Structural and functional determination of peptide versus small molecule ligand binding at the apelin receptor

Ethics declaration

Ethical approval and written informed consent was obtained from the participants by the National Institute for Health Research (NIHR) BioResource for Rare Diseases³⁴. All animal care and rodent experiments complied with the Home Office (United Kingdom) guidelines under the Animals (Scientific Procedures) Act 1986 Amendment Regulations (SI 2012/3,039) and were approved by the local ethics committee (University of Cambridge Animal Welfare and Ethical Review Body).

Identifying rare sequence variation in the apelin receptor (APLNR) gene

Whole genomes sequenced by the NIHR BioResource (NBR) project, a prospective component of the Genomics England 100,000 Genomes Project, were screened for rare sequence variants in the apelin receptor (APLNR) gene that had a minor allele frequency of <1 in 10,000 in the general population. The cohort comprised 13,037 patients diagnosed with one of several rare diseases or cancers. Following whole genome sequencing, case-control rare variant analysis was performed as described previously⁵².

We used predetermined exclusion criteria to identify nine apelin receptor single nucleotide missense variants that occurred at highly conserved sites, predicted using in silico conservation scores from GERP (Genomic Evolutionary Rate Profiling), phastCons, and phyloP (http://genome.ucsc.edu)^53,54, and that were likely to show pathogenicity using deleteriousness scores from SIFT (Sorting Intolerant From Tolerant)⁵⁵ and PolyPhen-2 (Polymorphism Phenotyping v2)⁵⁶. Additionally, these variants occurred at sites predicted to affect receptor binding and/or function when compared to corresponding sites in similar class A GPCRs using GPCRdb (https://gpcrdb.org/)³⁷. For example, apelin receptor residue R127^3.50 forms part of the E/DRY motif, which is highly conserved amongst class A GPCRs to regulate conformation and activity, and mutation would likely alter protein function⁵⁷. The exclusion criteria ensured only variants that were ultra-rare in the general population, likely under selective pressure, and that occur at amino acid sites that are highly evolutionarily conserved were retained. After initial filtering, subsequent analyses of genetic data in the gnomAD (version 4.1.0) and UK BioBank cohorts, comprising a total of 807,602 samples, confirmed that selected variants remained at an ultra-rare level of occurrence across these datasets. Overall, nine missense variants, and two single nucleotide deletion frameshift mutations, were subsequently characterised in vitro.

Apelin receptor plasmid DNA

Wild-type apelin receptor was PCR cloned into pcDNA3.1 (Invitrogen), using Phusion Hot Start II DNA polymerase (ThermoFisher Scientific), in accordance with the manufacturer’s instructions. Apelin receptor variants were made by site-directed mutagenesis, using Phusion Hot Start II DNA polymerase (ThermoFisher Scientific), in accordance with the manufacturer’s instructions. Receptor plasmid constructs were tagged C-terminally with an eGFP reporter. Cloned plasmid vectors were transformed in DH5α competent Escherichia coli, before isolation using a QIAprep Spin Miniprep Kit (QIAGEN), in accordance with the manufacturer’s instructions.

CHO-K1 cell culture and transfection

CHO-K1 (ECACC 85051005) cells were cultured in 25 mL DMEM/F-12 (1:1) nutrient mix + L-glutamine (ThermoFisher Scientific), supplemented with 10% FBS (ThermoFisher Scientific) in the presence of 0.1 mg/mL Normocin antibiotic formulation (InvivoGen) in T175 cell culture flasks (Corning). Cells were maintained at 37 °C, in humidified air containing 5% CO₂. Transfection with wild-type or variant apelin receptor plasmid DNA (1 ng/mL final concentration) was performed using a TransIT-CHO Transfection Kit (MIR 2174, Mirus), in serum-free Opti-MEM I media (11058021, Gibco), as per the manufacturer’s instructions.

High content screening

High content screening was used to characterise apelin receptor variant expression, membrane localisation, and fluorescent ligand binding. CHO-K1 cells transiently transfected with wild-type or variant apelin receptor construct tagged C-terminally with eGFP were plated at ~10,000/well in CellCarrier-96 Ultra Plates (PerkinElmer). On the day of experimentation, cells were washed with HBSS before treatment with well-validated apelin647 (300 nM) or ELA647 (1 µM) fluorescent ligand³⁸ for 90 min at room temperature, until equilibrium was reached. Non-specific binding was determined in the presence of a saturating 10 µM concentration of unlabelled [Pyr¹]apelin-13. Fluorescent ligand was washed out with HBSS, and cells were fixed with 4% formaldehyde for ~3 min. After another HBSS wash, cells were treated with Hoechst 33342 nuclear marker (10 µg/mL) for 15 min. After a final wash, cells were maintained in 100 µL HBSS.

Cells maintained in the 96-well plates were imaged on an Opera Phenix High Content Screening System (PerkinElmer), a highly sensitive, dual-camera spinning disk confocal fluorescent imaging platform, using a 40x/NA1.1 water immersion objective. The imaging field included a grid of 16 (4 × 4) regions, where each region is ~250 × 250 µm², at a focal depth of 0.5 µm above the bottom of the plate. Three fluorescent channels were used. The first channel (blue) used an excitation wavelength of 405 nm and emission filter of 435–480 nm for the Hoechst 33342 nuclear marker. The second channel (green) used an excitation wavelength of 488 nm and emission filter of 500–550 nm for the eGFP reporter. The third channel (red) used an excitation wavelength of 631 nm for the apelin647 and ELA647 fluorescent ligand. For all channels, excitation laser intensity was set at 50% power, with a 50 ms exposure time. Additionally, each imaging field was imaged using bright-field and digital phase contrast (DPC). The DPC images were used to identify outer boundaries of the cells as a surrogate of the membrane, and a collar extending ±2 µm from this membrane region was established to segment this cellular compartment.

Following acquisition, fluorescence was quantified using in-built Harmony High Content Imaging and Analysis Software (PerkinElmer), as described previously³⁸. Only cells that were positive for eGFP (i.e. mean whole cell eGFP fluorescence of >120 arbitrary fluorescence units, AFU) were included in the analysis, where background fluorescence was calculated as ~112 AFU. Mean membrane fluorescence was then quantified for the green channel as an indication of membrane protein expression, and for the red channel as an indication of fluorescent ligand binding.

Radioligand binding

Radioligand binding studies were performed using [¹²⁵I]-apelin-13 ([Glp⁶⁵,Nle⁷⁵,Tyr⁷⁷][¹²⁵I]-apelin-13, PerkinElmer)^10,30,38. The radiolabel has a specific activity of 2200 Ci/mmol. Plastic-ware used in radioligand binding experiments was coated with Sigmacote, a siliconising reagent that forms a covalent film to reduce sticking of the radiolabel.

Membrane protein preparations were made by harvesting CHO-K1 cells transiently transfected with wild-type or variant apelin receptor construct into tubes. Tubes were spun at 1000 xg for 10 min at 4 °C to pellet cells, before resuspending and triturating in ice-cold TRIS wash buffer (50 mM TRIS-HCl in deionised water, pH balanced to 7.4 at room temperature) to induce complete hypotonic lysis of cells. Cells were spun a second time at 21,000 × g for 20 min at 4 °C, to obtain a crude membrane fraction. Supernatant was discarded and membrane preparations resuspended in TRIS wash buffer. Cell membrane preparations were stored at –70 °C until use in experiments. Protein concentration of cell membrane preparations was determined using a DC Protein Assay (Bio-Rad), as per the manufacturer’s instructions.

For binding experiments, membrane preparations were diluted in binding buffer (50 mM TRIS-HCl and 5 mM MgCl₂, balanced to pH 7.4 at room temperature) to give a final protein concentration of 1 mg/mL. For saturation binding experiments, protein was treated with a concentration range (2 pM–1 nM) of [¹²⁵I]-apelin-13, diluted in binding buffer, and incubated for 90 min at room temperature to reach equilibrium. For competition binding experiments, protein was treated with a single 0.1 nM concentration of [¹²⁵I]-apelin-13, against a concentration range (1 pM–1 μM) of ELA-11 peptide. Non-specific binding was determined in the presence of a saturating 10 µM concentration of unlabelled [Pyr¹]apelin-13 in all instances. At the end-point, samples were spun in a centrifuge at 20,000 × g for 10 min at 4 °C to terminate equilibrium. Supernatant was aspirated, and pellets were resuspended and triturated in 500 μL ice-cold wash buffer, over ice. Samples were spun a second time at 20,000 × g for 10 min at 4 °C, before aspiration of the supernatant. Radioactivity in pellets was counted using a Cobra II model 5003 gamma counter (Packard).

Saturation binding data were analysed using the EBDA and LIGAND components of the KELL (Kinetic, EBDA, Ligand, Lowry) software package (Biosoft)⁵⁸ to generate receptor affinity (K_D) and receptor density (B_max) values. B_max values, in fmol/mg, were calculated using the known K_D value of the radiolabel (0.076 nM), known specific activity of the radiolabel (2200 Ci/mmol), and concentration of protein used (1 mg/mL). Saturation data were presented using the nonlinear one site-specific binding with Hill slope model fitted by GraphPad Prism version 6.07 for Windows (GraphPad Software). Competition binding data were presented using the nonlinear one site Fit K_i [2] model fitted by GraphPad Prism, with constraints for the radiolabel concentration (0.1 nM) and K_D (0.076 nM). The software was also used to calculate the binding affinity (K_i value) of competing ELA-11 ligand using the Cheng-Prusoff equation.

Dynamic mass redistribution functional assay

Functional responses of CHO-K1 cells transfected with wild-type or variant apelin receptor to [Pyr¹]apelin-13 were assessed using BIND reader technology (SRU Biosystems) dynamic mass redistribution assay. Cells were plated at ~25,000/well in proprietary 96-well biosensor plates that are covered with a nanostructured optical grate. Cells were subsequently washed with HBSS before acclimatising to room conditions for 20 min. The plate was inserted into the BIND reader and a 5 min baseline read was recorded before treatment with a concentration range (0.1 nM–1 µM final) of [Pyr¹]apelin-13. Maximum change in peak wavelength value (ΔPWV) from baseline after drug addition was used to plot concentration response curves, where data were fitted to four parameter logistic curves in GraphPad Prism. Maximum response (E_max) and pD₂ (–log10 EC₅₀) values were calculated for [Pyr¹]apelin-13 at wild-type and variant apelin receptors. Untransfected cells were unresponsive to [Pyr¹]apelin-13.

AlphaFold2 modelling

The active state of apelin receptor bound to peptide ligands was modelled with AlphaFold2 (https://www.readcube.com/library/5d1671d8-5c39-40fd-8d64-7a7d3a0c899e:ccb52d59-1461-4b7a-a85e-1e21631710a1). AlphaFold2 was run in multimer mode using the sequences of the apelin receptor, mini Gi, and either apelin or ELA. Five structures were generated for each language model and the output 25 models were ranked. All models were examined simultaneously and the best representative model by score was selected for analysis. AlphaFold2 was chosen for modelling these structures as the native apelin peptide bound to active apelin receptor has not as yet been determined. While structures of ELA bound to active apelin receptor do exist, the binding mode is variable and the resolution is low. The AlphaFold2 model of ELA at apelin receptor is consistent with the existing structures, is supported by the relevant SAR, and provides a comparative model to the apelin-apelin receptor predicted model. The mutation, T/M89^2.64, was modelled using Schrodinger’s Maestro suite (Schrödinger Release 2022-2: Maestro, Schrödinger, LLC, New York, NY, 2021.). The binding surface was calculated using GRID from Molecular Discovery (https://pubmed.ncbi.nlm.nih.gov/3892003/).

Generation of human apelin stabilised receptor (NxStaR®)

Full-length human apelin receptor was used as a background to generate a thermostabilised receptor using a combination of scanning mutagenesis^41,42 utilising tritiated [³H]-NXE’870 (RC TRITEC), and directed evolution^43,44 utilising [Cys(AF488)]-[PEG4]-QRPRLSHKGP-[Nle]-P-Y(OBn)-acid (Cambridge Research Biochemicals). For assessment of mutant thermostability, cells transfected with receptor were treated with the [³H]-NXE’870 radiolabel for 2 h at room temperature to reach equilibrium, before termination by incubation at 4 °C for 5 min. The compound chemotype is similar to CMF-019³⁰. Cells were then solubilised with 0.1% (w/v) n-dodecyl-β -d- maltopyranoside (DDM) supplemented with 0.12% cholesteryl hemisuccinate and 0.6% (w/v) 3-(3-cholamidopropyl) dimethylammonio-1-propanesulfonate (CHAPS) (Anatrace) for 1 h at 4 °C. Crude lysates were cleared by centrifugation at 16,000 × g for 15 min. Thermostability was measured by incubation of lysate samples at different temperatures for 30 min, followed by separation of unbound radioligand using gel filtration. Levels of ligand bound receptor were determined using liquid scintillation counting. Thermal stability (apparent Tm) was defined as the temperature at which 50% ligand binding was retained following plotting of the radioligand binding data against temperature using the sigmoidal dose-response (variable slope) equation in GraphPad Prism. The resultant apelin receptor NxStaR contained six thermostabilising mutations: T/V87^2.62, N/A112^3.35, T/M207^5.44, F/L214^5.51, I/A224^5.61, and S/A298^7.42. The NxStaR thermostabilising mutations were functionally assessed, showing comparable expression and binding compared to the wild-type apelin receptor template but overall improved thermostability (see Table S3).

To generate the minimal construct required for crystallisation, the apelin receptor NxStaR was truncated at the N- and C-termini (leaving receptor residues 7-330), and two palmitoylation sites (C/L325 and C/M326) and one N-linked glycosylation site (T/N177) were mutated. To identify the most optimal bRIL (E. coli b₅₆₂RIL domain) fusion construct, a matrix was designed between I228 and E238 in ICL3, with 30 constructs generated. These constructs were ranked using thermostability and fSEC analysis⁵⁹, and the most optimal fusion construct was found to comprise a bRIL between I228 and E235.

Apelin receptor NxStaR expression, membrane preparation, and protein purification

Towards protein preparation, we found a combination of single mutations and truncations that improves expression of the apelin receptor in insect cells, and its thermostability in the presence of the small synthetic agonist NXE’870 (Fig. S7). The compound chemotype is similar to CMF-019³⁰. For crystallisation trials in LCP, the stabilised receptor (NxStaR) was first fused with Clostridium pasteurianum rubredoxin inserted into ICL3⁶⁰, as reported previously for the crystal structure determination of the apelin receptor in complex with the synthetic peptide agonist AMG3054³¹. It could be solubilised and purified using the soft detergent dodecyl-β-D-maltopyranoside, in combination with cholesteryl hemi-succinate, in yield and homogeneity amenable to crystallisation experiments. This construct produced small crystals and more sizeable crystallisation hits were optimised by replacing the rubredoxin fusion with the E. coli soluble cytochrome b₅₆₂RIL domain inserted between residues 228 and 235, and by carrying out purification and crystallisation experiments in the presence of CMF-019 (see below).

The apelin receptor NxStaR bRIL baculovirus construct was generated using the bac-to-bac system (ThermoFisher Scientific), as per the manufacturer’s instructions. For expression in insect cells, Spodoptera frugiperda (Sf21) cells were grown at 27 °C with shaking at 135 rpm in 2.5 L total suspension of ESF921 medium (Expression Systems) supplemented with 10% (v/v) foetal bovine serum (Sigma-Aldrich) and 1% (v/v) of a Penicillin-Streptomycin mixture (PAA Laboratories) in 5 L Thomson flasks. Typically, cells were infected at a density of 2.5 to 3.0 × 10⁶ cells/ml with the recombinant virus at an approximate multiplicity of infection of 1, and grown further for 48 h. Cells were then harvested by centrifugation at 3500 rpm at 4 °C, washed in PBS supplemented with Complete EDTA-free protease inhibitor cocktail tablets (Roche), and harvested again by centrifugation at 3500 rpm at 4 °C, and finally stored as pellets at –80 °C.

For membrane preparation, one cell pellet corresponding to 5 L of culture was thawed in water at room temperature for 1 h and resuspended with 300 mL of 25 mM HEPES-NaOH, pH 7.5, 20 mM KCl, 10 mM MgCl₂, and Complete EDTA-free protease inhibitor cocktail tablets (Roche), and stirred for 15 min to remove clumps. Cells were then disrupted in one pass through a microfluidizer at ~15,000 psi (processor M-110L Pneumatic, Microfluidics) cooled with ice. Membranes were pelleted by centrifugation at 200,000 × g at 4 °C for 45 min, and resuspended using a glass dounce homogenizer in lysis buffer supplemented with 1 M NaCl, and centrifuged again at 200,000 × g for 45 min at 4 °C. After removal of the supernatant, membranes were resuspended using a glass dounce homogenizer again in lysis buffer (~150 mL final volume) and stored at –80 °C.

Prior to solubilisation, membranes were incubated with CMF-019 and, subsequently, with iodoacetamide. For solubilisation, the membrane solution was supplemented with buffer including n-Dodecyl-β-D-maltopyranoside (Anatrace) and cholesteryl hemisuccinate (Sigma). For purification, the solubilisation slurry was cleared by centrifugation and the supernatant was loaded onto Strep-Tactin Superflow resin (QIAGEN). After washing the resin, the protein was eluted using D-desthiobiotin (Sigma-Aldrich), and the fractions containing the apelin receptor NxStaR bRIL protein were pooled and concentrated on a 100 kDa concentrator for buffer exchange in buffer without D-desthiobiotin using a PD-10 column (Cytiva). The apelin receptor NxStaR bRIL protein pool was then supplemented with 3C PreScission Protease (Cytiva) and PNGase F protein (New England Biolabs) to remove the C-terminal purification tag and potential N-glycosylation, respectively. The apelin receptor NxStaR bRIL protein was further purified by incubation with Ni-NTA agarose (QIAGEN) and GST (Cytiva) resins to capture the C-terminal tags and the enzymes. After removal of the affinity resin beads, the protein was further concentrated and spun at 200,000 × g for further analysis on size exclusion chromatography and crystallization experiments.

Apelin receptor NxStaR crystallisation, structure determination, and refinement

Crystallisation experiments were carried out in LCP at 20 °C. Monoolein (Nu-Chek), supplemented with cholesterol (Sigma) (ratio of 9:1 w/w) and 100 µM CMF-019, was mixed with the protein at a concentration of ~70 mg/mL (based on extinction coefficient and molecular weight of construct) using the twin-syringe method at a ratio of 2:3 (v/v). Boli were dispensed onto Laminex glass bases (Molecular Dimensions Ltd) and overlaid with 800 nL of crystallization solution using a Mosquito LCP crystallization robot (STP Labtech), and sealed using Laminex film covers (Molecular Dimensions Ltd). Rod-shaped crystals used for data collection grew at 20 °C for ~5–6 days in conditions consisting of 25 to 35% polyethylene glycol (PEG) 400, 100 mM sodium citrate, pH 4.75, and 150 mM of ammonium phosphate, potassium phosphate, or sodium acetate, harvested in their mother liquor with a 5% higher PEG 400 concentration, and flash-frozen in liquid nitrogen.

X-ray diffraction data frames (0.25°/frame; ~400 frames per crystal) were collected on a Pilatus 6M detector at beamline I24 (Diamond Light Source). Data from individual crystals were integrated using XDS⁶¹, combined using POINTLESS⁶² within the CCP4 program suite⁶³, and further scaled together using AIMLESS, and corrected for anisotropy using NxSTARANISO within aP_scale⁶⁴ (Global Phasing Limited, Cambridge, UK).

The structure of the apelin receptor NxStaR bRIL-CMF-019 complex was determined by molecular replacement (MR) with Phaser⁶⁵, using the previously reported structure of the apelin receptor in complex with AMG3054 peptide³¹ as the search model (PDB code: 5VBL). Model refinement was performed first using phenix.refine⁶⁶ and further using BUSTER⁶⁷ (Global Phasing Limited, Cambridge, UK), including TLS refinement for two groups corresponding to the apelin receptor and the bRIL fusion. The final protein structure includes the bRIL fusion in ICL3 and apelin receptor residues 29 to 324, but excluding residues 55 to 61 in ICL1, and 135 to 139 in ICL2. Structure panels were generated using PyMOL. Crystallographic X-ray data collection & structure refinement statistics are provided in the Supplemtary Information (Table S4).

Surface plasmon resonance

SPR experiments investigating pH dependence of peptide and compound binding were performed using a Biacore 8K+ instrument (Cytiva) and experiments investigating binding of Apelin 13 isoforms were performed on a Biacore T200 instrument (Cytiva). Both were equipped with a NTA sensor chip (Cytiva). Apelin receptors and a negative control were immobilised at 25 °C by a combination of Ni-NTA capture and amine coupling. The buffer used comprised 0.01 M HEPES, pH 7.4, 0.15 M NaCl, 0.02% DDM. 0.5 M EDTA pH 8.0 was injected for 60 s at 5 µL/min followed by a 60 s injection of run buffer containing 0.5 mM NiCl₂. The surface was activated with a 1:1 mixture of 0.1 M NHS and 0.5 M EDC at a flow rate of 10 µL/min for 420 s. The receptors were injected at a concentration of 1 µM at 5 µL/min until a surface response of 5000 RU was achieved. The surface was equilibrated for 12 h in buffer before commencing experiments.

Ligand binding kinetics were analysed at 10 °C and had a data collection rate of 10 Hz. The buffer used for measuring peptide isoform binding was 0.01 M HEPES, pH 7.5, 0.15 M NaCl, 0.02% DDM, 5% DMSO. The analyses were run in single-cycle format with five concentration points in a twofold dilution series between 6.25 nM and 100 nM for Apelin-13 and Apelin-13 ΔC2. The concentration range was between 62.5 nM and 1 µM for all other isoforms. The association phase was 120 s for each concentration and the dissociation phase was 600 s. The flow rate was 50 µL/min.

Compound and peptide binding was measured in pH 7.5 buffer previously described and in buffer at pH 6.0 comprising of 0.025 M MES, pH 6.0, 0.15 M NaCl, 0.02% DDM, 5% DMSO. The analyses were run in single-cycle format with four concentration points in a twofold dilution series. The concentration range was 25–200 nM for NXE’992, NXE’145 and NXE’870 and 6.25–50 nM for NXE’686. The association phase was 60 s for each concentration and the dissociation phase was 300 s. There were 4 blank cycles between compound injections.

The data were processed using Biacore T200 Evaluation Software (Version 2.0, Cytiva) and Biacore Insight Evaluation Software (Version 5.0, Cytiva). Sensorgrams were reference and blank subtracted before fitting with 1:1 binding model which accounted for drift, bulk shift and mass transport. Sensorgram images were prepared using Microsoft Excel.

Human embryonic stem cell culture, differentiation, and purification

Pluripotent, undifferentiated H9 human embryonic stem cells (hESCs, WA09 from WiCell) were maintained as culture colonies on gelatin/MEF media (0.1% gelatin from porcine skin, advanced DMEM/F12, 10% Foetal Bovine Serum (FBS), 1% L-Glutamine, 100 μM β-mercaptoethanol, 100 μ/mL penicillin-streptomycin) coated 6-well plates in chemically defined medium (CDM-BSA: IMDM/F12 (1:1), 15 µg/mL transferrin, 7 µg/mL insulin, 450 mM monothioglycerol, 1% chemically defined concentrated lipids, 5 mg/mL BSA, 100 μ/mL penicillin-streptomycin), supplemented with fibroblast growth factor 2 (FGF2, 12 ng/mL) and Activin-A (10 ng/mL) (maintenance media) with daily media refreshment. Similar protocols have been used previously⁵¹.

For differentiation, cells were plated at ~80,000/well in CDM-BSA supplemented with FGF2 (12 ng/mL), Activin-A (30 ng/mL), and Rho-associated protein kinase inhibitor (ROCKi, 10 µM) on Matrigel-coated 6-well plates. Following incubation at 37 °C for 4 h, mesoderm initiation was started. Cells were treated with 2 mL/well CDM-BSA supplemented with FGF2 (20 ng/mL), Ly294002 (phosphoinositide 3-kinase inhibitor, 10 µM), Activin-A (50 ng/mL) and Bone Morphogenetic Protein 4 (BMP4, 10 ng/mL). Following a further incubation at 37 °C for 42 h, media was removed, and cells washed with PBS, before covering with 2 mL/well CDM-BSA supplemented with FGF2 (8 ng/mL), BMP4 (10 ng/mL), retinoic acid (1 µM), and the WNT signalling pathway inhibitor IWR1-endo (1 ng/mL). After 48 h, media was refreshed, and after a further 48 h, cells were washed with PBS and media was changed to 2 mL/well CDM-BSA supplemented with FGF2 (8 ng/ml) and BMP4 (10 ng/ml).

After 48 h, media was removed, and cells washed with PBS, before covering with 2 mL/well CDM-BSA. Differentiated cardiomyocytes (hESC-CMs) were maintained in CDM-BSA media, with media changes every other day until spontaneous contraction was observed.

For purification, cardiomyocytes were split and re-plated in Matrigel-coated 6-well plates, and covered with CDM-BSA supplemented with ROCKi (10 µM) to promote survival. Cells were incubated overnight to allow adherence and recovery before transferring to lactate selection media (DMEM no glucose, no sodium pyruvate, 1X MEM Non-Essential Amino Acids, 4 mM sodium L-lactate dissolved in HEPES) for 72 h, with one media refresh, in order to generate a pure population of cardiomyocytes. After 72 h, media was changed to CDM-BSA, and refreshed every second day.

Base editing to generate R/H168^4.64 variant apelin receptor hESC and hESC-CMs

Base editing was used, as described previously^68,69, to generate hESCs and hESC-CMs carrying the R/H168^4.64 apelin receptor variant. Here, a cytosine base editor was used, in combination with custom guide RNAs (gRNAs), to induce a change in sequence from G-C to A-T, resulting in a substitution mutation of an arginine to a histidine at apelin receptor amino acid position 168.

To design the gRNA, a 1 kb genomic target sequence was input and a ranked list of potential 20 bp gRNAs, based on computationally predicted on- and off-target effects was generated (https://www.benchling.com/). All gRNAs identified directly preceded the PAM motif 5’-NGG, where N represents any base. This PAM sequence is specific for the SpCas9 ortholog and the PAM sequence can occur in the positive or negative strand. For cloning gRNA into the expression plasmid (pGL3-U6-sgRNA-PGK-puromycin plasmid), the vector was cut with BsaI type II restriction enzyme, creating non-compatible sticky ends for the ligation of gRNA. For plasmid preparation, vector was expanded by plating on LB Agar plates containing ampicillin (100 µg/mL), followed by liquid culture in LB broth containing ampicillin (100 µg/mL). Plasmids were isolated using the Plasmid Plus Midi Kit (QIAGEN), as per the manufacturer’s instructions, and NanoDropped to determine DNA concentration.

The gRNA oligonucleotides were then phosphorylated and annealed to form double stranded fragments. Reaction mixture containing 1 µL of top and bottom oligonucleotide (100 µM), 1 µL of T4 Ligase 10X Buffer, 1 µL T4 Polynucleotide Kinase and 6 µL nuclease free water were incubated at 37 °C for 30 min, followed by 95 °C for 5 min, and then temperature ramped down to 25 °C at 5 °C per minute. Annealed R/H168 oligonucleotides were then diluted 1:200 in nuclease free water and reactions set up for ligation of oligonucleotide into the gRNA expression vector (pGL3-U6-sgRNA-PGK-puromycin). Reaction mix consisted of 2 µL diluted annealed oligonucleotides, 100 ng vector, 2 µL Tango Buffer (10X), 1 µL 10 mM DTT, 1 µL 10 mM ATP, 1 µL BsaI FastDigest restriction enzyme, 1 µL T4 Ligase and made up to 20 µL with nuclease free water. Samples were incubated for 6 cycles of 37 °C for 5 min, followed by 21 °C for 5 min.

To remove residual linearised DNA fragments, 11 µL of ligation reaction was mixed with 1.5 µL 10X PlasmidSafe Buffer, 1.5 µL 10 mM ATP and 1 µl PlasmidSafe ATP-dependent DNase, and incubated at 37 °C for 30 min, followed by 70 °C for a further 30 min. Next, 2 µL of the treated ligation reaction was transformed into 25 µL of α-Select Gold Efficiency Chemically Competent Cells and plated on LB Agar + ampicillin (100 µg/mL) plates. Single colonies were selected for liquid starter culture and plasmid DNA isolated using the GenElute Plasmid Miniprep Kit and analysed by Sanger Sequencing (Source Bioscience) using the U6-FOR primer (GACTATCATATGCTTACCGT) to verify presence of gRNA. Positive clones were then regrown in 50 mL liquid culture overnight and plasmid isolated using the Plasmid Plus Midi Kit with an elution volume of 100 µL nuclease free water and concentration determined by NanoDrop.

For nucleofection, hESCs were treated with BE4max base editor vector plus gRNA vector, using the Amaxa 4D Nucleofector (Lonza) and P3 Primary Cell 4D-Nucleofector X Kit, as per manufacturer’s instructions. To prepare hESCs for nucleofection, cells were incubated in E8 complete media (DMEM/F12, insulin/transferrin/selenium at 20/11/13.4 mg/mL, 0.05% sodium bicarbonate, 7 µM L-Ascorbic acid 2-phosphate, 100 μ/mL penicillin-streptomycin), containing ROCKi (10 µM) for 1 h. For each nucleofection reaction, 1 × 10⁶ cells/reaction were added to 100 µL Nucleofector solution.

For the base editor reaction, the two vectors were added to Nucleofector solution at a ratio of 1:3 gRNA:base editor, with a total DNA content of 8 µg (2 µg gRNA vector, 6 µg base editor). A control reaction was also set up, with 2 µg of pmaxGFP vector added to Nucleofector solution, in order to visualise successful nucleofection. Each reaction was then transferred to a nucleofector cuvette and placed into the Amaxa 4D Nucleofector, and pulse program CA137 used. Cells were then allowed to recover for 5 min, before adding 500 µL E8 complete media supplemented with CloneR (diluted 1:10, Stem Cell Technologies) and leaving for a further 5 min. Using the provided plastic transfer pipette, the entire cuvette contents were added to 12 mL of E8 complete +CloneR and plated onto 10 cm dishes. Plates were transferred to 37 °C and incubated overnight to allow attachment. After 24 h, media was changed to E8 complete +puromycin (1 µg/mL), which was maintained for 48 h with daily refresh, before transferring to normal E8 complete. Cells were cultured until visible colonies were present. Resistant colonies were selected manually for expansion and allowed to grow clonally.

To genotype positive colonies, genotyping primers were designed for the APLNR nucleotide targeted for genetic manipulation. Sequences ~1 kb up- and down-stream of the residue of interest were input into Primer BLAST online tool and primer pairs generating products with length ~1.5 kb with minimal off target binding and amplification selected. Annealing temperature was determined for the selected primer pair using the New England BioLabs Tm Calculator and performing gradient PCR. Clonal cells were collected and pelleted by centrifugation. Genomic DNA extraction was performed on pelleted cells using the GenElute Mammalian Genomic DNA Miniprep Kit. DNA concentration was determined by NanoDrop and 100 ng of genomic DNA used for each genotyping reaction. The Q5 Hot Start High Fidelity DNA Polymerase kit was used, in combination with the designed sequencing primers. For each reaction, 5 µL was taken for gel electrophoresis to check for appearance of a band at the appropriate size. The remaining 20 µL of reactions producing bands at the predicted size were then PCR purified, NanoDropped and sent for Sanger Sequencing (Source Bioscience). Positive clones were then expanded in culture and differentiated to hESC-CMs for use in further assays.

Characterisation of R/H168^4.64 hESCs/hESC-CMs in vitro

To assess apelin receptor mRNA expression in R/H168^4.64 variant hESCs/hESC-CMs, RNA extraction was performed using the GenElute Total RNA Purification Kit. Briefly, cells were lysed in 350 µL of RNA lysis buffer and RNA was precipitated with an equal volume of 70% ethanol. Samples were transferred to GenElute Columns for RNA binding, washing, and elution. Samples were eluted in 30 μL Nuclease Free Water, with RNA concentration determined using a NanoDrop 1000 (ThermoFisher). cDNA was produced from 1 µg of RNA using the Promega Reverse Transcription System in a 20 µL reaction, as per the manufacturer’s recommendation. To 1 µg RNA, 1 µL of each of Random Primers and Oligo(dT)15 primers were added, made up to 11.9 µL with nuclease free water and incubated at 70 °C for 10 min. A mastermix containing 4 µL MgCl₂ (25 mM), 2 µL Reverse Transcription 10X Buffer, 2 µL dNTPs (10 mM), 0.5 µL Recombinant RNasin Ribonuclease Inhibitor and 0.6 µL AMV Reverse Transcriptase per sample was made and 9.1 µL added to each. Samples were then run on a thermocycler. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed for 45 cycles using the ABI 7500 Real-Time PCR System (Applied Biosystems) to analyse mRNA expression using 96-Well TaqMan Gene Expression Assays or SYBR Green Based Assays. Human 18S rRNA or GAPDH were used as house-keeping genes, owing to their stable expression level across cells used. Relative expression was normalised to housekeeping gene expression using the 2(-∆CT) or the 2(-∆∆CT) method⁷⁰. CT values over 36 were excluded as non-specific amplicon.

For protein expression, wild-type and R/H168^4.64 hESC-CMs were harvested and plated at ~25,000/well in CellCarrier-96 Ultra Plates (PerkinElmer). To detect apelin receptor using immunocytochemistry, cells were fixed with 4% formaldehyde for 3 min at room temperature. Cells were washed with PBS and blocked with 5% donkey sera in PBS for 1 h. Cells were then treated with primary apelin receptor antibody (Sigma-Aldrich, SAB2700205, 1:50) in PBS containing 3% donkey serum and 1% Tween for 24 h at 4 °C. Subsequently, cells were washed, before treatment with the secondary antibody (abcam, ab150073, 1:200) in PBS containing 3% donkey serum, 1% Tween, and 1:10,000 DAPI nuclear dye for 1 h at room temperature. For fluorescent ligand (apelin647) binding, cells were treated as described for CHO-K1 cells above. Cells were also imaged as described above using the Opera Phenix High Content Screening System (PerkinElmer).

Saturation radioligand binding, using [¹²⁵I]-apelin-13, was performed in hESC and hESC-CM membrane preparations as described above for CHO-K1 cells.

The differentiation efficiency of hESCs harbouring the R/H168^4.64 variant into hESC-CMs was determined using flow cytometry, as described previously⁵¹. Populations of cardiomyocytes were harvested and pelleted by centrifugation at 300 × g for 3 min. Pellets were resuspended in PBS supplemented with 0.1% BSA and 2 mM EDTA (PBE), with CD90 (Thy-1) Monoclonal Antibody directly conjugated to PE diluted at 1:50, for 1 h at 4 °C. Cells were then washed with PBE and resuspended in Fixation/Solubilisation solution (BD Cytofix/Cytoperm Fixation/Permeabilization Kit, Biosciences) for 20 min at 4 °C. Following incubation, cells were washed using 1X BD Perm/Wash Buffer (Biosciences) and then resuspended in 1X BD Perm/Wash Buffer containing directly conjugated Anti-Cardiac Troponin T-APC antibody diluted at 1:50 and incubated for 2 h at 4 °C. Cells were then washed in 1X BD Perm/Wash Buffer, resuspended in PBE and transferred to flow tubes. Samples were run on the LSRFortessa Cell Analyzer (BD Biosciences) and analysis performed using FlowJo v10.8.1 software (Fig. S13).

Apelin peptide secretion was determined via sandwich ELISA, using the apelin-12 (Human, Rat, Mouse, Bovine) EIA kit (Phoenix), as per the manufacturer’s instructions. In brief, supernatant samples were added to the secondary antibody coated wells. Primary antibody directed against apelin was added, along with biotinylated apelin peptide, and incubated for 2 h at room temperature with orbital shaking (300 rpm). Wells were washed and blot dried, then streptavidin-conjugated horseradish peroxidase (SA-HRP) was added and incubated for 1 h with orbital shaking. Wells were then washed and dried, and substrate added for 1 h with orbital shaking. The reaction was stopped by the addition of 2 N HCl and absorbance measured at 450 nm using a FLUOstar Omega Microplate Reader. Absorbance is inversely proportional to the concentration of peptide in the sample. Concentrations were determined via extrapolation of a standard curve of known concentrations of BSA.

To assess voltage sensing in hESC-CMs, cells were loaded with FluoVolt Membrane Potential voltage sensitive dye (ThermoFisher, F10488) as per the manufacturer’s instructions. FluoVolt was diluted 1:1000 in 1X Tyrode’s solution supplemented with glucose (10X Tyrode’s Solution diluted with mH₂O, 5 mM glucose, pH 7.4), along with PowerLoad solution diluted 1:100. For hESC-CMs in 6-well plates, media was aspirated and replaced with 2 mL/well diluted FluoVolt solution and incubated for 30 min at 37 °C. FluoVolt was then aspirated and replaced with 3 mL/well Tyrode’s solution. Cells were imaged using an Axio Observer A1 Inverted Phase Contrast Fluorescence Microscope with LabCam adaptor mounted, and videos recorded in slow motion using an iPhone 7. Cells were paced at 1 Hz, 1.5 Hz and 2 Hz using the C-Pace EM fitted with 6-well plate adaptor (IonOptix). Generated videos were loaded into a custom MATLAB (R2021a) code designed to extract values for time-to-peak and decay time in groups of cells selected as visibly contracting.

Pharmacological characterisation of peptide agonists

Affinities were determined as described above; see Radioligand Binding section. cAMP (cAMP Hunter™ eXpress AGTRL1 CHO-K1 GPCR Assay; 95-0147E2CP2M) and β-arrestin recruitment (PathHunter® eXpress AGTRL1 CHO-K1 β-Arrestin-1 GPCR Assay; 93-1050E2CP2M) assays were carried out in cells expressing the human apelin receptor according to the manufacturer’s protocol to obtain values of pD₂. Cardiovascular actions of NXE’065 (30,100 and 150 µg/kg) and NXE’515 (100 and 150 µg/kg) in anaesthetised rat were determined for CMF-019³⁰. Animal experiments were performed in accordance with guidelines from the local ethics committee (University of Cambridge) and the Home Office (UK) under the Scientific Procedures Act (1986).

Diagrams and schematics

Schema were created with BioRender.com. GPCR snake plots were generated using the GPCRdb server³⁷. Chemical structures were drawn using ChemDraw (RRID:SCR_016768).

Statistical analysis

Data are expressed as mean ± standard deviation (SD). Raw data were handled using Microsoft Excel. Graphical presentation and statistical tests were performed using GraphPad Prism version 6.07. For saturation radioligand binding experiments, data were analysed using the EBDA and LIGAND components of the KELL (Kinetic, EBDA, Ligand, Lowry) software package (Biosoft). For competition binding experiments, Ki values were determined using Cheng-Prusoff methodology. For in vitro GPCR assays, EC₅₀, pD₂, and E_max values were calculated using GraphPad Prism version 6.07. Statistical tests are indicated in figure legends where used, as are experimental n numbers. A p-value of <0.05 was determined as significant.

Reporting summary

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

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s41467-024-55381-w