Arrestins as programmable integrators of GPCR signaling: structural microstates, spatiotemporal logic, and therapeutic control

Read post on nature.com

IntroductionG protein-coupled receptors (GPCRs) constitute the largest superfamily of cell-surface signaling proteins in mammals, mediating responses to diverse stimuli ranging from photons to peptide hormones and serving as targets for ~36% of approved drugs1,2. Traditionally, GPCRs transduce signals via two distinct pathways: G protein activation and arrestin recruitment3,4. The arrestin family comprises four isoforms: retina-specific visual arrestins (arrestin-1 and -4) and ubiquitously expressed non-visual arrestins (arrestin-2 and -3, commonly referred to as β-arrestin 1 and 2, or βarr1/2)5,6. Once viewed merely as signal terminators, arrestins are now recognized as versatile, spatiotemporal scaffolds that both terminate G protein signaling and initiate distinct pathways (MAPK/ERK, JNK, etc.) and perform noncanonical, GPCR-independent functions7,8,9,10,11,12,13,14. This duality underpins the concept of biased (functionally selective) signaling and motivates structure-guided pharmacology.The landmark crystal structure of the rhodopsin–arrestin-1 complex (PDB: 4ZWJ)15 provided the first definitive proof of the core-engaged state, revealing how the finger loop inserts into the receptor’s intracellular bundle to form a high-affinity interface and showing that arrestin engages the membrane directly through its C-edge. Importantly, earlier structural insights into phosphorylation-dependent β-arrestin activation were obtained from the crystal structure of βarr1 bound to a GPCR-derived phosphopeptide, which revealed how clustered phosphates engage basic pockets in the βarr1 N-domain to induce the active conformation16. Building on this framework, the concept of the “phosphorylation barcode”17,18,19,20 — distinct GPCR kinase (GRK)-mediated patterns on receptor C-terminal tails or intracellular loops — was subsequently structurally elucidated in the rhodopsin–arrestin-1 complex (PDB: 5W0P)21. This structure revealed the precise coordination of phosphoresidues by the arrestin-1 N-domain, demonstrating how these patterns regulate the binding geometry and activation.Building on these foundations, the recent “resolution revolution” in cryo-electron microscopy (cryo-EM) has further transformed the field. By utilizing lipid nanodiscs or conformation-selective stabilizers, researchers have moved beyond static snapshots to reveal that arrestin engagement is not a monolithic process but rather a diverse structural continuum dictated by receptor topology and the membrane context15,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41.This structural “multiverse” spans from the deep core conformation, which sterically blocks G proteins to drive desensitization, to the tail conformation29,42,43, which allows simultaneous G protein binding. The specific architecture is tuned by isoform-specific mechanics, such as the distinct C-edge membrane anchoring of βarr1 vs βarr244, and by the receptor class, with class B receptors capable of forming stable “megacomplexes” that drive sustained endosomal signaling42. Furthermore, novel “loop-engaged”33, “side-engaged”35, and membrane-anchored34 conformations have been identified in recent structures, indicating that arrestin can decode complex regulatory inputs to adopt highly specialized conformations.Mechanistically, arrestin recruitment and functional output are encoded by layered regulatory codes: phosphorylation barcodes (e.g., PxPP and PxxPxϕ) together with the receptor’s activation state determine the arrestin binding geometry, conformational activation, and complex stability16,21,45,46,47. Multivalent engagement of receptor-attached phosphates by arrestin phosphate-sensing residues (lysine/arginine) disrupts the interdomain polar core and the three-element autoinhibitory interaction, permitting the transition from the basal to the high-affinity, receptor-bound arrestin conformation16,48,49. This information is integrated with a complementary lipid code, in which membrane phosphoinositides such as phosphatidylinositol 4,5-bisphosphate (PIP2) and the cofactor inositol hexakisphosphate (IP6) allosterically orient and prime arrestin for receptor engagement and C-edge insertion50,51,52,53,54,55. Together, these regulatory layers define a continuum of arrestin microstates, each with distinct signaling capacities.These structural and mechanistic insights provide a molecular foundation for biased signaling, in which ligands preferentially stabilize receptor–G protein or receptor–β-arrestin microstates linked to therapeutic signaling while minimizing adverse effects56,57,58,59,60,61,62. Moreover, the recognition of megacomplex and compartmentalized endosomal signaling has reshaped our understanding of how internalized receptors contribute to physiological processes, from metabolism to neuroplasticity, motivating the development of strategies for generating spatially and temporally targeted therapies42,63,64,65,66,67,68.Together, these findings establish that GPCR–arrestin coupling spans a continuous structural and functional spectrum shaped by receptor topology, phosphorylation patterns, the lipid environment, and arrestin isoform-specific mechanics. Arrestin recruitment is therefore not a binary on–off switch but a conditional, allosterically regulated process that encodes distinct signaling outputs through discrete conformational microstates. This emerging structural multiverse provides the mechanistic foundation for biased signaling, endosomal signaling, and arrestin-mediated functions. In this review, we synthesize the results from recent high-resolution structural, biochemical, and mechanistic studies to map the diversity of GPCR–arrestin engagement modes, elucidate the molecular logic by which regulatory inputs are decoded, and highlight how these insights enable rational design of biased ligands, intracellular allosteric modulators (IAMs), and arrestin-directed therapeutic strategies.The structural landscape of GPCR–arrestin complexesThe transition from X-ray crystallography to cryo-EM has substantially advanced GPCR–arrestin structural biology, enabling the visualization of these dynamic complexes in near-native lipid environments and revealing that their coupling is not a monolithic process but a diverse continuum of engagement modes dictated by the receptor topology and isoform specificity.Capture of GPCR–arrestin complexesGPCR–arrestin complexes were historically more difficult to resolve than GPCR–G protein assemblies because they require receptor phosphorylation and exhibit pronounced structural heterogeneity and interdomain flexibility. Although early crystallographic studies established the foundational core conformation, recent cryo-EM studies have expanded this limited static snapshot into a much broader conformational landscape15,16,21,22,43,69. Methodological innovations have driven this progress, significantly expanding the structural repertoire23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40 (Table 1). Cryo-EM combined with lipid nanodiscs or lipid layers preserves the membrane contacts essential for arrestin binding25,44.Table 1 Structures of the GPCR‒arrestin complex.Full size tablePioneering stabilization strategies have been indispensable to overcome the profound inherent flexibility of these assemblies. These approaches include receptor engineering (e.g., appending a hyperphosphorylated surrogate C-tail, such as that of the vasopressin 2 receptor, V2R), direct receptor‒arrestin fusions, the use of pre-activated/truncated arrestins, and the application of conformationally selective binders (e.g., Nb3270, scFv3043, and Fab734) to trap transient intermediates. While these ingenious tools have profoundly advanced the field, their methodological caveats must be recognized. By design, surrogate tails, arrestin fusions, and rigid chaperones restrict natural interdomain flexibility to funnel the complex toward specific structural outcomes. Moreover, cryo-EM exquisitely captures the most stable, energetically favorable states within a broader dynamic continuum, providing critical high-resolution snapshots of discrete conformational states rather than the complete, transient spectrum of intermediates.Collectively, arrestins share a conserved bilobed architecture composed of N- and C-domains held in the inactive state by two key restraints: a polar core formed by a network of charged residues and a three-element interaction involving the arrestin C-terminal tail, β-strand I, and α-helix I16,48,49 (Fig. 1a–c). Engagement with phosphorylated GPCR tails or intracellular loops (especially intracellular loop 3 (ICL3)) disrupts these restraints, releasing the arrestin C-tail and inducing a characteristic interdomain rotation (~18°–25°)16,45,47,71. This rearrangement repositions critical functional elements, including the finger, lariat, middle, and C-loops, to enable receptor contact and membrane association27,43,44,72,73 (Fig. 1a). Importantly, arrestin activation is best described as conformationally heterogeneous and state-selective rather than strictly on–off.Fig. 1: Structural basis of arrestin activation and diversity of GPCR–arrestin interactions.Full size imagea Superposition of the structures of inactive βarr1 (PDB: 1G4M) and active V2R-bound βarr1 (PDB: 7R0C). Hallmark activation features of the coordinated rearrangement of central crest loops are highlighted. b, c Molecular restraints stabilizing arrestin in the inactive state. Rearrangement of the polar core network upon activation (b). Disruption of the three-element interaction (c) among the C-terminal tail, β-strand I, and α-helix I after activation. d Four distinct arrestin conformations observed in the GPR1–βarr1 complex (PDB: 9UYH, 9UYI, 9UYJ, and 9UYL), illustrating isoform-specific conformational heterogeneity and tilted engagement geometries. e Structural comparison of finger loop and C-edge loop configurations across distinct GPR1–βarr1 conformations, highlighting differences in membrane anchoring and engagement depth. f Comparison between GPR1–βarr1 (conformation 1) and GPR1–βarr2 (PDB: 9UYM). A cholesteryl hemisuccinate (CHS) molecule that mediates a cholesterol-dependent interaction between GPR1 and βarr2 is shown as sticks. g Representative class A GPCR–arrestin assemblies, including rhodopsin–arrestin-1 (left panel; PDB: 4ZWJ) and APJR–βarr1 dimeric complexes exhibiting asymmetric 2:1 (middle panel; PDB: 9KUX) and symmetric 2:2 (right panel; PDB: 9KUW) stoichiometries. h Class B1 GPCR–arrestin complexes, including the GCGR–βarr1 complex (left panel; PDB: 8JRV) and two distinct PTH1R–βarr1 conformations (middle and right panels; PDB: 9LXR and 9LXP), illustrating the dynamic equilibrium between tail-engaged and core-engaged architectures. i, j Comparison of global orientations (i) and local interface differences (j) between the PTH1R–βarr1 cryo-EM structures and an integrative model derived from live-cell crosslinking (PDB: 9A3I). The crosslinking-guided molecular dynamics model, which was built upon 136 validated intermolecular proximity points, captures highly flexible elements that are invisible in cryo-EM images and reveals a highly heterogeneous conformational ensemble. k Class C GPCR–arrestin complexes, including the mGlu8R–βarr1 dimer (top panel; PDB: 9MBA) and two distinct mGlu3R dimeric assemblies (bottom panels; PDB: 9II2 and 9II3), highlighting obligate dimerization, variable stoichiometry, and shallow peripheral arrestin engagement.Structural features of arrestin isoformsComparative structures show that arrestin recruitment is a logic-gated selection process: isoform mechanics, membrane coupling, and the receptor architecture together determine the geometry and function.Cryo-EM comparisons indicate that isoform selectivity stems from differences in the C-edge length, membrane anchoring, lipid dependence, and conformational flexibility — not a simple core vs tail dichotomy34,38,44,45,47. For example, βarr2 displays a significantly larger interdomain rotation (~23°–25°) than βarr1 (~18°–20°)45,47. In atypical chemokine receptor 3 (ACKR3), βarr1 forms a tightly membrane-anchored complex with deep finger-loop insertion and robust C-edge penetration, whereas βarr2 contacts the membrane via the finger loop but lacks stable C-edge anchoring because of a shorter loop34. Thus, both isoforms interact with the receptor tail and the membrane but differ in membrane coupling efficiency and conformational stabilization, leading to distinct scaffolding outcomes.In G protein-coupled receptor 1 (GPR1), isoform divergence is reversed in terms of conformational heterogeneity38. βarr1 has multiple binding conformations, including tilted, membrane-anchored states supported by its extended C-edge loops38 (Fig. 1d, e). In contrast, βarr2 adopts a single, well-defined core-engaged conformation with a non-membrane-anchored C-edge (Fig. 1f). Notably, the GPR1–βarr2 complex is stabilized by a cholesterol-mediated interface bridging ICL2 and βarr2 central loops (Fig. 1f), a feature that is absent in βarr1 complexes and selectively required for βarr2 recruitment and signaling38.Structural features of arrestin in distinct GPCRsBeyond isoform effects, GPCR class-specific architectures further dictate arrestin stoichiometry and geometry. Most resolved class A GPCRs typically support the formation of a 1:1 complex with variable degrees of finger-loop penetration15,21,23,24,25,26,27,28,30,31,33,34,35,36,38,39 (Fig. 1g). Structures of the apelin receptor (APJR) can assemble asymmetric 2:1 and symmetric 2:2 receptor–βarr1 complexes in which bridging or steric crowding favors tail-dominant or shallow-core engagement36 (Fig. 1g).Class B1 GPCRs exhibit dynamic β-arrestin engagement equilibria. The glucagon receptor (GCGR)‒βarr1 complex adopts a tail-engaged conformation29 (Fig. 1h). βarr1 stabilization relies on extensive C-tail phosphorylation and PIP2-mediated C-edge insertion, while the transmembrane domain (TMD) core remains accessible for simultaneous G protein binding29. This architecture permits megacomplex formation and sustained endosomal signaling. In contrast, parathyroid hormone 1 receptor (PTH1R)–βarr1 structures40 (PDB: 9LXR and 9LXP) reveal that class B1 receptors can stabilize both tail-engaged and core-engaged states (Fig. 1h). This structural plasticity enables both sustained endosomal signaling (tail state) and receptor desensitization (core state).In addition to cryo-EM snapshots, the conformational heterogeneity of the PTH1R–βarr1 interaction has been elucidated through orthogonal live-cell crosslinking74. When directly compared to cryo-EM structures, this crosslinking-derived model provides crucial complementary insights. While cryo-EM resolves distinct core-engaged states, the crosslinking data capture a dynamic conformational ensemble40,74 (Fig. 1i). Unbiased molecular dynamics simulations, which are guided by these extensive proximity constraints, confirm that this native ensemble dynamically bridges the discrete cryo-EM states through a 13º–25° fluctuation in the βarr1 orientation74. Furthermore, this approach revealed flexible elements that are intrinsically invisible in cryo-EM density maps, tracking highly mobile ICL3 and revealing unappreciated interfaces, such as the proximal phosphorylation cluster engaging the arrestin “N-edge” and extended interactions of the distal C-terminal tail with the N-domain40,74 (Fig. 1j). Together, integrating cryo-EM with in-cell conformational mapping provides a more complete and mechanistically robust understanding of PTH1R–βarr1 interactions.Class C GPCRs function as obligate dimers with stoichiometric versatility imposed by the dimer architecture. Unlike class A receptors, β-arrestin-bound metabotropic glutamate receptors (mGluRs) retain transmembrane helix 6 (TM6) in an inward, inactive-like conformation, thereby precluding deep arrestin core insertion32,37. As a result, arrestin engages mGluRs through a shallow, peripheral interface rather than through a canonical core pocket (Fig. 1k). The βarr1 finger loop occupies a cavity formed by TM3, ICL1, ICL2, and ICL3, while the primary binding energy is derived from extensive contacts with the phosphorylated C-tail. Metabotropic glutamate receptor subtype 3 (mGlu3R) supports the formation of an asymmetric 2:1 complex, in which a single βarr1 binds one protomer and allosterically stabilizes an asymmetric dimeric unit, whereas both mGlu3R and mGlu8R form symmetric 2:2 assemblies with one βarr1 bound to each protomer32,37. In both cases, βarr1 binding favors scaffolding-competent states while excluding the deep-core, desensitization-centric mechanisms characteristic of monomeric class A receptors.Engagement conformations: a dynamic continuumThe static duality of core- vs tail-engaged conformations has been superseded by the dynamic engagement continuum, where arrestin samples a landscape of metastable intermediates that tune the signaling output. Canonical core-engaged conformations (e.g., rhodopsin15,21,22, μ-opioid receptor (MOR)39, and PTH1R (PDB: 9LXR)40) involve deep finger-loop insertion to sterically block G proteins and drive desensitization (Fig. 2a), whereas the tail-engaged conformation (e.g., GCGR29) tethers β-arrestin primarily via the phosphorylated C-terminus and permits megacomplex formation and sustained signaling42 (Fig. 2b).Fig. 2: The structural continuum of GPCR–arrestin engagement conformations.Full size imagea Canonical core-engaged conformation. Representative structures of the core conformation are shown, including rhodopsin–arrestin-1 (PDB: 4ZWJ), MOR–βarr1 (PDB: 9WSV), and PTH1R–βarr1 (PDB: 9LXR) complexes. The receptor TMD core and C-terminal tail are highlighted by purple and red circles, respectively. Comparisons of intracellular arrestin orientations are shown on the right; the canonical state (O1, observed in rhodopsin) and the rotated state (O2) illustrate conformational plasticity within the core-engaged mode. b Tail-engaged conformation. The tail-engaged conformation is exemplified by the GCGR–βarr1 complex (PDB: 8JRV). Two perspectives highlight the exclusive tethering via the C-tail (top panel, red circle) and the unoccupied, sterically accessible TMD core (bottom panel, purple circle), which permits simultaneous G protein coupling. c Loop-engaged conformation. The loop-engaged conformation was captured in the NTSR1–βarr1–SBI-553 complex (PDB: 8ZYU). The remodeled ICL3 of NTSR1 (colored red and highlighted) inserts into the arrestin central crest, while the arrestin finger loop remains external to the receptor core. d Side-engaged conformation. The noncanonical side-engaged mode observed in the GPR52–βarr1 complex (PDB: 9IJR). Arrestin binds laterally via ICL1 and helix 8 without penetrating the receptor core. TM6 remains in an inactive-like conformation, supporting constitutive, agonist-independent arrestin recruitment. e Membrane-anchored conformation. The ACKR3–βarr1 complex (PDB: 9E82) is characterized by a membrane-anchored arrestin conformation. The arrestin finger loop (indicated by a red arrow) inserts directly into the membrane/micelle boundary rather than the receptor TMD core. Note that the TMD core (purple circle) remains unoccupied, while the complex is tethered by the C-tail (red circle).The loop-engaged conformation, which was first structurally defined in the neurotensin receptor 1 (NTSR1)–βarr1–SBI-553 complex33, represents a β-arrestin-selective receptor conformation (Fig. 2c, SBI-553 is a β-arrestin-biased allosteric modulator discussed in the section of “Structural mechanisms of biased signaling”). In this state, the βarr1 finger loop remains outside the transmembrane cavity, whereas the remodeled ICL3 loop of NTSR1 inserts into the central crest of βarr1 in a “hooking” manner. This architecture sterically excludes G proteins but preserves arrestin scaffolding, explaining modulator-driven β-arrestin bias.The side-engaged conformation is a distinct noncanonical mode observed in the orphan receptor G protein-coupled receptor 52 (GPR52)35 (Fig. 2d). βarr1 binds laterally via ICL1 and helix 8 of GPR52, without core penetration or the outward movement of TM6. The receptor remains in an inactive-like conformation stabilized by a conserved D3.49-R3.50-I5.62-A6.34 motif, enabling constitutive, agonist-independent βarr1 recruitment35. This mode functions as a regulatory mechanism rather than an activity-dependent switch.The membrane-anchored conformation, as observed in the ACKR3–βarr1 complex34, represents a unique variant of tail-type binding (Fig. 2e). Here, βarr1 is tethered to the phosphorylated C-terminus, yet its finger loop, which is rich in conserved hydrophobic residues, inserts into the lipid/micelle boundary rather than the receptor core34. Functionally, this lipid anchoring (involving both the finger loop and C-edge) promotes rapid capture dictated by the phosphorylation barcode, but importantly does not occlude the cytoplasmic cleft. This interaction leaves the TMD accessible for G protein or GRK access, establishing a membrane-coupled stabilized state that facilitates recruitment without enforcing canonical desensitization.Together, these structures establish that GPCR–arrestin coupling spans core, loop, side, membrane-anchored, and tail conformations, forming a continuous structural spectrum that decouples arrestin recruitment from receptor activation and provides a mechanistic framework for signaling bias.The logic of assembly: multi-dimensional modulationArrestin is not a passive adapter but a conditional signal integrator that decodes layered regulatory inputs into distinct functional outputs. The assembly of GPCR–arrestin complexes is specified by the coordinated actions of the GRK-written phosphorylation barcodes, lipid cofactors, and isoform-dependent structural features that bias membrane engagement and conformational stability, thereby shaping the arrestin microstate landscape.The expanded phosphorylation barcodeThe recruitment and activation of arrestin are governed by a sophisticated “phosphorylation barcode”17,18,19,20, a concept established by seminal work on the β2-adrenergic receptor (β2AR) and M3 muscarinic receptor (M3R). These studies demonstrated that distinct spatial phosphorylation patterns that are actively written by specific GRK isoforms function dually as a recruitment dock and an allosteric activation trigger. As described in the section of “Capture of GPCR–arrestin complexes”, multivalent phosphate engagement disrupts the polar core and three-element autoinhibitory interactions48,49, triggering the conformational switch required for productive receptor engagement16,45,47 (Fig. 1b, c). However, arrestin activation is not uniform; instead, the specific pattern and location of phosphorylation determine which functional conformation is stabilized. This pattern-dependent activation is validated by mass spectrometry-based phosphoproteomics, mutagenesis, and real-time BRET biosensors75,76,77,78.Zhou et al. fundamentally demonstrated that specific spatial arrangements of phosphates, rather than total negative charge, determine the fidelity of arrestin recruitment and activation through the X-ray structure21. Based on the crystal structure of the rhodopsin‒arrestin-1 complex (PDB: 5W0P) and extensive sequence analysis, they identified a recurrent consensus phosphorylation code, Px(x)PxxP/E/D, where “P” denotes a phosphorylated serine or threonine. In this mechanism, the precise spacing of phosphorylated residues or acidic amino acids (aspartate or glutamate) is positioned to engage discrete positively charged phosphate-sensor pockets in the arrestin-1 N-domain (including K15, K16, R30, K111, R172, and K301), displacing the arrestin-1 C-tail and initiating activation21 (Fig. 3a). In support of this phosphorylation barcode, the V2R C-terminus contains an extended PxxPxxP motif that tightly engages the βarr1 N-domain16 (Fig. 3b).Fig. 3: Diversity of phosphorylation barcodes in the GPCR C-terminus and ICL3.Full size imagea Classical C-terminal phosphorylation barcode “PxPxxP/E/D” in rhodopsin, as observed in the rhodopsin–arrestin-1 complex (PDB: 5W0P). b Classical C-terminal phosphorylation barcode “PxxPxxP/E/D” in V2R, as resolved in the V2Rpp–βarr1 complex (PDB: 4JQI). c, d The “PxPP” phosphorylation barcode located in the C-terminus of NTSR1 in the NTSR1‒βarr1 complex (PDB: 8ZYU) (c) and in the ICL3 of M2R in the M2Rpp‒βarr1 complex (PDB: 8JAF) (d). e Distinct “PxxPxϕ” phosphorylation barcodes in the proximal and distal C-tail of ACKR3 in the ACKR3–βarr1 complex (PDB: 9E82). The proximal barcode is preferentially phosphorylated by GRK5, whereas the distal barcode is selectively modified by GRK2. f Schematic illustration of proximal and distal phosphorylation barcodes in ACKR3 and CXCR4, highlighting differential GRK specificity and spatial organization (the figure was created with BioRender). g Structure of the NTSR1‒GRK2‒Gαq complex (PDB: 8JPC) showing the insertion of the N-terminal helix (αN) of GRK2 into the TMD core of NTSR1. KD, kinase domain; PHD, pleckstrin homology domain; RHD, regulatory G protein signaling homology domain. h Structure of the rhodopsin‒GRK1‒Fab6 complex (PDB: 7MTB) illustrating the insertion of the αN of GRK1 into the TMD core of rhodopsin.Full Px(x)PxxP/E/D motifs maximize binding affinity through multivalent electrostatic networks with the arrestin N-domain79,80,81. Structural and biochemical analyses of diverse GPCRs, including C-X-C chemokine receptor type 7 (CXCR7), have demonstrated that these canonical clusters coordinate tightly with the basic pockets of arrestin to dictate the conformation of the engaged receptor C-tail80,81. However, while extended tracks drive high-affinity assembly, they are not strictly required to encode specific functional outcomes. Indeed, even a single key phosphorylation site can act as a crucial structural anchor to direct interdomain rotation and regulate trafficking82. Recognizing this functional diversity, the structural focus of the field has naturally expanded to investigate how alternative phosphorylation barcodes also efficiently coordinate the activation of the complexes.Consequently, structural studies further refined the phosphorylation barcode concept. Structural analyses of β-arrestin complexes with C-C chemokine receptor type 5 (CCR5) and other GPCR phosphopeptides revealed that a PxPP/E/D motif45,47,83 in which two phosphorylated residues are separated by a single amino acid and followed by an additional phosphate or acidic residue acts as a privileged activating geometry when it is present in GPCR phosphotails (such as NTSR133) or ICL3 (such as M2 muscarinic receptor (M2R)45,47) (Fig. 3c, d). Structural snapshots of β-arrestin bound to diverse GPCR-derived phosphopeptides show that this PxPP pattern engages a conserved “K-K-R-R-K-K lock” (a network of positively charged residues) within the arrestin N-domain, promoting the displacement of the arrestin C-terminus45,47. However, the PxPP sequence is sufficient but not essential. Many GPCRs, including several atypical chemokine receptors (e.g., ACKR334), lack the PxPP sequence yet still recruit β-arrestin via alternative phosphorylation patterns that stabilize distinct active conformations. Thus, PxPP represents a high-efficiency barcode element embedded within a broader, combinatorial phosphorylation code in which the motif composition, spacing, and GRK-specific installation collectively tune the arrestin conformation and downstream signaling.Recent structural insights have expanded the classical phosphorylation code to include an alternative “hydrophobic anchor” motif that was identified in ACKR3 as x(pT)xx(pS/T)xϕ, where ϕ represents a medium-sized hydrophobic residue34 (Fig. 3e, f). In this model, β-arrestin activation is facilitated by the terminal hydrophobic residue docking into a complementary hydrophobic pocket on the arrestin surface, providing a critical anchoring point that stabilizes the active complex. A bioinformatic analysis revealed that while the minimal PxxP core is ubiquitous (~91% of class A GPCR C-tails, as recently revealed in GPR1 for the PxxE motif38), high-affinity β-arrestin engagement typically requires an additional “plus-one” element: either a third phosphate, as in the PxPP motif (~68%), or a hydrophobic anchor, as in PxxPxϕ (~63%)34. This expanded barcode repertoire allows arrestin to decode a broader spectrum of receptors using distinct chemical and stereochemical cues.The functional logic of these barcodes was illuminated by the mechanistic studies of Latorraca et al.84; by detailing the molecular effects of phosphorylation barcodes on βarr1, their work has key implications for arrestin function. Specifically, they demonstrated that arrestin binding and activation are separable processes encoded by distinct phosphorylation patterns. Some barcodes primarily act as binding codes, stabilizing arrestin at the receptor, whereas activation codes are required to drive interdomain rotation, finger-loop engagement, and other conformational changes required for arrestin-mediated signaling84. Consequently, strong binding does not necessarily equate to functional activation: a receptor may recruit β-arrestin without inducing downstream signaling or conversely induce partial or transient activation with limited binding, which is consistent with the catalytic activation characteristic of β-arrestin85. Importantly, barcode function depends less on the total number of phosphates than on their spatial arrangement and context, as individual sites can exert stimulatory or inhibitory effects depending on their combinatorial placement. Thus, GPCR phosphorylation operates as a combinatorial logic gate, independently tuning arrestin recruitment, conformational activation, and effector coupling.The modulation of GRKs: topological writers of the barcodePhosphorylation barcodes are not intrinsic properties of GPCRs alone but are actively written by distinct GRKs19,86. GRK isoform specificity determines the location, composition, and kinetics of receptor phosphorylation, thereby shaping arrestin recruitment, activation, and downstream functional outcomes. Accumulating biochemical, structural, and cellular evidence indicates that much of what is described as “arrestin bias” is, in practice, GRK-dependent barcode bias87,88,89. Structural work has shown how this process occurs in atomic detail: the structure of rhodopsin–GRK190 first established the mode of GRK engagement, and more recent structures reveal GRK2 assembling onto NTSR191, with its N-terminal helix inserting into the intracellular cavity to position the kinase domain for access to cytoplasmic phosphorylation sites — a process that is further promoted by the intracellular biased modulator SBI-553, demonstrating how ligands and modulators tune GRK engagement and thereby encode the phosphorylation barcode90,91 (Fig. 3g, h).Functionally, GRK2/3 are largely cytosolic and are recruited to activated receptors by Gβγ (and other adapters), which biases them toward phosphorylating more distal, high-density clusters on C-tails — a pattern associated with rapid β-arrestin binding and dynamic, internalization-prone complexes19,34,92,93. In contrast, GRK5/6 is more strongly associated with the membrane and preferentially phosphorylates proximal, site-specific motifs near the membrane, producing barcodes that stabilize long-lived, scaffolding-competent β-arrestin states19,34,92,93. These mechanistic contrasts underlie “barcode bias”: the same receptor can produce qualitatively different arrestin outcomes depending on which GRKs dominate. Furthermore, the use of a barcode system depends on the cellular context, as illustrated for ACKR3. Because it lacks canonical G protein coupling, ACKR3 cannot independently liberate Gβγ and thus relies primarily on GRK5-mediated phosphorylation92. However, it can dynamically recruit GRK2 through the release of Gβγ subunits during the co-activation of CXCR434,92. Through these mechanisms, the phosphorylation barcode extends beyond isolated intra-receptor regulation and instead functions as a sophisticated sensor of inter-receptor crosstalk and the broader cellular environment.Recent receptor-specific structures of ACKR334 and CXCR445,93 have provided structural proof of this principle (Fig. 3e, f). A comparative analysis of ACKR3 complexes revealed that the specific location of the phosphorylation barcode dictates the β-arrestin binding mode: proximal phosphorylation by GRK5 promotes robust C-edge membrane anchoring and stable finger-loop contacts with the lipid/micelle surface (PDB: 9E82), whereas GRK2-mediated phosphorylation of the distal C-tail yields a more transient interaction with minimal membrane penetration (e.g., PDB: 8TII)34. These distinct phospho-engagements alter the orientation of the finger loop of β-arrestin relative to that of the lipid micelle and modulate the stability of membrane anchoring. This mechanism directly links GRK identity to arrestin geometry and dwell time, determining the balance between stable scaffolding and transient trafficking. Hence, controlling GRK localization or activity is important for tuning arrestin-dependent signaling.Lipid cofactors as allosteric regulatorsWhile receptor phosphorylation patterns provide sequence-encoded signals for arrestin recruitment, its stabilization and activation are further modulated by a “lipid code” comprising membrane-anchored phosphoinositides and soluble inositol phosphates50,51,94,95,96. Together, these factors constitute a “lipid code” that functions not only as an anchor but also as an allosteric regulator that defines receptor class phenotypes and temporal signaling dynamics.PIP2, a membrane-bound lipid, plays a central role in regulating β-arrestin by binding basic patches on the arrestin C-lobe. Functional and structural analyses have demonstrated that PIP2 acts as an allosteric cofactor that stabilizes an active-like β-arrestin intermediate, increasing finger loop and gate loop dynamics and increasing reactivity toward active-state sensors such as Fab30, even in the absence of a GPCR50,94,95. By binding to the C-lobe basic patch, plasma membrane (PM) phosphoinositides (such as PI(4,5)P2 and PI(3,4,5)P3) — but notably not the lipid PI(3)P — lower the energetic barrier for GPCR core engagement and effectively shift the equilibrium toward an active state without requiring full C-tail displacement50. This lipid-dependent activation provides a molecular explanation for the distinction between class A and class B GPCRs, which are classified by their arrestin binding affinities and the duration of receptor–arrestin interactions97. Class A receptors typically require the coincident detection of phosphorylation and PIP2 for robust recruitment (confining activity to PIP2-rich PM domains), whereas class B receptors with strong phosphorylation barcodes can bypass the PIP2 requirement and sustain β-arrestin association in endosomes50.In addition to allosteric priming, Kuramoto et al. recently revealed that this regulation relies on a multivalent interaction; they identified a noncanonical PIP2 binding site at the tips of the β-arrestin C-domain that works synergistically with the known canonical site98. This multivalent binding precedes stable C-edge loop insertion into the lipid bilayer98. Together, these sequential processes ensure the formation of stable GPCR–β-arrestin complexes, driving their rapid accumulation into specialized membrane domains to promote efficient receptor desensitization and endocytosis.The stabilization of β-arrestin by PIP2 was verified by the GPCR–β-arrestin structures, such as NTSR1‒βarr124,33 and GCGR‒βarr129 (Fig. 4a, b). In the NTSR1–βarr1 complex, diC8-PtdIns(4,5)P2 acts as a molecular bridge: its 4,5-bisphosphate headgroup engages basic residues on the C-lobe of βarr1, while its lipid tail packs against the membrane-facing surfaces of TM1, TM2, and TM4 of NTSR124,33. In the GCGR–βarr1 complex, diC8-PtdIns(4,5)P2 adopts a related binding mode, with 4,5-bisphosphate similarly coordinating the βarr1 C-lobe, while an additional bridging phosphate interacts with residues in TM1, ICL1, and helix 8 of the GCGR, thereby reinforcing receptor–βarr1 coupling29.Fig. 4: Lipids as modulators of the arrestin structure and GPCR engagement.Full size imagea, b Binding modes of the phosphoinositide diC8-PtdIns(4,5)P2 within the C-edge of arrestin in the NTSR1–βarr1 complex (PDB: 8ZYU) (a) and GCGR–βarr1 complex (PDB: 8JRV) (b). c Comparative analysis of IP6 binding sites across the arrestin-1 (PDB: 7JTB), βarr1 (PDB: 1ZSH), and βarr2 (PDB: 5TV1) isoforms, highlighting the conserved polar pocket. d, e Structural characterization of the CHS binding site in the GPR1–βarr2 complex (PDB: 9UYM), showing an overall view of the lipid‒protein interface (d) and a detailed representation of the specific coordinating residues (e). f, g Structural characterization of the POA binding site in the GPR1–βarr1 complex (PDB: 9UYN), illustrating the global orientation of the fatty acid (f) and a close-up view of the hydrophobic interactions (g).While soluble diC8-PI(4,5)P2 provides vital structural insights, it inherently lacks the extended planar surface of a physiological lipid bilayer. This limitation prevents the simultaneous engagement of multiple membrane-binding elements distributed across both the N-domain and the C-domain of arrestin. Furthermore, highly curved environments (such as small-diameter nanodiscs) cause the arrestin C-edge in GPCR–β-arrestin complexes to “hang” without a stable membrane contact and destabilize the active complex25. To overcome these limitations, recent biophysical studies utilizing extended flat lipid nanodiscs have revealed that the two β-arrestins deploy distinct membrane-anchoring mechanisms99. Basal βarr1 selectively targets PI(4,5)P2-enriched nanodiscs via its C-edge with high affinity, shifting to finger-loop engagement upon activation to support further stabilization. In contrast, βarr2 binds more broadly to various anionic lipids (PI(4,5)P2, POPS, and POPG)99. Although the lipid bilayer alone is insufficient to fully activate either isoform, it synergistically increases the recruitment and activation of both β-arrestins in the presence of a phosphorylated GPCR C-tail. Ultimately, the lipid environment serves as a crucial, isoform-specific modulator that finely tunes arrestin recruitment and activation dynamics.In addition to membrane lipids, soluble IP6 functions as an isoform-specific allosteric regulator that differentially modulates arrestin conformational equilibria and oligomeric states. In βarr2, IP6 functions as a potent structural activator; by targeting the N-domain phosphate sensor to mimic a phosphorylated receptor C-tail, it disrupts the central polar core and induces the interdomain twist to stabilize active trimers52 (Fig. 4c). In contrast, IP6 serves primarily as a conformational priming agent for arrestin-1. Here, IP6 binding displaces the autoinhibitory C-tail but leaves the polar core intact, dissociating basal oligomers into functional monomers without inducing active twisting55,100,101 (Fig. 4c). Finally, IP6 serves as a negative regulator of βarr1. Through a unique two-site binding mechanism, IP6 drives the formation of inactive cytosolic βarr1 oligomers101,102 (Fig. 4c). By promoting this oligomeric state, IP6 effectively sequesters the protein in the cytoplasm, preventing free βarr1 monomers from translocating to the nucleus or interacting with PM receptors102.Finally, specific interfacial sterols and lipids can act as structural cofactors to tune isoform selectivity and constitutive activity38. Recent cryo-EM analyses of the GPR1‒βarr2 complex revealed that cholesterol wedges between ICL2 of the receptor and the central loops of βarr2, strengthening receptor‒βarr2 binding via extensive hydrophobic contacts38 (Figs. 1f and 4d, e). Furthermore, the endogenous fatty acid palmitoleic acid (POA) has been shown to be involved in constitutive recruitment. Structural data revealed that binding of the POA to GPR1 enhances agonist-independent capture of βarr138 (Fig. 4f, g).Structural mechanisms of biased signalingThe G protein and arrestin pathways constitute two fundamentally different modes through which GPCRs transduce extracellular information3. G proteins rapidly relay signals to classical second-messenger cascades, resulting in biochemical responses103,104; arrestins then lead to the desensitization of G protein signaling and the internalization of the receptors and simultaneously act as scaffolds for distinct downstream modules7,8,9,10,11,12,13,14,105,106. Mechanistically, the receptor-induced conformational activation of β-arrestin exposes binding interfaces that scaffold downstream kinase cascades5,13,107,108. This structural rearrangement enables the physical engagement of SH3 domains within Src family kinases and the coordinated activation of mitogen-activated protein kinase (MAPK) pathways, such as the c-Raf-MEK1-ERK1/2 axis and the JNK cascade11,108,109,110,111,112,113,114,115. By spatially restricting these signaling complexes, β-arrestin translates receptor engagement into precise, G protein-independent physiological responses11,107,108,116.The functional divergence of these two signaling branches — G protein-driven second messengers vs β-arrestin-mediated desensitization and kinase scaffolding — provides the biological foundation for biased agonism57,117. The concept of biased signaling captures the observation that chemically distinct ligands acting at the same receptor can differentially engage these outputs, namely, the ability of specific ligands to selectively stabilize receptor conformations that activate one cytosolic transducer (e.g., G protein or β-arrestin) while silencing the other58,118,119,120,121 (Fig. 5a). This paradigm shift offers a route to dissociate therapeutic efficacy from on-target side effects, a concept that has now been validated by high-resolution structural biology.Fig. 5: Structural basis of biased GPCR signaling toward G proteins or β-arrestins.Full size imagea Conceptual schematic illustrating ligand-directed signaling bias, in which GPCR ligands preferentially stabilize G protein-biased, β-arrestin-biased, or balanced signaling states, leading to distinct downstream signaling outputs. b Illustration of the Gs-biased conformation stabilized by the ligand PCO371 in the PTH1R‒Gs complex (PDB: 8JR9). c Detailed comparison of the Gs-bound state and βarr1-bound states of PTH1R, highlighting steric clashes between PCO371 and the βarr1 finger loop in both State 1 (top panel; PDB: 9LXR) and State 2 (bottom panel; PDB: 9LXP). d The canonical core-engaged conformation of the NTSR1–βarr1 complex stabilized by the biased allosteric modulator SBI-553 (PDB: 8ZYT). e Structural alignments of the SBI-553-bound NTSR1–βarr1 complex (PDB: 8ZYT) with the canonical NTSR1–Gq (top panel; PDB: 8FMZ) and NTSR1–Go (bottom panel; PDB: 8FN0) complexes. This superposition shows that the binding position of SBI-553 sterically clashes with the incoming Gα α5-helix, effectively blocking G protein coupling while seamlessly accommodating the βarr1 finger loop. f The loop-engaged conformation of the NTSR1–βarr1 complex stabilized by SBI-553 (PDB: 8ZYU). In this unique state, the arrestin finger loop is entirely excluded from the receptor core; instead, ICL3 of the receptor docks directly into the cavity of the βarr1 central crest. g Comparison of the loop-engaged NTSR1–βarr1 complex (PDB: 8ZYU) with the canonical NTSR1–Gq (top panel) and NTSR1–Go (bottom panel) complexes. The superposition illustrates that SBI-553 occupies the intracellular cavity to sterically occlude G protein coupling, forcing arrestin to engage via a distinct interaction interface mediated by ICL3 of the receptor. h Structural role of TM1 in determining signaling specificity for MOR. Superposition and comparison of the MOR‒Gz (PDB: 9WST), MOR‒Gi (PDB: 8EFQ), and MOR‒βarr1 (PDB: 9WSV) complexes highlight the conformational differences in TM1.The first major structural breakthrough in defining the atomic basis of bias arose from the study of class B1 GPCRs, specifically PTH1R bound to the small molecule PCO371 and the Gs protein122,123. Unlike endogenous peptide agonists that bind the extracellular domain, PCO371 engages a conserved intracellular pocket formed by TM2, TM3, TM6, TM7, and helix 8122 (Fig. 5b). This intracellular binding site acts as an allosteric wedge. By occupying the G protein coupling cavity directly, PCO371 stabilizes a distinct outward rotation of TM6 that is optimal for Gs activation but sterically incompatible with the broad interdomain interface required for β-arrestin recruitment (Fig. 5c). This finding proves that bias can be engineered by targeting conserved intracellular cavities to physically enforce a specific transducer-coupling geometry.With advancements in complex stabilization, the field subsequently resolved the mechanism of β-arrestin bias. The landmark structure of the NTSR1 complexed with the IAM SBI-553 provided the definitive mechanism for β-arrestin bias33,91. Mechanistically, SBI-553 achieves profound bias by occupying the intracellular transmembrane cavity to sterically occlude G protein coupling while simultaneously accommodating βarr1 recruitment through two distinct spatial modes33,124 (Fig. 5d–g). In the canonical core-engaged conformation, SBI-553 allows the βarr1 finger loop to seamlessly co-occupy the receptor pocket (Fig. 5d, e). Strikingly, SBI-553 also stabilizes an unprecedented loop-engaged conformation. In this state, the βarr1 finger loop does not insert into the receptor core; instead, ICL3 of the receptor docks directly into the cavity of the βarr1 central crest (Fig. 5f, g). This interface is created by significant conformational remodeling of the ICL3–TM6 and TM1–ICL1 regions, which form extensive new contacts with the βarr1 loops. Notably, a structural comparison with canonical complexes23,24 suggests that this loop engagement is not a naturally occurring intermediate but a synthetic state induced specifically by a biased allosteric modulator to bypass the steric constraints of the drug-occupied core.Most recently, structural efforts have redefined the understanding of orthosteric bias by mapping the complete effector landscape of the MOR39. Structures of the MOR complexed with its distinct transducers — Gi125, Gz, and βarr139 — reveal that signaling selectivity arises from conformational preferences encoded outside the canonical intracellular coupling pocket. Specifically, TM1 has been identified as a previously underappreciated allosteric regulator of this bias (Fig. 5h). A structural analysis demonstrated that TM1 acts as a transmission gear for transducer selectivity through distinct conformational dynamics: outward TM1 movement promotes the structural flexibility necessary for G protein coupling, whereas inward TM1 rearrangement stabilizes the intracellular binding pocket to facilitate βarr1 recruitment39. Residues within the newly identified “TM1 fusion pocket” (specifically Y751.39 and N861.50) modulate ligand binding and induce conformational fluctuations that propagate to the intracellular TM2–TM7–helix 8 interface, thereby differentially modulating the coupling efficiency of the G protein vs β-arrestin39. This mechanism establishes TM1 not as a static scaffold but as a dynamic allosteric switch driven by thermodynamic heterogeneity, potentially linking biased signaling to oligomerization interfaces where TM1 frequently governs receptor assembly.Noncanonical endosomal signaling and megacomplex formationNoncanonical endosomal signalingFor decades, the “canonical” model suggested that GPCR signaling is strictly confined to the PM, with arrestin recruitment functioning primarily as a termination mechanism that uncouples G proteins and drives internalization. This paradigm was revised by the discovery of endosomal signaling, which was first visualized in the pioneering works of Calebiro et al.63 and Ferrandon et al.126, who utilized Förster resonance energy transfer (FRET)-based biosensors to demonstrate persistent cAMP generation from internalized thyroid-stimulating hormone receptor and PTH1R. Endosomes serve as discrete signaling compartments that generate “cAMP waves” distinct from those generated by PM signaling (Fig. 6a). Endosomal signaling is now recognized as a general feature of multiple GPCRs, including β2AR127 and V2R128, with important implications for signal duration, location, and specificity129,130,131. Spatial encoding enables cells to distinguish transient PM-derived signals from sustained endosomal signaling programs, thereby selectively engaging gene expression pathways that are inaccessible to acute PM signaling132,133.Fig. 6: Endosomal GPCR signaling and the GPCR–G protein–β-arrestin megacomplex.Full size imagea Schematic illustration of biphasic cAMP signaling, consisting of a rapid, transient phase initiated at the plasma membrane followed by a sustained signaling phase originating from endosomes after receptor internalization (created with BioRender). b Overall architecture of the β2V2R‒Gs‒βarr1 megacomplex. The composite model is assembled from the β2AR‒Gs‒Nb35 (PDB: 6NI3) and V2R_tail‒βarr1‒Nb32‒Fab30 (PDB: 6NI2) complexes, illustrating the simultaneous engagement of the G protein and arrestin by a single receptor. c Close-up view of the Gαs α5-helix inserted into the TMD of β2V2R, highlighting canonical G protein coupling within the megacomplex configuration. d Detailed view of the interaction of the phosphorylated V2R C-tail with βarr1, demonstrating tail-mediated arrestin engagement that coexists with active G protein coupling and underlies sustained endosomal signaling. e Overall architecture of the atazanavir-stabilized β2AR_3muts‒Gs‒βarr1 megacomplex (PDB: 9L8L), demonstrating that simultaneous coupling is facilitated by the small molecule atazanavir. f Close-up of canonical Gαs α5-helix engagement within the β2AR core. g Detailed view of the atazanavir-mediated interface between β2AR_3muts and βarr1.Among these receptors, PTH1R represents the most mechanistically complete model of endosomal signaling, integrating the ligand residence time, β-arrestin scaffolding, G protein coupling, and spatial signal termination134,135,136,137,138,139. PTH1R signaling proceeds through a biphasic cAMP program: a rapid PM phase followed by a sustained endosomal phase that is essential for downstream transcriptional and physiological responses134,135,136.Following ligand binding at the PM, PTH1R activates Gs, inducing an acute increase in cAMP levels. Almost immediately, GRK-mediated phosphorylation of the receptor promotes β-arrestin recruitment and internalization while stabilizing a signaling-competent receptor state rather than enforcing immediate desensitization. Crucially, a long ligand residence time allows PTH1R to remain active during endocytosis, sustaining cAMP production from early endosomes126. Within endosomes, β-arrestin not only coexists with G protein but also actively scaffolds signaling complexes that support continued cAMP production135,140. In particular, β-arrestin-bound PTH1R can engage Gβγ subunits, forming a PTH1R–β-arrestin–Gβγ signaling node that contributes to downstream signaling and reinforces sustained cAMP generation140. This β-arrestin-mediated organization explains how endosomal PTH1R signaling remains robust even when PM G protein activation has terminated.While sustained endosomal signaling is critical for physiological responses, cells employ multiple mechanisms to actively terminate this signaling and prevent pathological overactivation. Local cAMP activates PKA, which phosphorylates signaling components to decrease receptor activity, whereas v-ATPase-driven endosomal acidification gradually destabilizes ligand–receptor interactions138. Finally, signal termination is executed by the retromer complex, which selectively sorts PTH1R from signaling endosomes into recycling pathways, resulting in the concomitant cessation of cAMP production and the establishment of the retromer complex as a molecular off-switch for sustained GPCR signaling139.The formation of a megacomplex: a structural basis for endosome signalingA key structural prerequisite for sustained endosomal signaling is the megacomplex (or super-complex), a ternary assembly of a GPCR, G protein, and β-arrestin. Conceptually, megacomplexes resolve the long-standing paradox of how G protein signaling can persist despite β-arrestin engagement. While the tail-engaged conformation, which leaves the TMD accessible to the G protein, is now recognized as a permissive architectural feature for megacomplex formation, the precise geometry and stability of GPCR–β-arrestin–G protein assemblies were initially unclear29,130.Direct evidence for megacomplex formation was first provided using a β2V2R chimera (β2AR fused to the V2R C-terminal tail)42,64. Thomsen et al. biochemically demonstrated that a single receptor species can form a ternary assembly with both heterotrimeric G protein and β-arrestin and mediate sustained G protein signaling in cells; notably, they presented representative negative-stain EM images of the formation of megacomplexes64. Subsequently, Nguyen et al. resolved the structure of the β2V2R‒Gs‒βarr1 megacomplex, revealing that the high-affinity phosphotail of V2R locks β-arrestin into a specific tail-engaged conformation while the intracellular core remains fully accessible to the Gs heterotrimer42 (Fig. 6b–d). This engagement is driven by a cluster of six specific phosphorylated residues on the V2R tail (pS357, pT359, pT360, pS362, pS363, and pS364), which form extensive electrostatic networks with conserved lysine/arginine clusters on the βarr1 N-domain42 (Fig. 6d). Crucially, molecular dynamics simulations further revealed that this “hanging” geometry is stabilized by the C-terminal loops of βarr1 (Fig. 6b), which make transient membrane contacts that orient βarr1 upright and prevent core engagement, thereby leaving the intracellular receptor cavity accessible for Gs coupling (Fig. 6c).Although early structures relied on chimeric phosphotails, the recent cryo-EM structure of the β2AR_3muts (K6.35R, G6.38S, and I6.39V)‒Gs‒βarr1 complex has redefined this paradigm by demonstrating that class A GPCRs can also form megacomplexes under appropriate allosteric conditions68 (Fig. 6e–g). Unlike tail-engaged models dominated by N-lobe engagement, this assembly adopts a C-lobe-dominant binding mode that is independent of the phosphorylated receptor C-tail. Here, βarr1 associates with the receptor primarily through C-lobe membrane anchoring, which is stabilized by the allosteric modulator atazanavir bridging the TM6/7 region of the receptor and the βarr1 197-loop68 (Fig. 6g). Structurally, βarr1 adopts a vertical “pendulum” conformation that occupies the interfacial groove between the Gα and Gβγ subunits, allowing it to align compatibly with the heterotrimer across multiple nucleotide-binding states (Fig. 6e). Conceptually, this structure identifies atazanavir as a molecular glue that converts β-arrestin from a competitive inhibitor of G protein coupling into a cooperative scaffold, highlighting a therapeutic strategy in which allosteric ligands stabilize noncanonical β-arrestin states to sustain signaling.Physiological imperatives and therapeutic frontiersThe elucidation of the GPCR–arrestin structural multiverse has reframed arrestins as central organizers of physiological signaling rather than passive terminators of G protein activity. Across tissues, arrestins function as tunable, spatiotemporal integrators that coordinate the signal strength, duration, and localization107,141. Consequently, physiological homeostasis depends not only on receptor activation but also on the precise engagement, conformation, and lifetime of GPCR–arrestin complexes. The disruption of this balance through altered phosphorylation patterns, impaired arrestin recruitment, or mislocalized signaling emerges as a unifying mechanism underlying diverse human diseases142,143,144.Clinical manifestations of an arrestin–G protein imbalanceMany clinical pathologies arise from distorted coupling between G proteins and arrestins rather than simply from receptor hyperactivity. In cardiovascular disease, chronic catecholamine exposure drives excessive beta-1 adrenoceptor (β1AR) phosphorylation, leading to receptor downregulation and maladaptive cardiac remodeling145. The clinical efficacy of carvedilol exemplifies how the selective engagement of β-arrestin microstates can restore balance: by suppressing toxic Gs/cAMP signaling while preserving a cardioprotective β-arrestin-EGFR-ERK axis, carvedilol improves outcomes despite attenuated canonical signaling146,147,148.In the nervous system, the MOR highlights a different dimension of imbalance. Here, G protein signaling mediates analgesia, whereas βarr2 recruitment has been implicated in the adverse effects of opioids, including respiratory depression, constipation, and tolerance149,150. Importantly, accumulating evidence suggests that these outcomes do not arise simply from β-arrestin recruitment per se but rather from the stabilization of specific β-arrestin conformations and signaling complexes downstream of MOR151,152,153. Therapeutic strategies that indiscriminately suppress arrestin engagement may therefore be suboptimal, highlighting the need for ligands that selectively modulate arrestin signaling states rather than globally eliminating arrestin coupling.Endocrine disorders further illustrate how the loss of arrestin regulation disrupts physiological timing. Skeletal homeostasis depends on tightly controlled PTH1R signaling. Constitutively active receptor mutations (Jansen’s metaphyseal chondrodysplasia, H223R154) drive persistent Gs activation, whereas C-terminal truncations (Eiken syndrome, R485X155,156) abolish the phosphorylation barcode required for β-arrestin recruitment. In both cases, a failure to engage β-arrestin-mediated desensitization traps PTH1R at the PM, resulting in unchecked surface cAMP signaling. Together, these examples illustrate a shared principle: disease emerges from the inappropriate persistence, localization, or geometry of GPCR–arrestin signaling complexes.Refining therapeutic inputs: from ligand bias to structural controlEarly efforts to exploit biased signaling focused on orthosteric ligands designed to favor either G protein or arrestin pathways. While conceptually transformative, this strategy has proven difficult to translate to the clinic. The G protein-biased MOR agonist oliceridine (TRV130) was developed to retain analgesic effects while reducing β-arrestin-associated side effects157. Although oliceridine successfully minimized β-arrestin recruitment, its clinical benefit was more modest than anticipated, suggesting that biased signaling outcomes may be context dependent and likely rely on the precise modulation of specific β-arrestin microstates rather than global pathway suppression158,159,160. Similarly, the β-arrestin-biased angiotensin II type 1 receptor (AT1R) ligand TRV027 showed promise in preclinical models but failed to outperform the placebo in phase II trials of patients with heart failure, underscoring the challenge of extrapolating system-level bias across species and disease contexts161,162,163. These outcomes highlight a key limitation of first-generation biased ligands: they modulate signaling indirectly from the extracellular orthosteric site, lacking precise control over the intracellular transducer interface. However, the clinical efficacy of carvedilol in heart failure patients demonstrates that therapeutic success extends beyond simple binary pathway selection. By simultaneously antagonizing Gs/cAMP signaling and stimulating cardioprotective β-arrestin signaling148, carvedilol exemplifies the need for nuanced signal integration. Consequently, rational drug design should evolve from empirical pathway suppression toward precise structural control, directly targeting the intracellular interface to stabilize specific, therapeutically beneficial receptor–transducer microstates.A decisive conceptual advance has emerged with the development of IAMs, facilitated by the growing number of resolved structures depicting GPCR–transducer interaction interfaces. Unlike orthosteric ligands, IAMs act directly on the intracellular pocket to enforce defined coupling geometries. The small molecule PCO371 binds within the intracellular cavity of PTH1R, stabilizing a Gαs-compatible conformation while sterically hindering β-arrestin engagement, thereby sustaining signaling and disfavoring internalization122,123. Conversely, SBI-553 occupies the Gαq binding pocket of NTSR1, simultaneously blocking G protein coupling and inducing a noncanonical β-arrestin loop-engaged conformation24,33,91. While conventional allosteric modulators can drive signaling bias by inducing receptor-wide conformational changes similar to those induced by orthosteric agonists, these distinct IAMs show that signaling bias can also be dictated directly at the transducer interface.In addition to modulating the receptor conformation, recent discoveries reveal that β-arrestin itself can be directly reprogrammed. The atazanavir-stabilized GPCR–G protein–βarr1 megacomplex exemplifies a new therapeutic paradigm in which small molecules function as molecular glues68. By bridging the βarr1 C-lobe to the receptor, atazanavir converts βarr1 from a competitive antagonist of G protein coupling into a cooperative signaling scaffold. Unlike traditional biased agonists, which struggle to completely abolish β-arrestin recruitment, this noncanonical “pendulum” conformation allows G protein and β-arrestin to coexist on the receptor68. By circumventing the steric competition that typically drives receptor desensitization, this molecular glue strategy sustains therapeutic G protein signaling even in the presence of β-arrestin, offering a direct mechanistic solution to drug tolerance.Together, intracellular modulators and molecular glues extend the structural multiverse framework from descriptive biology to actionable pharmacology. Rather than targeting receptors alone, next-generation therapeutics can selectively sculpt arrestin microstates, manipulate the signaling geometry, and rewire the spatiotemporal signaling logic. This shift from ligand bias to structural control of transducer assemblies positions arrestin-centered pharmacology as a powerful frontier for achieving durable efficacy with reduced side effects.Concluding remarks and future directionsAdvances in cryo-EM and complementary structural approaches have fundamentally reshaped our understanding of GPCR–arrestin coupling. What was once viewed as a simple binary interaction — receptor on, arrestin off — has now emerged as a continuous, context-dependent spectrum of conformational states. High-resolution structures captured in native-like membrane environments reveal that arrestin can engage GPCRs through deep-core insertion, shallow tail tethering, and a range of intermediate loop-, side-engaged, and membrane-anchored conformations. Collectively, these architectures establish that arrestin recruitment is not exclusively synonymous with signal termination but instead encodes graded functional outputs through discrete conformational microstates.A central conclusion of this review is that arrestin functions as a programmable signal integrator rather than a passive adapter. Arrestin isoforms decode identical phosphorylation inputs differently through variations in the C-edge architecture, membrane anchoring, lipid sensitivity, and conformational plasticity, yielding isoform-selective scaffolding outcomes. The GPCR class further constrains this decoding logic: class A receptors favor dynamic 1:1 assemblies that toggle between desensitizing and scaffolding states; class B1 receptors stabilize tail-engaged conformations compatible with megacomplex formation and sustained endosomal signaling; and obligate dimeric class C receptors enforce shallow, peripheral arrestin interfaces that bias toward scaffolding rather than steric G protein exclusion. Thus, arrestin signaling emerges from the intersection of receptor topology, isoform mechanics, and cellular context.The phosphorylation barcode provides a dominant, but not exclusive, layer of control over arrestin engagement. The motif composition and spatial arrangement — rather than merely the absolute number of phosphates — govern arrestin activation by selectively disrupting autoinhibitory restraints and promoting interdomain rearrangements. These barcodes are actively written by GRKs, whose isoform-specific localization and kinetics determine whether arrestin engagement is transient or long-lived. Superimposed on this code is a lipid-dependent regulatory layer in which membrane phosphoinositides and soluble inositol phosphates act as allosteric cofactors that prime, stabilize, and spatially restrict arrestin activation. Together, phosphorylation and lipid codes define a multi-dimensional regulatory logic that governs arrestin microstate selection.These structural principles provide a mechanistic foundation for biased signaling that extends beyond ligand efficacy alone. Biased agonism arises from selective stabilization of receptor–transducer geometries, often through subtle rearrangements outside the canonical transducer-binding pocket. IAMs and molecular glue exemplify this concept by directly reshaping the transducer interface and converting β-arrestin from a competitive terminator into a cooperative signaling scaffold. These mechanisms underscore that bias is a geometrically encoded property of the receptor–transducer complex, not merely a pharmacological abstraction.Spatial regulation represents a further critical dimension of arrestin biology. The discovery of GPCR–G protein–β-arrestin megacomplexes and sustained endosomal signaling challenges the long-standing view that arrestin solely terminates G protein signaling at the PM. Instead, β-arrestin can stabilize G protein signaling-competent complexes within endosomes, encoding temporal and spatial information essential for transcriptional and physiological responses. How distinct β-arrestin microstates are selectively generated, maintained, and terminated across cellular compartments remains a central unanswered question.Looking forward, several priorities emerge for the field. First, capturing arrestin dynamics requires integrating time-resolved cryo-EM to capture millisecond intermediates with single-molecule FRET and molecular dynamics simulations spanning microsecond-to-second timescales. Second, expanding structural coverage to underexplored GPCR classes, orphan receptors, and βarr2-specific complexes will test the generality of current models. Third, linking defined structural microstates to in vivo physiology remains a major challenge; a combination of structure-guided perturbations with genetically encoded biosensors and spatially resolved signaling assays is needed to bridge this gap. Finally, the translation of the structural logic of arrestins into therapeutics will require precision strategies that target not only receptors but also GRKs, lipids, and arrestin interfaces themselves.In summary, arrestins should now be viewed as programmable, spatiotemporal integrators of GPCR signaling that operate across conformational, biochemical, and spatial dimensions. Decoding how regulatory inputs sculpt arrestin microstates provides a unifying framework for interpreting GPCR physiology and a roadmap for designing next-generation therapeutics that exploit signaling bias, endosomal signaling, and arrestin-directed modulation with unprecedented precision.ReferencesPierce, K. L., Premont, R. T. & Lefkowitz, R. J. Seven-transmembrane receptors. Nat. Rev. Mol. Cell Biol. 3, 639–650 (2002).Article  CAS  PubMed  Google Scholar Lorente, J. S. et al. GPCR drug discovery: new agents, targets and indications. Nat. Rev. Drug Discov. 24, 458–479 (2025).Article  CAS  PubMed  Google Scholar Hilger, D., Masureel, M. & Kobilka, B. K. Structure and dynamics of GPCR signaling complexes. Nat. Struct. Mol. Biol. 25, 4–12 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Liu, S., Anderson, P. J., Rajagopal, S., Lefkowitz, R. J. & Rockman, H. A. G protein-coupled receptors: a century of research and discovery. Circ. Res. 135, 174–197 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Gurevich, V. V. & Gurevich, E. V. The structural basis of arrestin-mediated regulation of G-protein-coupled receptors. Pharmacol. Ther. 110, 465–502 (2006).Article  CAS  PubMed  PubMed Central  Google Scholar Gurevich, V. V. & Gurevich, E. V. GPCR signaling regulation: the role of GRKs and arrestins. Front. Pharmacol. 10, 125 (2019).Article  CAS  PubMed  PubMed Central  Google Scholar Lohse, M. J., Benovic, J. L., Codina, J., Caron, M. G. & Lefkowitz, R. J. beta-Arrestin: a protein that regulates beta-adrenergic receptor function. Science 248, 1547–1550 (1990).Article  CAS  PubMed  Google Scholar Ferguson, S. S. et al. Role of beta-Arrestin in mediating agonist-promoted G protein-coupled receptor internalization. Science 271, 363–366 (1996).Article  CAS  PubMed  Google Scholar Lefkowitz, R. J. Arrestins come of age: a personal historical perspective. Prog. Mol. Biol. Transl. Sci. 118, 3–18 (2013).Article  CAS  PubMed  Google Scholar Daaka, Y. et al. Essential role for G protein-coupled receptor endocytosis in the activation of mitogen-activated protein kinase. J. Biol. Chem. 273, 685–688 (1998).Article  CAS  PubMed  Google Scholar Luttrell, L. M. et al. Activation and targeting of extracellular signal-regulated kinases by beta-Arrestin scaffolds. Proc. Natl. Acad. Sci. USA 98, 2449–2454 (2001).Article  CAS  PubMed  PubMed Central  Google Scholar Lefkowitz, R. J. & Shenoy, S. K. Transduction of receptor signals by beta-Arrestins. Science 308, 512–517 (2005).Article  CAS  PubMed  Google Scholar Shukla, A. K., Xiao, K. & Lefkowitz, R. J. Emerging paradigms of beta-Arrestin-dependent seven transmembrane receptor signaling. Trends Biochem. Sci. 36, 457–469 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Gurevich, V. V. & Gurevich, E. V. The new face of active receptor bound arrestin attracts new partners. Structure 11, 1037–1042 (2003).Article  CAS  PubMed  Google Scholar Kang, Y. et al. Crystal structure of rhodopsin bound to arrestin by femtosecond X-ray laser. Nature 523, 561–567 (2015).Article  CAS  PubMed  PubMed Central  Google Scholar Shukla, A. K. et al. Structure of active beta-Arrestin-1 bound to a G-protein-coupled receptor phosphopeptide. Nature 497, 137–141 (2013).Article  CAS  PubMed  PubMed Central  Google Scholar Ren, X. R. et al. Different G protein-coupled receptor kinases govern G protein and beta-Arrestin-mediated signaling of V2 vasopressin receptor. Proc. Natl. Acad. Sci. USA 102, 1448–1453 (2005).Article  CAS  PubMed  PubMed Central  Google Scholar Kim, J. et al. Functional antagonism of different G protein-coupled receptor kinases for beta-Arrestin-mediated angiotensin II receptor signaling. Proc. Natl. Acad. Sci. USA 102, 1442–1447 (2005).Article  CAS  PubMed  PubMed Central  Google Scholar Nobles, K. N. et al. Distinct phosphorylation sites on the beta(2)-adrenergic receptor establish a barcode that encodes differential functions of beta-Arrestin. Sci. Signal. 4, ra51 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Butcher, A. J. et al. Differential G-protein-coupled receptor phosphorylation provides evidence for a signaling bar code. J. Biol. Chem. 286, 11506–11518 (2011).Article  CAS  PubMed  Google Scholar Zhou, X. E. et al. Identification of phosphorylation codes for arrestin recruitment by G protein-coupled receptors. Cell 170, 457–469.e13 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Zhou, X. E. et al. X-ray laser diffraction for structure determination of the rhodopsin-arrestin complex. Sci. Data 3, 160021 (2016).Article  CAS  PubMed  PubMed Central  Google Scholar Yin, W. et al. A complex structure of arrestin-2 bound to a G protein-coupled receptor. Cell Res. 29, 971–983 (2019).Article  CAS  PubMed  PubMed Central  Google Scholar Huang, W. et al. Structure of the neurotensin receptor 1 in complex with beta-Arrestin 1. Nature 579, 303–308 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Staus, D. P. et al. Structure of the M2 muscarinic receptor–β-arrestin complex in a lipid nanodisc. Nature 579, 297–302 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Lee, Y. et al. Molecular basis of beta-Arrestin coupling to formoterol-bound beta(1)-adrenoceptor. Nature 583, 862–866 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Bous, J. et al. Structure of the vasopressin hormone-V2 receptor-beta-Arrestin1 ternary complex. Sci. Adv. 8, eabo7761 (2022).Cao, C. et al. Signaling snapshots of a serotonin receptor activated by the prototypical psychedelic LSD. Neuron 110, 3154–3167.e7 (2022).Article  CAS  PubMed  PubMed Central  Google Scholar Chen, K. et al. Tail engagement of arrestin at the glucagon receptor. Nature 620, 904–910 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Wang, Y. et al. Cryo-EM structure of cannabinoid receptor CB1-beta-Arrestin complex. Protein Cell 15, 230–234 (2024).Article  PubMed  PubMed Central  Google Scholar Liao, Y. Y. et al. Snapshot of the cannabinoid receptor 1-arrestin complex unravels the biased signaling mechanism. Cell 186, 5784–5797.e17 (2023).Article  CAS  PubMed  Google Scholar Wen, T. et al. Molecular basis of beta-Arrestin coupling to the metabotropic glutamate receptor mGlu3. Nat. Chem. Biol. 21, 1262–1269 (2025).Article  CAS  PubMed  Google Scholar Sun, D. et al. Molecular mechanism of the arrestin-biased agonism of neurotensin receptor 1 by an intracellular allosteric modulator. Cell Res. 35, 284–295 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Chen, Q. et al. Effect of phosphorylation barcodes on arrestin binding to a chemokine receptor. Nature 643, 280–287 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Lin, X. et al. Constitutive arrestin recruitment by orphan GPR52 via an atypical binding mode. Cell Res. 35, 913–916 (2025).Article  PubMed  PubMed Central  Google Scholar Yue, Y. et al. Mechanistic insights into the versatile stoichiometry and biased signaling of the apelin receptor-arrestin complex. Nat. Commun. 16, 7403 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Zhao, J. et al. Structural characterization of five functional states of metabotropic glutamate receptor 8. Mol. Cell 85, 3460–3473.e6 (2025).Article  CAS  PubMed  Google Scholar Cai, H. et al. Noncanonical agonist-dependent and -independent arrestin recruitment of GPR1. Science 390, eadt8794 (2025).Article  CAS  PubMed  Google Scholar Zhang, H. et al. The molecular basis of mu-opioid receptor signaling plasticity. Cell Res 35, 1021–1036 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Zhao, L. H. et al. Structural basis of PTH1R-beta-Arrestin core engagement reveals design principles for G-protein-biased therapeutics. Nat. Struct. Mol. Biol. 33, 868–881 (2026).Article  CAS  PubMed  Google Scholar Zhai, X. Core conformation of arrestin coupling to parathyroid hormone type 1 receptor. Nat. Commun. 17, 5707 (2026).Article  CAS  PubMed  PubMed Central  Google Scholar Nguyen, A. H. et al. Structure of an endosomal signaling GPCR–G protein–β-arrestin megacomplex. Nat. Struct. Mol. Biol. 26, 1123–1131 (2019).Article  CAS  PubMed  PubMed Central  Google Scholar Shukla, A. K. et al. Visualization of arrestin recruitment by a G-protein-coupled receptor. Nature 512, 218–222 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Lally, C. C. M., Bauer, B., Selent, J. & Sommer, M. E. C-edge loops of arrestin function as a membrane anchor. Nat. Commun. 8, 14258 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Maharana, J. et al. Structural snapshots uncover a key phosphorylation motif in GPCRs driving beta-Arrestin activation. Mol. Cell 83, 2091–2107.e7 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Eiger, D. S. et al. Phosphorylation barcodes direct biased chemokine signaling at CXCR3. Cell Chem. Biol. 30, 362–382.e8 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Maharana, J. et al. Molecular insights into atypical modes of beta-Arrestin interaction with seven transmembrane receptors. Science 383, 101–108 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Hirsch, J. A., Schubert, C., Gurevich, V. V. & Sigler, P. B. The 2.8 A crystal structure of visual arrestin: a model for arrestin’s regulation. Cell 97, 257–269 (1999).Article  CAS  PubMed  Google Scholar Chen, Q., Iverson, T. M. & Gurevich, V. V. Structural basis of arrestin-dependent signal transduction. Trends Biochem. Sci. 43, 412–423 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Janetzko, J. et al. Membrane phosphoinositides regulate GPCR-beta-Arrestin complex assembly and dynamics. Cell 185, 4560–4573.e19 (2022).Article  CAS  PubMed  PubMed Central  Google Scholar Gomes, A. A. S. et al. Lipids modulate the dynamics of GPCR:beta-Arrestin interaction. Nat. Commun. 16, 4982 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Chen, Q. et al. Structural basis of arrestin-3 activation and signaling. Nat. Commun. 8, 1427 (2017).Article  PubMed  PubMed Central  Google Scholar Hedger, G. & Yen, H. Y. The influence of phosphoinositide lipids in the molecular biology of membrane proteins: recent insights from simulations. J. Mol. Biol. 437, 168937 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Chen, Q. et al. An eight amino acid segment controls oligomerization and preferred conformation of the two non-visual arrestins. J. Mol. Biol. 433, 166790 (2021).Article  CAS  PubMed  Google Scholar Sander, C. L. et al. Structural evidence for visual arrestin priming via complexation of phosphoinositols. Structure 30, 263–277.e5 (2022).Article  CAS  PubMed  Google Scholar Wootten, D., Christopoulos, A., Marti-Solano, M., Babu, M. M. & Sexton, P. M. Mechanisms of signalling and biased agonism in G protein-coupled receptors. Nat. Rev. Mol. Cell Biol. 19, 638–653 (2018).Article  CAS  PubMed  Google Scholar Rajagopal, S., Rajagopal, K. & Lefkowitz, R. J. Teaching old receptors new tricks: biasing seven-transmembrane receptors. Nat. Rev. Drug Discov. 9, 373–386 (2010).Article  CAS  PubMed  PubMed Central  Google Scholar Gurevich, V. V. & Gurevich, E. V. Biased GPCR signaling: possible mechanisms and inherent limitations. Pharmacol. Ther. 211, 107540 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Gan, T. J. & Wase, L. Oliceridine, a G protein-selective ligand at the mu-opioid receptor, for the management of moderate to severe acute pain. Drugs Today 56, 269–286 (2020).Article  CAS  Google Scholar Hill, D. M. & Todor, L. A. A prospective, historical-controlled evaluation of oliceridine for moderate or severe pain in patients with acute burn injuries (RELIEVE). Burns 51, 107343 (2025).Article  PubMed  Google Scholar Elisi, G. M. & Bottegoni, G. Impact of G protein-coupled receptor conformation on signaling bias: integrating simulations and biophysical experiments. Pharmacol. Ther. 274, 108905 (2025).Article  CAS  PubMed  Google Scholar Wingler, L. M. & Lefkowitz, R. J. Conformational basis of G protein-coupled receptor signaling versatility. Trends Cell Biol. 30, 736–747 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Calebiro, D. et al. Persistent cAMP-signals triggered by internalized G-protein-coupled receptors. PLoS Biol 7, e1000172 (2009).Article  PubMed  PubMed Central  Google Scholar Thomsen, A. R. B. et al. GPCR-G protein-β-arrestin super-complex mediates sustained G protein signaling. Cell 166, 907–919 (2016).Article  CAS  PubMed  PubMed Central  Google Scholar Nguyen, A. H. & Lefkowitz, R. J. Signaling at the endosome: cryo-EM structure of a GPCR-G protein-beta-Arrestin megacomplex. FEBS J 288, 2562–2569 (2021).Article  CAS  PubMed  PubMed Central  Google Scholar Sokrat, B. et al. Role of the V2R-βarrestin-Gβγ complex in promoting G protein translocation to endosomes. Commun. Biol. 7, 826 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Pena, K. A., Savransky, S. & Lewis, B. Endosomal signaling via cAMP in parathyroid hormone (PTH) type 1 receptor biology. Mol. Cell. Endocrinol. 581, 112107 (2024).Article  CAS  PubMed  Google Scholar He, G. et al. A GPCR-G protein-beta-Arrestin megacomplex enabled by a versatile allosteric modulator. Cell 189, 1434–1450.e22 (2026).Article  CAS  PubMed  Google Scholar Xie, W., Lai, J., Cai, H., Eric Xu, H. & Yin, W. Structural features of arrestin-mediated GPCR signaling. Med. Drug Discov. 24, 100201 (2024).Article  CAS  Google Scholar Cahill, T. J. 3rd et al. Distinct conformations of GPCR-beta-Arrestin complexes mediate desensitization, signaling, and endocytosis. Proc. Natl. Acad. Sci. USA 114, 2562–2567 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Kang, Y. et al. A structural snapshot of the rhodopsin–arrestin complex. FEBS J 283, 816–821 (2015).Article  PubMed  PubMed Central  Google Scholar Han, M., Gurevich, V. V., Vishnivetskiy, S. A., Sigler, P. B. & Schubert, C. Crystal structure of beta-Arrestin at 1.9 A: possible mechanism of receptor binding and membrane translocation. Structure 9, 869–880 (2001).Article  CAS  PubMed  Google Scholar Latorraca, N. R. et al. Molecular mechanism of GPCR-mediated arrestin activation. Nature 557, 452–456 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Aydin, Y. et al. Structural details of a Class B GPCR-arrestin complex revealed by genetically encoded crosslinkers in living cells. Nat. Commun. 14, 1151 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Yang, F. et al. Phospho-selective mechanisms of arrestin conformations and functions revealed by unnatural amino acid incorporation and 19F-NMR. Nat. Commun. 6, 8202 (2015).Article  PubMed  PubMed Central  Google Scholar Butcher, A. J. et al. Concomitant action of structural elements and receptor phosphorylation determines arrestin-3 interaction with the free fatty acid receptor FFA4. J. Biol. Chem. 289, 18451–18465 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Yang, Z. et al. Phosphorylation of G protein-coupled receptors: from the barcode hypothesis to the flute model. Mol. Pharmacol. 92, 201–210 (2017).Article  CAS  PubMed  Google Scholar Bouzo-Lorenzo, M. et al. Distinct phosphorylation sites on the ghrelin receptor, GHSR1a, establish a code that determines the functions of β-arrestins. Sci. Rep. 6, 22495 (2016).Article  CAS  PubMed  PubMed Central  Google Scholar Guillien, M. et al. Phosphorylation motif dictates GPCR C-terminal domain conformation and arrestin interaction. Structure 31, 1394–1406.e7 (2023).Article  CAS  PubMed  Google Scholar Min, K. et al. Crystal structure of beta-Arrestin 2 in complex with CXCR7 phosphopeptide. Structure 28, 1014–1023.e4 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Gareri, C. et al. Phosphorylation patterns in the AT1R C-terminal tail specify distinct downstream signaling pathways. Sci. Signal. 17, eadk5736 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Dwivedi-Agnihotri, H. et al. Distinct phosphorylation sites in a prototypical GPCR differently orchestrate beta-Arrestin interaction, trafficking, and signaling. Sci. Adv. 6, eabb8368 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Isaikina, P. et al. A key GPCR phosphorylation motif discovered in arrestin2⋅CCR5 phosphopeptide complexes. Mol. Cell 83, 2108–2121.e7 (2023).Article  CAS  PubMed  Google Scholar Latorraca, N. R. et al. How GPCR phosphorylation patterns orchestrate arrestin-mediated signaling. Cell 183, 1813–1825.e8 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Eichel, K. et al. Catalytic activation of beta-Arrestin by GPCRs. Nature 557, 381–386 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Komolov, K. E. & Benovic, J. L. G protein-coupled receptor kinases: past, present and future. Cell. Signal. 41, 17–24 (2018).Article  CAS  PubMed  Google Scholar Chen, Q. & Tesmer, J. J. G. G protein-coupled receptor interactions with arrestins and GPCR kinases: the unresolved issue of signal bias. J. Biol. Chem. 298, 102279 (2022).Article  CAS  PubMed  PubMed Central  Google Scholar Choi, M. et al. G protein-coupled receptor kinases (GRKs) orchestrate biased agonism at the beta(2)-adrenergic receptor. Sci. Signal. 11, eaar7084 (2018).Article  PubMed  Google Scholar Matthees, E. S. F. et al. GRK specificity and Gβγ dependency determines the potential of a GPCR for arrestin-biased agonism. Commun. Biol. 7, 802 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Chen, Q. et al. Structures of rhodopsin in complex with G-protein-coupled receptor kinase 1. Nature 595, 600–605 (2021).Article  CAS  PubMed  PubMed Central  Google Scholar Duan, J. et al. GPCR activation and GRK2 assembly by a biased intracellular agonist. Nature 620, 676–681 (2023).Article  CAS  PubMed  Google Scholar Schafer, C. T., Chen, Q., Tesmer, J. J. G. & Handel, T. M. Atypical chemokine receptor 3 “senses” CXC chemokine receptor 4 activation through GPCR kinase phosphorylation. Mol. Pharm. 104, 174–186 (2023).Article  CAS  Google Scholar Busillo, J. M. et al. Site-specific phosphorylation of CXCR4 is dynamically regulated by multiple kinases and results in differential modulation of CXCR4 signaling. J. Biol. Chem. 285, 7805–7817 (2010).Article  CAS  PubMed  PubMed Central  Google Scholar Janetzko, J. et al. Membrane phosphoinositides allosterically tune beta-Arrestin dynamics to facilitate GPCR core engagement. bioRxiv. https://doi.org/10.1101/2025.06.06.658200 (2025).Kim, K. & Chung, K. Y. Molecular mechanism of beta-Arrestin-2 pre-activation by phosphatidylinositol 4,5-bisphosphate. EMBO Rep 25, 4190–4205 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Grimes, J. et al. Plasma membrane preassociation drives beta-Arrestin coupling to receptors and activation. Cell 186, 2238–2255.e20 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Oakley, R. H., Laporte, S. A., Holt, J. A., Caron, M. G. & Barak, L. S. Differential affinities of visual arrestin, βarrestin1, and βarrestin2 for G protein-coupled receptors delineate two major classes of receptors. J. Biol. Chem. 275, 17201–17210 (2000).Article  CAS  PubMed  Google Scholar Kuramoto, R. et al. Membrane-domain compartmentalization of active GPCRs by beta-Arrestins through PtdIns(4,5)P(2) binding. Nat. Chem. Biol. 21, 1927–1937 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Killeen, T. D. et al. Distinct membrane binding properties of the two non-visual arrestins. Commun. Biol. 9, 150 (2026).Article  CAS  PubMed  PubMed Central  Google Scholar Zhuang, T., Vishnivetskiy, S. A., Gurevich, V. V. & Sanders, C. R. Elucidation of inositol hexaphosphate and heparin interaction sites and conformational changes in arrestin-1 by solution nuclear magnetic resonance. Biochemistry 49, 10473–10485 (2010).Article  CAS  PubMed  PubMed Central  Google Scholar Hanson, S. M., Vishnivetskiy, S. A., Hubbell, W. L. & Gurevich, V. V. Opposing effects of inositol hexakisphosphate on rod arrestin and arrestin2 self-association. Biochemistry 47, 1070–1075 (2008).Article  CAS  PubMed  Google Scholar Milano, S. K., Kim, Y. M., Stefano, F. P., Benovic, J. L. & Brenner, C. Nonvisual arrestin oligomerization and cellular localization are regulated by inositol hexakisphosphate binding. J. Biol. Chem. 281, 9812–9823 (2006).Article  CAS  PubMed  Google Scholar Wettschureck, N. & Offermanns, S. Mammalian G proteins and their cell type specific functions. Physiol. Rev. 85, 1159–1204 (2005).Article  CAS  PubMed  Google Scholar Calebiro, D., Koszegi, Z., Lanoiselee, Y., Miljus, T. & O’Brien, S. G protein-coupled receptor-G protein interactions: a single-molecule perspective. Physiol. Rev. 101, 857–906 (2021).Article  CAS  PubMed  Google Scholar Hausdorff, W. P., Caron, M. G. & Lefkowitz, R. J. Turning off the signal: desensitization of beta-adrenergic receptor function. FASEB J 4, 2881–2889 (1990).Article  CAS  PubMed  Google Scholar Coffa, S., Breitman, M., Spiller, B. W. & Gurevich, V. V. A single mutation in arrestin-2 prevents ERK1/2 activation by reducing c-Raf1 binding. Biochemistry 50, 6951–6958 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Peterson, Y. K. & Luttrell, L. M. The diverse roles of arrestin scaffolds in G protein-coupled receptor signaling. Pharmacol. Rev. 69, 256–297 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar DeWire, S. M., Ahn, S., Lefkowitz, R. J. & Shenoy, S. K. Beta-Arrestins and cell signaling. Annu. Rev. Physiol. 69, 483–510 (2007).Article  CAS  PubMed  Google Scholar Luttrell, L. M. et al. Beta-Arrestin-dependent formation of beta2 adrenergic receptor-Src protein kinase complexes. Science 283, 655–661 (1999).Article  CAS  PubMed  Google Scholar McDonald, P. H. et al. Beta-Arrestin 2: a receptor-regulated MAPK scaffold for the activation of JNK3. Science 290, 1574–1577 (2000).Article  CAS  PubMed  Google Scholar Kahsai, A. W. et al. Signal transduction at GPCRs: Allosteric activation of the ERK MAPK by beta-Arrestin. Proc. Natl. Acad. Sci. USA 120, e2303794120 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Pakharukova, N. et al. Mechanism of beta-Arrestin 1 mediated Src activation via Src SH3 domain revealed by cryo-electron microscopy. Nat. Commun. 17, 2973 (2026).Article  CAS  PubMed  PubMed Central  Google Scholar Kaya, A. I., Perry, N. A., Gurevich, V. V. & Iverson, T. M. Phosphorylation barcode-dependent signal bias of the dopamine D1 receptor. Proc. Natl. Acad. Sci. USA 117, 14139–14149 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Pakharukova, N., Masoudi, A., Pani, B., Staus, D. P. & Lefkowitz, R. J. Allosteric activation of proto-oncogene kinase Src by GPCR–beta-Arrestin complexes. J. Biol. Chem. 295, 16773–16784 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Yang, F. et al. Allosteric mechanisms underlie GPCR signaling to SH3-domain proteins through arrestin. Nat. Chem. Biol. 14, 876–886 (2018).Article  CAS  PubMed  Google Scholar Ahn, S., Shenoy, S. K., Wei, H. & Lefkowitz, R. J. Differential kinetic and spatial patterns of β-arrestin and G protein-mediated ERK activation by the angiotensin II receptor. J. Biol. Chem. 279, 35518–35525 (2004).Article  CAS  PubMed  Google Scholar Whalen, E. J., Rajagopal, S. & Lefkowitz, R. J. Therapeutic potential of beta-Arrestin- and G protein-biased agonists. Trends Mol. Med. 17, 126–139 (2011).Article  CAS  PubMed  Google Scholar Smith, J. S., Lefkowitz, R. J. & Rajagopal, S. Biased signalling: from simple switches to allosteric microprocessors. Nat. Rev. Drug Discov. 17, 243–260 (2018).Article  CAS  PubMed  PubMed Central  Google Scholar Oliveira de Souza, C., Sun, X. & Oh, D. Metabolic functions of G protein-coupled receptors and beta-Arrestin-mediated signaling pathways in the pathophysiology of type 2 diabetes and obesity. Front. Endocrinol. 12, 715877 (2021).Article  Google Scholar Suomivuori, C. M. et al. Molecular mechanism of biased signaling in a prototypical G protein-coupled receptor. Science 367, 881–887 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Wingler, L. M. et al. Angiotensin analogs with divergent bias stabilize distinct receptor conformations. Cell 176, 468–478.e11 (2019).Article  CAS  PubMed  PubMed Central  Google Scholar Zhao, L. H. et al. Conserved class B GPCR activation by a biased intracellular agonist. Nature 621, 635–641 (2023).Article  CAS  PubMed  Google Scholar Kobayashi, K. et al. Class B1 GPCR activation by an intracellular agonist. Nature 618, 1085–1093 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Krumm, B. E. et al. Neurotensin receptor allosterism revealed in complex with a biased allosteric modulator. Biochemistry 62, 1233–1248 (2023).Article  CAS  PubMed  Google Scholar Zhuang, Y. et al. Molecular recognition of morphine and fentanyl by the human mu-opioid receptor. Cell 185, 4361–4375.e19 (2022).Article  CAS  PubMed  Google Scholar Ferrandon, S. et al. Sustained cyclic AMP production by parathyroid hormone receptor endocytosis. Nat. Chem. Biol. 5, 734–742 (2009).Article  CAS  PubMed  PubMed Central  Google Scholar Irannejad, R. et al. Conformational biosensors reveal GPCR signalling from endosomes. Nature 495, 534–538 (2013).Article  CAS  PubMed  PubMed Central  Google Scholar Feinstein, T. N. et al. Noncanonical control of vasopressin receptor type 2 signaling by retromer and arrestin. J. Biol. Chem. 288, 27849–27860 (2013).Article  CAS  PubMed  PubMed Central  Google Scholar Calebiro, D., Miljus, T. & O’Brien, S. Endomembrane GPCR signaling: 15 years on, the quest continues. Trends Biochem. Sci. 50, 46–60 (2025).Article  CAS  PubMed  Google Scholar Flores-Espinoza, E. & Thomsen, A. R. B. Beneath the surface: endosomal GPCR signaling. Trends Biochem. Sci. 49, 520–531 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Pavlos, N. J. & Friedman, P. A. GPCR signaling and trafficking: the long and short of it. Trends Endocrinol. Metab. 28, 213–226 (2017).Article  CAS  PubMed  Google Scholar Godbole, A., Lyga, S., Lohse, M. J. & Calebiro, D. Internalized TSH receptors en route to the TGN induce local G(s)-protein signaling and gene transcription. Nat. Commun. 8, 443 (2017).Article  PubMed  PubMed Central  Google Scholar Tsvetanova, N. G. & von Zastrow, M. Spatial encoding of cyclic AMP signaling specificity by GPCR endocytosis. Nat. Chem. Biol. 10, 1061–1065 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Vilardaga, J. P. et al. Molecular mechanisms of PTH/PTHrP class B GPCR signaling and pharmacological implications. Endocr. Rev. 44, 474–491 (2023).Article  PubMed  PubMed Central  Google Scholar Pena, K. A. Endosomal parathyroid hormone receptor signaling. Am. J. Physiol. Cell Physiol. 323, C783–C790 (2022).Article  CAS  PubMed  PubMed Central  Google Scholar Sutkeviciute, I., Clark, L. J., White, A. D., Gardella, T. J. & Vilardaga, J. P. PTH/PTHrP receptor signaling, allostery, and structures. Trends Endocrinol. Metab. 30, 860–874 (2019).Article  CAS  PubMed  PubMed Central  Google Scholar White, A. D. et al. Gq/11-dependent regulation of endosomal cAMP generation by parathyroid hormone class B GPCR. Proc. Natl. Acad. Sci. USA 117, 7455–7460 (2020).Article  CAS  PubMed  PubMed Central  Google Scholar Gidon, A. et al. Endosomal GPCR signaling turned off by negative feedback actions of PKA and v-ATPase. Nat. Chem. Biol. 10, 707–709 (2014).Article  CAS  PubMed  PubMed Central  Google Scholar Feinstein, T. N. et al. Retromer terminates the generation of cAMP by internalized PTH receptors. Nat. Chem. Biol. 7, 278–284 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Wehbi, V. L. et al. Noncanonical GPCR signaling arising from a PTH receptor–arrestin–Gβγ complex. Proc. Natl. Acad. Sci. USA 110, 1530–1535 (2013).Article  CAS  PubMed  PubMed Central  Google Scholar Shenoy, S. K. & Lefkowitz, R. J. beta-Arrestin-mediated receptor trafficking and signal transduction. Trends Pharmacol. Sci. 32, 521–533 (2011).Article  CAS  PubMed  PubMed Central  Google Scholar Sharma, D. & Parameswaran, N. Multifaceted role of beta-Arrestins in inflammation and disease. Genes Immun 16, 499–513 (2015).Article  CAS  PubMed  PubMed Central  Google Scholar Kee, T. R. et al. The multifaceted functions of beta-Arrestins and their therapeutic potential in neurodegenerative diseases. Exp. Mol. Med. 56, 129–141 (2024).Article  CAS  PubMed  PubMed Central  Google Scholar Wess, J., Oteng, A. B., Rivera-Gonzalez, O., Gurevich, E. V. & Gurevich, V. V. beta-Arrestins: structure, function, physiology, and pharmacological perspectives. Pharmacol. Rev. 75, 854–884 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Noma, T. et al. Beta-Arrestin-mediated beta1-adrenergic receptor transactivation of the EGFR confers cardioprotection. J. Clin. Invest. 117, 2445–2458 (2007).Article  CAS  PubMed  PubMed Central  Google Scholar Wang, J. et al. beta-Arrestin-biased allosteric modulator potentiates carvedilol-stimulated beta adrenergic receptor cardioprotection. Mol. Pharmacol. 100, 568–579 (2021).Article  CAS  PubMed  PubMed Central  Google Scholar Kim, I. M. et al. Beta-blockers alprenolol and carvedilol stimulate beta-Arrestin-mediated EGFR transactivation. Proc. Natl. Acad. Sci. USA 105, 14555–14560 (2008).Article  CAS  PubMed  PubMed Central  Google Scholar Wisler, J. W. et al. A unique mechanism of beta-blocker action: carvedilol stimulates beta-Arrestin signaling. Proc. Natl. Acad. Sci. USA 104, 16657–16662 (2007).Article  CAS  PubMed  PubMed Central  Google Scholar Raehal, K. M., Walker, J. K. & Bohn, L. M. Morphine side effects in beta-Arrestin 2 knockout mice. J. Pharmacol. Exp. Ther. 314, 1195–1201 (2005).Article  CAS  PubMed  Google Scholar Bohn, L. M. et al. Enhanced morphine analgesia in mice lacking beta-Arrestin 2. Science 286, 2495–2498 (1999).Article  CAS  PubMed  Google Scholar Kliewer, A. et al. Phosphorylation-deficient G-protein-biased mu-opioid receptors improve analgesia and diminish tolerance but worsen opioid side effects. Nat. Commun. 10, 367 (2019).Article  CAS  PubMed  PubMed Central  Google Scholar Gillis, A. et al. Critical assessment of G protein-biased agonism at the mu-opioid receptor. Trends Pharmacol. Sci. 41, 947–959 (2020).Article  CAS  PubMed  Google Scholar Che, T., Dwivedi-Agnihotri, H., Shukla, A. K. & Roth, B. L. Biased ligands at opioid receptors: Current status and future directions. Sci. Signal. 14, eaav0320 (2021).Article  CAS  PubMed  PubMed Central  Google Scholar Schipani, E., Kruse, K. & Juppner, H. A constitutively active mutant PTH-PTHrP receptor in Jansen-type metaphyseal chondrodysplasia. Science 268, 98–100 (1995).Article  CAS  PubMed  Google Scholar Portales-Castillo, I. et al. Altered signaling and desensitization responses in PTH1R mutants associated with Eiken syndrome. Commun. Biol. 6, 599 (2023).Article  CAS  PubMed  PubMed Central  Google Scholar Duchatelet, S., Ostergaard, E., Cortes, D., Lemainque, A. & Julier, C. Recessive mutations in PTHR1 cause contrasting skeletal dysplasias in Eiken and Blomstrand syndromes. Hum. Mol. Genet. 14, 1–5 (2005).Article  CAS  PubMed  Google Scholar Volkow, N. D. & Collins, F. S. The role of science in addressing the opioid crisis. N. Engl. J. Med. 377, 391–394 (2017).Article  PubMed  Google Scholar Ju, J., Li, Z., Liu, J., Peng, X. & Gao, F. Biased opioid receptor agonists: balancing analgesic efficacy and side-effect profiles. Int. J. Mol. Sci. 26, 1862 (2025).Article  CAS  PubMed  PubMed Central  Google Scholar Gillis, A. et al. Low intrinsic efficacy for G protein activation can explain the improved side effect profiles of new opioid agonists. Sci. Signal. 13, eaaz3140 (2020).Article  CAS  PubMed  Google Scholar Altarifi, A. A. et al. Effects of acute and repeated treatment with the biased mu opioid receptor agonist TRV130 (oliceridine) on measures of antinociception, gastrointestinal function, and abuse liability in rodents. J. Psychopharmacol. 31, 730–739 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Singh, K. D. & Karnik, S. S. Implications of beta-Arrestin biased signaling by angiotensin II type 1 receptor for cardiovascular drug discovery and therapeutics. Cell. Signal. 124, 111410 (2024).Article  CAS  PubMed  Google Scholar Mathieu, N. M., Nakagawa, P., Grobe, J. L. & Sigmund, C. D. Insights into the role of angiotensin-II AT(1) receptor-dependent beta-Arrestin signaling in cardiovascular disease. Hypertension 81, 6–16 (2024).Article  CAS  PubMed  Google Scholar Pang, P. S. et al. Biased ligand of the angiotensin II type 1 receptor in patients with acute heart failure: a randomized, double-blind, placebo-controlled, phase IIB, dose ranging trial (BLAST-AHF). Eur. Heart J. 38, 2364–2373 (2017).Article  CAS  PubMed  PubMed Central  Google Scholar Download referencesAcknowledgementsThis work was supported by the National Natural Science Foundation of China (32130022, 82495184 and 82121005 to H.E.X., 32371255 and 32071203 to L.-H.Z.); the Natural Science Foundation of Shanghai (23ZR1475200 to L.-H.Z.); Ruijin Hospital Youth Cultivation Program (2024JC002); the National Key R&D Program of China (2022YFC2703105 to H.E.X.); CAS Strategic Priority Research Program (XDB37030103 to H.E.X.); Shanghai Municipal Science and Technology Major Project (2019SHZDZX02 to H.E.X.); the Lingang Laboratory (LG-GG-202204-01 to H.E.X.); State Key Laboratory of Drug Research (SKLDR-2023-TT-04 to H.E.X.).Author informationAuthors and AffiliationsState Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, ChinaQian He & H. Eric XuUniversity of Chinese Academy of Sciences, Beijing, ChinaQian He & H. Eric XuState Key Laboratory of Medical Genomics, Research Center for Medicinal Structural Biology, National Research Center for Translational Medicine at Shanghai, Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, ChinaLi-Hua Zhao & H. Eric XuAuthorsQian HeView author publicationsSearch author on:PubMed Google ScholarLi-Hua ZhaoView author publicationsSearch author on:PubMed Google ScholarH. Eric XuView author publicationsSearch author on:PubMed Google ScholarContributionsQ.H. wrote the manuscript and prepared figures; H.E.X. and L.-H.Z. supervised and revised the manuscript. All authors reviewed and approved the final version.Corresponding authorsCorrespondence to Qian He, Li-Hua Zhao or H. Eric Xu.Ethics declarationsConflict of interestThe authors declare no competing interests.Declaration of the use of generative AI and AI-assisted technologies in the writing processDuring the preparation of this review, we utilized AI-assisted tools, including Gemini3 and ChatGPT, to aid in language refinement and writing clarity. All the content generated with these tools was carefully reviewed and edited by the authors, who take full responsibility for the final version of the manuscript.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Rights and permissionsOpen Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.Reprints and permissionsAbout this article