Subtle Structural Modification of a Synthetic Cannabinoid Receptor Agonist Drastically Increases its Efficacy at the CB1 Receptor

Abstract

The emergence of synthetic cannabinoid receptor agonists (SCRAs) as illicit psychoactive substances has posed considerable public health risks, including fatalities. Many SCRAs exhibit much higher efficacy and potency compared with the phytocannabinoid Δ⁹-tetrahydrocannabinol (THC) at the cannabinoid receptor 1 (CB1R), leading to dramatic differences in signaling levels that can be toxic. In this study, we investigated the structure-activity relationships of aminoalkylindole SCRAs at CB1Rs, focusing on 5F-pentylindoles containing an amide linker attached to different head moieties. Using in vitro bioluminescence resonance energy transfer assays, we identified a few SCRAs exhibiting significantly higher efficacy in engaging the G_i protein and recruiting β-arrestin than the reference CB1R full agonist CP55940. Importantly, the extra methyl group on the head moiety of 5F-MDMB-PICA, as compared to that of 5F-MMB-PICA, led to a large increase in efficacy and potency at the CB1R. This pharmacological observation was supported by the functional effects of these SCRAs on glutamate field potentials recorded in hippocampal slices. Molecular modeling and simulations of the CB1R models bound with both of the SCRAs revealed critical structural determinants contributing to the higher efficacy of 5F-MDMB-PICA and how these subtle differences propagated to the receptor-G protein interface. Thus, we find that apparently minor structural changes in the head moiety of SCRAs can cause major changes in efficacy. Our results highlight the need for close monitoring of the structural modifications of newly emerging SCRAs and their potential for toxic drug responses in humans.

Keywords: bioluminescence resonance energy transfer; cannabinoid receptor 1; molecular dynamics; synthetic cannabinoids.

Figure 1

G_i1 engagement BRET drug-induced engagement BRET between CB1R-Rluc and G_i1-Venus (A) measured in response to CP55940 (B), AM2201 (C), 5F-NNEI (D), 5F-SDB-006 (E), 5F-CUMYL-PICA (F), 5F-MMB-PICA (G), and 5F-MDMB-PICA (H) at various time points (2, 16, 30, 44 min light to dark blue). Concentration–response curves are plotted as a percentage of maximal response by CP55940 at each time point and presented as means ± SEM of n ≥ 3 independent experiments.

Figure 2

β-Arrestin 2 recruitment BRET drug-induced recruitment BRET between CB1R-Rluc and β-arrestin 2-Venus (A) measured in response to CP55940 (B), AM2201 (C), 5F-NNEI (D), 5F-SDB-006 (E), 5F-CUMYL-PICA (F), 5F-MMB-PICA (G), and 5F-MDMB-PICA (H) at various time points (2, 16, 30, and 44 min light to dark blue). Concentration–response curves are plotted as a percentage of maximal response by CP55940 at each time point and presented as means ± SEM of n ≥ 3 independent experiments.

Figure 3

Tested SCRAs exhibited similar efficacy and potency trends in the G_i1 protein engagement and β-arrestin 2 recruitment assays E_max-pEC₅₀ comparison between CB1R-G_i1 engagement and β-arrestin 2 recruitment ^#Y-axis cutoff value at the highest concentration is shown for its E_max.

Figure 4

Brain slice electrophysiology concentration-dependent inhibition of hippocampal glutamate release by 5F-MDMB-PICA and 5F-MMB-PICA. (A) Representative averaged traces for 5F-MDMB-PICA (upper) and 5F-MMB-PICA (lower). (B) Summary time course of recordings (n ≥ 3 recordings) demonstrating the effect of SCRAs. (C) Concentration–response curve for 5F-MDMB-PICA and 5F-MMB-PICA (n ≥ 3 recordings per concentration). The pEC₅₀ was calculated to be 7.20 and 6.72 for 5F-MDMB-PICA and 5F-MMB-PICA, respectively. Data points are presented as means ± SEM.

Figure 5

Ligands’ dihedral angle and binding pocket. (A) Zoom-out view of the receptor (gray) bound with Gαi (dark gray) and MDMB-FUBINACA (magenta). (B) Binding mode of 5F-MDMB-PICA (green), MDMB-FUBINACA (magenta), and 5F-MMB-PICA (yellow and orange) and their associated binding site residues that show differences in contact frequency between 5F-MDMB-PICA, MDMB-FUBINACA, and 5F-MMB-PICA. The residues that show differences are color-coded correspondingly. (C) Zoom-in view of the superimposed ligands in the binding site. They occupy the same region in the binding pocket, but 5F-MMB-PICA picks two distinct conformers for the head moiety. (D) 2D map of head moiety dihedral angles (Supporting Information Figure S1).

Figure 6

PIA-GPCR analysis and TMs’ COM distribution (A) PIA-GPCR COM results for TMs’ extracellular domain (ΔTMe(COM) = TMe(COM)_5F-MDMB-PICA – TMe(COM)_5F-MMB-PICA). (B) 5F-MDMB-PICA in the binding pocket. The highlighted transmembrane subsegments are extracellular parts of TM1, TM2, TM3, and TM7 (denoted as TM1e, TM2e, TM3e, and TM7e). The blue dashed arrows show the COM distances between TM1e–TM3e and TM2e–TM7e. (C) 2D distribution of the COM distances. 5F-MMB-PICA shows two distinct distributions for TM2e–TM7e COM distance, which seems to be the result of the changes in head moiety seen for this ligand.

Figure 7

Correlation-based network pathway. (A) CB1R structure bound with 5F-MDMB-PICA with final correlated pairs mapped on the structure. The green pairs are representative pairs for 5F-MDMB-PICA and the yellow pairs are representative for 5F-MMB-PICA. There is a noticeable difference between these two ligand-signaling pathway. The CB1R/5F-MDMB-PICA network includes a pathway consisting of eight strongly correlated pairs (solid green lines) which connect the E133^1.49–L399^7.55 (TM1i–TM7i) changes to the changes at the extracellular-middle domain, F174^2.61–D176^2.63 (TM2m–TM2e), the intracellular domain, A160^2.47–A398^7.54 (TM2i–TM7i), H143^1.59–F408^H8 (TM1i–H8), T344^6.36–L399^7.55 (TM6i–TM7i), H406^H8–R409^H8 (H8–H8), and the changes between the intracellular domain and Gα,T313^ICL3–S265^Gα (ICL3–Gα), T313^ICL3–E318^Gα (ICL3–Gα), and R400^7.56–K349^Gα (TM7i–Gα). E133^1.49–L399^7.55 (TM1i–TM7i) is at the core of these correlated pairs. There is no such pathway for CB1R/5F-MMB-PICA. (B) TM2e key residue H178^2.65 contributes to the 5F-MDMB-PICA signaling pathway. (C) The TM2i residue (F155^2.42) and TM7i NPxxY residues (P394^7.50 and Y397^7.53) are unique to the 5F-MDMB-PICA network pathway.