Molecular mechanism of the wake-promoting agent TAK-925

Abstract

The OX₂ orexin receptor (OX₂R) is a highly expressed G protein-coupled receptor (GPCR) in the brain that regulates wakefulness and circadian rhythms in humans. Antagonism of OX₂R is a proven therapeutic strategy for insomnia drugs, and agonism of OX₂R is a potentially powerful approach for narcolepsy type 1, which is characterized by the death of orexinergic neurons. Until recently, agonism of OX₂R had been considered 'undruggable.' We harness cryo-electron microscopy of OX₂R-G protein complexes to determine how the first clinically tested OX₂R agonist TAK-925 can activate OX₂R in a highly selective manner. Two structures of TAK-925-bound OX₂R with either a G_q mimetic or G_i reveal that TAK-925 binds at the same site occupied by antagonists, yet interacts with the transmembrane helices to trigger activating microswitches. Our structural and mutagenesis data show that TAK-925's selectivity is mediated by subtle differences between OX₁ and OX₂ receptor subtypes at the orthosteric pocket. Finally, differences in the polarity of interactions at the G protein binding interfaces help to rationalize OX₂R's coupling selectivity for G_q signaling. The mechanisms of TAK-925's binding, activation, and selectivity presented herein will aid in understanding the efficacy of small molecule OX₂R agonists for narcolepsy and other circadian disorders.

Fig. 1. Overall structures of the OX₂R-G protein complexes.

a Cryo-EM reconstruction of the complex OX₂R-mG_sqiN. Receptor density is in purple, TAK-925 is yellow, G_αsqiN is orange, G_β1 is green, G_γ2 is pink. The cryo-EM density from Relion was displayed in Chimera as colored surfaces (contour level 0.033, 5 sigma), where different colored zones corresponded to the different polypeptides. b Model of complex as a cartoon, with the TAK-925 agonist as spheres with yellow carbons, blue nitrogens and red oxygens. c Cryo-EM density surrounding the ligand TAK-925 in the complex structure OX₂R-mG_sqiN, with the TAK-925 agonist as sticks with colored by heteroatom (yellow carbons, blue nitrogens and red oxygens). Cryo-EM density was displayed in Pymol (4.2 sigma) within 2 Å around TAK-925. d Cryo-EM reconstruction of the complex OX₂R-G_i1. Receptor density is in purple, TAK-925 is yellow, G_αi1 is blue, G_β1 is green, G_γ2 is pink. The cryo-EM density from cryoSPARC was displayed in Chimera as colored surfaces (contour level 0.5, 5 sigma), where different colored zones corresponded to the different polypeptides. e Model of complex as a cartoon, with the TAK-925 agonist as spheres with yellow carbons, blue nitrogens and red oxygens. f Cryo-EM density surrounding the ligand TAK-925 in the complex structure OX₂R-G_i1, with the TAK-925 agonist as sticks with colored by heteroatom (yellow carbons, blue nitrogens and red oxygens). Cryo-EM density was displayed in Pymol (5 sigma) within 2 Å around TAK-925.

Fig. 2. Binding of TAK-925 to OX₂R.

a Overlay of contact residues (sticks with purple carbons for mG_sqiN-coupled OX₂R and sticks with blue carbons for G_i1-coupled OX₂R) within 4 Å of TAK-925 (yellow carbons) when superimposing OX₂R- mG_sqiN and OX₂R-G_i1. The OX₂R backbone (silver) is from OX₂R-mG_sqiN. The hydrogen bond from Gln134^3.32 to TAK-925’s sulfonamide is not shown because this residue is behind the ligand from this viewpoint (same as in Fig. 2c). b Overlay of TAK-925 (sticks with yellow carbons), Compound 1 (sticks with orange carbons) and suvorexant (sticks with cyan carbons) when superimposing the OX₂R polypeptides from the OX₂R-mG_sqiN complex (this work), the OX₂R-mini-G_sqi complex (PDB 7L1V) and the antagonist-bound inactive conformation (PDB 4S0V). c Overlay of contact residues within 4 Å of TAK-925 when superimposing OX₂R-mG_sqiN (this work, magenta), OX₂R-mini-G_sqi/Compound 1 (PDB 7L1V, orange) and the suvorexant-bound inactive conformation of OX₂R (PDB 4S0V, cyan). TAK-925 is shown as transparent spheres. d Stimulation of G_q by OX₂R wild type (WT) and mutants when bound to TAK-925 (top) and orexin B (bottom). Each data point represents an average from n ≥ 3 independent experiments (each performed in duplicate), where n is shown in Supplementary Table 2. Error bars are ±SD. Data were normalized to the WT E_max and fitted to the three-parameter model ‘log(agonist) vs response’ in GraphPad Prism 9. Source data are provided as a Source Data file.

Fig. 3. Structural basis for high selectivity of TAK-925 for OX₂R.

a Overlay of contact residues (sticks with purple carbons for mG_sqiN-coupled OX₂R and sticks with wheat carbons for inactive OX₁R) within 4 Å of TAK-925 (transparent spheres with yellow carbons) when superimposing OX₂R-G_q and the antagonist-bound inactive conformation of OX₁R (PDB 4ZJ8). Divergent residues are circled by red dotted lines. b Stimulation of G_q by OX₁R wild type (WT) and subtype-swap single and double mutants when bound to TAK-925. Each data point represents an average from n = 3 independent experiments (each performed in duplicate). Error bars are ±SD. OX₁R data were normalized to S103T/A127T E_max. Data were fitted to the three-parameter model ‘log(agonist) vs response’ in GraphPad Prism 9. Source data are provided as a Source Data file. c Stimulation of G_q by OX₂R wild type (WT) and subtype-swap single and double mutants. Each data point represents an average from n ≥ 3 independent experiments (each performed in duplicate), where n is shown in Supplementary Table 2. Error bars are ±SD. OX₂R data were normalized to OX₂R WT E_max. Data were fitted to the three-parameter model ‘log(agonist) vs response’ in GraphPad Prism 9. Source data are provided as a Source Data file.

Fig. 4. Propagated changes in OX₂R activation.

a Conformational changes of the side chains of key residues at the orthosteric site when bound to agonist, comparing the active conformation (purple sticks, transparent gray cartoon) and the inactive conformation (cyan sticks and transparent cyan cartoon). TAK-925 is shown as yellow sticks and suvorexant is in cyan sticks. b Overlay of PIF transmission switch in the active conformation of OX₂R (purple sticks and transparent spheres) and the inactive conformation of OX₂R (PDB 4S0V, cyan sticks and transparent spheres). c Rewiring of micro switches on the intracellular side of OX₂R when bound to agonist, comparing the active conformation (purple sticks, gray cartoon) and the inactive conformation (cyan sticks). TAK-925 is shown as yellow sticks. d Conformational changes of DRY motif when coupled to G protein, comparing the active conformation (gray cartoon and purple stick) and the inactive conformation (cyan sticks). The H5 helices of G_q and G_i are shown as orange and blue cartoon, respectively. Hydrogen bonds are shown as dotted lines.

Fig. 5. Interfaces of OX₂R-G protein complexes and activation of G_i and G_q.

a Interactions within 4 Å between OX₂R and mG_sqiN (gray cartoon and purple sticks for OX₂R and orange cartoon and sticks for H5 helix of G_q α subunit). Hydrogen bonds are dotted lines. b Interactions within 4 Å between OX₂R and G_i1 (gray cartoon and purple sticks for OX₂R and blue cartoon and sticks for H5 helix of G_i α subunit). Hydrogen bonds are dotted lines. c G_q signaling mediated by OX₂R with TAK-925 or orexin B. Each data point represents an average from n ≥ 3 independent experiments (each performed in duplicate), where n is shown in Supplementary Table 2. Error bars are ±SD. Data were normalized to the orexin B E_max and fitted to the three-parameter model ‘log(agonist) vs response’ in GraphPad Prism 9. Source data are provided as a Source Data file. d G_i signaling mediated by OX₂R with TAK-925 or orexin B. Each data point represents an average from 3 independent experiments (each performed in duplicate). Error bars are ±SD. Source data are provided as a Source Data file.