A Single Point Mutation Blocks the Entrance of Ligands to the Cannabinoid CB2 Receptor via the Lipid Bilayer

Abstract

Molecular dynamic (MD) simulations have become a common tool to study the pathway of ligand entry to the orthosteric binding site of G protein-coupled receptors. Here, we have combined MD simulations and site-directed mutagenesis to study the binding process of the potent JWH-133 agonist to the cannabinoid CB₂ receptor (CB₂R). In CB₂R, the N-terminus and extracellular loop 2 fold over the ligand binding pocket, blocking access to the binding cavity from the extracellular environment. We, thus, hypothesized that the binding pathway is a multistage process consisting of the hydrophobic ligand diffusing in the lipid bilayer to contact a lipid-facing vestibule, from which the ligand enters an allosteric site inside the transmembrane bundle through a tunnel formed between TMs 1 and 7 and finally moving from the allosteric to the orthosteric binding cavity. This pathway was experimentally validated by the Ala282^7.36Phe mutation that blocks the entrance of the ligand, as JWH-133 was not able to decrease the forskolin-induced cAMP levels in cells expressing the mutant receptor. This proposed ligand entry pathway defines transient binding sites that are potential cavities for the design of synthetic modulators.

Figure 1

Ligand diffusion to CB₂R. (a) Cluster of tunnels, as calculated with CAVER, from the orthosteric cavity toward the extracellular domain (blue) and toward the lipid bilayer via either TMs 2 and 3 (orange) or TMs 1 and 7 (green). The probabilities of ligand transportation (throughput) of each tunnel, calculated from the ensemble of structures collected during the aggregated 100 μs of MD simulation, are shown. (b) In three of the 100 MD simulations, the ligand diffused from the lipid bilayer to the receptor surface (10 snapshots per simulation are shown) and remained bound to either TMs 5 and 6 (one simulation in green) or TMs 1 and 7 (two simulations in orange and yellow). (c–f) Comparison of the trajectories (Figure S3) in which the ligand spontaneously bound the lipid-facing part of TMs 1 and 7 (orange in panel b, surface-bound) and a trajectory in which the ligand did not bind the receptor surface (surface-unbound). Detailed view of key amino acids located at the entrance of the TMs 1 and 7 tunnel (c). Evolution of the Phe283^7.37 side chain and the χ₁ dihedral angle along the surface-unbound (left) and surface-bound (right) trajectories (d). Representative structures obtained during the MD simulations (Figure S3) in which the cavity is closed (in white) or open (in blue). The conformations of Phe283^7.37 in panel d and TM 1 in panel f that resemble these closed or open structures are shown in white or blue, respectively. Evolution of TM 1 and the distance between the top of TMs 1 (Cα atom of Thr34^1.33) and 7 (Cα atom of Val277^7.31) along the surface-unbound (left) and surface-bound (right) trajectories (f).

Figure 2

Well-tempered metadynamics simulations of the pathway of ligand entry to the orthosteric binding site of CB₂R. (a) The ligand pathway from the initial position of JWH-133 (in orange, see Figure 1b) to the final reference position bound to the orthosteric site of CB₂R (in black), obtained during the well-tempered metadynamics, is represented by dots (the dot color gradient, from red to blue, corresponds to simulation time, from the beginning to the end, respectively). Five independent replicas are shown. (b) The free-energy profile of the five replicas (the last 10 snapshots before convergence are plotted). The collective variable (CV) is the distance between the center of mass of JWH-133 in the initial and final positions. Three energetic minima are observed when the ligand is bound to the lipid-facing part of the tunnel (green rectangle), to the bundle-facing part of the tunnel (blue), or to the orthosteric site (pink). These three positions in the ligand pathway are depicted in panel a by an ellipse with a color matching to the colors in panel b. The initial state of the binding process is on the right part of the graph and the final state on the left part. (c–e) Representative structures of JWH-133 and the interacting side chains, along the ligand pathway, at the lipid-facing part (c), the bundle-facing part (d), and the orthosteric site (e). The color of JWH-133 corresponds to the color of the minima, whereas the final reference position is shown in black.

Figure 3

A single point mutation blocks the entrance of the ligand. (a) Extracellular view of CB₂R (ECLs are omitted for clarity) in which the position of Val36^1.35 (green surface) and Ala282^7.36 (red), located between the lipid-facing (Figure 2c) and bundle-facing (Figure 2d) minima, and Phe91^2.61, Phe94^2.64, and His95^2.65 side chains in TM 2 (blue) are shown. (b) Decrease of forskolin-induced cAMP (normalized to 100%), in HEK-293T cells, upon stimulation of nonmutated (black line), Val36^1.35Phe (green) and Ala282^7.36Phe (red) mutations (mutation to Phe is shown as a solid line), and Val36^1.35 Met (green) and Ala282^7.36 Met (red) mutations (mutation to Met is shown as a dotted line) with the JWH-133 agonist. (c,d) Ligand pathways and free-energy profiles (the last 10 snapshots before convergence are plotted) of five replicas, obtained in well-tempered metadynamics of JWH-133 entry to the Val36^1.35Phe (c) or Ala282^7.36Phe (d) mutant receptors. The collective variable (CV) is the distance between the center of mass of JWH-133 in the initial (orange) and final (black) conformations. See legend of Figure 2 for additional details.

Figure 4

The binding pathway of JWH-133 to the orthosteric site of CB₂R. JWH-133 (in orange) diffuses through the membrane to reach a minimum in the free-energy profile in which JWH-133 (in green) contacts a membrane vestibule or lipid-facing cavity. Subsequently, JWH-133 (in blue) moves to a second allosteric minimum or bundle-facing cavity in the trajectory. Finally, JWH-133 (in pink) moves to the most stable minimum in the pathway at the orthosteric binding site. Ligand positions during the pathway are shown by dots. RMSD values (light gray from unbiased MD simulations, dark gray from well-tempered metadynamics) relative to the reference docked binding mode of JWH-133 at the orthosteric site and a representative free-energy profile obtained in well-tempered metadynamics are shown. Color bars indicate the different locations of the ligand. Comparison of our previously reported binding mode of the negative allosteric modulator cannabidiol (CBD, gray) and JWH-133 at the allosteric minimum or bundle-facing cavity (in blue), and the binding mode of a bitopic ligand (compound 22, gray) and JWH-133 at the vestibule or lipid-facing cavity (in green) and the orthosteric binding site (in pink).