Neuropilin 1 and its inhibitory ligand mini-tryptophanyl-tRNA synthetase inversely regulate VE-cadherin turnover and vascular permeability

Abstract

The formation of a functional blood vessel network relies on the ability of endothelial cells (ECs) to dynamically rearrange their adhesive contacts in response to blood flow and guidance cues, such as vascular endothelial growth factor-A (VEGF-A) and class 3 semaphorins (SEMA3s). Neuropilin 1 (NRP1) is essential for blood vessel development, independently of its ligands VEGF-A and SEMA3, through poorly understood mechanisms. Grounding on unbiased proteomic analysis, we report here that NRP1 acts as an endocytic chaperone primarily for adhesion receptors on the surface of unstimulated ECs. NRP1 localizes at adherens junctions (AJs) where, interacting with VE-cadherin, promotes its basal internalization-dependent turnover and favors vascular permeability initiated by histamine in both cultured ECs and mice. We identify a splice variant of tryptophanyl-tRNA synthetase (mini-WARS) as an unconventionally secreted extracellular inhibitory ligand of NRP1 that, by stabilizing it at the AJs, slows down both VE-cadherin turnover and histamine-elicited endothelial leakage. Thus, our work shows a role for NRP1 as a major regulator of AJs plasticity and reveals how mini-WARS acts as a physiological NRP1 inhibitory ligand in the control of VE-cadherin endocytic turnover and vascular permeability.

Fig. 1. Generation and expression of a functional HaloTag-NRP1 construct in ECs.

a HaloTag protein tag fused at the N-terminus of mouse NRP1 (HT-NRP1); sp, signal peptide; HT, HaloTag; a1/a2, complement C1r/C1s, Uegf, Bmp1 (CUB) domain 1/2; b1/b2, coagulation factor VIII/V (FVIII/FV) domain 1/2; c, meprin/A5/µ-phosphatase (MAM) domain. b Western blot analysis of exogenous HT-NRP1 construct and endogenous α-Tubulin in ECs transduced with empty pCCL lentivirus (Ctrl) or carrying HT-NRP1 construct. A representative experiment out of three is shown. c, d HaloTag NRP1 localizes at the cell membrane when endocytosis is inhibited, whereas it co-localizes with the EEA1 endocytic marker during internalization. Fluorescent confocal microscopy on ECs transduced with pCCL HT-NRP1, incubated for 15 min at 4 °C (to inhibit endocytosis) with the cell impermeable HaloTag-Alexa 660-labeled chloroalkane (magenta) fluorescent ligand (c), and then allowed to internalize membrane proteins by a shift to 37 °C (d). ECs were simultaneously stained for the endogenous early endosomal marker EEA1 (green). HaloTag-Alexa 660 bound HT-NRP1 localized on the surface of ECs kept at 4 °C (c) and is efficiently endocytosed only upon incubation at 37 °C for 3 min, where HT-NRP1 co-localized in EEA1-positive early endosomes (d yellow arrows). Bottom panels are magnifications of top panels. Scale bar is 10 µm. A representative experiment out of three is shown. e VEGF-A165 promotes the association of HT-NRP1 to VEGFR-2. Empty pCCL (Ctrl) or pCCL-HaloTag-NRP1 transduced ECs (HT-NRP1) were labeled with 3 µM HaloTag cell impermeable Biotin ligand for 15 min at 4 °C (to inhibit endocytosis) and stimulated or not with VEGF-A165 (30 ng/ml, 10 min at 37 °C), then immunoprecipitated on streptavidin-beads. A representative experiment out of two is shown. f HT-NRP1 rescues the defective adhesion of siNRP1 ECs to fibronectin. Results are the mean of three independent experiments (each in technical triplicate) ±SEM. Statistical analysis: parametric one-way analysis of variance (ANOVA) with Bonferroni correction; *P value = 0.0199, **P value = 0.0041. Source data are provided as Source Data file.

Fig. 2. NRP1 interactome.

a Schematic drawing summarizing the role of NRP1 as a pro-endocytic receptor^,,, that promotes the internalization and recycling of associated plasma membrane receptors, such as integrins^, and other still to be identified receptors (in gray). b Cell adhesion receptors as preferential potential transmembrane NRP1 interactors. Chord diagram of transmembrane proteins found to interact with NRP1. Each protein was assigned to one or more subcellular locations based on UniProt annotation. In addition, to associate with transmembrane proteins mediating or regulating EC adhesion, NRP1 also associated with the endoplasmic reticulum reticulon 4 (RTN4) transmembrane protein that promotes cell adhesion by regulating integrin traffic. Moreover, our identification of the transmembrane metallopeptidase endothelin converting enzyme-1 (ECE1) as a NRP1 partner on the EC surface may suggest that the TBX1-independent control of cardiac OFT morphogenesis by endothelial NRP1 may involve ECE1^,, which functions by degrading receptor-bound extracellular ligands upon endocytosis (Supplementary Data 1). c Aminoacyl-tRNA proteins were found to be significantly enriched in the NRP1 interactome, after excluding transmembrane (b) and endosomal (Supplementary Fig. 2a) NRP1 interactors. STRING was used to perform enrichment analysis of KEGG pathways and to determine the physical and functional protein-protein interactions that define the edges between nodes. The plots was generated with R. Source data are provided as a Supplementary Data file.

Fig. 3. Endogenous NRP1 localizes at intercellular contacts and interacts with cell adhesion receptors in ECs.

a, b NRP1 co-localizes with VE-cadherin and PECAM1 at endothelial intercellular contacts. Confocal microscopy analysis of endogenous NRP1 (magenta) and VE-cadherin (a) and PECAM1 (b) (green) in cultured ECs. Scale bar is 10 µm. c, d NRP1 co-immunoprecipitates with VE-cadherin and PECAM1 intercellular adhesion receptors. Co-immunoprecipitation of endogenous NRP1 and VE-cadherin (c) and PECAM1 (d) from lysates of cultured ECs that were previously incubated live with the mouse mAb (R&D Systems—MAB3870) recognizing the extracellular domain of human NRP1 or control mouse IgGs (Ctrl). A representative experiment out of five (c) or four (d) is shown. e Pull-down experiments further confirm VE-cadherin as a NRP1 binding partner. Representative western blot analysis (out of three biological replicates) of in vitro pull-down assay in which NRP1-Fc-tagged bound to an agarose-conjugated anti-NRP1 antibody was incubated with recombinant VE-cadherin-Fc-tagged or Fc alone. Source data are provided as Source Data file.

Fig. 4. NRP1 promotes VE-cadherin turnover and histamine-elicited endothelial permeability.

a NRP1 impairs VE-cadherin localization at endothelial cell-to-cell contacts. Confocal microscopy on control (siCtrl) or NRP1 (siNRP1) silenced ECs rescued with empty vector (Ctrl/siCtrl and Ctrl/siNRP1) or with mouse Nrp1 (Nrp1/siNRP1) and stained for NRP1 (magenta) and VE-cadherin (green) and quantification of VE-cadherin fluorescence intensity in AJs. Scale bar 10 µm. Mean ± SEM of three independent experiments (depicted in different colors). One-way ANOVA with Bonferroni correction; ***P value =  0.0004. b NRP1 silencing does not affect VE-cadherin protein levels. A representative experiment of three is shown. c FRAP analysis reveals NRP1 promotes VE-cadherin turnover at AJs in ECs transduced with VE-cadherin-mCherry and silenced or not for NRP1. Above, VE-cadherin fluorescence recovery = normalized signal in each cell before and after photobleaching to 100% and 0%. Three independent experiments (8 cells each) mean ± SEM is shown. Below, mobile fraction = fluorescence recovered post-bleach —after photobleaching/pre-bleach ROI fluorescence—after photobleaching. Mean of three independent experiments ± SEM is shown. Scale bar 5 µm. Two-way ANOVA with Bonferroni correction; ns = P value ≥ 0.05, *P value ≤ 0.05, **P value ≤ 0.01, ***P value = 0.000028. d NRP1 promotes VE-cadherin internalization. VE-cadherin endocytosis in Ctrl/siCtrl, Ctrl/siNRP1 or Nrp1/siNRP1 ECs measured by capture ELISA. Mean ± SEM of three independent experiments is shown. One-way ANOVA with Bonferroni correction; **P value = 0.0037, ***P value = 0.0003. e NRP1 sustains histamine-elicited endothelial permeability. Quantification of Dextran through Ctrl/siCtrl, Ctrl/siNRP1 or Nrp1/siNRP1 EC monolayer after histamine. Mean ± SEM of three independent experiments is shown. Two-way ANOVA with Bonferroni correction; *P value ≤ 0.05, **P value ≤ 0.01. f Vascular permeability reduction in Nrp1^EC−/− mice. Left, confocal microscopy on Nrp1^+/+ or Nrp1^EC−/− mice ear sections stained for NRP1 (magenta), MECA32 (green), and DAPI (blue). Scale bar 75 µm. Right, histamine permeability in Nrp1^+/+ or Nrp1^EC−/− mice (Evans Blue/skin weight). Data are mean ± SD (n = 5 per group). Two-tailed heteroscedastic Student’s t test; *P value = 0.0321. Source data are provided as Source Data file.

Fig. 5. Mini-WARS is an extracellular inhibitory NRP1 ligand that impairs VE-cadherin turnover.

a Schematic showing the different domains in human Full Length (FL, red), mini (green) and T2 (blue) WARS isoforms; WHEP, vertebrate-specific extension; ESE, eukaryote-specific extension. b Endogenous NRP1 and mini-WARS co-immunoprecipitation. An anti-WARS Ab recognizing the catalytic Rossmann fold and the anticodon recognition domain detected only the mini-WARS band (green arrow) co-immunoprecipitating with NRP1. A representative experiment out of five is shown. c Pull-down experiment identifies mini-WARS as a preferential binding partner of NRP1-Fc. A representative western blot from three independent experiments is shown. d Mini-WARS interaction with NRP1 does not rely on its cytoplasmic domain. A HA antibody used to detect WT or ∆cyto or ∆b1/b2 NRP1-HA co-immunoprecipitated with FL or T2 or mini-WARS-V5 co-transfected in ECs. The image is representative of three independent experiments. e S100A10 silencing increases the secretion of V5-mini-WARS in the EC extracellular medium. Control (siCtrl) or S100A10 (siS100A10) silenced ECs were transfected with mini-WARS V5 tagged and their medium loaded on V5 beads. A representative western blot from three independent experiments is shown. f Mini-WARS fosters NRP1 and VE-cadherin localization at endothelial cell-to-cell contacts. Confocal microscopy on pCCL empty lentivirus (Ctrl) or pCCL mini-WARS transduced ECs stained for NRP1 (magenta) and VE-cadherin (green); quantification of NRP1 (above) and VE-cadherin (below) fluorescence intensity in AJs. Data are the mean ± SEM of three independent experiments (depicted in different color shades). Scale bar is 10 µm. Statistical analysis: two-tailed heteroscedastic Student’s t test; *P value = 0.0164, ***P value = 0.0005. g Mini-WARS overexpression does not affect NRP1 and VE-cadherin protein levels in ECs. A representative experiment out of three is shown. h Mini-WARS overexpression decreases NRP1 and VE-cadherin internalization in ECs. Percentage of NRP1 (left) and VE-cadherin (right) endocytosis in Ctrl or mini-WARS-transduced ECs as measured by surface labeling followed by capture ELISA. Data are the mean ± SEM of three independent experiments. Statistical analysis: two-tailed heteroscedastic Student’s t test; *P value = 0.0201 (left), *P value = 0.0498 (right). Source data are provided as a Source Data file.

Fig. 6. Mini-WARS is an extracellular inhibitory NRP1 ligand that impairs histamine-elicited endothelial permeability.

a Mini-WARS overexpression decreases histamine-elicited endothelial permeability. Quantification of 70 kDa Dextran-FITC transendothelial flux over time through a confluent Ctrl (empty lentivirus) or mini-WARS-transduced EC monolayer after histamine stimulation. Data are the mean ± SEM of three independent experiments. Statistical analysis: two-way ANOVA with Bonferroni correction; *P value = 0.0126, **P value = 0.0020, 0.0016. Western blot analysis shows mini-WARS overexpression in ECs. b Mini-WARS silencing impairs histamine-elicited endothelial permeability. Quantification of 70 kDa Dextran-FITC transendothelial flux over time through a confluent control (siCtrl) or WARS (siWARS) silenced ECs rescued for empty vector (Ctrl/siCtrl and Ctrl/siWARS) or for mini-WARS (mini-WARS/siWARS) ECs monolayer after histamine stimulation. Data are the mean ± SEM of three independent experiments. Statistical analysis: two-way ANOVA with Bonferroni correction; **P value ≤ 0.01. Western blot analysis shows WARS silencing and mini-WARS overexpression. c Mini-WARS overexpression decreases histamine-elicited endothelial permeability via NRP1. Quantification of 70 kDa Dextran-FITC transendothelial flux over time through a confluent control (siCtrl, left) or NRP1 silenced (siNRP1, right) and transduced with empty lentivirus (Ctrl) or mini-WARS EC monolayer after histamine stimulation. Data are the mean ± SEM of three independent experiments. Statistical analysis: two-way ANOVA with Bonferroni correction; ns = P value ≥ 0.05, **P value = 0.0047, 0.0070, 0.0010. Western blot analysis shows NRP1 silencing and mini-WARS overexpression in ECs. Source data are provided as a Source Data file.