Novel sst₂-selective somatostatin agonists. Three-dimensional consensus structure by NMR

Abstract

The three-dimensional NMR structures of six octapeptide agonist analogues of somatostatin (SRIF) in the free form are described. These analogues, with the basic sequence H-DPhe/Phe²-c[Cys³-Xxx⁷-DTrp⁸-Lys⁹-Thr¹⁰-Cys¹⁴]-Thr-NH₂ (the numbering refers to the position in native SRIF), with Xxx⁷ being Ala/Aph, exhibit potent and highly selective binding to human SRIF type 2 (sst₂) receptors. The backbone of these sst₂-selective analogues have the usual type-II’ β-turn reported in the literature for sst_2/3/5-subtype-selective analogues. Correlating biological results and NMR studies led to the identification of the side chains of DPhe², DTrp⁸ and Lys⁹ as the necessary components of the sst₂ pharmacophore. This is the first study to show that the aromatic ring at position 7 (Phe⁷) is not critical for sst₂ binding and that it plays an important role in sst₃ and sst₅ binding. This pharmacophore is therefore different from that proposed by others for sst_2/3/5 analogues.

Introduction

Somatostatin (SRIF, H-Ala¹-Gly²-c[Cys³-Lys⁴-Asn⁵-Phe⁶-Phe⁷-Trp⁸-Lys⁹-Thr¹⁰-Phe¹¹-Thr¹²-Ser¹³-Cys¹⁴]-OH), a cyclic tetradecapeptide, isolated from the hypothalamus as a growth hormone inhibitor, is now known to be a multifunctional peptide located in most of the brain regions and in peripheral organs.^1,2 Cells containing SRIF are typically neurons or endocrine-like cells which are found in high density throughout the central and peripheral nervous systems, in the endocrine pancreas, and in the gut and in small numbers in the thyroid, adrenals, submandibular glands, kidneys, prostrate and placenta.^1,2 The activities of SRIF are mediated through a family of five different high affinity membrane receptors, sst₁-sst₅ (sst_s). Due to its broad spectrum of physiological activities and its short duration of action due to rapid proteolytic degradation in vivo,³ SRIF continues to be a target for the development of subtype-specific analogues.^{4-7 (and references therein)}

Indeed, over the last three decades, hundreds of SRIF analogues have been reported and tested in different biological systems including affinity and selectivity towards the five receptor-subtypes.^8-10 Correspondingly, early structure-activity relationship (SAR) studies ruled out the specific involvement of the side chains of all residues but Phe⁷-DTrp⁸-Lys⁹ for biological recognition.^11-13 (Note: the numbering of residues follows that of native SRIF.) Extensive structural studies including NMR and x-ray diffraction¹⁴ were carried out to elucidate the pharmacophore and the consensus structural motif of analogues binding predominantly to sst_2/5 (and sst₃) receptors. On the basis of these 3D structures of the free peptides, a pharmacophore model was proposed for SRIF analogues binding to sst_2/5 (and sst₃).^15-18 In this model, the side chains and the relative spatial arrangement of DPhe², Phe⁷, DTrp⁸ and Lys⁹ constitute the most essential elements necessary for binding (Figure 3C). The side chain of DTrp⁸ is in close proximity to the side chain of Lys⁹ (∼5 Å), whereas the side chain of Phe⁷ is about 7 − 9 Å away from the side chain of DTrp⁸ and 9 − 11 Å from the side chain of Lys⁹. The aromatic side chain at position 2 outside the disulfide bond was not conserved in its position in most of these analogues. The distance between DPhe² and DTrp⁸, Lys⁹, Phe⁷ were 11−15 Å, 12−15 Å, 5−11 Å respectively.¹⁹ In this model of the pharmacophore, the rotamer of the side chains of DTrp⁸ and Phe⁷ are in the trans configuration and that of Lys⁹ is in the gauche configuration.

Figure 3.

Consensus structural motif of sst₂-selective SRIF analogues. (A) Stereo view of the consensus structural motif for the sst₂-selective analogues 1 (red), 2 (green), 3 (navy blue) and 4 (cyan). Only the side chains of the residues DPhe², DTrp⁸ and Lys⁹ for 1, 2, 3 and 4 are shown. The distances between the Cγ atoms of DTrp⁸, Lys⁹ and DPhe² are displayed. For each analogue, the conformer with the lowest target function is displayed. (B) Schematic drawing of the pharmacophore for the sst₂-selective SRIF analogues. (C) Schematic drawing of the pharmacophore for the sst_2/3/5-selective SRIF analogues.¹⁶ (D) Schematic drawing of the pharmacophore for the sst₁-selective SRIF analogues.⁶² (E) Schematic drawing of the pharmacophore for the sst₄-selective SRIF analogues.²⁰ In all of the pharmacophores, the distance between the Cγ atoms of the respective residues is displayed.

Recently, we have proposed models of the sst₄ and the sst₁ pharmacophores derived from the NMR consensus structures of several receptor selective analogues.^20,21 These models of the pharmacophores are different from that of the reported sst_2/3/5 selective pharmacophore in the following manner. The sst₄ pharmacophore has one aromatic side chain, either Phe⁶ or Phe¹¹ involved in binding in addition to the DTrp⁸ and Lys⁹ pair. The distance between DTrp⁸ and Lys⁹ is 4.5−6.5 Å, between DTrp⁸ and Phe^6/11 is 5.5−9.5 Å and between Lys⁹ and Phe^6/11 is 4.5−6.5 Å. The aromatic side chain of Phe^6/11 is closer to the side chain of Lys⁹ than that of DTrp⁸ (Figure 3E). On the other hand, the sst₁ pharmacophore has two aromatic side chains, at position 6 or 7 and 11, involved in binding in addition to the DTrp⁸ and IAmp⁹ pair. The positioning of the two aromatic side chains at the back of the peptide plane has to be noted, which is opposite to that of the sst₄ and sst_2/3/5 pharmacophores. The sst₁ receptor selectivity is mainly achieved by the amino acid IAmp, which replaces Lys at position 9. The distance between DTrp⁸ and IAmp⁹ is 7−8 Å, between DTrp⁸ and Phe^6/7 is 6−7.5 Å and between IAmp⁹ and Phe¹¹ is 8−10 Å (Figure 3D).

In a similar manner, we present a model of the sst₂ pharmacophore based on the 3D NMR structures of the free peptides. The three-dimensional structures of six analogues in DMSO-d₆ are presented including the structure of octreotide amide. These analogues have the typical octreotide scaffold with a type-II’ β-turn backbone conformation similar to that of the reported sst_2/3/5 selective analogues, yet their pharmacophore has one aromatic side chain (Phe⁷) less than that of the reported sst_2/3/5 selective analogues.

Results

Peptide synthesis and characterization are described in Table 1 and experimental section. The proton resonance assignment and the structure determination by NMR for each of the sst₂-selective analogues 1 − 6 (Table 1) are presented in Tables 2 and 3.

Table I.

Physico-chemical properties and sst₁₋₅ binding affinities (IC₅₀s, nM) of analogues studied by NMR.

ID #	Compound	Purity (%)		MSc		IC50 nMd
		HPLCa	CZEb	M(mono) calc.	MH+(mono) obs.	sst1	sst2	sst3	sst4	sst5
1	H-DPhe-c[Cys-Ala-DTrp-Lys-Thr-Cys]-Thr-NH2	99	99	955.40	956.38	>1K (3)	8.6;4.5;3.6	772;∼1472;543	>1K (3)	158;807;226
2	H2N-CO-DPhe-c[Cys-Aph(CONH2)-DTrp-Lys-Thr-Cys]-Thr-NH2	98	97	1132.45	1133.17	>1K (3)	6.4;9;6.2	210;725;394	>1K;905;494	400;77;100
3	H2N-CO-Phe-c[Cys-Aph(CONH2)-DTrp-Lys-Thr-Cys]-Thr-NH2	98	99	1132.46	1133.44	>1K (2)	7.5;20	942;1094	957;872	109;260
4	H2N-CO-DPhe-c[Cys-Aph(CONHOCH3)-DTrp-Lys-Thr-Cys]-Thr-NH2	99	99	1119.46	1120.42	>1K (2)	3.1;4.4	>1K;958	>1K (2)	808;539
5	H-c[Cys-Phe-DTrp-Lys-Thr-Cys]-OH	97	98	78430	785.27	>1K	378	>1K	>1K	>1K
6	H-DPhe-c[Cys-Phe-DTrp-Lys-Thr-Cys]-Thr-NH2	95	99	1031.43	1032.00	>1K	2.7;2.2	80;38	>10K	2.7;3.9

^aPercent purity determined by HPLC (Hewlett-Packard Series II 1090 Liquid Chromatograph) using buffer system: A = TEAP (pH 2.5) and B = 60% CH₃CN/40% A with a gradient slope of 1% B/min, at flow rate of 0.2 mL/min on a Vydac C₁₈ column (0.21 × 15 cm, 5-μm particle size, 300 Å pore size). Detection at 214 nm.

^bCapillary zone electrophoresis (CZE) was done using a Beckman P/ACE System 2050 controlled by an IBM Personal System/2 Model 50Z and using a ChromJet integrator. Field strength of 15 kV at 30 °C, mobile phase: 100 mM sodium phosphate (85:15, H₂O:CH₃CN) pH 2.50, on a Supelco P175 capillary (363 μm OD x 75 μm ID X 50 cm length). Detection at 214 nm.

^cMass spectra (MALDI-MS) were measured on an ABI-Perseptive DE-STR instrument. The instrument employs a nitrogen laser (337 nm) at a repetition rate of 20 Hz. The applied accelerating voltage was 20 kV. Spectra were recorded in delayed extraction mode (300 ns delay). All spectra were recorded in the positive reflector mode. Spectra were sums of 100 laser shots. Matrix α-cyano-4-hydroxycinnamic acid was prepared as saturated solutions in 0.3% trifluoroacetic acid and 50% acetonitrile. The observed monoisotopic (M + H)⁺ values of each peptide corresponded with the calculated (M + H)⁺ values.

^dThe IC₅₀ values (nM) were derived from competitive radioligand displacement assays reflect the affinities of the analogues for the cloned somatostatin receptors using the non-selective ¹²⁵I-[Leu⁸,DTrp²²,Tyr²⁵]SRIF-28, as the radioligand.

Table 2.

Proton chemical shifts of analogues 1−6^*

Residue	1H	Analogues
		1	2	3	4	5	6
H2N-CO			5.42	5.66	7.59
dPhe/Phe2	NH	7.99	6.33	6.42	8.01
	αH	4.20	4.55	4.53	4.19		4.15
	βH	3.22,2.96	3.08,2.72	3.00,2.80	3.25,2.95		3.24,2.95
	H2,6	7.38	7.28	7.14	7.38		7.38
	H3,5	7.29	7.23	7.21	7.30		7.29
Cys3	NH	9.22	8.72	8.59	9.23		9.22
	αH	5.27	5.22	5.16	5.27	4.14	5.28
	βH	2.76,2.82	2.77,2.77	2.81,2.81	2.82,2.82	3.07,2.98	2.81,2.81
		Ala	Aph(CONH2)	Aph(CONH2)	Aph(CONHOCH3)	Phe	Phe
Aph/Ala7	NH	8.44	8.49	8.45	8.52	8.36	8.55
	αH	4.47	4.58	4.60	4.65	4.67	4.66
	βCH3	1.17
	βH		2.77,2.77	2.81,2.81	2.79,2.79	2.81,2.81	2.84,2.84
	H2,6		6.96	6.99	7.03	7.03	7.08
	H3,5		7.22	7.28	7.44	7.16	7.17
	HT		8.46	8.54	8.84
	NH2		5.86	5.92	9.55
	CH3				3.62
dTrp8	NH	8.92	8.82	8.80	8.81	8.74	8.77
	αH	4.30	4.20	4.19	4.19	4.58	4.19
	βH	3.07,2.99	2.98,2.73	2.98,2.72	2.98,2.73	3.14,2.86	2.96,2.71
	HD1	7.19	6.93	6.91	6.97	6.99	6.95
	HE1	10.87	10.76	10.77	10.74	10.80	10.81
	HZ2	7.32	7.29	7.30	7.29	7.32	7.32
	HH2	7.06	7.04	7.06	7.04	7.08	7.06
	HZ3	6.99	6.98	7.00	6.99	7.00	6.99
	HE3	7.54	7.42	7.41	7.41	7.47	7.43
Lys9	NH	8.43	8.36	8.39	8.40	8.50	8.46
	αH	3.99	4.01	4.00	3.97	3.89	3.99
	βH	1.70,1.25	1.72,1.27	1.73,1.29	1.70,1.26	1.69,1.37	1.72,1.30
	γH	0.72,0.72	0.77,0.74	0.80,0.76	0.76,0.72	0.86,0.86	0.79,0.79
	δH	1.29,1.29	1.30,1.30	1.32,1.32	1.32,1.32	1.35,1.35	1.35,1.35
	εH	2.55,2.55	2.56,2.56	2.58,2.58	2.55,2.55	2.60,2.60	2.57,2.57
	εNH	7.60	7.56	7.62	7.58	7.63	7.62

Residue	1H	Analogues
		1	2	3	4	5	6
Thr10	NH	7.68	7.58	7.58	7.63	7.72	7.62
	αH	4.48	4.49	4.51	4.48	4.29	4.50
	βH	3.99	3.97	3.97	4.02	4.18	4.00
	γH	1.05	1.06	1.06	1.08	1.09	1.07
	OH	4.88	4.80	4.83	4.87
Cys14	NH	8.42	8.46	8.47	8.49	8.08	8.49
	αH	5.15	5.30	5.21	5.15	4.58	5.17
	βH	2.86,2.86	2.85,2.85	2.84,2.84	2.86,2.86	3.14,2.86	2.86,2.86
Thr15	NH	8.06	8.17	8.13	8.05		8.07
	αH	4.22	4.21	4.23	4.23		4.21
	βH	4.02	4.06	3.93	4.03		4.03
	γH	1.05	1.07	0.99	1.07		1.06
	OH	5.15	5.13	5.07	5.15
NH2		7.53,7.36	7.42,7.26	7.43,7.35	7.53,7.36		7.35,7.52

^*The chemical shifts were measured at 298K in DMSO in ppm with the internal reference of the DMSO signal at 2.49 ppm.

Table 3.

Characterization of the NMR structures of analogues 1−6^*

ID#	NOE distance restraints	Angle restraints***	CYANA Target function**	Backbone RMSD (Å)	Overall RMSD (Å)	CFF91 energies (Kcal/mol)			Residual restraint violations on
						Total energy	Van der Waals	Electrostatic	Distances		Dihedral Angles
									No. ≥ 0.1 Å	Max (Å)	No. ≥ 1.5 deg	Max (deg)
1	131	24	0.005	0.52 ± 0.09	1.03 ± 0.20	178 ± 7	89 ± 4	89 ± 7	0.7 ± 0.1	0.11 ± 0.01	0 ± 0	0 ± 0
2	135	17	0.001	0.44 ± 0.24	0.96 ± 0.30	211 ± 10	118 ± 5	93 ± 8	0.2 ± 0.1	0.05 ± 0.01	0 ± 0	0 ± 0
3	130	25	0.003	0.47 ± 0.28	1.07 ± 0.45	204 ± 14	124 ± 11	81 ± 7	0.3 ± 0.1	0.07 ± 0.03	0 ± 0	0 ± 0
4	130	19	0.008	0.73 ± 0.25	1.21 ± 0.31	188 ± 11	116 ± 7	72 ± 6	0.5 ± 0.1	0.10 ± 0.02	0 ± 0	0 ± 0
5	80	14	0.0002	0.08 ± 0.09	0.62 ± 0.19	158 ± 10	86 ± 7	73 ± 13	0.1 ± 0.0	0.03 ± 0.00	0 ± 0	0 ± 0
6	115	24	0.002	0.56 ± 0.14	1.06 ± 0.28	210 ± 14	121 ± 10	89 ± 8	0.4 ± 0.1	0.11 ± 0.02	0 ± 0	0 ± 0

^*The bundle of 20 conformers with the lowest residual target function was used to represent the NMR structures of each analogue.

^**The target function is zero only if all the experimental distance and torsion angle constraints are fulfilled and all non-bonded atom pairs satisfy a check for the absence of steric overlap. The target function is proportional to the sum of the square of the difference between calculated distance and isolated constraint or van der Waals restraints and similarly isolated angular restraints are included in the target function. For the exact definition see reference.³²

^***Meaningful NOE distance restraints may include intra-residual and sequential NOEs.³²

Assignment of proton resonances, collection of structural restraints, and structure determination

The nearly complete chemical shift assignment of proton resonances (Table 2) for analogues 1 − 6 (Table 1) has been carried out using two-dimensional (2D) NMR experiments applying the standard procedure described in the Experimental Section. The N-terminal amino protons for analogues 5 and 6 were not observed due to fast exchange with the solvent. The amide proton of DPhe²/Phe for analogues 2 and 3 were observed at high field due to the carbamoylation of the N-terminus. Chemical shifts can provide insight into the structure of the peptides, which is particularly true for SRIF analogues. The ring current of the indole ring leads to a distinct shift of the C^γ protons of the sequential Lys side chain to lower frequencies when these groups are closer in space, as in a β-turn. The downfield shift of C^γ protons has been observed in all of these sst₂-selective analogues, which is similar to the observations found in the non-selective analogues binding to sst_2/3/5.^17,22-30 This indicates the presence of a β-turn around these residues in these analogues.

A large number of experimental NOEs is observed for all of the six analogues in the NOESY spectrum measured with a mixing time of 100 ms, leading to over 100 to 120 meaningful distance restraints per analogue and concomitantly ∼15 restraints per residue, which is a typical number for folded proteins (Table 3). These structural restraints are used as input for the structure calculation with the program CYANA³¹ followed by restrained energy minimization using the program DISCOVER.³² The resulting bundle of 20 conformers per analogue represents the 3D structure of each analogue in solution. For each analogue, the small residual constraint violations in the distances for the 20 refined conformers (Table 3) and the coincidence of the experimental NOEs and short interatomic distances (data not shown) indicate that the input data represent a self-consistent set, and that the restraints are well satisfied in the calculated conformers (Table 3). The deviations from ideal geometry are minimal, and similar energy values are obtained for all of the 20 conformers for each analogue. The quality of the structures determined is reflected by the small backbone RMSD values relative to the mean coordinates of ∼0.5 Å (see Table 3 and Figure 2).

Figure 2.

NMR structures of analogues 1−6. For each analogue, twenty energy-minimized conformers with the lowest target function are used to represent the 3D NMR structure. The bundle is obtained by overlapping the Cα atoms of the residues 2−9. The backbone and the side chains are displayed including the disulfide bridge. The following color code is used: red (1) H-DPhe-c[Cys-Ala-DTrp-Lys-Thr-Cys]-Thr-NH₂, dark green (2) H₂NCO-DPhe-c[Cys-Aph(CONH₂)-DTrp-Lys-Thr-Cys]-Thr-NH₂, navy blue (3) H₂NCO-Phe-c[Cys-Aph(CONH₂)-DTrp-Lys-Thr-Cys]-Thr-NH₂, cyan (4) H₂NCO-DPhe-c[Cys-Aph(CONHOCH₃)-DTrp-Lys-Thr-Cys]-Thr-NH₂, yellow (5) H-c[Cys-Phe-DTrp-Lys-Thr-Cys]-OH, magenta (6) H-DPhe-c[Cys-Phe-DTrp-Lys-Thr-Cys]-Thr-NH₂. The side chains that are involved in sst₂-binding are highlighted; these are DTrp at position 8 in light green, Lys at position 9 in blue and DPhe or Phe at position 2 in yellow.

Three-dimensional structure of H-DPhe²- c[Cys³-Ala⁷-DTrp⁸-Lys⁹-Thr¹⁰-Cys¹⁴]-Thr-NH₂(1)

Analogue 1 binds to receptor 2 with high affinity (IC₅₀ < 10 nM) and has an Ala at position 7. The quality of the structure is reflected by the small RMSD (Table 3), which can also be visually depicted from Figure 2 showing the bundle of 20 conformers representing the 3D structure. From the backbone torsion angles (Table 4), it can be seen that the backbone contains a β-turn of type-II’ around DTrp⁸-Lys⁹. The β-turn is supported by the strong sequential d_αN(i,i+1) NOEs and medium-range d_αN(i,i+2) NOE (Figure 1) between DTrp⁸ and Thr¹⁰, as well as the hydrogen bond between Thr¹⁰NH-O’Ala⁷ in all the 20 structures. The low temperature coefficient observed for the amide proton of Thr¹⁰ (−1.1 ppb/K) confirms the presence of the hydrogen bond.³³ From the torsion angles listed in Table 4, it can be seen that the side chain of DTrp⁸ is in the trans rotamer, Lys⁹ is in the gauche⁺ rotamer and that of DPhe² is in the gauche⁻ rotamer.

Table 4.

Torsion angles ϕ, ψ and χ₁ (in degrees) of the bundle of 20 energy minimized conformers.

Analogue	Angle	DPhe2	Cys3	Ala7	DTrp8	Lys9	Thr10	Cys14	Thr15
1	ϕ	-	−33 ± 60	−156 ± 13	72 ± 3	−91 ± 0	−140 ± 4	−156 ± 16	−11 ± 46
	ψ	67 ± 17	87 ± 6	132 ± 3	−146 ± 2	26 ± 1	−149 ± 23	87 ± 8	121 ± 56
	χ1	108 ± 6	132 ± 3	170 ± 95	−151 ± 3	−128 ± 2	−22 ± 86	176 ± 3	173 ± 54
2				Aph(CONH2)7
	ϕ	121 ± 45	−84 ± 35	−178 ± 19	97 ± 14	−103 ± 7	−74 ± 5	87 ± 2	−19 ± 37
	ψ	172 ± 9	117 ± 16	88 ± 9	−113 ± 14	2 ± 6	−137 ± 8	81 ± 8	110 ± 76
	χ1	115 ± 32	−121 ± 40	−175 ± 14	150 ± 6	−83 ± 7	−43 ± 7	−143 ± 24	−163 ± 58
3		Phe2		Aph(CONH2)7
	ϕ	−149 ± 9	174 ± 38	179 ± 22	95 ± 22	−103 ± 12	−63 ± 5	83 ± 28	−109 ± 59
	ψ	−145 ± 51	110 ± 11	98 ± 20	−108 ± 7	−8 ± 1	−140 ± 5	78 ± 8	48 ± 87
	χ1	−132 ± 30	−144 ± 24	−126 ± 18	−171 ± 6	−105 ± 1	−39 ± 2	−110 ± 34	−129 ± 76
4		DPhe2		Aph(CONHOCH3)7
	ϕ	175 ± 6	43 ± 2	−111 ± 8	44 ± 5	−141 ± 4	−143 ± 10	167 ± 71	−83 ± 83
	ψ	41 ± 4	53 ± 5	59 ± 5	−104 ± 3	45 ± 8	165 ± 52	78 ± 14	105 ± 78
	χ1	98 ± 6	−76 ± 34	−177 ± 5	160 ± 5	−45 ± 8	−52 ± 7	−155 ± 13	118 ± 49
5				Phe7
	ϕ		−	−154 ± 8	156 ± 3	−93 ± 3	−93 ± 3	−54 ± 5
	ψ		168 ± 2	23 ± 6	−128 ± 3	12 ± 0	2 ± 4	−
	χ1		−28 ± 3	160 ± 38	166 ± 1	−116 ± 0	−62 ± 8	−156 ± 4
6				Phe7
	ϕ	−	−71 ± 28	−173 ± 11	102 ± 2	−93 ± 3	−104 ± 5	−168 ± 5	−89 ± 64
	ψ	−160 ± 0	−153 ± 66	98 ± 6	−121 ± 0	1 ± 3	45 ± 5	81 ± 3	98 ± 70
	χ1	32 ± 0	−117 ± 5	−150 ± 29	−168 ± 1	−126 ± 4	140 ± 8	−148 ± 12	175 ± 4

Figure 1.

Survey of characteristic NOEs used in CYANA for structure calculation for analogues 1-6. Thin, medium and thick bars represent weak (4.5 to 6 Å), medium (3 to 4.5 Å) and strong (< 3 Å) NOEs observed in the NOESY spectrum. The medium-range connectivities d_NN(i,i+2), d_αN(i,i+2), and d_βN(i,i+2) are shown by lines starting and ending at the positions of the residues related by the NOE. Residues designated with X and X’ correspond to the amino acid Aph(CONH₂) and Aph(CONHOCH₃), respectively.

Three-dimensional structure of H₂N-CO-DPhe²-c[Cys³-Aph(CONH₂)⁷-DTrp⁸-Lys⁹-Thr¹⁰-Cys¹⁴]-Thr-NH₂ (2)

Analogue 2 differs from analogue 1 at position 7 by the Aph(CONH₂) group as well as the N-terminal carbamoylation (Table 1). The backbone torsion angles (Table 4) indicate a β-turn of type-II’ conformation around DTrp⁸-Lys⁹, which is supported by the medium range d_αN(i,i+2) NOE between DTrp⁸-Thr¹⁰ (Figure 1). The turn structure is further stabilized by the observed hydrogen bond between Thr¹⁰NH-O’Aph⁷ in most of the structures. The presence of the hydrogen bond is further confirmed by the unshifted amide proton resonance of Thr¹⁰ when the temperature was raised from 298K to 313K. The long range medium NOE observed between DPhe²(NH) and Thr¹⁰(HN) as well as the strong αH NOE observed between the cysteins stabilize the structure. The side chain of DPhe² is in the gauche⁺ rotamer, Aph⁷ and DTrp⁸ are in the trans rotamer and that of Lys⁹ is in the gauche⁻ rotamer (Table 4).

Three-dimensional structure of H₂N-CO-Phe²-c[Cys³-Aph(CONH₂)⁷-DTrp⁸-Lys⁹-Thr¹⁰-Cys¹⁴]-Thr-NH₂(3)

Analogue 3 binds selectively with high affinity to the sst₂ receptor and is different from analogue 2, with a Phe at position 2. From the backbone torsion angles (Table 4), it can be seen that the backbone contains a β-turn of type-II’ around DTrp⁸-Lys⁹. The β-turn is supported by the medium-range d_αN(i,i+2) NOE and the weak d_NN(i,i+2) NOE (Figure 1) between DTrp⁸ and Thr¹⁰, as well as the hydrogen bond Thr¹⁰NH-O’Aph⁷ observed in most of the 20 structures. The low temperature coefficient observed for the amide proton of Thr¹⁰ (−0.4 ppb/K) confirms the presence of the hydrogen bond.³³ In addition long range medium NOEs are observed between the amide protons of Phe² and Thr¹⁰ and the cysteins, which stabilizes the structure. The side chains of Phe², Aph⁷ and Lys⁹ are in the gauche⁺ rotamer and that of DTrp⁸ is in the trans rotamer (Table 4).

Three-dimensional structure of H₂N-CO-DPhe²-c[Cys³-Aph(CONHOCH₃)⁷-DTrp⁸-Lys⁹-Thr¹⁰-Cys¹⁴]-Thr-NH₂(4)

Analogue 4 binds selectively with nanomolar affinity to the sst₂ receptor. It differs from analogues 2 and 3 by the longer chain of Aph(CONHOCH₃) at position 7, which inhibits the binding to receptors 3 and 5 completely (Table 1). The backbone torsion angles indicate a β-turn of type-II’ conformation around DTrp⁸-Lys⁹ (Table 4), which is supported by the medium-range d_αN(i,i+2) NOE and the weak d_NN(i,i+2) NOE observed between DTrp⁸ and Thr¹⁰ (Figure 1). The unshifted amide proton resonance of Thr¹⁰ at 7.63 ppm (from 298K to 313K) suggests the presence of a hydrogen bond involving this amide proton, which has been observed in all the 20 structures. The side chain of DPhe² is in the gauche⁺ rotamer, Aph⁷ and DTrp⁸ are in the trans rotamer and that of Lys⁹ is in the gauche⁻ rotamer (Table 4). The other long range NOEs, which stabilize the structure, are observed between the amide proton of DPhe² and αH proton of Cys¹⁴ and the αH protons of the cysteins.

Three-dimensional structure of H-c[Cys³-Phe⁷-DTrp⁸-Lys⁹-Thr¹⁰-Cys¹⁴]-OH (5)

Analogue 5 is the shortest analogue, a cyclic hexapeptide containing only the core residues. Analogue 5 was synthesized to identify the role of _DPhe² in sst₂ binding. Indeed, it did not bind to all the five receptors (Table 1). From the backbone torsion angles it can be seen that it has a type-II’ β-turn around DTrp⁸ and Lys⁹(Table 4) and the turn is supported by the presence of the medium range d_αN(i,i+2) NOE observed between DTrp⁸ and Thr¹⁰ (Figure 1) as well as the hydrogen bond observed between Thr¹⁰NH-O’Phe⁷ in all of the 20 structures. The low temperature coefficient observed for the amide proton of Thr¹⁰ (−1.4 ppb/K) confirms the presence of the hydrogen bond.³³ The side chain of Phe⁷ and DTrp⁸ are in the trans rotamer and that of Lys⁹ is in the gauche⁺ rotamer (Table 4).

Three-dimensional structure of H-DPhe²-c[Cys³-Phe⁷-DTrp⁸-Lys⁹-Thr¹⁰-Cys¹⁴]-Thr-NH₂ (6)

Analogue 6, which is very similar to octreotide (Thr-ol is substituted by Thr-NH₂), binds to the sst_2/3/5 receptors with moderately high affinity (Table 1). The 3D structure of this analogue was obtained in order to compare the structure of octreotide with the other analogues under identical conditions. Melacini et al. reported that the backbone of Sandostatin (octreotide) could be in a conformational equilibrium between β-turn and helical structures in solvents different from DMSO.^19,28 From the torsion angles, we observe a type-II’ β-turn conformation for the backbone, which is supported by the presence of the medium range d_αN(i,i+2) NOE observed between DTrp⁸ and Thr¹⁰ (Figure 1) as well as the hydrogen bond observed between Thr¹⁰NH-O’Phe⁷ in all of the 20 structures. The unshifted amide proton resonance of Thr¹⁰ at 7.58 ppm (from 298K to 313K) confirms that this amide proton is involved in a hydrogen bond. The side chain of Phe⁷ and DTrp⁸ are in the trans rotamer and that of Lys⁹ is in the gauche⁺ rotamer (Table 4).

Discussion

We have recently proposed models of the sst₁ and sst₄ pharmacophores based on the 3D NMR structures of several receptor selective analogues. The sst₁ and sst₄ receptors have maximum sequence similarity and hence were assumed to be part of one close family, different from the family of sst_2/3/5 receptors. Yet, the proposed sst₁ and sst₄ pharmacophores are completely distinct from each other. For example, the position and the number of aromatic rings involved in binding are different for the two ligands, suggesting that these hydrophobic residues interact with different regions of the sst₁ and sst₄ receptors, respectively (Figure 3D and 3E). On the other hand, most of the previously published analogues binding with high affinity to sst₂ receptor were also binding to sst₅ receptors with nM affinity and sometimes to sst₃ receptors as well. Melacini et al., proposed a pharmacophore model for these analogues binding non-selectively to all of the three receptors (Figure 3C).¹⁹ Here, we have identified the crucial residues involved in selective sst₂ binding, revealing subtle differences in the pharmacophores between the only sst₂-selective analogues and the less selective sst_2/3/5-selective analogues. In the sst_2/3/5-selective pharmacophore found in octreotide and homologues, we identified two aromatic side chains (at position 2 and 7) in addition to the DTrp⁸ and Lys⁹ residues. Removal of the aromatic amino acid at position 2 (analogue 5) resulted in loss of binding to all of the three receptors (compare with analogue 6). This clearly confirms that residue 2 is important for binding to sst_2/3/5 receptors (Table 1). On the other hand, replacement of Phe at position 7 by Ala (analogue 1) resulted in selective binding (nM) to receptor 2. Hence, it can be concluded that this aromatic ring is not crucial for sst₂ binding, and is crucial for sst₃ and sst₅ binding, as the binding to receptors 3 and 5 decreased (Table 1). Replacement of Phe at position 7 by an amino acid having a longer side chain, such as Aph has also decreased the binding to sst₃ and sst₅ receptors, retaining high binding affinity to sst₂ receptors.

Three-dimensional structures of the sst₂-selective analogues and the pharmacophore model

Most of the bioactive analogues of SRIF reported so far have a β-turn type either of type II’ or type VI for the backbone conformation.^{13,15,17,18,24,34-39} The structures of all of the presented four sst₂-selective analogues have also a β-turn of type-II’ (Figure 2, Table 4). In these analogues, the side chain of DTrp⁸ is in the trans conformer and the side chain of Lys⁹ is either in the gauche⁺ or gauche⁻ conformer, bringing the two side chains adjacent to each other in a close proximity. The side chain of Phe or DPhe at position 2 is either in the gauche⁺ or gauche⁻conformer, far from the DTrp-Lys pair and their position is conserved in all of the four analogues studied here (Figure 3A). Hence, we propose a pharmacaphore model for sst₂-selective analogues involving the three side chains, namely, the indole (DTrp) at position 8, the amino-alkyl (Lys) at position 9 and an aromatic ring at position 2 (Figure 3B). In addition, the right spatial arrangement of the indole ring, the lysine side chain and the aromatic ring of a phenylalanine is also important for sst₂-selective binding. The proposed distances between the Cγ of residue 8 and Cγ of Phe² is 12−13.5 Å; Cγ of residue 8 and Cγ of Lys ⁹ is 4−5 Å and Cγ of Phe² and Cγ of Lys⁹ is 12.5−15 Å. Conservative replacements of these residues do not change the binding affinities and receptor selectivity evidently.

Comparison of the sst₂-selective versus sst_2/3/5-selective pharmacophore

With the consensus structural motif for the sst₂-selective SRIF analogues (Figures 3A and B), comparison between the sst₂ and the pharmacophore for sst_2/3/5-selective analogues (Figure 3C) is possible. Interestingly, Goodman and coworkers illustrated that in SRIF analogues, which bind selectively to sst_{2 2/3/5}, the side chains of DTrp⁸, Lys⁹, Phe⁷ and Phe² constitute the most essential elements necessary for binding.^15,16 In their pharmacophore model, DTrp⁸ and Lys⁹ were at a close proximity of ∼5 Å, which is similar to the sst₂ pharmacophore. The aromatic ring at position 7, Phe⁷ is farther away from DTrp⁸ (7 to 9 Å) and Lys⁹ (9 to 11 Å), and is not present in the sst₂ pharmacophore. Hence Phe⁷ is not crucial for sst₂ receptor binding. The position of DPhe² in the sst₂ pharmacophore is highly conserved whereas it varies in the sst_2/3/5 pharmacophore. This difference can easily be seen from the superposition of the sst₂-selective analogue 1 and the sst_2/3/5-selective octreotide (Figure 4).¹⁶ They differ mostly in the position of the phenylalanine at position 2. Based on our studies reported here and the observation of Melacini et al.,¹⁹ octreotide undergoes a conformational exchange between β-turn and helical backbone conformation, it can be proposed that octreotide undergoes conformational change to fit into both the sst₂ and sst₅ pharmacophores. It should prefer the β-turn conformation, where the DPhe² is close to the sst₂ pharmacophore (Figure 4A) and must prefer the helical conformation to fit the sst₅ pharmacophore (Figure 4B), where DPhe² is further away from the sst₂ pharmacophore (this hypothesis is further supported by unpublished results of sst₅ selective analogues). This conformational flexibility is therefore necessary to explain the non-selective binding of the octreotide type of analogues to sst_2/3/5 receptors simultaneously.

Figure 4.

Comparison of the 3D structure of sst₂-selective (analogue 1) and sst_2/3/5-selective (octreotide and analogue 6) SRIF analogues. Stereo view of the superposition of the 3D structure of the sst₂-selective analogue 1 (red) with the 3D structure of the sst_2/3/5-selective octreotide (green)¹⁶ (A) in β-turn conformation and (B) in helical conformation. It must be noted that DPhe² labeled, in both of the analogues is important for selective binding, but differs in its spatial orientation relative to the DTrp⁸-Lys⁹ pair.

Comparison of the sst₂-selective versus sst₁- and sst₄-selective pharmacophores

Since three different pharmacophores are available for the subtype-selective analogues, namely, 1, 2 and 4, the differences among these subtype-selective pharmacophores can be discussed with respect to sst₂ pharmacophore (Figure 3B). The sst₁ pharmacophore has two aromatic side chains important for binding and they are closer to DTrp-IAmp pair and are present on the back side of the peptide, compared to the sst₂ pharmacophore, where one aromatic ring important for binding is far from DTrp-Lys pair. Also IAmp at position 9 crucial for sst₁ selectivity is further away from DTrp⁸, compared to the other pharmacophores, where Lys at position 9 is closer to DTrp⁸. The sst₄ pharmacophore is somewhat similar to sst₂ pharmacophore in the number of interacting residues with the receptors, namely DTrp⁸, Lys⁹ and one aromatic side chain either at position 6 or 11. But the position of the aromatic side chain is closer to the DTrp-Lys pair in the sst₄ pharmacophore, whereas it is farther away in the sst₂ pharmacophore. Hence, these subtype-selective pharmacophores explain why the sst₂-selective analogues did not bind to other receptors of SRIF, specifically to 1 and 4.

Conclusions

The 3D conformations of four cyclic SRIF octapeptide analogues with a hexapeptide core having high binding affinity and selectivity to the sst₂ receptor have been presented. These studies indicate that these analogues have a β-turn of type-II’ for the backbone conformation which orients the side chains of the essentially important residues, namely indole at position 8, amino alkyl group at position 9 and an aromatic ring outside the cycle, in their respective positions for effective receptor-ligand binding. Based on this, we have proposed the SRIF binding motif for the sst₂ receptor consisting of these three side chains, namely DTrp⁸, Lys⁹ and DPhe². This binding motif differs from the binding motif for sst₂/sst₃/sst₅-selective receptors by one less aromatic side chain at position 7 (Figure 3). Furthermore, the model proposed also explains the selective binding of the non-peptoid analogues of SRIF agonists.^40-43 The pharmacophore models proposed so far enable us to understand the binding of several other SRIF analogues to somatostatin receptors and will play an important role in designing highly selective peptides as well as non-peptide ligands of SRIF.

Experimental Section

Peptide synthesis, purification and characterization

Starting Materials

4-methylbenzhydrylamine resin (MBHA) with the capacity of 0.4 mmol/g and Boc-Cys(Mob)-CM resin with a capacity of 0.3−0.4 mmol/g were used. All Boc-Nα–protected amino acids were commercially available (Chem Impex, Wood Dale, IL, Reanal Finechemical Co., Budapest, Hungary), except Boc-Aph(Fmoc) which was synthesized in our laboratory. ⁴⁴

Synthesis and Purification

Peptides were synthesized by SPPS methodology following the Boc strategy and purified as we published earlier.⁴⁵ The -CO-NH₂ (Cbm) ureido group at the 4-amino function of Aph and at the N-terminus of the peptides was introduced simultaneously on the resin. The 4-amino function of Aph was freed with 20% piperidine in NMP and the N-terminus Boc was deprotected with 50% TFA in DCM after completion of the synthesis, then the carbamoylation was carried out with NaOCN (200 mg, 1.3 mmol) in NMP (8 mL) and glacial acetic acid (6 mL) per gram of initial resin. The ureido group -CO-NHOCH₃ (Cbm-OMe) at the 4-amino function of Aph (analogue 4) was also introduced on the resin. The Nα-Boc protected resin-bound hexapeptide Boc-Aph(Fmoc)-DTrp-Lys[Z(2Cl)]-Thr(Bzl)-Cys(Mob)-Thr(Bzl)-MBHA was treated with 20% piperidine in NMP to free the 4-amino function of Aph and CH₃O-NH-COOC₆H₄-NO₂ (10 fold excess) in the presence of DIPEA in NMP was added to the resin which was shaken at RT for 4 h as we described earlier.⁴⁶ After completion of the synthesis, N-terminus Boc was deprotected with 50% TFA in DCM and the carbamoylation was carried out as described above. The completed peptides were cleaved from the resin by HF containing the scavengers anisole (10% v/v) and methyl sulfide (5% v/v) for 60 min at 0 °C. The diethyl ether precipitated crude peptides were cyclized in 75% acetic acid (200 mL) by addition of iodine (10% solution in methanol) until the appearance of a stable orange color. Forty minutes later, ascorbic acid was added to quench the excess of iodine. The crude, lyophilized peptides were purified by preparative RP-HPLC.⁴⁷

Characterization

The purity of the final peptides was determined by analytical RP-HPLC, CZE analysis.⁴⁸ (See legend of Table 1 for details.) Each peptide showed a purity of >95% by these methods. The observed monoisotopic (M + H)⁺ values of each peptide corresponded with the calculated (M + H)⁺ values.

Sample preparation and NMR Experiments

Peptides were synthesized on solid-phase with Boc chemistry on a CS-Bio Peptide Synthesizer (Model CS536).^4-6 Anhydrous HF cleaved the products from the resin support with simultaneous side chain deprotection. The diethyl ether precipitated crude peptides were cyclized in 75% acetic acid by the addition of iodine (10% solution in methanol). All peptides were purified by preparative RP-HPLC. The purity of the final products was determined by analytical RP-HPLC and CZE. They were >95% pure by both methods. The products were also characterized by mass spectroscopy; the observed (M+H)⁺ values coincided with the calculated (M+H)⁺ values.

NMR samples were prepared by dissolving 2 mg of the analogue in 0.5 mL of DMSO-d₆. The ¹H NMR spectra were recorded on a Bruker 700 MHz spectrometer operating at proton frequency of 700 MHz. Chemical shifts were measured using DMSO (δ = 2.49 ppm) as an internal standard. The 1D spectra were also acquired at temperatures between 298 and 318 K and were utilized to measure the temperature coefficients of the amide resonances. All the 2D spectra were acquired at 298 K. Resonance assignments of the various proton resonances have been carried out using total correlation spectroscopy (TOCSY);^49,50 double-quantum filtered spectroscopy (DQF-COSY)⁵¹ and nuclear Overhauser enhancement spectroscopy (NOESY).^52-54 The TOCSY experiments employed the MLEV-17 spin-locking sequence suggested by Davis and Bax,⁴⁹ applied for a mixing time of 50 ms. The NOESY experiments were carried out with a mixing time of 100 ms. The TOCSY and NOESY spectra were acquired using 800 complex data points in the ω₁ dimension and 1024 complex data points in the ω₂ dimension with t_1max = 47 ms and a t_2max = 120 ms and were subsequently zero-filled to 1024 × 2048 before Fourier transformation. The DQF-COSY spectra were acquired with 1024 × 4096 data points and were zero-filled to 2048 × 4096 before Fourier transformation. The TOCSY, DQF-COSY and NOESY spectra were acquired with 8, 8 and 16 scans, respectively, with a relaxation delay of 1 s. The signal from the residual water of the solvent was suppressed using pre-saturation during the relaxation delay and during the mixing time. The TOCSY and NOESY data were multiplied by 75° shifted sine-function in both dimensions. All the spectra were processed using the software PROSA.⁵⁵ The spectra were analyzed using the software X-EASY.⁵⁶

Structure Determination

The chemical shift assignment of the major conformer (the population of the minor conformer was < 10%) was obtained by the standard procedure using DQF-COSY and TOCSY spectra for intra-residual assignment and the NOESY spectrum was used for the sequential assignment.⁵⁷ The collection of structural restraints was based on the NOEs and vicinal ³J_NHα couplings. Dihedral angle constraints were obtained from the ³J_NHα couplings, which were measured from the 1D ¹H NMR spectra and from the intra-residual and sequential NOEs along with the macro GRIDSEARCH in the program CYANA.³¹ The calibration of NOE intensities versus ¹H-¹H distance restraints and appropriate pseudo-atom corrections to the non-stereo specifically assigned methylene, methyl and ring protons were performed using the program CYANA. On an average, approximately 100 NOE constraints and 20 angle constraints were utilized while calculating the conformers (Table 3). A total of 100 conformers were initially generated by CYANA and a bundle containing 20 CYANA conformers with the lowest target function values were utilized for further restrained energy minimization, using the CFF91 force field⁵⁸ with the energy criteria fit 0.1 kCal/mol/Å⁵⁹ in the program DISCOVER with steepest decent and conjugate gradient algorithms.⁶⁰ The resulting energy minimized bundle of 20 conformers was used as a basis for discussing the solution conformation of the different SRIF analogues. The structures were analyzed using the program MOLMOL.⁶¹

Additional abbreviations

Amp

4-aminomethylphenylalanine

Aph

4-amino-phenylalanine

CYANA

Combined assignment and dynamics algorithm for NMR applications

DMSO

dimethylsulfoxide

DQF-COSY

double quantum filtered correlation spectroscopy

IAmp

4-(N-Isopropyl)-aminomethyl-phenylalanine

NMR

nuclear magnetic resonance

NOESY

nuclear Overhauser enhancement spectroscopy

3D

three-dimensional

PROSA

Processing algorithms

RMSD

root mean square deviation

SAR

structure activity relationships

SRIF

somatostatin

TOCSY

total correlation spectroscopy.