G protein-coupled receptors (GPCRs) mediate information transfer to cells from the surrounding environment. In most cases, signaling is initiated or amplified when the receptor binds to an agonist, an event that alters the conformational profile of the receptor. Signal transduction results from interaction between the agonist-receptor complex and cytosolic partners such as G proteins, GPCR kinases (GRKs), and β-arrestins. Changes in agonist structure can lead to "signal bias", i.e., changes in the relative strength of signaling involving different partners. Some GPCRs, including those activated by long peptide hormones, continue to signal after internalization. In these cases, changes in agonist structure can lead to changes in the relative extent of signaling from different sites, e.g., cell surface vs endosomes ("location bias"). Many GPCRs are targets of approved drugs or drug candidates, and tuning signal bias and/or location bias is widely considered to be important for optimizing therapeutic profiles. Here we report another mechanism of modulating outcome via agonist modification: alteration of intracellular trafficking. The synthetic peptide agonist designated SPT, which contains five β-amino acid residues, was previously shown to activate the parathyroid hormone receptor-1 (PTH1R) and cause prolonged signaling in mice by an unknown mechanism. The SPT-PTH1R complex continues to stimulate cAMP production after internalization. We now find that the SPT-PTH1R complex impairs the sorting of early endosomes into recycling endosomes relative to the receptor complexed to the drug teriparatide. These findings suggest that altering intracellular GPCR trafficking patterns represents an unappreciated strategy for achieving prolonged action in vivo.

Maladaptive consummatory behaviors can arise from dysregulated circuits, like the extended amygdala that governs motivation and feeding. Neurotensin (NTS) is expressed throughout the central, peripheral, and enteric nervous systems with well-established roles in energy balance and feeding. SBI-553, a β-arrestin-biased allosteric modulator of NTSR1, recruits β-arrestin while attenuating G q -mediated signaling. We used SBI-553 to examine NTS modulation of extended amygdala GABAergic signaling, and probed its effects on food consumption in mice. Ex vivo , we found that NTS and SBI-553 differentially modulates GABAergic neurotransmission across extended amygdala subregions. In vivo , SBI-553 reduces palatable food consumption in both fed and food-deprived mice, with greater reductions under fasted conditions. SBI-553 alters activation across CeA subregions in a sex- and feeding-state-dependent manner: SBI-553 increases cFos immunofluorescence in the CeA L and CeA C , but not the CeA M . This work supports neurotensinergic modulation as a compelling target for further investigation into the neural substrates of consummatory behaviors.

Second generation antipsychotics (SGAs) are widely used clinical tools; yet, they often cause negative side effects and take weeks to become effective, leading to poor patient compliance. The effect/side effect profile of individual SGAs is highly variable, and the mechanisms that underlie this variability are not well understood. Here, we identify a role of type 3 dopamine receptor (D3R) neurons in the Nucleus Accumbens (NAc) that mimics the aversive effects of quetiapine. Using single-nucleus RNA sequencing, we show that D3R is expressed in a subpopulation of D1R neurons and defines a distinct NAc cell type. We found that both clozapine and quetiapine cause acute conditioned place aversion in mice, but aversion subsides only after 21 days of treatment with quetiapine, a drug characterized as an arrestin-biased agonist at D3R. We provide evidence at both the cellular and population level that quetiapine inhibits D3R-expressing neurons in the lateral shell (LatSh) of the NAc. We also demonstrate that local injection of quetiapine into LatSh NAc is sufficient to cause place aversion. Selective optogenetic inhibition of D3R-neurons in the LatSh produces real time place aversion in mice, correlating this cell type to the aversive effects of quetiapine. These findings suggest a cell-type-specific mechanism underlying quetiapine-induced aversion and its potential attenuation with chronic treatment, offering insight into the cell types and circuitry that shape the quetiapine side effect profile.

The apelin receptor (AplnR) is a prominent peptide-binding Class A G protein-coupled receptor (GPCR) involved in the regulation of cardiovascular, gastrointestinal, and immune functions, as well as skeletal development. Activation of the AplnR-mediated β-arrestin pathway has been associated with adverse effects including pathological cardiac hypertrophy and heart failure. Therefore, the development of G protein-biased small-molecule agonists is expected to provide an improved safety profile for long-term therapeutic use. In this study, a series of 1H-1,2,4-triazole-3-carboxamide derivatives were designed and synthesized as small-molecule AplnR agonists. Among them, the most promising compound B-007 exhibited potent agonist activity with an EC50 value of 11.6 nM. Notably, compared with an endogenous ligand apelin-13, B-007 demonstrated a pronounced G protein bias with abolished β-arrestin signaling. These results identified a new promising scaffold for the development of G protein-biased agonists.

Bitter taste receptors (TAS2Rs) are GPCRs functionally expressed in extraoral organs including the lung, heart, brain, and gastrointestinal system and are candidate targets for novel drugs. It is unclear whether TAS2R5 can be stabilized by structurally distinct agonists resulting in pathway selectivity (biasing) to achieve optimal outcomes. We screened TAS2R5 agonists for signaling toward or away from β-arrestin or G protein ([Ca[2+]]i). Agonists T5-1 (1,10-phenanthroline) and T5-6 (4,7-dimethyl-1,10-phenanthroline) coupled equally to G protein but showed marked differences in homologous desensitization (∼21% and ∼91%, respectively). This desensitization was due to dissimilar degrees of β-arrestin engagement compared to the reference agonist T5-7: T5-6 evoked up to ∼86% enhancement of β-arrestin1 or β-arrestin2 recruitment, while T5-1 was biased in the opposite direction, up to ∼78% less, compared to T5-7. T5-1 elicited little receptor internalization compared to T5-6 consistent with the β-arrestin findings. For the initial response, T5-1 and T5-6 represent one-pathway differences (decreased or increased β-arrestin) leading to bias in different directions, while G protein signaling was equivalent. However, under desensitizing conditions T5-6 was decoupled from G protein representing two-pathway biasing in opposing directions for T5-6 (increased β-arrestin and decreased G protein), imposing extreme biasing of β-arrestin over G protein. Our finding that agonists can affect β-arrestin biasing in either direction suggests that TAS2R5 is sufficiently pliable for specific signals to be engineered depending on the desired therapeutic outcome. The dynamic nature of the directionally contrasting changes of the two pathways over time accentuated bias and is due to the interaction of the two signals.

Arrestins regulate G protein-coupled receptor (GPCR) signaling by undergoing large-scale conformational rearrangements, yet the solution-state equilibria that underlie arrestin pre-activation remain poorly defined. While prior studies identified slow conformational exchange at the interdomain interface, these minor states could not be structurally linked to activation because their resonances broaden beyond detection upon receptor binding. Here, we use multinuclear NMR spectroscopy to characterize the intrinsic conformational landscape of full-length, human arrestin-2 in solution. We identify two distinct, pre-existing conformational equilibria that mirror key steps of the receptor-driven activation process. First, a slow exchange process populates a receptor-bound-like, interdomain-twisted conformation at physiological temperatures. In parallel, a faster, globally distributed equilibrium populates a state consistent with C-terminal tail release. Dynamic analyses reveal localized rigidification in the receptor-bound-like minor states despite arrestin's overall flexibility, while backbone relaxation data indicate widespread μs-ms conformational exchange. Together, these results demonstrate that arrestin-2 acts as a preorganized scaffold that intrinsically samples receptor-binding relevant conformations in the absence of binding partners. This provides a solution-state framework for arrestin pre-activation and establishes a dynamic fingerprint for future ligand-dependent studies.

A central feature of G protein-coupled receptor (GPCR) desensitization is the direct competition between heterotrimeric G proteins and β-arrestins (βarrs) for an overlapping binding site within the intracellular receptor cavity. Although numerous high-resolution structures of GPCR-transducer complexes exist, the exact nature of this shared site and the molecular basis of transducer competition remain unclear. To investigate this, we employed an interdisciplinary approach integrating systemic mutational mapping, bioinformatics, and structural analysis across multiple classes of GPCRs and their transducers and regulators. We identified two highly conserved leucine residues within both the βarr finger loop and the Gα C-terminal α-helix, which engage a hydrophobic patch on GPCRs formed by TM3, TM5, and TM6 in a nearly identical manner, thereby stabilizing the complexes. Notably, the GPCR kinase N-terminal α-helix also contains hydrophobic residues that associate with this same receptor patch and are vital for the GPCR-GRK engagement. Our findings reveal a conserved hydrophobic interface that mediates direct competition among GPCR transducers and regulators suggesting a universal mechanism that governs receptor access and desensitization.

The extent to which local GPCR surface density governs engagement of downstream signaling and trafficking pathways remains unclear. Using single-particle tracking of the μ-opioid receptor (MOR), we show that receptor density differentially regulates G protein signaling and GRK2/3-β-arrestin-dependent receptor trafficking. At low surface density, MORs activate G proteins but fail to enter clathrin-coated structures despite the presence of endogenous GRK2/3 and β-arrestin. Increasing MOR density, co-expressing other class A GPCRs, or elevating GRK2 or β-arrestin abundance rescues agonist-induced MOR trafficking. In contrast, the class B GPCR V2R blocks MOR trafficking at both low and high MOR densities. These results support a model in which increasing class A GPCR density, despite worsening effector-to-receptor stoichiometry, promotes trafficking by forming an affinity matrix that enables reversible GRK2/3 and β-arrestin interactions to be productively used by neighboring receptors in a density-dependent manner, whereas class B GPCRs sequester β-arrestin and block trafficking.

G-protein-coupled receptors (GPCRs) transmit cellular signals through both G protein and arrestin pathways and biased signaling offers potential therapeutic advantages through selective activation. Although GPCR-G protein complexes are well characterized, structural understanding of class B GPCR-arrestin interactions remains limited. Here we show the cryo-electron microscopy structure of parathyroid hormone receptor 1 in core engagement with β-arrestin 1, revealing the molecular basis of arrestin coupling. The structure shows a rearrangement in which inward movement of extracellular transmembrane helix 5 (TM5) and extracellular loop 3 (ECL3) drives outward displacement of cytoplasmic TM5, forming a configuration required for arrestin binding. Guided by comparison with the Gs-coupled state, we designed peptide analogs that prevent these TM5/ECL3 conformational changes, producing G-protein-biased agonists that preserve agonist efficacy while reducing arrestin recruitment. In an ovariectomized mouse model, a lead compound shows comparable therapeutic efficacy, providing a framework for structure-guided design of biased therapeutics targeting class B GPCRs.

The arrestin-mediated regulation of signaling through the metabotropic glutamate receptors (mGlus) is fundamental mechanism modulating excitatory transmission and synaptic plasticity. However, molecular details of arrestin engagement are elusive for these dimeric receptors. Here we report the structures of mGlu5 and mGlu7 bound to β-arrestin 1 (βarr1) and their endogenous agonist glutamate. The structures reveal a symmetric interaction pattern of two βarr1 molecules coupling solely to the membrane-proximal C-terminal regions of the active mGlu5, and an asymmetric binding manner with one βarr1 interacting with both transmembrane domains in the inactive mGlu7. Supported by mass spectrometry and functional data, the mGlus adopt multiple subtype-dependent arrestin binding modes that are determined by both the receptor core and C terminus. Furthermore, the activities of arrestin-mediated desensitization and endocytosis of the mGlus are likely associated with the distinct arrestin binding patterns, which may account for the differential signaling profiles of different mGlu subtypes.

Among the vehicles investigated as delivery agents for antisense RNAs are extracellular vesicles (EVs) released from cultured cells. Arrestin domain-containing 1 (ARRDC1)-mediated microvesicles (ARMMs) are naturally occurring EVs that uniquely are formed by outward budding of cytoplasmic membranes. Previous work has shown that ARMMs production is quantitatively controlled by the ARRDC1 protein and that ARRDC1 and macromolecules attached to it are loaded into nascent ARMMs during ARMM formation. Here, we report a strategy for constructing cell factories that biologically manufacture both ARMMs and short hairpin RNA (shRNA)-like antisense RNA precursors, designated as shT-RNAs. Human HEK293T cells mutated in the endoribonuclease DICER1 were engineered to express a fusion protein containing components of ARRDC1 and the trans-activator of transcription (Tat) peptide encoded by the HIV-1 virus. Prodrug RNAs were constructed by replacing the canonical loop regions of shRNAs with a 24-nucleotide segment derived from the Tat-binding trans-activation response (TAR) element. We show that a truncated TAR sequence embedded within the structural framework of shT-RNAs enables their linkage to the ARRDC1-Tat fusion protein and consequent loading of the RNAs into nascent ARMMs, and that RNA is protected by ARMMs membranes from attack by external ribonucleases. We further show that prodrug modules consisting of shT-RNAs that target different genomic sequences of SARS-CoV-2 virus and are manufactured as components of a single transcript can, when delivered by ARMMs, be activated to suppress virus RNA production. Our results indicate that the ARMMs-based platform we have designed can deliver gene-silencing RNAs in the form of biologically produced shRNA-like prodrugs.