|Home | About | Journals | Submit | Contact Us | Français|
The novel phenylalanine analogues 4'-[N-((4'-phenyl)-phenethyl)carboxamido]phenylalanine (Bcp) and 2',6'-dimethyl-4'-[N-((4'-phenyl)-phenethyl)carboxamido]phenylalanine (Dbcp) were substituted for Tyr1 in the δ opioid antagonist TIPP (H-Tyr-Tic-Phe-Phe-OH; Tic = tetrahydroisoquinoline-3-carboxylic acid). Unexpectedly, [Bcp1]TIPP was a potent, selective δ opioid agonist, whereas [Dbcp1]TIPP retained high δ antagonist activity. Receptor docking studies indicated similar binding modes for the two peptides, except for the biphenylethyl moiety which occupied distinct receptor subsites. The dipeptide H-Dbcp-Tic-OH was a highly selective δ antagonist with subnanomolar δ receptor affinity.
On the basis of early structure-activity relationship (SAR) studies performed with Met- and Leu-enkephalin (for a review, see ref.1) it was prematurely concluded that all structural modifications of the Tyr1 phenolic ring in opioid peptides would be detrimental to opioid activity, a belief that was held for nearly two decades. In a key-discovery, Mosberg and colleagues showed in 1992 that dimethylation at the 2',6'-positions of the Tyr1 residue of a cyclic enkephalin analogue, as achieved by substitution of 2',6'-dimethyltyrosine (Dmt), increased opioid agonist potency by 1–2 orders of magnitude.2 The most plausible explanation for this potency increase is that the two methyl groups may engage in additional binding interactions with residues in the receptor binding site. A more recent, interesting finding, independently reported by two groups,3,4 was that substitution of a carboxamido (−CONH2) group for the Tyr1 phenolic hydroxyl group in opioid peptides, as achieved by replacement of the tyrosine with p-carboxamidophenylalanine, resulted in compounds that retained high opioid activity. Replacement of the phenolic hydroxyl group in non-peptide opiates with the −CONH2 group had previously been shown to result in compounds retaining high μ receptor binding affinity5,6 and introduction of a (4'-phenyl)-phenethyl substituent at the −CONH2 group of 8-carboxamido-cyclazocine did not diminish μ receptor binding affinity.7 The latter finding prompted the design and synthesis of opioid peptide analogues in which the Tyr1 hydroxyl function is replaced by a ((4'-phenyl)-phenethyl)carboxamido group, which was achieved by substitution of 4'-[N-((4'-phenyl)-phenethyl)carboxamido]phenylalanine (Bcp; Fig. 1).8 The cyclic enkephalin analogue H-Bcp-c[D-Cys-Gly-Phe(pNO2)-D-Cys]NH2 turned out to be a potent, μ-selective agonist, indicating that the large biphenylethyl substituent not only was tolerated at the μ opioid receptor but likely enhanced binding affinity by interacting with a receptor subsite.
Delta opioid peptide antagonists containing 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid (Tic) in the 2-position of the peptide sequence, the so-called TIP(P) peptides, were first reported in 1992.9 The two prototypes, TIPP (H-Tyr-Tic-Phe-Phe-OH) and TIP (H-Tyr-Tic-Phe-OH) displayed high δ antagonist activity and high δ receptor selectivity. Subsequent SAR studies led to a number of TIPP-derived δ antagonists with subnanomolar δ receptor binding affinity (for a review, see ref.10), including H-Dmt-Tic-Phe-Phe-OH (DIPP). The dipeptide H-Tyr-Tic-OH was shown to be a weak δ opioid antagonist, whereas H-Dmt-Tic-OH showed low nanomolar δ antagonist activity and high δ receptor selectivity.11 Here, we describe the synthesis and opioid activity profiles of analogues of TIP(P) peptides, in which the Tyr1 residue was replaced by Bcp or by the novel Tyr analogue 2',6'-dimethyl-4'-[N-((4'-phenyl)-phenethyl)carboxamido]phenylalanine (Dbcp) (Fig. 1).
Boc-Bcp-OH was synthesized by activating the hydroxyl group of Boc-Tyr-OMe as the triflate, followed by carbonylation with carbon monoxide in the presence of potassium acetate, bis-(diphenylphosphinyl)ferrocene (DPPF) and palladium acetate, as described by Wang et al.,12 affording Boc-Phe(4'-COOH)-OMe in good yield. Subsequent reaction with 2-(4-biphenyl)ethylamine using O-(benzotriazol-1-yl)-N,N,N',N'-tetramethyluronium hexafluorophosphate (HBTU) as coupling agent, followed by ester hydrolysis with 2N NaOH yielded the target compound. Boc-Dbcp-OH was prepared in an analogous manner, starting with Boc-Dmt-OEt. Peptides were synthesized by standard solid-phase methods, were purified by preparative HPLC and were obtained in >95% purity.
The Bcp1-tetrapeptide analogue H-Bcp-Tic-Phe-Phe-OH (1) showed about 2-fold higher δ receptor binding affinity (Kiδ = 0.605 ± 0.058 nM) than the TIPP parent peptide (δ antagonist) (8) and retained high δ receptor binding selectivity (Table 1). Surprisingly, this compound turned out to be a potent full δ opioid agonist (IC50 = 3.42 ± 0.36 nM) in the MVD assay (Table 2). In the GPI assay, it displayed quite weak μ partial agonist properties. The C-terminally truncated tripeptide H-Bcp-Tic-Phe-OH (2) also showed subnanomolar δ receptor binding affinity, as well as high δ vs. μ and δ vs. κ binding selectivity. Unlike peptide 1, however, the Bcp1-tripeptide was a δ antagonist (Keδ = 18.7 ± 1.2 nM) with low efficacy δ partial agonist activity (maximal inhibition of contractions = 15%) in the MVD assay. The low efficacy partial agonist component may explain the discrepancy between the Keδ and Kiδ values determined for this compound. In the GPI assay, peptide 2 was characterized as a weak μ partial agonist. Further C-terminal truncation led to the dipeptide H-Bcp-Tic-OH (3) which also displayed subnanomolar δ receptor binding affinity and extraordinarily high δ receptor selectivity. In the MVD assay, compound 3 was a δ antagonist with a similar low efficacy δ partial agonist component as peptide 2. The 375-fold higher δ receptor binding affinity of H-Bcp-Tic-OH (Kiδ = 0.646 ± 0.088 nM) as compared to H-Tyr-Tic-OH (9, Kiδ = 242 ± 28 nM), indicates that the biphenylethyl-carboxamido group contributes in a major way to the binding interaction energy of compound 3.
The two pseudopeptides containing a reduced peptide bond between the 2- and 3- position residues, H-Bcp-TicΨ[CH2NH]Phe-Phe-OH (4) and H-Bcp-TicΨ[CH2NH]Cha-Phe-OH (5, Cha = cyclohexylalanine) displayed 6- and 13- fold lower δ receptor binding affinity than their respective Tyr1-containing parent pseudopeptides, which are δ opioid antagonists,11 and turned out to be nearly full δ agonists (e = 0.8) in the MVD assay with potencies 4–9-fold lower than those of tetrapeptide 1. These results indicate that the altered rotational mobility around the reduced peptide bond in these two pseudopeptides may position the biphenylethyl carboxamido group in a location at the δ receptor binding site that is less favorable for optimal δ receptor binding and activation.
The tetrapeptide containing Dbcp (Fig. 1) in place of Tyr1, H-Dbcp-Tic-Phe-Phe-OH (6), showed subnanomolar δ receptor binding affinity comparable to that of the Bcp1-tetrapeptide 1 and similarly high δ vs. μ and δ vs. κ receptor binding selectivity. However, unlike agonist 1, compound 6 retained high δ antagonist activity in the MVD assay and was a 2-fold more potent antagonist than the TIPP parent (8). H-Dbcp-Tic-Phe-Phe-OH (6) displayed 3-fold lower δ receptor binding affinity than H-Dmt-Tic-Phe-Phe-OH (10) and 1.5-fold lower δ affinity than H-Cdp-Tic-Phe-Phe-OH (12; Cdp = 4’-carboxamido-2’,6’-dimethylphenylalanine). Similarly, in the MVD assay, 6 also showed 15-fold and 5-fold decreases in δ antagonist activity (Ke value) as compared to 10 and 12, respectively. These comparisons indicate that the presence of the biphenylethyl substituent in H-Dbcp-Tic-Phe-Phe-OH does not cause an increase in δ receptor binding affinity. It is also possible that the interaction of the biphenylethyl substituent with the receptor may slightly alter the positioning of the 2’,6’-dimethylphenylalanine moiety at the receptor, resulting in the observed decrease in δ receptor binding affinity and δ antagonist activity. H-Dbcp-Tic-Phe-Phe-OH (6) has the same low μ receptor binding affinity as H-Dmt-Tic-Phe-Phe-OH (10) but, interestingly, 15-fold higher μ affinity than H-Cdp-Tic-Phe-Phe-OH (12). In agreement with the result of the μ receptor binding assay, 6 also showed about 8-fold higher μ antagonist activity than 12 in the GPI assay. This result indicates that the biphenylethyl substituent of 6 does strengthen binding to the μ receptor in the inactive state to some extent. On the other hand, 10 behaves as a weak agonist at the μ receptor.
The dipeptide H-Dbcp-Tic-OH (7) also displayed subnanomolar δ opioid receptor binding affinity (Kiδ = 0.825 ± 0.107 nM), very high δ receptor binding selectivity and subnanomolar δ antagonist activity (Keδ = 1.76 ± 0.30 nM) in the MVD assay. In comparison with H-Dmt-Tic-OH (11)10,11, it showed 2-fold higher δ receptor binding affinity and 4-fold higher δ antagonist activity. Thus, unlike in the case of tetrapeptide 6, the biphenylethyl moiety of 7 strengthens binding to the δ receptor to some extent. This result can be explained with the different binding mode of the biphenylethyl moiety in the dipeptide as compared to the tetrapeptide (see below). Compound 7 also displayed about 2-fold higher μ receptor binding affinity than 11 and, consequently, these two dipeptides have about he same high δ vs. μ receptor selectivity (Table 1). With the exception of a slightly elevated molecular weight (MW = 576), the physico-chemical properties of H-Dbcp-Tic-OH are in agreement with Lipinski’s rule of five, and this compound may turn out to be of interest as a pharmacological tool. Unlike peptides 2 and 3, which have a weak δ partial agonist component, 6 and 7 showed no δ opioid agonist activity in the MVD assay at concentrations up to 10 µM.
Studies of flexible docking of ligands to the δ opioid receptor were performed using Mosberg's models of the receptor in the inactive and in the activated state. The tetrapeptides H-Bcp-Tic-Phe-Phe-OH (1, δ agonist) and H-Dbcp-Tic-Phe-Phe-OH (6, δ antagonist) were docked to the activated and the inactive form of the δ receptor, respectively (Fig. 2). In general, a comparison of the ligand-receptor interactions of agonist 1 bound to the active form of the receptor with those of antagonist 6 bound to the inactive receptor form revealed that many of the interactions involved the same receptor residues, with notable exceptions. In both cases, the N-terminal amino group of the ligand formed a salt bridge with Asp128 of the receptor. The Phe4 residue of the two peptides bound to their respective receptor forms occupied only slightly different receptor regions, 2–3 Å apart. While in both cases interactions with Leu200 and Trp284 are seen, the Phe4 side chain of 1 also interacts with Gln201 and Ser204 in the active form of the receptor. The Phe3 residue in both peptides interacts with Val281, Trp284 and Leu300 and, additionally, with Cys303 in the case of agonist 1 and with Ile304 in the case of antagonist 6. The Tic2 residue in the two ligands also occupies a very similar receptor site, interacting with Met199 and Leu300 and, in addition, this residue in the agonist interacts with Cys303, while in the antagonist it is in close proximity to Ile304. In both ligands the phenyl ring of the phenylalanine portion of the 1-position residue interacts with Tyr129, Met132, Phe133 and Trp274, and additionally with Ile277 in the case of the antagonist.
Interestingly, the biphenylethyl substituent in agonist 1 and antagonist 6 interacts with different receptor regions and are seen to be about 4–5 Å apart from each other when the two receptor complexes are compared. While the biphenyl moiety of both ligands interacts with Lys214 and Phe218 of the receptor, it differentially interacts with Glu201, Ser204, Asp210 and Thr211 in the case of the agonist and with Ile183, Thr213 and Val217 in the case of the antagonist. The different positioning of the biphenyl moiety in the agonist- vs. antagonist-receptor complexes appears to be due to additional hydrophobic interactions of the 2',6'-dimethyl groups of Dbcp1 in the antagonist with Tyr127 and Trp274 of the receptor in the inactive state and is reflected in different dihedral angles for receptor-bound peptide 1 (Ψ1 = 89°, χ1= 177°, χ2= 69°) vs. receptor-bound peptide 6 (Ψ1 = 132°, χ1= 161°, χ2= 89°).
Flexible docking of the dipeptide antagonists H-Bcp-Tic-OH (3) and Dbcp-Tic-OH (7) to the δ receptor in the inactive state was also performed. A superposition of the receptor-bound conformations of the dipeptide antagonists 3 and 7 and the tetrapeptide antagonist 6 is shown in Fig. 3. As in the case of the tetrapeptide, the N-terminal amino group of the two dipeptides forms a salt bridge with Asp128 of the receptor. The Tic2 residue and the phenyl ring of the phenylalanine portion of the 1-position residue occupy the same receptor binding subsites as the corresponding moieties in the tetrapeptide. However, in comparison with the 2’,6’-dimethylphenyl moiety in docked tetrapeptide 6, that same moiety in the docked dipeptides 3 and 7 has a twisted orientation. As a consequence of this twist, the interaction of one of the 2’,6’-dimethyl groups in dipeptide 7 with Trp274, as seen in the case of the tetrapeptide, is lost, whereas one of the methyl groups still interacts with Tyr129 as in the case of the tetrapeptide. Most interestingly, the biphenyl moiety of the dipeptides assumes a completely different orientation as compared to the one in the tetrapeptide and interacts with the same receptor region as the phenyl rings of Phe3 and Phe4 in the tetrapeptide, particularly with residues Leu200 and Trp284. These interactions may explain the increased δ receptor binding affinity of H-Dbcp-Tic-OH as compared to H-Dmt-Tic-OH.
Replacement of the Tyr1 residue in TIPP peptides with a Bcp or Dbcp residue resulted in a general increase in δ opioid receptor binding affinity. All Bcp1- or Dbcp1- tetra-, tri- and dipeptide analogues displayed similar subnanomolar δ receptor binding affinity and high δ receptor binding selectivity. The dipeptide H-Dbcp-Tic-OH (7) is a low molecular weight δ opioid antagonist with subnanomolar δ receptor binding affinity and extraordinarily high δ receptor selectivity, and represents a novel pharmacological tool. The observation that [Bcp1]TIPP (1) is a δ opioid agonist, whereas [Dbcp1]TIPP (6) is a δ antagonist is of great significance. The flexible docking studies performed with these two tetrapeptides indicated that the large biphenylethyl group contained in the Bcp residue of the δ agonist interacts with an accessory binding site of the activated receptor distinct from the binding site of the biphenylethyl group of the Dbcp residue of the δ antagonist bound to the receptor in the inactive state. This finding demonstrates that the differential orientation of a large hydrophobic substituent introduced at an appropriate site of a ligand may result in the induction of or interaction with a distinct active or inactive receptor conformation. In the two docked dipeptide antagonists, the biphenylethyl moiety assumes yet another very different orientation, permitting it to interact with a receptor region similar to that with which the Phe3 and Phe4 residues of the tetrapeptide antagonist (6) interact. These results indicate that the δ opioid receptor can accommodate the biphenylethyl moiety contained in the 1-position residue of TIPP peptides in a number of different binding modes that promote the interaction of the ligand with the receptor in either an activated or an inactive state. In future studies we will replace the biphenylethyl group in the Bcp1- and Dbcp1-TIPP analogues described here with a variety of substituents containing large aromatic moieties that might interact with different hydrophobic subsites at the δ receptor, in an effort to further explore the agonist vs. antagonist behavior of this class of δ opioid receptor ligands.
Peptides were prepared by the manual solid-phase technique using Boc-protection for the α-amino group and DIC/HOBt or HBTU/DIEA as coupling agents. Peptides 1, 2, 4, 5, 6, and 12 were assembled on a polystyrene-divinylbenzene (1%) resin (100–200 mesh) (Boc-Phe-resin, 0.65 equiv./g, Bachem Bioscience, King of Prussia). For the synthesis of dipeptides 3 and 7 a Boc-Tic-resin (0.35 equiv./g) was prepared according to the Gisin method13, using a Merrifield resin (D-2120, 100–200 mesh, 1.05 mM/g Bachem Bioscience). Peptide assembly was carried out according to a published protocol.14 To introduce the reduced peptide bond between the Tic2 and Phe3 (or Cha3) residues, a reductive alkylation reaction15 between 2-Boc-1,2,3,4-tetrahydro-isoquinoline-3-aldehyde14 and the amino group of the resin-bound H-Phe-Phe- or H-Cha-Phe- dipeptide was performed as follows. 2-Boc-1,2,3,4-tetrahydroisoquinoline-3-aldehyde (2 equiv.) in DMF containing 1% AcOH was added to the resin. Sodium cyano-borohydride (5.0 equiv.) was then added portionwise over a period of 40 min and the reaction was allowed to continue for 3 h. After washing, deprotection and coupling of the Boc-Bcp-OH residue were performed according to the standard protocol. Peptides were cleaved from the resin and deprotected by HF/anisole treatment in the usual manner. After evaporation of the HF, the resin was extracted three times with Et2O and subsequently three times with glacial AcOH. The peptides were obtained in solid form through lyophilization of the acetic acid extracts. Peptides were purified by preparative reversed phase HPLC. Analytical parameters are listed in Supporting Information.
All calculations were performed using the molecular modeling software SYBYL version 7.0 (Tripos Associates). The standard Tripos force field was used for energy calculations and a dielectric constant of 1 was used. The TIPP analogues were constructed as previously described,16,17 using the standard fragment library in SYBYL, and were subjected to 300 ps of molecular dynamics simulation. Conformations were sampled every 10 picoseconds and were minimized, and the lowest-energy conformation from the dynamics simulation was used as the starting structure for flexible docking to the receptor. The resulting overall conformations of the Bcp1- and Dbcp1-peptide analogues were similar to the previously proposed δ receptor-bound conformations of TIPP peptides.16,17 Models of the δ opioid receptor in the inactive and in the activated state, constructed by Mosberg et al. by homology modeling based on the crystal structure of rhodopsin (http://mosberglab.phar.umich.edu/resources/), were used in the docking studies. Flexible docking was performed using the software program GLIDE (Schrödinger LLC). Each of the resulting ligand-receptor complexes was minimized using the conjugate gradient approach.18 Molecular dynamics simulations of 100 ps at 300 K were performed in order to assess the stability of each complex. In each case no significant change in the complex structure was observed during the simulation.
This work was financially supported by the U.S. National Institutes of Health (DA-004443) and the Canadian Institutes of Health Research (MOP-89716).
aAbbreviations: Bcp, 4'-[N-((4'-phenyl)-phenethyl)carboxamido]phenylalanine; Cha, cyclohexylalanine; DAMGO, H-Tyr-D-Ala-Gly-Phe(NMe)-Gly-ol; Dbcp, 2',6'-dimethyl-4'-[N-((4'-phenyl)-phenethyl)carboxamido]phenylalanine; DIC, 1,3-diisopropylcarbodiimide; DIEA, N,N-diisopropylethylamine; DIPP, H-Dmt-Tic-Phe-Phe-OH; Dmt, 2',6'-dimethyltyrosine; DPDPE, H-Tyr-c[D-Pen-Gly-Phe-D-Pen]OH; DPPF, 1,1’-bis-(diphenylphosphino)ferrocene; DSLET, H-Tyr-D-Ser-Gly-Phe-Leu-Thr-OH; GPI, guinea pig ileum; HBTU, O-(benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexafluorophosphate; HOBt, 1-hydroxybenzotriazole; MVD, mouse vas deferens; SAR, structure-activity relationships; Tic, tetrahydroisoquinoline-3-carboxylic acid; TAPP, H-Tyr-D-Ala-Phe-Phe-NH2; TIPP, H-Tyr-Tic-Phe-Phe-OH; TMH, transmembrane helix.
Supporting Information Available: Experimental details and refs 19–28. This material is available free of charge via the Internet.