Structure-Based Evolution of G Protein-Biased μ-Opioid Receptor Agonists

Abstract

The μ-opioid receptor (μOR) is the major target for opioid analgesics. Activation of μOR initiates signaling through G protein pathways as well as through β-arrestin recruitment. μOR agonists that are biased towards G protein signaling pathways demonstrate diminished side effects. PZM21, discovered by computational docking, is a G protein biased μOR agonist. Here we report the cryoEM structure of PZM21 bound μOR in complex with G_i protein. Structure-based evolution led to multiple PZM21 analogs with more pronounced G_i protein bias and increased lipophilicity to improve CNS penetration. Among them, FH210 shows extremely low potency and efficacy for arrestin recruitment. We further determined the cryoEM structure of FH210 bound to μOR in complex with G_i protein and confirmed its expected binding pose. The structural and pharmacological studies reveal a potential mechanism to reduce β-arrestin recruitment by the μOR, and hold promise for developing next-generation analgesics with fewer adverse effects.

Keywords: Drug Discovery; GPCRs; Medicinal Chemistry; Opioids; cryoEM Structure.

Figure 1

A) Structure of the μOR‐G_i complex bound to PZM21 colored by subunit. Green, μOR; orange, PZM21; gold, Gα_i; cyan, Gβ; purple Gγ; gray scFv16. B) View of PZM21 in the binding pocket. PZM21 forms polar contacts to D147^3.32 and Y326^7.43 and the phenol group is near H297^6.52 suggesting that PZM21 forms a water‐mediated interaction with H297^6.52, as observed in previous structures of the μOR. C) Surface presentation of the thiophene group of PZM21 in the lipophilic vestibule of the μOR.

Scheme 1

Synthesis of PZM21 analogs. Reagents and conditions: a) formaldehyde, sodium triacetoxyborohydride, water, acetonitrile, −10 °C, 5 min (>99 %); b) borane–THF complex, THF, 0 °C to reflux, 6 h (43 %), c) carboxylic acid, BOP or PyBOP, triethylamine, DMF, r.t., 0.5–24 h (56–95 %).

Figure 2

Functional activity of morphine, fentanyl, PZM21 and FH210 at the μOR. A) G‐protein signaling measured by IP accumulation assay in cells transfected with the chimeric G protein G_qi. B), C) β‐arrestin‐2 recruitment measured by PathHunter assay, with (C) and without (B) GRK2 co‐transfection. D) G_i/o/z subtypes activation measured by BRET‐based assay.

Figure 3

A) Structure of the μOR‐G_i complex bound to FH210 colored by subunit. Blue, μOR; beige, FH210; gold, Gα_i; cyan, Gβ; purple Gγ; gray scFv16. B) View of FH210 in the binding pocket. Blue, μOR; beige, FH210. FH210 forms polar contacts with D147^3.32 and Y326^7.43 and shows proximity of the phenol group to H297^6.52 suggesting that the ligand forms a water mediated interaction to H297^6.52. C) Surface representation of the naphthyl group of FH210 in the lipophilic vestibule of the μOR.

Figure 4

Comparison of μOR bound to PZM21 and FH210. A), B) Contact surface area (red) between the thiophene of PZM21 (orange) and the PZM21‐μOR cryoEM complex (green), a=124 Ų; and between the naphthyl of FH210 (beige) and the FH210‐μOR cryoEM complex (blue), a=155 Ų. Contact surface area was calculated with a cut‐off value of 2 Å using UCSF Chimera. C) Representative structures extracted from MD simulations of PZM21‐μOR and FH210‐μOR. The simulations show a more compact lipophilic vestibule for FH210 bound μOR. The distances are measured between the Cα atoms of the labeled residues. The histogram plots on the right shows the distance distribution over all simulations. The bar chart at the bottom of each plot shows the mean and SEM, with the dots representing the individual values. These statistics are based on 6 individual simulations and the first 500 ns of each simulation was not included in these analyses.