The First Selective Agonist for the Neuropeptide YY5Receptor Increases Food Intake in Rats*

The first Y5 receptor-selective analog of neuropeptide Y (NPY), [Ala31,Aib32]NPY, has been developed and biologically characterized. Using competition binding assays on cell lines that express different Y receptors, we determined the affinity of this analog to be 6 nm at the human Y5 receptor, >500 nm at the Y1 and Y2 receptors, and >1000 nm at the Y4 receptor. Activity studies performed in vitro using a cAMP enzyme immunoassay, and in vivousing food intake studies in rats, showed that the peptide acted as an agonist. Further peptides obtained by the combination of the Ala31-Aib32 motif with chimeric peptides containing segments of NPY and pancreatic polypeptide displayed the same selectivity and even higher affinity (up to 0.2 nm) for the Y5 receptor. In vivoadministration of the new Y5 receptor-selective agonists significantly stimulated feeding in rats. The NMR solution structures of NPY and [Ala31,Aib32]NPY showed a different conformation in the C-terminal region, where the α-helix of NPY was substituted by a more flexible, 310-helical turn structure.

Neuropeptide Y (NPY), 1 a 36-residue peptide amide, is a member of the pancreatic polypeptide (PP) hormone family that also includes PP and peptide YY (PYY) (1). NPY is expressed in the central and peripheral nervous systems and is one of the most abundant neuropeptides in the brain. Several important physiological activities, such as induction of food intake, inhibition of anxiety, increase in memory retention, presynaptic inhibition of neurotransmitter release, vasoconstriction, and regulation of ethanol consumption, have been attributed to NPY (2,3). Especially, the role of NPY in feeding is of major interest because NPY receptor antagonists are po-tential anti-obesity drug candidates. Many studies have established the strong central influence of NPY in feeding behavior; for example, injection of NPY into the hypothalamus increases food intake (4,5), and high NPY levels are correlated with leptin deficiency (6), the hormone that is secreted by adipocytes and regulates body weight and energy balance (7,8). Furthermore, NPY knockout can reduce obesity in leptin-deficient mice (named ob/ob mice) (6).
Five distinct Y receptor subtypes that bind NPY, PYY, and PP with different affinities have been identified, cloned, and characterized (9). They all belong to the superfamily of the G-protein-coupled receptors and are referred to as Y 1 , Y 2 , Y 4 , Y 5 , and Y 6 . From studies conducted using partly selective agonists and antagonists, antisense approaches, and knockout techniques, the Y 1 and Y 5 receptors have been suggested to mediate the stimulatory effect of NPY on food intake (10,11). NPY, PYY, [Leu 31 ,Pro 34 ]NPY, and three N-terminally truncated analogs, NPY-(2-36), NPY- , and PYY- , have been shown to increase food intake. The rank order of potency of the agonists suggests that the Y 5 receptor is most likely involved in food intake (10,12). However, it has also been shown that both Y 1 and Y 5 receptor antagonists can inhibit NPY-induced food intake (13,14). One limitation of the agonists currently used for in vivo studies is the lack of receptor selectivity; NPY and PYY bind to the receptors Y 1 , Y 2 , and Y 5 with high affinity, the analog [Leu 31 ,Pro 34 ]NPY has high affinity for the Y 1 , Y 4 , and Y 5 receptors, whereas N-terminal fragments are potent at the Y 2 as well as at the Y 5 receptor (9,12,15). [D-Trp 32 ]NPY has been described in literature as a weak Y 5 receptor-selective agonist with orexigenic properties (12), but antagonism against NPY-induced increase in food intake has been observed as well (16,17). Furthermore, binding affinity studies on [D-Trp 32 ]NPY at the Y receptors have shown significant Y 2 receptor affinity in addition to its affinity for the Y 5 receptor (18).
Because highly specific tools to investigate the Y 5 receptor activity are still missing, we have focused our work on the design of NPY receptor agonists with both high affinity and selectivity for the Y 5 subtype. It is well established that the C-terminal part of NPY represents the interaction site with the Y receptors and that amino acid exchange is poorly tolerated in the region 33-36 (49); therefore, we induced a conformational change within the peptide region that mediates receptor binding by introducing the ␤-turn-inducing dipeptide Ala-Aib (aminoisobutyric acid) (19) into positions 31-32 of NPY and some peptides that contain segments of NPY and PP (NPY/PP chimeras). The [Ala 31 ,Aib 32 ]-modified peptides showed high selectivity for the Y 5 receptor. Furthermore, in vitro and in vivo studies clearly proved their NPY receptor agonism as well as stimulation of food intake.
The solution structure of [Ala 31 ,Aib 32 ]pNPY was investigated by CD and two-dimensional NMR. By comparison with the NMR structure of human NPY (hNPY) determined by Monks and co-workers (20), a significant conformational change of the C-terminal fragment 28 -36 was observed; although in the native peptide the ␣-helix extends up to residue 36, the ␣-helical motif of [Ala 31 ,Aib 32 ]pNPY ends with a 3 10helical turn between residues Ile 28 and Ala 31 followed by a flexible C terminus. Considering the binding properties of this NPY analog, this structural modification seems to be the key for Y 5 receptor selectivity.
Peptide Synthesis and Purification-The peptides were synthesized by Fmoc/tert-butyl solid-phase strategy with an automated multiple peptide synthesizer (Syro MultiSynTech, Bochum, Germany), starting from 4-(2Ј,4Ј-dimethoxyphenyl-fluorenylmethoxycarbonyl-aminomethyl)-phenoxyacetamido-norleucylaminomethyl resin (30 mg, 0.45 mol/mg) (21). Each Fmoc-amino acid (10-fold excess) was introduced by double coupling (twice for 36 min) using in situ activation with diisopropylcarbodiimide and hydroxybenzotriazole. Fmoc removal was carried out with piperidine in dimethylformamide (15 min). Cleavage from the resin and amino acid side chain deprotection were accomplished in one step with 90% trifluoroacetic acid in the presence of 10% scavengers (thioanisole and thiocresol 1:1) for 3 h. For cleavage of the methionine-containing peptides, ethanedithiol was added to the cleavage mixture. The peptides were precipitated from ice-cold diethyl ether and collected by centrifugation, resuspended in ether, and centrifuged again. This procedure was repeated four times. The crude peptides were dissolved in tert-butanol/water (3/1 w/w), frozen, and lyophilized. The methionine-containing peptides were subsequently treated with trifluoroacetic acid/trimethylbromosilane/ethanedithiol (96/2.4/1.6 v/v, 30 min) to reduce the methionine sulfoxide (22) and then precipitated from ice-cold ether as described above. The lyophilized peptides were purified by preparative HPLC using a Nucleosil C-18 column (6 m, 25 ϫ 30 mm, Waters), 0.1% trifluoroacetic acid in water, and 0.1% trifluoroacetic acid in acetonitrile as the eluting system. The pure products were characterized by analytical HPLC performed on a LiChrospher RP-18 column (5 m, 3 ϫ 125 mm, Merck KG, Darmstadt, Germany) and by electrospray ionization mass spectrometry (Finnigan). The peptides were dissolved in diluted HCl (0.5 mM), frozen, and lyophilized; this procedure was repeated twice to obtain the corresponding hydrochloride salts.
Binding Assays-Cells were resuspended in incubation buffer (Minimum Essential Medium with Earl's salts containing 0.1% bacitracin, 50 M pefabloc SC, and 1% bovine serum albumin). 200 l of the suspension containing ca. 440,000 cells were incubated with 25 l of a 10 nM solution of [ 3 H]propionyl-NPY and 25 l of a 10 M solution of NPY or analog. Nonspecific binding was defined in the presence of 1 M cold NPY. After 1.5 h at room temperature, the incubation was terminated by centrifugation at 2,000 ϫ g and 4°C for 5 min. The pellets were then washed once with phosphate-buffered saline by centrifugation, resuspended in phosphate-buffered saline, and mixed with the scintillation mixture. Radioactivity was determined using a beta-counter.
cAMP Enzyme Immunoassay-Cells grown to confluency were resuspended in cAMP buffer (145 mM NaCl, 1 mM MgSO 4 , 5 mM KCl, 10 mM HEPES, 0.5% bovine serum albumin, 10 mM glucose, 0.1 mM 3-isobutyl-1-methylxanthine, pH 7.4). 1.5 million cells were incubated with 40 l of a 1.5 mM solution of forskolin and different concentrations of peptides for 1 h at 37°C. Incubation was stopped by adding 100 l of a 1 M solution of HCl. Cell lysis was done by freezing followed by centrifugation, and the supernatant was diluted 1:30. The intracellular amount of cAMP was determined using a cAMP enzyme immunoassay (Biotrak). Reactions were performed according to the protocol of the manufacturer, and optical density was determined at 450 nm.
Food Intake Studies-Adult male rats weighing between 340 and 400 g were housed individually and maintained on a 12:12 h light-dark cycle beginning at 6 a.m. (13). Tap water and standard laboratory chow were available throughout. After 1 week of habituation to their new housing conditions, the animals were anesthetized with sodium pentobarbital for the placement of stainless steel guide cannulae. Bilateral guide cannulae were placed 1 mm above the paraventricular nucleus according to the stereotaxic coordinates (48): AP, Ϫ1.8; L, 0.5; V, 7.0. Guide cannulae were maintained in place on the skull with small metal screws and dental acrylic cement. Cannulae were closed with a stainless steel stylet when not in use. Rats were allowed to recover for 1 week and were adapted to the injection procedure. On the day of the experiments, drugs were injected between 8 and 9 a. m. For each experiment, eight rats were used, and for each dose a different group of rats was used.
CD Spectroscopy-The CD spectra were recorded on a JASCO model J720 spectropolarimeter over the range of 185-250 nm at 20°C under a N 2 atmosphere. The peptides were dissolved in water at a concentration of 60 M. The pH was lowered to 3.2 by the addition of small aliquots of a 0.12 M solution of HCl followed by dilution with water to obtain a concentration of 30 M. Each measurement was performed four times using a sample cell with constant temperature and with a path of 0.02 cm. The response time was set to 2 s at a scan speed of 20 nm/min, a sensitivity range of 10 millidegrees, and a step resolution of 0.2 nm. The CD spectrum of the solvent was subtracted from the CD spectra of the peptide solutions to eliminate the interference from cell, solvent, and optical equipment. High frequency noise was reduced by means of a low-path Fourier transform filter.  Structure Calculation-Upper limits for the structure calculation were taken from the volume integrals of NOESY peaks from all experiments recorded at 2 mM concentration. 3 J NH␣ coupling constants were determined from the splitting of the in-phase doublets of NOESY peaks involving amide protons (25), but only those couplings were included that were indicative of non-rotationally averaged torsion angles. From all unambiguously assigned NOESY peaks, 14 3 J ␣␤ coupling constants and 5 3 J NH␣ coupling constants (Ͻ6 Hz), 274 meaningful upper distance limits, as well as 158 , , 1 , and 2 torsion angle restraints, were derived. The calculation was performed by restrained molecular dynamics in torsion angle space using a simulated annealing protocol as implemented in the program DYANA (26). Figs. 4 and 5 were generated using MOLMOL (27).

RESULTS
[Ala 31 ,Aib 32 ]pNPY: Selectivity and Agonism at The Y 5 Receptor-The peptide was synthesized by solid-phase technique and purified by preparative HPLC. Electrospray ionization mass spectrometry and analytical HPLC confirmed peptide purity and identity. The amino acid sequence is shown in Table I.
[Ala 31 ,Aib 32 ]pNPY bound to the Y 5 receptor with an IC 50 of 6 nM, whereas values in the range of 500 to 1000 nM were observed at the other receptors (Table II). Accordingly, [Ala 31 ,Aib 32 ]pNPY turned out to be highly selective as well as to have high affinity for the Y 5 receptor.
To further characterize the newly developed Y 5 receptorselective ligand, we investigated its ability to activate the receptor. Signal transduction of NPY-activated Y receptors is mediated by G i protein coupling, which leads to the inhibition of the enzyme adenylyl cyclase (9). As shown in Table II and Fig. 1, [Ala 31 ,Aib 32 ]pNPY inhibited the forskolin-stimulated production of intracellular cAMP in a dose-dependent manner.
Taking into account that the analog was 10-fold less potent in binding and 5-fold less potent in activating the Y 5 receptor than NPY, a good correlation between affinity and activity was found for NPY and [Ala 31 ,Aib 32 ]pNPY.
The pharmacological profile of [Ala 31 ,Aib 32 ]pNPY has all the features that are required for a peptide to be referred to as a highly selective, full agonist at the Y 5 receptor, which makes it a very interesting tool for studying the biological functions that have been attributed or speculated to be related to the Y 5 receptor.
Ala 31 -Aib 32 : Key Motif for Y5 Receptor Selectivity-To further improve receptor affinity, we designed several new peptides based on sequences of PP/NPY chimeras that have been identified to bind to the Y 5 receptor with picomolar affinity (50). The PP/NPY chimeras were modified by the introduction of the Ala-Aib dipeptide at positions 31-32. These analogs will be referred to as the Ala-Aib-containing PP/NPY chimeras hereafter. All peptides were prepared by solid-phase peptide synthesis; the sequences are shown in Table I Signal transduction assays confirmed that the ligands described above were agonists at the Y 5 receptor. In particular, [cPP 1-7 ,NPY 19 -23 ,Ala 31 ,Aib 32 ,Gln 34 ]hPP was at least as efficient as NPY itself in activating the receptor (Fig. 1).
Accordingly, a new class of Y 5 receptor selective ligands has been developed in which receptor selectivity is provided by the sequence motif Ala 31 -Aib 32 .
Stimulation of Food Intake in Rats by Selective Y 5 Receptor Activation-To investigate the in vivo potency of [Ala 31 ,Aib 32 ]pNPY in stimulating food intake through activation of the Y 5 receptor, the selective agonist was administered centrally to rats at three different doses (0.2, 2.0, and 6.0 nmol), and food intake was subsequently monitored over an 8-h span (first, at 1 h and then every 2 h). The results showed a dosedependent stimulatory effect on food intake in rats treated with the NPY analog compared with the control animals (Fig. 2a). One hour after administration, the peptide, although inactive at the low dose of 0.2 nmol, induced an 8-and 10-fold increase in food intake at the higher doses of 2.0 and 6.0 nmol, respectively. After 4 h, even the dose of 0.2 nmol provoked a significant increase in food intake. The stimulation of food intake induced by [Ala 31 ,Aib 32 ]pNPY remained significant even 8 h after administration; accordingly, the in food intake increased 3-fold at a dose of 0.2 nmol, 4-fold at 2.0 nmol, and 5-fold at 6.0 nmol as compared with unstimulated consumption. After 24 h, the rats treated with [Ala 31 ,Aib 32 ]pNPY ate like the control animals.
The results of the in vivo feeding experiments clearly show that selective activation of the Y 5 receptor influences food intake in a positive manner. Furthermore, Y 5 receptor mediated stimulation of food intake depends on the administered dose of the orexigenic agent. The analog [cPP [1][2][3][4][5][6][7] ,NPY 19 -23 , Ala 31 ,Aib 32 ,Gln 34 ]hPP, which shows a significantly higher affinity than hNPY itself, turned out to increase feeding ϳ2.5fold more effectively than hNPY, which correlates nicely with the increase in affinity.
Structural Characterization of [Ala 31 ,Aib 32 ]pNPY-The solution structure of [Ala 31 ,Aib 32 ]pNPY was investigated by CD and two-dimensional NMR spectroscopy. Fig. 3 shows the CD spectrum of the pNPY analog compared with the wild-type peptide at pH 3.2. The CD spectrum of pNPY displayed the typical features of an ␣-helix with two negative bands at 220 and 208 nm and a positive band at 186 nm. The CD spectrum of the analog was characterized by a decrease in intensity, especially of the positive band and of the negative band at 220 nm. By measuring the ellipticity at 220 nm, using the equation of Chen et al. (28), the fractional helix content was calculated to be approximately 16 and 20% for the modified and the native pNPY, respectively. Accordingly, the substitutions of Ile 31 with Ala and Thr 32 with Aib led to a reduction of helicity of about 20%. Furthermore, the change in the shape of the CD curve indicates a partially different conformation of [Ala 31 , Aib 32 ]pNPY. A comparison with the CD spectra of known structures suggests an increase in ␤-turn or 3 10 -helix in the analog (29,30).
To gain a more detailed understanding of the structural differences between [Ala 31 ,Aib 32 ]NPY and NPY, we used twodimensional NMR. Distance constraints used for the structure calculation were generated from 200-ms NOESY spectra of the 2 mM sample at pH 3.2 and 37°C. Based on the hydrogen bonding pattern, it was concluded that the NPY analog is ␣-helical in peptide region 15-31 with a mean pairwise rootmean-square difference of 0.95 Å for the backbone heavy atoms. The N terminus showed no preferred conformation in solution (Fig. 4). In addition, the following observations support the view that the helix is significantly destabilized toward the C terminus.  (20), with the C ␣ H resonance of His 26 very close to its random coil value. (d) All of the 3 J NH␣ coupling constants of residues 24 -31 and 33-36 were around 7 Hz, which was the value found for rotationally averaged -backbone angles. In contrast, five 3 J NH␣ coupling constants in peptide region 14 -23 were found to be smaller than 6 Hz, clearly showing that the N-terminal part of the helix is more stable. Interestingly, the intermolecular NOEs, as previously proposed for the dimer structure of hNPY (20, 32), were not detected in this work. DISCUSSION Obesity has become one of the most common health disorders over the past two decades (33)(34)(35). Several medical problems, such as hypertension, cardiovascular diseases, type II diabetes and some forms of cancer, have been correlated with increased body weight . In the past few years, important advances have been made in understanding the mechanisms involved in food intake and energy homeostasis. It became evident that a key role in the regulation of food intake is played by the adipocytederived peptide hormone leptin and its receptor (36). Leptin levels are correlated with fat mass; accordingly, high levels of this hormone reduce food intake and body weight and stimulate metabolism. Lack of leptin or its receptor in mutant mice (ob/ob and db/db mice) induces hyperphagia, obesity, decreased energy expenditure, diabetes, and infertility (37)(38)(39). Leptin regulates the hypothalamic expression of several neuropeptides (34), among them NPY. It has been shown that ob/ob mice are characterized by an overproduction of NPY in the hypothalamus (6), whereas administration of leptin inhibits NPY secretion. Because of its stimulatory effect on food intake, NPY and its receptor are interesting targets for drug design against obesity. Accordingly, it is important to understand the mechanisms that underlie NPY-stimulated food intake. Among the currently identified NPY receptor subtypes, the Y 1 and Y 5 receptors are the most likely candidates to mediate the orexigenic action of NPY (10,11). However, many doubts still remain, as several results obtained with different methods are contradictory. Wahlestedt and co-workers (40) observed no effect on food intake after administration of Y 1 receptor antisense oligonucleotides, in contrast to the data of Lopez-Valpuesta and co-workers (41) who found inhibition of NPY-induced feeding. On the other hand, administration of antisense oligonucleotides directed against the Y 5 receptor reduced NPY-induced food intake (42). The knockout studies conducted so far have given unexpected results. Mice with both Y 1 or Y 5 receptor knockout gene exhibited mild obesity, although some differences were observed; the Y 1 receptor-deficient animal showed increased body weight without a change in food intake, mild hyperinsulemia, an elevated basal level of plasma insulin, and an absence of NPY-mediated vasoconstriction (43,44), whereas the Y 5 receptor-deficient mice presented hyperphagia (45). Inhibition of NPY-induced feeding was produced by both Y 1 and Y 5 receptor-selective nonpeptide antagonists (13,14,46,47). On the other hand, NPY receptor agonists like [Leu 31 ,Pro 34 ]NPY, NPY-(2-36), or NPY-(3-36) stimulated food intake (12). However, these peptides are not selective for one receptor, with [Leu 31 ,Pro 34 ]NPY displaying equal affinity to the Y 1 and the Y 5 receptors. In addition, N-terminally truncated analogs are ligands for both Y 2 and Y 5 receptors. Although the Y 2 receptor does not seem to mediate the NPYinduced stimulation of food intake, other actions regulated by this receptor can be elicited, which might interfere with the activation of the Y 5 receptor, thereby providing a complex and unclear physiological response. Unfortunately, the use of [D-Trp 32 ]NPY did not give unambiguous results. Gerald and co-workers (12) observed cAMP inhibition and stimulation of food intake mediated by the Y 5 receptor and concluded that this peptide is a weak Y 5 receptor-selective agonist in vitro and in vivo. In contrast, antagonism of NPY-induced stimulation of food intake by [D-Trp 32 ]NPY was found by Balasubramaniam and co-workers (16) and by Small and co-workers (17). In view of these observations, it must be stated that the in vivo profile of this analog of NPY is not well defined. Furthermore, binding potency at the Y receptors was in the order of Y 2 Ӎ Y 5 Ͼ Ͼ Y 1 ,Y 4 (Table II).
In this work, we have presented the first Y 5 receptor-selective ligands with subnanomolar affinity: [Ala 31 ,Aib 32 ]pNPY and the Ala-Aib-containing PP/NPY chimeras. The common sequence element Ala 31 -Aib 32 represents the key motif for receptor selectivity. The nonproteinogenic amino acid Aib has already been reported able to change the conformation within a peptide sequence, because no C␣-H is available to form appropriate hydrogen bonds (19). Circular dichroism studies on [Ala 31 ,Aib 32 ]pNPY revealed reduced ␣-helical content compared with NPY. The solution structure obtained by two-dimensional NMR and molecular dynamics confined the destabilization of the helix at the C-terminal end encompassing residues 32-36, for which the peptide conformation is apparently not well defined. Interestingly, the presence of an i, I ϩ 3 hydrogen bond between residues 28 and 31 and the absence of the corresponding i, i ϩ 4 bond suggests that the ␣-helix motif ends with a 3 10 -helical turn.
The in vivo feeding profile of the newly developed Y 5 receptor-selective agonists confirms that this receptor subtype is involved in the stimulation of food intake, and the potency in stimulation of feeding correlates nicely with the in vitro affinity to the Y 5 receptor. The most effective analog, [cPP 1-7 ,NPY 19 -23, Ala 31 ,Aib 32 ,Gln 34 ]hPP, turned out to be ϳ2.5-fold more potent that hNPY and accordingly thus far is the most potent stimulator of food intake. Although more aspects remain to be investigated, our results show that the Y 5 receptor certainly plays a decisive role in the complex system that controls hunger. For NPY, as well as for the equally or less potent analogs, the feeding behavior turned back to normal after 24 h, and no difference was found in the food intake of mice with and without injections; this may be because of the proteolytic cleavage of peptides or because of the uptake of the ligand receptor complex by internalization. In contrast, the effect of [cPP 1-7,NPY 19 -23 ,Ala 31 ,Aib 32 ,Gln 34 ]hPP with respect to feeding was still strong even 24 h after injection.
Until now, combined Y 1 /Y 5 or Y 2 /Y 5 ligands have been used to characterize Y 5 receptor activity. Our new compounds provide highly interesting tools and allow, for the first time, this receptor to be selectively targeted with affinities that are as potent as or even 2.5-fold more active than hNPY. Furthermore, radio-or fluorescence-labeled analogs of [Ala 31 , Aib 32 ]pNPY might be useful for Y 5 receptor specific assays, for receptor localization and investigation of receptor expression in normal and pathogenic brain structures.
The Y 5 receptor is speculated to be involved in epilepsy, sexual behavior, and circadian rhythm by activation or deactivation of special neurons in the hypothalamus or hippocampus (11). The Ala-Aib-containing agonists, with high affinity and selectivity for the Y 5 receptor, can contribute to the elucidation of the (patho)physiological relevance of this receptor subtype. Accordingly, we have already demonstrated that the Y 5 receptor is important for food intake, and further anticipated activities, such as epileptic seizures, are currently subject to investigation.