Identification and Characterization of Two G Protein-coupled Receptors for Neuropeptide FF*

The central nervous system octapeptide, neuropeptide FF (NPFF), is believed to play a role in pain modulation and opiate tolerance. Two G protein-coupled receptors, NPFF1 and NPFF2, were isolated from human and rat central nervous system tissues. NPFF specifically bound to NPFF1 (K d = 1.13 nm) and NPFF2 (K d = 0.37 nm), and both receptors were activated by NPFF in a variety of heterologous expression systems. The localization of mRNA and binding sites of these receptors in the dorsal horn of the spinal cord, the lateral hypothalamus, the spinal trigeminal nuclei, and the thalamic nuclei supports a role for NPFF in pain modulation. Among the receptors with the highest amino acid sequence homology to NPFF1 and NPFF2 are members of the orexin, NPY, and cholecystokinin families, which have been implicated in feeding. These similarities together with the finding that BIBP3226, an anorexigenic Y1 receptor ligand, also binds to NPFF1 suggest a potential role for NPFF1 in feeding. The identification of NPFF1 and NPFF2 will help delineate their roles in these and other physiological functions.

The octapeptide neuropeptide FF (NPFF 1 or F-8-F-amide) and the related octadecapeptide neuropeptide AF (NPAF or A-18-F-amide) were originally isolated from bovine brain (1) and later determined to be encoded by the same gene and cleaved from a common precursor protein (2). There is a large body of evidence suggesting that NPFF is involved in nociception and in the modulation of opiate-induced analgesia, morphine tolerance, and morphine abstinence (3)(4)(5)(6)(7)(8)(9)(10)(11). Interestingly, NPFF possesses both anti-opioid and pro-opioid actions in animal models of pain. The intracerebroventricular administration of NPFF reverses morphine-induced analgesia in rats, and administration of anti-NPFF antibodies increases opiateinduced analgesia (Reviewed in Ref. 12). Conversely, intrathecal administration of NPFF analogs induces a long-lasting, opioid-induced analgesia and potentiates morphine-ionduced analgesia (12). Other reports have also implicated NPFF in physiological processes such as insulin release, food intake, memory, blood pressure regulation, and electrolyte balance (3). Binding of the NPFF analog [ 125 I]YLFQPQRF-amide to rat spinal cord membranes has revealed a high affinity binding site for which opioid receptor ligands do not compete (13), and the autoradiographic distribution of [ 125 I]YLFQPQRF-amide binding sites indicates high density binding in various regions throughout the rat CNS (14).
The exact mechanism underlying the anti-and pro-opioid effects of NPFF is currently unknown, but these seemingly opposing physiological effects could be accounted for by the existence of multiple receptor subtypes. Until now, the cloning of NPFF receptors has remained elusive. NPFF has been shown to activate adenylyl cyclase in mouse olfactory bulb membranes (15), and NPFF binding to rat brain and spinal cord membranes is inhibited by guanine nucleotides (16), suggesting that NPFF elicits its actions through a G proteincoupled receptor (GPCR).
A peptide related to NPFF, FMRF-amide, activates a cation channel (FaNaCh) in the mollusc Helix aspersa (17), which is a member of the DEG/ENaC family of channels. Although an FMRF-amide-gated channel homologous to FaNaCh has not been identified in vertebrates, both FMRF-amide and to a lesser extent, NPFF, can potentiate responses to acid at members of the related ASIC (acid-sensing ion channel) family of acid-sensing channels (34). This action of NPFF is clearly distinct from the effects observed in the CNS which exhibit a considerably higher potency.
Utilizing a GPCR-targeted degenerate PCR methodology, we have identified a novel GPCR that can specifically bind and be activated by neuropeptide FF and the related peptides PQRFamide and A-18-F-amide, which we have named NPFF1. In addition, we have identified and isolated a second GPCR, structurally related to NPFF1, that can also bind and be activated by NPFF. We have named this second receptor NPFF2. Although NPFF binding sites have been identified in the literature in isolated membranes or in situ, this is the first report identifying a specific receptor system for NPFF.

EXPERIMENTAL PROCEDURES
Materials-NPFF and other commercially available peptides were purchased from Bachem (Torrance, CA). All other peptides and peptoids * 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 nucleotide sequence (s) were synthesized manually or by using an Advanced Chemtech 396-9000 automated peptide synthesizer (Advanced Chemtech, Louisville, KY). Oligonucleotides were synthesized on an Expedite 8909 oligonucleotide synthesizer (PerkinElmer Life Sciences).
Degenerate PCR Cloning-100 ng of rat genomic DNA was subjected to PCR with primers corresponding to the sixth (5Ј-GYNTWYRYNNT-NWSNTGGHTNCC-3Ј) and seventh (5Ј-AVNADNGBRWAVANNA-NNGGRTT-3Ј) transmembrane domains of the rhodopsin GPCR family. Conditions were as follows: 94°C for 3 min; 10 cycles of 94°C for 1 min, 44°C for 1 min, 45 s, and 72°C for 2 min; 30 cycles of 94°C for 1 min, 49°C for 1 min, 45 s, and 72°C for 2 min; 72°C for 4 min. Products were subcloned into the TA cloning kit (Invitrogen, Carlsbad, CA) and sequenced using the ABI Big Dye cycle sequencing protocol and ABI 377 sequencers (Applied Biosystems Inc., Foster City, CA). Nucleotide and amino acid sequence analyses were performed using the Wisconsin Package (GCG, Genetics Computer Group, Madison, WI).
5Ј/3Ј Rapid Amplification of cDNA Ends (RACE)-To determine the full-length coding sequence of AA449919, 5Ј/3Ј RACE was performed on human spleen Marathon Ready cDNA (CLONTECH, Palo Alto, CA). Nested primers specific to AA449919 were used according to the manufacturer's instructions. The products were sequenced as described above. The Wisconsin Package and Sequencher 3.0 (Gene Codes Corp., Ann Arbor, MI) were used to assemble the full-length contiguous sequence of human NPFF2 (hNPFF2) from the AA449919 EST and the RACE products. The full-length clone was amplified from human spinal cord cDNA using primers flanking the initiating methionine and the stop codon in six independent PCR reactions with the Expand Long Template PCR System (Roche Molecular Biochemicals), and subcloned into pcDNA3.1(ϩ). Each of the six products was fully sequenced, and the construct that agreed 100% with the consensus of the six reactions was used for pharmacological analysis.
cDNA Library Screening-Primers specific to the rat receptor fragment were used to isolate a clone representing the full-length BN6 (rNPFF1) receptor from a rat hypothalamic cDNA library (18) using the following PCR protocol: 94°C, hold for 3 min; 40 cycles of 94°C for 1 min, 68°C for 2 min; 4-min hold at 68°C. Positive library pools were subsequently diluted and rescreened by PCR using the same protocol. Positive sub-pools were plated for colony hybridization with 32 P-labeled oligonucleotide probes. Isolated positive colonies were chosen, and the respective plasmids were sequenced as described above. Similarly, the full-length hNPFF1 receptor was isolated by PCR screening of pools of a human spinal cord cDNA library.
Electrophysiology-Chimeric G␣ q/i3 , G␣ q/z , and G␣ q/s were generated by PCR using primers encoding human G␣ q and the C-terminal five amino acids of G␣ i3 , G␣ z , and G␣ q/s (19). Xenopus oocytes were prepared and injected with mRNA as described (18,20). Unless otherwise specified, oocytes were voltage clamped at Ϫ80 mV. Drugs were applied by superfusion in a solution containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , and 5 mM HEPES, pH 7.5.
Receptor Binding Studies-Membranes from COS-7 or HEK-293 cells expressing hNPFF1 or hNPFF2 were isolated and subjected to equilibrium binding assays. In equilibrium saturation binding assays, isolated membranes were incubated in binding buffer (50 mM Tris-HCl, 60 mM NaCl, 1 mM MgCl, 33 M EDTA, 33 M EGTA, pH 7.4, supplemented with 0.2% bovine serum albumin, 2 g/ml aprotinin, and 20 M bestatin) with increasing concentrations of [ 125 I]D-Tyr-Leu-(N-methyl)Phe-Gln-Pro-Glu-Arg-Phe-NH 2 ([ 125 I]1DMeNPFF). In equilibrium competition binding assays, isolated membranes were incubated with 50 pM [ 125 I]1DMeNPFF in the presence of 10 -12 different concentrations of competing ligand for 2 h at 25°C, after which the reaction was stopped by filtration through a double layer of glass fiber filters treated with 0.1% polyethyleneimine using a cell harvester. Radioactivity was measured by scintillation counting. Nonspecific binding is defined as the amount of radioactivity remaining after incubation of membrane protein in the presence of 1 M final concentration of unlabeled NPFF.
Phosphoinositide Turn-over-COS-7 cells were transiently transfected in 96-well plates with G␣ q/z5 and either hNPFF1 or hNPFF2. The day before the assay, the growth medium was changed to 100 l of medium containing 1% serum and 0.5 Ci of [ 3 H]myo-inositol, and the plates were incubated overnight at 37°C in 5% CO 2 . Immediately before the assay, the medium was removed and replaced with 200 l of phosphate-buffered saline containing 10 mM LiCl. The [ 3 H]inositol phosphate accumulation from inositol phospholipid metabolism was started by the addition of increasing concentrations of NPFF, after which the plates were incubated for 1 h in a CO 2 incubator. Reactions were terminated by addition of 15 l of 50% v/v trichloroacetic acid, followed by a 40-min incubation at 4°C. After neutralizing trichloroacetic acid with 40 l of 1 M Tris, the contents of the wells were transferred to a multiscreen HV filter plate (Millipore, Bedford, MA) containing Dowex AG1-X8 (200 -400 mesh, formate form). Each well was washed two times with 200 l water, followed by 2ϫ 200 l of 5 mM sodium tetraborate/60 mM ammonium formate. The [ 3 H]inositol phosphates were eluted with 200 l of 1.2 M ammonium formate/0.1 M formic acid, and samples were counted by liquid scintillation counting.
cAMP Formation Assay-COS-7 cells were transiently transfected with G␣ z and either hNPFF1 or hNPFF2 and incubated in phosphatebuffered saline supplemented with 10 mM HEPES, 5 mM theophylline, 2 g/ml aprotinin, 0.5 mg/ml leupeptin, and 10 g/ml phosphoramidon for 20 min at 37°C, in 5% CO 2 . Test compounds were added, and the cells were incubated for an additional 10 min at 37°C, after which the reaction was stopped by the addition of 100 mM HCl. The plates were then incubated at 4°C for 15 min, and the cAMP content in the stopping solution was measured by radioimmunoassay. Radioactivity was measured using a gamma counter equipped with data reduction software.
Intracellular Calcium Mobilization Assay-COS-7 cells expressing the chimeric G protein G␣ q/z5 , G␣ q/i3 , or G␣ q/s and either rNPFF1 or hNPFF2 were plated in 96-well plates and grown to confluence. After incubation with Fluo 3-AM, cells were washed with Hanks' balanced salt solution and equilibrated for 20 min. The fluorescence emission caused by intracellular calcium mobilization elicited by agonists of the expressed receptor was determined with a fluorescence imaging plate reader (FLIPR TM , Molecular Devices Corp., Sunnyvale, CA).
Receptor Autoradiography-These methods have been described in detail previously (21). [ 125 I]1DMeNPFF (specific activity, 2200 Ci/ mmol) was synthesized by iodination with chloramine-T (PerkinElmer Life Sciences). Adjacent tissue sections were incubated in the presence of 300 nM BIBP3226 (Research Biochemicals Inc., Natick, MA), to selectively displace NPFF1 receptor binding sites or 300 nM frog pancreatic polypeptides (PP) (frog PP, Peninsula, Belmont, CA) to selectively displace binding to the NPFF2 receptor binding sites. Nonspecific binding was determined in the presence of 1 M NPFF in the incubation buffer.
mRNA Localization-For the detection of RNA encoding NPFF receptors, quantitative RT-PCR was performed on mRNA extracted from multiple tissue samples. RNA was prepared using Trizol (Life Technologies, Inc.) or was purchased (CLONTECH). Reverse transcription and PCR reactions were carried out in 50-l volumes using rTth DNA polymerase (PerkinElmer Life Sciences). The following primer sets were synthesized: hNPFF1 forward, 5Ј-CTGGTCACCGTCTACGCCTT-3Ј, reverse, 5Ј-CCGCGGCGGAAGTTCT-3Ј; hNPFF2 forward, 5-CCTG-ATTGTGGCCCTGCT-3Ј, reverse, 5Ј-CATTTGGAGAAAGGTCAGCGT-AG-3Ј; rNPFF1 forward, 5Ј-GCTGTGGAAAGGTTCCGCT-3Ј, reverse, 5Ј-CGCCTTCCGAAGGGTCA-3Ј; rNPFF2 forward, 5Ј-GAGGATCTAC-ACCACCGTGCTATT-3Ј, reverse, 5Ј-GAAGCCCCAATCCTTGCATAC-3Ј. Fluorogenic probes were synthesized using 6-carboxyfluorescein as the reporter at the 5Ј end and 6-carboxy-4,7,2,7Ј-tetramethylrhodamine as a quencher at the 3Ј end of the oligonucleotide (Synthegen, LLC). Each RT-PCR reaction contained 100 ng of total RNA. RNA was quantified using spectroscopy (A 260 ) and RiboGreen (Molecular Probes) assays. All reagents for RT-PCR (except mRNA and oligonucleotide primers) were obtained from PerkinElmer Life Sciences, and the manufacturer's protocols were used for RT-PCR. Each 96-well plate contained RNA extracted from tissue (in triplicate), controls, and standard curves to facilitate relative quantification of NPFF1 and NPFF2 RNA. Standard curves for quantification of human and rat NPFF1 and NPFF2 were constructed using varying amounts of RNA extracted from whole brain. To confirm that RNA was not contaminated with genomic DNA, PCR reactions were carried out without reverse transcription using Taq DNA polymerase. The integrity of the RNA was assessed by amplification of RNA coding for cyclophilin or glyceraldehyde 3-phosphate dehydrogenase. Following reverse transcription and PCR amplification, data were analyzed using PerkinElmer sequence detection software. The fluorescent signal from each well was normalized using an internal passive reference, and data were fitted to a standard curve to obtain the relative quantities of NPFF RNA expression.
Chromosomal Localization-Chromosomal localization for human NPFF1 and NPFF2 receptor genes was established using a panel of radiation hybrids prepared by the Stanford Human Genome Center (SHGC) and distributed by Research Genetics, Inc. The "Stanford G3" panel of 83 radiation hybrids was analyzed by PCR using the same primers, probes, and thermal cycler profiles as used for localization. 20 ng of DNA was used in each PCR reaction. Data were submitted to the Radiation Hybrid Server (SHGC), which linked the NPFF1 and NPFF2 gene sequences to specific markers. NCBI LocusLink and NCBI GeneMap '99 were used in further analyses of gene localization.

RESULTS
Cloning and Identification of NPFF1-Utilizing a GPCRtargeted degenerate PCR methodology on rat genomic DNA, we identified a novel GPCR fragment most closely related to several peptide-ligand GPCRs. The full-length rat receptor, BN6, was isolated from a rat hypothalamic cDNA library, and the human ortholog, BO102, was subsequently isolated from a human spinal cord cDNA library. Sequence analysis of the rat and human receptors revealed coding sequences of 1296 bp and 1290 bp, and predicted proteins of 432 and 430 amino acids, respectively, which share 87% identity (Fig. 1). Amino acid comparison of BO102 with known GPCRs indicates that it is most similar to human orexin1 (37% identity), human orexin2 (35%), human neuropeptide Y (NPY) Y2 (34%), human cholecystokinin A (CCKA) (34%), human NPY Y1 (32%), mouse GIR (32%), human prolactin-releasing hormone receptor (32%), and human NPY Y4 (31%).
To determine the ligand specificity of this receptor, Xenopus oocytes expressing BN6 and a G␣ q/z5 chimeric G␣ G-protein were used to screen a library of peptidic neurotransmitters. Within this collection only NPFF (1 M) elicited reliable and robust responses (Fig. 2). Current amplitudes averaged 459 Ϯ 81 nA (n ϭ 13) and these exhibited a concentration dependence with an EC 50 of 163 nM (n ϭ 8 oocytes; data not shown). The NPFF-related ligands, A-18-F-amide, Y-18-F-amide, Y-8-Famide (1 M) and the C-terminal tetrapeptide PQRF-amide (10 M), also activated the receptor (data not shown). These results suggested that BN6 encoded a receptor for NPFF, herein called rNPFF1.
Cloning and Identification of NPFF2-A search of Genbank data bases revealed a related human expressed sequence tag (EST) fragment of 532 bp (accession #AA449919), which encoded an amino acid sequence with 59% identity to hNPFF1 and 50% identity to rNPFF1. RACE was used on human spleen cDNA to clone the 5Ј and 3Ј ends of AA449919, and the fulllength receptor, BO89, was amplified from spinal cord cDNA.
The rat ortholog of BO89 was cloned by PCR, and named BO119. The coding regions of BO89 and BO119 are 1260 and 1251 bp, encoding proteins with predicted lengths of 420 and 417 amino acids, respectively (Fig. 1). BO89 and BO119 share 78% amino acid identity and are 49 -50% identical to rat and human NPFF1.
Oocytes expressing BO89 were robustly activated by NPFF (1 M, Fig. 2). Mean current amplitudes were 528 Ϯ 99 nA (n ϭ 18). This finding suggested that BO89 was an additional member of the NPFF receptor family; therefore, it was named hNPFF2. Both rNPFF1 and hNPFF2 receptor responses were dependent upon co-expression of the chimeric G␣ q/z5 G-protein.
Subsequent to the identification of BO89 as NPFF2, a report was published describing the cloning of an orphan receptor sequence named NPGPR (22). NPGPR is nearly identical to NPFF2 except that the N terminus of NPGPR is longer by 102 amino acids. Although it is possible that the more N-terminal initiating methionine could be used for translation of this receptor, the second methionine (the initiating methionine of NPFF2) is surrounded by a good kozak consensus sequence (atcATGaat) and would code for a protein of approximately the same length as rNPFF2, rNPFF1, and hNPFF1.
Binding Properties of NPFF1 and NPFF2-To further assess the pharmacological identity of the human NPFF1 and NPFF2 receptors, the binding properties of the cloned receptors were explored using [ 125 I]1DMeNPFF as a radioligand. The specific binding of [ 125 I]1DMeNPFF with membranes harvested from COS-7 cells transfected with NPFF1 or NPFF2 receptors at 25°C reached a maximum by 60 min and remained unchanged for up to 120 min (data not shown). Membranes from transiently transfected COS-7 cells exhibited high affinity, saturable [ 125 I]1DMeNPFF binding for both NPFF1 and NPFF2 receptors (Fig. 3, A and B). Nonlinear analysis of [ 125 I]1DMeNPFF saturation data yielded an equilibrium dissociation constant (K d ) of 1.13 Ϯ 0.16 and 0.37 Ϯ 0.03 nM (S.D. n ϭ 2) for NPFF1 and NPFF2 receptors, respectively (Fig. 3, C  and D). Untransfected host cells did not display specific [ 125 I]1DMeNPFF binding.
Transient expression of hNPFF1 and hNPFF2 receptors in 293 human embryonic kidney cells (HEK-293 cells) yielded similar K d values from saturation studies and a more robust expression (Bmax ϭ 1592 and 510 fmoles/mg protein for hNPFF1 and hNPFF2, respectively) as compared with the COS-7 cells (B max ϭ 543 and 47 fmol/mg protein for hNPFF1 and hNPFF2, respectively). Therefore, the HEK-293 cells were used to measure the binding affinities (pK i ) of various NPFFrelated peptides in a competition binding assay using [ 125 I]1DMeNPFF as the radioligand. The C-terminal RF-amide peptide, PQRF-amide, displaced [ 125 I]1DMeNPFF binding to both NPFF1 and NPFF2. In addition, NPFF receptors showed high binding affinity for FMRF and lower binding affinity for its D-Met analog, suggesting that the binding domain of the receptors recognizes the C-terminal RF-amide of NPFF. Other C-terminal RF-amide peptides such as frog PP, an NPY Y4 receptor agonist (23), showed greater affinity for the rat (125fold) and human (300-fold) NPFF2 receptors compared with the rat and human NPFF1 receptors (see Table I). Conversely, human PP, human NPY, and peptide YY, which contain a C-terminal RY-amide (24 -26), did not bind to either NPFF1 or NPFF2 (data not shown). Interestingly, the synthetic C-terminal RY-amide peptoid BIBP3226, an NPY Y1-selective compound (27), displayed 10 -60-fold higher affinity for the human and rat NPFF1 receptor as compared with NPFF2 receptors. These findings question the pharmacological selectivity of this peptoid for NPY Y1 receptors, suggesting that BIBP3226 and related compounds may mediate some of their in vivo effects through NPFF receptors rather than through NPY Y1 receptors (28,29).
NPFF1 and NPFF2 Coupling to Heterotrimeric G Proteins-The ability of NPFF1 and NPFF2 receptors to couple functionally to heterotrimeric G proteins was tested using intact COS-7 cells transiently expressing these receptors. NPFF (1 M) had no effect on either basal or forskolin-stimulated cAMP formation or PI turn-over in untransfected COS-7 cells, indicating that endogenous adenylate cyclase-or PI-coupled NPFF receptors are not expressed in untransfected cells. In COS-7 cells transfected with the rat NPFF1 receptor, NPFF elicited a small (2-fold) increase in total inositol phosphate release with an EC 50 of 239 nM (Fig. 4A), which most likely reflects a minor activation of this pathway. Pretreatment of cells expressing the rat NPFF1 receptor with 100 ng/ml pertussis toxin (PTX) for 18 h prevented the NPFF-mediated activation of PI turn-over, suggesting that the activation of the phospholipase C pathway in native cells transiently expressing NPFF1 is most likely secondary to the activation of endogenous PTX-sensitive G proteins and not G␣ q . When NPFF1 was co-expressed with the G␣ q/z5 chimera, NPFF stimulation resulted in a much more robust inositol phosphate release response, which was not sensitive to PTX treatment, with an EC 50 that was left-shifted 2 log units relative to transfection with NPFF1 alone (Fig. 4B). The PTX insensitivity of the response in cells co-expressing NPFF1 and the G␣ q/z5 chimera suggests that the PI response in cells co-expressing the chimera is mediated by the activation of phospholipase C by the G␣ q domain of the chimera and not secondary to activation of endogenous PTX-sensitive G proteins. In COS-7 cells expressing the human NPFF2 alone, we could not detect a PI turn-over response to NPFF.
To further characterize the functional activity of the receptors, the ability of rNPFF1 and hNPFF2 to stimulate intracellular Ca 2ϩ mobilization when co-expressed with different chimeric G proteins was tested in COS-7 cells. Co-transfection of rat NPFF1 or human NPFF2 receptors with either G␣ q/i3 or G␣ q/z5 led, in both cases, to the activation by NPFF of intracellular Ca 2ϩ mobilization in a concentration-dependent manner (Fig. 5). The EC 50 values for the NPFF-mediated stimulation of intracellular Ca 2ϩ release were in good agreement with the binding affinities of NPFF at NPFF1 and NPFF2 receptors as measured in equilibrium binding assays. However, when G␣ q/s was co-transfected with NPFF1, the activation of intracellular

FIG. 3. Binding of [ 125 I]1DMeNPFF to crude membranes of cells expressing human NPFF1 and human NPFF2 receptors. Saturation isotherm at human NPFF1 (A) and human NPFF2 receptors (B) expressed in Cos-7 cells.
Competition analysis for the binding of 0.1-0.2 nM [ 125 I]1DMeNPFF at hNPFF1 (C) and hNPFF2 receptors (D) expressed in HEK-293 cells. Affinities (IC 50 values) obtained for inhibition of radioligand binding were converted to K i values using the Cheng-Prussoff equation (39). Mean K i values calculated from these studies are summarized in Table I Ca 2ϩ mobilization by NPFF was right-shifted and displayed a weaker maximal response (Fig. 5A) . Furthermore, co-transfection of G␣ q/s with NPFF2 did not permit intracellular Ca 2ϩ mobilization by NPFF (Fig. 5B). No response was detected in cells expressing G␣ q/s or G␣ q/z alone. These results would suggest that although NPFF1 can couple to cyclase-stimulatory G proteins, NPFF1 and NPFF2 may couple more efficiently to cyclase-inhibitory G proteins in this heterologous system. Subsequent functional studies monitoring intracellular Ca 2ϩ fluxes with hNPFF1 and hNPFF2 were conducted, using a fluorescence imaging plate reader, with transiently transfected COS-7 cells co-expressing either NPFF1 or NPFF2 and G␣ q/z5 . NPFF elicited an increase in intracellular Ca 2ϩ when either hNPFF1 or hNPFF2 were transfected, whereas there was no response observed in cells transfected with only the G␣ q/z chi-mera. As shown in Table II, PQRF amide acted as a full agonist in cells expressing either the NPFF1 or NPFF2 receptors. In contrast, only cells expressing hNPFF2, but not hNPFF1, responded with an intracellular Ca 2ϩ response to frog PP, suggesting that frog PP is an NPFF2-selective agonist.
Anatomical Distribution of NPFF1 and NPFF2 in the Rat CNS-The anatomical distribution of NPFF1 and NPFF2 receptor binding sites was revealed throughout the rat CNS by receptor autoradiography using 0.5 nM [ 125 I]1DMeNPFF and making use of the subtype-selective displacers, frog PP and BIBP3226 (Fig. 6). The distribution of the rNPFF1 and rNPFF2 receptor binding sites correlated well with the reported distribution of NPFF-like immunoreactive neurons and terminals (30). Additionally, the distribution of rNPFF1 and rNPFF2 receptors was concordant with previous reports of the

FIG. 5. NPFF-stimulated activation of intracellular Ca 2؉ mobilization in COS-7 cell expressing recombinant human NPFF receptors and chimeric G proteins.
Changes in intracellular Ca 2ϩ mobilization were measured in intact COS-7 cells loaded with Fluo 3 as described under "Experimental Procedures." COS-7 cells were co-transfected with G␣ q/i3 , G␣ q/z , or G␣ q/s alone or in combination with the cDNAs encoding the rat NPFF1 receptor (A) or the human NPFF2 receptor (B). The intracellular Ca 2ϩ response was measured from the maximum fluorescence intensity after the addition of NPFF. Results are expressed as mean Ϯ S.E. of the percentage of E max calculated from the maximum response elicited by NPFF (maximum response Ϫ basal response) in cells transfected with a combination of NPFF receptor and G␣q/z. The data shown are from two independent experiments performed in duplicate. anatomical distribution of NPFF binding sites obtained using [ 125 I]1DMeNPFF and [ 125 I]Y-8-F-amide (31).
The highest density of NPFF1 binding sites in the rat CNS was observed in the lateral septum, the anterodorsal thalamic nucleus, and the superior colliculus. Moderate NPFF1 binding was detected in the septum, accumbens nucleus, the bed nucleus of the stria terminalis, and the interpeduncular nucleus. Lower densities of NPFF1 binding sites were observed in thalamic and brain stem nuclei, the amygdala, and hippocampus. The highest densities of NPFF2 binding sites were detected in the parafascicular and reticular thalamic nuclei, the anterior pretectal nucleus, lateral hypothalamus, medial mammillary nucleus, ventral hippocampus, the A5 noradrenergic cell group, spinal trigeminal nucleus, and spinal cord dorsal horn. In the spinal cord, only NPFF2 binding sites were detected in the dorsal horn of the spinal cord (arrow in Fig. 6, DЈЈ). Moderate NPFF2 binding was observed in the subiculum, whereas the central gray and various brain stem nuclei displayed lower binding densities for NPFF2.

Localization of NPFF1 and NPFF2 mRNA in Human and
Rat Tissues-Quantitative RT-PCR was used to assay NPFF1 and NPFF2 mRNA in 24 human tissues and 41 rat tissues (Table III). NPFF1 receptor mRNA is expressed at higher levels in the CNS than in peripheral tissues. The highest levels of hNPFF2 mRNA were detected in the placenta, whereas the highest levels of rNPFF2 mRNA were found in CNS structures such as the spinal cord, medulla oblongata, hypothalamus, substantia nigra, and amygdala. The chromosomal localization of human NPFF1 and NPFF2 was determined by radiation hybrid mapping using the Stanford G3 panel. The hNPFF1 gene maps to SHGC-30283, which is localized to chromosome 10q21, and the hNPFF2 receptor gene maps to SHGC-24728, which is localized to chromosome 4q13.2-q13.3. This chromosomal localization is markedly similar to the genes encoding the NPY family of receptors, which share 31-51% identity between subtypes and 30 -34% identity with the NPFF receptors (32). DISCUSSION Neuropeptide FF has been shown to modulate a variety of physiological processes such as insulin release, blood pressure regulation, food intake, electrolyte balance, nociception, opiateinduced analgesia, and morphine abstinence syndrome (3,4,6,12), although its role in pain modulation is the most well defined (3,6,12). Involvement of NPFF in such a diverse array of processes suggests that NPFF may interact with more than one receptor subtype or that stimulation of NPFF receptors in various target tissues triggers different processes that are tissue-dependent. In this study, we have isolated and characterized two NPFF receptor subtypes, NPFF1 and NPFF2, that are capable of binding NPFF and related peptides in the nanomolar and subnanomolar range. The evidence obtained by coexpression of chimeric proteins suggests that NPFF receptors may couple more efficiently to the activation of the adenylate cyclase inhibitory class of G proteins (G␣ i , G␣ o , G␣ z ). Unfortunately, receptor-mediated adenylate cyclase inhibition was not  detectable in COS-7 cells expressing either NPFF1 or NPFF2 receptors through the endogenous repertoire of G proteins in this cell line, indicating that further work is needed to fully characterize the endogenous signaling characteristics of these receptors. The finding that NPFF1 can couple to G␣ q/s in this heterologous expression system may explain the results in the report by Gherardi and Zajac (15) that demonstrates NPFF activation of adenylate cyclase in olfactory bulb membranes. Alternatively, the observations of Gherardi and Zajac may be explained by the possibility of the existence of other NPFF receptors that have yet to be discovered or other NPFF receptor systems (33). The distribution of rat and human NPFF1 and NPFF2 mRNA as measured by RT-PCR is consistent with a broad modulatory action in the periphery as well as the CNS (Tables  III and IV). However, there are considerable species differences in the distribution of NPFF1 and NPFF2 between the human and rat. NPFF1 mRNA is more abundant in the human spinal cord, which is in sharp contrast to the rat, where NPFF2 mRNA is more predominant. Although NPFF2 mRNA is detectable in the human spinal cord, its abundance relative to other regions is much lower than seen in rat; this implies that in the human, the NPFF1 receptor might play a more prominent role in sensory modulation than NPFF2. Peripheral organs also demonstrated substantial species differences in the distribution of NPFF receptor mRNA. Although rat heart expressed high levels of NPFF2 mRNA, human heart expressed little of either NPFF1 or NPFF2 mRNA. The human spleen contained transcripts for both NPFF1 and NPFF2 receptors at higher levels than found in rats.
Although the difference in mRNA localization between species complicates the interpretation of the data, these results are not surprising. Dupuy et al. (33) demonstrated significant species differences in the localization of NPFF binding in rodent and lagomorph brain and spinal cord, even between rats and mice. Possible explanations for these observations are that NPFF may have different functions in different species, that there may be significant differences in the central NPFF pathways between species (33), that a related peptide such as A-18-F-amide may be the endogenous ligand for NPFF1 or NPFF2 in certain tissues, or that the relative amounts of NPFF1 and NPFF2 vary between other species in the same manner as they do between human and rat. Of course, there is also the possibility of the existence of other unidentified NPFF receptor subtypes or other NPFF receptor systems. Recently, Askwith et al. (34) demonstrated that neuropeptide FF and FMRF-amide activated proton-gated currents from cultured sensory neurons and heterologously expressed acid-sensing or dorsal root acidsensing ion channels. This activation, however, required a much higher concentration of NPFF and FMRF-amide (EC 50 ϭ 33 M) than is required for NPFF1 or NPFF2 activation (EC 50 ϭ 16 and 2 nM, respectively).
Taken together, the results from the RT-PCR and receptor binding studies using [ 125 I]1DMeNPFF have examined distribution of NPFF1 and NPFF2 mRNA and receptor binding sites in the rat CNS. The anatomical distribution of NPFF-like immunoreactivity, NPFF2 mRNA, and NPFF2 receptor binding sites in various CNS regions such as the dorsal root ganglia, spinal cord, spinal trigeminal, parafascicular, and raphe nuclei, and lateral hypothalamus, supports a role for the NPFF2 receptor in nociception. NPFF has been shown to attenuate the analgesic effects of morphine after intrathecal and intracerebroventricular injection (35), and the localization suggests that this effect may be mediated by the NPFF2 receptor. However, a contribution of NPFF1 in nociception should not be ruled out, especially since NPFF1 mRNA is present in very high levels in   Trace  Trace  Pineal  Trace  4  Pituitary  24  34  Retina  14  40  Salivary gland  Trace  33  Spinal cord  24  100  Spleen  Trace  Trace  Stomach  Trace  14  Skeletal muscle  Trace  Trace  Striatum  17  16  Substantia nigra  49  67  Testes  43  4  Thalamus  3  15  Thymus  Trace  12  Trigeminal ganglia  16  57  Urinary bladder  Trace  16  Uterus  Trace  Trace  Vas deferens  Trace  Trace  Whole brain  21  24 the human spinal cord. Although in the rat there is some discordance between the expression of NPFF1 mRNA and binding sites in the dorsal horn of the spinal cord. The inability to observe NPFF1 binding sites might be explained by limitations of detection of the autoradiographic technique or possibly that NPFF1 receptors are expressed on terminal projections of spinal cord neurons outside of the spinal cord. The localization of mRNA and binding sites for NPFF1 and NPFF2 in various components of the basal ganglia, such as the nucleus accumbens, substantia nigra, and the caudate putamen, suggests that NPFF receptors might play an in the regulation of the central dopaminergic system, albeit and indirect one, because previous studies have shown that NPFF binding sites are not found in dopaminergic cell bodies (36). Some of the highest levels of NPFF-like immunoreactivity have been observed in the rat hypothalamus (3,37). The localization of NPFF2 receptor binding sites and NPFF2 mRNA in the hypothalamus, a region involved in catecholaminergic and serotonergic feeding systems, circadian feeding, and spontaneous activity, suggests that NPFF2 may be involved in the regulation of ingestive behavior. NPFF itself has been shown to reduce food intake in rats (38). Likewise, the presence of NPFF1 mRNA in the hypothalamus raises the possibility that it may also have a role in hypothalamic function. The neuropeptide Y Y1 receptor antagonist, BIBP3226, is an NPFF1selective ligand (Table I) which has been shown to block feeding through a nonspecific mechanism, not secondary to inhibition of Y1 (28). Therefore, it is possible that the inhibition of NPYinduced feeding by BIBP3226 and the inhibition of feeding by NPFF, as demonstrated by Murase et al. (38), are mediated through NPFF1.
In addition to BIBP3226 binding to both NPFF receptors with preferential binding to rat and human NPFF1, frog PP, an NPY Y4 agonist, also binds to both NPFF receptor subtypes, displaying a higher affinity for rat and human NPFF2. Furthermore, the levels of amino acid identity and similarity between the two NPFF receptor subtypes and the four known NPY receptor subtypes are compatible with the idea that all six receptors could belong to the same evolutionary lineage. Finally, the human NPFF2 gene is localized in the vicinity of the Y5-Y1-Y2 gene cluster on chromosome 4q31 and the human NPFF1 gene maps to chromosome 10q21, close to the NPY Y4 gene on 10q11-21. This leads us to believe that the NPFF1 and NPFF2 genes may have been generated by gene duplication of ancestral NPY receptor genes (32).
In conclusion, we have identified and isolated two members of a receptor family for NPFF and NPFF-related hormones. The fact that NPFF1 and NPFF2 mRNA transcripts are not always found in the same tissues may explain the varied effects of NPFF in different tissues, as well as the seemingly paradoxical pro-and anti-opiate effects of NPFF. The discovery and characterization of these two NPFF receptors provide the means of identifying receptor-selective pharmaceutical agents necessary to further probe and understand the physiological roles and potential therapeutic applications of NPFF action.