Signaling profiles and structural characterization of GPR174

To investigate the signaling profiles of GPR174 in response to LysoPS, we employed NanoLuc Binary Technology (NanoBiT), a robust method for quantifying the engagement of multiple G proteins by GPCRs [23,27]. We observed that LysoPS activation of GPR174 exhibits a strong coupling preference for G_s over G_i, with no detectable response through G_q and a weak response through G₁₃ (Fig 1A and 1B; S1 and S2 Tables). These results indicate that LysoPS activates G_s, consistent with previous pharmacological studies [24,28]. Notably, the significantly stronger signaling through the G_s pathway compared to G_i supported efficient assembly and structural determination of the GPR174-G_s complex [20,26,29]. To further characterize the structural basis of LysoPS-triggered activation and G protein selectivity, we applied the same NanoBiT tethering strategy to assemble both GPR174-G_s and GPR174-G_i complexes [27]. Finally, the structures of LysoPS-bound GPR174-G_s and GPR174-G_i complexes were solved at resolutions of 2.0 Å and 3.4 Å, respectively (Figs 1C, 1D, S1, and S2; S3 Table). These high-resolution density maps enabled clear visualization of LysoPS binding in the GPR174-G_s complex and the overall architecture of the seven-transmembrane domain of GPR174 in both G_s- and G_i-bound structures.

Identification of hydration-mediated network in GPR174

The high-resolution cryo-EM structure of the LysoPS (18:0)-bound GPR174-G_s complex reveals a well-defined internal hydration network comprising 14 water molecules, which are primarily distributed across three functional regions: (i) five water molecules (W_L1–W_L5) located in the orthosteric binding pocket stabilize the polar headgroup of LysoPS, initiating signal transduction; (ii) eight water molecules (W_S1–W_S8) participate in the formation and stabilization of key activation motifs, including the canonical DRY and NPxxY motifs, as well as other less conserved elements; (iii) a single water molecule (W_G1) mediates an interaction network at the interface between the intracellular pocket of GPR174 and the C-terminal α5 helix of Gα_s (Figs 2A, 2B, and S3A).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Hydration-mediated signaling network of the GPR174-G_s complex.

(A) The overall structure of GPR174 with the LysoPS ligand (purple), with ordered internal water molecules shown as red spheres. (B) Structural depiction of the hydration-mediated signaling network in the GPR174-G_s complex. Hydrogen bonds are shown as black dashed lines; hydration-associated residues are depicted as pale green sticks. (C) Effects of mutations at key residues involved in the hydration-mediated signaling network on the logEC₅₀ of LysoPS (18:0)-induced Gα_s-Gγ dissociation. Values are mean ± SEM from independent experiments (n ≥ 3), each performed in triplicate. Exact n is indicated on the panel. NA, not applicable; ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical significance was assessed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs. WT. Statistical analysis results are summarized in S4 Table. (D) Effects of mutations at key residues involved in the hydration-mediated signaling network on the span of LysoPS (18:0)-induced Gα_s-Gγ dissociation. Values are mean ± SEM from independent experiments (n ≥ 3), each performed in triplicate. Exact n is indicated on the panel. NA, not applicable; ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical significance was assessed using one-way ANOVA followed by Dunnett’s multiple comparisons test vs. WT. Statistical analysis results are summarized in S4 Table. The data used to generate graphs in Fig 2C and 2D are available in S1 Data.

https://doi.org/10.1371/journal.pbio.3003447.g002

The orthosteric pocket of GPR174, framed by ECL2 and the extracellular halves of TM1-TM7, exhibits an amphipathic organization comprising a polar headgroup-binding region and a distal hydrophobic groove. Five structured waters stabilize the zwitterionic headgroup of LysoPS, facilitating conformational priming for receptor activation (Fig 2B). The lipid tail of LysoPS is embedded in a deep hydrophobic valley formed by TM3, TM4, TM5, and TM6, where it is stabilized by aromatic stacking interactions with Y99^3.33, Y103^3.37, F152^4.50, Y246^6.51, and F250^6.55 (S3B and S3C Fig). This hydrophobic subpocket contributes to ligand anchoring and defines the spatial boundary of the polar hydration cluster.

On the intracellular side, a structured hydration-mediated signaling network, composed of W_S1-W_S8 and W_G1, forms a hydrogen-bond “necklace” that bridges the sodium ion-binding pocket, the NPxxY motif, the DRY motif, and the α5 helix of Gα_s (Fig 2B). This network can be subdivided into three clusters. The first cluster comprises W_S1 and W_S2 in the sodium-binding pocket, where they form hydrogen bonds with D65^2.50, S105^3.39, Q68^2.53, and N284^7.45. Interactions of N284^7.45 and D288^7.49 with W_S2 and W_S3 provide a link between the first and second clusters. The second cluster consists of W_S3 and W_S4 adjacent to the NPxxY motif, where N284^7.45 and Y292^7.53 act as coordinating residues. Y292^7.53 further engages W_S4 and W_S5, extending the network toward the third cluster. The third cluster, formed by W_S5-W_S8 proximal to the DRY motif, connects R116^3.50 with T205^5.58 and T208^5.61, propagating the hydration pathway from the NPxxY motif to the cytoplasmic end of TM5. Collectively, these three clusters establish a continuous water-mediated pathway through the receptor core (S3D–S3F Fig). Through these contacts, W_G1 contributes to both stabilizing the active-state conformation of GPR174 and anchoring the Gα_s protein (S3G Fig). This structured hydration network reinforces the active-state conformation and supports G protein engagement.

To assess the functional relevance of the structured hydration network resolved in the cryo-EM structure, we mutated polar residues that coordinate internal water molecules along the W_S1-W_S8 and W_G1 pathway (Fig 2C and 2D; S4 Table). Two key residues, D65^2.50 and R116^3.50, are centrally positioned within the hydrogen-bonded network. D65^2.50 coordinates W_S2 near the receptor core, whereas R116^3.50 interacts with W_S6 and W_G1 at the G protein interface. Substitution of D65^2.50 with asparagine and R116^3.50 with glutamine replaced charged side chains with uncharged polar groups and severely impaired LysoPS-induced receptor activity. We next tested D288^7.49 and Y292^7.53, two conserved residues within the NPxxY motif. In the cryo-EM structure, D288^7.49 coordinates W_S2 and W_S3, and Y292^7.53 interacts with W_S4 and W_S5. Substitution of either residue reduced LysoPS-induced receptor activity. We also evaluated additional water-coordinating residues by introducing hydrophobic substitutions that preserve side chain volume while disrupting hydrogen bonding. These included Q68^2.53L, S105^3.39A, T205^5.58V, T208^5.61V, and N284^7.45L. Each mutation resulted in reduced receptor activity, consistent with a role in stabilizing internal waters and maintaining the geometry of the hydrogen-bond network.

To investigate the dynamic stability of the signaling-associated hydration network, we further performed molecular dynamics simulations specifically focused on nine internal waters (W_S1–W_S8 and W_G1) in GPR174. These water molecules exhibit prolonged residence times (0.19–1.34 ns), far exceeding the ~10 ps exchange rate of bulk solvent and approaching that of deeply buried structural waters reported in other class A GPCRs (S4A Fig). Upon removal and rehydration, these waters spontaneously returned to their original positions within 50 ns and reformed the hydrogen-bonding network with lifetimes comparable to the native configuration (S4B Fig). RMSD analysis revealed that coordinating residues, particularly those within the DRY and NPxxY motifs, remained conformationally stable during 1000-ns simulations (RMSD < 3 Å), reinforcing the structural and functional significance of the hydration network (S4C–S4E Fig; S5 and S6 Tables). Our structural comparison of GPR174 in complexes with both G_s and G_i proteins reveals a high similarity in the positions of key residues along the hydration network (average RMSD < 0.6 Å). In the GPR174-G_i complex, the G_i simulations exhibited shorter water-residue hydrogen-bond lifetimes relative to G_s (S5A Fig and S7 Table). The overall complex remained stable across three independent 1 µs trajectories (S5B Fig). These simulations further showed rapid reoccupation of the internal hydration sites (W_S1–W_S8 and W_G1) within ~50 ns (S5C Fig). Consistent with the reduced hydrogen-bond lifetimes, coordinating residues displayed increased RMSD values in the G_i simulations relative to G_s (S5D Fig), which likely reduces the visibility of internal waters at 3.4 Å resolution in the current reconstruction. Consistent with this, the same set of hydration-network mutations also reduced LysoPS-induced G_i signaling (S5E and S5F Fig; S8 and S9 Tables).

Together, these results support a model in which the W_S1–W_S8 and W_G1 network in GPR174 forms a continuous hydration network linking conserved activation motifs to the G protein interface, reinforcing the active-state conformation and enabling intracellular signal propagation.

Hydration cavities and allosteric water networks across class A GPCRs

While GPR174 illustrates how ordered internal water molecules can form a functional hydration network, the presence and conservation of such networks across class A GPCRs are not well established. We therefore sought to define the structural basis for hydration compatibility across receptors by analyzing water-accessible cavities that may facilitate internal water retention. To investigate the conservation of internal hydration features across class A GPCRs, we first defined three principal water-enriched cavities in GPR174 based on the cryo-EM structure and parKVFinder analysis (Fig 3A) [30]. These include (i) the CWC, located proximal to the sodium-binding pocket and shaped primarily by residues 2.50, 3.39, and 7.45; (ii) the JWC, situated adjacent to Y^7.53 and linking conserved motifs; and (iii) the EWC, defined by surrounding residues 3.50, 5.58, and 5.61.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Conserved and divergent hydration cavities across class A GPCRs.

(A) Sequence conservation analysis of residues forming three water-associated cavities: conserved water cavity (CWC), junctional water cavity (JWC), and extended water cavity (EWC). Logos indicate residue frequency at positions 2.50, 2.53, 3.39, 3.50, 5.58, 5.61, 7.45, and 7.53 across class A GPCRs. Green pie charts represent the property conservation frequency of corresponding residues in GPR174. Insets show examples of polar interactions within these cavities observed in GPR174 (green) and P2Y₁R (orange, PDB ID: 7XXH), with cavity regions identified by parKVFinder and visualized in PyMOL. (B) Distribution and functional impact of residue properties at position 5.58. Across class A GPCRs, this position is occupied mainly by either short side-chain polar uncharged residues (e.g., Ser, Thr, Cys) or by bulky hydrophobic/aromatic residues (e.g., Tyr, Phe). To assess structural consequences, EWC volumes were quantified from available active-state structures, including GPR174 (this study), P2Y₁₀R (PDB ID: 8KGG), GPR55 (PDB ID: 9GE2), PF₂αR (PDB ID: 8IUK), GPR20 (PDB ID: 8HS3), H₄R (PDB ID: 7YFC), GAL₂R (PDB ID: 7WQ4), β₂AR (PDB ID: 3SN6), P2Y₁R (PDB ID: 7XXH), GPR52 (PDB ID: 6LI3), FFA₂R (PDB ID: 8J24), and GPR119 (PDB ID: 7WCM). Receptors with short side-chain polar uncharged residues generally maintain enlarged EWCs, whereas those with bulky hydrophobic/aromatic residues exhibit restricted cavities. (C) Structural comparison of GPR55 (C^5.58, PDB ID: 9GE2) and P2Y₁R (Y^5.58, PDB ID: 7XXH). In GPR55, ordered water molecules are observed in the EWC based on cryo-EM density. In P2Y₁R, water occupancy detected by MD simulations is limited to the CWC and JWC.

https://doi.org/10.1371/journal.pbio.3003447.g003

To assess the conservation of hydration-associated cavities, we performed sequence alignments of residues involved in shaping the CWC, JWC, and EWC across class A GPCRs using GPCRdb [31]. For each cavity, we calculated the frequency of residues with physicochemically similar side chain properties to those observed in GPR174 (Fig 3A). Among the residues shaping the CWC, positions 2.50, 3.39, and 7.45 are highly conserved. Over 80% of class A GPCRs contain residues with physicochemically similar side chains at these positions. In contrast, position 2.53 exhibits low conservation, suggesting that it may play a modulatory role in determining the volume of the CWC. We next examined the sequence profile of the JWC, which is located adjacent to Y^7.53. Tyrosine is conserved at position 7.53 in 91% of class A GPCRs, which suggests that the position and geometry of the JWC may be structurally preserved across these receptors. We then analyzed the EWC, which is defined by residues 3.50, 5.58, and 5.61, located close to the cytoplasmic end of TM5 and TM6. Among these residues, R^3.50 is highly conserved, present in over 90% of class A GPCRs. In contrast, positions 5.58 and 5.61 show marked variability. Position 5.58, which represents the core determinant of the EWC and is the primary driver of cavity formation, falls into two distinct categories: small, uncharged polar residues (e.g., Asn, Thr, Cys) are observed in roughly 15% of receptors, whereas bulky aromatic residues (e.g., Tyr, Phe) occupy this position in approximately 75% of receptors, imposing steric constraints that limit hydration (Fig 3B). By contrast, 5.61 lies at the periphery of the cavity and plays a minor role in shaping the EWC. This distribution suggests that the key regulatory position 5.58 governs the size and hydration capacity of the EWC. Receptors with small, uncharged polar residues at this site allow water molecules to occupy the cavity and substitute for side-chain packing, whereas those with bulky aromatic residues provide steric support during TM6 outward movement.

To evaluate whether side-chain composition at cavity-facing residues correlates with cavity size, we performed cavity volume measurements across a panel of class A GPCRs (S6 and S7 Figs; S10 Table). CWC and JWC volumes remained relatively stable across receptors, consistent with the peripheral position of 2.53 relative to the sodium-binding pocket and the high conservation of Y^7.53 (S8 Fig). In contrast, the EWC volume exhibits a marked dependence on the residue at position 5.58 (Fig 3B). Receptors with small polar residues at 5.58 (e.g., GPR174, P2Y₁₀R, GPR55) consistently exhibit significantly larger EWC volumes compared to those with bulky hydrophobic residues (e.g., GAL₂R, β₂AR, P2Y₁R) [25,32–41], consistent with our sequence-based analysis. To further validate the key regulatory role of residue composition at position 5.58 in modulating EWC volume, we focused on two representative receptors, GPR55 and P2Y₁R, which feature distinct side-chain properties at this position (Fig 3C). In GPR55, where residue 5.58 is cysteine, two water molecules are resolved in the EWC of the active-state GPR55-G₁₃ structure determined by cryo-EM [32]. In P2Y₁R, where residue 5.58 is tyrosine, we analyzed the active-state cryo-EM structure of the P2Y₁R-G₁₁ complex and performed molecular dynamics simulations, which did not reveal the presence of water molecules in the EWC [38]. These observations indicate that the side chain properties at 5.58 play a central role in modulating EWC architecture and hydration.

Collectively, these results define a modular hydration network in class A GPCRs, where conserved cavities (CWC, JWC) stabilize core activation motifs, whereas the EWC provides receptor-specific flexibility. Position 5.58 plays a key role, where small, uncharged polar residues permit water molecules to occupy the cavity and act as a scaffold, while bulky residues substitute for this role by providing steric support. Thus, TM6 outward movement enlarges the TM5-TM6 cavity, which is stabilized either by water-mediated scaffolding or by bulky residues at 5.58, thereby maintaining the active conformation and promoting signaling. To further provide functional support for this conserved hydration-linked transmission scaffold, we quantified cAMP signaling output in P2Y₁R mutants at structurally equivalent positions, which showed that mutations at these positions reduced signaling efficiency relative to WT (S9 Fig; S11 and S12 Tables).

Structural and hydration determinants of G protein selectivity in GPR174

More than half of GPCRs are capable of engaging either G_s or G_i proteins, and a subset exhibits bifunctional coupling to both subtypes, as observed in GPR4, GPR120, CCK₁R, 5-HT₄R, and GCGR [3,42–45]. GPR174 couples to G_s, G_i, and the G₁₃ family of G proteins, with preferential activation of G_s over G_i signaling (Fig 1A) [43].

We determined cryo-EM structures of GPR174 in complex with both G_s and G_i, which enabled a comparative analysis of structural features underlying G protein selectivity. The receptor adopts a highly similar overall conformation in both states (Cα RMSD 0.77 Å; S10A and S10B Fig), and the orthosteric binding pocket remains nearly identical (all-atom RMSD < 0.5 Å; S10C Fig). The G_i complex was assembled in the presence of LysoPS, but the pocket density was insufficient for reliable ligand modeling, and the final G_i coordinates were deposited without a modeled ligand. Alanine substitutions of ligand-interacting residues impaired both G_s- and G_i-mediated signaling under our assay conditions (S10D–S10G Fig and S13–S16 Tables), indicating that these residues are involved in receptor activation and are unlikely to be the primary determinants of G protein subtype selectivity.

In contrast, the G_s complex features an enlarged G protein-binding interface, increasing from 890 Ų in the G_i-coupled complex to 1,560 Ų in the G_s-coupled complex, and involves significantly more contact residues (Figs 4A and S10H). To assess the functional relevance of these contact residues, we performed alanine substitutions at multiple interface positions. Most of these mutations impaired G_s or G_i signaling, indicating that both interfaces rely on extensive polar and nonpolar contacts for productive G protein engagement (S10I–S10L Fig and S13–S16 Tables). These structural differences are associated with distinct configurations of the G protein α5 helix. In the G_s complex, the α5 helix of Gα_s inserts deeply into a hydrophobic pocket formed by TM2 and TM3, whereas in the G_i complex, the α5 helix remains more surface-exposed and does not occupy this cavity (Fig 4B). Within this context, three features contribute to the selectivity of G_s engagement. First, in the G_s complex, the α5 helix of Gα_s inserts into a hydrophobic cavity formed by TM2 and TM3 (V55^2.40, M58^2.43, Y293^7.54, F299^8.50), where it is stabilized by extensive van der Waals interactions (Fig 4C). The α5 helix of Gα_i adopts a more surface-exposed configuration and does not engage this hydrophobic pocket (Fig 4D). Mutations that introduced bulky side chains into this cavity (V55^2.40F/M58^2.43F) abolished G_s signaling and produced a much weaker effect on G_i coupling, which remained measurable in our assay (Fig 4E and 4F; S13–S16 Tables), indicating that this site is critical for G_s engagement. Second, a structured water molecule (W_G1), resolved only in the G_s-bound complex, forms a hydrogen-bond bridge between R116^3.50 (of the conserved DRY motif) and the backbone carbonyls of Y391^H5.23 and D392^H5.24 on the α5 helix (Fig 4C and 4D). Substitution of R116^3.50 with glutamine or alanine attenuated G_s coupling and exerted a more modest impact on G_i coupling (Fig 4G and 4H; S13–S16 Tables). Third, a salt bridge between R115^3.49 and D134^4.42, which is not present in the G_i-bound complex, stabilizes the active-state conformation in the G_s structure (Fig 4I). Disruption of this interaction through R115^3.49A or R115^3.49Q mutations substantially reduced G_s signaling while enhancing G_i activity (Fig 4J and 4K; S13–S16 Tables), suggesting that this interaction plays a dual role by promoting G_s selectivity and simultaneously constraining G_i activation.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Hydration-mediated structural mechanisms underlying G protein selectivity in GPR174.

(A) Residues involved in selective interactions with Gα_s (pale green circles) or Gα_i (teal circles) at the receptor-G protein interface. Residues shared between Gα_s and Gα_i interfaces are shown in gray. (B) Side view of superposed GPR174-G_s (pale green) and GPR174-G_i (teal) complexes. (C) and (D) Conformational details of the α5-helix upon GPR174 coupling with G_s (C, light steel blue) or G_i (D, light salmon). (E) and (F) LysoPS (18:0) concentration-response curves for GPR174 activation via G_s (E) and via G_i (F), shown for WT as well as the hydrophobic-cavity mutants V55^2.40F and M58^2.43F. (G) and (H) LysoPS (18:0) concentration-response curves for GPR174 activation via G_s (G) and via G_i (H), shown for WT as well as the R116^3.50 mutants R116^3.50A and R116^3.50Q. (I) Structural comparison of the DRY motif region in GPR174-Gα_s (pale green) or GPR174-Gα_i (teal) complexes. (J) and (K) LysoPS (18:0) concentration–response curves for GPR174 activation via G_s (J) and via G_i (K), shown for WT as well as the DRY-motif mutants R115^3.49A and R115^3.49Q. For all functional curve panels, data are presented as mean ± SEM from at least three independent experiments, each performed in triplicate, and fitted potency and response parameters are reported in S13–S16 Tables. The data used to generate graphs in Fig 4E–4H, 4J, and 4K are available in S1 Data.

https://doi.org/10.1371/journal.pbio.3003447.g004

These findings indicate that GPR174 achieves G protein selectivity primarily through conformational and hydration-mediated mechanisms, rather than ligand recognition. The combined contributions of the hydrophobic cavity, the water-mediated bridge involving W_G1 at the DRY motif, and the R115^3.49 – D134^4.42 salt bridge promote G_s-specific engagement while permitting conditional coupling to G_i.

Constructs

For structural determination, wild-type (WT) human GPR174 (UniProt ID: Q9BXC1) was cloned into a modified pFastBac1 vector. An N-terminal hemagglutinin (HA) signal peptide and FLAG tag (DYKDDDDA) were introduced, followed by a thermostabilized BRIL to enhance receptor expression. The C terminus was fused to a LgBiT tag, a TEV protease cleavage site, and tandem maltose-binding protein (MBP) tags. Human dominant-negative (DN) Gα_s and DN Gα_i1 were generated by site-directed mutagenesis as previously described to decrease the nucleotide-binding affinity and increase G protein stability [48]. All three G_s subunits, including human DNGα_s, WT Gβ₁ fusion to HiBiT, and Gγ₂, were cloned separately into the pFastBac vector. Similarly, the three G_i subunits, including human DNGα_i, WT Gβ₁ fusion to HiBiT, and Gγ₂, were cloned into the pFastBac vector separately [27].

Expression and purification of Nb35 and scFv16

Nanobody-35 (Nb35) with a C-terminal 6× His tag was expressed in the periplasm of Escherichia coli BL21, as described previously [37]. scFv16 containing a C-terminal 6× His tag was constructed and expressed using the Bac-to-Bac system, as described previously [49]. Briefly, the overexpressed Nb35 and scFv16 protein were purified by Ni-NTA resin and eluted by using 500 mM imidazole, respectively. The eluted proteins were then cleaved by TEV protease and further purified by gel filtration chromatography using a Superdex 200 Increase 10/300 GL column. The purified Nb35 and scFv16 fractions were concentrated to approximately 5 mg/mL and stored at −80 °C until use.

Expression and purification of LysoPS-GPR174-G protein complex

Spodoptera frugiperda (Sf9) cells were cultured in ESF921 medium at 27 °C with shaking at 120 rpm. HEK293T cells were maintained in a humidified incubator at 37 °C with 5% CO₂ in DMEM (VWR) supplemented with 10% fetal bovine serum (FBS, VWR), 100 I.U./mL penicillin, and 100 μg/mL streptomycin (Invitrogen). GPR174 (1–333)-LgBiT, DNGα_s or DNGα_i1, Gβ₁-HiBiT fusion, and Gγ₂ were co-expressed in Sf9 insect cells using the Bac-to-Bac baculovirus expression system (Thermo Fisher). The cells were cultured and grown to a density of 2.4 × 10⁶ cells mL⁻¹and infected with four separate baculoviruses (GPR174 (1-333)-LgBiT fusion, DNGα_s or DNGα_i1, Gβ₁-HiBiT fusion, and Gγ₂) at a ratio of 1:1:1:1. After culturing for 48 h at 27 °C, the cells were harvested by centrifugation and the cell pellets were stored at −80 °C. To purify the GPR174–G_s and GPR174–G_i complexes, cell pellets were resuspended in buffer containing 20 mM HEPES (pH 7.5), 150 mM NaCl, and 10 mM MgCl₂, supplemented with protease inhibitor cocktail tablets (Roche). The suspension was homogenized 60 times using a Dounce homogenizer to disrupt cell membranes. Complex formation was initiated by adding 10 μM lysophosphatidylserine (LysoPS) and 25 mU/mL apyrase (Sigma), followed by solubilization with 0.5% (w/v) lauryl maltose neopentyl glycol (LMNG, Anatrace) and 0.1% (w/v) cholesterol hemisuccinate (CHS, Anatrace) for 2 h at 4 °C. The lysate was clarified by centrifugation at 17,000 rpm for 30 min, and the supernatant was incubated with amylose resin for 1 h. After washing with buffer containing 20 mM HEPES (pH 7.5), 150 mM NaCl, 10 mM MgCl₂, 10 μM LysoPS, 0.01% LMNG, and 0.002% CHS, the complexes were eluted with the same buffer supplemented with 10 mM maltose. TEV protease was added to cleave the MBP tags. The eluted complex was then concentrated and further purified by size-exclusion chromatography on a Superose 6 Increase 10/300 GL column (GE Healthcare) pre-equilibrated with buffer containing 20 mM HEPES (pH 7.5), 150 mM NaCl, 10 mM MgCl₂, 0.00225% LMNG, 0.00075% GDN, 0.0006% CHS, and 10 μM LysoPS. Peak fractions of the GPR174–G_s or GPR174–G_i complexes were collected and concentrated to 10 mg/mL using a 100-kDa cutoff Amicon Ultra Centrifugal Filter (Millipore) for cryo-EM analysis.

Cryo-EM grid preparation and data collection

Holey carbon grids (Quantifoil R1.2/1.3, 300 mesh) were glow-discharged at 25 mA for 1 min using a PELCO easiGlow system. Purified GPR174–G_s or GPR174–G_i complexes were applied to the glow-discharged grids. Excess sample was blotted for 3.5 s with a blot force of 8 using Ted Pella filter paper at 4 °C and 100% humidity, and the grids were then vitrified in liquid ethane using a Vitrobot Mark IV (Thermo Fisher). Cryo-EM data were collected on a Titan Krios microscope operated at 300 kV, located at the Core Facilities of Zhejiang University Medical Center and Liangzhu Laboratory. For the GPR174–G_s complex, data were acquired using EPU software in super-resolution mode with a calibrated pixel size of 0.74 Å and a defocus range of −0.6 to −1.2 μm. Each movie consisted of 30 frames with a total exposure dose of 80 e⁻/Ų. For the GPR174–G_i complex, micrographs were acquired under similar conditions with a pixel size of 0.93 Å, a defocus range of −1.0 to −2.0 μm, and 40 frames per movie with a total dose of 52 e⁻/Ų. In total, 8,814 and 8,490 movies were collected for the GPR174–G_s and GPR174–G_i complexes, respectively.

Cryo-EM data processing

Movies were aligned using RELION 4.0 [50], and contrast transfer function (CTF) parameters were estimated with Gctf v1.18 [51]. Subsequent data processing was performed using RELION 4.0 and CryoSPARC v3.3.2 [52]. For the GPR174-G_s complex, micrographs were initially processed using RELION 4.0 and CryoSPARC v4.0.3. Template-based particle picking in RELION yielded 4,203,634 particle projections. These were imported into CryoSPARC and subjected to multiple rounds of 2D classification. Selected high-quality particles were used to generate ab initio models, followed by several rounds of heterogeneous refinement in CryoSPARC. A subset of particles was re-extracted and further classified by two additional rounds of 3D classification in RELION to remove poorly defined particles. The resulting 170,885 particles were refined using 3D refinement, CTF refinement, and Bayesian polishing in RELION, yielding a final map with a reported global resolution of 2.0 Å. For the GPR174-G_i complex, 3,495,640 particles were initially selected. After similar processing steps, particles were re-extracted into RELION for 3D classification. A final subset of 553,019 particles was used for 3D refinement, CTF refinement, and Bayesian polishing, resulting in a cryo-EM map with a global resolution of 3.4 Å. These reconstructions were used for subsequent model building and analysis.

Model building and structure refinement

Initial models were generated based on the previously determined structure of LysoPS-bound GPR174 (PDB ID: 7XV3) and AlphaFold3 predictions [53]. The models were manually fitted into the density maps using UCSF Chimera, followed by flexible fitting with Rosetta. Manual model rebuilding was performed in Coot, and real-space refinement was carried out in Phenix. Final refinement statistics were assessed using the Comprehensive Validation (cryo-EM) module in Phenix [54–56]. Model statistics are summarized in S3 Table. Structural figures were prepared using UCSF ChimeraX and PyMOL [57,58].

Molecular dynamics simulation

Analysis.

Hydration-associated waters W_S1–W_S8 and W_G1 were defined from the cryo-EM structure and monitored throughout the trajectories. Hydrogen-bond lifetimes between these waters and coordinating residues were calculated using a custom Python script. The average lifetime was calculated according to the following equation:

where t_{H-bond, i} represents the lifetime of each hydrogen bond formation, N_H-bond means the total number of the bond formations, and t_avg means the average lifetime. The hydration-mediated transmission network for the GPR174–G_s–rWat system was reformed within ~50 ns and remained stable throughout the simulations. Overall system stability was then assessed by calculating the RMSD of heavy atoms, with transmembrane helix Cα atoms. GPR174-G_i and P2Y₁R-G_s simulations were analyzed using the same workflow, and internal water occupancy was quantified. All structural analyses and visualizations were performed using UCSF ChimeraX and PyMOL [57,58].