In Search of NPY Y4R Antagonists: Incorporation of Carbamoylated Arginine, Aza-Amino Acids, or d-Amino Acids into Oligopeptides Derived from the C-Termini of the Endogenous Agonists.

The cross-linked pentapeptides (2R,7R)-diaminooctanedioyl-bis(Tyr-Arg-Leu-Arg-Tyr-amide) ((2R,7R)-BVD-74D, (2R,7R)-1) and octanedioyl-bis(Tyr-Arg-Leu-Arg-Tyr-amide) (2) as well as the pentapeptide Ac-Tyr-Arg-Leu-Arg-Tyr-amide (3) were previously described as neuropeptide Y Y4 receptor (Y4R) partial agonists. Here, we report on a series of analogues of (2R,7R)-1 and 2 in which Arg2, Leu3, or Arg4 were replaced by the respective aza-amino acids. The replacement of Arg2 in 3 with a carbamoylated arginine building block and the extension of the N-terminus by an additional arginine led to the high-affinity hexapeptide Ac-Arg-Tyr-Nω-[(4-aminobutyl)aminocarbonyl]Arg-Leu-Arg-Tyr-amide (35), which was used as a precursor for a d-amino acid scan. The target compounds were investigated for Y4R functional activity in assays with complementary readouts: aequorin Ca2+ and β-arrestin 1 or β-arrestin 2 assays. In contrast to the parent compounds, which are Y4R agonists, several ligands were able to suppress the effect elicited by the endogenous ligand pancreatic polypeptide and therefore represent a novel class of peptide Y4R antagonists.

Introduction

Among the neuropeptide Y (NPY) receptors, designated Y1R, Y2R, Y4R, and Y5R,[1] the Y4R plays a special role because it preferentially binds pancreatic polypeptide.[2] Compared to the Y1R, Y2R, and Y5R subtypes, fewer ligands by far (e.g., see Figure ) have been reported for the Y4R, in particular, high-affinity Y4R antagonists are still lacking.[3−6]

Figure 1

Amino acid sequence of human PP and structures of the described NPY Y4R agonists 1–3 and the Y1R/Y4R antagonist UR-MK188 (4).

Y4R agonists are considered to be potential antiobesity agents.[7,8] The diastereomeric mixture of the cross-linked (“dimeric”) pentapeptide 1 (BVD-74D), a mimic of the C-terminus of pancreatic polypeptide, was described as a high-affinity Y4R agonist,[4] having an effect on food intake in mice.[9] As reported previously, (2R,7R)-1 (note: in the following, positions 2 and 7 refer to the stereo centers in the 2,7-diaminosuberic acid moiety) was, by a factor of 3–10, more potent than (2S,7S)-1, depending on the type of assay.[10,11] Even though Y4R agonists should have higher clinical potential, antagonists are of interest as well, in particular, as pharmacological tools. The pentapeptide 3,[6] a Y4R agonist with an affinity in the two-digit nanomolar range (Ki = 50 nM), shares the same amino acid sequence with the peptide moieties of 1. The incorporation of cyclic β-amino acids such as (1R,2S)-2-aminocyclopentane-1-carboxylic acid resulted in partial agonism.[6]

To investigate the impact of backbone modifications in Y4R agonists on functional activity in more detail, additional analogues of (2R,7R)-1, 2,[11] and 3 were prepared. Here, we report on the synthesis of Y4R ligands in which aza-amino acids or d-amino acids were introduced. The title compounds were characterized in binding and functional cellular assays with complementary readouts.

Results and Discussion

Chemistry

Aza-peptides contain at least one amino acid in which the α-carbon atom is replaced by nitrogen.[12] The semicarbazide substructure reduces conformational flexibility.[13,14] Several studies have revealed that the incorporation of aza-amino acids into bioactive peptides may result in a longer duration of action or higher potencies compared to the parent peptide.[15−17] Aza-peptides can be prepared by the reaction of an N-protected, N′-substituted hydrazine with an isocyanate.[12] However, the synthesis of aza-peptides on a solid phase is compromised by an intramolecular side-reaction of resin-bound isocyanates that results in the formation of hydantoins (Figure A), lowering the yields of the target compounds and requiring time-consuming purification.[18] To circumvent hydantoin formation, N-Fmoc-aza1-dipeptides were prepared from N-protected, N′-substituted hydrazines and benzyl ester-protected amino acids, followed by hydrogenolytic debenzylation in solution (Figure B).[19]

Figure 2

Synthesis of aza-peptides on a solid phase (A) or in solution (B).

Two different Fmoc-protected hydrazines were synthesized to mimic the side chains of Orn or Leu, respectively (Scheme ). 3-Aminopropan-1-ol was Boc-protected and subsequently oxidized under Parikh–Döring conditions using dimethyl sulfoxide (DMSO) as an oxidant to form the aldehyde 7. Condensation of (9H-fluoren-9-yl)methyl hydrazinecarboxylate 8 and the aldehyde 7 or isobutyric aldehyde resulted in hydrazones 9 and 11. Subsequent reduction using sodium cyanotrihydridoborate gave access to the substituted hydrazines 10 and 12.

Scheme 1

Synthesis of the Hydrazines 10 and 12

Reagents and conditions: (a) Boc2O, Na2CO3, tetrahydrofuran (THF), 0 °C, 10 min, followed by room temperature (rt), 14 h; (b) sulfur trioxide pyridine complex, DMSO, triethylamine (TEA), CH2Cl2, 0 °C, 1 h, followed by rt, 3 h; (c) Fmoc-Cl, MeCN/H2O (1:1), 0 °C, 10 min, followed by rt, 12 h; (d) 7, CH2Cl2, 14 h; (e) isobutyraldehyde, CH2Cl2, 50 h; (f) sodium cyanotrihydridoborate, CH2Cl2/MeOH (4:3), 2 M HCl (pH 2–3), rt, 2–6 h.

The benzyl ester-protected amino acids 14–16 were treated with triphosgene to give the respective isocyanates in situ (Scheme ). Without purification, the isocyanates were treated with hydrazines 10 and 12 affording the benzyl-protected N-Fmoc-aza1-dipeptides in yields >70%. Subsequent ester cleavage by hydrogenation gave access to the N-Fmoc-aza1-dipeptides Fmoc-aza-Leu-Arg(Pbf)-OH (17), Fmoc-aza-Orn(Boc)-Tyr(tBu)-OH (18), and Fmoc-aza-Orn(Boc)-Leu-OH (19). The dipeptides 17–19 were employed in solid phase synthesis under the same conditions as those for the standard Fmoc-protected amino acids, resulting in quantitative coupling efficiencies.

Scheme 2

Synthesis of the N-Fmoc-aza1-dipeptides 17–19

Reagents and conditions: (a) Benzyl alcohol, 4-dimethylaminopyridine (4-DMAP), dicyclohexylcarbodiimide (DCC), CH2Cl2, rt, 14 h, followed by DEA/CH2Cl2 (1:8), rt; (b) triphosgene, N,N-diisopropylethylamine (DIPEA), CH2Cl2, rt, 90 min, followed by Pd/C, H2, MeOH, rt, 1 h.

To optimize the coupling of a resin-bound aza1-peptide, the model peptide H-Ala-aza-Leu-Arg(Pbf)-NH2 (20) was synthesized under various conditions (Scheme ). When 3-[bis(dimethylamino)methyliumyl]-3H-benzotriazol-1-oxide hexafluorophosphate/hydroxybenzotriazole (HBTU/HOBt) was used as coupling reagent at 35 °C, only 5–10% of the desired products were formed. The same holds for the application of the succinimidyl ester, Fmoc-Ala-OSu. The coupling efficiency was increased to approximately 50% when the 7-aza-benzotriazole derivative HATU was used as coupling reagent. The best results (>50% coupling efficiency) were achieved by in situ activation of Fmoc-Ala-OH to the corresponding acyl chloride using triphosgene in THF. Surprisingly, when CH2Cl2 was used as a solvent instead of THF, only minor amounts of 20 were detected.

Scheme 3

Incorporation of aza-Leu into the Model Peptide 20 via SPPS

Reagents and conditions: (a) SPPS (Fmoc strategy), 17/HBTU/HOBt/DIPEA (3/3/3/6 equiv), solvent DMF, 35 °C, 14 h; Fmoc deprotection was carried out with 20% piperidine in DMF/N-methyl-2-pyrrolidone (NMP) (8:2), rt, 2 × 10 min; (b) variation of the conditions according to the table; Fmoc deprotection as under (a); (c) Fmoc deprotection, followed by CH2Cl2/trifluoroacetic acid (TFA) (97:3), 10 × 6 min.

The dimeric aza-peptide (2R,7R)-26, an aza-Leu analogue of (2R,7R)-1, was prepared from the protected pentapeptide 21 and the diacid (R,R)-24 (Scheme ). For the preparation of the aza-peptides 27–31, octanedioic acid disuccinimidyl ester (25) was used for cross-linking. Cross-linking of 21 with 25 yielded the dimeric aza-peptide 27, an aza-Leu analogue of 2. Dimerization of the protected peptides 22 and 23 resulted in the aza-peptides 28 and 30 containing ornithine instead of arginine. Subsequently, guanidinylation was performed with N,N′-di-Boc-1H-pyrazole-1-carboxamidine resulting in 29 and 31, which were aza-Arg-containing analogues of 2.

Scheme 4

Synthesis of “Homodimeric” Aza-Peptides 26–31

Reagents and conditions: (a) SPPS (Fmoc strategy), Fmoc-aa/HBTU/HOBt/DIPEA (5/5/5/10 equiv), solvent DMF/NMP (8:2), “double” coupling at rt, 60 min or dipeptide (17, 18, or 19)/HBTU/HOBt/DIPEA (3/3/3/6), solvent: DMF, 35 °C, 14 h; Fmoc deprotection as under (a) in Scheme ; (b) triphosgene, DIPEA, THF, 35 °C, 14 h; Fmoc deprotection as under (a) in Scheme ; (c) CH2Cl2/TFA (97:3), 10 × 6 min; (d) HBTU, HOBt, DIPEA, anhydrous DMF, 35 °C, 16 h; (e) 1 equiv of 25 in 1% DIPEA in anhydrous DMF, followed by 2.5 equiv of 21, 22, or 23, 35 °C, 16 h; (f) TFA/H2O (95:5), rt, 2.5 h; (g) N,N’-di-Boc-1H-pyrazole-1-carboxamidine, DMF, DIPEA, rt, 4 h; Boc-deprotection, TFA/CH2Cl2/H2O (10/10/1), rt, 3 h.

Very recently, we reported that the replacement of Arg2 with the Nω-carbamoylated arginine 32 (Figure ) in one of the pentapeptide chains of 2 led to increased Y4R affinity.[11] Similarly, 32 was used for the preparation of peptides 33–42 (Figure ), which are analogues of the pentapeptide 3, on an Fmoc-Sieber-PS resin by manual SPPS according to the Fmoc strategy.

Figure 3

Structure of the Nω-carbamoylated arginine building block 32.[20]

Figure 4

Primary structures of all investigated peptides.

Finally, the applicability of 32 to the preparation of cyclic peptides via head-to-side-chain cyclization was investigated. Cyclic peptides are structurally more constrained than their linear counterparts and can mimic secondary structural elements of native peptides or proteins. Additionally, cyclic peptides are less prone to enzymatic degradation. The formation of head-to-tail cyclic peptides is often compromised by rather rigid backbones. However, the side-chain of 32 is very flexible, facilitating an approximation of the activated C-terminus and the terminal amino group. The hexapeptide 43 was prepared on an H-l-Tyr(tBu)-2CT resin followed by global side-chain deprotection (Scheme ). Cyclization was achieved at a peptide concentration of 1 mM using benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate (PyBOP) as coupling reagent. Under these conditions, the formation of peptide dimers was not observed.

Scheme 5

Synthesis of the Head-to-Side-Chain Cyclic Peptide 43

Reagents and conditions: (a) SPPS (Fmoc strategy), Fmoc-aa/HBTU/HOBt/DIPEA (5/5/5/10 equiv), solvent DMF/NMP (8:2), double coupling at rt, 60 min or 32/HBTU/HOBt/DIPEA (3/3/3/6), solvent: DMF, 35 °C, 14 h; Fmoc deprotection as under (a) in Scheme ; (b) CH2Cl2/TFA (97:3), 10 × 6 min; (c) TFA/H2O (95:5), rt, 2.5 h; (d) PyBOP, HOBt, DIPEA, DMF, rt, 24 h.

Competition Binding Studies at NPY Receptor Subtypes. Y4R Binding

The Ki values of all of the target compounds were determined in competition binding studies on live cells expressing Y4R (CHO-hY4R-Gqi5-mtAEQ cells) using the radioligand [3H]UR-KK193 (45, Figure ) (Table , Figure ).[11] The affinities of the aza-peptides 27 (Ki = 134 nM), 29 (Ki = 52.7 nM), and 31 (Ki = 113 nM) were lower than the affinity of the respective reference compound 2 (Ki = 3.5 nM). (R,R)-2,7-Diaminosuberic acid was superior to octanedioic acid as a linker in the dimeric aza-peptides (cf. (2R,7R)-26: Ki = 30.8 nM), as in the case of peptide 1 compared to 2. The aza-Orn precursors of 29 and 31, compounds 28 (Ki = 130 nM) and 30 (Ki = 385 nM), bind to the Y4R as well, though with lower affinities compared to their aza-Arg counterparts.

Figure 5

Structures of the Y4R radioligand [3H]UR-KK193 (45) and the Y1R radioligand [3H]UR-MK136 (46).

Table 1

Binding Data at NPY Receptor Subtypes

	Y1R	Y2R	Y4R	Y5R
compd	Ki [nM]a	Ki [nM]b	Ki [nM]c	Ki [nM]d
hPP	440e	>5000e	0.65f	17e
(2R,7R)-1	440f	830f	0.45f	1500f
2	720f	1700f	3.5f	280f
3	>5000	>5000	337 ± 110	>5000
4	24g	920g	660f	>5000g
(2R,7R)-26	1840 ± 380	>5000	30.8 ± 7.5	>5000
27	574 ± 49	>5000	134 ± 20	2700 ± 490
28	n.d.	n.d.	130 ± 39	n.d.
29	3730 ± 350	>5000	52.7 ± 7.1	3560 ± 440
30	n.d.	n.d.	385 ± 127	n.d.
31	53 ± 12	3270 ± 870	113 ± 14	2100 ± 240
33	589 ± 120	>5000	56.9 ± 5.3	3440 ± 210
34	>5000	>5000	19.5 ± 2.6	>5000
35	4830 ± 1400	>5000	2.87 ± 0.78	>5000
36	1180 ± 380	>5000	3.43 ± 1.3	>5000
37	1120 ± 270	>5000	501 ± 200	>5000
38	>5000	>5000	724 ± 140	>5000
39	>5000	>5000	568 ± 76	>5000
40	>5000	>5000	248 ± 23	>5000
41	>5000	>5000	10.4 ± 3.3	>5000
42	>5000	>5000	40.5± 15	>5000
43	>5000	>5000	>5000	>5000
44	>5000	>5000	1650 ± 290	>5000

Radioligand competition binding assay with 46(22) (Kd = 6.2 nM, c = 4 nM) using MCF-7-hY1 cells.[23]

Radioligand competition binding assay with [3H]propionyl-pNPY[24] (Kd = 1.4 nM, c = 1 nM) using CHO-hY2-Gqi5-mtAEQ cells.[25]

Radioligand competition binding assay with 45(11) (Kd = 0.67 nM, c = 0.6 nM) using CHO-hY4R-Gqi5-mtAEQ cells.[26]

Radioligand competition binding assay with [3H]propionyl-pNPY (Kd = 4.83 nM, c = 4 nM) using HEC-1b hY5R cells.[27]

Ki value reported by Berlicki et al.[6]

Ki value reported by Kuhn et al.[11]

Ki value reported by Keller et al.[5] Presented are mean values ± SEM from at least three independent experiments (performed in triplicate). n.d.: not determined.

Figure 6

Competition binding performed with the radioligand 45 at CHO-hY4R-Gqi5-mtAEQ cells. (A–C) Displacement of 45 (Kd = 0.67 nM, c = 0.6 nM). Data represent mean values ± SEM of at least three independent experiments, each performed in triplicate.

Among the structural modifications of pentapeptide 3, replacement of the N-terminal acetic acid with octanoic acid (compound 33) resulted in an increase in Y4R affinity by a factor of almost six. The replacement of Arg2 in 3 with the carbamoylated arginine 32 was even more favorable: the affinity of 34 (Ki = 19.5 nM) was 17-fold higher than the affinity of its parent compound 3. Extension of the N-terminus by an additional arginine led to a further increase in affinity: the Ki value of compound 35 was in the single-digit nanomolar range (Ki = 2.87 nM). Up to now, comparable binding data were achieved only in the case of cross-linked or significantly longer linear peptides. The extension of the N-terminus of 3 by an additional arginine alone had almost the same impact on Y4R affinity, that is, 36 (Ki = 3.43 nM) showed nearly the same affinity as peptide 35. Interestingly, the replacement of the arginine adjacent to the C-terminus with the modified arginine building block 32 was not tolerated (37: Ki = 501 nM).

Following the structural optimization of 3, a d-amino acid scan was performed with the hexapeptide 35. Whereas replacement of Arg1 or Tyr2 with d-Arg or d-Tyr, respectively, was tolerated (cf. 41, 42), replacement of Leu4, Arg5, or Tyr6 by the corresponding d-amino acid led to a marked decrease in Y4R affinity (cf. 38–40).

Previous studies revealed that C-terminal amidation is crucial for Y4R binding of the endogenous ligand hPP.[21] Similarly, the affinity of the carboxylic acid 43 (IC50 > 5000 nM) was much lower than the affinity of the amide 35. Cyclization of 43 resulted in an increase in Y4R affinity. However, the cyclic peptide 44 (Ki = 1650 nM) was only a weak binder compared to the linear analogue 35.

NPY Receptor Subtype Selectivity

Ki values of all of the target compounds were also determined at the Y1, Y2, and Y5 receptors using the radioligands [3H]propionyl-pNPY (Y2R and Y5R) or [3H]UR-MK136 (46, Y1R, Figure ). Neither the linear peptides 34–43 nor the cyclic peptide 44 displayed remarkable affinity to the Y1, Y2, or Y5 receptors. Having a more than 1000-fold selectivity for the Y4R over the other NPY receptors, peptide 35 proved to be superior to (2R,7R)-1 and 2. Only the pentapeptide 33 bound to the Y1R with a Ki value in the submicromolar range (Ki (Y1R) = 589 nM), that is, the lipophilic N-terminal octanoyl residue is disadvantageous with respect to Y4R selectivity.

The cross-linked aza-peptides (2R,7R)-26, 27, and 29 displayed only low Y1R, Y2R, and Y5R affinities, whereas the affinity of 31 (Ki = 53 nM) at the Y1R was markedly higher than that of the peptide analogue 2. In a Fura-2 Ca2+ assay on human erythroleukemia (HEL) cells, 31 revealed Y1R antagonism (Figure ) with a Kb value of 11.5 nM.

Figure 7

Y1R antagonism of 31. Inhibition of pNPY (c = 10 nM, EC50 = 0.87 nM)-induced intracellular Ca2+ mobilization in HEL cells by 31 (Kb = 11.5 ± 1.1 nM), determined in a Fura-2 assay. Mean ± SEM from three independent experiments.

Functional Studies at the Y4R

The target compounds were investigated for Y4R agonism and antagonism in an aequorin Ca2+ assay as well as β-arrestin 1 and β-arrestin 2 recruitment assays on genetically engineered CHO or HEK293T cells, respectively (for data see Table , Figures –10, SI Figures 1–8).

Table 2

NPY Y4R Agonist Potencies (EC50) and Intrinsic Activities (α) or Antagonism (IC50 Values)

	aequorin assaya		β-arrestin 1 assayb		β-arrestin 2 assayb
compd	EC50 or IC50 [nM]	α	EC50 or IC50 [nM]	α	EC50 or IC50 [nM]	α
hPP	9.7c	1	3.54 ± 0.59	1	2.74 ± 0.23	1
(2R,7R)-1	6.9c	0.62	40.4 ± 11	0.54	31.8 ± 12	0.58
2	49c	0.73	n.d.	n.d.	n.d.	n.d.
3	6390 ± 450	0.55	n.d.	n.d.	n.d.	n.d.
4	226d	n.a.	187 ± 72	n.a.	202 ± 78	n.a.
(2R,7R)-26e	388 ± 51	n.a.	5430 ± 350	n.a.	1320 ± 230	n.a.
27f	1430 ± 290	n.a.	1090 ± 350	n.a.	712 ± 87	n.a.
29	542 ± 110	0.07	875 ± 260	0.09	473 ±65	0.12
	572 ± 59	n.a.	n.d.	n.d.	n.d.	n.d.
31	1340 ± 530	n.a.	n.d.	n.d.	n.d.	n.d.
33	2050 ± 84	0.44	2200 ± 130	0.56	1790 ± 1070	0.61
34	1460 ± 83	0.61	1070 ± 520	0.62	568 ± 120	0.60
35	187 ± 23	0.58	144 ± 28	0.48	124.1 ± 24	0.55
36	303 ± 68	0.78	187 ± 40	0.51	140 ± 45	0.54
37	9580 ± 2000	n.a.	6640 ± 780	n.a.	8050 ± 120	n.a.
38	18 800 ± 6500	n.a.	1310 ± 390	n.a.	1120 ± 250	n.a.
39	8340 ± 2900	n.a.	n.d.	n.d.	n.d.	n.d.
40	6000 ± 2600	n.a.	n.d.	n.d.	n.d.	n.d.
41	608 ± 140	0.61	2050 ± 780	0.46	710 ± 200	0.49
42	756 ± 230	0.63	2160 ± 600	0.45	624 ± 150	0.56
44	12 400 ± 5500	n.a.	5390 ± 1630	n.a.	3940 ± 660	n.a.

Aequorin calcium mobilization assay on CHO-hY4-Gqi5-mtAEQ cells.[26] Antagonism (IC50 values given in italics) was determined in the presence of 100 nM hPP (EC50 = 9.7 nM).

β-Arrestin 1,2 recruitment assay on HEK293T cells stably expressing the Y4R and ARRB1 or ARRB2. Antagonism (IC50 values given in italics) was determined in the presence of 3 nM hPP (EC50 (ARRB1) = 3.54 nM; EC50 (ARRB2) = 2.74 nM).

EC50 value reported by Kuhn et al.[11]

Reported as Kb = 20 nM by Keller et al.[5]

Y4R agonism on HEK293-hY4R-CRE Luc cells (luciferase reporter gene assay): EC50 = 176 ± 50 nM, α = 0.81.

Y4R agonism on HEK293-hY4R-CRE Luc cells (luciferase reporter gene assay): EC50 = 220 ± 73 nM, α = 0.75. Presented are mean values ± SEM from at least three independent experiments (performed in triplicate). n.a.: not applicable; n.d.: not determined.

Figure 8

Y4R agonism of (2R,7R)-1, 33–36, 41, and 42 determined in a calcium (aequorin) assay, a β-arrestin recruitment assay, and a luciferase reporter gene assay. (A, B) Induced intracellular Ca2+ mobilization in CHO-hY4R-mtAEQ-Gqi5 cells. (C, D) Induced β-arrestin 1 recruitment in HEK293T-ARRB1-Y4R cells. (E, F) Induced β-arrestin 2 recruitment in HEK293T-ARRB2-Y4R cells. Data points shown for (A)–(F) are the mean ± SEM of at least three independent experiments performed in triplicate.

Figure 10

Y4R functional activity of 38 and 44 determined in a calcium (aequorin) assay and a β-arrestin recruitment assay. (A) Inhibition of hPP (EC50 = 9.7 nM, c = 100 nM)-induced intracellular Ca2+ mobilization in CHO-hY4R-mtAEQ-Gqi5 cells. (B) Inhibition of hPP (EC50 (Arr1) = 3.54 nM, EC50 (Arr2) = 2.74 nM, c = 3 nM)-induced β-arrestin 1,2 recruitment in HEK293T-ARRB1-Y4R cells (β-arrestin 1, solid lines) and HEK293T-ARRB2-Y4R cells (β-arrestin 2, dashed lines). Data points shown for (A) and (B) are the mean ± SEM of at least three independent experiments performed in triplicate. Note: Inhibition of hPP-induced response appears incomplete, however, concentrations of antagonists higher than those given in the curves could not be applied due to a shortage of the test compounds. There was no intrinsic activity of both compounds in the respective assays when performed in the agonist mode (SI Figures 3, 5, and 7).

Y4R Agonism

(2R,7R)-1 was a partial agonist in both the Ca2+ assay and the β-arrestin recruitment assays (α = 0.62, (Ca2+ assay), α = 0.54 (β-arrestin 1) or 0.58 (β-arrestin 2), Figure A,C,E). The pentapeptides 33 and 34, and the hexapeptides 35 and 36, were partial agonists achieving between 44 and 78% of the maximal response of hPP in both the calcium and the arrestin assay (Figure ). Y4R agonism was retained, though at a lower level, when d-amino acids were introduced into the N-terminal part of 35 (cf. 41, 42; Figure ). In contrast, the aza-peptides (2R,7R)-26, 27, and 31, the carbamoyl-Arg modified peptides 37–40, including those bearing d-amino acids in the C-terminus (38–40), as well as the cyclic hexapeptide 44 were devoid of Y4R agonism in the aequorin Ca2+ assay and the β-arrestin recruitment assays (SI Figures 2–7). The aza-peptide 29 displayed weak partial agonism with extremely low intrinsic activity (EC50 = 542 nM, α = 0.07, (Ca2+ assay); EC50 = 875 nM, α = 0.09 (β-arrestin 1); EC50 = 473, α = 0.12 (β-arrestin 2)). Compounds (2R,7R)-26 and 27 were also investigated for Y4R agonism in a luciferase reporter gene assay. Surprisingly, both aza-peptides displayed partial agonism in this assay (α = 0.81, EC50 = 176 nM ((2R,7R)-26), α = 0.75, EC50 = 220 nM (27)). Functional assays with distal readouts such as the luciferase reporter gene assay can reflect pronounced signal amplifications. It should be taken into account that the incubation period in the luciferase reporter gene assays was 4.5 h compared to 60 min for the arrestin assays and a few seconds in the case of the aequorin assay. Therefore, compounds displaying negligible agonistic activity in assays with proximal readouts may show much higher efficacies in reporter gene assays, as demonstrated, for example, for histamine H4 receptor ligands.[28] Likewise, (2R,7R)-1, a partial agonist in the Ca2+ assay and the arrestin assay, appeared as a full agonist in the luciferase assay.[11]

Y4R Antagonism

The incorporation of aza-amino acids into the cross-linked peptides (2R,7R)-1 and 2 had an impact on the quality of action. The dimeric aza-peptides (2R,7R)-26, 27, 29, and 31 were able to suppress the response elicited by the endogenous ligand hPP (Figure , SI Figure 8). The introduction of d-amino acids into the C-terminal tripeptide of 35 led to Y4R antagonism, with compound 40, in which Leu4 was replaced by d-Leu, displaying the strongest antagonism in the Ca2+ assay (Table , SI Figure 8). Compound 37, in which Arg5 was replaced by the carbamoylated arginine 32, and the cyclic hexapeptide 44 were identified as weak Y4R antagonists (Figure , SI Figure 1).

Figure 9

Y4R antagonism of (2R,7R)-26 and 27 determined in a calcium (aequorin) assay, a β-arrestin recruitment assay, and a luciferase reporter gene assay. (A) Inhibition of hPP (EC50 = 9.7 nM, c = 100 nM)-induced Ca2+ mobilization in CHO-hY4R-mtAEQ-Gqi5 cells. (B) Inhibition of hPP (EC50 (Arr1) = 3.54 nM, EC50 (Arr2) = 2.74 nM, c = 3 nM)-induced β-arrestin 1,2 recruitment in HEK293T-ARRB1-Y4R cells (β-arrestin 1, solid lines) and HEK293T-ARRB2-Y4R cells (β-arrestin 2, dashed lines). (C) Inhibition of forskolin-stimulated (2 μM) luciferase activity (corresponding to 100%) in hY4R expressing HEK293 cells with the maximum inhibitory effect of the endogenous ligand hPP, which was set to 0% luciferase activity and corresponds to full agonism (α = 1.0). Data points shown for (A)–(C) are the mean ± SEM of at least three independent experiments performed in triplicate.

In summary, modification of the C-terminal part of the monomeric and cross-linked peptide partial agonists by the introduction of aza-amino acids or d-amino acids as well as rigidization of the backbone by head-to-side-chain cyclization changed the quality of action to antagonism in the aequorin and β-arrestin assays. Unfortunately, these modifications went along with a decrease in Y4R affinity. The EC50 values of the reported Y4R partial agonists (33–36, 41, 42) were substantially higher than the Ki values. This effect was previously discussed and can be attributed to the presence or absence of sodium ions in the buffers.[11,29]

Conclusions

The affinity and potency of the Y4R partial agonist 3 was considerably increased by the replacement of Arg2 with the carbamoylated arginine 32 and the introduction of an additional arginine at the N-terminus. Several backbone modifications of the C-terminal tripeptide (e.g., the introduction of aza-amino acids or d-amino acids) of 35 as well as head-to-side-chain cyclization changed the quality of action of the linear or dimeric Y4R peptide ligands from partial agonism to antagonism. The resulting compounds displayed weaker Y4R antagonism than the reported Y4R antagonist 4. However, due to their extremely facile synthetic accessibility, linear peptides derived from 35 containing d-amino acids (e.g., 40 (Ki = 248 nM)) are promising building blocks for further structural modifications and might pave the way for the development of peptide Y4R antagonists with increased affinity.

Experimental Section

Chemistry: General Conditions

Chemicals and solvents were purchased from commercial suppliers and used without further purification unless otherwise indicated. DMF for peptide synthesis, NMP for peptide synthesis, 4-DMAP, sodium cyanotrihydroborate, and HOBt hydrate were from Acros Organics/Fisher Scientific (Nidderau, Germany). Fmoc-Sieber-PS resin (0.61 mmol/g), Fmoc-Arg(Pbf)-OH, Fmoc-Tyr(tBu)-OH, H-l-Tyr(tBu)-2CT resin (0.68 mmol/g), PyBOP, and HBTU were from Iris Biotech (Marktredwitz, Germany). MeCN for HPLC (gradient grade), Fmoc-Leu-OH, DCC, sulfur trioxide pyridine complex, hydrazine hydroxide, and methanol were from Merck (Darmstadt, Germany). Trifluoroacetic acid, TEA, triphosgene, Fmoc chloride, isobutyraldehyde, DEA, CH2Cl2, diethyl ether, and Triton X-100 were from Sigma-Aldrich (Deisenhofen, Germany); di-tert-butyl dicarbonate (>97%) and 3-aminopropan-1-ol were from Alfa Aesar (Karlsruhe, Germany). DIPEA (99%) was from ABCR (Karlsruhe, Germany), and Fmoc-Ala-OSu was from Advanced Chemtech (Louisville, KY). Bovine serum albumin (BSA) and bacitracin were from Serva (Heidelberg, Germany) and coelenterazin h was obtained from Biotrend (Cologne, Germany). Human pancreatic polypeptide and porcine NPY were from Synpeptide (Shanghai, China). The synthesis of (2R,7R)-1,[11]3,[6]4,[5]24,[11]25,[30]32,[20]45,[11]46,[22] and [3H]propionyl-pNPY[24] was previously described. Millipore water was used throughout for the preparation of buffers and HPLC eluents. Stock solutions of test compounds were prepared in Millipore water containing 0.1% TFA. Polypropylene reaction vessels (1.5 or 2 mL) with a screw cap (Süd-Laborbedarf, Gauting, Germany) were used for small scale reactions (e.g., the preparation of 44) and for the storage of stock solutions. Thin-layer and column chromatography, NMR spectroscopy, mass spectrometry, preparative and analytical HPLC, as well as freeze-drying were performed as described previously.[11]

Compound Characterization

Compounds were characterized as described previously.[11] Purities determined by reversed-phase high-performance liquid chromatography (RP-HPLC) were >95%.

Chemistry: Experimental Protocols and Analytical Data. General Procedure for Solid Phase Peptide Synthesis

Peptides were synthesized by manual SPPS using the Fmoc strategy on an Fmoc-Sieber-PS resin or an H-l-Tyr(tBu)-2CT resin (43) as described with minor modifications. Five milliliters Discardit II syringes (Becton Dickinson, Heidelberg, Germany) were equipped with 35 μm polyethylene frits (Roland Vetter Laborbedarf, Ammerbuch, Germany) and used as reaction vessels. For the coupling of standard d- or l-amino acid to the N-terminus of an amino acid or the amino group of the Fmoc-Sieber-PS resin, DMF/NMP (8:2) was used as solvent. Fmoc amino acids (5-fold excess) were preactivated with HBTU/HOBt/DIPEA (5/5/10 equiv) for 5 min and added to the resin. Double coupling at rt was performed for all standard amino acids for 45 min. The coupling of the arginine building block 32 and the dipeptides 17–19 (3-fold excess, preactivated with HBTU/HOBt/DIPEA (3/3/6 equiv)) was performed at 35 °C for 14 h (“single coupling”) using anhydrous DMF as solvent. For the coupling of Fmoc amino acids to a resin-bound aza-amino acid, triphosgene was used as the coupling reagent. Triphosgene (1.75 equiv) was dissolved in anhydrous THF and cooled to 0 °C. A solution of Fmoc amino acid (5 equiv) and DIPEA (15 equiv) in anhydrous THF was added dropwise causing the formation of a white precipitate. After addition, stirring was continued for 5 min. The suspension was centrifuged and the supernatant was added to the resin. Coupling was performed at 35 °C for 12 h.

After coupling was completed, the resin was washed with DMF/NMP and treated with 20% piperidine in DMF/NMP (8:2) at rt (2×) for 10 min to remove the Fmoc group, followed by thorough washing of the resin.

Benzyl (S)-2-Amino-3-[4-(tert-butoxy)phenyl]propanoate (15)

Fmoc-Tyr(tBu)-OH (2 g, 4.35 mmol, 1 equiv), 4-DMAP (53.2 mg, 0.1 equiv), and benzyl alcohol (1 mL, 2.2 equiv) were dissolved in CH2Cl2 (25 mL) and a solution of DCC (943 mg, 1.05 equiv) in CH2Cl2 (6 mL) was added dropwise under ice-cooling. The mixture was stirred at rt for 14 h. After completion of the reaction (monitored by thin-layer chromatography (TLC) (light petroleum/EtOAc 2:1): R = 0.76), the mixture was filtered, and the filtrate was washed with 0.1 M HCl solution (30 mL) and brine (20 mL). The volume of the organic layer was adjusted to 70 mL with CH2Cl2, and diethylamine (12 mL) was added for removal of the Fmoc group. After all of the starting material had been consumed, the volatiles were removed on a rotary evaporator, and the residue was taken up in light petroleum/EtOAc (1:1, 10 mL) and subjected to column chromatography (eluent: light petroleum/EtOAc 1:1 → EtOAc/MeOH 4:1). The desired compound was obtained as a yellowish highly viscous oil (1.05 g, 73.7%). 1H NMR (300 MHz, DMSO-d6): δ (ppm) 1.26 (s, 9H), 1.80 (br s, 2H), 2.71–2.89 (m, 2H), 3.60 (t, 2H, J 6.9 Hz), 5.05 (s, 2H), 6.84 (d, 2H, J 8.5 Hz), 7.06 (d, 2H, J 8.4 Hz), 7.25–7.39 (m, 5H). 13C NMR (100.6 MHz, DMSO-d6): δ (ppm) 29.0 (3 carb), 56.4, 66.0, 78.0, 123.8 (2 carb), 128.40 (2 carb), 128.43, 128.8 (2 carb), 130.2 (2 carb), 132.9, 136.5, 153.9, 175.4. HRMS (ESI): m/z [M + H]+ calcd for [C20H26NO3]+: 328.1907, found: 328.1915. C20H25NO3 (327.42).

N2-(2-{[(9H-Fluoren-9-yl)methoxy]carbonyl}-1-isobutylhydrazine-1-carbonyl)-Nω-[(2,2,4,6,7-pentamethyl-2,3-dihydrobenzofuran-5-yl)sulfonyl]-l-arginine (17)

Triphosgene (152 mg, 0.4 equiv) was dissolved in CH2Cl2 (5 mL). A solution of 14 (145 mg, 0.442 mmol) and DIPEA (450 μL, 2 equiv) in CH2Cl2 (5 mL) was added dropwise under ice-cooling. The mixture was allowed to warm to ambient temperature under stirring for 30 min. A solution of 12 (200 mg, 1.1 equiv) in CH2Cl2 (2 mL) was added and stirring was continued for 90 min. The mixture was diluted with EtOAc (60 mL) and washed with water. The volatiles were removed under reduced pressure, the residue was taken up in EtOAc/hexane (2:1, 10 mL) and subjected to column chromatography (eluent: EtOAc/hexane (2:1)). The protected intermediate was dissolved in MeOH (10 mL), and Pd/C-catalyst (50 mg) was added. The mixture was stirred vigorously under hydrogen. After completion of the hydrogenation, the solids were removed by filtration through celite, the volume was reduced on a rotary evaporator. The desired compound was purified via column chromatography (eluent: CH2Cl2/MeOH (10:1) → CH2Cl2/MeOH (1:1) + 0.5% TFA). Removal of the volatiles under reduced pressure and by lyophilization afforded 14 as a white lyophilisate (370 mg, 88.5%). TLC (CH2Cl2/MeOH (10:0.6) + 1% AcOH): R = 0.28. 1H NMR (400 MHz, methanol-d4): δ (ppm) 0.85 (s, 6H), 1.41 (s, 6H), 1.47–1.69 (m, 3H), 1.69–1.79 (m, 1H), 1.79–1.92 (m, 1H), 2.06 (s, 3H), 2.49 (s, 3H), 2.56 (s, 3H), 2.95 (s, 2H), 3.15 (s, 2H), 4.12–4.30 (m, 2H), 4.52 (s, 2H), 7.28 (t, 2H, J 7.3 Hz), 7.37 (t, 2H, J 7.3 Hz), 7.54–7.70 (m, 2H), 7.77 (d, 2H, J 7.5 Hz). HRMS (ESI): m/z [M + H]+ calcd for [C39H51N6O8S]+: 763.3484, found: 763.3494. C39H50N6O8S (762.92).

(S)-5-({[(9H-Fluoren-9-yl)methoxy]carbonyl}amino)-2-[4-(tert-butoxy)benzyl]-12,12-dimethyl-4,10-dioxo-11-oxa-3,5,9-triazatridecanoic Acid (18)

Triphosgene (52.5 mg, 0.4 equiv) was dissolved in CH2Cl2 (2 mL). A solution of 15 (145 mg, 0.442 mmol) and DIPEA (150 μL, 2 equiv) in CH2Cl2 (2 mL) was added dropwise under ice-cooling. The mixture was allowed to warm to ambient temperature under stirring for 30 min. A solution of 10 (200 mg, 1.1 equiv) in CH2Cl2 (2 mL) was added, and stirring was continued for 90 min. The mixture was diluted with EtOAc (40 mL) and washed with water. The volatiles were removed under reduced pressure, the residue was taken up in EtOAc/hexane (1:1, 10 mL) and subjected to column chromatography (eluent: EtOAc/hexane (1:1), R = 0.57). The protected intermediate was dissolved in MeOH (10 mL), and Pd/C-catalyst (30 mg) was added. The mixture was stirred vigorously under hydrogen. After completion of the hydrogenation, the solids were removed by filtration through celite, the volume was reduced on a rotary evaporator. The desired compound was purified via column chromatography (eluent: CH2Cl2/MeOH (10:0.5) → CH2Cl2/MeOH (10:0.6) + 1% AcOH). Removal of the volatiles under reduced pressure and by lyophilization afforded 18 as a white lyophilisate (221 mg, 74.1%). TLC (CH2Cl2/MeOH (10:0.6) + 0.5% AcOH): R = 0.35. 1H NMR (400 MHz, methanol-d4): δ (ppm) 1.20 (s, 9H), 1.42 (s, 9H), 1.51–1.66 (m, 2H), 2.81–3.20 (m, 5H), 3.69 (br s, 1H), 4.21 (s, 1H), 4.35–4.60 (m, 3H), 6.79 (s, 2H), 7.04 (s, 2H), 7.24–7.35 (m, 2H), 7.39 (t, 2H, J 7.2 Hz), 7.54–7.73 (m, 2H), 7.80 (d, 2H, J 7.5 Hz). HRMS (ESI): m/z [M + H]+ calcd for [C37H47N4O8]+: 675.3388, found: 675.3403. C37H46N4O8 (674.80).

{[1-(9H-Fluoren-9-yl)-12,12-dimethyl-3,10-dioxo-2,11-dioxa-4,5,9-triazatridecan-5-yl]carbonyl}-l-leucine (19)

Triphosgene (52.5 mg, 0.4 equiv) was dissolved in CH2Cl2 (2 mL). A solution of 16 (97.7 mg, 0.442 mmol) and DIPEA (150 μL, 2 equiv) in CH2Cl2 (2 mL) was added dropwise under ice-cooling. The mixture was allowed to warm to ambient temperature under stirring for 30 min. A solution of 10 (200 mg, 1.1 equiv) in CH2Cl2 (2 mL) was added and stirring was continued for 90 min. The mixture was diluted with EtOAc (40 mL) and washed with water. The volatiles were removed under reduced pressure and the residue was taken up in EtOAc/hexane (1:1, 10 mL) and subjected to column chromatography (eluent: EtOAc/hexane (1:1), R = 0.71). The protected intermediate was dissolved in MeOH (10 mL), and Pd/C-catalyst (30 mg) was added. The mixture was stirred vigorously under hydrogen. After completion of the hydrogenation, the solids were removed by filtration through celite, and the volume was reduced on a rotary evaporator. The desired compound was purified via column chromatography (eluent: CH2Cl2/MeOH (10:0.4) → CH2Cl2/MeOH (10:0.6) + 0.1% TFA). Removal of the volatiles afforded 19 as a sticky solid (132 mg, 52.5%). TLC (CH2Cl2/MeOH (10:0.6) + 0.5% AcOH): R = 0.55. 1H NMR (400 MHz, methanol-d4): δ (ppm) 0.90 (d, 6H, J 4.4 Hz), 1.42 (s, 9H), 1.52–1.77 (m, 5H), 3.03 (s, 2H), 3.35–3.95 (m, 2H), 4.23 (t, 1H, J 6.3 Hz), 4.29 (t, 1H, J 7.0 Hz), 4.52 (s, 2H), 7.31 (t, 2H, J 7.3 Hz), 7.39 (t, 2H, J 7.4 Hz), 7.59–7.71 (m, 2H), 7.79 (d, 2H, J 7.5 Hz). HRMS (ESI): m/z [M + H]+ calcd for [C30H41N4O7]+: 569.2970, found: 569.2969. C30H40N4O7 (568.67).

Tyr(tBu)-Arg(Pbf)-aza-Leu-Arg(Pbf)-Tyr(tBu)-amide Hydrotrifluoroacetate (21)

Compound 21 was synthesized according to the general procedure (100 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 42:58–78:22, tR = 14.9 min) afforded 21 as a white solid (39 mg, 42.6%). HRMS (ESI): m/z [M + 2H]2+ calcd for [C69H105N13O13S2]2+: 693.8693, found: 693.8708. C69H103N13O13S2·C2HF3O2 (1386.06 + 114.02).

Tyr(tBu)-Arg(Pbf)-Leu-aza-Orn(Boc)-Tyr(tBu)-amide Hydrotrifluoroacetate (22)

Compound 22 was synthesized according to the general procedure (100 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–20 min MeCN/0.1% aq TFA 42:58–78:22, tR = 14.2 min) gave 22 as a white solid (42 mg, 52.7%). HRMS (ESI): m/z [M + 2H]2+ calcd for [C60H95N11O12S]2+: 596.8436, found: 596.8445. C60H93N11O12S·C2HF3O2 (1192.53 + 114.02).

Tyr(tBu)-aza-Orn(Boc)-Leu-Arg(Pbf)-Tyr(tBu)-amide Hydrotrifluoroacetate (23)

Compound 23 was synthesized according to the general procedure (100 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–20 min MeCN/0.1% aq TFA 42:58–78:22, tR = 14.6 min) afforded 23 as a white solid (36 mg, 45.2%). HRMS (ESI): m/z [M + 2H]2+ calcd for [C60H95N11O12S]2+: 596.8436, found: 596.8446. C60H93N11O12S·C2HF3O2 (1192.53 + 114.02).

(2R,7R)-Diaminooctanedioyl-bis(Tyr-Arg-aza-Leu-Arg-Tyr-amide) Hexakis(hydrotrifluoroacetate) ((2R,7R)-26)

Compound 24 (2.08 mg, 5.1 μmol), HBTU (4.1 mg, 2.1 equiv), and HOBt (1.57 mg, 2 equiv) were dissolved in anhydrous DMF (400 μL). DIPEA (10 μL, 5 equiv) was added. Stirring was continued at rt for 5 min, followed by addition of a solution of 21 (19.3 mg, 12.8 μmol) in DMF. The mixture was stirred at 35 °C for 16 h. Water (10 mL) was added and the protected intermediate was extracted with CH2Cl2 (2 × 10 mL). The combined organic phases were concentrated, and the residue was dried in vacuo. TFA/water (95:5, 2 mL) was added, and the mixture was stirred at rt for 2.5 h. Water (100 mL) was added followed by lyophilization. The product was purified by preparative HPLC (gradient: 0–20 min: MeCN/0.1% aq TFA 3:97–42:58, tR = 15.6 min). Freeze-drying of the eluate gave (2R,7R)-26 as a white fluffy solid (3.45 mg, 28.3%). HRMS (ESI): m/z [M + 4H]4+ calcd for [C78H126N28O16]4+: 427.7471, found: 427.7485. RP-HPLC (220 nm): 99% (tR = 17.72 min, k = 5.2). C78H122N28O16·C12H6F18O12 (1708.01 + 684.12).

Octanedioyl-bis(Tyr-Arg-aza-Leu-Arg-Tyr-amide) Tetrakis(hydrotrifluoroacetate) (27)

Compound 21 (28.8 mg, 19.2 μmol) was dissolved in anhydrous DMF/DIPEA (99:1 v/v) (600 μL). Compound 25 (2.83 mg, 7.7 μmol) was added, and the mixture was stirred at 35 °C for 16 h. Water (10 mL) was added, and the protected intermediate was extracted with CH2Cl2 (2 × 10 mL). The combined extracts were evaporated, and the residue was dried in vacuo. TFA/water (95:5 v/v) (2 mL) was added, and the mixture was stirred at rt for 2.5 h. Water (100 mL) was added followed by lyophilization. The product was purified by preparative HPLC (gradient: 0–20 min: MeCN/0.1% aq TFA 3:97–42:58, tR = 16.5 min). Lyophilization of the eluate afforded 27 as a white fluffy solid (6.3 mg, 38.3%). HRMS (ESI): m/z [M + 4H]4+ calcd for C78H124N26O16: 420.2417, found: 420.2433. RP-HPLC (220 nm): 95% (tR = 19.13 min, k = 5.7). C78H120N26O16·C8H4F12O8 (1677.98 + 456.08).

Octanedioyl-bis(Tyr-Arg-Leu-aza-Orn-Tyr-amide) Tetrakis(hydrotrifluoroacetate) (28)

Compound 22 (16.7 mg, 12.8 μmol) was dissolved in anhydrous DMF/DIPEA (99:1 v/v) (600 μL). Compound 25 (1.9 mg, 5.1 μmol) was added, and the mixture was stirred at 35 °C for 16 h. After addition of water (10 mL), the protected intermediate was extracted with CH2Cl2 (2 × 10 mL). The combined extracts were evaporated, and the residue was dried in vacuo. TFA/water (95:5 v/v) (2 mL) was added, and the mixture was stirred at rt for 2.5 h. Water (100 mL) was added, followed by lyophilization. The product was purified by preparative HPLC (gradient: 0–20 min: MeCN/0.1% aq TFA 3:97–42:58, tR = 16.1 min). Lyophilization of the eluate afforded 28 as a white fluffy solid (4.2 mg, 40.2%). HRMS (ESI): m/z [M + 4H]4+ calcd for C76H120N22O16: 399.2308, found: 399.2327. RP-HPLC (220 nm): 99% (tR = 19.55 min, k = 5.8). C76H116N22O16·C8H4F12O8 (1593.90 + 456.08).

Octanedioyl-bis(Tyr-Arg-Leu-aza-Arg-Tyr-amide) Tetrakis(hydrotrifluoroacetate) (29)

Compound 28 (3.7 mg, 2.05 μmol) was dissolved in anhydrous DMF/DIPEA (99:1 v/v) (300 μL). A solution of N,N′-di-Boc-1H-pyrazole-1-carboxamidine (1 mg/10 μL, 16 μL, 2.5 equiv) was added, and the mixture was stirred at rt for 4 h. After addition of water (5 mL), the protected intermediate was purified by preparative HPLC (gradient: 0–25 min: MeCN/0.1% aq TFA 3:97–52:48, tR = 26.4 min). The eluates were subjected to lyophilization, and the residue was taken up in TFA/CH2Cl2/H2O (5:5:0.5, 3 mL). The resulting mixture was stirred for 3 h. The volatiles were removed on a rotary evaporator and by lyophilization. 29 was obtained as a white fluffy solid (1.62 mg, 37.0%). HRMS (ESI): m/z [M + 4H]4+ calcd for C78H124N26O16: 420.2417, found: 420.2434. RP-HPLC (220 nm): >99% (tR = 20.15 min, k = 6.0). C78H120N26O16·C8H4F12O8 (1677.98 + 456.08).

Octanedioyl-bis(Tyr-aza-Orn-Leu-Arg-Tyr-amide) Tetrakis(hydrotrifluoroacetate) (30)

Compound 23 (16.7 mg, 12.8 μmol) was dissolved in anhydrous DMF/DIPEA (99:1 v/v) (600 μL). Compound 25 (1.88 mg, 5.1 μmol) was added, and the mixture was stirred at 35 °C for 16 h. After addition of water (10 mL), the protected intermediate was extracted with CH2Cl2 (2 × 10 mL). The combined extracts were evaporated, and the residue was dried in vacuo. TFA/water (95:5 v/v) (2 mL) was added, and the mixture was stirred at rt for 2.5 h. Water (100 mL) was added followed by lyophilization. The product was purified by preparative HPLC (gradient: 0–20 min: MeCN/0.1% aq TFA 6:94–37:63, tR = 17.4 min). Lyophilization of the eluate afforded 30 as a white fluffy solid (3.9 mg, 37.3%). HRMS (ESI): m/z [M + 4H]4+ calcd for C76H120N22O16: 399.2308, found: 399.2317. RP-HPLC (220 nm): 98% (tR = 20.13 min, k = 6.0). C76H116N22O16·C8H4F12O8 (1593.90 + 456.08).

Octanedioyl-bis(Tyr-aza-Arg-Leu-Arg-Tyr-amide) Tetrakis(hydrotrifluoroacetate) (31)

Compound 30 (3.0 mg, 1.66 μmol) was dissolved in anhydrous DMF/DIPEA (99:1 v/v) (300 μL). A solution of N,N′-di-Boc-1H-pyrazole-1-carboxamidine (1 mg/10 μL, 13 μL, 2.5 equiv) was added, and the mixture was stirred at rt for 4 h. After addition of water (5 mL), the protected intermediate was purified by preparative HPLC (gradient: 0–25 min: MeCN/0.1% aq TFA 3:97–52:48, tR = 26.1 min). The solvents were removed by lyophilization, the residue was taken up in TFA/CH2Cl2/H2O (5:5:0.5, 3 mL), and the resulting mixture was stirred for 3 h. The volatiles were removed on a rotary evaporator and by lyophilization. The title compound was obtained as a white fluffy solid (1.50 mg, 42.3%). HRMS (ESI): m/z [M + 4H]4+ calcd for C78H124N26O16: 420.2417, found: 420.2437. RP-HPLC (220 nm): 96% (tR = 20.86 min, k = 6.3). C78H120N26O16·C8H4F12O8 (1677.98 + 456.08).

Octanoyl-Tyr-Arg-Leu-Arg-Tyr-amide Bis(hydrotrifluoroacetate) (33)

The pentapeptide 33 was synthesized according to the general procedure (82 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 22:78–51:49, tR = 13.2 min) afforded 33 as a white solid (33.4 mg, 59.5%). HRMS (ESI): m/z [M + 2H]2+ calcd for [C44H72N12O8]2+: 448.2793, found: 448.2800. RP-HPLC (220 nm): 99% (tR = 28.04 min, k = 8.8). C44H70N12O8·C4H2F6O4 (895.13 + 228.04).

Ac-Tyr-Nω-[(4-aminobutyl)aminocarbonyl]Arg-Leu-Arg-Tyr-amide Tris(hydrotrifluoroacetate) (34)

The pentapeptide 34 was synthesized according to the general procedure (18.8 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 14.4 min) gave 34 as a white solid (7.7 mg, 53.0%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C43H71N14O9]3+: 309.1837, found: 309.1850. RP-HPLC (220 nm): 97% (tR = 16.96 min, k = 4.9). C43H68N14O9·C6H3F9O6 (925.12 + 342.06).

Ac-Arg-Tyr-Nω-[(4-aminobutyl)aminocarbonyl]Arg-Leu-Arg-Tyr-amide Tetrakis(hydrotrifluoroacetate) (35)

The hexapeptide 35 was synthesized according to the general procedure (18.8 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.6 min) yielded 35 as a white solid (9.3 mg, 52.7%). HRMS (ESI): m/z [M + 4H]4+ calcd for [C49H84N18O10]4+: 271.1649, found: 217.1657. RP-HPLC (220 nm): 97% (tR = 15.22 min, k = 4.3). C49H80N18O10·C8H4F12O8 (1081.31 + 456.08).

Ac-Arg-Tyr-Arg-Leu-Arg-Tyr-amide Tris(hydrotrifluoroacetate) (36)

The hexapeptide 36 was synthesized according to the general procedure (35 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.7 min) afforded 36 as a white solid (18.3 mg, 65.5%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C44H70N16O9]3+: 323.1910, found: 323.1927. RP-HPLC (220 nm): 99% (tR = 15.98 min, k = 4.6). C44H70N16O9·C6H3F9O6 (967.15 + 342.06).

Ac-Arg-Tyr-Arg-Leu-Nω-[(4-aminobutyl)aminocarbonyl]Arg-Tyr-amide Tetrakis(hydrotrifluoroacetate) (37)

The hexapeptide 37 was synthesized according to the general procedure (35 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.6 min) gave 37 as a white solid (7.2 mg, 21.9%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C49H83N18O10]3+: 361.2174, found: 361.2187. RP-HPLC (220 nm): 99% (tR = 15.17 min, k = 4.3). C49H80N18O10·C8H4F12O8 (1081.31 + 456.08).

Ac-Arg-Tyr-Nω-[(4-aminobutyl)aminocarbonyl]Arg-Leu-Arg-d-Tyr-amide Tetrakis(hydrotrifluoroacetate) (38)

The hexapeptide 38 was synthesized according to the general procedure (35 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.6 min) afforded 38 as a white solid (5.7 mg, 17.4%). HRMS (ESI): m/z [M + 4H]4+ calcd for [C49H84N18O10]4+: 271.1649, found: 217.1666. RP-HPLC (220 nm): 98% (tR = 15.25 min, k = 4.3). C49H80N18O10·C8H4F12O8 (1081.31 + 456.08).

Ac-Arg-Tyr-Nω-[(4-aminobutyl)aminocarbonyl]Arg-Leu-d-Arg-Tyr-amide Tetrakis(hydrotrifluoroacetate) (39)

The hexapeptide 39 was synthesized according to the general procedure (35 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.3 min) yielded 39 as a white solid (6.0 mg, 18.3%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C49H83N18O10]3+: 361.2174, found: 361.2190. RP-HPLC (220 nm): 95% (tR = 14.77 min, k = 4.1). C49H80N18O10·C8H4F12O8 (1081.31 + 456.08).

Ac-Arg-Tyr-Nω-[(4-aminobutyl)aminocarbonyl]Arg-d-Leu-Arg-Tyr-amide Tetrakis(hydrotrifluoroacetate) (40)

The hexapeptide 40 was synthesized according to the general procedure (35 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.4 min) afforded 40 as a white solid (6.6 mg, 20.1%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C49H83N18O10]3+: 361.2174, found: 361.2189. RP-HPLC (220 nm): 96% (tR = 14.33 min, k = 4.0). C49H80N18O10·C8H4F12O8 (1081.31 + 456.08).

Ac-Arg-d-Tyr-{Nω-[N-(4-aminobutyl)aminocarbonyl]}Arg-Leu-Arg-Tyr-amide Tetrakis(hydrotrifluoroacetate) (41)

The hexapeptide 41 was synthesized according to the general procedure (35 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.5 min) gave 41 as a white solid (6.7 mg, 20.5%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C49H83N18O10]3+: 361.2174, found: 361.2190. RP-HPLC (220 nm): 95% (tR = 15.1 min, k = 4.3). C49H80N18O10·C8H4F12O8 (1081.31 + 456.08).

Ac-d-Arg-Tyr-{Nω-[N-(4-aminobutyl)aminocarbonyl]}Arg-Leu-Arg-Tyr-amide Tetrakis(hydrotrifluoroacetate) (42)

The hexapeptide 42 was synthesized according to the general procedure (35 mg Fmoc-Sieber-PS resin (loading: 0.61 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 13.5 min) afforded 42 as a white solid (7.0 mg, 21.4%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C49H83N18O10]3+: 361.2174, found: 361.2190. RP-HPLC (220 nm): 99% (tR = 14.94 min, k = 4.2). C49H80N18O10·C8H4F12O8 (1081.31 + 456.08).

Ac-Arg-Tyr-{Nω-[N-(4-aminobutyl)aminocarbonyl]}Arg-Leu-Arg-Tyr-OH Tetrakis(hydrotrifluoroacetate) (43)

The hexapeptide 43 was synthesized according to the general procedure (44.1 mg, H-l-Tyr(tBu)-2CT resin (loading: 0.68 mmol/g)). Purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 13:87–42:58, tR = 10.7 min) yielded 43 as a white solid (15.2 mg, 32.9%). HRMS (ESI): m/z [M + 4H]4+ calcd for [C49H83N17O11]4+: 271.4109, found: 271.4124. RP-HPLC (220 nm): 95% (tR = 16.28 min, k = 4.7). C49H79N17O11·C8H4F12O8 (1082.28 + 456.08).

(S)-2-Acetamido-N-((S)-1-{[(12S,15S,18S,21S)-2-amino-15-(3-guanidinopropyl)-12-(4-hydroxybenzyl)-18-isobutyl-4,11,14,17,20-pentaoxo-1,3,5,10,13,16,19-heptaazacyclotetracos-2-en-21-yl]amino}-3-(4-hydroxyphenyl)-1-oxopropan-2-yl)-5-guanidinopentanamide Tris(hydrotrifluoroacetate) (44)

Compound 43 (9.0 mg, 5.85 μmol), HOBt (1.79 mg, 2 equiv), and DIPEA (0.9 μL, 8 equiv) were dissolved in anhydrous DMF (5850 μL) to give a 1 mM solution. A solution of PyBOP (3.65 mg, 1.2 equiv) was added dropwise. After stirring at rt for 24 h, 0.1% aq TFA (20 mL) was added, and purification by preparative HPLC (gradient: 0–18 min MeCN/0.1% aq TFA 3:97–42:58, tR = 15.9 min) afforded 44 as a white solid (3.9 mg, 47.4%). HRMS (ESI): m/z [M + 3H]3+ calcd for [C49H80N17O10]3+: 355.5419, found: 355.5434. RP-HPLC (220 nm): 95% (tR = 18.4 min, k = 5.4). C49H77N17O10·C6H3F9O6 (1064.27 + 342.06).

Cells

HEC-1B (HTB-113) human endometrial cancer cells were from the American Type Culture Collection (Rockville, MD). MCF-7-Y1 cells were previously established in our laboratory.[23] HEL cells were kindly provided by Dr. M. C. Michel (Universitätsklinikum Essen, Germany), and CHO-K1 (ACC-110) cells and HEK293T (ACC 635) cells were from Deutsche Sammlung für Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany).

Routinely performed examinations for mycoplasma contamination using the Venor GeM Mycoplasma Detection Kit (Minerva Biolabs, Berlin, Germany) were negative for all cell types.

HEK293T-ARRB1-hY4R and HEK293T-ARRB2-hY4R Cells

The β-arrestin recruitment was quantified by a luciferase complementation assay based on the emerald luciferase (ELuc) from the brazilian click-beetle Pyrearinus termitilluminans. The fusion construct of the human hY4R (the hY4R cDNA was from cDNA Resource Center, Bloomsburg, PA) and the C-terminal luciferase fragment (hY4R-ELucC) was generated using the previously described construct SSTR2-ELucC[31,32] by replacing the cDNA of SSTR2 with the cDNA of the hY4R. HEK293T cells were stably transfected with the pcDNA3.1/myc-HIS (B) vector encoding the β-arrestin isoform 1 or 2, N-terminally fused with the N-terminus of the luciferase (ELucN-ARRB1 or ELucN-ARRB2, respectively)[31] and the pcDNA4/V5-HIS (B) vector encoding hY4R-ELucC using G418 (600 μg/mL) and zeocin (40 μg/mL) as selection antibiotics, as previously described for the H1R.[33] To determine the hY4R-ELucC expression, saturation binding studies were performed with both HEK293T-ARRB1-hY4R cells and HEK293T-ARRB2-hY4R cells using the radioligand 45 (cf. SI Figures S6 and S7).

Cell Culture

Cells were maintained in 25 or 75 cm2 flasks (Sarstedt, Nümbrecht, Germany) in a humidified atmosphere (95% air, 5% CO2) at 37 °C. MCF-7-Y1 cells,[22] HEL cells,[34] CHO-hY2-Gqi5-mtAEQ cells,[26] CHO-hY4-Gqi5-mtAEQ cells,[26] and HEK293T-hY4R-CRE Luc cells[11] were cultured as described previously. HEK293T-ARRB1-hY4R and HEK293T-ARRB2-hY4R cells were maintained in Dulbecco’s modified Eagle’s medium containing 10% FCS, 600 μg/mL G418, and 40 μg/mL zeocin.

Radioligand Binding Assays

Radioligand binding assays at all NPY receptor subtypes were performed at 22 ± 1 °C according to the experimental protocols described in detail previously.[11] Radioactivity (dpm) was measured with a MicroBeta2 plate counter (PerkinElmer, Rodgau, Germany). Cells, buffers, radioligands, and the respective Kd values and concentrations, as well as the assay conditions are summarized in Table .

Table 3

Conditions of the Radioligand Competitions Binding Assays at the NPY Receptor Subtypes

	Y1R	Y2R	Y4R	Y5R
receptor source	MCF-7-Y1 cells[23]	CHO-hY2R-Gqi5-mtAEQ cells[25]	CHO-hY4R-Gqi5-mtAEQ cells[26]	HEC-1B-hY5 cells[27]
buffer	N-(2-hydroxyethyl)piperazine-N′-ethanesulfonic acid (HEPES) buffer containing 150 mM NaCla	sodium-free HEPES bufferb	sodium-free HEPES bufferb	HEPES buffer containing 150 mM NaCla
radioligand	compd 46	[3H]propionyl-pNPY	compd 45(11)	[3H]propionyl-pNPY
Kd	6.2 nM	1.4 nM	0.67 nM	4.8 nM
concentration	4 nM	1.0 nM	0.6 nM	4 nM
incubation period	90 min	90 min	90 min	120 min
separation	suction	suction	filtration	suction
scintillation cocktail	Optiphase Supermix	Optiphase Supermix	Rotiscint eco plus	Optiphase Supermix

Isotonic HEPES buffer pH 7.4 (150 mM NaCl, 10 mM HEPES, 25 mM NaHCO3, 2.5 mM CaCl2, 1.2 mM KH2PO4, 1.2 mM MgSO4, 5 mM KCl).

Hypotonic, sodium-free HEPES buffer pH 7.4 (25 mM HEPES, 2.5 mM CaCl2, 1 mM MgCl2). Binding buffers were supplemented with BSA (1%) and bacitracin (0.1 mg/mL).

Fura-2 Calcium Assay

The assay was performed with HEL cells as previously described.[34,35]

Aequorin Calcium Assay

The assay was performed on CHO-hY4-Gqi5-mtAEQ cells as previously described[26] using a GENios Pro plate reader (Tecan, Salzburg, Austria). Areas under the curve were calculated using SigmaPlot 12.5 software (Systat Software Inc., Chicago, IL).

Luciferase Assay

The Luciferase assay was performed on HEK293-hY4-CRE Luc cells as previously described.[11]

β-Arrestin Recruitment Assay

The recruitment of β-arrestin was measured via the split-luciferase complementation technique.[32] Agonist potencies were determined on HEK293T-ARRB1-Y4R and HEK293T-ARRB2-Y4R cells using the GENios Pro microplate reader (Tecan, Salzburg, Austria) as previously described for HEK293T-ARRB1-H1R and HEK293T-ARRB2-H2R, respectively.[33] For the determination of antagonism, the cells were preincubated in the presence of the antagonist for 15 min. Ten microliters of an hPP solution (30 nM, final concentration 3 nM) was added and incubation was continued at 25 °C for 60 min. The plates were further processed as in the case of the agonist mode.

Data Analysis

Concentration response curves from functional assays and displacement curves from radioligand competition binding were analyzed by four-parameter sigmoidal fits (GraphPad Prism 5.0, San Diego, CA). Agonist potencies are given as EC50 values, intrinsic activities are expressed as α values with respect to the effect of 1 μM hPP (maximal response α = 1.0). Antagonistic activities were determined in the presence of 3 nM hPP (β-arrestin 1/2 assay, EC50 = 3.54 nM (ARRB1), EC50 = 2.74 nM (ARRB2)), 100 nM hPP (aequorin assay, EC50 = 9.7 nM) or 10 nM pNPY (Fura-2 Ca2+ assay, EC50 = 0.87 nM (Y1R)). Ki and Kb (Fura-2 Ca2+ (Y1R)) values were calculated from IC50 values using the Cheng–Prusoff equation.[36]