Conformation-activity relationships of cyclo-constrained μ/5 opioid agonists derived from the N-terminal tetrapeptide segment of dermorphin/deltorphin

Abstract

The N-terminal tetrapeptide segments of dermorphin (Tyr-D-Ala-Phe-Gly-Tyr-Pro-Ser-NH2) and deltorphin (Tyr-D-Ala-Phe-Asp/Glu-Val-Val-Gly-NH2) are agonists at the opioid receptors µ and δ, respectively. (D-Arg2,Lys4)-dermorphin(1–4) amide (Tyr-D-Arg-Phe-Lys-NH2, DALDA) and [Dmt1]DALDA (Dmt is 2′,6′-dimethyl-tyrosine) are among the most potent and selective µ-agonists reported to date, both in vitro (the latter one having picomolar µ receptor affinity) and in vivo. In this communication, conformation-activity studies of cyclic tetrapeptide analogs of dermorphin/deltorphin are presented and discussed. They include the peptide Tyr-c[D-Cys-Phe-Cys]NH2, constrained via an SγS'γ disulfide between Cys2 and Cys4, and its dicarba analogs, the CγC'γ-saturated and -olefinic ones. They are potent nonselective or moderately µ -selective opioid agonists in vitro [1] (Table 1). With a major structural constraint imposed by the 11-membered ring spanning residues 2–4, they are expected to manifest well-defined conformations of the backbone. Given the small size and confirmed bioactivity of the peptides [1], there is a good chance that their conformations determined in solution would correspond to the receptor-bound ones. We used 2D-NMR in H2O/D2O supported with molecular dynamics (MD) in this study. Table 1 The analogs studied and their biological ativities a. Results and Discussion NMR spectra were recorded in H2O/D2O solution at 305 K with peptide concentrations about 3 mM, on a Varian Unity 500 Plus spectrometer with sodium 3-trimethylsilyltetradeuteriumpropionate, TSP, as internal standard. The assignment of proton chemical shifts was accomplished using 2D proton spectra TOCSY (80 ms), ROESY (150 ms), and DQF-COSY. Due to problems with solubility, productive sets of 2D spectra with usable cross-peaks could be collected only for analogs 1 and 3. Data were processed using ACDLabs [2] and XEASY [3] software. The 3JHNαH coupling constants were obtained from the DQF-COSY spectra. The structures of 1 and 3 were refined using 2.9 ns productive MD runs, utilizing the Time-Averaged Constraints procedure (TAV-MD) dedicated to structure deter-minations from NMR of small flexible peptides [4]. Prior to the productive MD, routine operations, including parameterization of new moieties [5], TAV-constrained model-building, energy minimization, thermalization, etc., were performed. MD simulations were carried out using the AMBER, ver. 8.0 software [5]. Each set of conformations from the TAV-MD trajectories was clustered into 5 to 6 families of conformations. Analogs 2 and 4 were submitted to a similar 2.9 ns MD, non-restrained by NMR data, and with starting structures taken at random. Remarkably, the structures of these two essentially diverse MDs converged into two clearly different types of structures: analogs 1 and 4 merged into one of these and analogs 2 and 3 into the other, as shown in Figure 1. Fig. 1 Left: 103 refined structures of 1; Right: 109 structures of 3. The aromatic rings of Tyr1 and Phe3 in 1 and 4 are on the same side of the heterodetic ring at an interacting distance, while in 2 and 3 the two rings are located far apart from one another. The remarkable merge of analogs 1 and 4 into one and analogs 2 and 3 into the other structural type correlates with the higher µ agonist activities and μ vs. δ selectivities of the former as compared to the latter. Thus, it is tempting to hypothesize that the enhanced μ agonist potency and µ receptor selectivity of analogs 1 and 4 may be attributed to the location of the two aromatic rings in close proximity and engaged in a stacking interaction. Of course, the exocyclic Tyr1 residue and the Phe3 side chain still enjoy considerable orientational freedom and it is possible that they may change their orientation upon binding to the receptor.