Opioid drugs reach deep into their G proteins, single-molecule study reveals

Every dose of an opioid drug, from morphine to fentanyl, begins with a molecular handshake at the surface of a nerve cell. The drug binds the μ-opioid receptor, a seven-helix protein embedded in the membrane, and the receptor in turn grips a G protein waiting on the intracellular side. For decades, pharmacologists have ranked these drugs by efficacy, the strength of the cellular response they provoke, and have largely attributed those differences to how well each ligand nudges the receptor into a shape that recruits and activates its G protein partner. A new study published in Nature Structural & Molecular Biology by Jonathan Deutsch, Rabindra Shivnaraine, Brian Kobilka, Steven Chu and colleagues now shows that the story does not end at the receptor. The drug’s identity is transmitted, allosterically, more than 70 angstroms across the entire signaling complex, reshaping the internal dynamics of the G protein itself.

The target of the new work is the α subunit of the Gi protein, a two-lobed molecular machine. One lobe, the Ras-like domain, carries the guanine nucleotide pocket that holds GDP in the resting state. The other, the α-helical domain, is a six-helix bundle that clamps down over that pocket, burying GDP at the interface between the two lobes. Activation by a receptor requires the helical domain to swing open by tens of angstroms, exposing the nucleotide site so GDP can escape and GTP can flood in. That opening step, long inferred from crystallography and double electron–electron resonance spectroscopy, had never been watched in real time on single molecules, and its dependence on the drug occupying the receptor was unknown.

To capture those motions, the team engineered a minimal-cysteine version of the Gαi1 α subunit and attached two fluorophores, LD555 and JFX673, at positions R90C on the helical domain and E238C on the Ras-like domain. When the helical domain opens, the dyes drift apart and the Förster resonance energy transfer signal between them drops; when the domain closes, the signal rises. The labeled heterotrimers were coupled to purified μ-opioid receptors, immobilized on quartz slides through an anti-Flag antibody fragment, and imaged with total internal reflection fluorescence microscopy at 50-millisecond resolution. The construct behaved like the wild-type protein in GTP turnover assays, preserving the canonical rank order of ligand efficacies, so the dynamics observed could be interpreted with confidence.

Individual complexes flickered reversibly between a low-FRET, open state centered near 0.27 and a high-FRET, closed state near 0.62, exchanging roughly once per second. The crucial finding emerged when the researchers compared complexes bound to different opioids. Full agonists such as lofentanil, BU72 and DAMGO shifted the equilibrium toward the open state, while partial agonists such as morphine, buprenorphine and mitragynine pseudoindoxyl left the protein spending more time closed. Strikingly, the fraction of time spent in the open state correlated almost perfectly with GTP turnover activity, the standard biochemical readout of efficacy, with a coefficient of determination of 0.97. The unliganded receptor was least able to stabilize the open conformation, matching its sluggish basal activity.

Hidden Markov modeling of the trajectories revealed where the ligand exerts its leverage. Dwell times in the open state were essentially identical across all ligands, averaging about 1.4 seconds, but dwell times in the closed state depended strongly on efficacy. Lofentanil-bound complexes opened roughly 2.25-fold faster than unliganded ones, at about 0.9 per second versus 0.4 per second. In other words, high-efficacy drugs do not hold the G protein open; they destabilize its closed, GDP-clamped configuration, making domain separation more likely. A transient mid-FRET state near 0.47, often lasting barely 100 milliseconds, appeared to serve as an unstable intermediate on the path between the extremes.

The generality of the mechanism was tested with a second receptor. Neurotensin receptor 1, coupled to the same labeled Gi protein, produced an analogous correlation between open-state occupancy and GTP turnover, with R² of 0.89. Adding SBI-553, an allosteric modulator known to bias neurotensin signaling, shifted the helical domain further toward the open state in proportion to its positive effect on nucleotide exchange. Together with earlier bioluminescence resonance energy transfer studies showing ligand-specific G protein conformations at calcitonin and α2-adrenergic receptors, the results suggest that long-range allosteric encoding of efficacy may be a widespread feature of GPCR signaling rather than an opioid peculiarity.

The team also dissected the nucleotide chemistry underlying these motions. Because GTP or GDP delivery causes the complex to fall apart too quickly for equilibrium imaging, the researchers used a mixture of GMP and pyrophosphate, which together mimic the triphosphate in the nucleotide pocket without triggering dissociation. Titrating this pair onto DAMGO-bound complexes drove a concentration-dependent closure of the helical domain, confirming that the closed state binds nucleotide more tightly than the open state. Morphine-bound complexes were roughly fourfold more sensitive to the nucleotide mimic than DAMGO-bound ones, with half-maximal effective concentrations of 0.34 and 1.53 micromolar respectively, exactly as expected if partial agonists leave the G protein in a higher-affinity, more closed ensemble. Synchronization analysis caught many closure events passing through the partially open intermediate, implying that nucleotide contacts are broken and formed progressively rather than all at once, an energetically economical route for exchange.

These dynamics resolve a long-standing pharmacological puzzle. In isolation, buprenorphine-bound complexes actually bound the fluorescent GTP analog GTPγS faster than lofentanil-bound complexes, which seems backwards for a weaker agonist. But the cytoplasm contains both GDP and GTP, with GTP about tenfold more abundant, so activation is limited by GDP release, not GTP binding. When physiological GDP was added, the ranking reversed: GDP competitively slowed GTP loading onto buprenorphine-bound complexes far more than onto lofentanil-bound ones. Full agonists pry the domains apart, lower nucleotide affinity and let GDP escape quickly; partial agonists favor closure, prolong GDP residence and delay activation. Structurally, the authors trace this allostery to the α5 helix of Gα, which inserts into the receptor’s cytosolic vestibule and disrupts the α1 helix and the interdomain interface, with full agonists promoting deeper α5 engagement through their larger displacement of transmembrane helix 6.

The implications reach well beyond basic mechanism. Full agonists dominate pain management, while partial agonists such as buprenorphine anchor the treatment of opioid use disorder, and a mechanistic account of why partial agonism arises, kinetic trapping of GDP-bound intermediates rather than mere failure of receptor coupling, gives medicinal chemists a new dial to tune. If the G protein itself is a locus where ligand-specific signaling behavior is encoded, then safer analgesics might be designed not only by shaping receptor conformations but by controlling how those conformations propagate into the transducer. The study also cautions structural biologists: cryo-electron microscopy snapshots, which capture the lowest-energy states of a heterogeneous ensemble, often show minimal differences between ligands, whereas the single-molecule trajectories reveal that subtle changes in interdomain energetics can transform the signaling outcome. Watching the machine move, it turns out, matters as much as photographing where it rests.

Subject of Research: Allosteric control of G protein conformational dynamics by opioid ligands at the μ-opioid receptor

Article Title: Ligand-dependent G protein dynamics underlying opioid efficacy

Article References: Ligand-dependent G protein dynamics underlying opioid efficacy. (n.d.). https://doi.org/10.1038/s41594-026-01893-6

Image Credits: AI Generated

DOI: 10.1038/s41594-026-01893-6

Keywords: μ-opioid receptor, G protein, GPCR, single-molecule FRET, allostery, opioid efficacy, GDP release, nucleotide exchange, buprenorphine, pharmacology, conformational dynamics, structural biology