Structure-Activity Relationships of 1-Benzoylazulenes at the OX1 and OX2 Orexin Receptors.

We previously demonstrated the potential of di- or trisubstituted azulenes as ligands (potentiators, weak agonists, and antagonists) of the orexin receptors. In this study we investigated 27 1-benzoylazulene derivatives, uncovering seven potentiators of the orexin response on OX1 and two weak dual orexin receptor agonists. For potentiators, replacement of the azulene scaffold by indole retained the activity of four out of six compounds. The structure-activity relationships for agonism and potentiation can be summarized into a bicyclic aromatic ring system substituted with two hydrogen-bond acceptors (1-position, benzoyl; 6-position, carboxyl/ester) within 7-8 Šof each other; a third acceptor at the 3-position is also well tolerated. The same pharmacophoric signature is found in the preferred conformations of the orexin receptor agonist Nag26 from molecular dynamics simulations. Subtle changes switch the activity between weak agonism and potentiation, suggesting overlapping binding sites.

Determinants of Orexin Receptor Binding and Activation-A Molecular Dynamics Study.

We assess the stability of two previously suggested binding modes for the neuropeptide orexin-A in the OX2 receptor through extensive molecular dynamics simulations. As the activation determinants of the receptor remain unknown, we simulated an unliganded receptor and two small-molecular ligands, the antagonist suvorexant and the agonist Nag26 for comparison. Each system was simulated in pure POPC membrane as well as in the 25% cholesterol-POPC membrane. In total, we carried out 36 µs of simulations. Through this set of simulations, we report a stable binding mode for the C-terminus of orexin-A. In addition, we suggest interactions that would promote orexin receptor activation, as well as others that would stabilize the inactive state.

Stapled truncated orexin peptides as orexin receptor agonists.

The peptides orexin-A and -B, the endogenous agonists of the orexin receptors, have similar 19-amino-acid C-termini which retain full maximum response as truncated peptides with only marginally reduced potency, while further N-terminal truncations successively reduce the activity. The peptides have been suggested to bind in an α-helical conformation, and truncation beyond a certain critical length is likely to disrupt the overall helical structure. In this study, we set out to stabilize the α-helical conformation of orexin-A15-33 via peptide stapling at four different sites. At a suggested hinge region, we varied the length of the cross-linker as well as replaced the staple with two α-aminoisobutyric acid residues. Modifications close to the peptide C-terminus, which is crucial for activity, were not allowed. However, central and N-terminal modifications yielded bioactive peptides, albeit with decreased potencies. This provides evidence that the orexin receptors can accommodate and be activated by α-helical peptides. The decrease in potency is likely linked to a stabilization of suboptimal peptide conformation or blocking of peptide backbone-receptor interactions at the hinge region by the helical stabilization or the modified amino acids.

Pharmacophore Model To Discover OX1 and OX2 Orexin Receptor Ligands.

Small molecule agonists and antagonists of the orexinergic system have key implications for research and therapeutic purposes. We report a pharmacophore model trained on ∼200 antagonists and prospectively validated by screening a collection of ∼137,000 compounds. The resulting hit list, 395 compounds, was tested for OX1 and OX2 receptor activity using calcium mobilization assay in recombinant cell lines. Validation was conducted using both calcium mobilization and [(125)I]-orexin-A competition binding. Compounds 4-7 have weak agonist activity and Ki's in the 1-30 µM range; compounds 8-14 are antagonists with Ki's in the 0.1-10 µM range for OX2 and 1-50 µM for the OX1 receptor. Docking simulations were used to devise a working hypothesis where two subpockets are important for activation, one between TM5 and TM6 lined by Phe5.42, Tyr5.47, and Tyr6.48 and another above the orthosteric pocket lined by Asp2.65 and Tyr7.32.

Modeling of the OX1R-orexin-A complex suggests two alternative binding modes.

BACKGROUND: Interactions between the orexin peptides and their cognate OX1 and OX2 receptors remain poorly characterized. Site-directed mutagenesis studies on orexin peptides and receptors have indicated amino acids important for ligand binding and receptor activation. However, a better understanding of specific pairwise interactions would benefit small molecule discovery. RESULTS: We constructed a set of three-dimensional models of the orexin 1 receptor based on the 3D-structures of the orexin 2 receptor (released while this manuscript was under review), neurotensin receptor 1 and chemokine receptor CXCR4, conducted an exhaustive docking of orexin-A16-33 peptide fragment with ZDOCK and RDOCK, and analyzed a total of 4301 complexes through multidimensional scaling and clustering. The best docking poses reveal two alternative binding modes, where the C-terminus of the peptide lies deep in the binding pocket, on average about 5-6 Å above Tyr(6.48) and close to Gln(3.32). The binding modes differ in the about 100° rotation of the peptide; the peptide His26 faces either the receptor's fifth transmembrane helix or the seventh helix. Both binding modes are well in line with previous mutation studies and partake in hydrogen bonding similar to suvorexant. CONCLUSIONS: We present two binding modes for orexin-A into orexin 1 receptor, which help rationalize previous results from site-directed mutagenesis studies. The binding modes should serve small molecule discovery, and offer insights into the mechanism of receptor activation.