Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 May 23.
Published in final edited form as: Nat Struct Mol Biol. 2022 Mar 3;29(3):274–281. doi: 10.1038/s41594-022-00728-4

Oxytocin signaling complex reveals a molecular switch for cation dependence

Justin G Meyerowitz 1,2,3,*, Michael J Robertson 1,3,*, Ximena Barros-Álvarez 1,3, Ouliana Panova 1,3, Robert M Nwokonko 1,3, Yang Gao 1,3, Georgios Skiniotis 1,3,4
PMCID: PMC12101619  NIHMSID: NIHMS2077289  PMID: 35241813

Summary

Oxytocin (OT) and vasopressin (AVP) are conserved peptide signaling hormones that are critical for diverse processes including osmotic homeostasis, reproduction, lactation, and social interaction. OT acts through the oxytocin receptor (OTR), a magnesium-dependent G protein-coupled receptor that is a therapeutic target for treatment of postpartum hemorrhage, dysfunctional labor, and autism. However, the molecular mechanisms that underly OTR activation by OT and dependence on magnesium remain unknown. Here we present the wild-type active state structure of human OTR bound to OT and miniGq/i determined by cryo-EM. The structure reveals a unique activation mechanism adopted by OTR involving both the formation of a Mg2+ coordination complex between OT and the receptor, as well as the disruption of transmembrane helix 7 (TM7) by OT. Our functional assays demonstrate the role of TM7 disruption and provide the mechanism of full agonism by OT and partial agonism by oxytocin analogs. Furthermore, we find that the identity of a single cation-coordinating residue across vasopressin family receptors determines whether the receptor is cation-dependent. Collectively, these results demonstrate how the Mg2+-dependent oxytocin receptor is activated by oxytocin, provide essential information for structure-based drug discovery efforts, and shed light on the molecular determinants of cation dependence of vasopressin family receptors throughout animals.

Introduction

Oxytocin (OT) and vasopressin (AVP) are nine-residue, cyclized signaling peptides that act as both hormones and neurotransmitters. These hormones are present throughout the animal kingdom, where they regulate diverse processes such as osmotic homeostasis, reproduction, social bonding, lactation, and labor. OT acts through the oxytocin receptor (OTR), a G protein-coupled receptor (GPCR) that primarily couples to Gq/11-family of G proteins, leading to numerous physiologic processes depending on the biological context. For example, in humans, activation of OTR in the myometrium causes uterine contractions during labor, while in the central nervous system leads to social recognition and the initiation of maternal nurturing behavior1-3, and in myoepithelial cells of the mammary gland triggers milk ejection during suckling4. Additionally, OTR signaling plays an important role in the modulation of pain5, reduction in addiction-related behaviors6, and fundamental social behaviors such as trust and risk-acceptance7.

OTR is the target of clinically approved drugs for labor induction and augmentation (oxytocin), first-line treatment of postpartum hemorrhage (oxytocin and carbetocin), and is a target of great interest for the treatment of spontaneous preterm labor, infertility, migraines, and autism spectrum disorder (ASD)8,9. Studies in both animal models and patients have demonstrated that intranasal administration of OT briefly relieves core symptoms in ASD, schizophrenia, and anxiety disorders10,11. However, as a peptide, OT has unfavorable pharmacokinetic properties, and there exist no small-molecule full agonists of OTR. Discovery of small non-peptide agonists has been hampered by the large binding site and lack of structural information for the receptor. Furthermore, while a crystal structure of OTR was recently determined12, this structure represents the receptor in an inactive and mutationally thermostabilized state bound to an antagonist and lacks identification of the critical bound magnesium.

Here we present the cryo-electron microscopy (cryo-EM) structure of wild-type human OTR in complex with OT and G protein. We characterize the mechanism of full agonism by OT and partial agonism by oxytocin analogs, determine the architecture of magnesium (Mg2+) coordination required for receptor activation, and show that the identity of a single conserved residue determines whether activation of a vasopressin-family receptor requires Mg2+ as a cofactor.

Results

The oxytocin receptor signaling complex

OTR has been reported to signal through multiple different G protein partners, including both Gq/11- and Gi/o-family G proteins13. To pursue structural studies of the active OTR-G protein signaling complex, we first sought to identify an optimal G protein partner for stable complex formation. To this end, we used the TRUPATH platform of optimized G protein-specific BRET-based biosensors to determine the extent of OTR-dependent G protein activation in response to OT, as compared to a canonical “full activator” of any given G protein (Fig. 1a, Supplementary Fig. 1a, Supplementary Table 1). Our results show that OTR is a strong activator of Gq, G11, and G15 and a weak activator of Gi/o-family proteins. Thus, we decided to pursue cryo-EM studies of OTR in complex with Gq/11. We expressed full-length human wild-type OTR in human embryonic kidney cells (HEK293) and purified the receptor in the presence of oxytocin and magnesium (Supplementary Fig. 1b, c). For complex formation we employed an engineered miniGq/11-family heterotrimer (miniGq/i) that enables binding of the single-chain variable fragment scFv16, which has been shown to aid in GPCR-G-protein complex stabilization for structural studies14,15 (Supplementary Fig. 1d, e). A three-dimensional cryo-EM map of the OTR-Gq complex was obtained at a nominal global resolution of 2.9 Å, with local refinements of the G protein and transmembrane domain (TMs) yielding improved density maps with resolutions of 2.9 Å and 3.3 Å, respectively (Supplementary Fig. 2a-e, Supplementary Table 2), allowing for the unambiguous modeling of the complex (Fig. 1b, Supplementary Fig. 3a-f).

Figure 1. Structure of the oxytocin receptor-miniGq complex bound to oxytocin.

Figure 1.

Open in a new tab

a, Survey of Emax of OT-dependent OTR activation of 12 different G protein BRET-based biosensors using benchmark receptors NTSR1 and ADRB2 (error bars are S.E.M.; nd: no coupling detected), b, Overall map and model for the oxytocin-oxytocin receptor-miniGqi complex, c, Cutaways of identical receptor-aligned views of binding pockets for antagonist retosiban (PDB: 6TPK) and agonist oxytocin bound to OTR with transmembrane (TM) helices and N-terminal domain (NTD) labeled. Dashed lines serve as reference for alignment of receptors, d, Extracellular ligand-capping structure formed by transmembrane helix 1 (TM1), the N-terminal domain (NTD), and extracellular loops 1 and 2 (ECL1, ECL2). e, Dose-response curves of OT-dependent activation of OTRWT (WT; orange boxes), OTRR34 (R34A; purple circles), OTRF103A (F103A; blue diamonds), and OTRR104A (R104A; teal circles) using a G11 biosensor with matched surface expression levels. Error bars are S.E.M.; ranges in inset are asymmetric 95% CI from simultaneous curve fitting of 3 biological replicates. f, Overlay of receptor model (blue ribbons) and putative cholesterol- and lipid-like densities from cryoEM map (gold) with labeled cholesterol network clusters (CNCs).

On the extracellular side of the receptor, while extracellular loop 2 (ECL2) has the beta-finger motif and disulfide bond of other Family A receptors, a distinct ligand-capping complex is formed by interactions between the N-terminal domain (NTD), ECL1 and ECL2. This extracellular complex, absent in the inactive state (Fig. 1c), is formed by rotameric changes of ECL1 and NTD to form a cation-π-cation interaction between R34NTD, F103ECL1, and R104ECL1, with a concomitant opening of ECL2 (Fig 1d). The significance of these interactions in stabilizing the active binding site architecture is supported by prior observations that removal of R34 by truncation or substitution results in a 2,000-fold decrease in OT binding affinity16. Additionally, we observe impaired OT-dependent activation of G11 by OTRR34A, OTRF103A (18 nM, 14 – 24 nM 95% CI), and OTRR104A (88 nM, 50 – 150 nM 95% CI) compared to wild-type (3.7 nM, 2.7 – 5.3 nM 95% CI; Fig. 1e, Supplementary Fig. 3g, Supplementary Table 4). The modest deficiency in activation by OTRF103A compared to OTRWT may be due to the ability of hydration to partially compensate for the absence of the phenyl. As a result of these interactions, the functional binding pocket of OTR is extended towards the extracellular space as compared to the inactive state (Fig. 1c, d).

Cholesterol has been shown to be critical for OTR function and plays a role more generally in GPCR allosteric modulation17-19. We observe several potential cholesterol- and lipid-like densities in positions that correspond to a known cholesterol network cluster (CNC) found in cryoEM and X-ray structures of human GPCRs20. These include an unambiguous density for cholesteryl hemisuccinate (CHS) in an extrahelical pocket formed by TM4 and TM5 at the outer leaflet (CNC45o, numbering from Taghon et al.) (Fig. 1f), a position where a cholesterol was also observed in the OTR inactive state crystal structure12. Additionally, we observe potential cholesterol- or lipid-like densities corresponding to CNC345i and CNC234i at the inner leaflet and CNC6o at the outer leaflet.

The C-terminal amidated, disulfide-cyclized nonapeptide OT adopts a near-planar conformation in the OTR binding pocket that is oriented perpendicular to the plane of the membrane, in which Tyr2 of OT sits deepest within the receptor core alongside TM7 and TM2, and Ile3 occupies a hydrophobic pocket formed by Y2005.38, I2015.39 and I2045.42 (Fig. 2a, Supplementary Fig. 4a) [Ballesteros-Weinstein numbering denoted in superscript21]. Additionally, Gln4 makes a polar interaction with Q2956.55, consistent with the prior finding that mutagenesis of the corresponding residue in vasopressin 1a receptor (V1aR) (Q311A6.55) decreases the binding affinity of both AVP and OT22. Overall, this binding pose is similar to that of vasopressin 2 receptor (V2R)-bound AVP, with both Phe3 of AVP and Ile3 of OT occupying a hydrophobic pocket and W188 of ECL2 capping the ligand (Supplementary Fig. 4b).

Figure 2. Oxytocin binding mode and G protein interaction.

Figure 2.

Open in a new tab

a, View from TM6 of oxytocin (OT) in the oxytocin receptor binding pocket with overlay of cryoEM map at threshold level 0.105 in ChimeraX; receptor in blue, oxytocin in orange, and Mg2+ in green. Superscript is Ballesteros-Weinstein notation. b, Key interactions between the oxytocin receptor and the alpha-5 helix of miniGqi (Gq α5).

The oxytocin receptor binds miniGq/i in a manner similar to that of the 5-HT serotonin receptor HTR2A23, although the C-terminal alpha-helix of Gαq sits shallower in the core of the OTR transmembrane bundle (Supplementary Fig. 4c, d). Similar to prior observed receptor-Gq complexes, key interactions between the receptor and the α5 helix of miniGq/i are polar contacts with TM3 and hydrophobic interactions between I2235.61, I2275.65, I1413.54 on the receptor and L236, L240, L245, and V246 of helix α5 (Fig. 2b, Supplementary Fig. 4e). Notably, the angular orientation and interactions are more similar between OTR-miniGq/i and V2R-Gs24 than they are between OTR-miniGq/i and HTR2A-miniGq/i (Supplementary Fig. 4c-e).

Disruption of TM7 helix underlies oxytocin receptor activation

One particularly notable feature of OT binding is a hydrogen bond formed by Tyr2 across TM7 to the backbone of L3167.40 and disruption of the alpha helical nature of a stretch of TM7 (L3167.40 to S3197.43) as defined by loss of +4 backbone H-bonding and helical phi-psi angles (Fig. 3a, Supplementary Fig. 5a, e, f). While this interaction is also observed in the AVP-V2R-Gs complex24,25, Tyr2 is not oriented to make this contact in one of the multiple observed states of the active V2R-Gs complex (Supplementary Fig. 5b)26, suggesting that TM7 helical disruption by AVP may not be required for full activation of V2R. Additionally, in the inactive-state structure of OTR, the indanyl of retosiban is oriented towards the receptor core along the fully helical TM7—however, the retosiban-bound inactive state structure includes N325K7.49 as a thermostabilizing mutation which may lock TM7 in a continuous helical conformation through a polar interaction with N3217.45 (Supplementary Fig 4c, d), confounding interpretation of the disrupted helicity found only in the active state. Taken together, it is unclear from these structural data alone whether TM7 helical disruption is required for receptor activation.

Figure 3. Disruption of TM7 helix underlies receptor activation.

Figure 3.

Open in a new tab

a, Overlay of oxytocin-bound OTR (sky blue) vs antagonist-bound inactive state (mustard; PDB: 6TPK) models. TM7 alpha-helix disruption via H-bond formation between Tyr2 of OT and backbone of L316 of OTR (top right inset), causing rotation and displacement of bottom of TM7 (bottom inset), and outward movement of TM6 to accommodate the N-terminal helix of Gq (bottom left inset). b, Measurement of TM7 reformation as percent helicity of L3167.40 to S3197.43 in all-atom molecular dynamics simulations of 3 replicates each of active state with (blue) and without (mustard) OT. c, Chemical structures of oxytocin derivatives carbetocin, methyl-(O-Tyr)-oxytocin (meOT), and Phe2-OT compared to oxytocin (changes from OT in insets). d, Emax of meOT, Phe2-OT, and carbetocin as compared to oxytocin for effector-specific BRET-based biosensors. Error bars are S.E.M., each dot represents a biological replicate, each replicate corresponding to technical duplicate of 11-point dose response curves.

To explore the role of the hydrogen bond interaction between Tyr2 and the backbone carbonyl of L3167.40 in receptor activation (Fig. 3 a), we performed all-atom molecular dynamics (MD) simulations on both the oxytocin-bound receptor in the active state, and the identical active state in the absence of ligand. We found that in the apo state, the helical nature of TM7 rapidly reformed (Fig. 3b), while TM7 remained disrupted in the presence of OT. These results, in combination with the observation that the helicity of TM7 is intact in the antagonist-bound state, suggest that TM7 disruption may play a role in differentiating receptor agonism from antagonism.

To experimentally interrogate the role of the hydrogen bond between Tyr2 and the backbone carbonyl of L3167.40, we tested three analogs of OT for their ability to activate Gq/11-family signaling and recruit β-arrestin. Phe2-oxytocin (Phe2-OT) lacks the tyrosine hydroxyl and is thus unable to act as an H-bond donor to TM7, but can still be sterically accommodated in the active-state conformation (Fig. 3c). In contrast, 2-(O-methyltyrosine)-oxytocin (meOT) is unable to act as an H-bond donor, and the methylated tyrosine would sterically clash with TM7 in the same rotameric conformation of OTR-bound OT. We also tested carbetocin, which is a clinically approved partial agonist of OTR with a methylated tyrosine (Fig 2c). All three OT analogs retain high potency for receptor activation as measured by both Black-Leff analysis and the G15 biosensor, with KA ranging from 1.8 – 38 nM (Supplementary Table 3). However, Phe2-OT is a partial agonist for Gq, G11, G15, β-arrestin 1 (β1) and β-arrestin 2 (β2) (Fig. 3d, Supplementary Fig. 6a, b, Supplementary Table 3), demonstrating that the tyrosine hydroxyl is required for full agonism. Additionally, meOT and carbetocin behave as a neutral antagonist (meOT) or weak agonist (carbetocin) for Gq, weak agonists for G11, β1, and β2, and moderate agonists for G15 (Fig. 3d, Supplementary Fig. 6a-c, Supplementary Table 3). Additionally, the clinical compound atosiban, with an O-ethyltyrosine, is a neutral antagonist of OTR27.

In contrast, Phe2-AVP behaves as a full agonist for V1aR and vasopressin 1b receptor (V1bR) activation of Gq, and V2R activation of Gs (Supplementary Fig. 6d), suggesting that the importance of the Tyr2-TM7 hydrogen bond for vasopressin receptor activation is less clear, and consistent with the observation of an alternative Tyr2 orientation in the V2R-Gs complex. Taken together, these data demonstrate that hydrogen bond disruption of the helical nature of TM7 by OT is a key component of full OTR activation, in which the bottom half of TM7, containing N3257.49, P3267.50 and Y3297.53 of the NPxxY activation motif, is rotated and translated outward towards TM6, stabilizing a receptor conformation in which TM6 is displaced from the core to accommodate G protein engagement (Fig 3a, Supplementary Fig. 5d).

The OT-OTR cation coordination complex

Our cryo-EM map of OTR reveals a strong spheroid density in the ligand binding pocket between OT, the NTD, TM1, and TM2 (Fig. 2a, Supplementary Fig. 3d). While no such density is present in the OTR antagonist-bound crystal structure, it has been proposed that E421.35 and D1002.65 are responsible for cation coordination12, suggesting that the observed density corresponds to a cation. Here we observe a roughly octahedral coordination geometry with interatomic distances from this density of 2.28 Å to D1002.65, 4.03 Å to E421.35, 4.34 Å to the backbone carbonyl of Pro7 of OT, and 4.50 Å to R34NTD (Fig. 4a). Typical Mg2+ coordination distance in protein structures is ~2.1-2.3 Å, suggesting that D1002.65 directly coordinates the cation, whereas E421.35, Pro7, and R34NTD are water-mediated interactions28. To explore the role of water molecules in mediating coordination, we ran JAWS, a molecular simulation methodology, to calculate the occupancies and absolute binding free energies of water molecules around Mg2+ 29. In three independent replicates, we found that water molecules are highly favored in positions that mediate the coordination of cation by the above residues, in an octahedral geometry (Fig. 4b).

Figure 4. Architecture of the magnesium cofactor coordination site.

Figure 4.

Open in a new tab

a, Coordination site geometry of Mg2+ by E421.35, D1002.65, R34NTD, and backbone carbonyl of Pro7 of OT with interatomic distances displayed. b, Locations of highly favorable water sites averaged from three independent replicates of JAWS calculations. c, Magnesium but not calcium potentiates OT-dependent OTR-G11 signaling in BRET dose-response assay in the presence of 500 nM OT (error bars are S.E.M.). Dose-response curves of OT-dependent OTR activation of G11 in the presence or absence of Mg2+ with d, wild-type OTR, e, OTRD100A, or f, OTRD100K. g, Bar graph of EC50 values of OT-dependent OTR activation of G11 in the presence (black) or the absence (grey) of magnesium. Error bars are 95% CI, statistical comparison made by extra sum of squares F test. ****P<0.0001; n.s.: P>0.05.

Agonist but not antagonist binding to OTR requires Mg2+, and both OT binding affinity30 and OT-dependent OTR activation (Fig. 4c) is potentiated by Mg2+ but not Ca2+. To validate the role of these residues in coordinating Mg2+, we tested the ability of OT to activate OTR mutants in the presence or absence of Mg2+. OT demonstrated potent activation of OTR as measured by activation of Gq or G11, and recruitment of beta-arrestin 1 or beta-arrestin 2 in the presence of Mg2+, with poor efficacy and potency in the presence of 0.5 mM EDTA (Fig. 3d, Supplementary Fig. 7a, Supplementary Table 4). In contrast, activation of OTRD100A by OT was ~300 nM and similar in both the presence and absence of Mg2+ (Fig. 4e, g; Supplementary Table 4). OTRE42A was incapable of activating G protein in either the presence or absence of Mg2+ (Supplementary Table 4). This is consistent with recent observations that OTRD100A and OTRE42A exhibit weaker binding affinities to OT compared to OTRWT, and that those affinities are unaffected by the absence of Mg2+ 12. We noted that the interatomic distances were such that replacement of D1002.65 with a lysine (D1002.65K) should place the positively charged amine in the position of Mg2+ and hypothesized that lysine could functionally replace the cation. Indeed, we found that OTRD100K was activated equipotently by OT (~10 nM) in the presence or absence of Mg2+ (Fig. 4f, g, Supplementary Fig. 7b, Supplementary Table 4).

Empirical observations and quantum mechanical calculations have shown that cation coordination sites with multiple charged cation-interacting residues and rigid coordination geometries favor Ca2+, whereas sites with suboptimal geometries and polar, non-ionized (e.g., water) interacting groups favor Mg2+ 31. Thus, the water-mediated nature of this magnesium coordination site between OTR and OT, which is required for receptor activation, may explain the structural basis of OTR selectivity for Mg2+ over Ca2+.

Vasopressin receptor cation sensitivity is defined by a single amino acid

The vasopressin family of receptors in humans includes OTR, V1aR, and vasopressin 1b receptor (V1bR), which predominantly signal through Gq/11-family G proteins, and V2R, which predominantly signals through Gs32,33. The cation-coordinating residues R34NTD, E421.35, and D1002.65 in OTR are fully conserved across human oxytocin and vasopressin receptors except for V2R where the aspartate is replaced by a lysine (K1002.65; Fig. 6a). The presence of a positively charged side chain in this position raised the hypothesis that V2R is insensitive to Mg2+. Indeed, we found that while the potency of AVP-dependent activation of both V1aR and V1bR was sensitive to the presence of Mg2+, V2R activation by AVP is not affected by the absence of Mg2+ (Fig. 5b-d). Taken together with the observation that OTRD100K activation is Mg2+-independent, this result suggests that K1002.65 of V2R functionally replaces the cation in the same manner as the lysine in OTRD100K.

Figure 5. An aspartate to lysine transition reverses magnesium dependence in the oxytocin/vasopressin receptor family.

Figure 5.

Open in a new tab

a, Alignment of human oxytocin/vasopressin family receptors illustrating the identity of cation-coordinating residues. Reference number above residue corresponds to the residue number in OTR. b, Bar graph of EC50 for AVP-dependent V1aR activation of G11 signaling in the presence (black) and absence (grey) of Mg2+ for wild-type OTR, OTRD112K, and OTRD112A. c, Bar graph of EC50 for AVP-dependent activation of G11 (V1bR) or Gs (V2R) signaling in the presence (black) and absence (grey) of Mg2+ for wild-type V1bR, V1bRD95K, wild-type V2R, and V2RK100D. Error bars are 95% C.I., statistical comparison with extra sum of squares F test. ****P<0.0001, ***P<0.0005, n.s.: P>0.05. d, Dose-response curves in the presence and absence of Mg2+ for V1aR-G11, V1bR-G11, and V2R-GS. e, Dose-response curves in the presence and absence of Mg2+ for V1aRD112K-G11, V1bRD95K-G11, and V2RK100D-GS. Error bars represent 95% confidence intervals from simultaneous fitting of 3-5 biological replicates. f, Sequence alignment of cation-coordinating residues in vasotocin family receptors across vertebrates. *Receptor not present in species.

To verify the role of the lysine in functional replacement of Mg2+, we also found that AVP potency against V1aRD112A is ~150 nM and independent of Mg2+, and that AVP potency against V1aRD112K is ~10 nM and independent of Mg2+ (Fig. 5b, e, Supplementary Table 5). Like OTR, V1aRE54A was deficient in AVP-dependent G protein activation (Supplementary Table 5). Reciprocally, V2RK100D gains sensitivity to Mg2+ as a cofactor, with AVP-activated EC50 of 0.48 nM in the presence and 11 nM in the absence of Mg2+ (Fig. 5c, e, Supplementary Fig. 7b, Supplementary Table 5).

Vasopressin family receptors are present throughout the animal kingdom, and their evolution in vertebrates is well-characterized34. Given that the Mg2+-independent V2R and Mg2+-sensitive OTR/V1aR/V1bR all share a common protogene, we interrogated the identity of this conserved cation binding site across vasopressin family receptors in vertebrates using a recently proposed framework and nomenclature for vasopressin receptor evolution33. The cation-coordinating residues R34NTD, E421.35 and D1002.65 are widely conserved, with D1002.65 replaced by glutamate in certain homologs of some species (Fig. 5f, Supplementary Table 6). Notably, we found that the lysine corresponding to cation insensitivity is present in V2R (VTR2C in newly proposed nomenclature) in Cercopithecoidea (Old World monkeys, including humans) and in V1aR (VTR1A) in lampreys and sharks (Fig. 5f, Supplementary Fig. 7c). These observations reveal that this cation binding site has undergone evolutionary selection, suggesting the functional relevance of cation sensitivity or insensitivity of vasopressin family receptors. As positively selected structural sites are found in functionally important regions of proteins35, these data suggest that the cation sensitivity or insensitivity of vasopressin/oxytocin receptors has been guided by distinct environmental pressures throughout evolution.

Discussion

The cryo-EM structure of OT-bound OTR presented here, in combination with our pharmacologic data, reveal the molecular basis of OTR activation. Upon ligand binding, disruption of the helical character of TM7 by OT drives the transition from inactive to active arrangement of the TMs to accommodate the G protein. Hydrogen bonding between the peptide and the disrupted TM7 appears to underline, at least in part, the mechanism of full versus partial agonism of OTR. This Tyr2-TM7 H-bond interaction appears to be less critical for the vasopressin receptors (Supplementary Fig. 6d), thus further exploration of the activation mechanism of the vasopressin receptors is required.

Moreover, we find that a water-mediated cation coordination complex between the backbone of OT and several residues of OTR is responsible for the Mg2+ dependence of OTR activation. We observe that cation coordination is required not just for OT binding, as previously observed, but for OTR activation (Fig. 4d, Fig. 5d, Supplementary Fig. 7a, b). More critically, our data suggest that the efficacy of OT can be limited by physiologically relevant concentrations of Mg2+: For example, even at supraphysiologic plasma concentrations of OT (500 nM), a low but realistic serum level of Mg2+ such as 0.75 mM would correspond to only 70% of maximum activation (Fig. 4c). This phenomenon may underly certain clinical findings, such as the observation that Mg2+ is effective in attenuating menstrual-related migraine attacks, in which the oxytocin system is mechanistically implicated36. That Mg2+ may in some cases be pharmacodynamically limiting OT efficacy could have important implications for physiologic and pathophysiologic processes involving OTR.

Finally, we report that the architecture of the Mg2+ coordination site, which is similar across vasopressin receptors and across vertebrates, has, surprisingly, evolved in select clades to function in a cation-independent manner. The nature of the selective pressures that act on the Mg2+-dependence of this receptor family requires further exploration.

Methods

Cloning and mutagenesis

All mutagenesis cloning was performed using the Q5 Site-Directed Mutagenesis kit (New England Biolabs), and all fragment insertion cloning was performed using PrimeSTAR Max DNA Polymerase (Takara Biosciences) and either Gibson Cloning (New England Biolabs) or In-Fusion HD (Takara Biosciences). Human OTR cDNA was obtained from the cDNA Resource Center (www.cdna.org), and V1aR, V1bR, and V2R from the Presto-Tango Library (Addgene) which were modified for use in BRET assays with the addition of a stop codon at the end of the receptor sequence. All plasmids used for cloning, expression, and assays were sequence-verified (Elim Biopharma).

Expression and purification of OTR

Full-length human OXTR was cloned into pEZT-BM38 with an N-terminal hemagglutinin (HA) signal sequence, N-terminal sequence of the beta-2 adrenergic receptor, FLAG epitope tag, and TEV cleavage site, and C-terminal eGFP and 6xHis tag. The BacMam system (Thermo Fisher) was used to generate baculovirus for infection of HEK-293S mammalian cells grown in suspension at 37°C. Cells were harvested at 48 h post transfection, washed with cold PBS with 2.5 mM MgCl2, flash frozen, and stored at −80°C until further use. All purification steps were carried out at 4°C, and all buffers were supplemented with 20 μM oxytocin acetate (OT, MedChemExpress) unless otherwise noted. Cells were thawed on ice, resuspended by loose Dounce in a hypotonic buffer containing 20 mM HEPES pH 7.45, 10 mM NaCl, 5 mM MgCl2, protease inhibitors, and 1 μM OT, lysed for 2 h by stirring, and centrifuged at 100,000 g for 35 min. Membranes were collected, resuspended by loose Dounce in a buffer containing 20 mM HEPES pH 7.45, 500 mM NaCl, 5mM MgCl2, 20% glycerol, protease inhibitors, and 1 μM OT, and flash frozen and stored at −80°C until further use. Frozen membranes were thawed by stirring into the same buffer, into which detergents were added dropwise while stirring to a final concentration of 1% lauryl maltose neopentyl glycol (LMNG, Anatrace), 0.2% cholesteryl hemisuccinate (CHS, Anatrace), and 0.2% sodium cholate (Anatrace) and allowed to stir for 3 h. Solubilized membrane was clarified by centrifugation at 100,000 g for 45 min and further clarified using 0.45 μm syringe filters (CELLTREAT). Imidazole was added to 20 mM before sample was loaded at 1 ml/min onto a 5 ml HisTrap HP affinity purification column (Cytiva), washed with buffer containing 20 mM HEPES pH 7.45, 250 mM NaCl, 10 mM MgCl2, 15% glycerol, 2 μM OT, 0.01%/0.002% LMNG/CHS, 50 mM imidazole, and eluted with the same buffer containing 200 mM imidazole. Fractions containing sample were pooled, spiked with 5 mM CaCl2, loaded onto a column packed with M1 FLAG affinity resin, washed with buffer containing 20 mM HEPES pH 7.45, 150 mM NaCl, 5 mM MgCl2, 4 μM OT, 0.01%/0.002% LMNG/CHS, 1 mM CaCl2, and eluted using the same buffer with 200 ug/ml FLAG peptide (ApexBio) and without CaCl2. Fractions containing sample were pooled, concentrated using a 50 kD MWCO spin column (Vivaspin), and further purified using size exclusion chromatography with a Superose 6 Increase 10/300 GL column (Cytiva) in a buffer containing 20 mM HEPES pH 7.45, 100 mM NaCl, 5 mM MgCl2, 0.002%/0.0004% LMNG/CHS, and 10 μM OT. Monomeric fractions were pooled, spiked with 10% glycerol and an additional 10 μM OT, and concentrated to 2.4 mg/ml before flash freezing for further use.

Expression and purification of heterotrimeric mini-Gq/i protein complex

The miniGq/i chimeric construct was expressed and purified as reported previously23. Briefly, Sf9 insect cells were infected with virus encoding for all G protein subunits (mini-GαqiN and Gβ1/γ2), with histidine tag at the amino terminus of the β subunit and harvested 72 hours post infection. Cells were lysed in buffer containing 20 mM HEPES pH 7.5, 100 mM NaCl, 5% glycerol, 1 mM MgCl2, 100 μM GDP, 30 mM imidazole, 0.2 % (v/v) triton X-100, 5 mM β-mercaptoethanol, protease inhibitors and benzonase. The soluble fraction was isolated by centrifugation at 100,000 g for 40 min at 4°C. The heterotrimer was affinity purified using Ni-NTA resin, and the histidine tag was cleaved by HRV3C protease overnight. The heterotrimer was further purified by size exclusion chromatography on an Enrich SEC 650 column (Bio-Rad) in buffer containing 20 mM HEPES pH 7.5, 100 mM NaCl, 0.001% LMNG, 0.00025% GDN, 0.0001% CHS, 5% glycerol, 1 mM MgCh2, 20 μM GDP and 100 μM TCEP. Fractions representing the G protein heterotrimer were collected and concentrated to ~15 mg/mL and stored at −80 °C until further use.

Expression and purification of scFv16

The secreted single chain construct of Fab16 (scFv16) recognizes an epitope composed of the terminal part of the αN helix of Gαi1 as well as part of the Gβ1 subunit14 and was generated as previously described39. Purified scFv16 was concentrated to 13 mg/ml and flash frozen in buffer containing 20 mM HEPES pH 7.5, 100 mM NaCl, and 15% glycerol for later use.

Formation of the OTR-G-protein complex

Purified OTR was combined with 20 μM OT and heterotrimeric mini-Gq at a molar ration of 1:1.3 and allowed to incubate at room temperature for 45 min. Sample was then chilled to 4°C before the addition of 0.1 U/ml apyrase (Sigma-Aldrich) and 2.5 U/ml 3C protease (Sigma-Aldrich). Complexation was allowed to continue for 45 min before the addition of scFv16 at a molar ratio of 1:2 receptor:scFv16, and the sample was incubated on ice for 12 h before purification by size exclusion chromatography using a Superose 6 Increase 10/300 column. Fractions containing complex were pooled and concentrated to 11.2 mg/ml using a 100 kDa MWCO spin column for freezing.

Cryo-EM grid preparation and data collection

Purified complex (3.5 ul of ~11.2 mg/mL) was applied to UltraAuFoil gold grids (Quantifoil, Au300-R1.2/1.3) that were freshly glow-discharged at 10 mA for 45 s. Grids were vitrified on a Vitrobot Mark IV (ThermoFisher Scientific) at 22°C and 100% humidity plunged into liquid ethane with a wait time of 10 s and blotting time of 3 s. Cryo-EM imaging was performed on a Titan Krios G2 (ThermoFisher Scientific) transmission electron microscope operated at 300 kV with a K3 Summit direct electron detector (Gatan) at a magnification of 29,000x with a pixel size of 0.8521 Å/pix, defocus ranging from −0.8 to −1.8 μm, and with a total dose of 68.44 e−/Å2 in super-resolution mode using SerialEM40.

Cryo-EM data processing and model refinement

Dose-fractionated image stacks from 7,727 movie files at 50 frames each were imported to RELION v3.141 and subjected to beam-induced motion correction with dose weighting using MotionCor2 v1.2.642, followed by CTF estimation using CTFFIND43. Micrographs with estimated CTF <3.5Å were used in further processing. Laplacian autopicking was used to pick 5.67M particles from the resulting 6,913 micrographs, which were then extracted and imported to cryoSPARC v2.144. These particles were down-sampled 4x and subjected to 3 rounds of 2D classification. The remaining 1.74M particles were then subjected to heterogeneous refinement at 2x down-sampling using three 15 Å lowpass-filtered maps made in ChimeraX representing receptor-G protein complex, G protein alone, and receptor alone. Particles corresponding to receptor-G protein complex (985k) were then used for an ab initio reconstruction and 3D classification to yield 436,362 particles, which were imported back to RELION using pyEM v0.545 to perform Bayesian particle polishing46. The particle stack was then transferred to cryoSPARC for an additional round of 2D classification to yield 423,703 final particles, which were CTF refined47 before reconstruction using homogeneous refinement, followed by a global non-uniform refinement48 and local refinements of the transmembrane domain and G protein alone. The map resulting from the local refinement of TMs was sharpened using DeepEMhancer49. The composite map was generated through normalization of thresholds in ChimeraX by combining the local refinement of G protein, local refinement of TMs (without deepEMhancer), and scFv16 from the global non-uniform refinement. All maps were deposited in EMDB in both their original and combined form.

The initial model was fit into the composite map manually in ChimeraX v1.1 (UCSF) using the miniQ heterotrimer (PDB: 6WHA) and an oxytocin receptor homology model from gpcrdb.org50. Extracellular elements (NTD, ECL1, ECL2) were modeled based on the local refinement of TMs that was sharpened by deepEMhancer. Real-space refinement was performed iteratively using Phenix v1.1851 and Coot v0.8.9.352. The final model was inspected for fit to the maps and the geometry was analyzed using MolProbity53 as part of the Phenix software suite. All molecular graphics of both maps and models were generated using UCSF ChimeraX v1.354 or Blender v2 (blender.org).

Bioluminescence resonance energy transfer (BRET) assays

BRET assays were performed and analyzed as previously described for both G protein55 and arrestins23 with the following modifications: HEK-293S cells grown in FreeStyle 293 suspension media (Thermo Fisher) were transfected at a density of 1 million cells/mL in 2 mL volume using 1200 ng total DNA at 1:1:1:1 ratio of receptor:Gα:Gβ:Gγ and a DNA:PEI ratio of 1:5, and incubated in a 24 deep well plate at 220 rpm, 37°C for 48 hours. Cells were harvested by centrifugation, washed with Hank’s Balanced Salt Solution (HBSS) without Calcium/Magnesium (Gibco), and resuspended in assay buffer (HBSS with 20 mM HEPES pH 7.45) with 5 μg/mL freshly prepared coelenterazine 400a (GoldBio). Cells were then placed in white-walled, white-bottom 96 well plates (Costar) in a volume of 60 μl/well and 60,000 cells/well. Drug dilutions were prepared in drug buffer (assay buffer with 0.1% BSA, 6 mM CaCl2, 6 mM MgCl2), of which 30 μl were immediately added to plated cells. For magnesium-sensitivity assays, drug dilutions were prepared in assay buffer with either 1.5 mM EDTA or 6 mM MgCl2. Ten minutes after the addition of ligand, plates were read using a SpectraMax iD5 plate reader using 585 nm and 525 nm emission filters with a one second integration time per well. The computed BRET ratios (GFP2/RLuc8 emission) were normalized to ligand-free control (Net BRET) prior to further analysis.

Ligands

Phe2-oxytocin and 2-(O-methyltyrosine)-oxytocin were custom synthesized by Elim Biopharm, Inc. and verified by liquid chromatography-mass spectrometry. Oxytocin (CAS 50-56-6, ApexBio) and arginine vasopressin (CAS 113-79-1, MedChem Express), and oxytocin analogs were dissolved to 10 mM in citrate buffer pH 4.8 and stored at −20°C until use. Neurotensin(8-13) (CAS 60482-95-3, MedChem Express) and (−)-isoproterenol hydrochloride (CAS 5984-95-2, Sigma-Aldrich) were dissolved to 10 mM in water and stored at −20°C until use.

Molecular dynamics simulations

System set up for molecular dynamics simulations started from the structure determined in this work with the G-protein removed, in the presence and absence of OTR peptide. Receptor was oriented in a lipid bilayer with the OPM webserver56 and solvated in a box containing 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine POPC/CHS lipid bilayer, TIP3P water57, and 150 mM NaCl with the CHARMM-GUI58. PSF and PDB files for CHARMM36m with hydrogen mass repartitioning were taken from the CHARMM-GUI and used for molecular dynamics simulations. MD simulations were run in the NAMD software59. All simulations employed a Langevin thermostat and a Nosé-Hoover Langevin piston barostat at 1 atm with a period of 50 fs and decay of 25 fs; periodic boundary conditions were used with nonbonded interactions smoothed starting at 10 Å to 12 Å with long-range interactions treated with particle mesh Ewald (PME). A 2 fs timestep with SHAKE and SETTLE algorithms was used during equilibration and a 4 fs timestep during production60,61. The system was minimized for 1,500 steps and then gradually heated from 0 to 303.15K in increments of 20K simulating for 0.4 ns at each interval. A 1 kcal/mol/ Å2 harmonic restraint was applied to all non-hydrogen, non-water, and non-ion atoms during heating and for an additional 10 ns of equilibration. The restraints were then applied to only non-hydrogen protein atoms for 10 ns of equilibration, followed by only CA atom restraints for a final 10 ns of equilibration. The first 30 ns of unrestrained simulation was also considered to be equilibration; production simulations were performed for 1.5 μs in triplicate with different initial velocity seeds for each condition. Dihedral angles were measured in VMD for the four residues considered to be part of the TM7 kink (L3167.40-L3177.41-A3187.42-S3197.43), and helicity was measured as a phi angle between −90 and −35 and a psi angle between −70 and −15 degrees as evaluated with a custom python script.

JAWS calculations

JAWS simulation input files were prepared from the structure determined in this work by generating a pdb file of all atoms within 25Å of the magnesium ion and running the JAWS preparation scripts of the GemSpot pipeline62. A 15Å sphere around magnesium was solvated with theta waters to be sampled in the simulation and treated as flexible. The protein was simulated with the OPLS-AA/M force field63, and the TIP4P model was used for the water. MCPRO was used to execute the Monte Carlo simulations64; 5 million steps were employed for solvent equilibration, 10 million in identifying hydration sites, and 50 million in the production phase. Displayed waters are those where there was strong agreement between triplicate simulations in position and binding energy was estimated to be favorable (<0 kcal/mol).

Sequence alignments

Alignments were performed using Geneious Prime v2021.1.1 (Biomatters Ltd.) with MUSCLE v3.8.42565 using the fully annotated protein sequences retrieved from NCBI for the species indicated in the figure.

Cell surface expression testing

Transfection of HEK-293S cells were performed with identical conditions as for BRET assays. Cells were washed with FACS buffer (1% BSA in PBS), stained with mouse anti-FLAG primary antibody (Sigma) and Alexa Fluor 647 Goat Anti-Mouse IgG secondary antibody (Abcam) and read on a NovoCyte Quanteon running NovoExpress v1.3.0 (Agilent) with >50,000 cells counted.

Statistical analysis

Eleven-point dose-response curves in technical duplicate were analyzed by simultaneous curve-fitting of at least 3 biological replicates (minimum 66 data points/curve) using a log(dose) vs. response model in Prism 9.1.0 (GraphPad Software) as previously described55. All 95% confidence intervals for EC50 and Emax were asymmetrically calculated. Comparisons between Emax of stimulation-response curves from different receptors were performed with extra sum-of-squares F Test. Partial agonism was calculated using the Black-Leff operational model66. Any additional details of analysis are described in figure and table legends.

Supplementary Material

Supplementary material

Acknowledgments

The authors would like to thank Constantina Theophanopoulou and Erich Jarvis at The Rockefeller University for their advice and discussions, Elizabeth Montabana for support in data collection, and the Stanford cryo-Electron Microscopy center (cEMc). We thank Ryan Hibbs for gifting the pEZT vector. This work was supported, in part, by National Institute of Health T32-GM089626 (J.G.M.), and the Extreme Science and Engineering Discovery Environment (XSEDE)37 resource comet-gpu and expanse through sdsc-comet/sdsc-expanse allocation TG-MCB190153, which is supported by National Science Foundation grant number ACI-1548562.

Footnotes

Competing interests

The authors declare no competing interests.

Data availability

All data generated or analyzed during this study are included in the Article and its Supplementary Information. Cryo-EM maps of the OT-OTR-Gq complex have been deposited in the Electron Microscopy Data Bank under accession code EMD-24733, including the composite map, global non-uniform refinement, local refinement of TMs with deepEMhancer, and local refinement of G protein. The atomic coordinates have been deposited in the Protein Data Bank under the accession code 7RYC. All other relevant data is available from the corresponding author upon request.

References

  • 1.Fuchs AR, Fuchs F, Husslein P, Soloff MS & Fernström MJ Oxytocin receptors and human parturition: a dual role for oxytocin in the initiation of labor. Science 215, 1396–1398 (1982). [DOI] [PubMed] [Google Scholar]
  • 2.Ferguson JN et al. Social amnesia in mice lacking the oxytocin gene. Nat. Genet 25, 284–288 (2000). [DOI] [PubMed] [Google Scholar]
  • 3.Carcea I. et al. Oxytocin neurons enable social transmission of maternal behaviour. Nature (2021) doi: 10.10.1038/s41586-021-03814-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Augustine RA, Seymour AJ, Campbell RE, Grattan DR & Brown CH Integrative neuro-humoral regulation of oxytocin neuron activity in pregnancy and lactation. J. Neuroendocrinol (2018) doi: 10.10.1111/jne.12569. [DOI] [PubMed] [Google Scholar]
  • 5.Eliava M. et al. A New Population of Parvocellular Oxytocin Neurons Controlling Magnocellular Neuron Activity and Inflammatory Pain Processing. Neuron 89, 1291–1304 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.King CE, Gano A & Becker HC The role of oxytocin in alcohol and drug abuse. Brain Res. 1736, 146761 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Kosfeld M, Heinrichs M, Zak PJ, Fischbacher U & Fehr E Oxytocin increases trust in humans. Nature 435, 673–676 (2005). [DOI] [PubMed] [Google Scholar]
  • 8.Patel SS & Ludmir J Drugs for the Treatment and Prevention of Preterm Labor. Clin. Perinatol 46, 159–172 (2019). [DOI] [PubMed] [Google Scholar]
  • 9.Peñagarikano O. et al. Exogenous and evoked oxytocin restores social behavior in the Cntnap2 mouse model of autism. Sci. Transl. Med 7, 271ra8 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Martins D, Paduraru M & Paloyelis Y Heterogeneity in response to repeated intranasal oxytocin in schizophrenia and autism spectrum disorders: A meta-analysis of variance. Br. J. Pharmacol (2021) doi: 10.10.1111/bph.15451. [DOI] [PubMed] [Google Scholar]
  • 11.Yoon S & Kim Y-K The Role of the Oxytocin System in Anxiety Disorders. Adv. Exp. Med. Biol 1191, 103–120 (2020). [DOI] [PubMed] [Google Scholar]
  • 12.Waltenspühl Y, Schöppe J, Ehrenmann J, Kummer L & Plückthun A Crystal structure of the human oxytocin receptor. Sci. Adv 6, eabb5419 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Busnelli M. et al. Functional selective oxytocin-derived agonists discriminate between individual G protein family subtypes. J. Biol. Chem 287, 3617–3629 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Maeda S. et al. Development of an antibody fragment that stabilizes GPCR/G-protein complexes. Nat. Commun 9, 3712 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Israeli H. et al. Structure reveals the activation mechanism of the MC4 receptor to initiate satiation signaling. Science 372, 808–814 (2021). [DOI] [PubMed] [Google Scholar]
  • 16.Wesley VJ, Hawtin SR, Howard HC & Wheatley M Agonist-specific, high-affinity binding epitopes are contributed by an arginine in the N-terminus of the human oxytocin receptor. Biochemistry 41, 5086–5092 (2002). [DOI] [PubMed] [Google Scholar]
  • 17.Klein U, Gimpl G & Fahrenholz F Alteration of the myometrial plasma membrane cholesterol content with beta-cyclodextrin modulates the binding affinity of the oxytocin receptor. Biochemistry 34, 13784–13793 (1995). [DOI] [PubMed] [Google Scholar]
  • 18.Lemel L. et al. The ligand-bound state of a G protein-coupled receptor stabilizes the interaction of functional cholesterol molecules. J. Lipid Res 62, 100059 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Jafurulla M, Aditya Kumar G, Rao BD & Chattopadhyay A A Critical Analysis of Molecular Mechanisms Underlying Membrane Cholesterol Sensitivity of GPCRs. Adv. Exp. Med. Biol 1115, 21–52 (2019). [DOI] [PubMed] [Google Scholar]
  • 20.Taghon GJ, Rowe JB, Kapolka NJ & Isom DG Predictable cholesterol binding sites in GPCRs lack consensus motifs. Struct. Lond. Engl. 1993 29, 499–506.e3 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Ballesteros JA & Weinstein H [19] Integrated methods for the construction of three-dimensional models and computational probing of structure-function relations in G protein-coupled receptors. in Methods in Neurosciences vol. 25 366–428 (Elsevier, 1995). [Google Scholar]
  • 22.Mouillac B. et al. The binding site of neuropeptide vasopressin V1a receptor. Evidence for a major localization within transmembrane regions. J. Biol. Chem 270, 25771–25777 (1995). [DOI] [PubMed] [Google Scholar]
  • 23.Kim K. et al. Structure of a Hallucinogen-Activated Gq-Coupled 5-HT2A Serotonin Receptor. Cell 182, 1574–1588.e19 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Wang L. et al. Cryo-EM structure of the AVP-vasopressin receptor 2-Gs signaling complex. Cell Res. 31, 932–934 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Zhou F. et al. Molecular basis of ligand recognition and activation of human V2 vasopressin receptor. Cell Res. 31, 929–931 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Bous J. et al. Cryo-electron microscopy structure of the antidiuretic hormone argininevasopressin V2 receptor signaling complex. Sci. Adv 7, eabg5628 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Lamont RF The development and introduction of anti-oxytocic tocolytics. BJOG Int. J. Obstet. Gynaecol 110 Suppl 20, 108–112 (2003). [PubMed] [Google Scholar]
  • 28.Zheng H, Chruszcz M, Lasota P, Lebioda L & Minor W Data mining of metal ion environments present in protein structures. J. Inorg. Biochem 102, 1765–1776 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Michel J, Tirado-Rives J & Jorgensen WL Prediction of the water content in protein binding sites. J. Phys. Chem. B 113, 13337–13346 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Antoni FA & Chadio SE Essential role of magnesium in oxytocin-receptor affinity and ligand specificity. Biochem. J 257, 611–614 (1989). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Jing Z, Liu C, Qi R & Ren P Many-body effect determines the selectivity for Ca2+ and Mg2+ in proteins. Proc. Natl. Acad. Sci. U. S. A 115, E7495–E7501 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Zhu X, Gilbert S, Birnbaumer M & Birnbaumer L Dual signaling potential is common among Gs-coupled receptors and dependent on receptor density. Mol. Pharmacol 46, 460–469 (1994). [PubMed] [Google Scholar]
  • 33.Theofanopoulou C, Gedman G, Cahill JA, Boeckx C & Jarvis ED Universal nomenclature for oxytocin-vasotocin ligand and receptor families. Nature 592, 747–755 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Donaldson ZR & Young LJ Oxytocin, vasopressin, and the neurogenetics of sociality. Science 322, 900–904 (2008). [DOI] [PubMed] [Google Scholar]
  • 35.Slodkowicz G & Goldman N Integrated structural and evolutionary analysis reveals common mechanisms underlying adaptive evolution in mammals. Proc. Natl. Acad. Sci 117, 5977–5986 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Bharadwaj VN et al. A new hypothesis linking oxytocin to menstrual migraine. Headache J. Head Face Pain 61, 1051–1059 (2021). [DOI] [PubMed] [Google Scholar]
  • 37.Towns J. et al. XSEDE: Accelerating Scientific Discovery. Comput. Sci. Eng 16, 62–74 (2014). [Google Scholar]
  • 38.Morales-Perez CL, Noviello CM & Hibbs RE Manipulation of Subunit Stoichiometry in Heteromeric Membrane Proteins. Struct. Lond. Engl. 1993 24, 797–805 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Koehl A. et al. Structure of the μ-opioid receptor-Gi protein complex. Nature 558, 547–552 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Mastronarde DN Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol 152, 36–51 (2005). [DOI] [PubMed] [Google Scholar]
  • 41.Zivanov J. et al. New tools for automated high-resolution cryo-EM structure determination in RELION-3. eLife 7, e42166 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Zheng SQ et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods 14, 331–332 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Rohou A & Grigorieff N CTFFIND4: Fast and accurate defocus estimation from electron micrographs. J. Struct. Biol 192, 216–221 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Punjani A, Rubinstein JL, Fleet DJ & Brubaker MA cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14, 290–296 (2017). [DOI] [PubMed] [Google Scholar]
  • 45.Asarnow D, Palovcak E & Cheng Y asarnow/pyem: UCSF pyem v0.5. (Zenodo, 2019). doi: 10.10.5281/ZENODO.3576630. [DOI] [Google Scholar]
  • 46.Zivanov J, Nakane T & Scheres SHW A Bayesian approach to beam-induced motion correction in cryo-EM single-particle analysis. IUCrJ 6, 5–17 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Zivanov J, Nakane T & Scheres SHW Estimation of high-order aberrations and anisotropic magnification from cryo-EM data sets in RELION-3.1. IUCrJ 7, 253–267 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Punjani A, Zhang H & Fleet DJ Non-uniform refinement: adaptive regularization improves single-particle cryo-EM reconstruction. Nat. Methods 17, 1214–1221 (2020). [DOI] [PubMed] [Google Scholar]
  • 49.Sanchez-Garcia R. et al. DeepEMhancer: a deep learning solution for cryo-EM volume post-processing. Commun. Biol 4, 874 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Kooistra AJ et al. GPCRdb in 2021: integrating GPCR sequence, structure and function. Nucleic Acids Res. 49, D335–D343 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Adams PD et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr 66, 213–221 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Emsley P & Cowtan K Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr 60, 2126–2132 (2004). [DOI] [PubMed] [Google Scholar]
  • 53.Williams CJ et al. MolProbity: More and better reference data for improved all-atom structure validation. Protein Sci. Publ. Protein Soc 27, 293–315 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Pettersen EF et al. UCSF ChimeraX: Structure visualization for researchers, educators, and developers. Protein Sci. Publ. Protein Soc 30, 70–82 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Olsen RHJ et al. TRUPATH, an open-source biosensor platform for interrogating the GPCR transducerome. Nat. Chem. Biol 16, 841–849 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Lomize MA, Pogozheva ID, Joo H, Mosberg HI & Lomize AL OPM database and PPM web server: resources for positioning of proteins in membranes. Nucleic Acids Res. 40, D370–376 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW & Klein ML Comparison of simple potential functions for simulating liquid water. J. Chem. Phys 79, 926–935 (1983). [Google Scholar]
  • 58.Lee J. et al. CHARMM-GUI Input Generator for NAMD, GROMACS, AMBER, OpenMM, and CHARMM/OpenMM Simulations Using the CHARMM36 Additive Force Field. J. Chem. Theory Comput 12, 405–413 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Phillips JC et al. Scalable molecular dynamics with NAMD. J. Comput. Chem 26, 1781–1802 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Ryckaert J-P, Ciccotti G & Berendsen HJC Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys 23, 327–341 (1977). [Google Scholar]
  • 61.Miyamoto S & Kollman PA Settle: An analytical version of the SHAKE and RATTLE algorithm for rigid water models. J. Comput. Chem 13, 952–962 (1992). [Google Scholar]
  • 62.Robertson MJ, van Zundert GCP, Borrelli K & Skiniotis G GemSpot: A Pipeline for Robust Modeling of Ligands into Cryo-EM Maps. Structure 28, 707–716.e3 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.van Zundert GCP, Moriarty NW, Sobolev OV, Adams PD & Borrelli KW Macromolecular refinement of X-ray and cryoelectron microscopy structures with Phenix/OPLS3e for improved structure and ligand quality. Structure (2021) doi: 10.10.1016/j.str.2021.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Jorgensen WL & Tirado-Rives J Molecular modeling of organic and biomolecular systems using BOSS and MCPRO. J. Comput. Chem 26, 1689–1700 (2005). [DOI] [PubMed] [Google Scholar]
  • 65.Edgar RC MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66.Black JW, Leff P, Shankley NP & Wood J An operational model of pharmacological agonism: the effect of E/[A] curve shape on agonist dissociation constant estimation. Br. J. Pharmacol 84, 561–571 (1985). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material
NIHMS2077289-supplement-Supplementary_material.pdf (26.1MB, pdf)

Data Availability Statement

All data generated or analyzed during this study are included in the Article and its Supplementary Information. Cryo-EM maps of the OT-OTR-Gq complex have been deposited in the Electron Microscopy Data Bank under accession code EMD-24733, including the composite map, global non-uniform refinement, local refinement of TMs with deepEMhancer, and local refinement of G protein. The atomic coordinates have been deposited in the Protein Data Bank under the accession code 7RYC. All other relevant data is available from the corresponding author upon request.

RESOURCES