Chemical genetics, in the sense Stuart Schreiber and others used the phrase, meant asking a genetic-style question with a small molecule: remove a function, quickly, and only that function. GPCRs were an obvious place to try to build the tool on purpose. In 1998 Coward, Wada, and colleagues described Ro1, a κ-opioid receptor mutant that ignored dynorphin and answered spiradoline. They called it a RASSL. It proved the principle and advertised the failure mode. Spiradoline is not inert in a normal animal. The mutant was not perfectly inactive. Through the early 2000s a handful of other RASSLs were made, and almost none of them became a method other laboratories could trust in vivo. The field needed a receptor evolved against a ligand that did nothing else.

Bryan Roth’s laboratory stopped drawing the mutations and let yeast evolution find them. Armbruster, Li, Pausch, Herlitze, and Roth reported in 2007 a family of human muscarinic receptors, altered at two orthosteric positions (Ballesteros–Weinstein Y3.33 and A5.46), potently activated by clozapine-N-oxide (CNO) and essentially insensitive to acetylcholine. CNO was chosen because it was regarded as an inert metabolite of the antipsychotic clozapine. They named them DREADDs. Two years later Alexander and colleagues put hM3Dq into mouse neurons with a transgenic cross and changed behavior with a systemic injection. The same year Guettier and colleagues retargeted the scaffold to Gs and controlled insulin release. Chemogenetics became a systems tool. The bargain was attractive next to optogenetics, which had arrived in 2005 and owned the millisecond: no fiber optic, both hemispheres at once, a manipulation that lasts a session, repeatable across days, and in principle writable as a gene therapy plus a pill.

The toolkit thickened. Cre- and Flp-dependent AAVs put DREADDs under cell-type control. Vardy and colleagues added KORD in 2015, a κ-opioid DREADD opened by salvinorin B, so excitation and inhibition no longer had to share a ligand. By 2016 Roth’s Neuron primer could treat DREADDs as standard equipment.

The PSAM/PSEM (Pharmacologically Selective Actuator Module/Pharmacologically Selective Effector Molecule) chemogenetic technology was pioneered by Scott Sternson and colleagues at the Janelia Research Campus and first reported in Science in 2011. The researchers used a protein-engineering strategy based on a modified ligand-binding domain of the α7 nicotinic acetylcholine receptor (α7 nAChR), introducing mutations that enabled selective recognition of synthetic ligands (PSEMs) while minimizing activation by endogenous acetylcholine. These engineered receptor domains (PSAMs) were then coupled to ion-channel pore domains from other receptors, including the 5-HT₃ receptor and glycine receptor, creating modular tools capable of either directly exciting or inhibiting genetically targeted neurons. The team demonstrated selective neuronal activation and silencing in vitro and, importantly, established in vivo efficacy by reversibly suppressing hypothalamic neuron-driven feeding behavior in mice. This work introduced a complementary chemogenetic platform offering direct control of neuronal excitability through ligand-gated ion channels rather than G-protein-coupled receptor signaling.

A pivotal advance in the development of DREADD chemogenetics came in 2017, when Gomez, Bonaventura, Michaelides and colleagues made the unexpected discovery that clozapine, rather than its purported actuator clozapine-N-oxide (CNO), is the principal pharmacologically active ligand responsible for central DREADD activation in vivo. Published in Science, this study demonstrated that CNO has poor brain penetration and is converted metabolically to clozapine, which exhibits substantially higher affinity and potency at the engineered muscarinic receptors hM3Dq and hM4Di. Using PET imaging, receptor occupancy studies and behavioral experiments, the investigators established that systemically administered clozapine readily occupies brain-expressed DREADDs and that subthreshold doses can produce preferential DREADD-mediated effects. This discovery fundamentally changed the understanding of DREADD pharmacology and identified an FDA-approved medication as a viable actuator with potential for clinical translation. Building on this finding, Michaelides and colleagues subsequently developed a new generation of high-potency DREADD agonists, reported in Nature Communications in 2019. Through structure-guided medicinal chemistry based on clozapine, the team developed JHU37152 and JHU37160, novel agonists with nanomolar affinity, excellent brain penetration and high in vivo potency at both excitatory and inhibitory DREADDs. The study also introduced [¹⁸F]JHU37107, the first dedicated fluorine-18 PET radiotracer for DREADDs, enabling noninvasive visualization of receptor expression, including at distant neuronal projection sites. Critically, the compounds were validated across rodents and nonhuman primates, establishing a pharmacological and molecular-imaging toolkit that substantially advanced the potential of DREADD technology for translational neuroscience and therapeutic applications.

A major advance in chemogenetics came in 2019, when Magnus, Lee, Bonaventura, Michaelides, Sternson and colleagues reported the development of PSAM4 in Science. Through structure-guided receptor engineering, they introduced three mutations (L131G, Q139L, Y217F) into the α7 nAChR ligand-binding domain, producing a receptor with dramatically enhanced sensitivity to varenicline, an FDA-approved smoking-cessation medication. The resulting PSAM4 platform supported both neuronal inhibition through PSAM4-GlyR and excitation through PSAM4-5HT3. Importantly, the study also developed a new generation of ultrapotent PSEMs (uPSEMs), including compounds with subnanomolar potency and exceptional selectivity, such as uPSEM817 (EC₅₀ approximately 0.3 nM). The technology was validated through electrophysiology, behavioral studies, calcium imaging and PET in rodents, as well as electrophysiological demonstration of neuronal silencing in a rhesus macaque. Together, these advances substantially expanded the pharmacological capabilities of PSAM/PSEM chemogenetics, improving potency, selectivity, and translational potential while establishing a versatile platform for bidirectional control of neuronal activity.

An important subsequent milestone was the development of cocaine-responsive chemogenetics, reported by Gomez, Magnus, Bonaventura, Michaelides, Sternson and colleagues in Nature in 2025. This work represented a conceptual extension of PSAM/PSEM technology: rather than relying on an externally administered synthetic ligand to activate an engineered receptor, the investigators engineered ion channels that are directly activated by cocaine itself. Using protein engineering of the α7 nicotinic acetylcholine receptor ligand-binding domain, the team generated cocaine-gated channels, including coca-5HT3, with high sensitivity to cocaine and selectivity over endogenous molecules and other addictive drugs. The resulting receptor was expressed in neurons of the lateral habenula (LHb), a brain region normally inhibited by cocaine. Cocaine exposure therefore activated the engineered channels, producing an artificial neuronal feedback signal that opposed cocaine's normal effects on brain reward circuitry. In rats, this approach suppressed cocaine self-administration, particularly under high-effort conditions, without impairing motivation for food. Using PET imaging with [¹⁸F]fallypride, alongside complementary neurochemical approaches, the investigators demonstrated that cocaine-gated activation of LHb neurons attenuated cocaine-induced increases in extracellular dopamine in the nucleus accumbens. This study introduced the concept of synthetic physiology: engineering artificial, drug-dependent feedback mechanisms that selectively counteract the physiological effects of addictive substances. It also established a potential foundation for developing chemogenetic gene therapies that respond directly to addictive drugs while sparing natural reward processes. Importantly, this was a distinct extension of the broader receptor-engineering strategy rather than simply another PSAM4/PSEM variant.