Flipping the GPCR Switch: Structure-Based Development of Selective Cannabinoid Receptor 2 Inverse Agonists

Abstract

We report a blueprint for the rational design of G protein coupled receptor (GPCR) ligands with a tailored functional response. The present study discloses the structure-based design of cannabinoid receptor type 2 (CB₂R) selective inverse agonists (S)-1 and (R)-1, which were derived from privileged agonist HU-308 by introduction of a phenyl group at the gem-dimethylheptyl side chain. Epimer (R)-1 exhibits high affinity for CB₂R with K_d = 39.1 nM and serves as a platform for the synthesis of a wide variety of probes. Notably, for the first time these fluorescent probes retain their inverse agonist functionality, high affinity, and selectivity for CB₂R independent of linker and fluorophore substitution. Ligands (S)-1, (R)-1, and their derivatives act as inverse agonists in CB₂R-mediated cAMP as well as G protein recruitment assays and do not trigger β-arrestin-receptor association. Furthermore, no receptor activation was detected in live cell ERK_1/2 phosphorylation and Ca²⁺-release assays. Confocal fluorescence imaging experiments with (R)-7 (Alexa488) and (R)-9 (Alexa647) probes employing BV-2 microglial cells visualized CB₂R expressed at endogenous levels. Finally, molecular dynamics simulations corroborate the initial docking data in which inverse agonists restrict movement of toggle switch Trp258^6.48 and thereby stabilize CB₂R in its inactive state.

Scheme 1. Novel Structure-Based Design of HU-308-Derived CB₂R-Selective Inverse Agonists That Actuate Trp258^6.48 Toggle Switch

Figure 1

Comparison of active (A, PDB 8GUS, ligand HU-308) and inactive (B, PDB 5ZTY, ligand AM10257) CB₂R conformations. (C) Docking study of HU-308-derived putative inverse agonist (R)-1 in the inactive CB₂R conformation (PDB 5ZTY). (R)-1 reaches into the secondary pocket occupied by the toggle switch responsible for CB₂R activation, Trp258^6.48 (orange), and shares binding interactions virtually identical with those of AM10257.

Figure 2

Design of inverse agonists (S)-1 and (R)-1.

Scheme 2. Synthesis of Novel CB₂R-Selective HU-308-Derived Inverse Agonists

Reagents and conditions: (a) 1-chloro-5-iodopentane, t-BuLi, n-pentane, Et₂O, −78 °C to rt, 83%; (b) KHMDS, CS₂, THF, −78 °C to rt and then MeI, 40 °C, 95%; (c) BBr₃, CH₂Cl₂, 0 °C, 97%; (d) Pd/C, H₂, EtOAc, rt, 97%; (e) semipreparative SFC, (S)-14, 25%, >99% ee, (R)-14, 20%, 96% ee; (f) 15, pTsOH·H₂O, CH₂Cl₂, rt, 64–71%; (g) (MeO)₂SO₂, K₂CO₃, acetone, rt, 78–82%; (h) NaN₃, DMF, 50 °C, 88–96%; (i) N₂H₄·H₂O, (E)/(Z)-crotyl alcohol, EtOH, 75 °C, 76–94%; (j) for conjugation conditions and details, see Supporting Information; (k) H₂O, microwave irradiation, 150 °C, 99%; (l) 4-nitrobenzoyl chloride, NEt₃, DMAP, CH₂Cl₂, rt, 85%.

Figure 3

TR-FRET-based saturation binding profile of (R)-7 (Alexa488) at CB₂R determined at 37 °C. Nonspecific binding was determined in the presence of SR-144,528 (10 μM). Data shown as a mean ± SEM, N = 3.

Figure 4

BRET-based Gi-CASE membrane assay to characterize G protein recruitment at CB₂R. Efficacy, E_max, of the compounds is shown as a mean ± SEM, N = 3–4.

Figure 5

BRET-based assay to characterize β-arrestin recruitment at CB₂R. L = ligand. Data are a representative of N = 3.

Figure 6

Live cell AlphaScreen SureFire phospho-ERK assay with CB₂R inducible breast cancer HCC1954 cell line. Cells were optionally induced with doxycycline (DOX) for 24 h to stimulate expression of CB₂R followed by incubation with a vehicle (0.1% DMSO), agonist JWH133 (1 μM), or (R)-2 (1 μM) for 30 min. Statistical significance was examined by one-way ANOVA followed by Tukey’s post hoc test. ns = nonsignificant, ∗∗, p < 0.01; ∗∗∗, p < 0.001. Data are an average of three independent biological replicates.

Figure 7

Live cell fluorescent Ca²⁺ imaging in rat CB₂R overexpressing AtT-20 cells [AtT-20(rCB₂R)] loaded with Fluo-4AM (2 μM). After initial equilibration, (S)-3/(R)-3 (20 μM) was added, followed by HU-308 (20 μM) and ionomycin (10 μM). Shown are the average responses of 200 cells (A, left), individual traces of five representative cells (A, right), and representative fluorescence images from different time points (B). Averaged data plotted as mean ± SEM, T = 4.

Figure 8

Confocal imaging of (R)-7 in live cell lines. (A) AtT-20(SNAP-hCB₂R) cells were labeled for 15 min with (R)-7 (Alexa488, 625 nM, green), SNAP-JF549i (JF549i, 500 nM, red), and Hoechst33342 (Hoechst, 1 μM, blue) to visualize CB₂R, SNAP-tags, and nuclei, respectively. Fluorescence intensity profiles across the white line for Alexa488, JF549i, and Hoechst33342 are shown on the right. (B) AtT-20(rCB₂R) cells (left) and AtT-20(WT) cells (right) were treated with (R)-7 (625 nM, green) and Hoechst33342 (1 μM, blue) for 15 min and imaged by confocal microscopy. (C) Live BV-2 microglial cells that endogenously express CB₂R were incubated with (R)-7 (2.5 μM, green) and Hoechst33342 (1 μM, blue) for 15 min and imaged by confocal microscopy.

Figure 9

Representative frames from molecular dynamics (MD) simulations of CB₂R (PDB 5ZTY) in complex with (S)-3 or (R)-3. Superimposition at level of protein backbone of (A) CB₂R X-ray structure with AM10257 (light gray) and (R)-3 MD complex (ligand in gold and protein in light yellow) and of (B) the two inverse agonist complexes (S)-3 (ligand in blue and protein light blue) and (R)-3 (ligand in gold and protein in light yellow).