Membranes. Rat spinal cord membrane preparations and binding assays were performed using essentially the method for ¹²⁵I-YLFQPQRF-amide (¹²⁵I-Y8Fa) binding as described by Allard et al. (1989) and modified by Payza and Yang (1993). Briefly, membranes (100–160 μg/sample) were incubated with 0.02 to 0.07 nM ¹²⁵I-(1DMe)Y8Fa, and various concentrations of ligands in 50 mM Tris-HCl, pH 7.5 at RT, 60 mM NaCl, 1 mM MgCl₂, 3 μg/ml aprotinin, 7.5 g/l BSA, and 30 μM bestatin. Total binding (TB) and nonspecific binding (NSB) were determined in the absence and presence, respectively, of 1 μM (1DMe)Y8Fa. After 45 min at RT, incubations were terminated by rapid filtration (Tomtec Harvester96; Tomtec Inc., Hamden, CT) through GF/B glass fiber filter mats (presoaked at 4°C in 200 ml of 50 mM Tris-HCl, 60 mM NaCl, 1 mM MgCl₂, 1 g/l BSA, 5 g/l polyethyleneimine, pH 7.4 at 4°C for 45 min). Filters were washed four times with 5 ml of ice-cold wash buffer (50 mM Tris-HCl, 60 mM NaCl, 1 mM MgCl₂, pH 7.4 at 4°C). Specific binding (SB) was calculated as TB - NSB. The SB in the presence of various concentrations of test compounds was expressed as percentage of control SB, which was obtained as the TB in the absence of any competing compound minus NSB. Analysis of competition binding experiments was carried out by nonlinear least squares curve fitting with Hill slopes (n_H) being set to unity (Devillers et al., 1994). Affinity constants (K_i) were calculated from the IC₅₀ values according to the Cheng-Prusoff equation (Cheng and Prusoff, 1973). The K_D for this purpose was 0.19 nM, as determined in a pilot study.

¹²⁵I-(1DMe)Y8Fa Binding Assay on CHO-hNPFF2 Membranes. Recombinant CHO cells expressing the hNPFF2 receptor (Euroscreen S.A., Brussels, Belgium) were grown in Ham's F12 medium (Nutrient Mixture Ham's F12; Invitrogen, Glasgow, UK) supplemented with 10% fetal bovine calf serum and 400 μg/ml G418. Cells were harvested in phosphate-buffered saline and frozen at -70°C. To prepare membranes, thawed cell pellets were homogenized on ice in a Potter-Elvehjem homogenizer in 10 mM Tris-HCl, 0.1 mM EDTA, pH 7.5 at RT supplemented with 320 mM sucrose. The nuclear pellet obtained by centrifugation of the homogenate at 1,000g for 15 min at 4°C was discarded. From the supernatant, a membrane fraction was collected by centrifugation at 48,000g_max for 30 min at 4°C. The 48,000g pellet was re-suspended in 10 mM Tris-HCl, 0.1 mM EDTA, pH 7.5 at RT without sucrose and centrifuged again at 48,000g_max for 30 min. Binding of ¹²⁵I-(1DMe)Y8Fa to membranes (2.5 μg/sample) of hNPFF2 receptor expressing CHO cells and analysis of the binding experiments were carried out as described above for rat spinal cord membranes. The K_D used in the analysis of the competition binding experiments was 0.10 nM, as determined in a pilot study.

¹²⁵I-PrRP Binding Assay on CHO-hPrRP-R Membranes. Recombinant CHO cells expressing the human PrRP (hPrRP) receptor hGR3/GPR10 (Euroscreen S.A.) were grown, and membranes from these cells were prepared by the same method as described above for hNPFF2 expressing CHO cells.

¹²⁵I-PrRP20 competition binding experiments were performed by incubating membranes (0.75 μg/sample) with 0.02 to 0.05 nM iodinated peptide and various concentrations of ligands in 25 mM Hepes, pH 7.4 at RT, 1 mM CaCl₂, 5 mM MgCl₂, 3 μg/ml aprotinin, 30 μM bestatin, and 7.5 mg/ml BSA. NSB was determined in the presence of 1 μM hPrRP20. Incubations were terminated after 45 min at RT as described above for rat spinal cord membranes except that 25 mM Hepes, 1 mM CaCl₂, 5 mM MgCl₂, and 0.5 M NaCl, pH 7.4, at 4°C was used as the washing buffer. The analysis of the binding data was also carried out as described above using a K_D value of 0.3 nM, as determined in saturation binding experiments.

[³⁵S]GTPγS Binding Assay. The agonist activity of various ligands was determined by their ability to stimulate the receptor-mediated binding of [³⁵S]GTPγS to G proteins in membranes of CHO-hNPFF2 or CHO-hPrRP-R cells. Membranes (2 and 10 μg/sample of CHO-hNPFF2 and CHO-hPrRP-R cells, respectively) were incubated in 50 mM Tris-HCl, 5 mM MgCl₂, 1 mM dithiothreitol, 1 mM EDTA, 1 μM GDP, 20 mM NaCl (CHO-hNPFF2 assay), or 100 mM NaCl (CHO-hPrRP-R assay), pH 7.4 at RT with 6 to 12 concentrations of test ligands and a tracer concentration of [³⁵S]GTPγS (0.07–0.16 nM). After a 60-min incubation at RT (30-min preincubation without label followed by a 30-min stimulation after addition of label in CHO-hNPFF2 assay or 60 min stimulation in CHO-hPrRP-R assay without preincubation), the reaction was terminated by rapid vacuum filtration through glass fiber filters. Filters were washed four times with 5 ml of ice-cold wash buffer (20 mM Tris-HCl, 5 mM MgCl₂, 1 mM EDTA, pH 7.4 at RT), dried, and counted for radioactivity in a scintillation counter. Experimental results were calculated with nonlinear least squares curve fitting. Agonist effects were normalized against the stimulation obtained with reference compounds (1DMe)Y8Fa, in the case of the hNPFF2 receptor, and hPrRP20, in the case of the hPrRP receptor. The response of the reference agonists was set as 100%. Experiments were repeated at least three times, unless indicated otherwise.

Quantitative Receptor Autoradiography. Autoradiography was performed on rat spinal cord sections with ¹²⁵I-(1DMe)Y8Fa and was carried out essentially as described earlier for ¹²⁵I-Y8Fa (Allard et al., 1992). Coronal 20-μm cryosections at the cervical level of the rat spinal cord were collected onto poly-L-lysine-coated slides (Menzel-Gläser; Merck, Darmstadt, Germany) and stored dry at -70°C. The mounted sections were rehydrated in 50 mM Tris-HCl (pH 7.5), 140 mM NaCl and 0.5% BSA for 20 min at RT. The slides were washed twice with ice-cold 50 mM Tris-HCl (pH 7.5) for 2 min before the incubation with 0.05 nM ¹²⁵I-(1DMe)Y8Fa in 50 mM Tris-HCl (pH 7.5), 120 mM NaCl, 0.5% BSA, 0.1 mM bestatin, 1 mM EDTA, and 3 μM aprotinin. Displacement of ¹²⁵I-(1DMe)Y8Fa was analyzed by including unlabeled competing agents at the following concentrations: rNPSF (500 nM, 5 nM), NPFF (20 nM, 0.2 nM), (1DMe)Y8Fa (20 nM, 0.2 nM), bPrRP20 (200 nM, 2 nM), hRFRP-1 (1 μM, 10 nM), hRFRP(6G)-3 (1 μM, 10 nM), BIBP3226 (1 μM, 10 nM), and substance P (1 μM). The concentrations of the ligands [except hRFRP(6G)-3 and BIBP3226, which were tested at 10 nM and 1 μM] were chosen to represent approximately the K_i and the 100-fold K_i value (Table 1). After 60 min of incubation at RT, the slides were washed three times for 2 min in ice-cold Tris-HCl (pH 7.5) and were finally dipped in ice-cold distilled water, air-dried, and exposed on film (Biomax MR; Eastman Kodak, Rochester, NY). Autoradiographic films were analyzed using the MCID image analysis system (Imaging Research, St. Catherines, ON, Canada). The optical density of the area of the upper dorsal horn (laminae I–II) was measured and the integrated optical density values were determined based on a calibration curve derived from ¹⁴C standards (American Radiolabeled Chemicals, St. Louis, MO) exposed simultaneously with the samples to the films. The means ± S.E.M. of three animals and three sections in each area of every animal are given as the quantitative assessment of ligand binding activity. Statistical analysis was performed by one-way analysis of variance of grouped data (Newman-Keuls test; MicroComputer Specialists, Wynnewood, PA).

Results

Affinity of RFamide Peptide at NPFF Receptors. The affinities of various ligands for the hNPFF2 receptor expressed in CHO cells and the rat spinal cord NPFF receptor (rNPFF) were determined in competition binding assays with ¹²⁵I-(1DMe)Y8Fa (Table 1). ¹²⁵I-(1DMe)Y8Fa has been reported to have a slightly higher or comparable affinity for the hNPFF2 versus the hNPFF1 receptor (Bonini et al., 2000). In accordance with previous results, the high affinity of NPFF is lost in its nonamidated analog, NPFF-OH (Payza et al., 1993; Mazarguil et al., 2001). rNPSF had a 40-fold and 7-fold lower affinity than NPFF for the rNPFF and hNPFF2 receptors, respectively. hNPY and the C-terminal fragment 18–36 of hNPY (hNPY18-36) had an affinity of about 1 μM for the rNPFF receptor and a slightly lower affinity for the hNPFF2 receptor. BIBP3226, which has been designed to mimic the C terminus of the NPY molecule (Rudolf et al., 1994) bound to the rNPFF and hNPFF2 receptors with affinities of 390 ± 50 and 640 ± 80 nM, respectively.

The human endogenous RFamide peptides hPrRP20, hPrRP31, hRFRP-1, and hRFRP-3 had affinities better than 20 nM for the rNPFF receptor, whereas the affinities of these RFamide peptides for the hNPFF2 receptor were better than 60 nM. The affinity of hPrRP20 was comparable to the affinity obtained for the invertebrate neuropeptide FMRFamide. Interestingly, even a peptide containing only eight C-terminal amino acids of hPrRP31, hPrRP24–31, was able to bind to the NPFF receptors with high affinity.

Agonist Activity of RFamide Peptides at the hNPFF2 Receptor. [³⁵S]GTPγS binding experiments on membranes of CHO-hNPFF2 cells were performed in the presence of 20 mM NaCl and 1 μM GDP. The agonist potencies are directly proportional to the binding affinities (Tables 1 and 2: r = 0.9549, p < 0.0001). The efficacy of NPFF [79 ± 10% compared with (1DMe)Y8Fa] was lower than that for the metabolically stable analog (1DMe)Y8Fa (100% stimulation; Table 2). Interestingly, all other tested RFamide peptides had efficacy approximately similar to or even greater than that of (1DMe)Y8Fa. The most efficacious among the RFamide peptides was hPrRP31, for which an agonist response 1.7-fold that of (1DMe)Y8Fa was obtained (Table 2, Fig. 1A). The dose-dependent increase in [³⁵S]GTPγS binding caused by the RFamide peptides was clearly mediated through hNPFF2 receptors, since membranes from CHO cells not expressing this receptor failed to give rise to any significant increase in [³⁵S]GTPγS binding when challenged with micromolar concentrations of hPrRP20, (1DMe)Y8Fa, hPrRP31, or hRFRP-1 (data not shown). Concerning agonist potencies, the following rank order was observed in the [³⁵S]GTPγS binding assay on the hNPFF2 receptor: (1DMe)Y8Fa≥ NPFF > FMRFamide > hRFRP-1 ≥ hPrRP20 ≥ hPrRP24–31 ≥ hPrRP31 > rNPSF > hRFRP-3. The dose-response curves for NPFF-OH, hNPY, and hNPY18-36 did not reach plateau and did not, therefore, allow for calculation of reliable estimates of agonist potency (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Stimulation of [³⁵S]GTPγS binding on membranes of CHO cells expressing the human NPFF2 receptor (CHO-hNPFF2; A) or the human PrRP receptor (CHO-hPrRP-R; B). [³⁵S]GTPγS (0.07–0.16 nM) was added to membranes of CHO-hNPFF2 cells (2 μg/sample) that had been preincubated with the indicated concentration of (1DMe)Y8Fa (•) NPFF (○), hRFRP-1 (▵), and hPrRP31 (▾) (A) or to membranes of CHO-hPrRP-R cells (5–10 μg/sample) and the indicated concentration of hPrRP24–31 (•) NPFF (○), hRFRP-1 (▵), and hPrRP20 (▾), and hRFRP-3 (X) (B). The percentage of [³⁵S]GTPγS binding was normalized against the maximal effect of the reference compound (1DMe)Y8Fa (in A) or hPrRP20 (in B), which was set as 100%. The combined data of three different experiments performed in duplicate are shown.

Affinities and Activities of RFamide Peptides at the hPrRP Receptor. The affinities of various ligands for the hPrRP receptor hGR3/GPR10 expressed in CHO cells were determined in ¹²⁵I-PrRP competition binding assays. The affinities of the two naturally occurring PrRPs, the full-length peptide hPrRP31 and the 20-amino acid-long C-terminal fragment hPrRP20, at the hPrRP receptor hGR3/GRP10 were 0.68 nM ± 0.06 nM and 1.0 ± 0.3 nM, respectively (Table 3). The C-terminal octapeptide hPrRP24-31 displayed considerably lower affinity, with a K_i of 47 ± 11 nM. This result is in line with the findings published by Roland et al. (1999). In contrast to the generally quite high affinities of RFamide peptides at the NPFF receptor, the hPrRP receptor showed strong discrimination among the related mammalian RFamide peptides; NPFF, hRFRP-1, and hRFRP-3 did not display any noticeable interaction with the hPrRP receptor even in the micromolar range (Table 3). However, the C-terminal fragment of hNPY (hNPY18-36) and BIBP3226 bound to the hPrRP receptor with micromolar affinity.

[³⁵S]GTPγS binding was also performed with membranes of CHO-hPrRP-R cells. Both hPrRP20 and hPrRP31 dose dependently increased [³⁵S]GTPγS binding, and the signal over basal was about 50% for both peptides (Fig. 2). PTX uncouples G_i/o proteins from GPCRs and thereby disrupts the corresponding signaling pathways. Following PTX pretreatment, over 80% of the [³⁵S]GTPγS binding stimulated by hPrRP20 or hPrRP31 and over 50% of the basal [³⁵S]GTPγS binding was abolished (Fig. 2). The PrRP analog hPrRP24-31 showed reduced potency to stimulate [³⁵S]GTPγS binding (EC₅₀ values were 230 ± 140 and 1.7 ± 0.7 nM for hPrRP24–31 and hPrRP20, respectively), which is consistent with an earlier report, where the heptapeptide PrRP25-31 was described to be able to elicit a calcium response albeit with reduced potency compared with PrRP20 (Roland et al., 1999). However, in agreement with the binding results, we did not observe any activation of the hPrRP receptor in response to NPFF, hRFRP-1, or hRFRP-3 (Fig. 1B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Stimulation of [³⁵S]GTPγS binding on membranes of CHO cells expressing the human PrRP receptor (CHO-hPrRP-R), with and without PTX pretreatment of the cells. [³⁵S]GTPγS (0.07–0.16 nM) was added to membranes of transfected cells (10 μg/sample) that already contained hPrRP31 (•, ○) or hPrRP20 (▴, ▵). The membranes were prepared from non-PTX-pretreated control cells (•, ▴) or from cells that had been pretreated with 100 ng/ml PTX for 16 h (○, ▵). The percentage of [³⁵S]GTPγS binding was normalized against the basal binding in control membranes (which was set as 100%). The combined data of three different experiments performed in duplicate are shown.

Quantitative Receptor Autoradiography on Rat Spinal Cord Sections. NPFF binding sites in the superficial layers of the dorsal horn in the rat spinal cord represent mainly NPFF2 receptors (Bonini et al., 2000). A panel of ligands was analyzed by quantitative receptor autoradiography to visualize the displacement of ¹²⁵I-(1DMe)Y8Fa in rat spinal cord sections. The ligands were tested at two concentrations on cryostat sections from three different animals on the level of the cervical spinal cord (See Materials and Methods). Substance P was included as a negative control, because it did not inhibit specific ¹²⁵I-(1DMe)Y8Fa binding in the competition binding assay on rat spinal cord membranes (data not shown). In accordance with previous reports, we observed a high density of binding sites in the superficial layers of the dorsal horn and around the central canal (Allard et al., 1989, 1992). The receptor density around the central canal was approximately 50% of that in the dorsal horn (Fig. 3, A and B). At the highest concentration tested, only substance P did not significantly affect ¹²⁵I-(1DMe)Y8Fa binding, whereas NPFF, (1DMe)Y8Fa, hRFRP-1, rNPSF, bPrRP20, BIBP3226, and hRFRP(6G)-3 competed for the same binding sites as ¹²⁵I-(1DMe)Y8Fa in the superficial layers of the dorsal horn and in the central canal (Fig. 3, A and B). Change of the C-terminal amino acid 6Q to 6G resulted in significant reduction of the binding of hRFRP(6G)-3 in comparison to other RFamides, demonstrating that the C-terminal structure beyond the RFamide motif is also critical for binding. It should be noted that the NPY₁ antagonist BIBP3226 at a concentration of 1 μM displaced ¹²⁵I(1DMe)Y8Fa binding by roughly 50% (as calculated in percentage of total binding in the absence of competing ligand) both in the dorsal horn and around the central canal.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

A, autoradiography of coronal sections of rat cervical spinal cord after 1 week's exposure to film showing binding of 0.05 nM ¹²⁵I-(1DMe)Y8Fa in the absence (total binding) or presence of a competing ligand as indicated in the figure (I,II = laminae I,II of the dorsal horn of the spinal cord; X = laminae X of the spinal cord). B, films (illustrated in A) were quantified by measuring optical densities bilaterally of the superficial layers of the upper dorsal horn (laminae I-II) or central canal of rat spinal cord sections by MCID image analyzer. By using a commercially available standard coexposed with the tissue sections, the values were converted to nanocuries per milligram. Averages ± S.E.M. of six readings were obtained from each rat (n = 3). One-way analysis of variance with the Newman-Keuls test, ★★★, p < 0.001 compared with control [total ¹²⁵I-(1DMe)Y8Fa in the absence of competing ligand]; ★★, p < 0.01 compared with control; ++, p < 0.01 compared with NPFF; and +++, p < 0.001 compared with NPFF.

Discussion

Although the role of NPFF is best defined in pain modulation (Panula et al., 1996; Roumy and Zajac, 1998; Panula et al., 1999; Vilim et al., 1999), NPFF also modulates a variety of other physiological processes such as insulin release, prolactin release, blood pressure, food intake, electrolyte balance, and morphine abstinence syndrome (Panula et al., 1996; Aarnisalo et al., 1997). PrRP, which was initially found to stimulate the release of prolactin from anterior pituitary lactotrophs (Hinuma et al., 1998), has been suggested to have a broader role in the central nervous system, and its ability to stimulate prolactin release might not represent its primary biological function (Samson et al., 2000). RFRPs have also been reported to exhibit regulatory effects on prolactin secretion, although by mechanisms other than PrRP (Hinuma et al., 2000). It therefore seems that RFamide peptides and their receptors may have evolved in close association with regulatory mechanisms for prolactin secretion in mammals (Hinuma et al., 2000). Nonetheless, the possibility that RFRPs might have other, still unknown, important functions cannot be ruled out.

In the present study, we found that the RFamide peptides hPrRP20, hPrRP31, hRFRP-1, and hRFRP-3 bind with high affinity to the rNPFF receptor as well as to the hNPFF2 receptor, suggesting that different RFamide peptides may exert their in vivo functions through the NPFF receptors. In contrast, the hPrRP receptor seems to require an additional motif beyond the C-terminal RFamide, since none of the related mammalian RFamide peptides, NPFF, hRFRP-1, and hRFRP-3, were able to compete with ¹²⁵I-PrRP for binding to recombinant hPrRP receptor. The affinities of hNPY and BIBP3226, a NPY₁ receptor antagonist, at hNPFF2 and rNPFF receptors were moderate. NPY and NPFF possess similar C-terminal sequences containing an arginine and a C-terminally amidated aromatic residue. Despite the only moderate affinities of hNPY and BIPB3226 at hNPFF2 and rNPFF receptors, it is thus conceivable that NPY ligands might cross-react with NPFF receptors, as has already been suggested by Mollereau et al. (2001). BIBP3226 has previously been reported to display 10- to 60-fold higher affinity for the hNPFF1 and rNPFF1 receptors as compared with the hNPFF2 and rNPFF2 receptors (Bonini et al., 2000).

NPFF and NPY, as well as their receptors, share common structural features that could explain the interaction of some NPY ligands with NPFF receptors. 1) the C-terminal sequences of NPY (RYamide) and the RFamide peptides (RF-amide), which are essential for interaction with their receptors, are very similar, with a difference of only a single hydroxyl group; and 2) the receptor sequences are 30 to 35% identical. However, based on the results presented in this paper, PrRP ligands also have to be considered as likely candidates for such cross-reactivity with the NPFF receptor.

Compared to the NPFF receptor, the hPrRP receptor is more clearly capable of differentiating between different RF-amide peptides; none of the tested related mammalian RF-amide peptides were able to bind to the PrRP receptor with affinity comparable to that of the PrRP peptides. Likewise, BIBP3226 and the C-terminal fragment of hNPY bound to the hPrRP receptor with only very low affinity.

We do note that in our assay conditions, the affinities of NPFF and BIBP3226 for the hNPFF2 receptors expressed in CHO cells were clearly lower than the affinities reported by Mollereau et al. (2001) for the same cell line using a novel radioligand, ¹²⁵I-EYF (¹²⁵I-EYWSLAAPQRF-NH₂). This discrepancy might be due to differences in the binding characteristics of ¹²⁵I-EYF and ¹²⁵I-(1DMe)Y8Fa. However, our affinity value for NPFF at the hNPFF2 receptor is in good agreement with earlier reports in which ¹²⁵I-(1DMe)Y8Fa or ¹²⁵I-Y8Fa was used as radioligand (Bonini et al., 2000; Liu et al., 2001).

Our quantitative receptor autoradiography results are in good agreement with the binding assay on rat spinal cord membranes. All tested RFamide peptides, as well as BIBP3226, were able to displace the binding of ¹²⁵I-(1DMe)Y8Fa at concentrations about 100-fold above their estimated K_i value or at 1 μM (hRFRP(6G)-3 and BIBP3226), both in the dorsal horn and around the central canal, where specific binding of ¹²⁵I-(1DMe)Y8Fa has previously been reported (Allard et al., 1989, 1992). The results indicate that NPFF, rNPSF, bPrRP20, and hRFRP-1 are able to compete for the same NPFF binding sites in rat spinal cord. Even though our autoradiographic data do not allow us to discriminate between different NPFF receptor subtypes, earlier reports have indicated that NPFF2 is the predominant receptor subtype in rat spinal cord (Bonini et al., 2000). Expression of rNPFF2 mRNA has also been reported both in the laminae I–II in the dorsal horn and the central canal region of the rat spinal cord (Liu et al., 2001). We cannot, however, exclude the possibility that additional, still uncharacterized, receptor subtypes exist in rat spinal cord.

The functional properties of the RFamide peptides were tested for their ability to stimulate [³⁵S]GTPγS binding. It was recently shown that an analog of NPFF dose dependently enhanced the binding of [³⁵S]GTPγS to membranes of CHO-hNPFF2 cells, and this stimulation was abolished by PTX pretreatment (Kotani et al., 2001). In our study more than 80% of the hPrRP20 or hPrRP31-induced [³⁵S]GTPγS binding via the hPrRP receptor was prevented by PTX. Taken together, both the hNPFF2 and the hPrRP receptor seem to couple to the G_i/o class of G proteins in CHO cells. It has previously been suggested that the hPrRP receptor hGR3/GRP10 signals through the G_q pathway in both GH3 rat pituitary tumor cells and primary cultures of rat anterior pituitary, but the coupling clearly depends on the specific cellular system in which the receptor is expressed (Kimura et al., 2000; Langmead et al., 2000).

The observation that hPrRP31 and hPrRP20 can interact with the hNPFF2 receptor was also confirmed in functional experiments. Both PrRP ligands caused clear increases in [³⁵S]GTPγS binding over basal. Unexpectedly, we found that hPrRP31 consistently produced significantly greater responses than NPFF or (1DMe)Y8Fa at the hNPFF2 receptor. This is an important finding, since the NPFF2 receptor has been suggested to be involved in mediating the analgesic effect of NPFF in the rat (Bonini et al., 2000; Yang et al., 2001). The higher efficacy of a ligand other than the putative endogenous ligand is surprising, but not unprecedented. A publication by Narita et al. (1998) represents a convincing example of a situation in which endogenous ligands do not produce the highest agonist response in a given test system. They show that the opioid peptides endomorphin-1 and -2 have lower efficacy in the [³⁵S]GTPγS binding assay on mouse spinal cord membranes than has the synthetic μ-opioid agonist [D-Ala²,NHPhe⁴,Gly-ol]enkephalin (Narita et al., 1998). Perret et al. (2002) reported that two vasoactive intestinal peptide (VIP) analogs are more efficacious than the natural agonist, VIP, at mutated VIPAC receptors and referred to these synthetic compounds producing a larger response than the endogenous ligand as “superagonists”. More surprising, however, and to our knowledge not previously reported, is the observation that the endogenous agonist of one GPCR possesses a higher efficacy on another GPCR than the purported endogenous agonist for the second receptor itself. To exclude the possibility that the higher response by hPrRP31 relative to that of NPFF was caused by the activation of an additional receptor in the cellular background of CHO cells, we also tested the effect of hPrRP31 in membranes of CHO-K1 cells that did not express the hNPFF2 receptor. No significant agonist effect of hPrRP31 could be detected in these experiments (data not shown). Our results therefore strongly suggest that, at least on the hNPFF2 receptor subtype, NPFF is not the most efficacious agonist, and hPrRP31 acts as a superagonist. In addition, this report provides a quantitative examination of the relative efficacies of several RFamide peptides on the hNPFF2 receptor. On the hPrRP receptor, NPFF was unable to elicit any functional response, as expected from the observation that NPFF interacts poorly with the hPrRP receptor.

Our observation that hPrRP31 binds to NPFF receptors with relatively high affinity and activates at least the hNPFF2 receptor subtype with high efficacy points to the possibility that PrRP might exert physiological effects through NPFF receptors. A physiologically significant interaction that might regulate, for example, autonomic functions could occur in the nucleus of the solitary tract, where both the NPFF2 receptor (Liu et al., 2001) and PrRP (Roland et al., 1999; Maruyama et al., 1999) have been shown to be present.

Very low numbers of PrRP immunoreactive fibers and nerve terminals have been observed in the laminae I–II of the dorsal horn and around the central canal in the rat spinal cord (P. Panula and K. Kuokkaner, unpublished observations). Tissue distribution studies indicate that PrRP mRNA, but no RFRP mRNA, is present in the rat spinal cord as detected by reverse transcription-polymerase chain reaction (Fujii et al., 1999; Hinuma et al., 2000). However, there is the possibility that the RFRP peptides could be transported to the spinal cord after translation. In future studies it will be very important to examine in detail the coexpression of the NPFF2 receptor and RFamide peptides, although a close physical association between the ligand-storing neurons and receptors is not necessary for interactions

Thus, it is conceivable that the RFamide peptides, NPFF, PrRP, and possibly RFRP, may compete for NPFF receptor under physiological and/or pathological conditions. However, whether the observed in vitro efficacy of hPrRP31 at the hNPFF2 receptor does indeed translate into an in vivo efficacy remains to be established.