Structure-based design of subtype-selective psychedelic analogs

Huiqiong Li; Lingjie Tang; Jinfeng Zhang; Bing Meng; Dongmei Cao; Yujin Chen; Jing Yu; Huan Wang; Zhi-Jie Liu; Sheng Wang; Jianjun Cheng

doi:10.1038/s41467-026-77658-y

. 2026 Sep 17;17:9874. doi: 10.1038/s41467-026-77658-y

Structure-based design of subtype-selective psychedelic analogs

Huiqiong Li ^1,^2,^#, Lingjie Tang ^3,^#, Jinfeng Zhang ^1,^#, Bing Meng ^1,^#, Dongmei Cao ^3,^#, Yujin Chen ^1,^2,^#, Jing Yu ³, Huan Wang ¹, Zhi-Jie Liu ^1,^2,^✉, Sheng Wang ^3,^4,^5,^✉, Jianjun Cheng ^1,^2,^✉

PMCID: PMC13586271 PMID: 42754591

Abstract

Classical psychedelics exert hallucinogenic and therapeutic effects primarily through activation of serotonin 2 A receptor (5-HT_2AR), offering promise as transformative treatments for neuropsychiatric disorders. However, their concurrent activation of 5-HT_2BR—associated with cardiac valvulopathy—raises serious safety concerns, underscoring the need for subtype-selective psychedelics. To address this, we determine the cryo-EM structure of 5-HT_2AR and perform a comparative structural analysis of the orthosteric binding pockets (OBPs) of 5-HT_2AR and 5-HT_2BR. Guided by key residue differences, we develop a trigonal pharmacophore model to inform the design of 5-HT_2AR-selective agonists that avoid 5-HT_2BR activation. Using this model, we design and synthesize two compound series that selectively activate 5-HT_2AR while antagonizing 5-HT_2BR. Molecular basis of subtype selectivity is confirmed by five additional cryo-EM structures of receptor-ligand complexes. Selected compounds also exhibit antidepressant-like efficacy in animal models. Our findings provide a strategy for the development of safer, subtype-selective psychedelic analogs with therapeutic potential.

Subject terms: Small molecules, Medicinal chemistry, Cryoelectron microscopy, High-throughput screening

The authors present a rational design strategy for 5-HT_2AR-selective agonists that avoid 5-HT_2BR activation, providing lead compounds for developing safer psychedelic analogs with subtype selectivity.

Introduction

Aminergic neurotransmitters—dopamine, serotonin, adrenaline, acetylcholine, and histamine—act through multiple receptor subtypes, the majority of which belong to the G protein-coupled receptor (GPCR) superfamily^1,2. These receptor subtypes usually have different expression patterns and play differential physiological roles. Within physiological systems, the activation of specific receptor subtypes from the same family is precisely regulated through spatio-temporal neurotransmitter release. In contrast, exogenous therapeutic drugs must rely on chemical structure-based receptor subtype selectivity to minimize off-target effects and adverse reactions^1,3. However, aminergic GPCRs have high sequence conservation especially within the orthosteric binding pocket (OBP), making the discovery of subtype-selective drugs challenging⁴. To circumvent this, the extended binding pocket (EBP) is usually targeted for subtype-selectivity^5,6. The design of bitopic agents, which feature a primary pharmacophore binding at the OBP covalently bonded to a secondary pharmacophore targeting the EBP or an allosteric site, has been frequently employed for this purpose^7,8. However, this is normally achieved at the expense of significantly increasing the molecular weight of the drug, which can be detrimental to the drug-like properties of central nervous system (CNS) therapeutics^9,10. The identification of subtype-selective GPCR ligands within the OBP primarily depends on extensive high-throughput screening and comprehensive structure-activity relationship (SAR) analyses, while structure-based rational design approaches remain relatively uncommon.

Serotonergic hallucinogens, or psychedelics, have recently regained attention as potential therapeutics for neuropsychiatric diseases, including depression, anxiety, substance use disorder and others^11,12. Classical psychedelics such as psilocybin (the active ingredient of “magic mushrooms”), lysergic acid diethylamide (LSD) and N,N-dimethyltryptamine (DMT) are promiscuous drugs while the serotonin 2A receptor (5-HT_2AR) has been demonstrated as their primary target^13,14. The 5-HT_2AR belongs to the family of G protein-coupled serotonin receptors, and it shares high sequence homology with the other serotonin receptors¹⁵, especially with the other two subtypes within the 5-HT₂ subfamily, 5-HT_2BR and 5-HT_2CR⁴. The 5-HT_2CR is mainly expressed in the CNS and has been demonstrated as a drug target for treating obesity, epilepsy, schizophrenia, and substance use disorder^16–20, while its role in the overall effects of existing psychedelics remains to be fully elucidated^21,22. The 5-HT_2BR is, however, highly expressed in the cardiovascular system, and chronic activation of 5-HT_2BR has been demonstrated to be associated with cardiac valvular disease^23,24. This has led to the market withdrawal of the drug combination fenfluramine-phentermine (Fen-Phen) which was used as appetite suppressant^25,26, and the ergoline drug pergolide, a dopaminergic agonist that was used for the treatment of Parkinson’s disease²⁷. A recent comprehensive pharmacological profiling of 41 classical psychedelics revealed that all of them activate the 5-HT_2BR, making it a critical anti-target²². Therefore, subtype selectivity against the 5-HT_2BR has been suggested by the U.S. Food and Drug Administration (FDA) as a crucial criterion in developing psychedelic-based therapeutics²⁸.

Existing 5-HT_2AR agonists, including psychedelics, show varied subtype-selectivity for the receptor¹². While this is a result of medicinal chemistry efforts across several decades, recent studies mostly continue pursuing subtype-selectivity using the trial-and-error approach through SAR studies^29,30. Structural study on the psychedelic compound 25CN-NBOH revealed a side-pocket besides the OBP, which is responsible for the observed selectivity for a limited number of phenethylamine psychedelics³¹. Recent structural and computational analyses of serotonin GPCR-ligand interactions have identified “selectivity hotspot” residues that could facilitate subtype-selective drug design, in particular in the EBP¹⁵. However, targeting the OBP for rational design of subtype-selective 5-HT_2AR agonists, especially against 5-HT_2BR, has not been reported.

In this work, we perform a systematic comparative analysis of the OBP structural features distinguishing 5-HT_2AR from 5-HT_2BR. Leveraging these insights, we develop a structure-based strategy for the rational design of 5-HT_2AR-selective agonists, supported by extensive structural biology results and behavioral pharmacology.

Results

Structural insights into OBP features of 5-HT_2AR and 5-HT_2BR

Most existing psychedelics exhibit nonselective activation of both 5-HT_2AR and 5-HT_2BR, exemplified by the prototype compound LSD. Sequence analysis identified only five differential residues within the binding pocket of LSD (Fig. 1a). While we have obtained the crystal structure of LSD bound to 5-HT_2AR in a previous study³², the structure was solved in the inactive state. When this study was initiated, the active-state structure of 5-HT_2AR bound to LSD had not yet been reported. Therefore, we solved the structure of the LSD/5-HT_2AR/miniG_q complex using single-particle cryo-electron microscopy (cryo-EM), at a resolution of 3.15 Å (Fig. 1b, and Supplementary Fig. S1). The structure was obtained in complex with a mini-Gα_q-βγ heterotrimer stabilized by a single-chain variable fragment (scFv16), using identical constructs as reported recently³¹. The structure is almost identical to the reported structure (Protein Data Bank (PDB): 9AS4; with an RMSD of 0.757 Å for C-α)³³. A comparison between our LSD/5-HT_2AR/miniG_q complex and the reported LSD/5-HT_2BR/miniG_q structure shows a key difference in LSD binding pose–the N(1)-H of LSD forms a hydrogen bond with S242^5.46 (the superscript represents the Ballesteros–Weinstein numbering scheme³⁴) in 5-HT_2AR (Fig. 1c) but with the backbone carbonyl of G221^5.42 in 5-HT_2BR (Fig. 1d).

Fig. 1 — a Comparison of amino acids in the orthosteric binding pocket of 5-HT_2AR and 5-HT_2BR. Differential residues were colored in orange, and conserved residues were colored in grey. b Overall cryo-EM map of 5-HT_2AR bound to LSD in a complex with miniG_q and electron density for LSD. c, d Binding poses of the nonselective agonist LSD in 5-HT_2AR and 5-HT_2BR (PDB: 7SRR). Different hydrogen bonding interactions of LSD in 5-HT_2AR and 5-HT_2BR were marked by red circles. e Schematic representation of the trigonal design model. 3.32 was the anchor point on TM3 and 5.39 and 5.46 are the differential amino acids on TM5 in 5-HT_2AR and 5-HT_2BR. The purple circle of the small molecule model represents a nitrogen-containing ring, the green circle represents an aromatic ring, and orange circle represents an aliphatic ring. f Strategies for computational and structure-based compound design. Series A compounds were designed by conformational entropy constraint strategy (left) and series B compounds were obtained by structure-guided modification of blonanserin (right).

The highly conserved D155^3.32 (D135^3.32 in 5-HT_2BR) serves as the anchor for the basic nitrogen atom of LSD in both receptors. Key differential residues within the lower OBP involved in the recognition of LSD include I152^3.29, V235^5.39, S242^5.46 (5-HT_2AR), in comparison to L132^3.29, M218^5.39, A225^5.46 (5-HT_2BR) (Fig. 1c, d). It has been previously demonstrated that steric interaction with A225^5.46 at 5-HT_2BR, which precludes the inward movement of TM5, is responsible for the antagonist activity of the ergoline derivative methysergide³⁵. This was supported by the finding that the A225^5.46G mutant creates a spatial tolerance for the N-methyl group of methysergide, leading to potent partial agonist activity of the compound³⁵. Thus, it becomes evident that strict steric constraints at the trigonal region at the lower OBP, formed by residues 3.32, 5.46 and 5.39, could serve as a differentiator between 5-HT_2AR and 5-HT_2BR agonists (Fig. 1e). Our hypothesis is, therefore, while residue D155^3.32 serves as an anchor of monoamine ligands through a conserved salt bridge, receptor recognition at site 5.39 of additional steric substitutions to the aromatic core of ligands would push the ligands toward residue 5.46, maintaining the activation for 5-HT_2AR but leading to antagonism of 5-HT_2BR (Fig. 1e).

Structure-based design of two chemical types

Based on this hypothesis, two approaches were adopted for the rational design of chemical types (Fig. 1f). In the first approach, we used a virtual screening (VS) method combined with fragment-based drug design (FBDD) based on our 5-HT_2AR structure bound with LSD. The OBP site depicted by LSD was dissected into two sub-pockets, one mainly involving residues V235^5.39 and S242^5.46 to accommodate a hydrophobic/aromatic moiety; the other is mainly around site D155^3.32 for recognizing the amino head group (Supplementary Fig. S2a). Based on this, we screened an Enamine fragment library which includes about 3800 drug-like fragments for 5.39–5.46 sub-pocket (Supplementary Fig. S2b), and then conformational restrained grafting was used to generate compounds with drug-like properties for CNS therapeutics (molecular weight <300; tPSA <60 Å²), which were further ranked through docking strain energy calculation (Supplementary Fig. S2c). A total of 17 compounds with the general structure Series A were selected for chemical synthesis and pharmacological profiling (“Methods”; Supplementary Table S1).

The second approach involved the examination of existing drugs targeting 5-HT_2AR for which binding poses have been reported, in order to accurately analyze their interactions with the trigonal region residues. A total of 7 drugs were analyzed. Among them, the chemical scaffolds of risperidone, lumateperone, cariprazine and aripiprazole are distant from TM5 (Supplementary Fig. S3a–d) and therefore unable to prevent the inward movement of TM5, which is not in line with our proposed design principle; while zotepine, pimavanserin and blonanserin are closer to TM5 and form hydrophobic interactions (Supplementary Fig. S3e–j). Blonanserin, a 5-HT_2AR antagonist approved for the treatment of schizophrenia³⁶, caught our attention. The eight-membered aliphatic ring fused to the pyridine core of blonanserin makes extensive hydrophobic interactions with TM5 of 5-HT_2AR, with a direct contact with V235^5.39 (Fig. 1f). Therefore, we selected blonanserin for structural modification. It has been previously demonstrated that the 4-fluoro-phenyl group of blonanserin resides in the deep binding pocket (DBP) and is responsible for the antagonist activity, a conserved mechanism of many 5-HT_2AR and dopamine D2 antagonists^32,37. To transform blonanserin into agonists of 5-HT_2AR, we removed the 4-fluoro-phenyl moiety and designed compounds of Series B (“Methods”; Supplementary Table S2).

Pharmacological profiling

Compounds from both chemical series were synthesized and their structures characterized analytically (“Methods”; Supplementary Information). Their pharmacological properties were characterized by assessing binding affinity at 5-HT_2AR and 5-HT_2BR, as well as functional activity in the G_q and β-arrestin2 signaling pathways, using the bioluminescence resonance energy transfer (BRET) method (“Methods”). Remarkably, a large portion of the compounds demonstrate subtype-selective activation of 5-HT_2AR, without activating 5-HT_2BR (Supplementary Tables S1 and S2, and Supplementary Fig. S4). The pharmacological profiles of representative compounds from both series are presented (Fig. 2). Compounds of series A show more potent 5-HT_2AR activity, likely due to the presence of the phenol group which makes a presumptive hydrogen-bond with S242^5.46; while series B compounds are predicted to make only hydrophobic interactions with TM5 and the conserved salt bridge with D155^3.32.

Fig. 2 — The G_q activation and β-arrestin2 recruitment assays (BRET) were used to measure functional activities of series A (a) and series B (b) compounds at 5-HT_2AR and 5-HT_2BR. Data are represented as mean ± SEM from n = 3 independent experiments. Source data are provided as a Source Data file.

Among them, compound IHCH-1906 from series A showed potent agonist activity at 5-HT_2AR in the G_q (EC₅₀ = 12.6 nM, E_max = 85%) and β-arrestin2 recruitment (EC₅₀ = 75.9 nM, E_max = 78%) assays (Fig. 2a, and Supplementary Table S1). For 5-HT_2BR, no activation was observed in the G_q signaling pathway. Shrinking the seven-member aliphatic ring to six (IHCH-1436) or five (IHCH-1480) led to minimal changes in potency and efficacy at 5-HT_2AR, while preserving the absence of 5-HT_2BR activation (Fig. 2a, and Supplementary Table S1). For Series B, cutting off the bottom phenyl moiety of blonanserin led to IHCH-6082, a 5-HT_2AR agonist with moderate potency (G_q, EC₅₀ = 871 nM, E_max = 98%; β-arrestin2, EC₅₀ = 813 nM, E_max = 60%), and encouragingly, with antagonist activity at 5-HT_2BR (G_q, IC₅₀ = 331 nM; β-arrestin2, IC₅₀ = 1230 nM) (Fig. 2b, and Supplementary Table S2). Slightly shrinking the aliphatic ring gave improved potency (IHCH-6108, G_q, EC₅₀ = 501 nM, E_max = 105%; β-arrestin2, EC₅₀ = 355 nM, E_max = 55%) (Fig. 2b, and Supplementary Table S2). Alterations of the nitrogen position on the pyridine ring resulted in compound IHCH-6122, which showed significantly enhanced potency at 5-HT_2AR (G_q, EC₅₀ = 97.7 nM, E_max = 109%; β-arrestin2, EC₅₀ = 186 nM, E_max = 73%), while maintaining antagonist activity at 5-HT_2BR (Fig. 2b, and Supplementary Table S2). These results showed that our design strategy could lead to 5-HT_2AR agonists with subtype-selectivity against 5-HT_2BR.

Binding pose validation with cryo-EM structures

To gain molecular insights into the structural basis of our designed subtype-selective 5-HT_2AR agonists, we solved the structures of IHCH-1906 and IHCH-6122 bound to both active-state 5-HT_2AR and inactive-state 5-HT_2BR using the single-particle cryo-EM method (Fig. 3a, and Supplementary Figs. S5–S8). The active-state 5-HT_2AR structures were obtained in complex with a mini-Gα_q-βγ heterotrimer stabilized by scFv16, using identical constructs that were reported previously³¹; while the Fab-stabilized 5-HT_2BR construct was used for 5-HT_2BR structural determination³⁸. The structures were solved at global resolutions between 3.18 and 3.40 Å (Supplementary Figs. S5–S8), with the protein structures highly aligned with the reported ones^31,38. Critically, density maps of the ligands allowed the assignment of binding poses of both compounds in all four structures (Fig. 3a). Binding orientation of the ligands was verified using the Gemspot pipeline³⁹.

As expected, compound IHCH-1906 binds in the OBP of 5-HT_2AR, with the putatively protonated nitrogen atom forming a salt bridge with D155^3.32 (Fig. 3b). The benzene ring forms edge-to-face π-π interaction with F340^6.52, and the phenol group hydrogen-bonds with S242^5.46. The aliphatic ring is located in the upper part of OBP, making contacts with hydrophobic residues, in particular V235^5.39 (Fig. 3b). A remarkable difference in compound orientation was observed when IHCH-1906 binds to 5-HT_2BR. The compound rotates about 45° counterclockwise in the OBP, with its phenol group pointing toward T140^3.37 (Fig. 3c, f). The aliphatic ring, which is at the interface with TM5, seems responsible for slightly pushing TM5 outward (Fig. 3f).

Compound IHCH-6122 binds to 5-HT_2AR and 5-HT_2BR in almost identical orientations (Fig. 3d, e, g). The molecule lies horizontally in the OBP, with the piperazine nitrogen forming a salt bridge with position 3.32, and the aliphatic ring pointing towards TM5. However, the TM5 of 5-HT_2BR moves outward slightly, which may be attributed to the steric repulsion by the aliphatic ring of compound IHCH-6122 (Fig. 3g). The almost identical binding poses of IHCH-6122 in 5-HT_2AR and 5-HT_2BR highlight a striking case of functional selectivity. This observation implies that TM5 contributes differently to the activation mechanism of the two receptors, where 5-HT_2BR is more sensitive to steric repulsion on TM5, and 5-HT_2AR is more tolerant.

We also compared the binding poses of IHCH-1906 and IHCH-6122 with that of LSD within the OBP of 5-HT_2BR (Fig. 3h). Compared with LSD, the hydrophobic aliphatic rings of IHCH-1906 and IHCH-6122 push closer to TM5 in 5-HT_2BR, with the distance shortened by approximately 1–2 Å. Taken together, our structural observations are consistent with a model in which steric repulsion with TM5 contributes to the 5-HT_2BR antagonist activity of our subtype-selective 5-HT_2AR agonists.

Validation through receptor mutagenesis and SAR

To further validate the triangle model for designing subtype-selective 5-HT_2AR agonists over 5-HT_2BR, we mutated key divergent residues (positions 5.39 and 5.46) on TM5 within the OBPs. In 5-HT_2AR, these positions are occupied by Val^5.39 and Ser^5.46, whereas 5-HT_2BR contains Met^5.39 and Ala^5.46. We performed reciprocal mutations in both receptors: single mutants V235^5.39M, S242^5.46A and double mutant V235^5.39M & S242^5.46A in 5-HT_2AR; single mutants M218^5.39V, A225^5.46S and double mutant M218^5.39V & A225^5.46S in 5-HT_2BR. For comparative purposes, the same mutations (V215^5.39M, A222^5.46S, and V215^5.39M & A222^5.46S) were also generated in 5-HT_2CR. Expression levels of the single mutants were similar to wild-type, but reduced for double mutants (Supplementary Fig. S9a). For the 5-HT_2AR mutants, functional assays revealed that Series A compound IHCH-1906 exhibited markedly reduced activity, whereas the Series B compound IHCH-6122 maintained comparable efficacy as the wild type (Fig. 4a, b). In contrast, for the 5-HT_2BR mutants, both compounds demonstrated a shift from antagonism to partial activation of 5-HT_2BR, in particular for IHCH-6122 (Fig. 4c–f, Supplementary Fig. S9b, and Supplementary Tables S3 and S4). Therefore, the swap of amino acids between both receptors led to the exchange of receptor response to the compounds. These findings suggested that the 5.39 and 5.46 residues are responsible for subtype-selectivity between 5-HT_2AR and 5-HT_2BR. Mutagenesis of residues 5.39 and 5.46 in 5-HT_2CR impaired the potency/efficacy of serotonin and IHCH-6122, but slightly enhanced the partial agonist activity of IHCH-1906 (Supplementary Fig. S9b, and Supplementary Table S3).

Fig. 4 — a–d Effects of point mutations at 5-HT_2AR and 5-HT_2BR on the functional activity (as measured by BRET) of compounds IHCH-1906 and IHCH-6122. e, f Functional activity of IHCH-1906 and IHCH-6122 at position 5.39, 5.46, and 5.39/5.46 double mutants of 5-HT_2BR in antagonist mode, measured by BRET assay. g–i Effects of structural modifications on the functional activity (as measured by BRET) of IHCH-1491, IHCH-1492 and IHCH-6101. Data are represented as mean ± SEM from n = 3 independent experiments. Source data are provided as a Source Data file.

The roles of the amino acids at both positions were also verified by SAR studies with structural modifications of the compounds. Removal of the hydroxyl group of the Series A compounds, which would abolish the interaction with S242^5.46 at 5-HT_2AR, led to reduced agonist activity, as exemplified by compounds IHCH-1491 and IHCH-1492. Notably, both compounds also exhibited weak activation of 5-HT_2BR (Fig. 4g, h). Reducing the size of the aliphatic ring in Series B compounds was expected to attenuate steric repulsion against residue 5.39. Consistent with our hypothesis, the exemplar compound IHCH-6101 began to exhibit weak activation of 5-HT_2BR, suggesting that alleviating steric hindrance at position 5.39 may contribute to the gain of agonist activity at this receptor (Fig. 4i).

We then turned to the right vertex of the triangle, where the conserved D155^3.32 anchors the basic nitrogen atoms of the compounds. It was observed that changing the position of the basic nitrogen atom also affects the compound’s selectivity between 5-HT_2AR and 5-HT_2BR (Supplementary Tables S1 and S2). When the nitrogen atom was re-located from the para to the meso position on the six-membered ring, which would slightly shorten the distance from this nitrogen atom to the aromatic ring, the compounds (IHCH-1904 and IHCH-6079) started to show 5-HT_2BR agonism (Fig. 5a). For IHCH-6079, it became a more potent agonist for 5-HT_2BR than 5-HT_2AR.

Subsequently, the structure of the IHCH-1904/5-HT_2AR/miniG_q complex was also resolved (Fig. 5b, c, and Supplementary Fig. S10). The binding pattern of IHCH-1904 in 5-HT_2AR was found to be highly similar to that of IHCH-1906, with the aliphatic rings exhibiting nearly identical positioning (Fig. 5d). To further understand the molecular features of subtype selectivity, molecular dynamics simulations were then performed for three ligand/receptor complexes: IHCH-1906/5-HT_2BR, IHCH-1904/5-HT_2BR and IHCH-6122/5-HT_2BR, each for 500 ns (“Methods”). The Rg and RMSD values indicated that all simulations maintained steady states (Supplementary Fig. S11a, b). By the affinity propagation algorithm, each frame conformation of the trajectory was clustered, and the largest cluster was selected for analysis. Compared to the IHCH-1906/5-HT_2BR complex, TM5 of the IHCH-1904-bound 5-HT_2BR shifted inward by 1.14 Å, and TM6 and TM7 also shifted inward (Fig. 5e). This is consistent with the established consensus that the extracellular binding pocket shrinks upon activation in class A GPCRs⁴⁰. In addition, substantial difference was observed between the binding poses of IHCH-1904 and IHCH-1906 in 5-HT_2BR (Fig. 5f). Relocation of the basic nitrogen atom allows a rotation of the tetrahydropyridine ring around D135^3.32, leading to its contacts with TM6 (Fig. 5f). The distances between TM3 and TM5, and TM3 and TM6 in the OBP of 5-HT_2BR were analyzed for the simulations with IHCH-1904, IHCH-1906 and IHCH-6122, which demonstrated that both distances were shorter when IHCH-1904 binds to 5-HT_2BR (Fig. 5g). This is in agreement with the pharmacological result that IHCH-1904 shows agonist activity at 5-HT_2BR. Taken together, these results supported that our triangle model represents a critical molecular feature that can be used to differentiate the OBPs of 5-HT_2AR and 5-HT_2BR. Selective agonists of 5-HT_2AR can be rationally designed by first restraining the ligand core via interactions with residue 5.39, and then fine-tuning the distance between the basic amino group and the aromatic ring of the compound to match the spatial separation between positions 3.32 and 5.46.

Pharmacological profiling at other serotonin receptors

Prior to advancing the designed 5-HT_2AR agonists without 5-HT_2BR activation into in vivo studies, we further screened them against other serotonin GPCRs for their functional activity. Compounds IHCH-1906, IHCH-6082 and IHCH-6122 were selected, and the non-selective compound IHCH-1904 was also tested in parallel (Supplementary Table S5, and Supplementary Fig. S12). Compound IHCH-1906 showed the most potent agonist activity at 5-HT_2AR, antagonist activity at 5-HT_2BR, partial agonist at 5-HT_1DR (EC₅₀ = 72.4 nM, E_max = 56%), and minimal activity at most other receptors. In contrast, IHCH-1904 showed more potent agonist activity at most tested serotonin receptors except 5-HT_1AR and 5-HT_5AR. Compound IHCH-6082 showed moderate agonist activity for 5-HT_2AR but antagonism for 5-HT_2BR. However, it showed potent agonism of 5-HT_2CR (G_q, EC₅₀ = 182 nM, E_max = 126%; β-arrestin2, EC₅₀ = 66.1 nM, E_max = 61%). Compound IHCH-6122 exhibited higher potency at 5-HT_2AR as well as 5-HT_2BR antagonism, with also potent activity at 5-HT_2CR, but good selectivity against other receptors. Radioligand binding assays showed that all four compounds exhibited negligible binding affinities at dopamine receptors (Supplementary Table S6), further confirming their serotonin receptor selectivity. Taken together, compounds IHCH-1906, IHCH-6082 and IHCH-6122 displayed polypharmacological profiles similar to most existing psychedelics²². Critically, the safety-risk-associated 5-HT_2BR activation has been eliminated.

Hallucinogenic potential and antidepressant effects in animal models

Compounds IHCH-1906, IHCH-6082 and IHCH-6122 were then tested for their physicochemical properties and in vitro pharmacokinetic properties (Supplementary Table S7). All three compounds demonstrated moderate logD values and high aqueous solubility. They exhibited excellent permeability in the Caco-2 assay and good metabolic stability in human and mouse plasma and microsomes. All three compounds were then administered to male C57BL/6J mice intraperitoneally, and their concentrations in plasma, brain as well as cerebrospinal fluid were examined (“Methods”). Compound IHCH-1906 showed poor absorption and low brain permeability (Supplementary Fig. S13). This is likely due to the presence of the phenol group, which would necessitate a pro-drug modification. Compounds IHCH-6082 and IHCH-6122 demonstrated excellent brain permeability in male C57BL/6J mice and rather limited plasma exposures (Fig. 6a, and Supplementary Fig. S13). Intraperitoneal injection of IHCH-6082 and IHCH-6122 at a dose of 3 mg/kg resulted in C_max values of 3178 and 4398 ng/g in the brain, with half-lives of 1.49 and 1.35 h, respectively (Supplementary Fig. S13d). These results indicate that both compounds penetrate into the brain very efficiently.

Fig. 6 — a Pharmacokinetic analysis of IHCH-6082 and IHCH-6122. Concentration-time plot of IHCH-6082 and IHCH-6122 in male C57BL/6J mice (n = 3 samples) following i.p. dosing of 3 mg/kg. b Compound-induced head-twitch response (HTR) in male C57BL/6J mice (n = 6, 4 samples) during the first 30 min after injection (i.p.) of vehicle, LSD (0.2 mg/kg), IHCH-6082 (1, 3 or 10 mg/kg) and IHCH-6122 (0.1, 1 or 10 mg/kg). c Immobilization time in tail suspension test in male C57BL/6J mice (n = 8 samples) at 24 h after a single injection (i.p.) of saline, IHCH-6082 (3 mg/kg), IHCH-6122 (10 mg/kg), IHCH-6122 (10 mg/kg) + MDL 100907 (1 mg/kg), or MDL 100907 (1 mg/kg). d Immobilization time in tail suspension test in male C57BL/6J mice (n = 8 samples) at 24 h and 7 days after 6 consecutive days injection (i.p.) of saline, IHCH-6082 (1 mg/kg), IHCH-6082 (1 mg/kg) + MDL 100907 (1 mg/kg), or MDL 100907 (1 mg/kg). b–d Error bars represent SEM. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, and ^****p < 0.0001 (one-way Analysis of variance (ANOVA) (Dunnett’s multiple comparisons test)). Source data are provided as a Source Data file.

The favorable CNS exposures of IHCH-6082 and IHCH-6122 prompted us to test both compounds in behavioral pharmacology models. Head-twitch responses (HTRs) were first measured to demonstrate their hallucinogenic potential⁴¹. The HTR test has been widely used in psychedelic research⁴², despite its limitations as a translational marker⁴³. LSD induced robust HTRs at a dose of 0.2 mg/kg (Fig. 6b), consistent with previous studies³². Compound IHCH-6082 failed to produce HTRs at lower doses but induced significant effects at 10 mg/kg (Fig. 6b). IHCH-6122 produced significant HTRs at a dose of 1 mg/kg, while the low (0.1 mg/kg) and high (10 mg/kg) doses showed no significant effects, resulting in an inverted U-shaped (bell-shaped) dose–response curve (Fig. 6b), a characteristic often seen with other psychedelic compounds³².

We then tested the antidepressant effects of both compounds in the tail-suspension test (TST) using the corticosterone-induced depressive-like model (“Methods”). Drugs were administered intraperitoneally and the immobilization time was tested 24 h later to preclude any behavioral disturbances due to instant effects. A single dose of IHCH-6082 at 3 mg/kg showed a trend toward antidepressant efficacy, but without statistical significance. A single dose of 10 mg/kg IHCH-6122 completely reversed the depression-like behaviors, an effect that was abolished by the 5-HT_2AR antagonist MDL 100907 (Fig. 6c), indicating that the behavioral effect is mediated at least in part via 5-HT_2AR. Neither compound impaired motor coordination in the rotarod test at the doses tested, ruling out potential confounding effects on locomotor function in the TST (Supplementary Fig. S14).

To further test the potential antidepressant effects of IHCH-6082, we employed a consecutive dosing regimen at a dose of 1 mg/kg. This is one tenth of the high dose which produced significant HTR. Male C57BL/6 J mice were administered IHCH-6082 once daily for 6 days and TST was conducted 24 h or 7 days after the last dose. In this test, compound IHCH-6082 showed a significant decrease of immobilization time at both time points (Fig. 6d), indicating rapid and lasting antidepressant effects. Similarly, the antidepressant effect was diminished by the 5-HT_2AR antagonist MDL 100907. Taken together, these results show that the designed compounds are potential antidepressant agents that are worth further evaluations.

Discussion

Psychedelics hold great potential as transformative therapeutics for a variety of neuropsychiatric disorders, including depression, anxiety and substance use disorder. It is now widely acknowledged that psychedelics exert their therapeutic effects mainly through the activation of 5-HT_2AR. However, existing psychedelics tend to activate 5-HT_2BR due to its high homology with 5-HT_2AR, representing a key safety risk for the chronic use of such treatments. Through molecular comparison of the active-state conformations of 5-HT_2AR and 5-HT_2BR bound with the non-selective agonist LSD, we propose a triangle model for abolishing off-target activation of 5-HT_2BR. This model exploits a TM3-TM5 coupling strategy with amino acid residues at positions 3.32, 5.39, and 5.46 as the key points for the design of subtype-selective 5-HT_2AR agonists over 5-HT_2BR. The ionic interaction between the ligand amino group and D155^3.32 and the steric repulsion from position 5.39 could be used as a two-point anchor for the ligand, while the distance between the amino group and the aromatic core can be fine-tuned to generate compounds that activate 5-HT_2AR while antagonizing 5-HT_2BR.

The ligand binding pocket of aminergic GPCRs is a continuous channel that can be generally divided into the OBPs and EBPs. Due to the fact that the OBPs share high sequence identity across most aminergic receptors, rational design of subtype-selective GPCR ligands within the OBPs is challenging. Most drug discovery programs tend to achieve subtype selectivity through the additional targeting of the EBPs by enlarging the size of the drug molecules. However, such strategies often come at the expense of compromising the drug-like properties of compounds. In this study, we demonstrate that deep structural insights can facilitate the generation of a molecular model for the rational design of subtype-selective GPCR ligands within the OBP between highly conserved aminergic receptors.

While the subtype-selective 5-HT_2AR agonists reported here constitute promising lead compounds for further evaluation, we acknowledge several limitations that warrant future investigation. These include the need for broader off-target selectivity screening, structural characterization of agonist-bound 5-HT_2BR to better understand selectivity mechanisms, and comprehensive PK/PD profiling to assess their translational potential. Nevertheless, our work presents a design strategy for next-generation psychedelic-inspired medications that maintain therapeutic potential while eliminating the valvulopathic liabilities associated with current psychedelic agents, enabling safer chronic use.

Methods

Ethics statement

All research described in this study complies with ethical regulations. Animal behavioral tests were approved by the Institutional Animal Care and Use Committee (IACUC) of the Center for Excellence in Molecular Cell Science (CEMCS), Chinese Academy of Sciences. Officially approved protocol number: SIBCB-s375-2407-31.

Fragment-based design of series A compounds

Fragment assembly simulations were performed using the following workflow. The structure of 5-HT_2AR solved in this study (LSD-bound) was prepared using the Schrödinger 2023-1 protein preparation workflow. The pH setting in the protonation state prediction was set to 6.9 to simulate physiological conditions in the tissue fluid, while other parameters were kept at their default settings. The fragment library used in this work was obtained from the 3D-Shape and High-Fidelity fragment libraries in the Enamine commercial compound database (https://enamine.net/compound-libraries/fragment-libraries). The molecules were prepared using the LigPrep module from Schrödinger 2023-1, with a pH setting of 6.9 and default parameters. Molecular docking was performed using the Glide module⁴⁴ from Schrödinger 2023-1, with the precision mode set to standard precision (SP) docking. Strain energy correction was enabled during the docking process, and intramolecular hydrogen bonds were rewarded. When evaluating the ideal conformations of the designs, the strain energy was calculated using the Strain Energy Rescoring module⁴⁵, and a global conformational search was performed using the MCMM/Low-Mode mode^46,47.

Radioligand binding affinity

Binding assays were performed using HEK 293T (ATCC CRL-11268; mycoplasma free) membrane preparations transiently expressing receptors. HEK 293T cells were transfected using polyethylenimine (PEI, Polysciences), and membrane preparation and radioligand binding assays were performed in 96-well plates. Binding assays were conducted in 96-well plates in (50 mM Tris, 10 mM MgCl₂, 0.1 mM EDTA, 0.1% BSA, pH 7.4) standard binding buffer using [³H]-LSD (PerkinElmer) for 5-HT receptors, ³H-SCH23390 (PerkinElmer) for D₁ and D₅ receptor and ³H-N-Methylspiperone (PerkinElmer) for D₂, D₃ and D₄ receptor, respectively. For displacement experiments, membrane and radioligands (0.8–1.0 nM) were incubated with increasing compound concentrations for 2 h at room temperature in the dark. The reaction was rapidly vacuum-filtered onto chilled 0.3% polyethyleneimine-soaked GF/A filters (Filtermat A, PerkinElmer), followed by three quick cold washes with 50 mM Tris-HCl (pH 7.4) and overlaying with melted Meltilex (PerkinElmer). Counts per minute (CPM) were measured on a MicroBeta TriLux reader (PerkinElmer), and data were analyzed in GraphPad Prism 9.0 using “One-site-Fit K_i” to calculate K_i.

Bioluminescence resonance energy transfer assay (BRET)

The BRET assay was performed to characterize the pharmacological activity of compounds⁴⁸. HEK 293T cells (ATCC CRL-11268; mycoplasma free) in 6-cm dishes at 60–80% confluency were transfected using a 1:5:1 ratio of receptor-Rluc8:GFP2-β-arrestin2:GRK2 for β-arrestin2 recruitment assay, a 10:25:25:1:39 ratio of receptor:G_αq:GFP2-kras^CAAX:GRK2^RH-Rluc8:pcDNA3.1 for G_αq by Bystander assay⁴⁹, a 1:1:1:1 ratio of receptor:G_αi1-Rluc8:G_β3:G_γ9-GFP2 for G_αi1 dissociation assay, a 1:4 ratio of receptor-Rluc8:miniGs-GFP2 for miniGs recruitment assay, respectively. 5-HT_2A/2BR contained different mutations in several tests. The next day, the cells were reseeded into a white opaque bottom 96-well assay plate (Beyotime) containing DMEM (Thermo) supplemented with 2% dialyzed FBS (Omega Scientific). 18 h after reseeding, the medium was decanted and replaced with 40 μL of drug buffer (1 × HBSS, 20 mM HEPES, pH 7.4) containing 7.5 μM coelenterazine 400a (Goldbio). After 2 min of incubation, cells received 20 μL of serially diluted compounds in drug buffer. The plates were measured at 410 nm and 515 nm emission using a Tristar 3 Multimode Reader (Berthold). BRET ratios calculated as GFP2 (515 nm)/Rluc8 (410 nm) emission were analyzed in GraphPad Prism 10. Normalization used the reference ligand (5-HT) response as a divisor, with the BRET ratio at [reference ligand]₀ set to 0% and that at [reference ligand]_max to 100%.

Generation of 5-HT_2AR and 5-HT_2BR constructs for structure determination

Complexes of 5-HT_2AR in the activated state and 5-HT_2BR in the inactivated state were generated for structural determination^31,37,50. For cryo-EM, a truncated wild-type human ΔN65-5-HT_2AR-ΔC405 construct with an N-terminal FLAG tag, 10 × His tag, TEV protease site, maltose-binding protein (MBP), and HRV 3C protease site was subcloned into a pFASTBac1 vector (Invitrogen) for Sf9 cell expression. The miniG_αq-γ2 and G_β1 constructs were subcloned into the pFASTBac-Dual vector. The minG_αq, from the miniG_s/q system, had its first 31 AA replaced by the RRTL peptide from miniG_s/q and the first 30 AA of G_αi2 to facilitate scFv16 binding. Human G_γ2 was fused to miniG_αq’s N-terminus, and human G_β1’s N-terminus had a 6 × His tag and TEV protease site for purification.

The codon-optimized 5-HT_2BR expression construct, synthesized by DNA2.0, was subcloned via AscI and FseI sites into a modified pFastBac1 vector. The pFastBac vector’s expression cassette included an HA signal sequence and N-terminal FLAG tag, plus a C-terminal PreScission protease site and 10× His tag for purification. P2C2-Fab was expressed in a pRH2.2 vector for E. coli cell expression. The Fab heavy and light chains were cloned into a modified pRH2.2 vector. The expression cassette included a ST II single peptide sequence at the light chain’s and heavy chain’s N-terminus respectively, with a His-tag at the heavy chain’s C-terminus for purification.

Protein expression

The Bac-to-Bac baculovirus expression system (Invitrogen) was used to generate high-titer recombinant baculovirus (>10⁹ virus particles/ml). In a 12-well plate (Biofil®), 1 mL/well of Sf9 cells (0.5 × 10⁶ cells/mL, Expression Systems) were transfected with ~5 μg recombinant bacmid using 3 μL Cellfectin II reagent (Invitrogen). Cells were cultured with 500 μL Sf-900 II SFM medium (Invitrogen) for 12 h during transfection. The medium was changed to 1 mL ESF 921 (Expression Systems), and cells were cultured for 4 days at 27 °C to obtain P₀ virus. The P₀-containing supernatant was collected, diluted into 50 mL of 3 × 10⁶ cells/mL (Expression Systems), and incubated for 3 days to produce P₁ virus. For infection, Sf9 cells were infected with P₁ virus (multiplicity of infection = 5). 5-HT_2AR expression in Sf9 cells (3 × 10⁶ cells/mL) lasted 48 h; miniG_αq-γ2/G_β1 complex expression was 72 h at 3 × 10⁶ cells/mL; scFv16 expression was 72 h at 2 × 10⁶ cells/mL. Cells expressing 5-HT_2AR and the complex were stored at −80 °C until purification, while scFv16-containing supernatant was immediately harvested.

5-HT_2BR expression and purification were conducted according to using reported procedures^38,51. Virus production and protein expression used the Bac-to-Bac Baculovirus Expression System (Invitrogen) in Sf9 cells. Expression was carried out by infecting Sf9 cells at a density of 2–3 × 10⁶ cells/mL with P2 virus (multiplicity of infection = 5). Insect cells were harvested by centrifugation 48 h post-infection. The P2C2-Fab construct was expressed in E. coli BL21(DE3) (Weidibio) cells. Cells were transformed with a vector expressing Fab and cultured in LB medium containing 1‰ ampicillin. When cells grew to appropriated density with OD₆₀₀ = 0.6–0.8, expression was induced by 1 mM IPTG and cells were cultured at 20 °C for 20 h. Cells expressing Fab were stored at −80 °C until purification.

Purification of 5-HT_2AR/compound complex

Thawed Sf9 cells expressing 5-HT_2AR were swollen in hypotonic buffer (10 mM HEPES (pH 7.5), 10 mM MgCl₂, 20 mM KCl, 1 × protein inhibitor (PI) with 500 μM AEBSF, 1 μM E-64, 1 μM leupeptin, 0.15 μM aprotinin) and centrifuged at 47,000 × g for 15 min at 4 °C. The membrane was further purified with repeated high osmotic buffer (10 mM HEPES (pH 7.5), 10 mM MgCl₂, 20 mM KCl, 1000 mM NaCl, 1 × PI) to remove soluble and membrane-associated proteins. The purified membrane was resuspended in buffer containing 100 mM HEPES (pH 7.5), 150 mM NaCl, 10 mM MgCl₂, 20 mM KCl, 2 × PI, and 20 μM LSD or IHCH-1906 or IHCH-6122 or IHCH-1904, then incubated for 1 h at room temperature. 5-HT_2AR was extracted in solubilization buffer containing 2 mg/mL iodoacetamide (Sigma), 50 mM HEPES (pH 7.5), 150 mM NaCl, 1% (w/v) lauryl maltose neopentylglycerol (LMNG, Anatrace), 0.2% (w/v) cholesterol hemisuccinic acid (CHS, Anatrace), 10 μM LSD, IHCH-1906, IHCH-6122 or IHCH-1904, and 1 × PI at 4 °C for 2 h. After centrifugation at 90,000 × g for 30 min at 4 °C, the supernatant was incubated with 500 μL TALON IMAC resin (Clontech), an additional 20 mM imidazole, and 800 mM NaCl at 4 °C overnight. The resin was collected by centrifugation and washed with 15 columns of volume (cv) Buffer Wash I (50 mM HEPES, 800 mM NaCl, 0.1% (w/v) LMNG, 0.02% (w/v) CHS, 20 mM imidazole, 10% (v/v) glycerol, 10 μM LSD or IHCH-1906 or IHCH-6122 or IHCH-1904) and Buffer Wash II (50 mM HEPES (pH 7.5), 500 mM NaCl, 0.1% (w/v) LMNG, 0.02% (w/v) CHS, 10% (v/v) glycerol, 10 μM LSD or IHCH-1906 or IHCH-6122 or IHCH-1904). Proteins were eluted with 3cv of buffer (50 mM HEPES (pH 7.5), 500 mM NaCl, 0.04% (w/v) LMNG, 0.008% (w/v) CHS, 250 mM imidazole, 10% (v/v) glycerol, 10 μM LSD, IHCH-1906, IHCH-6122 or IHCH-1904), then concentrated to 500 μL using a 100 kDa Vivaspin 20 concentrator (Sartorius Stedim) for further purification via size-exclusion chromatography on a Superdex 200 10/300 GL column (GE Healthcare). The 5-HT_2AR-containing collections were concentrated to 1 mL with 5 EU His-tagged HRV-3C protease and 500 EU PNGase F (NEB) at 4 °C for 6 h to remove the C-terminal His tag and amino-linked glycosylation. To stabilize the complex, 10 μM LSD, IHCH-1906, IHCH-6122 or IHCH-1904 was added. The mixture was passed through equilibrated TALON IMAC resin (Clontech), and the flowthrough with cleaved protein was collected. Finally, the complex was concentrated to 20 mg/mL using a 100 kDa Vivaspin 20 concentrator (Sartorius Stedim) for 5-HT_2AR/miniG_αq-γ2/G_β1/scFv16 complex formation.

Purification of miniG_αq-γ2/G_β1 heterotrimeric complex

Thawed Sf9 cells expressing the miniG_αq-γ2/G_β1 protein complex were lysed using a buffer containing 20 mM HEPES (pH 7.5), 100 mM NaCl, 1 mM MgCl₂, 30 mM imidazole, 5 mM β-mercaptoethanol, 0.2% (w/v) Triton X-100, and 1 × PI. The lysate was then centrifuged at 117,500 × g for 50 min at 4 °C. Following centrifugation, the supernatant was incubated with 500 μl of Ni NTA beads 6FF (Smart-Lifesciences) for 2 h at 4 °C. The beads were washed with 10 cv of Buffer I (20 mM HEPES (pH 7.5), 100 mM NaCl, 5 mM β-mercaptoethanol, 30 mM imidazole, 1 × PI). Proteins were eluted using 3 cv of elution buffer (20 mM HEPES, 100 mM NaCl, 5 mM β-mercaptoethanol, 300 mM imidazole, 1 × PI), then concentrated to 500 μl via a 10 kDa Vivaspin 20 concentrator (Sartorius Stedim) for further purification by size-exclusion chromatography on a Superdex 75 10/300 GL column (GE Healthcare). The collected protein was further concentrated to ~10 mg/ml for 5-HT_2AR/miniG_αq-γ2/G_β1 complex formation.

Purification of scFv16

The scFv16-containing supernatant was supplemented with 20 mM HEPES (pH 7.5), 100 mM NaCl, 1 × PI, and 500 μL Ni NTA beads 6FF (Smart-Lifesciences) and incubated overnight at 4 °C. The beads were then washed with 10 cv of Buffer I (20 mM HEPES, 100 mM NaCl, 30 mM imidazole, 1 × PI) and 10 cv of Buffer II (20 mM HEPES, 100 mM NaCl, 1 × PI). Proteins were eluted using 3 cv of elution buffer (20 mM HEPES, 100 mM NaCl, 300 mM imidazole, 1 × PI). Further purification was done via size-exclusion chromatography on a Superdex 75 10/300 GL column (GE Healthcare), and the purified protein was concentrated to ~5 mg/ml for 5-HT_2AR/miniG_αq-γ2/G_β1/scFv16 complex formation.

Formation of 5-HT_2AR/compound/miniG_αq-γ2/G_β1/scFv16 complex

The purified 5-HT_2AR/compound (LSD or IHCH-1906 or IHCH-6122 or IHCH-1904) complex was mixed with 1.2-molar-excess miniG_αq-γ2/G_β1 and scFv16, incubated at 4 °C overnight, and further purified via size-exclusion chromatography on a Superdex 200 10/300 GL column (GE Healthcare) using buffer with 20 mM HEPES (pH 7.5), 100 mM NaCl, 0.00075% (w/v) LMNG, 0.000075% (w/v) CHS, 0.00025% (w/v) Glyco-Diosgenin (GDN), and 5 μM compound (LSD or IHCH-1906 or IHCH-6122 or IHCH-1904). The purified complex was concentrated to 5 mg/mL for cryo-EM structure resolution.

Purification of 5-HT_2BR/compound complex

Insect cells were thawed in hypotonic buffer and isolated membranes were extensively washed with hypertonic buffer (1 M NaCl) to remove membrane-associated proteins. Purified membranes were resuspended in buffer (10 mM HEPES, pH 7.5, 10 mM MgCl₂, 20 mM KCl, 150 mM NaCl, 10 μM IHCH-1906 or IHCH-6122, 2 mg/mL iodoacetamide, EDTA-free protease inhibitors (Roche)), incubated at room temperature for 1 h, then solubilized in buffer (10 mM HEPES, pH 7.5, 150 mM NaCl, 1% (w/v) n-dodecyl-β-d-maltopyranoside (DDM, Anatrace), 0.2% (w/v) CHS, 10 μM IHCH-1906 or IHCH-6122, protease inhibitors) for 2 h at 4 °C. After ultracentrifugation to remove cell debris, protein was bound to TALON IMAC resin (Clontech) overnight at 4 °C with 20 mM imidazole and 800 mM NaCl. The resin was washed with 10 cv Wash Buffer I (50 mM HEPES, pH 7.5, 800 mM NaCl, 0.1% (w/v) DDM, 0.02% (w/v) CHS, 20 mM imidazole, 10% (v/v) glycerol, 10 μM IHCH-1906 or IHCH-6122) and 5 cv Wash Buffer II (50 mM HEPES, pH 7.5, 150 mM NaCl, 0.05% (w/v) DDM, 0.01% (w/v) CHS, 10% (v/v) glycerol, 10 μM IHCH-1906 or IHCH-6122). Protein was eluted in 5 cv of Wash Buffer II + 250 mM imidazole, concentrated to 0.5 mL, and desalted with PD MiniTrap G-25 columns (GE Healthcare) to remove imidazole. His-tagged PreScission protease was added to remove the C-terminal 10× His-tag overnight, followed by reverse IMAC to eliminate protease, cleaved His-tags, and uncleaved protein. The protein was concentrated to ~5 mg/ml for 5-HT_2BR/Fab complex formation.

Purification of P2C2–Fab

His-tagged Fab was released from cells by sonic and collected via incubating with 500 μL Ni NTA beads 6FF (Smart-Lifesciences) for 3 h at 4 °C. After washing the resin with 20 mM Tris, 100 mM NaCl, 10 mM imidazole (pH 7.5), Fab was eluted using the same buffer with 250 mM imidazole, and further purified on a Superdex75 column. Fab-containing fractions were pooled and concentrated with a 30 kDa cutoff concentrator for 5-HT_2BR/Fab complex formation.

Formation of 5-HT_2BR/compound/Fab complex

Co-incubate purified 5-HT_2BR and P2C2 Fab at 4 °C for 2 h before subjected to size-exclusion chromatography on a Superdex 200 Increase 10/300 column (GE Healthcare) that was pre-equilibrated with 20 mM HEPES pH 7.5, 100 mM NaCl, 1 μM IHCH-1906 or IHCH-6122, 0.001% (w/v) LMNG, 0.00033% (w/v) GDN and 0.0001% (w/v) CHS. Peak fractions were then collected and concentrated to ~5 mg/ml to make the cryo-EM grids.

Sample preparation, Cryo-EM data collection and processing

A volume of 3.5 μL of purified protein complex at a concentration of ~5 mg/mL was applied onto glow-discharged (40 s) EM grids (Quantifoil R1.2/1.3) under 100% humidity at 8 °C using a Vitrobot Mark IV (Thermo Fisher Scientific). Each sample was left to stand for 20 s, blotted for 3 s, and subsequently plunge-frozen into liquid ethane and transferred into liquid nitrogen for storage.

All movies were collected on a Titan Krios G3 300 kV electron microscope (Thermo Fisher Scientific) with a Gatan K3 Summit direct electron camera and a Gatan quantum energy filter at a magnification of ×105,000 (0.832 Å px⁻¹ or 0.828 Å px⁻¹) and the defocus ranged from −0.8 to −1.5 μm in super-resolution mode. Each movie comprised 40 frames with a total dose of 60 e^- Å⁻², and the exposure time was 2 s with a dose rate of 20 e^- pixel⁻¹ s⁻¹ using SerialEM software⁵² (Supplementary Table S8).

Cryo-EM data were manually curated, inspected, and processed using CryoSPARC (v.4.5.3)⁵³. All movies were processed for beam-induced motion correction using patch motion correction⁵⁴ and fitted locally-variable CTF landscapes using patch CTF estimation⁵⁵. Particles were selected using the blob picker and the template picker. Multiple 2D and 3D classification rounds were performed; subsequently, the best 3D class was further applied for final homogenous refinement, non-uniform refinement and local refinement, generating a density map with an indicated global resolution. The half-maps were alternatively sharpened using deepEMhancer to assist in visualizing the maps. And the sharpened maps output from the cryoSPARC local refinement job were deposited and used for model building and subsequent structural analysis.

Model building and refinement

The initial model of each compound/5-HT_2AR/miniG_q complex was derived from PDB: 8UWL⁵⁶, and the initial model of each compound/5-HT_2BR/Fab complex was derived from PDB: 5TUD³⁸. The initial model was fitted into the EM density map using ChimeraX (v.1.7)⁵⁷. The complex starting model was subjected to iterative rounds of manual adjustment and rebuilding in Coot (v.0.9.8.93)⁵⁸ and refined by real-space refinement in Phenix (v.1.20.1)⁵⁹. The binding conformation of the ligands was verified using the GlideEM module in Schrödinger 2023-1. Model statistics were validated using Molprobity, and final model statistics were provided in Supplementary Table S8. The structure figures were generated using ChimeraX⁵⁷ and Pymol⁶⁰.

Molecular dynamics simulation and analysis

In order to explore the effects of IHCH-1906, IHCH-1904 and IHCH-6122 on the conformation of 5-HT_2AR and 5-HT_2BR, the initial complex structures of the compounds with 5-HT_2AR and 5-HT_2BR were constructed. To obtain a reasonable protein structure in the apo state, we used AlphaFold3 webserver⁶¹, Schrödinger’s homology modeling⁶² and protein structure preparation modules^63,64 to fill in the missing loops and build the corresponding models. The membrane positions of the 5-HT_2AR and 5-HT_2BR models were assigned according to the OPM structure⁶⁵ (PDB code: 7RAN⁶⁶ and 7SRR⁶⁷) of the cell membrane constructed in the molecular dynamics simulation system, and the POPC bilayer membrane was assembled onto the corresponding protein with reference to the assigned membrane position. Subsequently, the Desmond program was employed through the AutoMD (https://github.com/Wang-Lin-boop/AutoMD, https://doi.org/10.5281/zenodo.21715805) script to perform molecular dynamics simulations under the S-OPLS force field. The complexed structures were explicitly solvated with SPC water molecules under cubic periodic boundary conditions for 15 Å buffer region. The overlapping water molecules were deleted, 0.15 M NaCl was added, and the systems were neutralized by adding Na⁺ as counter ions. For each initial structure, four independent simulations of 500 ns each were performed using different random velocity seeds, and the resulting trajectories were merged for subsequent analysis. Brownian motion simulation was used to relax these systems into local energy minimum states separately. An NPT ensemble was then applied to maintain the constant temperature (310 K) and pressure (1.01325 bar) of the systems.

After the product molecular dynamics simulations, the AutoTRJ script⁶⁸ (available along with AutoMD) was used for automated trajectory merging, periodic boundary treatment, structural clustering, as well as RMSF, RMSD, radius of gyration (Rg), and distance calculations. The affinity propagation algorithm⁶⁹ was used to cluster the representative conformations, and the largest cluster was retained as the representative conformation for further analysis. The Maestro graphical interface (Schrödinger LLC.) was used for trajectory visualization, and the Cα RMSD of the binding site (for 5-HT_2AR and 5-HT_2BR) was used to analyze the conformational stability during the simulations.

ADME assays

Caco-2 permeability

A-to-B and B-to-A transport were performed using HBSS buffer (pH 7.4) with varied DMSO and lucifer yellow (LY) concentrations as donor/receiver media. Test compounds were incubated at 10 μM on Caco-2 cell monolayers, with erythromycin, metoprolol and atenolol as reference controls. After 90 min incubation at 37 °C, samples were detected by fluorometry (λ_ex = 485 nm, λ_em = 535 nm) and LC-MS/MS. Apparent permeability coefficients (P_app), efflux ratio and mass recovery were calculated from bidirectional transport data. All tests were performed in duplicate.

Liver microsomal stability

Metabolic stability in human and mouse liver microsomes was evaluated with NADPH as a cofactor in 0.1 M pH 7.4 potassium phosphate buffer containing 1.0 mM EDTA. DMSO stock solutions of test compounds were diluted to working concentrations, and incubation systems with microsomes were pre-equilibrated on ice. Reactions were initiated by NADPH addition (final concentrations: microsomes 0.5 mg/mL, substrate 1.0 μM, NADPH 1.0 mM) and incubated at 37 °C for 0, 5, 15, 30, 45 min. Reactions were terminated by acetonitrile with an internal standard. After shaking and centrifugation, supernatants were analyzed by LC-MS/MS. Based on linear regression of ln(remaining percentage) vs. time, half-life (t_1/2) and intrinsic clearance (Cl_int) were calculated.

Plasma stability

Non-microsomal degradation stability was determined in human and mouse plasma (pH 7.4–8.0). Compound stock solutions were sequentially diluted to a working spiking solution with a buffer containing 0.5% BSA. The reaction was started by mixing the spiking solution with pre-warmed plasma (final substrate concentration = 2 μM) and incubated at 37 °C. Sampling was conducted at 0, 5, 15, 30, 45, 60 min, followed by quenching with internal standard-containing acetonitrile. Post-centrifugation, supernatants were detected by LC-MS/MS. t_1/2 was derived from linear fitting of the logarithmic residual fraction over incubation time.

Octanol-buffer logD

Lipophilic distribution at pH 7.4 was measured via octanol/potassium phosphate partition assay. Compound stock solution, octanol and pH 7.4 buffer were mixed, phase-equilibrated under shaking at 25 °C and centrifuged. Aqueous and octanol phases were separately diluted serially, then analyzed by LC-MS. The logD7.4 value was calculated as the logarithm of the concentration ratio of analyte in octanol to aqueous buffer.

Kinetic solubility

The shake-flask method was adopted to determine kinetic solubility in 0.1 M pH 7.4 sodium phosphate buffer. DMSO stock of test compound was added to buffer (final DMSO = 1%, initial concentration = 100 μM) and shaken at room temperature. Undissolved particles were removed by centrifugation. Supernatants with serial dilutions and calibration standards were analyzed via LC-MS/MS. Kinetic solubility (μM) was quantified according to the calibration curve.

Pharmacokinetics

The pharmacokinetics studies were performed by Suzhou Kangrun Pharmaceutical Testing Service, Inc. (Suzhou, China). Studies were performed with approved mouse protocols from the institutional animal care and use committees. Male C57BL/6J mice (6–8 weeks old, ∼25 g body weight) were purchased from JOINN Laboratories, Inc. (Suzhou). IHCH-1906, IHCH-6122 and IHCH-6082 were dissolved in 5% DMSO and 95% saline and administered at 3 mg/kg (i.p.); IHCH-6122 and IHCH-6082 were dissolved in 10% DMSO and 90% saline and administered 1 mg/kg (i.v.) and dissolved in 5% DMSO, 20% PEG400 and 75% saline and administered 10 mg/kg (p.o.), with nine animals in each group. Blood samples (0.1 mL) were collected from the mouse orbit at 0, 0.5, 2 and 6 h, and were then centrifuged at 5000 r.p.m. at 4 °C for 10 min to collect plasma samples. Cerebrospinal fluids (CSF) was collected at 0.5, 2 and 6 h. Brain tissues were collected at 0.5, 2 and 6 h, homogenized in 50% methanol [brain weight (g)/50% methanol (mL), 1:4] to obtain brain tissue solutions. All samples were stored at −80 °C for analysis. Drug concentrations in the samples were determined using liquid chromatography–mass spectrometry.

Mice

All animal behavioral experiments were conducted in accordance with the guidelines approved by the IACUC of the Center for Excellence in Molecular Cell Science (CEMCS), Chinese Academy of Sciences. Mice were housed in a certified pathogen-free (SPF) facility with controlled temperature (23 ± 1 °C), humidity (45–55%), and a 12/12 h light/dark cycle. Only male mice were used in this study to minimize variability introduced by the estrous cycle and to allow direct comparison with existing pharmacological and behavioral data in the literature.

Head twitch response (HTR) detection system based on magnetic signals

Magnetometer-based detection of HTR in male C57BL/6J mice (10–12 weeks old) was performed³². After a rostrocaudal incision exposed the cranium, a custom neodymium magnet (4 × 4 × 2 mm) was implanted with dental resin, keeping its N-S poles parallel to the dorsoventral plane. Mice recovered for 1 week pre-test. For head movement recording, mice were placed in a 12 cm-diameter glass beaker surrounded by 150 turns of #30 enameled copper wire. Coil voltage was recorded by PowerLab/4SP with LabChart V8.1.16, amplified, low-pass filtered at 10 kHz, and sampled at 40 kHz. LabChart data were processed with a 40–200 Hz digital band-pass filter for HTR detection, which was identified by sinusoidal wavelets meeting: (1) >2 bipolar peaks; (2) amplitude exceeding background noise; (3) duration <120 ms.

Corticosterone-induced depression-like model

Male C57BL/6J mice (10–12 weeks old) were administered fresh-prepared 25 µg/mL corticosterone in drinking water, replenished every 48 h for 14 days to induce depression-like symptoms. Starting at the eight water change, corticosterone was tapered over a week: 12.5, 6.25, then 0 µg/mL. Three days after the final (0 µg/mL) water change, the model was complete.

Tail suspension test (TST)

In a quiet room, each mouse’s tail was suspended on a force transducer which connected to Powerlab/4SP installed with LabChart v 8.1.16 to record the change in tension. Data were sampled at 100 Hz with a digital band-pass filter at 5 Hz. The recording lasted 6 min, after which male C57BL/6J mice (10–12 weeks old) were returned to their cages. Immobility time in the last 4 min was identified by searching for a sinusoidal wavelet period that meets: (1) Peak amplitude was less than 1% of animal body weight; (2) The wavelet period was continued for at least 2 s. This immobility time was used to measure depression-like behavior.

Rotarod test

Male C57BL/6J mice (10–12 weeks old) were placed in separate lanes on the rod rotating at 15 RPM immediately after i.p. injected with compounds or saline. Duration of the test is 10 min. The number of falls was counted.

Quantification and statistical analysis

All statistical analyses were done in GraphPad Prism 10.0. Dose–response data were fit using log(agonist) vs. response function with hill slope fixed at 1, and E_max/EC₅₀ were estimated. Data were normalized to reference ligands (serotonin for 5-HT_2A/2BR) and averaged across experiments. Using GraphPad Prism 10.0, reference ligand dose-response data were fit with the Black and Leff operational model⁷⁰.

Mouse treatments were randomized, and data were analyzed by experimenters blinded to treatment conditions. GraphPad Prism 10.0 was used for statistical analyzes. Sample sizes, similar to prior studies⁵⁰, were not determined by statistical methods. Data were assumed to follow a normal distribution without formal testing. Behavioral data were analyzed via one-way Analysis of variance (ANOVA) (Dunnett’s multiple comparisons test). All comparisons were pre-planned. Data are shown as mean ± SEM (standard error of the mean), with asterisks indicating significance levels (^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001 and ^∗∗∗∗P < 0.0001).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(8.5MB, pdf)}

Reporting Summary^{(105KB, pdf)}

Transparent Peer Review file^{(13.2MB, pdf)}

Source data

Source Data^{(7.5MB, zip)}

Acknowledgements

The cryo-EM data were collected at the Bio-Electron Microscopy Facility of ShanghaiTech University. We thank all staff at the facility for their assistance in cryo-EM data collection. We are grateful to Dr. Lingyun Yang at the NMR core facility for assistance in obtaining NMR data of all compounds.

Author contributions

J.C. conceived the project. J.C., S.W., and Z.-J.L. supervised the project. H.L., J.Z., and L.T. conducted most of the experiments unless specified otherwise; B.M. and Y.C. participated in cryo-EM data collection and structural determination; D.C. participated in protein preparation for structural determination; Y.C. and H.W. assisted in compound synthesis; J.Y. participated in animal behavior tests. J.C. and H.L. wrote the manuscript with input from all authors.

Peer review

Peer review information

Nature Communications thanks Marcin Kołaczkowski, Ellen Walker, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work was supported by grants from the National Natural Science Foundation of China (grant no. 82525061, 22377077 to J.C.; 82225045, 82530106 to S.W.); the Program of Shanghai Academic/Technology Research Leader (22XD1421900 to J.C.); Strategic Priority Research Program of the Chinese Academy of Sciences grant (XDB0990000 to S.W.); the Ministry of Science and Technology of China (2020YFA0509600 to S.W.); and the Shanghai Municipal Government and ShanghaiTech University (to J.C.).

Data availability

Source data are provided with this paper. The atomic coordinates and the cryo-EM density maps have been deposited in the Protein Data Bank (PDB): 9WPQ https://doi.org/10.2210/pdb9wpq/pdb (5-HT_2AR/LSD/miniGq); 9WQ6 https://doi.org/10.2210/pdb9wg6/pdb (5-HT_2AR/IHCH-1906); 9WQ5 https://doi.org/10.2210/pdb9wq5/pdb (5-HT_2AR/IHCH-6122/miniGq); 9WPX https://doi.org/10.2210/pdb9wpx/pdb (5-HT_2AR/IHCH-1904/miniGq); 9WQA https://doi.org/10.2210/pdb9wqa/pdb (5-HT_2BR/IHCH-1906/Fab); 9WQB https://doi.org/10.2210/pdb9wqb/pdb (5-HT_2BR/IHCH-6122/Fab); and the Electron Microscopy Data Bank (EMDB) EMD-66146 (5-HT_2AR/LSD/miniGq); EMD-66160 (5-HT_2AR/IHCH-1906); EMD-66159 (5-HT_2AR/IHCH-6122/miniGq); EMD-66152 (5-HT_2AR/IHCH-1904/miniGq); EMD-66164 (5-HT_2BR/IHCH-1906/Fab); EMD-66165 (5-HT_2BR/IHCH-6122/Fab). Source data are provided with this paper.

Competing interests

J.C., S.W., H.L., J.Z. and L.T. have filed patents (PCT/CN2025/132646; PCT/CN2025/076535) through their institutions for the compounds reported in this study. The remaining authors (B.M., D.C., Y.C., J.Y., H.W., and Z.-J.L.) declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Huiqiong Li, Lingjie Tang, Jinfeng Zhang, Bing Meng, Dongmei Cao, Yujin Chen.

Contributor Information

Zhi-Jie Liu, Email: liuzhj@shanghaitech.edu.cn.

Sheng Wang, Email: wangsheng@sibcb.ac.cn.

Jianjun Cheng, Email: chengjj@shanghaitech.edu.cn.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77658-y.

References

1.Hauser, A. S., Attwood, M. M., Rask-Andersen, M., Schioth, H. B. & Gloriam, D. E. Trends in GPCR drug discovery: new agents, targets and indications. Nat. Rev. Drug Discov.16, 829–842 10.1038/nrd.2017.178 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Vass, M. et al. Aminergic GPCR-ligand interactions: a chemical and structural map of receptor mutation data. J. Med. Chem.62, 3784–3839 10.1021/acs.jmedchem.8b00836 (2019). [DOI] [PubMed] [Google Scholar]
3.Lorente, J. S. et al. GPCR drug discovery: new agents, targets and indications. Nat. Rev. Drug Discov.24, 458–479 10.1038/s41573-025-01139-y (2025). [DOI] [PubMed] [Google Scholar]
4.Peng, Y. et al. 5-HT_2C receptor structures reveal the structural basis of GPCR polypharmacology. Cell172, 719–730 10.1016/j.cell.2018.01.001 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Fan, L. Y. et al. Haloperidol bound D2 dopamine receptor structure inspired the discovery of subtype selective ligands. Nat. Commun.11. 10.1038/s41467-020-14884-y. (2020). [DOI] [PMC free article] [PubMed]
6.Wang, S. et al. D4 dopamine receptor high-resolution structures enable the discovery of selective agonists. Science358, 381–386 10.1126/science.aan5468 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Arroyo-Urea, S. et al. A bitopic agonist bound to the dopamine 3 receptor reveals a selectivity site. Nat Commun15. 10.1038/s41467-024-51993-4. (2024). [DOI] [PMC free article] [PubMed]
8.Valant, C. et al. Separation of on-target efficacy from adverse effects through rational design of a bitopic adenosine receptor agonist. Proc. Natl. Acad. Sci. USA111, 4614–4619 10.1073/pnas.1320962111 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Rankovic, Z. CNS drug design: balancing physicochemical properties for optimal brain exposure. J. Med. Chem.58, 2584–2608 10.1021/jm501535r (2015). [DOI] [PubMed] [Google Scholar]
10.Ghose, A. K., Herbertz, T., Hudkins, R. L., Dorsey, B. D. & Mallamo, J. P. Knowledge-based, central nervous system (CNS) lead selection and lead optimization for CNS drug discovery. ACS Chem. Neurosci.3, 50–68 10.1021/cn200100h (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.McClure-Begley, T. D. & Roth, B. L. The promises and perils of psychedelic pharmacology for psychiatry. Nat. Rev. Drug Discov.21, 463–473 10.1038/s41573-022-00421-7 (2022). [DOI] [PubMed] [Google Scholar]
12.Duan, W., Cao, D. M., Wang, S. & Cheng, J. Serotonin 2A receptor (5-HT_2AR) agonists: psychedelics and non-hallucinogenic analogues as emerging antidepressants. Chem. Rev.124, 124–163 10.1021/acs.chemrev.3c00375 (2024). [DOI] [PubMed] [Google Scholar]
13.Slocum, S. T., DiBerto, J. F. & Roth, B. L. Molecular insights into psychedelic drug action. J. Neurochem.162, 24–38 10.1111/jnc.15540 (2022). [DOI] [PubMed] [Google Scholar]
14.Gumpper, R. H. & Roth, B. L. SnapShot: psychedelics and serotonin receptor signaling. Cell186, 232–232 10.1016/j.cell.2022.12.020 (2023). [DOI] [PubMed] [Google Scholar]
15.Simon, I. A. et al. Ligand selectivity hotspots in serotonin GPCRs. Trends Pharm. Sci.44, 978–990 10.1016/j.tips.2023.09.012 (2023). [DOI] [PubMed] [Google Scholar]
16.Higgins, G. A., Fletcher, P. J. & Shanahan, W. R. Lorcaserin: a review of its preclinical and clinical pharmacology and therapeutic potential. Pharm. Ther.205, 107417. 10.1016/j.pharmthera.2019.107417 (2020). [DOI] [PubMed] [Google Scholar]
17.He, Y. et al. Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior. Nat. Med.28, 2537–2546 10.1038/s41591-022-02106-5 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Dell’isola, G. B. et al. Evaluating bexicaserin for the treatment of developmental epileptic encephalopathies. Expert Opin. Pharmacother.25, 1121–1130 10.1080/14656566.2024.2373350 (2024). [DOI] [PubMed] [Google Scholar]
19.Pogorelov, V. M. et al. 5-HT_2C agonists modulate schizophrenia-like behaviors in mice. Neuropsychopharmacology42, 2163–2177 10.1038/npp.2017.52 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Cheng, J. & Kozikowski, A. P. We need 2C but not 2B: developing serotonin 2C (5-HT_2C) receptor agonists for the treatment of CNS disorders. ChemMedChem10, 1963–1967 10.1002/cmdc.201500437 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Canal, C. E. & Murnane, K. S. The serotonin 5-HT_2C receptor and the non-addictive nature of classic hallucinogens. J. Psychopharmacol.31, 127–143 10.1177/0269881116677104 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Jain, M. K. et al. The polypharmacology of psychedelics reveals multiple targets for potential therapeutics. Neuron 10.1016/j.neuron.2025.06.012 (2025). [DOI] [PubMed] [Google Scholar]
23.Roth, B. L. Focus on research-drug and valvular heart disease. N. Engl. J. Med356, 6–9 10.1056/NEJMp068265 (2007). [DOI] [PubMed] [Google Scholar]
24.Dumotier, B. M. & Urban, L. Preclinical mitigation of 5-HT_2B agonism-related cardiac valvulopathy revisited. J. Pharmacol. Tox. Met.128. 10.1016/j.vascn.2024.107542. (2024). [DOI] [PubMed]
25.Connolly, H. M. et al. Valvular heart disease associated with fenfluramine-phentermine. N. Engl. J. Med.337, 581–588 10.1056/NEJM199708283370901 (1997). [DOI] [PubMed] [Google Scholar]
26.Jick, H. et al. A population-based study of appetite-suppressant drugs and the risk of cardiac-valve regurgitation. N. Engl. J. Med.339, 719–724 10.1056/NEJM199809103391102 (1998). [DOI] [PubMed] [Google Scholar]
27.Schade, R., Andersohn, F., Suissa, S., Haverkamp, W. & Garbe, E. Dopamine agonists and the risk of cardiac-valve regurgitation. N. Engl. J. Med.356, 29–38 10.1056/NEJMoa062222 (2007). [DOI] [PubMed] [Google Scholar]
28.US Food and Drug Administration. Psychedelic Drugs: Considerations for Clinical Investigations. Guidance for Industry. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/psychedelic-drugs-considerations-clinical-investigations (2026).
29.Rorsted, E. M., Jensen, A. A., Smits, G., Frydenvang, K. & Kristensen, J. L. Discovery and structure-activity relationships of 2,5-dimethoxyphenylpiperidines as selective serotonin 5-HT2A receptor agonists. J. Med Chem.67, 7224–7244 10.1021/acs.jmedchem.4c00082 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Varty, G. B. et al. Synthesis and structure-activity relationships of 2,5-dimethoxy-4-substituted phenethylamines and the discovery of CYB210010: a potent, orally bioavailable and long-acting serotonin 5-HT2 receptor agonist. J. Med. Chem.67, 6144–6188 10.1021/acs.jmedchem.3c01961 (2024). [DOI] [PubMed] [Google Scholar]
31.Kim, K. et al. Structure of a hallucinogen-activated Gq-coupled 5-HT_2A serotonin receptor. Cell182, 1574–1588 10.1016/j.cell.2020.08.024 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Cao, D. M. et al. Structure-based discovery of nonhallucinogenic psychedelic analogs. Science375, 403–411 10.1126/science.abl8615 (2022). [DOI] [PubMed] [Google Scholar]
33.Gumpper, R. H. et al. The structural diversity of psychedelic drug actions revealed. Nat. Commun.16 10.1038/s41467-025-57956-7. (2025). [DOI] [PMC free article] [PubMed]
34.Ballesteros, J. A. & Weinstein, H. Integrated methods for the construction of three-dimensional models and computational probing of structure-function relations in G protein-coupled receptors. Methods Neurosci.25, 366–428 10.1016/S1043-9471(05)80049-7 (1995). [DOI]
35.McCorvy, J. D. et al. Structural determinants of 5-HT_2B receptor activation and biased agonism. Nat. Struct. Mol. Biol.25, 787–796 10.1038/s41594-018-0116-7 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Tenjin, T. et al. Profile of blonanserin for the treatment of schizophrenia. Neuropsych. Dis. Treat.9, 587–594 10.2147/Ndt.S34433 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Chen, Z. C. et al. Flexible scaffold-based cheminformatics approach for polypharmacological drug design. Cell187 10.1016/j.cell.2024.02.034. (2024). [DOI] [PubMed]
38.Ishchenko, A. et al. Structural insights into the extracellular recognition of the human serotonin 2B receptor by an antibody. Proc. Natl. Acad. Sci. USA114, 8223–8228 10.1073/pnas.1700891114 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Robertson, M. J., van Zundert, G. C. P., Borrelli, K. & Skiniotis, G. GemSpot: a pipeline for robust modeling of ligands into cryo-EM maps. Structure28, 707–716 10.1016/j.str.2020.04.018 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Zhou, Q. T. et al. Common activation mechanism of class A GPCRs. eLife8. 10.7554/eLife.50279. (2019). [DOI] [PMC free article] [PubMed]
41.Halberstadt, A. L. & Geyer, M. A. Characterization of the head-twitch response induced by hallucinogens in mice detection of the behavior based on the dynamics of head movement. Psychopharmacology227, 727–739 10.1007/s00213-013-3006-z (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Halberstadt, A. L., Chatha, M., Klein, A. K., Wallach, J. & Brandt, S. D. Correlation between the potency of hallucinogens in the mouse head-twitch response assay and their behavioral and subjective effects in other species. Neuropharmacology167, 107933. 10.1016/j.neuropharm.2019.107933 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Kehler, J. & Lindskov, M. S. T. Are the LSD-analogs lisuride and ergotamine examples of non-hallucinogenic serotonin 5-HT_2A receptor agonists?. J. Psychopharmacol.39, 889–895 10.1177/02698811251330741 (2025). [DOI] [PubMed] [Google Scholar]
44.Friesner, R. A. et al. Glide: a new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J. Med. Chem.47, 1739–1749 10.1021/jm0306430 (2004). [DOI] [PubMed] [Google Scholar]
45.Gu, S., Smith, M. S., Yang, Y., Irwin, J. J. & Shoichet, B. K. Ligand strain energy in large library docking. J. Chem. Inf. Model61, 4331–4341 10.1021/acs.jcim.1c00368 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Olanders, G., Alogheli, H., Brandt, P. & Karlén, A. Conformational analysis of macrocycles: comparing general and specialized methods. J. Comput. Aid Mol. Des.34, 231–252 10.1007/s10822-020-00277-2 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Labute, P. LowModeMD-implicit low-mode velocity filtering applied to conformational search of macrocycles and protein loops. J. Chem. Inf. Model50, 792–800 10.1021/ci900508k (2010). [DOI] [PubMed] [Google Scholar]
48.Olsen, R. H. J. et al. TRUPATH, an open-source biosensor platform for interrogating the GPCR transducerome. Nat. Chem. Biol.16, 841–849 10.1038/s41589-020-0535-8 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Janicot, R. et al. Direct interrogation of context-dependent GPCR activity with a universal biosensor platform. Cell187 10.1016/j.cell.2024.01.028. (2024). [DOI] [PMC free article] [PubMed]
50.Chen, Z. C. et al. Structure-based design of a novel third-generation antipsychotic drug lead with potential antidepressant properties. Nat. Neurosci.25, 39–49 10.1038/s41593-021-00971-w (2022). [DOI] [PubMed] [Google Scholar]
51.Wacker, D. et al. Structural features for functional selectivity at serotonin receptors. Science340, 615–619 10.1126/science.1232808 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Mastronarde, D. N. Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol.152, 36–51 10.1016/j.jsb.2005.07.007 (2005). [DOI] [PubMed] [Google Scholar]
53.Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods14, 290–296 10.1038/Nmeth.4169 (2017). [DOI] [PubMed] [Google Scholar]
54.Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods14, 331–332 10.1038/nmeth.4193 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Zhang, K. Gctf: Real-time CTF determination and correction. J. Struct. Biol.193, 1–12 10.1016/j.jsb.2015.11.003 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Lyu, J. K. et al. AlphaFold2 structures guide prospective ligand discovery. Science384 10.1126/science.adn6354. (2024). [DOI] [PMC free article] [PubMed]
57.Pettersen, E. F. et al. UCSF ChimeraX: structure visualization for researchers, educators, and developers. Protein Sci.30, 70–82 10.1002/pro.3943 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D66, 486–501 10.1107/S0907444910007493 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Liebschner, D. et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr. Sect. D Struct. Biol.75, 861–877 10.1107/S2059798319011471 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Rosignoli, S. & Paiardini, A. Boosting the full potential of PyMOL with structural biology plugins. Biomolecules12 10.3390/biom12121764. (2022). [DOI] [PMC free article] [PubMed]
61.Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature630, 493–500 10.1038/s41586-024-07487-w (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Muhammed, M. T. & Aki-Yalcin, E. Homology modeling in drug discovery: overview, current applications, and future perspectives. Chem. Biol. Drug Des.93, 12–20 10.1111/cbdd.13388 (2019). [DOI] [PubMed] [Google Scholar]
63.Jacobson, M. P. et al. A hierarchical approach to all-atom protein loop prediction. Proteins55, 351–367 10.1002/prot.10613 (2004). [DOI] [PubMed] [Google Scholar]
64.Sastry, G. M., Adzhigirey, M., Day, T., Annabhimoju, R. & Sherman, W. Protein and ligand preparation: parameters, protocols, and influence on virtual screening enrichments. J. Comput Aid Mol. Des.27, 221–234 10.1007/s10822-013-9644-8 (2013). [DOI] [PubMed] [Google Scholar]
65.Lomize, M. A., Pogozheva, I. D., Joo, H., Mosberg, H. I. & Lomize, A. L. OPM database and PPM web server: resources for positioning of proteins in membranes. Nucleic Acids Res.40, 370–376 10.1093/nar/gkr703 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Kaplan, A. L. et al. Bespoke library docking for 5-HT_2A receptor agonists with antidepressant activity. Nature610, 582–591 10.1038/s41586-022-05258-z (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Cao, C. et al. Signaling snapshots of a serotonin receptor activated by the prototypical psychedelic LSD. Neuron110, 3154–3167 10.1016/j.neuron.2022.08.006 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Liang, L. J. et al. A new variant of the colistin resistance gene MCR-1 with co-resistance to β-lactam antibiotics reveals a potential novel antimicrobial peptide. Plos Biol.21 10.1371/journal.pbio.3002433. (2023). [DOI] [PMC free article] [PubMed]
69.Frey, B. J. & Dueck, D. Clustering by passing messages between data points. Science315, 972–976 10.1126/science.1136800 (2007). [DOI] [PubMed] [Google Scholar]
70.Black, J. W. P. L. Operational models of pharmacological agonism. Proc. R. Soc. Lond. Ser. B Biol. Sci.220, 141–162 10.1098/rspb.1983.0093 (1983). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(8.5MB, pdf)}

Reporting Summary^{(105KB, pdf)}

Transparent Peer Review file^{(13.2MB, pdf)}

Source Data^{(7.5MB, zip)}

Data Availability Statement

[CR1] 1.Hauser, A. S., Attwood, M. M., Rask-Andersen, M., Schioth, H. B. & Gloriam, D. E. Trends in GPCR drug discovery: new agents, targets and indications. Nat. Rev. Drug Discov.16, 829–842 10.1038/nrd.2017.178 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Vass, M. et al. Aminergic GPCR-ligand interactions: a chemical and structural map of receptor mutation data. J. Med. Chem.62, 3784–3839 10.1021/acs.jmedchem.8b00836 (2019). [DOI] [PubMed] [Google Scholar]

[CR3] 3.Lorente, J. S. et al. GPCR drug discovery: new agents, targets and indications. Nat. Rev. Drug Discov.24, 458–479 10.1038/s41573-025-01139-y (2025). [DOI] [PubMed] [Google Scholar]

[CR4] 4.Peng, Y. et al. 5-HT_2C receptor structures reveal the structural basis of GPCR polypharmacology. Cell172, 719–730 10.1016/j.cell.2018.01.001 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Fan, L. Y. et al. Haloperidol bound D2 dopamine receptor structure inspired the discovery of subtype selective ligands. Nat. Commun.11. 10.1038/s41467-020-14884-y. (2020). [DOI] [PMC free article] [PubMed]

[CR6] 6.Wang, S. et al. D4 dopamine receptor high-resolution structures enable the discovery of selective agonists. Science358, 381–386 10.1126/science.aan5468 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Arroyo-Urea, S. et al. A bitopic agonist bound to the dopamine 3 receptor reveals a selectivity site. Nat Commun15. 10.1038/s41467-024-51993-4. (2024). [DOI] [PMC free article] [PubMed]

[CR8] 8.Valant, C. et al. Separation of on-target efficacy from adverse effects through rational design of a bitopic adenosine receptor agonist. Proc. Natl. Acad. Sci. USA111, 4614–4619 10.1073/pnas.1320962111 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Rankovic, Z. CNS drug design: balancing physicochemical properties for optimal brain exposure. J. Med. Chem.58, 2584–2608 10.1021/jm501535r (2015). [DOI] [PubMed] [Google Scholar]

[CR10] 10.Ghose, A. K., Herbertz, T., Hudkins, R. L., Dorsey, B. D. & Mallamo, J. P. Knowledge-based, central nervous system (CNS) lead selection and lead optimization for CNS drug discovery. ACS Chem. Neurosci.3, 50–68 10.1021/cn200100h (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.McClure-Begley, T. D. & Roth, B. L. The promises and perils of psychedelic pharmacology for psychiatry. Nat. Rev. Drug Discov.21, 463–473 10.1038/s41573-022-00421-7 (2022). [DOI] [PubMed] [Google Scholar]

[CR12] 12.Duan, W., Cao, D. M., Wang, S. & Cheng, J. Serotonin 2A receptor (5-HT_2AR) agonists: psychedelics and non-hallucinogenic analogues as emerging antidepressants. Chem. Rev.124, 124–163 10.1021/acs.chemrev.3c00375 (2024). [DOI] [PubMed] [Google Scholar]

[CR13] 13.Slocum, S. T., DiBerto, J. F. & Roth, B. L. Molecular insights into psychedelic drug action. J. Neurochem.162, 24–38 10.1111/jnc.15540 (2022). [DOI] [PubMed] [Google Scholar]

[CR14] 14.Gumpper, R. H. & Roth, B. L. SnapShot: psychedelics and serotonin receptor signaling. Cell186, 232–232 10.1016/j.cell.2022.12.020 (2023). [DOI] [PubMed] [Google Scholar]

[CR15] 15.Simon, I. A. et al. Ligand selectivity hotspots in serotonin GPCRs. Trends Pharm. Sci.44, 978–990 10.1016/j.tips.2023.09.012 (2023). [DOI] [PubMed] [Google Scholar]

[CR16] 16.Higgins, G. A., Fletcher, P. J. & Shanahan, W. R. Lorcaserin: a review of its preclinical and clinical pharmacology and therapeutic potential. Pharm. Ther.205, 107417. 10.1016/j.pharmthera.2019.107417 (2020). [DOI] [PubMed] [Google Scholar]

[CR17] 17.He, Y. et al. Human loss-of-function variants in the serotonin 2C receptor associated with obesity and maladaptive behavior. Nat. Med.28, 2537–2546 10.1038/s41591-022-02106-5 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Dell’isola, G. B. et al. Evaluating bexicaserin for the treatment of developmental epileptic encephalopathies. Expert Opin. Pharmacother.25, 1121–1130 10.1080/14656566.2024.2373350 (2024). [DOI] [PubMed] [Google Scholar]

[CR19] 19.Pogorelov, V. M. et al. 5-HT_2C agonists modulate schizophrenia-like behaviors in mice. Neuropsychopharmacology42, 2163–2177 10.1038/npp.2017.52 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Cheng, J. & Kozikowski, A. P. We need 2C but not 2B: developing serotonin 2C (5-HT_2C) receptor agonists for the treatment of CNS disorders. ChemMedChem10, 1963–1967 10.1002/cmdc.201500437 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Canal, C. E. & Murnane, K. S. The serotonin 5-HT_2C receptor and the non-addictive nature of classic hallucinogens. J. Psychopharmacol.31, 127–143 10.1177/0269881116677104 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Jain, M. K. et al. The polypharmacology of psychedelics reveals multiple targets for potential therapeutics. Neuron 10.1016/j.neuron.2025.06.012 (2025). [DOI] [PubMed] [Google Scholar]

[CR23] 23.Roth, B. L. Focus on research-drug and valvular heart disease. N. Engl. J. Med356, 6–9 10.1056/NEJMp068265 (2007). [DOI] [PubMed] [Google Scholar]

[CR24] 24.Dumotier, B. M. & Urban, L. Preclinical mitigation of 5-HT_2B agonism-related cardiac valvulopathy revisited. J. Pharmacol. Tox. Met.128. 10.1016/j.vascn.2024.107542. (2024). [DOI] [PubMed]

[CR25] 25.Connolly, H. M. et al. Valvular heart disease associated with fenfluramine-phentermine. N. Engl. J. Med.337, 581–588 10.1056/NEJM199708283370901 (1997). [DOI] [PubMed] [Google Scholar]

[CR26] 26.Jick, H. et al. A population-based study of appetite-suppressant drugs and the risk of cardiac-valve regurgitation. N. Engl. J. Med.339, 719–724 10.1056/NEJM199809103391102 (1998). [DOI] [PubMed] [Google Scholar]

[CR27] 27.Schade, R., Andersohn, F., Suissa, S., Haverkamp, W. & Garbe, E. Dopamine agonists and the risk of cardiac-valve regurgitation. N. Engl. J. Med.356, 29–38 10.1056/NEJMoa062222 (2007). [DOI] [PubMed] [Google Scholar]

[CR28] 28.US Food and Drug Administration. Psychedelic Drugs: Considerations for Clinical Investigations. Guidance for Industry. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/psychedelic-drugs-considerations-clinical-investigations (2026).

[CR29] 29.Rorsted, E. M., Jensen, A. A., Smits, G., Frydenvang, K. & Kristensen, J. L. Discovery and structure-activity relationships of 2,5-dimethoxyphenylpiperidines as selective serotonin 5-HT2A receptor agonists. J. Med Chem.67, 7224–7244 10.1021/acs.jmedchem.4c00082 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Varty, G. B. et al. Synthesis and structure-activity relationships of 2,5-dimethoxy-4-substituted phenethylamines and the discovery of CYB210010: a potent, orally bioavailable and long-acting serotonin 5-HT2 receptor agonist. J. Med. Chem.67, 6144–6188 10.1021/acs.jmedchem.3c01961 (2024). [DOI] [PubMed] [Google Scholar]

[CR31] 31.Kim, K. et al. Structure of a hallucinogen-activated Gq-coupled 5-HT_2A serotonin receptor. Cell182, 1574–1588 10.1016/j.cell.2020.08.024 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Cao, D. M. et al. Structure-based discovery of nonhallucinogenic psychedelic analogs. Science375, 403–411 10.1126/science.abl8615 (2022). [DOI] [PubMed] [Google Scholar]

[CR33] 33.Gumpper, R. H. et al. The structural diversity of psychedelic drug actions revealed. Nat. Commun.16 10.1038/s41467-025-57956-7. (2025). [DOI] [PMC free article] [PubMed]

[CR34] 34.Ballesteros, J. A. & Weinstein, H. Integrated methods for the construction of three-dimensional models and computational probing of structure-function relations in G protein-coupled receptors. Methods Neurosci.25, 366–428 10.1016/S1043-9471(05)80049-7 (1995). [DOI]

[CR35] 35.McCorvy, J. D. et al. Structural determinants of 5-HT_2B receptor activation and biased agonism. Nat. Struct. Mol. Biol.25, 787–796 10.1038/s41594-018-0116-7 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Tenjin, T. et al. Profile of blonanserin for the treatment of schizophrenia. Neuropsych. Dis. Treat.9, 587–594 10.2147/Ndt.S34433 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Chen, Z. C. et al. Flexible scaffold-based cheminformatics approach for polypharmacological drug design. Cell187 10.1016/j.cell.2024.02.034. (2024). [DOI] [PubMed]

[CR38] 38.Ishchenko, A. et al. Structural insights into the extracellular recognition of the human serotonin 2B receptor by an antibody. Proc. Natl. Acad. Sci. USA114, 8223–8228 10.1073/pnas.1700891114 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Robertson, M. J., van Zundert, G. C. P., Borrelli, K. & Skiniotis, G. GemSpot: a pipeline for robust modeling of ligands into cryo-EM maps. Structure28, 707–716 10.1016/j.str.2020.04.018 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Zhou, Q. T. et al. Common activation mechanism of class A GPCRs. eLife8. 10.7554/eLife.50279. (2019). [DOI] [PMC free article] [PubMed]

[CR41] 41.Halberstadt, A. L. & Geyer, M. A. Characterization of the head-twitch response induced by hallucinogens in mice detection of the behavior based on the dynamics of head movement. Psychopharmacology227, 727–739 10.1007/s00213-013-3006-z (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR42] 42.Halberstadt, A. L., Chatha, M., Klein, A. K., Wallach, J. & Brandt, S. D. Correlation between the potency of hallucinogens in the mouse head-twitch response assay and their behavioral and subjective effects in other species. Neuropharmacology167, 107933. 10.1016/j.neuropharm.2019.107933 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Kehler, J. & Lindskov, M. S. T. Are the LSD-analogs lisuride and ergotamine examples of non-hallucinogenic serotonin 5-HT_2A receptor agonists?. J. Psychopharmacol.39, 889–895 10.1177/02698811251330741 (2025). [DOI] [PubMed] [Google Scholar]

[CR44] 44.Friesner, R. A. et al. Glide: a new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J. Med. Chem.47, 1739–1749 10.1021/jm0306430 (2004). [DOI] [PubMed] [Google Scholar]

[CR45] 45.Gu, S., Smith, M. S., Yang, Y., Irwin, J. J. & Shoichet, B. K. Ligand strain energy in large library docking. J. Chem. Inf. Model61, 4331–4341 10.1021/acs.jcim.1c00368 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Olanders, G., Alogheli, H., Brandt, P. & Karlén, A. Conformational analysis of macrocycles: comparing general and specialized methods. J. Comput. Aid Mol. Des.34, 231–252 10.1007/s10822-020-00277-2 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Labute, P. LowModeMD-implicit low-mode velocity filtering applied to conformational search of macrocycles and protein loops. J. Chem. Inf. Model50, 792–800 10.1021/ci900508k (2010). [DOI] [PubMed] [Google Scholar]

[CR48] 48.Olsen, R. H. J. et al. TRUPATH, an open-source biosensor platform for interrogating the GPCR transducerome. Nat. Chem. Biol.16, 841–849 10.1038/s41589-020-0535-8 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Janicot, R. et al. Direct interrogation of context-dependent GPCR activity with a universal biosensor platform. Cell187 10.1016/j.cell.2024.01.028. (2024). [DOI] [PMC free article] [PubMed]

[CR50] 50.Chen, Z. C. et al. Structure-based design of a novel third-generation antipsychotic drug lead with potential antidepressant properties. Nat. Neurosci.25, 39–49 10.1038/s41593-021-00971-w (2022). [DOI] [PubMed] [Google Scholar]

[CR51] 51.Wacker, D. et al. Structural features for functional selectivity at serotonin receptors. Science340, 615–619 10.1126/science.1232808 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Mastronarde, D. N. Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol.152, 36–51 10.1016/j.jsb.2005.07.007 (2005). [DOI] [PubMed] [Google Scholar]

[CR53] 53.Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods14, 290–296 10.1038/Nmeth.4169 (2017). [DOI] [PubMed] [Google Scholar]

[CR54] 54.Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods14, 331–332 10.1038/nmeth.4193 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Zhang, K. Gctf: Real-time CTF determination and correction. J. Struct. Biol.193, 1–12 10.1016/j.jsb.2015.11.003 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Lyu, J. K. et al. AlphaFold2 structures guide prospective ligand discovery. Science384 10.1126/science.adn6354. (2024). [DOI] [PMC free article] [PubMed]

[CR57] 57.Pettersen, E. F. et al. UCSF ChimeraX: structure visualization for researchers, educators, and developers. Protein Sci.30, 70–82 10.1002/pro.3943 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D66, 486–501 10.1107/S0907444910007493 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Liebschner, D. et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr. Sect. D Struct. Biol.75, 861–877 10.1107/S2059798319011471 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Rosignoli, S. & Paiardini, A. Boosting the full potential of PyMOL with structural biology plugins. Biomolecules12 10.3390/biom12121764. (2022). [DOI] [PMC free article] [PubMed]

[CR61] 61.Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature630, 493–500 10.1038/s41586-024-07487-w (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Muhammed, M. T. & Aki-Yalcin, E. Homology modeling in drug discovery: overview, current applications, and future perspectives. Chem. Biol. Drug Des.93, 12–20 10.1111/cbdd.13388 (2019). [DOI] [PubMed] [Google Scholar]

[CR63] 63.Jacobson, M. P. et al. A hierarchical approach to all-atom protein loop prediction. Proteins55, 351–367 10.1002/prot.10613 (2004). [DOI] [PubMed] [Google Scholar]

[CR64] 64.Sastry, G. M., Adzhigirey, M., Day, T., Annabhimoju, R. & Sherman, W. Protein and ligand preparation: parameters, protocols, and influence on virtual screening enrichments. J. Comput Aid Mol. Des.27, 221–234 10.1007/s10822-013-9644-8 (2013). [DOI] [PubMed] [Google Scholar]

[CR65] 65.Lomize, M. A., Pogozheva, I. D., Joo, H., Mosberg, H. I. & Lomize, A. L. OPM database and PPM web server: resources for positioning of proteins in membranes. Nucleic Acids Res.40, 370–376 10.1093/nar/gkr703 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Kaplan, A. L. et al. Bespoke library docking for 5-HT_2A receptor agonists with antidepressant activity. Nature610, 582–591 10.1038/s41586-022-05258-z (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR67] 67.Cao, C. et al. Signaling snapshots of a serotonin receptor activated by the prototypical psychedelic LSD. Neuron110, 3154–3167 10.1016/j.neuron.2022.08.006 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR68] 68.Liang, L. J. et al. A new variant of the colistin resistance gene MCR-1 with co-resistance to β-lactam antibiotics reveals a potential novel antimicrobial peptide. Plos Biol.21 10.1371/journal.pbio.3002433. (2023). [DOI] [PMC free article] [PubMed]

[CR69] 69.Frey, B. J. & Dueck, D. Clustering by passing messages between data points. Science315, 972–976 10.1126/science.1136800 (2007). [DOI] [PubMed] [Google Scholar]

[CR70] 70.Black, J. W. P. L. Operational models of pharmacological agonism. Proc. R. Soc. Lond. Ser. B Biol. Sci.220, 141–162 10.1098/rspb.1983.0093 (1983). [DOI] [PubMed] [Google Scholar]

PERMALINK

Structure-based design of subtype-selective psychedelic analogs

Huiqiong Li

Lingjie Tang

Jinfeng Zhang

Bing Meng

Dongmei Cao

Yujin Chen

Jing Yu

Huan Wang

Zhi-Jie Liu

Sheng Wang

Jianjun Cheng

Abstract

Introduction

Results

Structural insights into OBP features of 5-HT2AR and 5-HT2BR

Fig. 1. Structural comparison between 5-HT2AR and 5-HT2BR and rational design of subtype-selective 5-HT2AR agonists.

Structure-based design of two chemical types

Pharmacological profiling

Fig. 2. Pharmacological profiling of subtype-selective 5-HT2AR agonists.

Binding pose validation with cryo-EM structures

Fig. 3. Structural basis of subtype-selective 5-HT2AR agonists.

Validation through receptor mutagenesis and SAR

Fig. 4. Effects of receptor mutants and compound structural modifications on 5-HT2AR versus 5-HT2BR activation.

Fig. 5. Comparison between non-selective and subtype-selective ligands.

Pharmacological profiling at other serotonin receptors

Hallucinogenic potential and antidepressant effects in animal models

Fig. 6. Pharmacokinetic profiling and behavioral pharmacology.

Discussion

Methods

Ethics statement

Fragment-based design of series A compounds

Radioligand binding affinity

Bioluminescence resonance energy transfer assay (BRET)

Generation of 5-HT2AR and 5-HT2BR constructs for structure determination

Protein expression

Purification of 5-HT2AR/compound complex

Purification of miniGαq-γ2/Gβ1 heterotrimeric complex

Purification of scFv16

Formation of 5-HT2AR/compound/miniGαq-γ2/Gβ1/scFv16 complex

Purification of 5-HT2BR/compound complex

Purification of P2C2–Fab

Formation of 5-HT2BR/compound/Fab complex

Sample preparation, Cryo-EM data collection and processing

Model building and refinement

Molecular dynamics simulation and analysis

ADME assays

Caco-2 permeability

Liver microsomal stability

Plasma stability

Octanol-buffer logD

Kinetic solubility

Pharmacokinetics

Mice

Head twitch response (HTR) detection system based on magnetic signals

Corticosterone-induced depression-like model

Tail suspension test (TST)

Rotarod test

Quantification and statistical analysis

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Funding

Data availability

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Structural insights into OBP features of 5-HT_2AR and 5-HT_2BR

Fig. 1. Structural comparison between 5-HT_2AR and 5-HT_2BR and rational design of subtype-selective 5-HT_2AR agonists.

Fig. 2. Pharmacological profiling of subtype-selective 5-HT_2AR agonists.

Fig. 3. Structural basis of subtype-selective 5-HT_2AR agonists.

Fig. 4. Effects of receptor mutants and compound structural modifications on 5-HT_2AR versus 5-HT_2BR activation.

Generation of 5-HT_2AR and 5-HT_2BR constructs for structure determination

Purification of 5-HT_2AR/compound complex

Purification of miniG_αq-γ2/G_β1 heterotrimeric complex

Formation of 5-HT_2AR/compound/miniG_αq-γ2/G_β1/scFv16 complex

Purification of 5-HT_2BR/compound complex

Formation of 5-HT_2BR/compound/Fab complex