Signaling snapshots of a serotonin receptor activated by the prototypical psychedelic LSD

Abstract

Serotonin (5-hydroxytryptamine [5-HT]) 5-HT2-family receptors represent essential targets for lysergic acid diethylamide (LSD) and all other psychedelic drugs. Although the primary psychedelic drug effects are mediated by the 5-HT_2A serotonin receptor (HTR2A), the 5-HT_2B serotonin receptor (HTR2B) has been used as a model receptor to study the activation mechanisms of psychedelic drugs due to its high expression and similarity to HTR2A. In this study, we determined the cryo-EM structures of LSD-bound HTR2B in the transducer-free, Gq-protein-coupled, and β-arrestin-1-coupled states. These structures provide distinct signaling snapshots of LSD's action, ranging from the transducer-free, partially active state to the transducer-coupled, fully active states. Insights from this study will both provide comprehensive molecular insights into the signaling mechanisms of the prototypical psychedelic LSD and accelerate the discovery of novel psychedelic drugs.

Keywords: Gq protein; HTR2B; LSD; functional selectivity; psychedelic; signaling transduction; structural biology; β-arrestin-1.

Figure 1.. Construct optimization and structures of HTR2B in different states.

(A) The C-tail of HTR2B. Serines and threonines are highlighted by green circles. C-tail truncations C426, C453 and C464 are highlighted by red lines. See Figure S1. (B) β-arrestin-1 recruitment by different C-tail and ICL3 truncations HTR2B constructs. The remaining C-tail residues of each HTR2B truncation construct are shown in the brackets. Data represent mean ± SEM of n = 6 biological replicates. See Table S2 for fitted parameter values. (C) K247^5.68 and E319^6.30 form an ionic lock to restrict HTR2B (PDB 5TVN) in a partial active state. Hydrogen bonds are depicted as red dashed lines. (D) Breakage of K247^5.68 and E319^6.30 ionic lock by double mutations K247^5.68V+E319^6.30L greatly strengthened both the basal and LSD-stimulated β-arrestin-1 recruitment to a level comparable to the Emax of the full-agonist 5-HT. Data represent mean ± SEM of n = 4 biological replicates. See Table S3 for fitted parameter values. (E) Cartoon representation of β-arrestin-1 (PDB 4JQI) with the additional hydrophobic C loop 2 presented in the isoform 1 colored green. The sequence alignment of β-arrestins (residues 330–313) is also shown here to highlight the difference. (F-H) cryoEM maps of LSD-bound transducer-free (F), Gq-coupled (G) and β-arrestin-1 (H) coupled HTR2B. The maps are arranged in a sequential manner to highlight the signaling process of HTR2B. See Figures S1–S4 and Table S1 for details of protein expression and cryoEM data-processing.

Figure 2.. Structural comparison of the transducer-free HTR2B cryoEM structure with the HTR2B crystal structure (PDB: 5VTN).

(A) Side view to show the differences in the intracellular tips of TM5 and TM6. The relatively conformational differences observed in the HTR2B cryoEM structures are indicated by red arrows. Several TM6 residues are shown as sticks to highlight the conformational changes between the cryoEM structure and the crystal structure. (B) Extracellular view of LSD binding pocket, highlighting a loose contact of LSD with surrounding helices in the transducer-free cryoEM structure of HTR2B. The relatively outward movements of helices observed in the HTR2B cryoEM structure are indicated by red arrows. (C) Side view of LSD binding pocket, showing a larger pocket in the HTR2B cryoEM structure.

Figure 3.. Gq engagement of HTR2B.

(A) The overall structure of HTR2B in complex with agonist LSD and miniGq. HTR2B and miniGq are coloured as red and cyan, respectively. (B) The interactions between the α5 helix of Gq and the cytoplasmic cavity of HTR2B. Key residues involved in interactions are shown as sticks. Polar interactions are highlighted by red dashed lines. (C) Interactions between the ICL2 of HTR2B and Gq, showing I161^34.51 of HTR2B forms strong hydrophobic interactions with Gq. (D) Structural comparison of the HTR2B-Gq complex with HTR2A-Gq complex reveals inward displacements of both the ICL2 of HTR2B and the αN helix of Gq in HTR2B-Gq complex structure. (E) Structural comparison of the HTR2B-Gq complex with HTR2A-Gq complex reveals almost identical conformations of Gq α5 helix.

Figure 4.. β-arrestin-1 coupling of HTR2B.

(A-C) Interactions of cytoplasmic core (A), ICL2 (B) and C-tail (C) of HTR2B with β-arrestin-1, respectively. HTR2B and β-arrestin-1 are colored as green and magenta, respectively. Key residues involved in interactions are shown as sticks. Polar interactions are depicted as red dashed lines. CryoEM Map for HTR2B C-tail is shown in panel C. See Figure S5. See also Table S5 for the phosphorylation probabilities of HTR2B C-tail residues. (D) Finger loop mutations impair of β-arrestin-1 recruitment of HTR2B. Data represent mean ± SEM of n = 3 biological replicates. See Table S4 for fitted parameter values. (E) Alanine substitutions of HTR2B ICL2 residue I161^34.51 and C-tail phosphorylation residues S455, S456 and S457 impair β-arrestin-1 recruitment of HTR2B. Data represent mean ± SEM of n = 3 biological replicates. See Table S4 for fitted parameter values.

Figure 5.. Structural comparison of the Gq- and β-arrestin-1-coupled states of HTR2B with its transducer-free state.

(A) Close-up view of LSD binding pocket in different states. The HTR2B in transducer-free state, Gq-coupled state and β-arrestin-1 coupled state are colored by gray, red and green, respectively. The relative movements of LSD and residues of β-arrestin-1-coupled HTR2B related to its Gq-coupled state are indicated by red arrows. (B) Intracellular view of HTR2Bs to show conformational changes from transducer-free state to transducer-coupling states. Movements of loop and helices are indicated by red arrows. (C) Structural comparison of the engagement of Gq and β-arrestin-1 to the cytoplasmic core of HTR2B. Conformational changes in TM5, TM6 and helix 8 of β-arrestin-1-coupled HTR2B related to its Gq-coupled state are depicted as red arrows. See Figure S7. (D) Side view of HTR2Bs to show the extra downward shift of the toggle switch W337^6.48 along with a more prominent TM6 outward displacement in β-arrestin-1 coupled HTR2B. Movements of residues and helices are indicated by red arrows. See Figure S7. (E-H) Conformational changes of PIF motif (E), DRY motif (F), NPxxY motif (G) and polar core (H) upon transducer coupling. Movements of HTR2B residues upon transducer coupling are indicated by red arrows. See Figure S7.

Figure 6.. Comparison of Gq- and β-arrestin-1-coupled HTR2B.

(A) Structural comparison of the overall binding mode of Gq and β-arrestin-1 to HTR2B. see Figure S7. (B-C) Intracellular cavities of HTR2B for Gq (B) and β-arrestin-1 (C) coupling. (D-E) Interactions of Gq α5 helix (D) and β-arrestin-1 finger loop (E) with the HTR2B helix 8 residue N384^8.47, showing N384^8.47 forms a strong hydrogen-bond with Gq but not with β-arrestin-1. (F-G) miniGq and β-arrestin-1 recruitment of WT HTR2B stimulated by 5-HT (F) and LSD (G). Data represent mean ± SEM of n = 3 biological replicates. See Table S6 for fitted parameter values. (H) miniGq recruitment of WT and N384^8.47 of HTR2B stimulated by LSD. Data represent mean ± SEM of n = 3 biological replicates. See Table S6 for fitted parameter values. (I) β-arrestin-1 recruitment of WT and N384^8.47 of HTR2B stimulated by LSD. Data represent mean ± SEM of n = 3 biological replicates. See Table S6 for fitted parameter values.

Figure 7.. Activation mechanism of HTR2B.

(A) Alignment of TM5 and TM6 of HTR2B in the transducer-free, Gq-coupled and β-arrestin-1-coupled states. The red arrows show the extensions and outward movements of TM5 and TM6. (B-D) Overall structures of the transducer-free state (B), Gq-coupled state (C) and β-arrestin-1-coupled state (D) of HTR2B, highlighting the further extension and outwards movement of the intracellular tips of TM5 and TM6 upon coupling to downstream transducers. (E) In simulation, TM6 of the transducer-free receptor samples intracellular conformations matching that seen in the G protein–bound structure. Traces show the distance between the Cα carbons of ionic lock residues E319^6.30 and R153^3.50 (black dashed lines in renderings) in a representative simulation. Thick traces represent smoothed values (i.e., moving averages); transparent traces represent original, unsmoothed values. Pink and gray dashed horizontal lines indicate values for the G protein–bound and transducer-free experimental structures, respectively. (F) Receptor conformations in transducer-free simulation overlap with those from transducer-bound simulations. Pink, green, and gray dashed horizontal lines indicate values for the G protein–bound, arrestin–bound, and transducer-free cryoEM structures, respectively. (G) In simulations initiated from transducer-bound structures but with the transducer removed, the receptor relaxes to conformations similar to the transducer-free cryo-EM structure. Pink, green, and gray dashed horizontal lines indicate values for the G protein–bound, arrestin–bound, and transducer-free cryoEM structures, respectively. (H-K) The conformation of ICL2 changes from the loop in both the crystal transducer-free state (H) and cryoEM transducer-free state (J) to α-helix in Gq-coupled state (J) and β-arrestin-1-coupled state (K), positioning HTR2B ICL2 residue I161^34.51 to interact with transducers.