Psychedelics are undergoing a renaissance as potential therapy for psychiatric disorders, with more than 200 clinical trials being studied across several countries1–3. However, the precise mechanisms by which these drugs bring about benefits and the potential clinical risks are not yet fully understood. The serotonin 2A receptor (5-HT2AR) was reported to be a Gq-coupled receptor and the primary interoceptive target of psychedelics4,5. Here we compared psychedelics and their non-hallucinogenic analogues (nHAs) using in vitro and in vivo approaches, finding that 5-HT2AR-mediated non-canonical Gi signalling is essential for hallucinogenic effect. We further presented five cryo-electron microscopy structures of 5-HT2AR–Gi/Gq in complex with psychedelics or nHAs. Structural analysis and pharmacological investigation revealed that a special contact between nHAs with 5-HT2AR mediated the signalling bias. Building on this insight, we identified a 2,5-dimethoxy-4-iodoamphetamine derivative, DOI-NBOMe, which exhibits potent and selective Gq-biased activity, and demonstrates promising therapeutic effects in mouse models without hallucinogenic effect. Our finding uncovers the functional mechanisms underlying the Gi signalling mediated by 5-HT2AR and provides valuable insights for designing psychedelic-based drugs with minimized risk from hallucinogenic effects.
