Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

[D-Pen²,D-Pen⁵]-Enkephalin (DPDPE) is considered the prototypical -selective opioid peptide. Tetrapeptides that are des-Gly³ analogs of DPDPE and include conformational restriction by cyclization via disulfide or dithioether groups have been developed with selectivity for either the - or µ-opioid receptor (Mosberg et al., 1988). These compounds have extremely low affinity for the -opioid receptor (Mosberg et al., 1998). The features that make up the pharmacophore in these tetrapeptides are the aromatic ring, hydroxyl group, and primary amine of the Tyr¹ residue, the aromatic ring of the Phe³ residue, and the C-terminal group (either carboxylic acid or carboxamide). The example shown in Fig. 1 is JOM-6 [Tyr-c[D-Cys-Phe-D-Pen]NH₂ (Et)], a cyclic tetrapeptide that exhibits high affinity and 75-fold selectivity for the µ-opioid receptor. As shown in Fig. 1, JOM-6 is cyclized via a dithioethane-bridging group between the side chain atoms of the D-Cys and D-Pen residues and contains a C terminus carboxamide. In contrast, the closely related JOM-5 (Tyr-c[D-Cys-Phe-D-Pen]NH₂) with a disulfide-bridging group exhibits approximately 21-fold reduced affinity for the µ-opioid receptor as a direct result of changes to the relative positions of the aromatic rings of the first and third amino acids (Mosberg et al., 1996).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Structure of JOM-6, Tyr-c[D-Cys-Phe-D-Pen]-NH2 (Et).

Traditionally, the para-hydroxyl substituent of the N-terminal residue has been highlighted as one of the critical pharmacophoric elements in opioid peptides (Casy and Parfitt, 1986). However, replacing the Tyr¹ residue of JOM-6 with Phe¹ gives the peptide Phe-c[D-Cys-Phe-D-Pen]NH₂ (Et) (JH-54), which we have recently reported to possess an only 4-fold reduced affinity at the µ-opioid receptor, compared with the parent compound, and to exhibit agonist activity in the guinea pig ileum preparation (Mosberg et al., 1998). This finding is surprising in view of the emphasis placed on the para-hydroxyl substituent of the Tyr residue in the traditional µ-opioid pharmacophore for peptides. Indeed these findings are more reminiscent of structure-activity requirements at the orphanin receptor (ORL₁), where the preferred N terminus is phenylalanine (Calo et al., 2000). However, several nonpeptide µ-ligands do lack a phenolic hydroxy group, including the prototypical µ-ligands fentanyl and methadone and related compounds (Casy and Parfitt, 1986; Subramanian et al., 2000).

In this work we further characterize the µ-agonist properties of the cyclic tetrapeptide JH-54 and show that in the [³⁵S]GTPS binding assay in membranes from cloned C₆ cells expressing the µ-opioid receptor, it possesses activity equivalent to that seen with the full µ-agonist fentanyl. In addition we report a series of analogs of JH-54 in which either the bridging group between amino acid residues 2 (Cys) and 4 (Pen) is changed or the N-terminal Phe¹ residue is replaced with a variety of synthetic amino acids that lack hydroxyl substitution. The results highlight the importance of the bridging group and the lack of a need for a tyrosine hydroxyl group and reveal the conformational requirements necessary for interactions of this series of peptides with the µ-opioid receptor. One especially notable finding is that replacement of the N-terminal residue by the nonaromatic cyclohexylalanine affords an analog of at least moderate affinity and potency in addition to good relative efficacy in the [³⁵S]GTPS binding assay, thus further challenging the structural requirements for binding to the µ-opioid receptor.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Peptide Synthesis. All peptides were prepared by standard solid-phase methods similar to those previously described for the synthesis of JOM-6 (Mosberg et al., 1988), using chloromethylated polystyrene (Merrifield) resin cross-linked with 1% divinylbenzene. All protected amino acids were obtained from commercial sources (Peptides International or Advanced Chemtech, both Louisville, KY), with the exception of those noted below. Trifluoroacetic acid (TFA) was used for deprotection, and dicyclohexylcarbodiimide (DCC) and 1-hydroxybenzotriazole (HOBt) were used to facilitate coupling. -Amino functions were tert-butyloxycarbonyl (Boc)-protected, and p-methylbenzyl protection was used for the labile sulfhydryl side chain groups of Cys and Pen. Simultaneous deprotection and cleavage from the resin were accomplished by treatment with anhydrous hydrogen fluoride in the presence of 5% p-cresol and 5% p-thiocresol for 45 min (Heath et al., 1986) at 0°C. Before cyclization, linear peptides were purified with reverse-phase high-performance liquid chromatography (RP-HPLC) on a Vydac (Hesperia, CA) 218TP C-18 column (2.5 × 22 cm), using the solvent system 0.1% TFA in water/0.1% TFA in acetonitrile by a gradient of 0 to 50% organic component in 50 min. Disulfide-containing analogs were prepared by treatment of an aqueous solution (pH 8.5) of the corresponding linear-free sulfhydryl-containing species with K₃Fe(CN)₆ (Mosberg et al., 1983). Cyclization to dithioether-containing analogs was accomplished by treatment of a dilute solution of the linear-free sulfhydryl-containing species in dimethyl formamide with potassium tert-butoxide, followed by the addition of the appropriate alkyl dibromide (Mosberg et al., 1987). All peptides were then purified with RP-HPLC, using the solvent system described above, and pure fractions were lyophilized. Final product confirmation was obtained using fast atom bombardment mass spectrometry. In all cases, the anticipated molecular weights were confirmed using fast atom bombardment mass spectrometry. For all peptides, final product purity as assessed through thin-layer chromatography (TLC) and analytical RP-HPLC, using the solvent system described above, was >95%.

c-PhPro. The racemic mixture of c-PhPro was converted to the corresponding methyl ester and subjected to treatment by chymotrypsin for 6 days to preferentially liberate the L stereoisomer of the free c-PhPro carboxylic acid (Mosberg et al., 1994). The free carboxylic acid was identified by liquid chromatography-mass spectrometry and purified by RP-HPLC. The resulting L-c-PhPro was subjected to chiral TLC [silica gel RP modification coated with Cu²⁺ and chiral reagent (Macherey-Nagel, Schweizerhall Inc., Piscataway, NJ) elution solvent H₂O:CH₃OH:CH₃CN (1:1:4), Rf = 0.62]. In a parallel analysis, samples of the c-PhPro¹-containing peptides 8a and 8b were subjected to acid hydrolysis in constant boiling HCl at 100°C for 24 h. Free cis-3-phenylproline was identified in both hydrolysates using liquid chromatography-mass spectrometry and was isolated using the RP-HPLC procedure described previously. c-PhPro from both hydrolyses was analyzed using chiral TLC. The resulting values [Rf (c-PhPro from 8a) = 0.62; Rf (c-PhPro from 8b) = 0.43] allow the assignment of stereochemistry of the c-PhPro in 8a as L and that in 8b as D. As an additional test, L-c-PhPro isolated from the chymotrypsin hydrolysis of the racemic methyl esters was subjected to optical rotation and compared with reported values for this amino acid (Belokon et al., 1988). These results confirmed the assignment.

t-PhPro. Chymotrypsin proved to be insufficiently selective toward the trans-3-L-phenylproline-OCH₃; thus, this procedure could not be used to resolve both stereoisomers, as was done in the case of the c-DL-PhPro mixture. Instead, t-DL-PhPro was converted to the corresponding N-acetyl-trans-3-phenylproline-methylbenzylamides and resolved using silica gel chromatography. Measurement of the optical rotation of the resolved isomers and comparison with the literature report of Chung et al. (1990) allowed unequivocal assignment. Each of the resultant stereochemically assigned N-acetyl-trans-3-phenylproline-methylbenzylamides was then hydrolyzed, using concentrated constant boiling HCl, purified, and compared with amino acids obtained from the hydrolysis of the peptides 7a and 7b, using chiral TLC as described above. This allowed the unequivocal assignment of -center stereochemistry of the t-PhPro residues as being D in 7a and L in 7b.

Chemicals and Drugs. [³H][D-Ala²,N-Me-Phe⁴,Gly⁵-ol]-Enkephalin (DAMGO; 54.5 Ci/mmol; 2.02 TBq/mmol) was obtained from Amersham Pharmacia Biotech (Piscataway, NJ). [³⁵S]GTPS (1250 Ci/mmol; 46.25 TBq/mmol), [³H]nociceptin (60 Ci/mmol; 2.2 TBq/mmol), and [³H][DPDPE (45 Ci/mmol; 1.7 TBq/mmol) were purchased from DuPont NEN (Boston, MA). Fentanyl HCl and naloxone HCl were generous gifts from the National Institute on Drug Abuse (Rockville, MD). Dulbecco's modified Eagle's medium (without sodium pyruvate; with 4500 mg l¹ glucose), minimum essential medium (with Earle's salts), fetal calf serum, penicillin/streptomycin, fungizone, trypsin, EDTA, and Geneticin were all from Life Technologies (Grand Island, NY). All other chemicals were of analytical grade and were purchased from Sigma Chemical Co. (St. Louis, MO).

Membrane Preparation for Biological Assays

Guinea Pig Brain Homogenates. For opioid binding assays, guinea pig brains (Pel-Freez Biologicals, Rogers, AR) were suspended in cold 50 mM Tris buffer, pH 7.4, and homogenized using a Polytron homogenizer (Brinkmann Instruments, Westbury, NY). The homogenate was centrifuged for 15 min at 14,000g at 4°C, and the supernatant was discarded. The pellet was resuspended in cold 50 mM Tris buffer, pH 7.4, homogenized, and recentrifuged. The pellet was suspended, and the homogenate was incubated at 37°C for 30 min to release endogenous opioids. After recentrifugation, the pellet was resuspended at a final tissue concentration of approximately 0.05 g/ml in cold 50 mM Tris buffer, pH 7.4, and stored in aliquots at 80°C.

Cultured Cells. For the [³⁵S]GTPS binding assay, rat glioma cells stably transfected with the rat µ-opioid receptor (C₆µ, passages 15-25) (Emmerson et al., 1996) were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and Geneticin at 1 mg/ml. SH-SY5Y cells (passages 75-85) for use in the [³H]nociceptin binding assays were cultured in minimum essential medium, supplemented with 10% fetal calf serum, 2.5 µg/ml amphotericin B (fungizone), 50 µg/ml penicillin/streptomycin, and 250 µg/ml L-glutamine at 37°C in a humidified 5% CO₂ atmosphere. C₆µ and SH-SY5Y cells were grown to confluency in monolayers at 37°C in a humidified 5% CO₂ atmosphere and harvested by agitation in HEPES (20 mM, pH 7.4)-buffered saline containing 1 mM EDTA. After centrifugation at 500g, the cell pellet was suspended in a buffer (pH 7.4) containing 20 mM HEPES, 100 mM NaCl, and 10 mM MgCl₂·6H₂O (buffer A) and homogenized using a Tissue Tearor (Biospec, Bartlesville, OK). The resultant homogenate was centrifuged at 40,000g, and the pellet was collected, washed in buffer A, and recentrifuged. The pellet was resuspended in buffer A to give a protein concentration of 1 to 2 mg/ml, then stored in aliquots at 80°C. All procedures were carried out at 4°C.

Radioligand Binding Assays. Opioid ligand binding assays were based on the displacement by the test compounds of radiolabeled (³H) ligands from opioid receptors in guinea pig brain membranes. The labeled ligands used were DAMGO (µ-ligand; 0.6 nM) and DPDPE (-ligand; 1.8 nM). The receptor binding assays were performed as described previously (Medzihradsky et al., 1984; Clark et al., 1988). The assay mixture, containing membrane suspension in 50 mM Tris buffer (pH 7.4), radiolabeled ligand, and test compound, was incubated at 25°C in triplicate for 1 h to allow binding to reach equilibrium. Subsequently, the samples were filtered rapidly, and the radioactivity retained was determined by liquid scintillation counting. Inhibition of radiolabeled ligand binding by the test compounds was determined from maximal specific binding, measured with an appropriate excess of unlabeled naloxone (10 µM).

[³⁵S]GTPS Binding Assay. Agonist stimulation of [³⁵S]GTPS binding in cell lines containing cloned receptors was measured as described previously (Traynor and Nahorski, 1995). Briefly, membranes prepared from C₆µ cells as described above were incubated for 60 min at 30°C in buffer A containing [³⁵S]GTPS (100 pM), GDP (10 µM), and varying concentrations of ligand in a total volume of 1 ml. Basal binding of [³⁵S]GTPS was determined in the absence of unlabeled ligand, and maximal stimulation was defined using fentanyl (10 µM). Bound and free [³⁵S]GTPS were separated by vacuum filtration through GF/B filters and quantified by liquid scintillation counting.

Data Analysis. IC₅₀ values were obtained by linear regression from plots relating inhibition of specific binding to the log of 12 different ligand concentrations, using the computer program LIGAND (Munson and Rodbard, 1980) (Biosoft Software, Milltown, NJ). K_i values were calculated using values for K_D of each radioligand previously determined from saturation binding assays (Cheng and Prusoff, 1973). EC₅₀ values for the [³⁵S]GTPS binding experiments were calculated using GraphPad Prism, version 2.01 (GraphPad, San Diego, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

The Phe¹-containing cyclic tetrapeptide JH-54, Phe-c[D-Cys-Phe-D-Pen]NH₂ (Et), exhibits high binding affinity (1.4 nM) and 730-fold selectivity for the µ-opioid receptor (Mosberg et al., 1998). In contrast, the closely related Tyr¹ analog JOM-6, Tyr-c[D-Cys-Phe-D-Pen]NH₂ (Et), shows only a 4.8-fold better affinity than its Phe¹ counterpart and an 8.5-fold reduction in selectivity (Mosberg et al., 1988; Table 1). To further characterize the structural requirements for binding at the µ-opioid receptor, we have synthesized and pharmacologically evaluated several analogs of JH-54. The relative positions of the aromatic rings in residues 1 and 3 have been shown to be important in determining the binding affinity and selectivity of these cyclic tetrapeptides (Mosberg et al., 1996; Wang et al., 1998). Consequently, the bridging group of JH-54 was reduced from dithioethane to a disulfide bond, giving Phe-c[D-Cys-Phe-D-Pen]NH₂ (1) (Fig. 2). To investigate the conformational requirements for binding, the following modified amino acids were incorporated in JH-54 as residue 1 substitutions: phenylglycine (Pgl) (2), homophenylalanine (Hfe) (3), diphenylalanine (Dip) (4), 3-(1-naphthylalanine) (1-Nal) (5), 3-(2-naphthylalanine) (2-Nal) (6), trans-phenylproline (t-PhPro) (7), cis-phenylproline (c-PhPro) (8), 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid (Tic) (9), 2-amino-2-carboxytetralin (Atc) (10), and cyclohexylalanine (Cha) (11). The structures of the N-terminal amino acids are shown in Fig. 2.

View this table: [in this window] [in a new window]   TABLE 1 Opioid receptor binding profiles of cyclic tetrapeptides Guinea pig brain homogenates were incubated with radioligand ([3H]DAMGO 0.6 nM for µ or [3H]DPDPE 1.8 nM for  studies) and varying concentrations of peptide in 50 mM Tris-HCl buffer (see Materials and Methods). Values represent means ± S.E.M. from three or more separate experiments performed in duplicate. Selectivities were calculated as the ratio of µ and  affinities.

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Structure of the N-terminal amino acids of analogs 1-11. The bridging group used is dithioethane in all cases, except where noted, i.e., in the analog 1 (disulfide).

Radioligand Binding. Binding affinities of the various synthesized peptides for µ- and -opioid receptors are given in Table 1 and compared with values for DPDPE and DAMGO. JH-54 has 730-fold selectivity for µ- over -receptors, whereas the disulfide-bridged analog 1 exhibited a 250-fold reduced affinity for the µ-receptor compared with JH-54 (K_i = 352 nM) and extremely low affinity for the -receptor (>10,000 nM).

[³⁵S]GTPS Binding. JH-54, containing an N-terminal Phe residue and a dithioether-bridging group, exhibited potency and relative efficacy less than the Tyr¹-containing and dithioether-bridged JOM-6. However, JH-54 did show relative efficacy equivalent to the Tyr¹-containing and disulfide-bridged JOM-5 and to the standard µ-opioids fentanyl and DAMGO, although the disulfide-bridged analog of JH-54 (1) was only a weak partial agonist (Table 2). N-terminal substituted analogs of JH-54 that exhibited affinity for the µ-opioid receptor of <300 nM were also examined in the [³⁵S]GTPS binding assay and compared with the full µ-agonists fentanyl, DAMGO, JOM-5, and JOM-6 (Table 2). Most of the compounds tested afforded EC₅₀ values between 1.0 and 4.2 times higher than their affinities (K_i) as measured by ligand binding assay. However JOM-6, JH-54, 1, and 9b had EC₅₀ values approximately 9- to 13-fold higher than their respective K_i values (Table 2).

View this table: [in this window] [in a new window]   TABLE 2 Potency and relative efficacy as measured in the [35S]GTPS assay Membranes from C6µ cells were incubated with [35S]GTPS (100 pM), GDP (10 µM), and varying concentrations of ligand in buffer A (see Materials and Methods). Values represent means ± S.E.M. from three or more separate experiments performed in duplicate.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results confirm that a hydroxyl moiety in the N-terminal residue is not an absolute requirement for the binding of this family of cyclic peptides to the µ-opioid receptor. Moreover, these compounds and the previously reported JH-54 have been shown to retain agonist properties, with potencies in many cases only 10-fold or less reduced compared with the prototypical µ-agonist DAMGO. Indeed, the ring present in the first residue need not even be aromatic because the peptide 11 exhibited reasonable affinity and potency and was a full agonist despite possessing a cyclohexyl ring. At the -receptor inclusion of a Phe¹ or Cha¹ residue was very deleterious to binding. Two other peptides that exhibit high affinity for the µ-opioid receptor despite lacking a para-hydroxyl group in the initial residue are the cyclic octapeptides CTAP and CTOP. These are related to a short stretch of the cyclic tetradecapeptide somatostatin (Pelton et al., 1985, 1986) with the general structure D-Phe-c[Cys-Tyr-D-Trp-X-Thr-Pen]-Thr-NH₂, X = ornithine (CTOP) or arginine (CTAP). However, unlike the tetrapeptides reported here, both are antagonists at the µ-opioid receptor. In addition Schiller et al. (2000) have reported on cyclic hexapeptides that cannot carry a positive charge. One of these compounds also lacks a phenolic hydroxy group. This compound binds to the -opioid receptor but is an antagonist.

The relative positions of the aromatic rings of the first and third residues, directly affected by the nature of the bridging group, are also important in binding to the µ-opioid receptor (Mosberg et al., 1988, 1996). The Tyr¹-containing and dithioethane-bridged peptide JOM-6 showed approximately 25-fold higher affinity than the corresponding disulfide analog JOM-5. Replacement of the Tyr¹ residue of JOM-6 with Phe¹, giving JH-54, resulted in only a 5-fold reduction in affinity and potency at the µ-opioid receptor. In contrast, loss of the tyrosyl hydroxyl group of JOM-5, giving 1, caused a drastic reduction in µ-receptor binding affinity and potency. Thus, only when cyclization is via a dithioether bridge as in JOM-6, and not when cyclized via a disulfide bond as in JOM-5, can the Tyr¹ residue be replaced with Phe¹ without abolishing affinity.

An assumption in the majority of structure-activity relationship studies is that all structurally related compounds with high affinity for a particular receptor bind in the same manner. This allows the use of conformationally restricted residues to define the space available within the binding pocket of the receptor. Optimal affinity for the µ-opioid receptor occurred when the side chain linking the  carbon and the aromatic ring of the first residue is -CH₂, as in phenylalanine. When the side chain is shortened as in phenylglycine (2) or expanded as in homophenylalanine (3), affinity is much reduced. Affinity at the -receptor is similarly affected. Thus, in the analogs 2 and 3, the conformational space accessible to the N-terminal residue phenyl ring does not allow it to assume a favorable position relative to that of the third residue. However, additional bulk in the N-terminal amino acid can be accommodated with only mild unfavorable interactions when naphthalene is incorporated into the first residue via attachment at either the 1 or 2 position (compounds 5 and 6). When the first residue incorporates an additional phenyl ring as a substitution at the  carbon, as in 4, µ-receptor binding affinity is greatly reduced. Thus, the cavity within the binding pocket that accommodates the tyrosine aromatic ring is large enough to accept the relatively compact naphthalene group, but not the much bulkier diphenylalanine.

The cis and trans isomers of phenylproline (7 and 8) were used to examine the effects of conformational restriction about the C-C bond without constraining rotation about the C-C bond in the first amino acid. Because a racemic mixture of each phenylproline isomer was used in the synthesis, four analogs resulted: a pair of trans-phenylprolines (L and D) and a pair of cis-phenylprolines (L and D). Within each pair, one isomer showed high affinity for the µ-opioid receptor (7b and 8b), but the complementary pair exhibited low affinity (7a and 8a). The difference in affinities between enantiomers was approximately 75-fold for the trans-phenylprolines and 225-fold for the cis-phenylprolines. The high-affinity analogs both exhibited moderate potencies in the [³⁵S]GTPS assay. As described above, we have unequivocally assigned the stereochemistry of the phenylproline residue in each of the four compounds, identifying the higher affinity analogs as the D-c-PhPro¹ (8b)- and L-t-PhPro¹ (7b)-containing peptides. Although the different  carbon stereochemistry of these two high-affinity analogs may at first appear surprising, superpositioning of the four phenylproline isomers provides an explanation. As seen in Fig. 3 the L-trans- and D-cis-phenylproline pair can assume identical orientations of the critically important amine and phenyl moieties, as can the D-trans and L-cis stereoisomer pair. However, in each of these pairs, superpositioning of the amine and phenyl groups results in differing orientations of the carboxyl moieties. Hence, at the µ-receptor site, which must accommodate the C-terminal tripeptide, the relative orientations of analogs 7b and 8b are slightly offset from the exact superposition of tyramine portions indicated in Fig. 3, to allow the C termini to assume better register.

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Overlap of L-trans-phenylproline, L-cis-phenylproline, D-trans-phenylproline, and D-cis-phenylproline. Heteroatoms are shown as indicated. For clarity, only the hydrogen of the secondary amino group is shown.

Restriction of rotation about the C-C and C-C bonds has been examined. First, formation of a bicyclic structure via cyclization to the amide nitrogen gave the N-terminal 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid analog 9, which had drastically reduced affinity for both the µ- and -opioid receptors. Second, formation of a bicyclic structure by cyclization to the  carbon gave the 2-amino-2-carboxytetralin analogs 10. Synthesis of 10 gave rise to the two stereoisomers (10a and 10b) that bound to the µ-opioid receptor with affinities of 7.2 and 44.5 nM, respectively; this represents 5-fold and 32-fold decreases compared with JH-54. Although both analogs are potent agonists as determined in the [³⁵S]GTPS binding assay, 10a showed significantly reduced ability to stimulate [³⁵S]GTPS binding. These findings contrast with L- and D-2-amino-2-carboxy-6-hydroxytetralin analogs of JOM-6, which exhibit affinity, potency, and efficacy identical with the parent at both µ- and -opioid receptors (McFadyen et al., 2000).

Perhaps the most surprising finding was that the Cha¹-containing analog 11 exhibited approximately 100-fold reduced, but still good, affinity when compared with JOM-6. Agonist potency was decreased almost 20-fold, but 11 did produce stimulation of [³⁵S]GTPS binding equivalent to fentanyl, indicating good efficacy. An aromatic ring is traditionally considered a crucial pharmacophoric element in opioid peptides (Morgan et al., 1976; Chang et al., 1976). Thus, removal of aromaticity in the N-terminal amino acid was expected to cause a drastic loss of affinity and efficacy. However, binding to the µ-receptor decreased, but was not abolished. As expected, 11 failed to exhibit affinity for the -receptor.

The compounds reported here are the first series of peptide agonists for the µ-opioid receptor that lack a para-hydroxyl group in the N-terminal residue. In common with the endogenous ligand for the related ORL₁ receptor, nociceptin (Meuneir et al., 1995) or orphanin FQ (Reinscheid et al., 1995), these tetrapeptides contain N-terminal phenylalanine or related residues lacking a para-hydroxyl group. However, neither the Tyr¹ peptide JOM-6, the Phe¹ analog JH-54, nor the Cha¹ compound 11 displaced bound [³H]nociceptin at concentrations up to 10 µM. These peptides are therefore not only selective for the µ- over the -opioid receptor, but also for the µ-opioid receptor over the ORL₁ receptor.

In conclusion, in the series of cyclic tetrapeptides with the general structure Tyr-c[D-Cys-Phe-D-Pen]NH₂ (Et), the Tyr¹ residue can be replaced with Phe¹ and a variety of related aromatic residues lacking a hydroxyl group without drastic reductions in affinity, potency, or relative efficacy at the µ-opioid receptor. Thus, although the tyrosyl hydroxyl group does play a role in the interaction of peptides with the µ-opioid receptor, this role is not critical. Modeling studies reveal that when the Tyr¹ peptide JOM-6 is docked to the µ-opioid receptor, the side chain of a Trp residue in transmembrane domain VII interacts with the cyclic system of the peptide. This causes a small shift in the orientation of the whole peptide relative to that assumed at the -opioid receptor. This moves the phenolic oxygen of the Tyr¹ residue from its presumed binding partner, His²⁹⁷ in transmembrane domain VI (Mosberg et al., 1998), such that this hydrogen bonding interaction makes only a minor a contribution. Hence, the Phe¹ (and related) analogs retain considerable affinity for the µ-receptor, but analogous substitutions in -receptor ligands lead to considerable losses in -receptor binding affinity. Moreover, even an N-terminal aromatic residue is not vital, because the cyclohexylalanine analog 11 is also an agonist at the µ-opioid receptor. Consequently, it is possible that any hydrophobic group will be partially able to substitute for the aromatic ring of the N-terminal amino acid by forming suitable van der Waals interactions with the appropriate region of the µ-receptor.