Edit · 2026 01 11 182653 Nature S41586 025 09302 6

Saving needs a token. Log in or register to get one. Edits are public and licensed CC-BY-SA-4.0.

# Structural and functional characterization of human sweet taste receptor | Nature

**URL:** https://www.nature.com/articles/s41586-025-09302-6

**Saved:** 2026-01-11T02:26:59.047Z

---

Article
Published: 24 June 2025
Structural and functional characterization of human sweet taste receptor
Zongjun Shi, Weixiu Xu, Lijie Wu, Xiaolei Yue, Shenhui Liu, Wei Ding, Jinyi Zhang, Bing Meng, Lianghao Zhao, Xiaoyan Liu, Junlin Liu, Zhi-Jie Liu & Tian Hua

Nature 
volume
 645, 
pages
801–808 (2025)Cite this article

16k Accesses

15 Citations

33 Altmetric

Metrics
details

Abstract

Sweet taste perception influences dietary choices and metabolic health. The human sweet taste receptor, a class C G-protein-coupled receptor (GPCR) heterodimer composed of TAS1R2 and TAS1R3 (refs. 1,2), senses a wide range of sweet compounds—including natural sugars, artificial sweeteners and sweet proteins—and affects metabolic regulation beyond taste. However, the lack of three-dimensional structures hinders our understanding of its precise working mechanism. Here we present cryo-electron microscopy structures of the full-length human sweet taste receptor in apo and sucralose-bound states. These structures reveal a distinct asymmetric heterodimer architecture, with sucralose binding exclusively to the Venus flytrap domain of TAS1R2. Combining mutagenesis and molecular dynamics simulations, this work delineates the sweetener-recognition modes in TAS1R2. Structural comparisons further uncover conformational changes upon ligand binding and a unique activation mechanism. These findings illuminate the signal transduction mechanisms of chemosensory receptors in the class C GPCR family and provide the molecular basis for the design of a new generation of sweeteners.

You have full access to this article via your institution.

Download PDF
Similar content being viewed by others
Biophysical and functional characterization of the human TAS1R2 sweet taste receptor overexpressed in a HEK293S inducible cell line
Article Open access
15 November 2021
Steviol rebaudiosides bind to four different sites of the human sweet taste receptor (T1R2/T1R3) complex explaining confusing experiments
Article Open access
18 October 2024
Prediction of dynamic allostery for the transmembrane domain of the sweet taste receptor subunit, TAS1R3
Article Open access
03 April 2023
Main

Taste is a fundamental sensory modality that has an important role in human life, guiding food choice and dietary behaviours. Humans perceive five basic taste stimuli: sweet, bitter, sour, salty and umami3,4,5,6, with sweet taste being essential for energy homeostasis. Three of these taste modalities—sweet, umami and bitter—are mediated by GPCRs7. Specifically, sweet and umami tastes are detected by type 1 taste receptors (TAS1Rs)8,9, which are class C GPCRs, whereas bitter tastes are sensed by type 2 taste receptors (TAS2Rs)10,11,12, which are categorized as class T GPCRs. Both receptor types are expressed in specialized taste receptor cells (TRCs) in the taste bud and signal primarily through the Gα subunit of gustducin (Gαgust), initiating downstream physiological responses13,14.

The TAS1Rs include three members—TAS1R1, TAS1R2 and TAS1R3 (refs. 8,9)—that form heterodimers to mediate sweet (TAS1R2–TAS1R3 heterodimer) or umami (TAS1R1–TAS1R3 heterodimer) taste perception1,2,15,16 (Fig. 1a). In addition to TAS1Rs, the class C GPCR family also includes metabotropic glutamate receptors (mGluRs)17, γ-aminobutyric acid receptors (GABABRs)18, calcium-sensing receptor (CaSR)19 and several orphan receptors20. TAS1Rs share the characteristic of class C GPCRs, which consist of an extracellular venus flytrap domain (VFTD), a cysteine-rich domain (CRD) and the seven-transmembrane domain (TMD).

Fig. 1: Cryo-EM structures of the human sweet taste receptor.

a, Schematic of sweet (TAS1R2–TAS1R3) and umami (TAS1R1–TAS1R3) taste receptor-mediated activation of the downstream gustducin signalling pathway by diverse ligands. GMP, guanosine monophosphate; IMP, inosine monophosphate. Created in BioRender. Hua, T. (2025) https://BioRender.com/jza1j0a. b, The activities of sucralose, neotame and sucrose on sweet taste receptor measured using the FLIPER microplate reader. Data are mean ± s.e.m. of three independent experiments performed in triplicate. c, Cryo-EM density maps and models of the apo-state TAS1R2–TAS1R3 heterodimer, shown from two different views. TAS1R2 is shown in medium turquoise and TAS1R3 is in purple. d, Cryo-EM density maps and models of sucralose-bound TAS1R2–TAS1R3 heterodimer, shown from two different views. TAS1R2 is shown in sandy brown and TAS1R3 is in green. Sucralose is shown as a magenta sphere.

Source Data

Full size image

The human sweet taste receptor recognizes diverse types of sweet-tasting resources, including natural sugars such as glucose and sucrose, artificial sweeteners such as sucralose and neotame21, and naturally occurring sweet proteins such as monellin, thaumatin, miraculin and brazzein22,23. Of note, the receptor exhibits low binding affinity for sugars, typically in the micromolar to millimolar range24,25 (Fig. 1b). Sweet taste perception is closely linked to energy metabolism because it serves as an early and crucial sensory signal that guides organisms towards energy-rich foods26,27. Beyond oral taste perception, the TAS1R2 and TAS1R3 receptors are ectopically expressed in the gut28, pancreas29 and brain30, where they are involved in regulating metabolic processes. Therefore, the ectopic TAS1Rs have emerged as potential therapeutic targets for obesity and metabolic disorders, including diabetes and hyperlipidemia29,31,32.

Despite its chemosensory and physiological importance, the structural characterization of the human sweet taste receptor remains a major challenge. Previous structural studies have only reported the crystal structures of the VFTD in TAS1Rs from medaka fish bound to amino acids33. However, the full-length structures of human TAS1Rs remain unknown, limiting our understanding of sweetener modulation and receptor activation mechanisms. Here we present cryo-electron microscopy (cryo-EM) structures of the full-length human TAS1R2–TAS1R3 heterodimer in both apo and sucralose-bound states, providing a structural template for understanding sweet taste perception at molecular level.

Structure of apo- and sucralose-bound TAS1R2–TAS1R3

TAS1Rs pose challenges for structure determination owing to their low exogenous recombinant overexpression and the inherent instability of their heterodimeric complexes. To enhance expression, we utilized a fluorescent protein fusion strategy, fusing mScarlet3 protein34 and mNeonGreen35 with flexible linkers to the C terminus of TAS1R2 and TAS1R3, respectively (Extended Data Fig. 1a and Methods), and co-expressed the modified TAS1R2 and TAS1R3 using the BacMam system (Methods). Using these strategies, we obtained stable TAS1R2–TAS1R3 heterodimers in the apo and sucralose-bound states (Extended Data Figs. 1b,c, 2 and 3) and determined the cryo-EM structures at overall resolutions of 3.4 Å (apo) and 3.5 Å (sucralose-bound) (Fig. 1c,d and Extended Data Table 1). The EM density maps were of sufficient quality to distinguish TAS1R2 from TAS1R3, aided by the well-resolved sidechain densities in most parts of the VFTD (Extended Data Fig. 4a,c), unique glycosylation sites in the CRD of TAS1R2 (TAS1R2-CRD) (Extended Data Fig. 4b), as well as the TMD in the apo structure (Extended Data Fig. 4e,f). These electron microscopy maps enabled us to build structure models for most receptor regions. Notably, the electron microscopy density for sucralose in the binding pocket was well resolved.

Additionally, we also determined the apo-TAS1R2–TAS1R3 structure in the absence of glycerol for the purpose of testing its potential activity on sweet taste receptor (Extended Data Fig. 5a–d and Methods), although our calcium mobilization assay showed that glycerol could not activate the sweet taste receptor even at 1.0 M concentration (Extended Data Fig. 1d). Consistently, this glycerol-free structure is identical to the apo structure (Extended Data Fig. 5e,f). Both heterodimer complex samples thus represent the true apo state of the receptor.

Overall structures of human sweet taste receptor

The overall architectures of the human sweet taste receptor in the apo and sucralose-bound states reveal an asymmetric heterodimer of TAS1R2 and TAS1R3 (Fig. 1c,d). Each subunit consists of an extracellular VFTD, connected to the TMD via a cysteine-rich region. Of note, the experimentally determined structures deviate significantly from the symmetric heterodimer model predicted by AlphaFold3 (ref. 36).

In the apo state, the TAS1R2–TAS1R3 heterodimer adopts an asymmetric conformation, characterized by an angle of approximately 70° between the VFTD and TMD dimer interfaces (Extended Data Fig. 6a). This angular offset results in twisted CRDs, contrasting with the straight architecture commonly observed in other apo class C GPCR dimers. Upon sucralose binding, the overall conformation of the TAS1R2–TAS1R3 heterodimer exhibits moderate changes, with a Cα root mean squared deviation (RMSD) of 2.97 Å relative to the apo state (Extended Data Fig. 6b). Of note, sucralose binds exclusively to the VFTD of TAS1R2 (TAS1R2-VFTD), triggering a conformational cascade involving the shift of VFTD and bending of CRD, ultimately reducing the overall height of the receptor in the sucralose-bound state (Extended Data Fig. 6b–d and Supplementary Video 1).

Sequence identity between the full-length human TAS1R2 and TAS1R3 is about 30%, with 32% identity in the VFTDs, 40% identity in the CRDs and 25% identity in the TMDs (Extended Data Fig. 7). Structural comparisons between the two subunits in apo state reveal substantial differences, with Cα RMSD of 9.30 Å. These differences occur mainly in the VFTD and CRD, further emphasizing the asymmetric architecture (Extended Data Fig. 4g).

While this manuscript was under review, a structure of the sucralose-bound human sweet taste receptor (Protein Data Bank (PDB): 9NOR) was reported37; this structure closely resembles our sucralose-bound structure and differs from the apo-state structure (Extended Data Fig. 5g).

Sucralose binding mode in TAS1R2

Sucralose is a widely used noncaloric sweetener and a potent agonist of the human sweet taste receptor, with a half-maximum effective concentration (EC50) of 43.5 µM as determined by our calcium mobilization assay (Fig. 1b, Extended Data Tables 2 and 3 and Methods), consistent with previous reports24,25. The VFTDs of TAS1R2 and TAS1R3 are structurally divided into the upper (LB1) and lower (LB2) lobes, with orthosteric agonists typically binding in the cleft formed between the lobes. A well-defined density was observed between LB1 and LB2 in TAS1R2-VFTD, consistent with the shape of sucralose (Fig. 2a), along with clear densities for surrounding key residues in the binding pocket (Extended Data Fig. 8a). By contrast, the VFTD of TAS1R3 remained in an open conformation with no additional ligand density identified (Fig. 2a,b).

Fig. 2: Sucralose binding mode in TAS1R2.

a, Cryo-EM density map of sucralose-bound VFTDs from the TAS1R2–TAS1R3 heterodimer. TAS1R2-VFTD is shown in sandy brown, TAS1R3-VFTD is in green and sucralose is in magenta. The zoomed-in view shows the cryo-EM density map of sucralose in the binding pocket. b, Sucralose binds in the cleft formed by LB1 and LB2 of the TAS1R2 VFTD. c, Structural comparison of TAS1R2-VFTD and TAS1R3-VFTD in the sucralose-bound state. Differences in the sucralose-binding pockets are highlighted. Green circles, residues from TAS1R3; orange circles, residues from TAS1R2. Residue numbers are labelled according to TAS1R2. d, Key residues involved in sucralose binding within the TAS1R2-VFTD are shown. e,f, Dose-response curves of key mutations in LB1 (e) and LB2 (f) of TAS1R2-VFTD in response to stimulation with sucralose in the calcium mobilization assay. Data are mean ± s.e.m. of three independent experiments performed in triplicate. Emax, maximum response; WT, wild type.

Source Data

Full size image

To precisely model the binding pose of sucralose, we integrated the cryo-EM density map, site-directed mutagenesis, published mutagenesis data38,39,40,41 and molecular dynamics simulations (Methods). The modelled sucralose is anchored in the inter-lobe cleft of TAS1R2-VFTD through a network of polar interactions, hydrogen bonds and hydrophobic contacts (Fig. 2d). Key contact residues in the binding pocket include I67, L71, Y103, D142, N143, S165, P277, D278, E302 and V384 (Fig. 2d). Among these, D142, D278, E302 and S303 form polar interactions or hydrogen bonds with sucralose. Consistently, mutagenesis studies on those residues show that D142A, E302A, S165A and S303A in TAS1R2-VFTD impaired or decreased sucralose-induced receptor activation (Fig. 2e,f and Extended Data Tables 2 and 3). Notably, Y103 functions as an anchor for sucralose binding and Y103A mutation abolished sucralose binding activity, yet the Y103F mutation retained receptor responsiveness (Fig. 2e). These residues are not conserved in TAS1R3, supporting the structural basis for the subunit-specific binding of sucralose to TAS1R2 (Fig. 2c). Furthermore, the binding pose of sucralose is stable during the molecular dynamics simulations, further validating our structural interpretation (Extended Data Fig. 8b). In addition, the sucralose binding mode is consistent with that of the recently reported structure (PDB: 9NOV) (Extended Data Fig. 8c).

Ligand specificity in the sweet taste receptor

To further explore the ligand specificity of TAS1R2-VFTD, we selected two additional representative sweeteners, neotame and natural sugar sucrose, to investigate their potential binding modes. In our calcium mobilization assay, both neotame and sucrose showed activity on the sweet taste receptor, with EC50 values of 1.5 µM and 2.5 mM, respectively (Fig. 1b and Extended Data Table 2). Molecular docking revealed that these two ligands adopt binding poses resembling that of sucralose, occupying the same inter-lobe cleft in the TAS1R2-VFTD (Extended Data Fig. 8d). However, each ligand forms different interactions with the surrounding Asp, Glu or Ser residues (Extended Data Fig. 8e,f). Our mutagenesis data and the prior reports38 show that mutations in those residues have different effects on sucralose, neotame and sucrose (Fig. 2e,f, Extended Data Fig. 8g,h and Extended Data Table 2).

Notably, the sucralose-binding pocket is relatively compact, imposing a size limitation on larger molecular weight orthosteric ligands. This suggests the possible existence of alternative binding sites in TAS1R2 or TAS1R3.

VFTD and CRD in the heterodimer

The dimerization interfaces of the VFTDs in the apo and sucralose-bound TAS1R2–TAS1R3 heterodimers are mediated primarily by lobe 1 and supported by lobe 2 of each subunit (Fig. 2b). Additionally, a loop (residues 511–519) from the CRD of TAS1R3 protrudes into the interface between lobe 2 of each subunit (Fig. 3a and Extended Data Fig. 4d). Consistently, alanine mutation of residue F514 or H515—located at the tip of this loop—impaired sucralose-induced receptor activation (Fig. 3e). Similar to other class C GPCRs, the VFTD and CRD are further linked by interdomain disulfide bonds, including C233–C513 in TAS1R2 and C236–C517 in TAS1R3 (Fig. 3b–d). Moreover, several glycosylation sites are observed in the VFTD and CRD regions of both subunits, as indicated by extra electron densities near Asn residues, and appear to be subunit-specific (Extended Data Fig. 4b).

Fig. 3: The VFTD and CRD in the TAS1R2–TAS1R3 heterodimer.

a, A loop region (residues 511–519) from TAS1R3-CRD inserts into the LB2 dimerization interface of TAS1R2-VFTD and TAS1R3-VFTD. Key interaction residues are shown in sticks. b, Conformational changes in TAS1R2-CRD from apo to sucralose-bound states, driven by LB2 rotation and movements, as indicated by red arrows. c, Interaction networks between LB2 and CRD loops in TAS1R2 and TAS1R3, highlighting conformational changes from apo to sucralose-bound states, indicated by red arrows. d, The rotation of LB2 in TAS1R2 induces a corresponding rotation and movement of LB2 in TAS1R3, driving conformational changes in the CRD of TAS1R3. e, Alanine substitution of F514 or H515 reduces sucralose activity on the sweet taste receptor in the calcium mobilization assay. Data are mean ± s.e.m. of three independent experiments performed in triplicate. f,g, VFTD conformations in the apo-TAS1R2–TAS1R3 structure. TAS1R3-VFTD (purple surface) adopts a more open conformation than TAS1R2-VFTD (cyan surface) (f), and the opening angle between LB1 and LB2 is larger in TAS1R3-VFTD, shown in cartoon representation (g). h, Overall structural comparison of VFTDs in apo- and sucralose-bound TAS1R2–TAS1R3 states. i, In the sucralose-bound state, TAS1R2-VFTD exhibits a minor closure compared with apo-TAS1R2-VFTD.

Source Data

Full size image

Structural comparison of the VFTDs in apo state reveals a Cα RMSD of about 2.08 Å, with TAS1R3-VFTD adopting a more open conformation compared with TAS1R2-VFTD (Fig. 3f,g). Furthermore, the distinct electrostatic charge distribution in the clefts between two subunits suggests different ligand recognition preferences between TAS1R2 and TAS1R3 (Extended Data Fig. 9a).

Upon sucralose binding, both VFTDs undergo conformational changes—the orthosteric binding pocket of TAS1R2 displays noticeable closure (Supplementary Video 2), whereas the cleft in TAS1R3 remains unchanged (Fig. 3h,i). The Cα RMSD between VFTDs of TAS1R2 and TAS1R3 is 2.51 Å, suggesting that the ligand-induced conformational changes are mainly confined to the TAS1R2-VFTD (Supplementary Video 3). These structural shifts in the VFTDs are subsequently propagated to CRDs (Fig. 3b–d).

The CRDs of TAS1R2 and TAS1R3 are stabilized by disulfide bonds and adopt a bent, asymmetric arrangement in the heterodimer (Fig. 4a). A unique glycosylation site, N527, in TAS1R2-CRD further emphasizes this subunit-specific structural divergence (Extended Data Fig. 4b). This asymmetric architecture may reflect functional specialization during receptor activation, particularly given that TAS1R3 also forms a heterodimer with TAS1R1 to mediate umami taste signalling. Notably, sweet proteins, which are reported to bind at the CRD region of the heterodimer42, may interact differently with these subunits. The distinct CRD structural configuration observed here provides a precise template for investigating binding modes of sweet proteins and their roles in modulating signal transmission in sweet taste perception.

Fig. 4: Asymmetric configuration of CRDs and dimerization interface of TMDs in TAS1R2–TAS1R3 heterodimer.

a, Asymmetric configuration of CRDs in apo and sucralose-bound states in the TAS1R2–TAS1R3 heterodimer. Conformational changes between the two states are indicated by red arrows. b,c, Conformations of ECL2 and its interaction with the CRD in the apo-state TAS1R2–TAS1R3 structure. ECL2–CRD interaction in TAS1R3 (b) and TAS1R2 (c). d, The dimerization interface is formed by TM5 and TM6 of TAS1R3 and TM6 of TAS1R2. Key interaction residues are shown in sticks. e, Structural comparison of the TMDs in apo and sucralose-bound states, shown in extracellular (left) and intracellular (right) views.

Full size image
TMDs in the sweet taste receptor

The cryo-EM density maps of the TMDs in the apo state are well-resolved, allowing accurate modelling (Extended Data Fig. 4f). The TMD dimerization interface is clearly defined, including the sidechains of key interacting residues (Extended Data Fig. 4f). This interface is mainly formed by TAS1R2-TM6 and the TM5 and TM6 helices of TAS1R3 (Fig. 4d). The closest distance between TAS1R2-TM6 (Cα of A782) and TAS1R3-TM6 (Cα of A783) is about 5.5 Å. The interface is stabilized mainly by hydrophobic and π–π interactions, involving residues from the extracellular region of TM6 in each subunit and intracellular parts of TM5 in TAS1R3 and TM6 in TAS1R2 (Fig. 4d).

Extracellular loop 2 (ECL2) adopts a conserved parallel β-sheet conformation that packs closely against a short β-strand at the C-terminal end of the CRD in each subunit (Fig. 4b,c). Despite the structural conservation, ECL2 engages the CRD in a subunit-specific manner. In TAS1R3, ECL2 extends obliquely and interacts with residues 542–546 in the TAS1R3-CRD (Fig. 4b). By contrast, TAS1R2-ECL2 is positioned more closely to ECL3, forming hydrophobic and polar interactions that may enhance the flexibility of the TAS1R2-CRD (Fig. 4c). Notably, ECL2 in both subunits maintains a conserved disulfide bond with TM3 (C633TM3–C722ECL2 in TAS1R3, C630TM3–C720ECL2 in TAS1R2) (Fig. 4b,c and Extended Data Fig. 4e). Notably, intracellular loop 2 (ICL2) in both TAS1R2 and TAS1R3 adopt a short helical conformation, resembling that observed in the bitter taste receptors TAS2R14 (ref. 43) and TAS2R46 (ref. 44) (Extended Data Figs. 4f and 9c).

In the sucralose-bound structure, 3D variability analysis reveals increased structural heterogeneity within the TMD regions, which accounts for the comparatively weaker electron microscopy density in these areas. Nevertheless, the overall TMD architecture remains largely consistent between the apo and sucralose-bound states. Guided by the apo structure, reliable modelling of the TMDs in the sucralose-bound receptor was achieved (Extended Data Fig. 4f). Comparing TMD dimer interfaces between apo and sucralose-bound states reveal only minor rearrangements, in which each TM6 undergoes a slight rotation towards the dimer interface (Fig. 4e). Of note, the major conformational changes occur in the intracellular region of TAS1R2-TMD, particularly within the cytoplasmic parts of TM3 and TM7, ICL2, ICL3 and helix 8 (H8), which may be relevant to the G-protein binding (Fig. 4e).

Structural diversity in class C GPCRs

The unique features that we found in the asymmetric TAS1R2–TAS1R3 heterodimer prompted us to compare its structural features with two representative class C GPCRs—the mGlu5 homodimer (PDB ID: 6N52)45 and the GABAB heterodimer (PDB ID: 6VJM)46, both of which have been resolved in full-length apo and agonist-bound states (Fig. 5). Structural comparison of the CRDs (or stalks) and TMDs in the apo state reveals that the TAS1R2–TAS1R3 heterodimer adopts a twisted and more compact packing of CRDs, with an interdomain distance about 23.0 Å (between Cα atoms of TAS1R2 C499 and TAS1R3 C541). This contrasts with the parallel and extended CRD (or stalks) conformation in mGlu5 and GABAB structures (Fig. 5a,c,e). Moreover, in the TAS1R2–TAS1R3 heterodimer, the TMD dimerization interface is primarily mediated by TM6, whereas in mGlu5 and GABAB, TM5 has the dominant role in mediating the TMD interfaces in their apo states (Fig. 5b,d,f).

Fig. 5: Schematic comparison of TAS1R2–TAS1R3 with mGlu5 and GABAB receptors.

a,c,e, Surface presentation illustrates the conformational changes of VFTD, CRD (stalk) and TMD from apo state to agonist-bound state in human sweet taste receptor (a), mGlu5 (c) and GABAB (e). b,d,f, Schematic comparison of TMD rotations, shown from the extracellular view, from apo state to agonist-bound state in human sweet taste receptor (b), mGlu5 (d) and GABAB (f). g, Cartoon illustration of sweet taste receptor in transitions, including inter-subunit and intra-subunit transitions, from the apo state to the sucralose-bound state.

Full size image

In the agonist-bound state, the sweet taste receptor exhibits an agonist recognition mode similar to that of GABAB (PDB: 6UO8)46, in which the agonist binds exclusively to one VFTD (TAS1R2 in sweet taste receptor, GABAB1 in GABAB receptor) (Fig. 5 and Extended Data Fig. 9d). In addition, TM6 was observed to rotate into the dimerization interface of the TMD upon agonist binding in GABAB and mGlu5 (Fig. 5d,f and Extended Data Fig. 9b). By contrast, in the sweet taste receptor, TM6 is already positioned at the dimerization interface in the apo state (Fig. 5b). These findings highlight the diverse mechanisms of ligand recognition and receptor activation among class C GPCRs.

Of note, the sucralose-bound VFTDs of human TAS1R2–TAS1R3 differ from the L-glutamine-bound VFTDs of medaka fish TAS1R2–TAS1R3 (ref. 33), in which L-glutamine binds both subunits, inducing closed conformations (Extended Data Fig. 9e). Additionally, the key residues in the ligand-binding pockets are distinct between human TAS1R2 and medaka fish TAS1R2 (Extended Data Fig. 9f), indicating diverse ligand-recognition mechanisms in TAS1Rs across different species.

Insights on sweet taste receptor activation

The apo and sucralose-bound structures provide important clues for dissecting the activation mechanism of the human sweet taste receptor. Notably, the vertical height of the full-length TAS1R2–TAS1R3 complex in the sucralose-bound state is about 6.0 Å shorter than that in the apo state (Extended Data Fig. 6b). This height difference was first apparent during cryo-EM data processing, and is clearly reflected in the original electron microscopy maps. To further explore the agonist-induced conformational transitions, we performed detailed structural comparisons across domains, revealing a coordinated cascade of conformational transformations from VFTD through CRD to TMD.

Sucralose binding to the TAS1R2-VFTD induces an upward shift in LB2, which is mechanically transduced to the CRD via a disulfide bond between C233 and C513 bridging LB2 and CRD (Fig. 5g and Supplementary Video 1). This upward pulling force from LB2 causes the CRD to bend and drives a lateral shift in both VFTDs (Fig. 5g). Concurrently, the movement of the VFTDs pushes the protruding loop (511–519) of TAS1R3 to the same direction. Together, these pulling and pushing forces promote cooperative conformational adjustment in the CRDs, which are subsequently transmitted to the TMDs (Fig. 5g). This synergistic mechanical coupling between the VFTDs and CRDs facilitates the propagation of small conformational changes across the receptors, which is distinct from the large-scale rearrangements observed in other class C GPCRs (Fig. 5a–f).

To further probe the G-protein coupling interface, we used AlphaFold3 to predict a model for coupling of TAS1R2–TAS1R3 with its cognate G protein, gustducin (Gαgustβ1γ2) (Extended Data Fig. 9g). In this predicted model, gustducin binds exclusively to the cytoplasmic region of TAS1R2 (Extended Data Fig. 9h). Comparison with the sucralose-bound TAS1R2-TMD reveals slight outward displacements of the intracellular loops and H8, accommodating the α5 helix of Gαgust (Fig. 6a). Gαgust docks into a shallow cavity formed by the intracellular ends of TM3, ICLs 1–3 and H8 of TAS1R2, with ICL2 and H8 serving as the primary contact sites (Fig. 6b). The functional relevance of this interface is supported by mutagenesis studies. Mutations of the key interacting residues in TAS1R2—S598ICL1A, F6593.58A, Y669ICL2A or Y756ICL3A—significantly impairs sucralose-induced receptor activation (Fig. 6c and Extended Data Table 2). In addition, the deletion of residues 822–839 in H8 of TAS1R2 abolishes activation of the receptor by sucralose (Fig. 6c and Extended Data Table 2). Of note, the predicted Gαgust binding mode in TAS1R2 is nearly identical to the reported interaction of miniGαs/gust with TAS1R2-TMD37 (Fig. 6d and Extended Data Fig. 9h).

Fig. 6: Predicted G-protein binding mode in TAS1R2.

a, Structural comparison of the TMDs in sucralose-bound TAS1R2 (sandy brown) and AlphaFold3-predicted TAS1R2 (AF3-TAS1R2)–G protein complex (cyan). The α5 helix inserts into a shallow cavity formed by ICL1–3 and H8 of TAS1R2, causing conformational changes in the intracellular region. The TMD in sucralose-bound TAS1R2 is represented as surface. b, Detailed interactions between α5 helix of Gαgust and intracellular regions of TAS1R2-TMD. Key residues are shown as sticks. c, Alanine substitutions of key residues in TAS1R2-TMD or deletion of residues 822–839 in H8 of TAS1R2 in G-protein coupling interface reduce sucralose activity on the sweet taste receptor in the calcium mobilization assay. Data are mean ± s.e.m. of three independent experiments performed in triplicate. d, Structural comparison of AlphaFold3-predicted TAS1R2–Gαgust interface and TAS1R2-TMD–miniGαs/gust interface (PDB: 9NOX). e,f, Structural comparison of the TMDs from TAS1R3 and TAS1R2 in the sucralose-bound state (e), highlighting distinct conformations in their intracellular regions (f).

Source Data

Full size image

Finally, structural alignment of TMDs between TAS1R2 and TAS1R3 highlights subunit-specific differences, particularly distinct conformational movements in ICLs and H8 (Fig. 6e,f). These differences are likely to contribute to the functional asymmetry of the heterodimer, where TAS1R2 serves as the major G-protein-coupling subunit in signal transduction by sweeteners.

Discussion

This study presents the cryo-EM structures of the full-length human sweet taste receptor (TAS1R2–TAS1R3 heterodimer) in both apo and sucralose-bound forms. These structures reveal a previously unknown asymmetric configuration of the sweet taste receptor, highlighting the mechanism of sweetener recognition and distinct conformational changes associated with activation. Our findings expand the understanding of the structural and functional mechanisms of class C GPCRs, particularly those in the chemosensory receptor subfamily.

The receptor has an asymmetric heterodimeric architecture, with TAS1R2 functioning as the primary ligand-binding subunit. Structural analysis, in combination with mutagenesis experiments and molecular dynamics simulations, identified key residues in the TAS1R2-VFTD that mediate sucralose binding. Further molecular docking and functional assays reveal the binding modes of sucrose and the artificial sweetener neotame, uncovering the ligand specificity of the orthosteric pocket of TAS1R2. These findings provide a structural framework for future efforts to modulate sweet taste receptor activity using a broader range of ligands47, including sweet proteins, positive allosteric modulators and negative allosteric modulators, to fully decode the principles of ligand recognition in TAS1R2–TAS1R3 heterodimer.

Notably, polymorphic variants in sweet taste receptor—particularly in TAS1R2, have been shown to influence sweet taste sensitivity or preference in human population48,49,50. For example, the I191V variant in TAS1R2 is associated with reduced glucose excursions51,52. Although I191 is not directly involved in sucralose binding, the I191V mutation may affect the stability of LB2 domain of TAS1R2, thereby altering receptor function. Future studies based on the 3D structures may provide deeper insights into how such polymorphisms influence receptor function and their potential links to metabolic diseases.

A particularly unexpected finding is the modest conformational changes in the receptor upon sucralose binding. Unlike other class C GPCRs, which typically undergo substantial rearrangements in the VFTD, CRD and TMD upon agonist binding, the sweet taste receptor exhibits only subtle shifts of VFTDs and bending of CRDs. This suggests that the receptor may exist in a pre-active conformation, a property also proposed for bitter taste receptors43,44. However, the recent study by Juen et al.37 reported a miniGs/gust-bound ‘inactive pre-coupled state’ model of the TAS1R2-TMD–TAS1R3-TMD complex, in which the TAS1R2-TMD undergoes a rotation of about 50° between the sucralose-bound and G-protein-coupled states. Structural determination of the full-length heterodimer in complex with G protein or other signalling partners will be essential to fully explain the activation conformation and signalling mechanism of the sweet taste receptor. Such studies could help validate the cis-activation model15, in which TAS1R2 mediates both ligand recognition and G-protein coupling. Nonetheless, our data also show that the shared subunit TAS1R3 undergoes conformational changes upon sucralose binding, suggesting a potential role in allosteric regulation. Several studies have reported ligand binding to TAS1R3 (refs. 21,52), and ectopically expressed TAS1R3 has been shown to mediate distinct physiological functions53,54. These findings highlight the need for deeper and broader investigations into the structural and functional roles of TAS1R3, preferably in conjunction with the related umami taste receptor.

Methods
Cell lines

Human embryonic kidney 293S (HEK293S, GnTI−) cells were obtained from American Type Culture Collection (ATCC) and cultured in FreeStyle 293 Expression Medium (Gibco) for protein expression. Sweet taste receptor was expressed through infection of HEK293S GnTI− with viral vectors derived from Spodoptera frugiperda (Sf9) cells. Sf9 cells were grown in ESF 921 medium (Expression systems) at 27 °C with shaking at 125 rpm.

Constructs

The codon-optimized human sweet taste receptor gene (Genewiz) was cloned into the pEG BacMam vector, with their original signal peptides, predicted with SignalP 6.0, were replaced by a haemagglutinin (HA) signal peptide, a glycine-serine linker (GS), a Flag tag and GS Linker (GSA). To further improve surface expression of both subunits, a fluorescent protein mScarlet3 (residues 2–221) was fused to the C terminus of TAS1R2 (residues 26–839) with an optimized GS linker (GSSG), a human rhinovirus 3 C protease cleavage site, and GS linker (SGG). Similarly, a fluorescent protein mNeonGreen (residues 2–236) was fused to the C terminus of TAS1R3 (residues 22–852) with the same construct, followed by a GS linker (GS), a tobacco etch virus (TEV) protease recognition and cleavage site, another GS linker (SG) and an extra 9× His tag for purification.

For functional studies, the native signal peptides of wild-type TAS1R2 and TAS1R3 were replaced with an N-terminal Flag tag to facilitate expression and detection. A chimeric G-protein construct was generated by replacing the C-terminal 44 amino acids of Gα16 with the corresponding sequence from Gαgust, yielding Gα16/gust44. The coding sequences for TAS1R2, TAS1R3 and Gα16/gust44 were subcloned into the pEG BacMam vector for baculovirus generation and subsequent infection of HEK293T cells.

Expression and purification of the human sweet taste receptor

Recombinant baculoviruses were produced following the manufacturer’s protocol using Sf9 cells (Bac-to-Bac expression system, Invitrogen). For protein expression, the HEK293S GnTI− cells were grown at a density of 2.0 × 106 per ml at 37 °C and coinfected with TAS1R2 and TAS1R3 BacMam viruses. 10 mM sodium butyrate was added to the culture medium to enhance the expression after 24 h post-infection. Cells were further cultured at 30 °C for 24 h and collected by centrifugation. The cell pellets were stored at −80 °C until use.

Cell pellets (4 l of cell cultures for each cryo-sample preparation) expressing human TAS1R2–TAS1R3 were lysed in a hypotonic buffer containing 10 mM HEPES pH 7.5, 10 mM MgCl2, 20 mM KCl, with EDTA-free complete protease inhibitor cocktail tablets (Roche). Extensive washes of membranes were performed by repeated Dounce homogenization and centrifugation (142,400g for 30 min) in a high osmotic buffer (hypotonic buffer supplemented with 1 M NaCl). The purified membranes were solubilized in the buffer of 25 mM HEPES pH7.5, 500 mM NaCl, 1.0% (w/v) lauryl maltose neopentyl glycol (LMNG, Anatrace), 0.2% (w/v) cholesterol hemisuccinate (CHS, Sigma-Aldrich), with or without 25 mM sucralose (Sinopharm Chemical Reagent Co.) for 3 h at 4 °C. The supernatant was isolated by ultracentrifugation (142,400g for 40 min), and then incubated with TALON IMAC resin and 10 mM imidazole, at 4 °C overnight. The resin was washed with 20 column volumes of washing buffer I containing 25 mM HEPES pH 7.5, 100 mM NaCl, 10% (v/v) glycerol, 0.05% (w/v) LMNG, 0.01% (w/v) CHS and 30 mM imidazole, with or without 25 mM sucralose, and 15 column volumes of washing buffer II containing 25 mM HEPES pH 7.5, 100 mM NaCl, 10% (v/v) glycerol, 0.03% (w/v) LMNG, 0.0006% (w/v) CHS and 50 mM imidazole, with or without 25 mM sucralose. The protein was eluted using 3 column volumes of elution buffer containing 25 mM HEPES pH 7.5, 100 mM NaCl, 10%(v/v) glycerol, 0.01% (w/v) LMNG, 0.002% (w/v) CHS and 250 mM imidazole, with or without 25 mM sucralose. The eluted protein was collected and concentrated using an Amicon ultra-centrifugal filter (molecular weight cutoff 100 kDa), then injected onto a Superose 6 increase 10/300 GL column pre-equilibrated with a buffer solution containing 25 mM HEPES pH 7.5, 100 mM NaCl, 0.00075% (w/v) LMNG, 0.00025% (w/v) CHS and 0.00025% GDN, with or without 25 mM sucralose. The peak fractions were pooled and concentrated to ~4–5 mg ml−1 for cryo-grid preparation. For the glycerol-free apo-state sample, glycerol was removed from all the related buffers. Typically, 1.0 l of HEK293S GnTI− cells produces about 30.0 µg of target protein.

Cryo-EM grid preparation and data collection

For the preparation of cryo-EM grids for apo (with or without glycerol) and sucralose bound samples, Quantifoil grids (Au 1.2/1.3, 400 mesh) were glow-discharged for 40 s. Three microlitres of concentrated protein was applied to each grid in the Vitrobot Mark IV (ThermoFisher Scientific), blotted for 2.5 s at 4 °C at 100% humidity and plunge-frozen into liquid ethane. Frozen grids were transferred to liquid nitrogen and stored for data acquisition.

For apo-state TAS1R2–TAS1R3 (with or without glycerol), all datasets were acquired using a Titan Krios microscope operated at 300 kV equipped with Falcon 4 Direct Electron Detector (ThermoFisher Scientific). EPU software (v.3.6.0.6389) was used for automated data collection according to standard procedures at a calibrated magnification of 130,000 corresponding to a magnified pixel size of 0.96 Å, employing a defocus range from −0.8 to −2.0 μm. Each movie comprised 40 frames with a total dose of 60 electrons per Å2.

For sucralose-bound TAS1R2–TAS1R3, all datasets were acquired using a Titan Krios microscope operated at 300 kV equipped with a Gatan K3 Summit counting camera and a Gatan Quantum energy filter. Movies were taken in EFTEM nanoprobe mode, with 70 mm C2 aperture, at a calibrated magnification of 105,000 corresponding to a magnified pixel size of 0.832 Å. Each movie was comprised of 40 frames with a total dose of 60 electrons per Å2. The data acquisition was executed through the utilization of SerialEM software (version 3.7)55, with a defocus range from −0.8 to −1.6 μm.

Cryo-EM data processing

Data processing was performed using CryoSPARC 4.6.0 (ref. 56). Dose-fractionated image stacks were performed using patch-based motion correction, with filtered frames used for further processing. CTF parameters for each micrograph were determined via patch-based estimation. Particles auto-picked by CryoSPARC’s Blob Picker from the first 1,000 micrographs were subjected to 2D classification, followed by particle extraction using CryoSPARC’s Template Picker and Topaz57 deep learning algorithm for improved particle picking.

For the apo-state human TAS1R2–TAS1R3 dataset, a total of 30,495 micrographs from two cryo grids were collected. As described previously, initial particles were selected using blob picker, followed by 2D classification to select good particles that were subsequently used as templates for template picker or to train a particle picking model of Topaz. Multiple candidate particle datasets were generated through iterative cycles of template picking and Topaz particle picking workflows. These datasets were used for iterative rounds of 2D classification and heterogeneous refinement to select good particles. A total of 394,712 high-quality particles were selected for ab initio model construction and the best-performing model was selected and subjected to heterogeneous refinement, homogeneous refinement, non-uniform refinement, and local refinement using the 189,062 high-quality particles, resulting in a global resolution of 3.41 Å at a gold-standard Fourier shell correlation (GSFSC) of 0.143. Subsequently, local refinement focused on VFTD was performed to generate a higher-quality VFTD map (partial CRD retention), which achieved a global resolution of 3.18 Å at a GSFSC of 0.143. For the TMD map (partial CRD retention), multiple rounds of 3D classification with local masks focused on the TMD region were performed, and the best particle sets were subjected to local refinement which achieved a resolution of 3.77 Å at a GSFSC of 0.143.

For the sucralose-bound TAS1R2–TAS1R3 dataset, a total of 120,009 movies from nine cryo-grids were collected, and yielding 614,404 non-redundant high-quality particles after a similar iterative particle picking workflow. These particles were subjected to multi-threaded and iterative rounds of ab initio model construction. Ultimately, the ab initio model exhibiting optimal local features in the VFTD region and the full-length model demonstrating optimal global features were selected for refinement, respectively. Following multiple rounds of 3D classification, a VFTD-focused model containing 84,470 particles was selected. A distinct mask was created in UCSF Chimera 1.17.3 for the generation of VFTD (partial CRD retention) map. After local refinement of the same dataset with its full-length map, the VFTD (partial CRD retention) map achieved a resolution of 3.33 Å at a GSFSC of 0.143. Simultaneously, the full-length reconstruction comprising 139,714 particles achieved a global resolution of 3.59 Å at a GSFSC of 0.143 after mask generation and local refinement. All above raw maps were processed with DeepEMhancer58 to facilitate model building and refinement.

For the glycerol-free apo-state TAS1R2–TAS1R3 dataset, a total of 6,412 movies from one cryo-grid were collected, yielding 160,751 non-redundant high-quality particles after a similar iterative particle picking workflow. These particles were subjected to iterative rounds of 2D classification, ab initio model construction and heterogeneous refinement for further selection. A total of 50,250 high-quality particles were selected for final ab initio model construction, non-uniform refinement, and local refinement, resulting in a global resolution of 4.10 Å at a GSFSC of 0.143.

Model building and refinement

The initial heterodimer models of apo-TAS1R2–TAS1R3 and sucralose bound TAS1R2–TAS1R3 were generated using AlphaFold2 (ref. 59) predictions. All models were docked into cryo-EM electron density maps using UCSF Chimera60 (v.1.17). Subsequent iterative manual rebuilding and adjustment were performed in Coot61 (v.0.9.8.5), followed by real-space refinement with secondary structure constraints implemented in PHENIX62 (v.1.20). The ligand binding pose within the pocket was further validated using Gemspot63. The model statistics were validated using MolProbity64. Final refinement statistics are presented in Extended Data Table 1 and UCSF Chimera (v.1.17), UCSF Chimera X (v.1.8) and PyMOL (v.3.0.0) ware used to generate the structural figures.

AlphaFold2 model and AlphaFold3 prediction

The monomer TAS1R2 (AF-Q8TE23-F1-v4) and TAS1R3 (AF-Q7RTX0-F1-v4) AlphaFold2 models used for initial model building were obtained from the AlphaFold Protein Structure Database (https://www.alphafold.ebi.ac.uk/)65. AlphaFold3 models of the TAS1R2–TAS1R3 heterodimer were generated using the AlphaFold Server (https://alphafoldserver.com)36. For the TAS1R2–TAS1R3 heterodimer model, full-length sequences of TAS1R2 and TAS1R3 were submitted as inputs. To model the TAS1R2–TAS1R3 complex coupled with the G-protein heterotrimer, we additionally included the sequences of Gαgust, Gβ1 and Gγ2. Among five generated models of each session, the highest-ranking output was selected for the related analysis.

Protein property analysis

GenScript ExpressPlus PAGE precast gels were employed and stained with Coomassie Brilliant Blue using the GenScript eStain LG automated gel staining equipment. For western blotting, proteins were electrophoretically transferred onto polyvinylidene fluoride membranes using the GenScript eBlot L1 rapid transfer system under standard protocol. The membranes were sequentially incubated in TBST (Tris-buffered saline with 0.1% Tween 20) containing 5% (w/v) skim milk for 1 h at room temperature. Subsequently, membranes were incubated with either Anti-TAS1R2 (Absin Bioscience, abs139799) (1:2,500 dilution) or Anti-TAS1R3 (Invitrogen, PA5-98522) (1:2,500 dilution) primary antibodies for 1 h at room temperature, followed by a horseradish peroxidase-conjugated anti-rabbit secondary antibody (Cell Signaling Technology, 7074S) (1:1,000 dilution) under the same conditions. The SDS–PAGE and western blot results are shown in Supplementary Fig. 1.

Analytical size-exclusion chromatography (aSEC or HPLC) was performed using an Agilent 1260 Bioinert HPLC system equipped with a Sepax Nanofilm SEC-250 column (4.6 × 250 mm) at a flow rate of 0.5 ml min−1. The column was pre-conditioned according to the manufacturer’s protocol. The protein sample was centrifuged at 15,871g for 10 min and loaded into the HPLC with running buffer consisted of 25 mM HEPES pH 7.5, 100 mM NaCl, 10% (v/v) glycerol, 0.01% (w/v) LMNG, and 0.002% (w/v) CHS. The protein was detected by the standard UV absorbance signals and the fluorescence detectors. Fluorescence detection parameters were set with excitation/emission wavelengths of 569/592 nm for TAS1R2–mScarlet3 and 506/517 nm for TAS1R3–mNeonGreen.

Intracellular calcium mobilization assay

Intracellular calcium mobilization assay was performed using a BacMam expression system, similar to that used for structural analysis. Baculoviruses were produced in S. frugiperda cells for wild type TAS1R2–TAS1R3 and mutants, and then used to infect mammalian cells for functional characterization. HEK293T cells (ATCC, CRL-11268; Mycoplasma-free) were seeded at a density of 1 × 106 cells in a 6-well plate and incubated overnight at 37 °C with 5% CO2 in DMEM medium (Gibco, C11995500CP) supplemented with 10% (v/v) fetal bovine serum (FBS; Gibco, 10091148).

On the day of infection, the medium was replaced with DMEM + GlutaMAX low glucose medium (Gibco, 10567014) supplemented with 10% (v/v) FBS and 2 mM lactisole. Cells were infected with baculovirus encoding wild-type TAS1R2 (or mutants), wild-type TAS1R3 (or mutants) and Gα16/gust44. After 24 h, cells were collected by trypsinization and seeded into black-sided, clear-bottom 384-well plates (Greiner Bio-one, 781091) at a density of 15,000 cells per well, with the same infection medium.

On the day of assay, the growth medium was removed and cells were incubated with 20 μl per well of 1× Calcium 6 dye (prepared in DPBS with 20 mM HEPES buffer) for 1.5 h at room temperature in the dark. Fluorescence was measured using the FLIPR microplate reader, which was programmed to collect 10 baseline readings (1 reading per second) prior to compound addition. Then, 10 μl of a 3× compound solution (prepared in DPBS with 20 mM HEPES buffer with 0.1% BSA) was recorded for an additional 2 min to assess agonist activity. Data were analysed by nonlinear regression using GraphPad Prism 8.0.

Fluorescence-activated cell sorting assay

The cell surface expression of wild-type and mutant variants was evaluated using a fluorescence-activated cell sorting assay. In brief, 10 μl of transfected cells were incubated with 10 μl of mouse anti-Flag (M2) fluorescein isothiocyanate (PE) conjugated antibody (Abcam) for 20 min at 4 °C. After incubation, 180 μl of PBS was added to the cells for dilution. The surface expression was subsequently analysed by measuring the fluorescence intensity using a Guava EasyCyte HT system (Millipore). The data were analysed by GuavaSoft 3.1.1, Guava ExpressPlus panel.

Molecular docking

The molecular docking study was performed using Schrödinger’s Glide module (2022-3 release)66. The target protein structure was prepared via Protein Preparation Wizard, including completion of missing side chains (Prime module)67, hydrogen bond optimization with PROPKA at pH 7.4, and energy minimization using OPLS4 force field68 (RMSD convergence criterion 0.3 Å). Ligands sucrose and neotame were processed through LigPrep with Epik state generation at pH 7.4 ± 2.0, enumerating possible stereoisomers and ionization states. The docking grid was centred on the binding pose of sucralose, with box dimensions validated through re-docking of the native ligand. For sucrose, flexible docking was executed in standard precision mode with default parameters. For the larger ligand neotame, induced fit docking69 was implemented using Schrödinger’s induced fit docking protocol to account for binding site flexibility. All docking results were evaluated by GlideScore (GScore) and Prime MM/GBSA binding free energy calculations.

Molecular dynamics simulations

Molecular dynamics simulations were performed using Schrödinger’s Desmond70 with optimized computational protocols. To balance accuracy and efficiency for the large protein-ligand system, we strategically truncated the structure while preserving functional domains critical for ligand binding: TAS1R2 (N terminus to Cys520) and TAS1R3 (N terminus to Cys524), with terminal residues capped by ACE/NME groups. The prepared complex, after hydrogen bond optimization and OPLS4 force field-based energy minimization, was solvated in an orthorhombic periodic box (SPC water model, ≥10 Å buffer distance) containing 0.15 M NaCl for neutralization, resulting in a final system size of ~110,000 atoms. Following Desmond’s default equilibration protocol, three independent 500 ns production simulations were performed under NPT ensemble (300 K, 1.01325 bar) using default parameters with a 2-fs timestep. Ligand binding modes were systematically assessed through Schrödinger’s simulation interaction diagram analysis, while ligand conformational stability was quantified via RMSD calculations implemented in VMD71.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The atomic coordinates and cryo-EM density maps for apo-TAS1R2–TAS1R3 heterodimer and sucralose-bound TAS1R2–TAS1R3 heterodimer have been deposited in the Protein Data Bank and Electron Microscopy Data Bank, respectively. The accession codes for TAS1R2–TAS1R3 heterodimer reported in this paper are 9UT8 and EMDB-64484 (apo full-length TAS1R2–TAS1R3); 9UT9 and EMDB-64485 (apo-TAS1R2-VFTD–TAS1R3); 9UTA and EMDB-64486 (apo TMD of TAS1R2–TAS1R3); 9UTB and EMDB-64487 (sucralose-bound full-length TAS1R2–TAS1R3); 9UTC and EMDB-64488 (sucralose-bound TAS1R2-VFTD–TAS1R3). Source data are provided with this paper.

References

Nelson, G. et al. Mammalian sweet taste receptors. Cell 106, 381–390 (2001).

Article
 
CAS
 
PubMed
 
Google Scholar

Zhao, G. Q. et al. The receptors for mammalian sweet and umami taste. Cell 115, 255–266 (2003).

Article
 
CAS
 
PubMed
 
Google Scholar

Lindemann, B. Taste reception. Physiol. Rev. 76, 719–766 (1996).

Article
 
CAS
 
PubMed
 
Google Scholar

Lindemann, B. Receptors and transduction in taste. Nature 413, 219–225 (2001).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Lee, H., Macpherson, L. J., Parada, C. A., Zuker, C. S. & Ryba, N. J. P. Rewiring the taste system. Nature 548, 330–333 (2017).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Beauchamp, G. K. Basic taste: a perceptual concept. J. Agric. Food Chem. 67, 13860–13869 (2019).

Article
 
CAS
 
PubMed
 
Google Scholar

Chandrashekar, J., Hoon, M. A., Ryba, N. J. & Zuker, C. S. The receptors and cells for mammalian taste. Nature 444, 288–294 (2006).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Max, M. et al. encoding a new candidate taste receptor, is allelic to the sweet responsiveness locus. Nat. Genet. 28, 58–63 (2001).

Article
 
CAS
 
PubMed
 
Google Scholar

Montmayeur, J. P., Liberles, S. D., Matsunami, H. & Buck, L. B. A candidate taste receptor gene near a sweet taste locus. Nat. Neurosci. 4, 492–498 (2001).

Article
 
CAS
 
PubMed
 
Google Scholar

Adler, E. et al. A novel family of mammalian taste receptors. Cell 100, 693–702 (2000).

Article
 
CAS
 
PubMed
 
Google Scholar

Chandrashekar, J. et al. T2Rs function as bitter taste receptors. Cell 100, 703–711 (2000).

Article
 
CAS
 
PubMed
 
Google Scholar

Matsunami, H., Montmayeur, J. P. & Buck, L. B. A family of candidate taste receptors in human and mouse. Nature 404, 601–604 (2000).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Tomonari, H. et al. Gα-gustducin is extensively coexpressed with sweet and bitter taste receptors in both the soft palate and fungiform papillae but has a different functional significance. Chem. Senses 37, 241–251 (2012).

Article
 
CAS
 
PubMed
 
Google Scholar

Yarmolinsky, D. A., Zuker, C. S. & Ryba, N. J. Common sense about taste: from mammals to insects. Cell 139, 234–244 (2009).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Damak, S. et al. Detection of sweet and umami taste in the absence of taste receptor T1r3. Science 301, 850–853 (2003).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Li, X. D. et al. Human receptors for sweet and umami taste. Proc. Natl Acad. Sci. USA 99, 4692–4696 (2002).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Reiner, A. & Levitz, J. Glutamatergic signaling in the central nervous system: ionotropic and metabotropic receptors in concert. Neuron 98, 1080–1098 (2018).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Heaney, C. F. & Kinney, J. W. Role of GABAB receptors in learning and memory and neurological disorders. Neurosci. Biobehav. Rev. 63, 1–28 (2016).

Article
 
CAS
 
PubMed
 
Google Scholar

Hannan, F. M., Kallay, E., Chang, W., Brandi, M. L. & Thakker, R. V. The calcium-sensing receptor in physiology and in calcitropic and noncalcitropic diseases. Nat. Rev. Endocrinol. 15, 33–51 (2018).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Liu, H., Li, Y. & Gao, Y. in G Protein-Coupled Receptors—Part B 1st edn, Vol. 195 (ed. Shukla, A. K.) Ch. 4 (Academic Press, 2023).

Nie, Y., Vigues, S., Hobbs, J. R., Conn, G. L. & Munger, S. D. Distinct contributions of T1R2 and T1R3 taste receptor subunits to the detection of sweet stimuli. Curr. Biol. 15, 1948–1952 (2005).

Article
 
CAS
 
PubMed
 
Google Scholar

Jiang, P. H. et al. The cysteine-rich region of T1R3 determines responses to intensely sweet proteins. J. Biol. Chem. 279, 45068–45075 (2004).

Article
 
CAS
 
PubMed
 
Google Scholar

Walters, D. E. & Hellekant, G. Interactions of the sweet protein brazzein with the sweet taste receptor. J. Agr. Food Chem. 54, 10129–10133 (2006).

Article
 
CAS
 
Google Scholar

Servant, G. et al. Positive allosteric modulators of the human sweet taste receptor enhance sweet taste. Proc. Natl Acad. Sci. USA 107, 4746–4751 (2010).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Belloir, C., Jeannin, M., Karolkowski, A., Scott, C. & Briand, L. A receptor-based assay to study the sweet and bitter tastes of sweeteners and binary sweet blends: the SWEET project. Chem. Senses 49, bjae041 (2024).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Zuker, C. S. Food for the brain. Cell 161, 9–11 (2015).

Article
 
CAS
 
PubMed
 
Google Scholar

Nieh, E. H. et al. Decoding neural circuits that control compulsive sucrose seeking. Cell 160, 528–541 (2015).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Margolskee, R. F. et al. T1R3 and gustducin in gut sense sugars to regulate expression of Na-glucose cotransporter 1. Proc. Natl Acad. Sci. USA 104, 15075–15080 (2007).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Kyriazis, G. A., Soundarapandian, M. M. & Tyrberg, B. Sweet taste receptor signaling in beta cells mediates fructose-induced potentiation of glucose-stimulated insulin secretion. Proc. Natl Acad. Sci. USA 109, E524–E532 (2012).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Ren, X., Zhou, L., Terwilliger, R., Newton, S. S. & de Araujo, I. E. Sweet taste signaling functions as a hypothalamic glucose sensor. Front. Integr. Neurosci. https://doi.org/10.3389/neuro.07.012.2009 (2009).

Janssen, S. et al. Bitter taste receptors and α-gustducin regulate the secretion of ghrelin with functional effects on food intake and gastric emptying. Proc. Natl Acad. Sci. USA 108, 2094–2099 (2011).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Lee, S. J., Depoortere, I. & Hatt, H. Therapeutic potential of ectopic olfactory and taste receptors. Nat. Rev. Drug Discov. 18, 116–138 (2019).

Article
 
CAS
 
PubMed
 
Google Scholar

Nuemket, N. et al. Structural basis for perception of diverse chemical substances by T1r taste receptors. Nat. Commun. 8, 15530 (2017).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Gadella, T. W. J. Jr et al. mScarlet3: a brilliant and fast-maturing red fluorescent protein. Nat. Methods 20, 541–545 (2023).

Article
 
PubMed
 
Google Scholar

Shaner, N. C. et al. A bright monomeric green fluorescent protein derived from Branchiostoma lanceolatum. Nat. Methods 10, 407–409 (2013).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493–500 (2024).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Juen, Z. et al. The structure of human sweetness. Cell https://doi.org/10.1016/j.cell.2025.04.021 (2025).

Masuda, K. et al. Characterization of the modes of binding between human sweet taste receptor and low-molecular-weight sweet compounds. PLoS ONE 7, e35380 (2012).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Zhang, F. et al. Molecular mechanism of the sweet taste enhancers. Proc. Natl Acad. Sci. USA 107, 4752–4757 (2010).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Assadi-Porter, F. M. et al. Key amino acid residues involved in multi-point binding interactions between brazzein, a sweet protein, and the T1R2–T1R3 human sweet receptor. J. Mol. Biol. 398, 584–599 (2010).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Liu, B. et al. Molecular mechanism of species-dependent sweet taste toward artificial sweeteners. J. Neurosci. 31, 11070–11076 (2011).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Servant, G., Kenakin, T., Zhang, L., Williams, M. & Servant, N. The function and allosteric control of the human sweet taste receptor. Adv. Pharmacol. 88, 59–82 (2020).

Article
 
CAS
 
PubMed
 
Google Scholar

Hu, X. et al. Bitter taste TAS2R14 activation by intracellular tastants and cholesterol. Nature 631, 459–466 (2024).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Xu, W. et al. Structural basis for strychnine activation of human bitter taste receptor TAS2R46. Science 377, 1298–1304 (2022).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Koehl, A. et al. Structural insights into the activation of metabotropic glutamate receptors. Nature 566, 79–84 (2019).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Shaye, H. et al. Structural basis of the activation of a metabotropic GABA receptor. Nature 584, 298–303 (2020).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

DuBois, G. E. & Prakash, I. Non-caloric sweeteners, sweetness modulators, and sweetener enhancers. Annu. Rev. Food Sci. Technol. 3, 353–380 (2012).

Article
 
CAS
 
PubMed
 
Google Scholar

Valente, C. et al. Genes from the TAS1R and TAS2R families of taste receptors: looking for signatures of their adaptive role in human evolution. Genome Biol. Evol. 10, 1139–1152 (2018).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Dias, A. G. et al. Variation in the TAS1R2 gene, sweet taste perception and intake of sugars. J. Nutrigenet. Nutrigenomics 8, 81–90 (2015).

CAS
 
PubMed
 
Google Scholar

Melis, M., Mastinu, M., Naciri, L. C., Muroni, P. & Tomassini Barbarossa, I. Associations between sweet taste sensitivity and polymorphisms (SNPs) in the TAS1R2 and TAS1R3 genes, gender, PROP taster status, and density of fungiform papillae in a genetically homogeneous Sardinian cohort. Nutrients 14, 4903 (2022).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Serrano, J. et al. The Ile191Val is a partial loss-of-function variant of the TAS1R2 sweet-taste receptor and is associated with reduced glucose excursions in humans. Mol. Metab. 54, 101339 (2021).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Dubovski, N., Ben-Shoshan Galeczki, Y., Malach, E. & Niv, M. Y. Sensitivity of human sweet taste receptor subunits T1R2 and T1R3 to activation by glucose enantiomers. Chem. Senses 48, bjad005 (2023).

Article
 
PubMed
 
Google Scholar

Yoshimura, A. et al. Taste receptor type 1 member 3 in osteoclasts regulates osteoclastogenesis via detection of glucose. J. Biol. Chem. 301, 108273 (2025).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Schmidt, P. et al. Tas1R3 dependent and independent recognition of sugars in the urethra and the role of tuft cells in this process. Adv. Biol. 8, e2400117 (2024).

Article
 
Google Scholar

Mastronarde, D. N. Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol. 152, 36–51 (2005).

Article
 
PubMed
 
Google Scholar

Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017).

Article
 
CAS
 
PubMed
 
Google Scholar

Bepler, T. et al. Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs. Nat. Methods 16, 1153–1160 (2019).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Sanchez-Garcia, R. et al. DeepEMhancer: a deep learning solution for cryo-EM volume post-processing. Commun. Biol. 4, 874 (2021).

Article
 
PubMed
 
PubMed Central
 
Google Scholar

Tunyasuvunakool, K. et al. Highly accurate protein structure prediction for the human proteome. Nature 596, 590–596 (2021).

Article
 
ADS
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D 60, 2126–2132 (2004).

Article
 
ADS
 
PubMed
 
Google Scholar

Adams, P. D. et al. Recent developments in the PHENIX software for automated crystallographic structure determination. J. Synchrotron Radiat. 11, 53–55 (2004).

Article
 
CAS
 
PubMed
 
Google Scholar

Robertson, M. J., van Zundert, G. C. P., Borrelli, K. & Skiniotis, G. GemSpot: a pipeline for robust modeling of ligands into cryo-EM maps. Structure 28, 707–716.e703 (2020).

Article
 
CAS
 
PubMed
 
PubMed Central
 
Google Scholar

Chen, V. B. et al. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D 66, 12–21 (2010).

Article
 
ADS
 
CAS
 
PubMed
 
Google Scholar

Varadi, M. et al. AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic Acids Res. 50, D439–D444 (2022).

Article
 
CAS
 
PubMed
 
Google Scholar

Yang, Y. et al. Efficient exploration of chemical space with docking and deep learning. J. Chem. Theory Comput. 17, 7106–7119 (2021).

Article
 
CAS
 
PubMed
 
Google Scholar

Jacobson, M. P. et al. A hierarchical approach to all-atom protein loop prediction. Proteins 55, 351–367 (2004).

Article
 
CAS
 
PubMed
 
Google Scholar

Lu, C. et al. OPLS4: improving force field accuracy on challenging regimes of chemical space. J. Chem. Theory Comput. 17, 4291–4300 (2021).

Article
 
CAS
 
PubMed
 
Google Scholar

Farid, R., Day, T., Friesner, R. A. & Pearlstein, R. A. New insights about HERG blockade obtained from protein modeling, potential energy mapping, and docking studies. Bioorg. Med. Chem. 14, 3160–3173 (2006).

Article
 
CAS
 
PubMed
 
Google Scholar

Bowers, K. J. et al. Scalable algorithms for molecular dynamics simulations on commodity clusters. In Proc. 2006 ACM/IEEE Conference on Supercomputing (ACM, IEEE, 2006).

Humphrey, W., Dalke, A. & Schulten, K. VMD: visual molecular dynamics. J. Mol. Graph. 14, 33–38 (1996).

Article
 
CAS
 
PubMed
 
Google Scholar

Download references

Acknowledgements

This work was supported by the National Natural Science Foundation of China grants 32271262 and 32471267 (T.H.), and 32230026 (Z.-J.L.), ShanghaiTech University School of Life Science Development Fund Project (Z.-J.L.), National Key Research and Development Program of China grant 2022YFA1302903 (T.H.) and 2023YFC2509100 (T.H.), Shanghai’s Top Priority Research Center 2022ZZ01017 (Z.-J.L.), AI4S Program at ShanghaiTech University (Z.-J.L.), Shanghai Frontiers Science Center for Biomacromolecules and Precision Medicine at ShanghaiTech University. We thank the Shanghai Municipal Government and ShanghaiTech University for financial support. We thank S.-W. Zhao, Y.-R. Wu and F. Zhao for discussions and suggestions for sucralose binding pose modelling; N. Chen, S.-W. Hu and P. Si at cell expression core; and Q.-Y. Shi at protein purification core of iHuman Institute for their support. The cryo-EM data were collected at the Bio-Electron Microscopy Facility of ShanghaiTech University, with the assistance of L. Wang and Q.-Q. Sun.

Author information
Author notes

These authors contributed equally: Zongjun Shi, Weixiu Xu, Lijie Wu, Xiaolei Yue

Authors and Affiliations

iHuman Institute, ShanghaiTech University, Shanghai, China

Zongjun Shi, Weixiu Xu, Lijie Wu, Xiaolei Yue, Shenhui Liu, Jinyi Zhang, Bing Meng, Lianghao Zhao, Xiaoyan Liu, Junlin Liu, Zhi-Jie Liu & Tian Hua

School of Life Science and Technology, ShanghaiTech University, Shanghai, China

Zongjun Shi, Xiaolei Yue, Shenhui Liu, Lianghao Zhao, Zhi-Jie Liu & Tian Hua

Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing, China

Wei Ding

Contributions

Z.-J.L. and T.H. initiated the project. Z.S., W.X. and X.Y. designed the expression constructs, purified the receptor complexes, prepared the final samples for cryo-EM data collection, participated in the electron microscopy data processing, and participated in the figure and manuscript preparation. L.W. performed the electron microscopy data processing and structure determination. W.X. performed the functional assays. S.L. performed the molecular docking and molecular dynamics simulations. W.D. assisted with the electron microscopy data processing. J.Z. and B.M. performed the initial constructs and sample optimization. L.Z. assisted with the construct design with fluorescent protein. X.L. performed the mammalian cell expression. J.L. performed the insect cell expression. Z.-J.L. and T.H. conceived and supervised the research, analysed the structures and wrote the manuscript with the input from all authors.

Corresponding authors

Correspondence to Zhi-Jie Liu or Tian Hua.

Ethics declarations
Competing interests

The authors declare no competing interests.

Peer review
Peer review information

Nature thanks Loïc Briand and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer review reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data figures and tables
Extended Data Fig. 1 Construct design and protein property analysis of human TAS1R2-TAS1R3.

a, Schematic diagram of the modified constructs for human TAS1R2 and TAS1R3 used for structure determination. b, The representative SDS-PAGE and western blotting of the TAS1R2-TAS1R3 heterodimer sample (≥ 4 independent experiments). The presence of TAS1R2 and TAS1R3 was confirmed by western blotting with anti-TAS1R2 and anti-TAS1R3 antibody, respectively. c, Analytical size-exclusion chromatograph (aSEC) analysis of the TAS1R2-TAS1R3 heterodimer sample, using UV280 absorbance and fluorescence detection. Left vertical axis: UV280 absorbance values (protein elution profile); Right vertical axis: normalized fluorescence values. Fluorescence detection parameters: TAS1R2-mScarlet3: Excitation 569 nm/ Emission 592 nm; for TAS1R3-mNeonGreen: Excitation 506 nm/ Emission 517 nm. d, Dose-response curves of glycerol on WT human sweet taste receptor measured in calcium mobilization assay. Data are mean ± s.e.m. of three independent experiments performed in triplicate.

Source Data

Extended Data Fig. 2 Structure determination of human TAS1R2-TAS1R3 heterodimer in apo state.

a, Size exclusion chromatogram profile of the apo TAS1R2-TAS1R3 heterodimer. b, Representative cryo-EM micrograph (30,495 micrographs) (scale bars, 50 nm). c, Two-dimensional (2D) class averages of apo TAS1R2-TAS1R3 particles, showing different particle view. d, Data processing pipeline for the apo TAS1R2-TAS1R3 complex, including 2D classification and 3D reconstruction workflow and final 3D density map colored according to local resolution (Å). e, Fourier shell correlation (FSC) curves for the final 3D density maps of the apo TAS1R2-TAS1R3 complex. Also shown are the Euler Angle Distributions of particles used for the final 3D reconstruction.

Extended Data Fig. 3 Structure determination of human TAS1R2-TAS1R3 heterodimer in sucralose bound state.

a, Size exclusion chromatogram profile of the TAS1R2-TAS1R3 in complex with sucralose. b, Representative cryo-EM micrograph (scale bars, 50 nm) (120,009 micrographs). c, Two-dimensional (2D) class averages of TAS1R2-TAS1R3 in complex with sucralose, showing different particle views. d, Data processing pipeline for the sucralose bound TAS1R2-TAS1R3 complex, including 2D classification and 3D reconstruction workflow and final 3D density map colored according to local resolution (Å). e, Fourier shell correlation (FSC) curves for the final 3D density maps of the TAS1R2-TAS1R3 in complex with sucralose. Euler Angle Distributions of the particles used for the final 3D reconstruction are also shown.

Extended Data Fig. 4 Representative density maps for apo- and sucralose-bound TAS1R2-TAS1R3 structures.

a, Cryo-EM density maps of representative regions in VFTD and CRD of apo TAS1R2-TAS1R3 heterodimer structure. b, Glycosylation sites in VFTD and CRD of the apo TAS1R2-TAS1R3 heterodimer, shown in ball sticks representation. Extra EM density around the side chains of these glycosylation sites is displayed. c, Cryo-EM density maps of representative regions in VFTD, CRD, TMD and ECL2 of sucralose-bound TAS1R2-TAS1R3 heterodimer structure. d, Cryo-EM density maps of loops in CRDs that insert into the VFTD dimerization interface in apo TAS1R2-TAS1R3 heterodimer structure. e, Cryo-EM density maps of extracellular regions of TMDs including the conserved disulfide bonds between ECL2 and TM3 in TAS1R2 and TAS1R3, respectively, in apo-state structure. f, Cryo-EM density maps for helices TM1-TM7 of transmembrane domain, ICL2 and ECL2 of TAS1R2 and TAS1R3, respectively, in apo-state structure. g, Structural comparison of TAS1R2 and TAS1R3 in apo state aligned with TMD.

Extended Data Fig. 5 Structure determination of human TAS1R2-TAS1R3 heterodimer in glycerol-free condition and structural comparisons with reported sucralose-bound sweet taste receptor.

a, Size exclusion chromatogram profile of the apo TAS1R2-TAS1R3 heterodimer in the glycerol-free buffer condition. b, Representative cryo-EM micrograph (scale bars, 50 nm) (6,412 micrographs). c, Two-dimensional (2D) class averages of apo TAS1R2-TAS1R3 particles in the glycerol-free buffer condition, showing different particle view. d, The raw map for apo TAS1R2-TAS1R3 in the glycerol-free buffer condition. e, The raw map for apo TAS1R2-TAS1R3 in the glycerol-free buffer condition aligns well with that of apo TAS1R2-TAS1R3 map. f, The apo TAS1R2-TAS1R3 structure fits well into the two raw maps. g, Structural comparisons of reported full-length TAS1R2-TAS1R3 (PDB: 9NOR) with apo TAS1R2-TAS1R3 (left) (RMSD: 2.07 Å) and sucralose-bound TAS1R2-TAS1R3 (right) (RMSD: 0.99 Å), respectively.

Extended Data Fig. 6 Structural features of full-length TAS1R2-TAS1R3 heterodimers.

a, Molecular surface of the apo-state TAS1R2-TAS1R3 complex, showing the plane of heterodimer interfaces for the VFTD and TMD. Structural elements involved in heterodimer formation are highlighted in cartoon representation. b, Structural comparison of TAS1R2 and TAS1R3 in apo- and sucralose-bound states. TAS1R3 is shown as cartoon and TAS1R2 is shown as transparent surface (left). TAS1R2 is shown as cartoon and TAS1R3 is shown as transparent surface (right). c, Conformational changes of VFTDs in apo- and sucralose-bound states viewed from the extracellular top. d, The bending of CRDs in TAS1R2 and TAS1R3 from apo- to sucralose-bound states.

Extended Data Fig. 7 Sequence alignment of human TAS1R2 and TAS1R3.

The secondary structure is shown.

Extended Data Fig. 8 Molecular dynamics simulations of sucralose binding pose in TAS1R2 and the docking poses of sucrose and neotame in sweet taste receptor.

a, Cryo-EM density maps for sucralose and surrounding residues in its binding pocket. b, RMSD of sucralose during 500 ns simulations of sucralose-TAS1R2-TAS1R3 (VFTDs). The colors denote for 3 independent runs. c, Structural comparisons with reported sucralose-bound TAS1R2-TAS1R3 VFTDs (PDB: 9NOV) and sucralose adopts a similar binding pose in TAS1R2-VFTD. d, Comparison of docking poses of sucrose and neotame with the experimental determined sucralose binding pose in TAS1R2-VFTD. e-f, The docking pose of neotame (e) and sucrose (f) in TAS1R2-VFTD. g-h, Dose-response curves of key residues mutations in LB1 and LB2 of TAS1R2-VFTD in response to stimulation with neotame (g) and sucrose (h) in calcium mobilization assay. Data are mean ± s.e.m. of three independent experiments performed in triplicate.

Source Data

Extended Data Fig. 9 Structural comparison of VFTD and TMD in TAS1R2-TAS1R3 structure with other class C GPCRs and AlphaFold 3 predicted model of TAS1R2-TAS1R3 in complex with G protein.

a, Electrostatic potential mapped onto the cleft formed by LB1 and LB2 in the VFTDs of TAS1R2 and TAS1R3. b, The structural comparison of TMD between agonist-bound mGlu5 homodimer and sucralose bound TAS1R2-TAS1R3 heterodimer. c, ICL2 in TAS1R2 and TAS1R3 forms a similar short helix conformation with that in bitter taste receptors TAS2R14 and TAS2R46. d, Conformational changes of VFTDs from apo- to agonist-bound states in TAS1R2, GABAB1, mGlu5 and CaSR. e-f, Structural comparison of VFTDs between L-Gln bound medaka fish TAS1R2-TAS1R3 (PDB: 5X2M) and sucralose bound human TAS1R2-TAS1R3 (e), the ligands binding modes in TAS1R2 (f). g, The overall structure model of TAS1R2-TAS1R3 in complex with Gustducin (Gαgustducin β1γ2) predicted by AlphaFold3. h, AlphaFold3 predicted TAS1R2-Gαgustducin complex model is similar with that of TAS1R2TMD-miniGαs/gust (PDB: 9NOX).

Extended Data Table 1 Cryo-EM data collection, model refinement, and validation statistics of TAS1R2-TAS1R3 heterodimer structures
Full size table
Extended Data Table 2 Sweetener molecules-induced intracellular calcium mobilization of wild-type (WT) and mutants of TAS1R2 or TAS1R3
Full size table
Extended Data Table 3 Relative surface expression of wild-type and mutants
Full size table
Supplementary information
Supplementary Figure

Uncropped SDS–PAGE and western blotting for Extended Data Fig. 1.

Reporting Summary
Peer Review File
Supplementary Video 1

The conformational changes between full-length apo-TAS1R2–TAS1R3 and sucralose-bound TAS1R2–TAS1R3 structures.

Supplementary Video 2

The binding mode of sucralose in TAS1R2-VFTD and the conformational changes of key residues from apo to sucralose-bound states.

Supplementary Video 3

The conformational changes of VFTDs from apo to sucralose-bound states in sweet taste receptor.

Source data
Source Data Figs. 1,2, 3 and 6 and Source Data Extended Data Figs. 1 and 8
Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article
Cite this article

Shi, Z., Xu, W., Wu, L. et al. Structural and functional characterization of human sweet taste receptor. Nature 645, 801–808 (2025). https://doi.org/10.1038/s41586-025-09302-6

Download citation

Received
28 April 2025

Accepted
17 June 2025

Published
24 June 2025

Version of record
23 July 2025

Issue date
18 September 2025

DOI
https://doi.org/10.1038/s41586-025-09302-6

Share this article

Anyone you share the following link with will be able to read this content:

Get shareable link

Provided by the Springer Nature SharedIt content-sharing initiative

Subjects
Electron microscopy
Receptor pharmacology
Taste receptors
This article is cited by
Structure and activation mechanism of human sweet taste receptor
Haolan WangXiao ChenChia-Hsueh Lee

Cell Research (2025)

Cancel

Preview · live