Physical interaction with Ephrin B1 promotes CXCR4 intracellular localization and oncogenic potential Authors: Alessandro Rabbito, Omolade Otun, Amos Fumagalli, Ana Alonso Bartolomé, Sonya Galant, Martial Séveno, Manuel Counson, Thierry Durroux, Cherine Bechara, Martine J Smit, Sébastien Granier, Martyna Szpakowska, Andy Chevigné, Séverine Chaumont-Dubel, Philippe Marin Published: 2026 ABSTRACT Chemokine receptor 4 (CXCR4) is a member of the chemokine receptor family, exclusively activated by the chemokine CXCL12. While CXCR4 regulates numerous physiological processes associated with cell migration and embryogenesis, its overexpression has been involved in various cancer types. Studies suggest that intracellular CXCR4 rather than CXCL12-induced signaling at the plasma membrane contributes to its pro-tumorigenic functions. Given the role of GPCR-interacting proteins in their trafficking and subcellular localization, we characterized the CXCR4 interactome using an affinity purification coupled to mass spectrometry (AP-MS) strategy. The most abundant protein identified in the CXCR4 interactome is Ephrin B1, a member of the Ephrin protein family that shares several functions with CXCR4, such as the regulation of cell migration and proliferation. Further studies showed that the interaction between CXCR4 and Ephrin B1 is direct and enhanced upon CXCR4 activation by CXCL12. They also indicated that Ephrin B1 prevents CXCR4 N-glycosylation, decreases CXCR4 cell surface expression, and consistently inhibits CXCL12-induced CXCR4 coupling to Gαi1−3 and recruitment of β-arrestins 1 and 2. Conversely, Ephrin B1 signals to Erk1/2 through CXCR4 activation and mediates the decrease in Death Receptor 5 expression elicited by intracellular CXCR4. Collectively, these findings identify Ephrin B1 as a potential regulator of CXCR4-driven tumorigenic signaling.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s00018-026-06281-3. Introduction The C-X-C chemokine receptor type 4 (CXCR4) is a conventional chemokine receptor that, unlike many chemokine receptors, binds to a single chemokine, CXCL12, also referred to as SDF-1. CXCR4 is expressed in numerous cell types, including endothelial cells, lymphocytes, fibroblasts, hematopoietic stem cells, neurons, and glial cells where it plays a pivotal role in migration, proliferation and differentiation [1]. CXCR4 is also involved in several pathological processes, including the Warts, Hypogammaglobulinemia, Immunodeficiency, and Myelokathexis (WHIM) syndrome and Human Immunodeficiency Virus-1 infection [2, 3]. It is overexpressed in numerous cancer types and is involved in cancer progression and metastasis [4]. CXCR4 is canonically coupled to Gαi proteins and inhibits adenylyl cyclase. It also activates the Erk1/2 pathway through several mechanisms involving either its phosphorylation by GRK3 and GRK6 and the recruitment of β-arrestins, or the activation of PI3Kγ heterodimers p110γ-p101 through Gβγ proteins [5, 6]. Despite the unequivocal role of CXCR4-operated signaling, such as PI3K-Erk signaling, in cancer growth and metastasis, the limited therapeutic efficacy of CXCR4 antagonists, including the FDA-approved compound AMD3100, suggests that some tumorigenic effects under the control of the receptor could be independent of the CXCL12-CXCR4 signaling axis [7]. This conundrum can in part be explained by the recent demonstration of the existence of agonist-independent, CXCR4 oligomer-mediated signaling, which promotes cell survival, migration and tumour growth in lymphoid neoplasms [8]. Another study has shown that intracellular CXCR4 rather than CXCL12-induced CXCR4 signaling promotes cancer cell survival through the downregulation of the Death Receptor 5 (DR5), thereby rendering cancer cells resistant to chemotherapeutic drugs like paclitaxel [9]. These findings suggest that future therapies targeting CXCR4 should not only consider receptor-operated signaling at the plasma membrane, but also mechanisms underlying the receptor localization in intracellular compartments. There is accumulating evidence indicating that the trafficking and targeting of G protein-coupled receptors (GPCRs) to specific cellular compartments depends on their association with protein partners. For instance, while the association of CXCR4 with Filamin A stabilizes the receptor at the plasma membrane by blocking its endocytosis [10], its interaction with Reticulon 3 (RTN3) promotes its translocation to the cytoplasm [11, 12]. These findings advocate for the in-depth characterization of the CXCR4 interactome to identify novel mechanisms underlying its tumorigenic effects, an issue we addressed thanks to a proteomic strategy combining affinity purification of receptor-interacting proteins and their identification by mass spectrometry. This interactomic screen identified Ephrin B1, a member of the single transmembrane domain Ephrin-B family overexpressed in numerous cancer types and involved in tumor progression and metastasis as well as resistance to chemotherapies [13]. A previous study showed that Ephrin B1 modulates G protein signaling induced by CXCR4 activation via the recruitment of the regulator of G protein signaling 3 (RGS3) protein through its C-terminal PDZ domain, thus establishing functional interactions between CXCR4 and Ephrin B1 [14]. In light of these findings, we further explored the impact of Ephrin B1 on CXCR4 post-translational modifications, cellular localization and signal transduction. We show that Ephrin B1 retains CXCR4 in intracellular compartments and contributes to CXCR4-dependent decrease in DR5 expression, suggesting that it might be a key player contributing to CXCR4’s oncogenic potential. Materials and methods Materials Plasmids and shRNAs The HA-CXCR4-pCDNA3.1 construct was obtained from cDNA.org (Bloomsburg University), and the FLAG-Ephrin B1-pcDNA3.1 plasmid was provided by the Montpellier Genomic Collections (MGC) platform (Biocampus, Montpellier). Both constructs encode N-terminally tagged CXCR4 and Ephrin B1, respectively. The constructs encoding Ephrin B1, B2 and B3 carrying N-terminal Myc tag (Myc-Ephrin B1-pcDNA3.1, Myc-Ephrin B2-pcDNA3.1 and Myc-Ephrin B3-pcDNA3.1) were obtained from the MGC platform. The Ephrin B1ΔCter construct (purchased from GeneCust) was designed by removing 84 amino acids from the C-terminus. Based on the predicted structure, this deletion eliminates the C-terminal domain of the protein, resulting in a truncated version with an approximate molecular weight of 28.5 kDa. The pcDNA3.1 plasmids encoding Venus-Tagged G protein subunit gamma 2 (Venus-γ2), FLAG-tagged G-protein subunit beta 2 (FLAG-β2), Gαi1-RLuc and Gαi3-RLuc were provided by Dr. D. Maurel (IGF, Montpellier, France). ACKR3-NeonGreen and CXCR4-NeonGreen were previously described [15]. HiBiT-CXCR4 plasmid was described in a previous study [16]. Ephrin B1-NeonGreen and Ephrin B1-Nluc were cloned as previously described [15]. Ephrin B2-Nluc and Ephrin B3-Nluc were cloned from Myc-Ephrin B2 and Myc-Ephrin B3 constructs, respectively. LYN-NeonGreen, ACKR3-Nluc, CXCR4-Nluc, LgBiT-β-arrestin1/2, and CXCR4-smBiT in pcDNA3.1 were previously described [17, 18]. ShRNAs were obtained from Sigma-Aldrich MISSION shRNA library clones. Specific EphrinB1 shRNA was selected amongst shRNAs targeting the coding sequence (ref NM_004429, target sequence AGCACCATGATTACTACATTA). For the control condition, we used the MISSION Non-Target shRNA Control sequence. This shRNA sequence does not target any known genes from any species. Both control and specific shRNAs were obtained in the pLKO.1-puro-CMV-tGFP backbone. Reagents CXCL12, purchased from R&D Systems (Ref 350-NS-CF), was dissolved in sterile PBS to a concentration of 10 µM, aliquoted, and stored at -80 °C. AMD3100 purchased from Tocris (Ref 3299) was dissolved in DMSO to a final concentration of 20 mM, aliquoted, and stored at -20 °C. Gallein was purchased from Tocris (Ref 3090) and dissolved in DMSO to a final concentration of 75 mM, aliquoted, and stored at -20 °C. Antibodies ProteinHost SpeciesDilution & ApplicationSupplierREF Primary antibodiesCXCR4Rabbit1/1,000 WBAbcamAB124824Ephrin B1Rabbit1/2,000 WBBiorbytorb395013HARat1/1,000 WBRoche11,867,423,001MycRabbit1/1,000 WBCell Signaling Technology2278CXCR4 pS324/325Rabbit1/1,000 WB7-TM7TM0071ACXCR4 pS330Rabbit1/1,000 WB7-TM7TM0071CCXCR4 pS346/347Rabbit1/1,000 WB7-TM7TM0071BDR5Rabbit1/1,000 WBCell Signaling Technology8074GAPDHRabbit1/5,000 WBCell Signaling Technology51,332p44/42 MAPKRabbit1/1,000 WBCell Signaling Technology9102p-p44/42 MAPKRabbit1/1,000 WBCell Signaling Technology9101LAMP1Rabbit1/250 ICCCell Signaling Technology9091CalnexinRabbit1/250 ICCTransduction laboratoriesC45520GM130Mouse1/250 ICCBD Transduction610,823LabelTarget SpeciesDilution & ApplicationSupplierREFSecondary antibodiesAlexa Fluor® 633Rabbit1/500 ICCInvitrogen210,071Alexa Fluor® 633Mouse1/500 ICCInvitrogen210,052HRPRat1/5,000 WBJackson ImmunoResearch112-035-003HRPRabbit1/5,000 WBCell Signaling Technology7074 ProteinHost SpeciesDilution & ApplicationSupplierREF Primary antibodiesCXCR4Rabbit1/1,000 WBAbcamAB124824Ephrin B1Rabbit1/2,000 WBBiorbytorb395013HARat1/1,000 WBRoche11,867,423,001MycRabbit1/1,000 WBCell Signaling Technology2278CXCR4 pS324/325Rabbit1/1,000 WB7-TM7TM0071ACXCR4 pS330Rabbit1/1,000 WB7-TM7TM0071CCXCR4 pS346/347Rabbit1/1,000 WB7-TM7TM0071BDR5Rabbit1/1,000 WBCell Signaling Technology8074GAPDHRabbit1/5,000 WBCell Signaling Technology51,332p44/42 MAPKRabbit1/1,000 WBCell Signaling Technology9102p-p44/42 MAPKRabbit1/1,000 WBCell Signaling Technology9101LAMP1Rabbit1/250 ICCCell Signaling Technology9091CalnexinRabbit1/250 ICCTransduction laboratoriesC45520GM130Mouse1/250 ICCBD Transduction610,823LabelTarget SpeciesDilution & ApplicationSupplierREFSecondary antibodiesAlexa Fluor® 633Rabbit1/500 ICCInvitrogen210,071Alexa Fluor® 633Mouse1/500 ICCInvitrogen210,052HRPRat1/5,000 WBJackson ImmunoResearch112-035-003HRPRabbit1/5,000 WBCell Signaling Technology7074 Cell cultures and transfection HEK293T cells, purchased from ATCC (Anassas, VI, ATCC, CRL-1573), were grown in Dulbecco’s Modified Eagle’s Medium (DMEM, Thermo Fisher Scientific, Ref 419960) supplemented with 10% heat-inactivated fetal bovine serum (Thermo Fisher Scientific, Ref 10099-133), 1% penicillin/streptomycin, and maintained in humidified atmosphere containing 5% CO2 at 37 °C. Cells were passed twice a week and used between passages 10 and 25. They were transfected using polyethylenimine (jetPEI®, Polyplus, Ref 101000053) and used 24/48 h after transfection. MCF7 cells, provided by Dr. Stéphan Jalaguier and Vincent Cavaillès (Institut de Recherche en Cancérologie de Montpellier), were cultured in /Dulbecco’s modified Eagle’s medium/F12 supplemented with 10% heat-inactivated bovine serum (Thermo Fisher Scientific, Ref 10099-133), 1% penicillin/streptomycin, and maintained in a humidified atmosphere containing 5% CO2 at 37 °C. Cells were passed twice a week and used between passages 10 and 25. They were transfected using Lipofectamine™ 2000 (Thermo Fisher Scientific, Ref 11668019) and used 48 h after transfection. Co-immunoprecipitation HEK293T cells transfected with empty vector or vectors expressing HA- or Flag-tagged proteins were lysed in ice-cold lysis buffer containing 1% n-Dodecyl-β-D-Maltopyranoside (DDM, Anatrace, Ref D310), 20 mM Tris-HCl (pH 7.5), 100 mM NaCl, 2.5 mM CaCl2, phosphatase inhibitors (NaF, 10 mM; Na+-vanadate, 2 mM; Na+-pyrophosphate, 1 mM; and β-glycerophosphate, 50 mM) and the cOmplete Protease Inhibitor Cocktail (Merck, Ref 11836145001). Samples were maintained under gentle agitation for 1 h, at 4 °C and centrifuged for 15 min at 15,000 × g to eliminate insoluble material. Soluble proteins were quantified by bicinchoninic acid assay (BCA, Merck, Refs B29643 and C2284) and equal protein amounts (5 or 0.5 mg for IP followed by mass spectrometry or Western blotting, respectively) were incubated with agarose-conjugated anti-HA antibody (Merck, Ref A2095) or with anti-FLAG M2 affinity gel (Sigma Aldrich, Ref A2220) overnight at 4 °C. Samples were then washed twice with an ice-cold solution of 0.5 M of NaCl and phosphatase inhibitors and three times with 0.25 M NaCl and phosphatase inhibitors. Immunoprecipitated proteins were then eluted in Laemmli sample buffer. Protein identification by mass spectrometry Immunoprecipitated proteins from cells transfected with an empty vector (Mock) or a vector HA-tagged CXCR4 were separated by SDS-PAGE and stained with Protein Staining Solution (Euromedex, Ref 10–0911). Gel lanes were cut into seven gel pieces and destained with 50 mM TriEthylAmmonium BiCarbonate (TEABC, Merck, Ref T7408) followed by three washes in 100% acetonitrile. After reduction (with 10 mM dithiothreitol in 50 mM TEABC at 60 °C for 30 min) and alkylation (with 55 mM iodoacetamide TEABC at room temperature for 60 min), proteins were digested in-gel using trypsin (500 ng/band, Gold, Promega, Ref V5280). Digest products were dehydrated in a vacuum centrifuge and reduced to 3 µL. The generated peptides were analyzed online by nano-flow liquid chromatography coupled to tandem-mass spectrometry (nanoLC-MS/MS) using an Orbitrap Elite mass spectrometer (Thermo Fisher Scientific, Waltham USA) coupled to an Ultimate 3000 HPLC (Thermo Fisher Scientific). Desalting and pre-concentration of samples were performed on-line on a Pepmap® pre-column (0.3 mm × 10 mm, Dionex). A gradient consisting of 0–40% B for 60 min and 80% B for 15 min (A = 0.1% formic acid, 2% acetonitrile in water; B = 0.1% formic acid in acetonitrile) at 300 nL/min was used to elute peptides from the capillary reverse-phase column (0.075 mm × 150 mm, Acclaim Pepmap 100® C18, Thermo Fisher Scientific). Eluted peptides were electrosprayed online at a voltage of 1.8 kV into the Orbitrap Elite mass spectrometer. A cycle of one full-scan mass spectrum (MS1, 400–2,000 m/z) at a resolution of 120,000 (at 400 m/z), followed by 20 data-dependent tandem-mass (MS2) spectra was repeated continuously throughout the nanoLC separation. All MS2 spectra were recorded using normalized collision energy (33%, activation Q 0.25 and activation time 10 ms) with an isolation window of 2 m/z. Data were acquired using the Xcalibur software (v 2.2). For all full scan measurements with the Orbitrap detector a lock-mass ion from ambient air (m/z 445.120024) was used as an internal calibrant. Mass spectra were processed using the MaxQuant software package (v 1.5.5.1) and MS2 using the Andromeda search engine [19] against the UniProtKB Reference proteome UP000005640 database for Homo sapiens (release 2017_10) and the contaminant database in MaxQuant. The following parameters were used: enzyme specificity set as Trypsin/P with a maximum of two missed cleavages, Oxidation (M) and Phosphorylation (STY) set as variable modifications and carbamidomethyl (C) as fixed modification, and a mass tolerance of 0.5 Da for-fragment ions. The maximum false peptide and protein discovery rate was specified as 0.01. Seven amino acids were required as minimum peptide length. When precursor peptides were present in MS1 spectra but not selected for fragmentation and identification by MS2 in given runs, peptide identification were based on accurate mass and retention times across LC-MS runs using the matching between runs tool in MaxQuant. Only proteins identified in all three biological replicates in at least one group were considered for further analysis. Relative protein quantification in IP from CXCR4-expressing cells and Mock cells was performed using the label-free quantification (LFQ) algorithm “[https://maxquant.net/maxquant/]”. For statistical analysis, missing values were defined using the imputation tool of the Perseus software (v. 1.5.6.072) “[https://maxquant.net/perseus/]”. Western blotting Proteins were separated by SDS-PAGE using 10% polyacrylamide gels and transferred to nitrocellulose membranes (Bio-Rad, Ref 1704271). Membranes were incubated in blocking buffer (Tris-HCl, 50 mM, pH 7.5; NaCl, 200 mM; Tween-20, 0.1% and skimmed dried milk, 5%) for 1 h at room temperature and overnight with primary antibodies in incubating buffer (Tris-HCl, 50 mM, pH 7.5; NaCl, 200 mM; Tween-20, 0.1% and Bovine Serum Albumin (Merck, Ref A2153), 3%) at 4 °C. Then membranes were immunoblotted with either or anti-rat (Jackson ImmunoResearch, Ref 112-035-003) or anti-rabbit (Cell Signaling Technology, Ref 7074) horseradish peroxidase (HRP)-conjugated secondary antibodies (1/5,000) in blocking buffer for 1 h at room temperature. Immunoreactivity was detected with an enhanced chemiluminescence method (Western lightning® Plus-ECL, Perkin Elmer, Ref NEL103E001EA) on a ChemiDoc™ Touch Imaging System (Bio-Rad). Quantification was performed using the Image Lab software (Bio-Rad). Production and purification of CXCR4 and Ephrin B1 For production in insect cells, the full-length genes encoding human CXCR4 or Ephrin B1 were subcloned into pFastBac1 to enable infection of Sf9 insect cells. The CXCR4 construct bore a hemagglutinin signal peptide followed by a Flag-tag preceding the receptor sequence. Mutations of glycosylated residues (N11Q, S18A and N33Q) were introduced to avoid N- or O-glycosylation for improved sample heterogeneity. The Ephrin B1 sequence was followed by a C-terminal 3 C human rhinovirus 3 C (HRV3C) protease cleavage site and a twin-Strep-tag (Strep-Ephrin B1). Flag-CXCR4 and Strep-Ephrin B1 were expressed in Sf9 cells using the pFastBac baculovirus system (Thermo Fisher Scientific). Cells were grown in suspension in EX-CELL 420 medium (Sigma-Aldrich) and infected at a density of 4 × 106 cells/ml with the recombinant baculovirus. Flasks were shaken for 48 h at 28 °C, subsequently harvested by centrifugation (3,000 × g, 20 min) and stored at -80 °C until usage. Cell pellets were first thawed and lysed by osmotic shock in a lysis buffer containing 10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 2 mg/ml iodoacetamide, and protease inhibitors: 50 µg/ml Leupeptin (Euromedex), 0.1 mg/ml Bensamidine (Sigma-Aldrich) and 0.1 mg/ml Phenylmethylsulfonyl fluoride (PMSF; Euromedex). Lysates were centrifuged (38,400 × g, 10 min), and the resulting pellet was solubilized in buffer containing 50 mM Tris (pH 7.5), 150 mM NaCl, 2 mg/ml iodoacetamide, 0.5% (w/v) dodecyl maltoside (DDM, Anatrace), Cholesteryl hemisuccinate (CHS, 0.1% w/v) and protease inhibitors (50 µg/ml Leupeptin, 0.1 mg/ml Bensamidine and 0.1 mg/ml PMSF) using a Dounce homogenizer. Resulting homogenates were then stirred for 1 h, at 4 °C and centrifuged (38,400 × g, 30 min). CXCR4 supernatant was loaded onto M2 anti-Flag affinity resin (Sigma-Aldrich Ref A2220) using gravity flow. Ephrin B1 supernatant was supplemented with BioLock (0.75 ml/L) and incubated with Streptactin resin (IBA-Lifesciences, Ref 2-1206-025) for 1 h. Resins were each washed with 10 column volumes (CV) of wash buffer containing 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.1% (w/v) DDM, and 0.02% (w/v) CHS. The resin was finally washed with a last wash buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, DDM 0.02% (w/v), CHS 0.002% (w/v) and eluted in last wash buffer supplemented with 0.4 mg/ml Flag peptide for elution of CXCR4 or 2.5 mM desthiobiotin for elution of Ephrin B1. Each eluate was concentrated using a 50 kDa molecular weight cutoff (MWCO) concentrator (Millipore). CXCR4 was then purified by size exclusion chromatography (SEC) using a Superdex 200 Increase (10/300 GL column) connected to an ÄKTA purifier system (GE Healthcare) and eluted in SEC buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.02% (w/v) DDM, 0.002% (w/v) CHS). Buffers used for CXCR4 purification were supplemented with the CXCR4 antagonist IT1t (1 µM in the lysis and solubilization buffers, 0.1 µM in the wash and elution buffers, Bio-Techne) to stabilize the receptor. Co-immunoprecipitation of purified CXCR4 and Ephrin B1 Complexes of the two purified proteins were formed by mixing them at a 2:1 ratio and incubating the samples at 4 °C overnight. The samples were then loaded onto M2 anti-Flag affinity resin, washed with 6 CV of SEC buffer, and eluted with SEC buffer supplemented with 0.4 mg/ml Flag peptide. The resulting eluate was loaded onto a column containing Streptactin resin, washed with 6 CV of SEC buffer, and eluted with SEC buffer supplemented with 2.5 mM Desthiobiotin. BRET assays assessing interactions between CXCR4 or ACKR3 and Ephrin Bs, recruitment of G proteins to CXCR4 and G protein dissociation HEK293T cells were co-transfected with the indicated constructs and seeded (25,000 cells/well) in white polyornithine-coated 96-well plates (SPL Life Sciences). Twenty-four hours after transfection, cells were washed with PBS. Coelenterazine H (Nanolight Technology, Ref 50909-86-9) was added at a final concentration of 5 µM for 10 min at 37 °C. Cells were then exposed to vehicle or CXCL12, and luminescence was measured using a Mithras LB 940 plate reader (Berthold Biotechnologies) that allows the sequential integration of light signals detected with two filter settings (Rluc/NLuc filter, 485 ± 20 nm; and YFP filter, 530 ± 25 nm). Data were collected using the MicroWin2000 software (Berthold Biotechnologies). Nano-BRET assay assessing plasma membrane localization of CXCR4 HEK293T cells were co-transfected with vectors encoding human CXCR4 C-terminally fused with Nano-luciferase (Nluc) and the first 11 residues of the human Lyn-kinase sequence, fused with NeonGreen, respectively, in the absence or presence of a vector encoding Ephrin B1. Twenty-four hours after transfection, cells were harvested, incubated for 20 min at 37 °C with coelenterazine H in Opti-MEM, and distributed into white 96-well plates (150,000 cells/well). The luminescence generated was measured with a GloMax Discover plate reader (Promega), equipped with 450/8 filter for donor luminescence emission and 530 LP filter for acceptor fluorescence emission [20]. CXCR4 cellular distribution monitoring using Nano-luciferase complementation (HiBiT) Receptor cellular distribution in basal conditions and upon ligand stimulation was monitored by using a Nano-luciferase complementation assay based on the NanoGlo HiBiT extracellular and Nano-Glo HiBiT lytic detection systems (Promega, Ref N2420 and N3030). HEK293T cells were co-transfected with pHiBiT vector encoding CXCR4 N-terminally fused to HiBiT and a vector encoding Ephrin B1. Forty-eight hours after transfection, cells were distributed in white 96-well plates (50,000 cells per well) and stimulated with CXCL12 (10 or 50 nM) for 30 min at 37 °C. After the addition of Nano-Glo® HiBiT reagent, containing soluble LgBiT protein, in ratio of 1:100 of the final volume, luminescence was recorded for 30 min with a GloMax Discover plate reader (Promega). In unstimulated conditions, surface and total receptor expression were determined using the Nano-Glo HiBiT extracellular detection (Promega) and Nano-Glo HiBiT lytic detection systems (Promega), respectively. β-arrestin recruitment to CXCR4 based on Nano-luciferase complementation (NanoBiT) β-arrestin-1 and β-arrestin-2 recruitment to CXCR4, was monitored using a Nano-luciferase complementation-based assay (NanoBiT, Promega). HEK293T cells were co-transfected with vectors encoding human β-arrestins N-terminally fused with LgBiT, CXCR4 C-terminally fused with SmBiT in the absence or presence of a vector encoding Ephrin B1. Twenty-four hours after transfection, cells were harvested, incubated for 20 min at 37 °C with coelenterazine H in Opti-MEM, and distributed into white 96-well plates (150,000 cells/well). CXCL12 was then added at the indicated concentration, and the luminescence generated upon Nano-luciferase complementation was measured with a GloMax Discover plate reader (Promega). Immunocytochemistry and confocal microscopy HEK293T cells grown on glass coverslips were fixed with a 4% solution of PFA in PBS for 10 min. Excess of PFA was quenched by washing cells in a 0.1 M solution of glycine in PBS for 10 min. Cells were permeabilized with a PBS solution containing 0.5% heat-inactivated bovine serum and 0.1% Triton X-100 for 15 min. Cells were washed three times and incubated with primary antibodies (anti-LAMP1 (rabbit or mouse) or anti-GM130 (mouse), or anti-Calnexin (rabbit)) in PBS containing 0.5% heat-inactivated bovine serum and 0.1% Triton X-100. After three washes in PBS, cells were incubated for 1 h at room temperature in PBS containing 0.5% bovine serum, 0.1% Triton X-100, the appropriate secondary antibody (Alexa Fluor® 633-conjugated anti-mouse antibody (1/250, Invitrogen, Ref 210052), Alexa Fluor® 633-conjugated anti-rabbit antibody (1/250, Invitrogen Ref 210071)) and DAPI (1 µg/ml). Cells were then washed three times with PBS and coverslips were mounted on Superfrost ultra plus glass slides using fluorescent mounting medium. Pictures were acquired with a LSM980 confocal microscope (Zeiss) equipped with a 40X oil-immersed lens, with a 2791 × 2791 resolution, and 1 μm between focus points. Statistics Statistical analysis of the CXCR4 interactome was performed using the Perseus software (v 1.5.6.072). Proteins were considered statistically significant using a t-test by setting the randomization number at 250, the False Discovery Rate at 0.01, and the S0 at 0.1. All other statistical analyses were performed using Prism (v.8.0, GraphPad Software Inc), and the statistical tests used are indicated in each legend. Concentration-response curves were fitted by the log (agonist) vs. response (four parameters) non-linear regression using Prism. Saturation BRET experiments were analyzed using Prism. The one-site specific binding model was used to obtain the curve fit. All data are presented as means ± SEM. Significance levels were defined as p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****). Results Characterization of the CXCR4 interactome reveals a physical interaction between CXCR4 and Ephrin B1 We analyzed the CXCR4 interactome in human embryonic kidney (HEK-293T) cells that endogenously express CXCR4 and where CXCR4-dependent signaling has been extensively investigated [5, 6, 21]. Due to the lack of a CXCR4 antibody providing receptor immunoprecipitation yields compatible with mass spectrometry analysis, we expressed hemagglutinin (HA)-tagged CXCR4 in the cells and CXCR4-interacting proteins were immunoprecipitated using an anti-HA monoclonal antibody immobilized onto agarose beads. Control immunoprecipitations were performed using cells transfected with an empty plasmid (Mock condition). Systematic analysis by LC-MS/MS of proteins in immunoprecipitates from both conditions in three biological replicates identified 1,203 proteins. Label-free quantification of their relative abundance in both conditions showed that 19 of them exhibited significant enrichment in immunoprecipitates from CXCR4-expressing cells, compared with immunoprecipitates from control cells (Fig. 1A and Table S1). As expected, CXCR4 (bait protein) was the most enriched one (log2LFQCXCR4/LFQMock = 8.13, -log10 p-value = 4.82), while the second ranked protein was Ephrin B1 (log2LFQCXCR4/LFQMock = 7.29, -log10 p-value = 4.96, Fig. 1A and Table S1). Ephrin B1 is a member of the Ephrin B family comprising three proteins (Ephrin B1-3) that act as ligands of the tyrosine kinase Eph receptors involved in numerous biological processes, including axon guidance and cell migration during neurodevelopment, and the proliferation of various cancer cell types [13]. Notably, Ephrin B1 was the only member of the Ephrin ligand family identified in the CXCR4 interactome. Fig. 1Physical interaction of Ephrin B1 with CXCR4. A Volcano plot representing proteins identified by nano-LC-MS/MS in immunoprecipitations from HEK293T cells transfected with HA-CXCR4 or empty plasmid. B Representative Western blots of HA and Flag immunoprecipitations performed from HEK293T cells expressing or not HA-CXCR4 or FLAG-Ephrin B1 or both proteins. C Normalized BRET values measured in HEK293T cells co-expressing CXCR4 or ACKR3 C-terminally fused to mNeonGreen (NG) and Ephrin B1 C-terminally fused to Nano-luciferase (Nluc), treated or not with CXCL12 (50 nM). BRET signal variations were fitted with the one-site total binding equation and background constraint to a constant value of 0 using Prism. The data are means ± SEM of values obtained in three independent experiments performed on different sets of cultured cells. D Representative Western blot of HA-immunoprecipitations performed from HEK293T cells expressing or not HA-CXCR4, or FLAG-Ephrin B1 or FLAG-Ephrin B1 deleted of its C-terminal domain (Ephrin B1 ΔC-ter) alone or in combination. In B and D, the illustrated blots are representative of three independent experiments performed on different sets of cultured cells. E Left: Coomassie blue staining of purified FLAG-CXCR4 and STREP-Ephrin B1 produced in the SF9 cell line. Right: Western blots of sequential Flag/Strep chromatography purification. The illustrated data are representative of three independent experiments. FT: flow-through Physical interaction of Ephrin B1 with CXCR4. A Volcano plot representing proteins identified by nano-LC-MS/MS in immunoprecipitations from HEK293T cells transfected with HA-CXCR4 or empty plasmid. B Representative Western blots of HA and Flag immunoprecipitations performed from HEK293T cells expressing or not HA-CXCR4 or FLAG-Ephrin B1 or both proteins. C Normalized BRET values measured in HEK293T cells co-expressing CXCR4 or ACKR3 C-terminally fused to mNeonGreen (NG) and Ephrin B1 C-terminally fused to Nano-luciferase (Nluc), treated or not with CXCL12 (50 nM). BRET signal variations were fitted with the one-site total binding equation and background constraint to a constant value of 0 using Prism. The data are means ± SEM of values obtained in three independent experiments performed on different sets of cultured cells. D Representative Western blot of HA-immunoprecipitations performed from HEK293T cells expressing or not HA-CXCR4, or FLAG-Ephrin B1 or FLAG-Ephrin B1 deleted of its C-terminal domain (Ephrin B1 ΔC-ter) alone or in combination. In B and D, the illustrated blots are representative of three independent experiments performed on different sets of cultured cells. E Left: Coomassie blue staining of purified FLAG-CXCR4 and STREP-Ephrin B1 produced in the SF9 cell line. Right: Western blots of sequential Flag/Strep chromatography purification. The illustrated data are representative of three independent experiments. FT: flow-through Consistent with the results of our interactomic screen, co-immunoprecipitation followed by Western blotting from cells co-expressing CXCR4 and Ephrin B1 confirmed that Ephrin B1 co-immunoprecipitates with CXCR4 and vice versa (Fig. 1B and Fig. S1A). In contrast, in cells co-expressing CXCR4 and Ephrin B2 or Ephrin B3, negligible amounts of Ephrin B2 and Ephrin B3 co-immunoprecipitated with the receptor (Fig. S1A). We also analyzed the specificity of CXCR4-Ephrin B1 interaction in living HEK-293T cells using bioluminescence resonance energy transfer (BRET). Under conditions of constant Ephrin B1-Nluc expression, the BRET signal increased hyperbolically as a function of the CXCR4-NeonGreen (NG) expression level (Fig. 1C). Further, the BRET signal was significantly increased upon CXCR4 activation by CXCL12 (Bmax = 802 ± 54.3 vs. 974 ± 29 mBRET in the absence and presence of CXCL12, respectively, p = 0.011, one-way ANOVA followed by Bonferroni’s multiple comparisons test), indicating that the interaction with Ephrin B1 is promoted by agonist stimulation of CXCR4. Consistent with the results of co-immunoprecipitation experiments, the BRET signal was much lower in cells co-expressing CXCR4-NeonGreen and Ephrin B2-Nluc or Ephrin B3-Nluc and was not increased by the exposure of cells to CXCL12 (Fig. S1B). Previous interactomics studies did not identify Ephrin Bs as protein partners of ACKR3 [22], an atypical chemokine receptor known to form heteromers and to be functionally linked with CXCR4 via their shared ligand CXCL12 [23–25]. Consistent with these findings, the BRET signal was markedly lower in cells co-expressing Ephrin B1-3-Nluc and increasing amounts of ACKR3-NeonGreen than in cells co-expressing Ephrin B1 and CXCR4 (Fig. 1C and Fig. S1B). Furthermore, no CXCL12-induced effect was observed. Collectively, these observations indicate a close and specific interaction between CXCR4 and Ephrin B1. Considering the strong enrichment of Ephrin B1 in the CXCR4 interactome, their role in common cellular and pathological processes [26–28] and the existence of functional interactions between both proteins [14], we further characterized the CXCR4-Ephrin B1 interaction. The C-terminal domain of GPCRs has been identified as a major site involved in their interaction with their protein partners [29]. Nevertheless, the co-immunoprecipitation of Ephrin B1 with CXCR4 was not affected by the deletion of CXCR4’s 15 C-terminal residues (major mutation responsible for the WHIM syndrome, CXCR4-WHIM [30]) nor of its entire C-terminal domain (CXCR4ΔCter, Fig. S2), indicating that the CXCR4 C-terminal domain is not involved in the recruitment of Ephrin B1. AlphaFold predictions suggest that, beyond the structured ephrin receptor-binding domain, Ephrin B1 contains extensive regions of intrinsic disorder, including the C-terminal cytosolic domain (Fig. S3). This intrinsically disordered C-terminal domain seems to be essential to the formation of the CXCR4-Ephrin B1 complex, as its deletion (Ephrin B1ΔCter) abolished the co-immunoprecipitation of Ephrin B1 with CXCR4 (Fig. 1D). Note that Ephrin B1ΔCter exhibits a similar subcellular distribution as the full-length protein in HEK-293T cells (Fig. S4). To establish whether the interaction between both proteins is direct, purified recombinant Flag-CXCR4 and Strep-Ephrin B1 were obtained from Sf9 cells, mixed at 2:1 protein ratio and subjected to sequential affinity purification on M2 anti-Flag affinity resin and, after elution of the retained material, on Streptactin resin. We found that Strep-Ephrin B1 was co-purified with Flag-CXCR4 on Flag resin while Flag-CXCR4 was co-copurified with Strep-Ephrin B1 on Streptactin resin, indicative of a direct physical interaction between both proteins (Fig. 1E). Impact of Ephrin B1 on CXCR4 glycosylation and phosphorylation Co-expression of Ephrin B1 with CXCR4 modified the pattern of CXCR4 migration in SDS-PAGE, leading to a single immunoreactive band at around 40 kDa, instead of two bands (the 40 kDa band and another band of higher apparent molecular weight) when the receptor was expressed alone, suggesting that Ephrin association with CXCR4 might affect CXCR4 post-translational modifications (see Fig. 2A and B). Treatment of protein extracts from cells expressing the receptor alone with N-glycosidase F, but not O-glycosidase, led to the disappearance of the higher molecular weight band and to a receptor migration pattern on SDS-PAGE identical to that observed in the presence of Ephrin B1 (single band of apparent molecular weight of 40 kDa, Fig. 2B). Moreover, N-glycosidase F treatment did not affect the receptor migration pattern in cells co-expressing Ephrin B1. Collectively, these results suggest that the co-expression of Ephrin B1 prevents CXCR4 N-glycosylation. Further supporting this hypothesis, Ephrin B1 expression did not affect the migration of CXCR4 mutated on N-glycosylated (positions 11 and 33) and O-glycosylated (position 18) residues (CXCR4ΔGlyc, Fig. 2B). Fig. 2Impact of Ephrin B1 on CXCR4 glycosylation and phosphorylation. A Snake plot of CXCR4. Glycosylation sites [43, 44], and phosphorylation sites on the C-terminal domain of the receptor [5, 45, 46] are highlighted. B Representative Western blots obtained from HEK293T cells expressing either wild-type HA-CXCR4 or CXCR4 carrying mutations on glycosylated residues (N11Q, S18A and N33Q, CXCR4ΔGlyc) in the presence or absence of Flag-Ephrin B1. Cell lysates were treated or not with PNGaseF (10 IU/100 µg) and/or O-glycosidase (20 IU/100 µg) for 30 min at 37 °C to remove N-glycosylation and/or O-glycosylation, respectively. The illustrated blots are representative of 3 independent experiments performed on different sets of cultured cells. C Representative Western blots performed from HEK293T cells, expressing or not HA-CXCR4 in the absence or presence of FLAG-Ephrin B1. Data are means ± SEM from 3 independent experiments performed from different sets of cultured cells. They were analyzed with a 2-way ANOVA with Bonferroni’s multiple comparison test. ns: non-significant Impact of Ephrin B1 on CXCR4 glycosylation and phosphorylation. A Snake plot of CXCR4. Glycosylation sites [43, 44], and phosphorylation sites on the C-terminal domain of the receptor [5, 45, 46] are highlighted. B Representative Western blots obtained from HEK293T cells expressing either wild-type HA-CXCR4 or CXCR4 carrying mutations on glycosylated residues (N11Q, S18A and N33Q, CXCR4ΔGlyc) in the presence or absence of Flag-Ephrin B1. Cell lysates were treated or not with PNGaseF (10 IU/100 µg) and/or O-glycosidase (20 IU/100 µg) for 30 min at 37 °C to remove N-glycosylation and/or O-glycosylation, respectively. The illustrated blots are representative of 3 independent experiments performed on different sets of cultured cells. C Representative Western blots performed from HEK293T cells, expressing or not HA-CXCR4 in the absence or presence of FLAG-Ephrin B1. Data are means ± SEM from 3 independent experiments performed from different sets of cultured cells. They were analyzed with a 2-way ANOVA with Bonferroni’s multiple comparison test. ns: non-significant We also examined the impact of Ephrin B1 expression on CXCR4 phosphorylation. CXCR4 is known to be sequentially phosphorylated on multiple residues located in its C-terminal domain, first on Ser346/347 by GRK2/3 and then on Ser330, Ser324/325 and Ser338/339 by GRK6 (Fig. 2A) [5, 21, 31]. Studies also suggested that PKC can phosphorylate Ser324/325 [5, 21]. Using phosphosite specific antibodies, we found that Ephrin B1 expression differentially affected the phosphorylation of Ser346/347, Ser330 and Ser324/325: while it enhanced Ser346/347 phosphorylation in cells exposed or not to CXCL12 for 5 min, it abolished the phosphorylation of both Ser330 and Ser324/325 in cells stimulated or not with CXCL12 (Fig. 2C). Impact of Ephrin B1 on CXCR4 subcellular localization Given the role of CXCR4 post-translational modifications in its trafficking, we investigated the effect of Ephrin B1 expression on the subcellular localization of YFP-tagged CXCR4 in HEK-293T cells by confocal fluorescent microscopy. While an important fraction of CXCR4-YFP was detected at the plasma membrane when it was expressed alone, in cells co-expressing Ephrin B1-RFP it showed a more pronounced distribution in intracellular compartments, where both proteins were colocalized (Fig. 3A, see also Fig. S5 for the cellular distribution of Ephrin B1-RFP when it was expressed alone). Co-immunostaining of cells co-expressing CXCR4-YFP and Ephrin B1-RFP with an antibody recognizing calnexin (endoplasmic reticulum) or GM130 (Golgi apparatus) showed an important fraction of the receptor in the endoplasmic reticulum and Golgi apparatus specifically in cells co-expressing CXCR4-YFP or Ephrin B1-RFP (Fig. 3A). Likewise, a fraction of the receptor was detected in the lysosomal compartment of cells co-expressing both proteins, as shown by co-immunostaining with a LAMP1 antibody (Fig. 3A). We next examined the effect of Ephrin B1 expression on plasma membrane localization of CXCR4-Nluc receptor using a Nano-BRET assay and LYN-NeonGreen as an acceptor anchored at the plasma membrane [17]. Confirming our confocal microscopy observations, the BRET signal decreased as a function of the amount of Ephrin B1 co-expressed with the receptor (Fig. 3B). Note that the co-expression of Ephrin B1 did not modify the total expression of CXCR4, as assessed by the total CXCR4-NLuc signal (Fig. S6). Given the limited level of CXCR4 internalization measured upon stimulation by CXCL12 for 60 min in our NanoBRET assay using NeonGreen-LYN (Fig. 3B), we assessed the impact of Ephrin B1 on CXCL12-induced CXCR4 internalization using a more sensitive cell surface detection approach based on the HiBiT Nano-luciferase complementation technology. Cells expressing N-terminally HiBiT-tagged CXCR4 were stimulated with CXCL12 and the remaining receptors at the plasma membrane were quantified by adding soluble LgBiT fragment. As expected, treatment of cells expressing CXCR4 alone with CXCL12 (10 nM, 30 min) already induced a significant internalization of the receptor that was further enhanced in cells exposed to 50 nM CXCL12, while co-expression of Ephrin B1 reduced CXCL12-induced receptor internalization (Fig. 3C). Consistent with these observations and with the well-described β-arrestin-dependent CXCR4 internalization elicited by its activation by CXCL12 [32, 33], co-expression of Ephrin B1 with CXCR4 inhibited the recruitment of both β-arrestin 1 and β-arrestin 2 by the receptor, as assessed using a Nano-luciferase complementation-based assay (Fig. 3D-F). Fig. 3Influence of Ephrin B1 on CXCR4 cell surface expression and internalization. A Stacked images of CXCR4-YFP (cyan) and Ephrin B1-RFP (magenta) fluorescent signals in different cellular compartments in HEK293T cells expressing CXCR4-YFP alone (left) or co-expressing CXCR4-YFP and Ephrin B1-RFP (right). The anti-calnexin immunostaining was used to label the endoplasmic reticulum, the anti-GM130 immunostaining the cis-Golgi, and the anti-LAMP1 immunostaining lysosomes (yellow). The illustrated fields are representative of three independent experiments performed on different sets of cultured cells. Scale bar: 10 μm. The line graphs were generated in Image J using the lines represented on the merged images, highlighting the colocalization of CXCR4 and Ephrin B1 in the three compartments. PM, plasma membrane signal. B Left: schematic representation of the NanoBRET assay assessing the interaction between CXCR4-NLuc and Lyn-mNeonGreen used to evaluate the plasma membrane localization of CXCR4 in cells expressing CXCR4-NLuc alone or coexpressing CXCR4-NLuc and FLAG-Ephrin B1 (transfected plasmid ratios, 1:1 and 1:3). Cells were exposed to vehicle or CXCL12 (10 or 50 nM) for 60 min. Right: the data represent the BRET ratio normalized to the vehicle condition in cells expressing CXCR4-NLuc alone. They are the means ± SEM of values obtained in duplicate in three independent experiments performed on different sets of cultured cells. C Schematic representation of the Nano-luciferase complementation assay used to evaluate the CXCL12-induced internalization of CXCR4. Right: the data represent CXCL12-induced internalization of CXCR4 normalized to vehicle in the corresponding condition, in cells expressing HiBIT-CXCR4 alone or coexpressing HiBIT-CXCR4 and FLAG-Ephrin B1 (plasmid ratios, 1:1 and 1:10). They are means ± SEM of duplicate determinations in three independent experiments performed on different sets of cultured cells. In B and C, the statistical analysis was the 2-way ANOVA with Bonferroni’s multiple comparison test. ns: non-significant. D Schematic representation of the Nano-luciferase complementation assay used to evaluate CXCL12-induced recruitment of β-arrestin 1,2 to CXCR4. E and F β-arrestin 1 (E) and β-arrestin 2 (F) recruitment to CXCR4 in response to exposure of HEK293T cells co-expressing CXCR4-LgBiT and SmBiT-β-arrestin1 or 2 in the absence or presence of FLAG-Ephrin B1 (CXCR4-LgBiT/FLAG-Ephrin B1 plasmid ratios, 1:1 and 1:10) to the indicated concentrations of CXCL12. The data are the means ± SEM of duplicate determinations performed in three independent experiments on different sets of cultured cells. Curves were fitted by the log[CXCL12] vs. luminescence non-linear regression using Prism Influence of Ephrin B1 on CXCR4 cell surface expression and internalization. A Stacked images of CXCR4-YFP (cyan) and Ephrin B1-RFP (magenta) fluorescent signals in different cellular compartments in HEK293T cells expressing CXCR4-YFP alone (left) or co-expressing CXCR4-YFP and Ephrin B1-RFP (right). The anti-calnexin immunostaining was used to label the endoplasmic reticulum, the anti-GM130 immunostaining the cis-Golgi, and the anti-LAMP1 immunostaining lysosomes (yellow). The illustrated fields are representative of three independent experiments performed on different sets of cultured cells. Scale bar: 10 μm. The line graphs were generated in Image J using the lines represented on the merged images, highlighting the colocalization of CXCR4 and Ephrin B1 in the three compartments. PM, plasma membrane signal. B Left: schematic representation of the NanoBRET assay assessing the interaction between CXCR4-NLuc and Lyn-mNeonGreen used to evaluate the plasma membrane localization of CXCR4 in cells expressing CXCR4-NLuc alone or coexpressing CXCR4-NLuc and FLAG-Ephrin B1 (transfected plasmid ratios, 1:1 and 1:3). Cells were exposed to vehicle or CXCL12 (10 or 50 nM) for 60 min. Right: the data represent the BRET ratio normalized to the vehicle condition in cells expressing CXCR4-NLuc alone. They are the means ± SEM of values obtained in duplicate in three independent experiments performed on different sets of cultured cells. C Schematic representation of the Nano-luciferase complementation assay used to evaluate the CXCL12-induced internalization of CXCR4. Right: the data represent CXCL12-induced internalization of CXCR4 normalized to vehicle in the corresponding condition, in cells expressing HiBIT-CXCR4 alone or coexpressing HiBIT-CXCR4 and FLAG-Ephrin B1 (plasmid ratios, 1:1 and 1:10). They are means ± SEM of duplicate determinations in three independent experiments performed on different sets of cultured cells. In B and C, the statistical analysis was the 2-way ANOVA with Bonferroni’s multiple comparison test. ns: non-significant. D Schematic representation of the Nano-luciferase complementation assay used to evaluate CXCL12-induced recruitment of β-arrestin 1,2 to CXCR4. E and F β-arrestin 1 (E) and β-arrestin 2 (F) recruitment to CXCR4 in response to exposure of HEK293T cells co-expressing CXCR4-LgBiT and SmBiT-β-arrestin1 or 2 in the absence or presence of FLAG-Ephrin B1 (CXCR4-LgBiT/FLAG-Ephrin B1 plasmid ratios, 1:1 and 1:10) to the indicated concentrations of CXCL12. The data are the means ± SEM of duplicate determinations performed in three independent experiments on different sets of cultured cells. Curves were fitted by the log[CXCL12] vs. luminescence non-linear regression using Prism Fig. 4Influence of Ephrin B1 on CXCR4 signaling. A Gαi1/3 recruitment to CXCR4 in HEK293T cells expressing CXCR4-YFP alone or co-expressing CXCR4-YFP and FLAG-Ephrin B1, in presence of either Rluc-Gαi1 or Rluc-Gαi3 and exposed or not to CXCL12 (10 nM). BRET values were normalized to the maximum BRET measured in cells expressing CXCR4 and challenged with CXCL12. The data are means ± SEM of triplicate determinations performed in three independent experiments on different sets of cultured cells. Normalized BRET variations were fitted with the one-site total binding equation and background constraint to constant value of 0 using Prism. B Gαi1/3-Gβγ dissociation elicited by exposure of HEK-293T cells expressing HA-CXCR4, alone or in combination with FLAG-Ephrin B1 (plasmid ratios 1:1 and 1:10), Rluc-Gαi1/3, Venus-γ2, and Flag-β2, to the indicated concentrations of CXCL12. The data were normalized to the values measured in vehicle-treated cells for in each condition. They are means ± SEM of triplicate determinations performed in three independent experiments on different sets of cultured cells. Curves were fitted by the log[CXCL12] vs. mBRET non-linear regression using Prism. C Representative Western blots performed from HEK293T cells expressing HA-CXCR4 or FLAG-Ephrin B1 or co-expressing HA-CXCR4 and FLAG-Ephrin B1, and treated with vehicle or CXCL12 (10 nM) for 5 min. When indicated, Gallein (25 µM) or AMD 3100 (1 µM) were added to cells 30 min before the vehicle/CXCL12 treatment. The data represent the pErk1/2/Total Erk1/2 ratio (in %). They are means ± SEM of values obtained in three independent experiments performed on different sets of cultured cells. They were analyzed with the 2-way ANOVA with Bonferroni’s multiple comparison test. ns: non-significant Influence of Ephrin B1 on CXCR4 signaling. A Gαi1/3 recruitment to CXCR4 in HEK293T cells expressing CXCR4-YFP alone or co-expressing CXCR4-YFP and FLAG-Ephrin B1, in presence of either Rluc-Gαi1 or Rluc-Gαi3 and exposed or not to CXCL12 (10 nM). BRET values were normalized to the maximum BRET measured in cells expressing CXCR4 and challenged with CXCL12. The data are means ± SEM of triplicate determinations performed in three independent experiments on different sets of cultured cells. Normalized BRET variations were fitted with the one-site total binding equation and background constraint to constant value of 0 using Prism. B Gαi1/3-Gβγ dissociation elicited by exposure of HEK-293T cells expressing HA-CXCR4, alone or in combination with FLAG-Ephrin B1 (plasmid ratios 1:1 and 1:10), Rluc-Gαi1/3, Venus-γ2, and Flag-β2, to the indicated concentrations of CXCL12. The data were normalized to the values measured in vehicle-treated cells for in each condition. They are means ± SEM of triplicate determinations performed in three independent experiments on different sets of cultured cells. Curves were fitted by the log[CXCL12] vs. mBRET non-linear regression using Prism. C Representative Western blots performed from HEK293T cells expressing HA-CXCR4 or FLAG-Ephrin B1 or co-expressing HA-CXCR4 and FLAG-Ephrin B1, and treated with vehicle or CXCL12 (10 nM) for 5 min. When indicated, Gallein (25 µM) or AMD 3100 (1 µM) were added to cells 30 min before the vehicle/CXCL12 treatment. The data represent the pErk1/2/Total Erk1/2 ratio (in %). They are means ± SEM of values obtained in three independent experiments performed on different sets of cultured cells. They were analyzed with the 2-way ANOVA with Bonferroni’s multiple comparison test. ns: non-significant Impact of Ephrin B1 on CXCR4 signaling Consistent with previous findings demonstrating that CXCR4 is canonically coupled with Gi proteins [34], saturation BRET analysis under conditions of constant RLuc-Gαi1 or RLuc-Gαi3 expression and increasing CXCR4-YFP expression level showed that CXCR4 recruits both Gαi1 and Gαi3 proteins in HEK-293T cells (Fig. 4A). Treatment of cells with CXCL12 did not markedly modify the recruitment of both G proteins by the receptor, whereas it was strongly reduced by the co-expression of Ephrin B1. Corroborating these observations, Ephrin B1 expression also decreased the ability of CXCL12 stimulation of CXCR4 to promote dissociation Gαi1 and Gαi3 from Gβγ, as assessed by the decrease in the BRET signal between venus-Gγ2 and Gαi1-RLuc or Gαi3-RLuc (Fig. 4B). We next investigated the effect of Ephrin B1 expression on the ability of CXCR4 to activate the Erk1/2 pathway, a key signaling cascade underlying tumor progression elicited by the CXCL12-CXCR4 axis. As previously shown [35], stimulation of CXCR4 endogenously expressed in HEK-293T cells with CXCL12 (10 nM, 5 min) induced an increase in Erk1/2 phosphorylation that was slightly but not significantly enhanced in cells transfected with a CXCR4 construct (Fig. 4C). In contrast, the CXCL12 response was strongly enhanced in cells transfected with an Ephrin B1 construct alone or co-transfected with the CXCR4 and Ephrin B1 constructs. The strong Erk1/2 activation induced by CXCL12 in cells overexpressing Ephrin B1 did involve CXCR4 activation, as it was abolished by pretreating cells with the CXCR4 antagonist AMD3100 (1 µM, 30 min, Fig. 4C). The activation of the Erk1/2 signaling cascade by GPCR ligands is often a complex process that involves G protein-dependent and β-arrestin-dependent mechanisms [24]. As shown on Fig. 3E and F, expression of Ephrin B1 inhibits β-arrestin 1 and β-arrestin 2 recruitment by CXCR4, making unlikely a role of β-arrestins in the enhanced CXCR4-operated Erk1/2 signaling observed in the presence of Ephrin B1. A previous study has demonstrated a role Gβγ and its translocation to the Golgi apparatus in Erk1/2 activation induced by CXCR4 stimulation in several cancer cell lines and HEK-293 cells [6]. Consistent with these findings and the presence of CXCR4 and Ephrin B1 in Golgi apparatus in cells expressing both protein partners (Fig. 3A), the activation of Erk1/2 induced by CXCL12 in cells expressing both protein partners (or overexpressing Ephrin B1 alone) was abolished by pretreating cells with the Gβγ pharmacological inhibitor Gallein (25 µM, 30 min, Fig. 4C). Role of Ephrin B1 in the regulation of Death Receptor 5 expression by CXCR4 in breast cancer cells Previous studies have shown that intracellular CXCR4 constitutively promotes the downregulation of Death Receptor 5 (DR5) independently of agonist stimulation of receptor signaling [9]. In light of these findings and the ability of Ephrin B1 to promote intracellular localization of the receptor, we next examined its influence on CXCR4-dependent regulation of DR5 expression level in MCF7 breast cancer cells. Expression of CXCR4 in MCF7 cells significantly reduced DR5 levels (Fig. 5A), consistent with the predominant intracellular localization of the receptor (Fig. 5B). This reduction in DR5 level was not prevented by treating cells with the CXCR4 antagonist AMD3100 (1 µM, added to the cells immediately after their transfection with the CXCR4 plasmid, Fig. 5A), confirming that it is independent of agonist receptor stimulation. In contrast, silencing Ephrin B1 expression in MCF7 cells using a shRNA, which reduced Ephrin B1 level by 61.90 ± 4.75% (n = 5, p < 0.0001, ANOVA followed by Dunnett’s test, Fig. S7), prevented the decrease in DR5 level elicited by CXCR4 overexpression (Fig. 5A). As expected, silencing Ephrin B1 expression also induced plasma membrane relocalization of a fraction of CXCR4 (Fig. 5B). These results suggest a key role of Ephrin B1 in the regulation of DR5 expression by CXCR4 in breast cancer cells. Fig. 5Impact of Ephrin B1 on CXCR4 regulation of Death Receptor 5 expression. A Representative Western blots assessing DR5, HA-CXCR4 and GAPDH expression in MCF7 cells transfected with empty vector (pCDNA3.1) or HA-CXCR4 constructs. Cells were also either transfected with control (scramble) shRNA or shRNA targeting Ephrin B1, or exposed to either vehicle or AMD3100 (1µM, added to cells after the transfection). The histogram shows immunoreactive signals of DR5 normalized to GAPDH signal. The data represent the means ± SEM of values obtained in five independent experiments performed on different sets of cultured cells. They were analyzed using multiple paired t-tests. B Left: Stacked images of CXCR4-YFP fluorescent signal in MCF7 cells co-transfected with the CXCR4-YFP construct and either the control shRNA or the Ephrin B1 shRNA. The illustrated fields are representative of three independent experiments performed on different sets of cultured cells. Scale bar: 10 μm. Right: Line graphs generated in Image J using the lines represented on the images highlight the subcellular localization of CXCR4. PM, plasma membrane signal Impact of Ephrin B1 on CXCR4 regulation of Death Receptor 5 expression. A Representative Western blots assessing DR5, HA-CXCR4 and GAPDH expression in MCF7 cells transfected with empty vector (pCDNA3.1) or HA-CXCR4 constructs. Cells were also either transfected with control (scramble) shRNA or shRNA targeting Ephrin B1, or exposed to either vehicle or AMD3100 (1µM, added to cells after the transfection). The histogram shows immunoreactive signals of DR5 normalized to GAPDH signal. The data represent the means ± SEM of values obtained in five independent experiments performed on different sets of cultured cells. They were analyzed using multiple paired t-tests. B Left: Stacked images of CXCR4-YFP fluorescent signal in MCF7 cells co-transfected with the CXCR4-YFP construct and either the control shRNA or the Ephrin B1 shRNA. The illustrated fields are representative of three independent experiments performed on different sets of cultured cells. Scale bar: 10 μm. Right: Line graphs generated in Image J using the lines represented on the images highlight the subcellular localization of CXCR4. PM, plasma membrane signal Discussion To identify mechanisms underlying tumorigenic effects of intracellular CXCR4 that are independent of CXCL12-induced signaling, we characterized the CXCR4 interactome and identified Ephrin B1 as a novel receptor-interacting protein. To our knowledge, this is the first demonstration of a physical interaction between a GPCR and a member of the Ephrin ligand family that comprises five glycosylphosphatidylinositol-anchored Ephrin A ligands (Ephrins A1-5) and three transmembrane Ephrin B ligands (Ephrins B1-3). Strikingly, Ephrin B1 was the most highly enriched protein identified in the CXCR4 interactome, suggesting a robust interaction between both partners. Corroborating this observation, we found that purified CXCR4 and Ephrin B1 can form a complex in vitro, indicative of a direct interaction between both proteins. We also demonstrated that the C-terminal domain of Ephrin B1 is essential to the interaction with CXCR4, suggesting that this intrinsically disordered region may facilitate protein–protein interactions through conformational flexibility. The ability of Ephrin B1 to associate with CXCR4 in living cells was validated by BRET experiments, which also demonstrated that agonist stimulation of CXCR4 promotes the recruitment of Ephrin B1 by the receptor, suggesting that CXCR4-Ephrin B1 interaction may be a dynamic process regulated by the receptor’s conformational state. Furthermore, BRET experiments showed a low ability of Ephrin B1 to interact with ACKR3, an atypical chemokine receptor sharing with CXCR4 the chemokine CXCL12 as ligand, reminiscent of a previous interactomic screen performed in the same cellular background (HEK-293T cells) that did not identify Ephrin B1 as an ACKR3 partner [22]. As ACKR3 and CXCR4 are known to form heteromers in HEK-293T cells [28], the low BRET signal measured in cells co-expressing ACKR3 and Ephrin B1 might reflect an indirect recruitment of Ephrin B1 through the heteromerization of ACKR3 with endogenously expressed CXCR4. Further supporting the specificity of CXCR4-Ephrin B1 interaction, Ephrin B1 was the only member of the Ephrin B ligand family identified in our interactomic screen. Furthermore, Ephrin B2 and Ephrin B3 did not co-immunoprecipitate with CXCR4 and the BRET signal was much lower in cells co-expressing CXCR4-Neon Green and Ephrin B2-NLuc or Ephrin B3-NLuc than that measured in cells expressing Ephrin B1-NLuc. The common roles of CXCR4 and Ephrin ligands and their receptors in cancer progression and resistance to chemotherapies prompted us to further investigate the reciprocal impact of Ephrin B1 and CXCR4 on their cellular localization and signaling. Using both confocal fluorescence microscopy and a Nano-BRET assay assessing plasma membrane localization of CXCR4, we provide converging evidence that Ephrin B1 promotes intracellular localization of the receptor in HEK-293T cells stimulated or not with CXCL12. Likewise, while Ephrin B1 was mainly detected at the plasma membrane of cells expressing RFP-Ephrin B1 alone, it was predominantly found in intracellular compartments of cells co-expressing CXCR4-YFP. These include the endoplasmic reticulum and the Golgi apparatus as well as lysosomes, where both proteins were colocalized. The increase in CXCR4 intracellular localization induced by Ephrin B1 expression might result from the retention of the receptor in intracellular compartments and/or the increase in its internalization. Using HiBiT surface Nanoluciferase complementation, we found that Ephrin B1 prevents rather than enhances CXCL12-induced CXCR4 internalization, an effect that correlates with the ability of Ephrin B1 to inhibit β-arrestin1 and β-arrestin2 recruitment to the receptor. This suggests that Ephrin B1 promotes CXCR4 intracellular localization through the inhibition of its targeting to the plasma membrane. Consistent with this hypothesis, the ability of Ephrin B1 to associate with non-glycosylated CXCR4 in vitro and to abolish its glycosylation suggests that CXCR4 and Ephrin B1 can associate early in the biosynthetic pathway, an effect leading to the inhibition of the forward trafficking and the intracellular sequestration of an important fraction of both proteins. The decreased CXCR4 plasma membrane localization in cells co-expressing Ephrin B1 might also be one of the mechanisms contributing to the inhibitory effect of Ephrin B1 on CXCL12-induced β-arrestin1 and β-arrestin2 recruitment to CXCR4. β-arrestin recruitment to GPCRs is known to depend on their phosphorylation on residues located in their C-terminal domain by GRKs and other protein kinases. Our results show that Ephrin B1 expression enhances CXCR4 phosphorylation on Ser346–347, two residues phosphorylated by GRK2/3 located in the cytoplasm [36, 37]. In contrast, Ephrin B1 expression abolishes the phosphorylation of residues (Ser324/325 and Ser330) phosphorylated by GRK6, which is constitutively localized at the plasma membrane [37, 38]. Collectively, these findings are consistent with the predominant intracellular localization of the receptor in cells expressing Ephrin B1. CXCR4 phosphorylation at these sites also induce contrasting effects on β-arrestin recruitment and receptor internalization [5, 21]. Whereas the phosphorylation of Ser346–347 seems to be prerequisite to β-arrestin recruitment, inhibition of Ser324/325, Ser330 and Ser339 phosphorylation by their mutation into alanine favors β-arrestin recruitment [5]. Our results demonstrating that Ephrin B1 expression increases Ser346–347 phosphorylation while it abolishes Ser324/325 and Ser330 phosphorylation suggest that its inhibitory effect on β-arrestin recruitment to CXCR4 (and receptor internalization) is independent of its modulation of the receptor phosphorylation state. This inhibitory effect might rather result from the decreased plasma membrane localization of CXCR4 and/or steric hindrance by Ephrin B1 bound to the receptor. A recent study has shown an interaction between CXCR4 and receptor activity-modifying protein3 (RAMP3), which inhibits both constitutive and agonist-dependent β-arrestin recruitment to the receptor [39]. This suggests that β-arrestin recruitment to CXCR4 and β-arrestin-dependent CXCR4 signaling are finely regulated by its association with various protein partners. A previous study has demonstrated that EphB2-Ephrin B1 reverse signaling inhibits the migration of cerebellar granule neurons triggered by CXCL12-induced activation of CXCR4 [14]. In fact, Ephrin B1 interacts with Regulator of G protein signaling 3 (RGS3) to negatively regulate Gi protein activation by CXCR4 through the GTPase activating protein activity of its RGS domain. It is likely that this process could be favored by the physical interaction between Ephrin B1 and CXCR4. The decrease in CXCR4 plasma membrane localization induced by Ephrin B1 co-expression might also be one of the mechanisms underlying its inhibitory effect on CXCL12-induced Gαi protein activation and chemoattraction. Ephrin forward and reverse signaling involve extensive crosstalk with major cytosolic signaling pathways known to be engaged by tyrosine kinase receptors that control cell survival, migration, and differentiation, such as Erk1/2 signaling [40]. Here, we show that Ephrin B1 expression, which already increases Erk1/2 activation in the absence of a CXCR4 agonist, strongly potentiates the CXCL12 response, while treatment of cells with the CXCR4 antagonist AMD3100 abolishes Ephrin B1-dependent Erk1/2 activation. These findings indicate that Ephrin B1 engages Erk1/2 signaling through active CXCR4. CXCR4-dependent Erk1/2 activation is known to strongly depend on Gβγ proteins [6]. Further supporting the implication of CXCR4 in Ephrin B1 stimulation of Erk1/2 signaling, the inhibition of Gβγ proteins also abolished the strong Erk1/2 activation measured in cells expressing Ephrin B1. Collectively, these observations are reminiscent of recent findings indicating that the EGF receptor and the HER3/HER2 receptors can signal to Erk1/2 through activated CXCR4/ACKR3 [41] and suggest that components of the Ephrin/Eph systems can likewise hijack the CXCR4 signaling machinery to activate the Erk1/2 pathway. The synergistic effects of CXCR4 and Ephrin B1 on Erk1/2 signaling might be one of the mechanisms underlying their common deleterious influence on cancer cell growth, invasion and metastasis. Besides CXCR4-operated Erk signaling, it has recently been proposed that the intracellular localization of CXCR4 in cancer cells plays a critical role in the receptor’s tumorigenic effects in vitro and in vivo through the downregulation of DR5 expression [9]. This prompted us to explore the influence of Ephrin B1 on DR5 expression in breast cancer cells. We found that silencing Ephrin B1 expression in MCF7 cells abolished the reduction of DR5 level induced by overexpressing CXCR4, while promoting relocalization of a fraction of CXCR4 to the plasma membrane, suggesting that Ephrin B1 regulates DR5 expression by promoting intracellular localization of CXCR4. Collectively, these findings indicate that Ephrin B1 can contribute to the tumorigenic potential of CXCR4 through the enhancement of its oncogenic signaling and its retention in intracellular compartments where it inhibits cell apoptosis and thus renders cancer cells resistant to chemotherapeutic drugs. Together with agonist-independent tumorigenic signaling elicited by CXCR4 oligomers [8], these observations may also explain the limited clinical efficacy of CXCR4 antagonists such as AMD3100. Although the role of the physical CXCR4-Ephrin B1 interaction could not be fully established in the absence of identification of binding motifs within the sequences of both partners, they suggest that targeting Ephrin B1 reverse signaling and/or its interaction with CXCR4 might be a relevant strategy in complement to CXCR4 antagonists to dampen the tumorigenic effects of this receptor. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1