Fig 1.
JAM-C is constitutively trafficked from endothelial cell junctions.
(A) Schematic of the endocytosis of endogenous JAM-C from the endothelial cell surface showing the extracellular immunoglobulin domains (blue) and the intracellular PDZ interacting domain (clear circle). (B) Untreated or (C) 100-nM-bafilomycin–treated HUVECs were fixed and costained for JAM-C (green) and VE-Cadherin (red). Boxed regions are shown magnified below in greyscale: (i) JAM-C and (ii) VE-Cadherin. Scale bar, 20 μm. (D) Quantification of the percent of untreated cells with a number of JAM-C–positive vesicles (551 cells from n = 3 experiments, error bars represent SEM). Underlying data are found in S1 Data. (E) Schematic of the endocytosis of JAM-C–HRPout from the endothelial cell surface showing the extracellular immunoglobulin domains (blue), the HRP tag (purple circle) and the intracellular PDZ interacting domain (clear circle). (F–I) HUVECs were transiently transfected with JAM-C–HRPout, fixed, and incubated with DAB and hydrogen peroxide for 30 min. Cells were then secondarily fixed and 70 nm sections prepared and imaged by transmission EM. Precipitated DAB can be clearly seen on (F) endocytic structures and early endosomes. (Fi) shows boxed area magnified, and red arrows highlight endocytic structures and endosomes: (G) junctions, (H and I) MVBs (red arrows). Scale bars, (F) 2 μm, (Fi) 500 nm, (G) 1 μm, (H) 2 μm, and (I) 200 nm. DAB, diaminobenzidine; EM, electron microscopy; HRP, horseradish peroxidase; HUVEC, human umbilical vein endothelial cell; JAM-C, junctional adhesion molecule-C; MVB, multivesicular body; PDZ, postsynaptic density protein/drosophila disc large tumour suppressor/zonula occludens protein 1; SEM, standard error of the mean; VE-Cadherin, vascular endothelial cadherin.
Fig 2.
JAM-C is turned over dynamically at the endothelial junction, and the rate of internalisation is increased during junctional disassembly and inflammation.
(A) HUVECs were fed with JAM-C antibody, excess antibody was removed, and the cells were incubated at 37°C before being fixed at 0 and 2 h and imaged by confocal microscopy. Scale bar, 20 μm; boxed regions are shown at higher magnification. (B) The difference in junctional intensity is quantified between 0 and 2 h (n = 4 experiments, error bars represent SEM, **p ≤ 0.01; t test). (C) HUVECs were treated with 4 mM EGTA in calcium-free, low-serum media for 30 min before washing and incubation at 37°C with conventional media. Cells were fixed before and during the time course of washout and recovery and stained for JAM-C (green) and VE-Cadherin (red). Images were acquired by confocal microscopy and the number of JAM-C–positive vesicles at each time point determined. Scale bar, 10 μm (5 fields of view/experiment, n = 3 experiments, error bars represent SD, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001; t test). (D) HUVECs were either left untreated or incubated with 10 ng/ml IL-1 or 50 ng/ml TNF-α for 4 h at 37°C. Cells were then fixed and stained for JAM-C (green) and VE-Cadherin (red). Images were acquired by confocal microscopy and (E) the number of JAM-C–positive vesicles at each time point determined. Scale bar, 10 μm (150 cells from n = 3 experiments, error bars represent SD, **p ≤ 0.01; t test). (F) Schematic of the endocytosis of JAM-C–GFPout from the endothelial cell surface showing the extracellular immunoglobulin domains (blue), the GFP tag (green circle), and the intracellular PDZ interacting domain (clear circle). (G) HUVECs were nucleofected with WT JAM-C–GFPout and imaged with a spinning-disk confocal microscope. Time indicates total time in media, and the boxed region at 0 s is shown magnified at later time points. JAM-C exists in vesicles, at the cell surface, and at the cell junctions. Large vesicles can be seen that tubulate and migrate away from the junctional region. Scale bar: 0 s, 20 μm; inset, 10 μm. Underlying data are found in S2 Data. GFP, green fluorescent protein; HUVEC, human umbilical vein endothelial cell; IL-1, interleukin-1; JAM-C, junctional adhesion molecule-C; n.s., not significant; PDZ, postsynaptic density protein/drosophila disc large tumour suppressor/zonula occludens protein 1; SEM, standard error of the mean; TNF, tumour necrosis factor; VE-Cadherin, vascular endothelial cadherin; WT, wild type.
Fig 3.
An HRP-based proximity-labelling approach reveals JAM-C cotraffics with NRP-1 and 2 and VE-Cadherin, but not with components of the LBRC.
(A) Schematic of the domain structure of the JAM-C–HRP construct. (B) In the presence of biotin tyramide and hydrogen peroxide, the HRP tag leads to the biotinylation of any proteins within 20 nm (labelled with an X) on tyrosine, tryptophan, cysteine, or histidine residues. (C) Endothelial cells expressing JAM-C–HRPout are fed with biotin tyramide, and then hydrogen peroxide is added for 1 min. Proteins within a 20-nm radius of JAM-C–HRP are labelled with biotin (shown in light purple). Addition of the membrane-impermeant HRP inhibitor ascorbate blocks the biotinylation reaction at the cell surface but not inside the cell. (D–F) HUVECs were transfected with JAM-C–HRPout and incubated in the presence or absence of bafilomycin A1 (100 nM) for 4 h to block lysosomal degradation. Cells were fed biotin tyramide for 30 min and then exposed to hydrogen peroxide for 1 min in the presence or absence of 50 mM ascorbate. (D) The cells were then fixed and labelled for streptavidin (green) and DAPI (blue) and images acquired by confocal microscopy or (E–F) the reaction was then quenched, and the cells were lysed. Biotinylated proteins were pulled down using neutravidin beads and lysate, and pulldown samples were analysed by (E) SDS-PAGE and western blot for the presence of JAM-C, tubulin, and biotinylated proteins or (F) following an on-bead tryptic digest mass spectrometry. A heat map of 4 independent mass spectrometry data sets is shown, with white meaning none and dark red a high signal. Individual experiments were carried out in duplicate, with each mass spectrometry run being repeated twice (to give a total of 4 analyses/experiment). p-Values are given across all 4 experiments (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001, t test). The + TNF-α experiment is shown in S3 Fig. Cotrafficked proteins appear in both ±ascorbate conditions, whilst proteins adjacent to JAM-C solely at the cell surface are only present in the −ascorbate condition. Scale bar, 20 μm. CD, cluster of differentiation; EGFR, epidermal growth factor receptor; ESAM, endothelial cell-selective adhesion molecule; HRP, horseradish peroxidase; HUVEC, human umbilical vein endothelial cell; ICAM, intercellular adhesion molecule; IL, interleukin; Jak-1, janus kinase 1; JAM-C, junctional adhesion molecule-C; LBRC, lateral border recycling compartment; LDL, low-density lipoprotein; NRP, neuropilin; n/s, not significant; PDZ, postsynaptic density protein/drosophila disc large tumour suppressor/zonula occludens protein 1; PECAM, platelet endothelial cell adhesion molecule; PLVAP, plasmalemma vesicle-associated protein; PTPR, receptor-type tyrosine phosphatase; Tie-2, tyrosine-protein kinase receptor 2; TM, transmembrane; TNF, tumour necrosis factor; VE-Cadherin, vascular endothelial cadherin; VEGFR, vascular endothelial growth factor receptor.
Fig 4.
A lysine residue in the cytoplasmic tail of JAM-C is required for its timely degradation.
(A) A schematic representation of the JAM-C–GFPout construct showing the residues in the cytoplasmic tail that were mutated by site-directed mutagenesis. (B) HUVECs were transfected with WT or mutant JAM-C–GFPout constructs, fixed, and stained for VE-Cadherin (red). The boxed area is magnified (bottom left of each panel) to show the Jam-C–GFP signal. All JAM-C mutants exhibit at least some junctional staining. The ΔPDZ construct exhibited the most marked change in localisation, existing additionally in a peri-Golgi region and expressing at higher levels. Scale bars, 20 μm. (C and D) HUVECs were transfected with WT or mutated JAM-C–GFPout constructs and treated with 10 μg/ml cycloheximide. (C) The rate of protein degradation was determined by SDS-PAGE and western blot over a 24-h period; a representative anti-GFP blot is shown. (D) Quantification of western blots normalised to the 0-h protein levels, data from n = 7 experiments, error bars represent SEM (*p ≤ 0.05; two-way ANOVA). Underlying data are found in S3 Data. a.a., amino acid; GFP, green fluorescent protein; HUVEC, human umbilical vein endothelial cell; JAM-C, junctional adhesion molecule-C; PDZ, postsynaptic density protein/drosophila disc large tumour suppressor/zonula occludens protein 1; SEM, standard error of the mean; TM, transmembrane; WT, wild type.
Fig 5.
JAM-C is ubiquitylated, and this governs its targeting to MVBs.
(A) A schematic representation of the JAM-C–EGFPout construct showing all the lysine residues in the cytoplasmic tail that can be potentially modified by ubiquitylation and the JAM-C mutant generated. (B) HUVECs were transfected with WT, Quad-K JAM-C–GFPout, or EGFR-GFP (as a positive control) and ubiquitin HA. Cells were lysed, and GFP-tagged proteins were pulled down using GFP trap agarose beads. Lysate and pulldown proteins were analysed by SDS-PAGE and western blot for the presence of GFP- and HA-tagged proteins. WT, but not Quad-K, JAM-C is modified with ubiquitin. (C) HUVECs were transfected with WT or Quad-K mutant JAM-C–GFPout and then fixed before imaging by confocal microscopy. Scale bar, 20 μm. (D and E) HUVECs were transfected with (D) WT or (E) Quad-K JAM-C–EGFPout and 100 nM bafilomycin was added (at t = 0 min) before imaging by time-lapse confocal microscopy. The number of vesicles over time in each cell or the average number of vesicles/cell is plotted (error bars represent SEM). A representative movie is shown of n = 3 separate experiments. Boxed regions are shown magnified as individual stills. Scale bars, 20 μm and 5 μm in magnified stills. (F) HUVECs were transiently transfected with Quad-K JAM-C–HRPout, fixed, and incubated with DAB and hydrogen peroxide for 30 min. Cells were then secondarily fixed, and 70-nm sections were prepared and imaged by transmission EM. Precipitated DAB can be clearly seen on early endosomes, but there is little evidence of labelling on MVBs (a, b, and c). Scale bar, 200 nm. Underlying data are found in S4 Data. a.a., amino acid; DAB, diaminobenzidine; EGFP, enhanced green fluorescent protein; EGFR, epidermal growth factor receptor; EM, electron microscopy; GFP, green fluorescent protein; HA, hemagglutinin; HRP, horseradish peroxidase; HUVEC, human umbilical vein endothelial cell; JAM-C, junctional adhesion molecule-C; MVB, multivesicular body; PDZ, postsynaptic density protein/drosophila disc large tumour suppressor/zonula occludens protein 1; QK, Quad-K; Quad-K, 4 lysine residues mutated to arginine residues; TM, transmembrane; Ub/Ubq, ubiquitin; WT, wild type.
Fig 6.
An APEX-2 proximity-labelling approach reveals potential JAM-C trafficking machinery.
(A) Schematic of the domain structure of the JAM-C–APEX-2 construct, showing the extracellular immunoglobulin domains (blue), the APEX-2 tag (green circle), and the intracellular PDZ interacting domain (clear circle). The APEX-2 tag, following feeding with biotin tyramide and incubation with hydrogen peroxide for 1 min, labels all neighbouring proteins within 20 nm on electron-rich amino acids (purple circle). (B–E) HUVECs were transfected with WT or Quad-K JAM-C–APEX-2. Cells were fed biotin tyramide for 30 min and then exposed to hydrogen peroxide for 1 min. The reaction was then quenched, and the cells were either (D) fixed and stained with streptavidin and analysed by confocal microscopy or (B, C, and E) lysed and biotinylated proteins pulled down using neutravidin beads for analysis by (B and C) SDS-PAGE and western blot for the presence of (B) JAM-C and tubulin or (C) biotinylated proteins or (E) following an on-bead tryptic digest, mass spectrometry. A heat map of n = 3 independent mass spectrometry data sets is shown, with white representing no signal and dark red a high signal. Each individual experiment was carried out in duplicate, with each mass spectrometry run being repeated twice (to give a total of 3 analyses/experiment). p-Values are given across all 3 experiments (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001; t test). (F and G) String analysis of proteins neighbouring (F) WT or (G) Quad-K JAM-C–APEX-2 proteins. *Lane removed of a JAM-C mutant not included in this analysis. APEX-2, ascorbate peroxidase 2; CBL, Casitas B-lineage lymphoma; ESCRT, endosomal sorting complex required for transport; Hrs, hepatocyte growth factor-regulated tyrosine kinase substrate; HUVEC, human umbilical vein endothelial cell; JAM-C, junctional adhesion molecule-C; KIF, kinesin-like protein; n/s, not significant; Quad-K, 4 lysine residues mutated to arginine residues; PDZ, postsynaptic density protein/drosophila disc large tumour suppressor/zonula occludens protein 1; siRNA, small interfering RNA; SNAP23, synaptosomal-associated protein 23; SNARE, soluble N-ethylmaleimide sensitive factor attachment protein receptor; STAM, signal transducing adapter molecule 1; TM, transmembrane; Tsg101, tumour susceptibility gene 101 protein; VPS, vacuolar protein sorting-associated protein; WASF2, Wiskott–Aldrich syndrome protein family member 2; WT, wild type.
Fig 7.
JAM-C is ubiquitylated by the E3 ligase CBL.
(A–G) HUVECs were transfected with 2 different CBL siRNAs over a 96-h period. (A) Cells were lysed and analysed by SDS-PAGE and western blot, detecting CBL and tubulin. Knockdown was typically ≥90%. (B–E) Untreated or CBL knockdown cells were transfected with JAM-C–GFPout and HA ubiquitin. Cells were lysed, and GFP-tagged proteins were pulled down using GFP trap agarose beads. Pulldown proteins were analysed by SDS-PAGE and western blots for the presence of GFP- and HA-tagged proteins. Quantification of (C) HA ubiquitin levels, (D) JAM-C–GFPout levels, or (E) ubiquitin levels normalised to levels of JAM-C. Results are from n = 3 experiments. Error bars represent SEM (*p ≤ 0.05; t test). (F) Mock and CBL knockdown cells were incubated with and without 100 nM bafilomycin for 4 h and then fixed and labelled for endogenous VE-Cadherin (red) and JAM-C (green). (G) Quantification of the number of JAM-C–positive vesicles present in mock and CBL knockdown cells treated with 100 nM bafilomycin (5 fields of view/condition). Results are shown normalised to the mean number of vesicles in the mock cells from n = 3 experiments. Error bars represent SD (**p ≤ 0.01; t test). Underlying data are found in S5 Data. CBL, Casitas B-lineage lymphoma; GFP, green fluorescent protein; HA, hemagglutinin; HUVEC, human umbilical vein endothelial cell; JAM-C, junctional adhesion molecule-C; kd, knockdown; SEM, standard error of the mean; siRNA, small interfering RNA; VE-Cadherin, vascular endothelial cadherin.
Fig 8.
Ubiquitylation of JAM-C is required for HUVEC migration.
(A–D) HUVECs were transfected with siRNA targeting luciferase or JAM-C over a 96-h period. The day after the second round of transfection, cells were additionally transduced with a GFP-, WT JAM-C-GFPout–, or a Quad-K JAM-C-GFPout–expressing lentivirus and allowed to reach confluency. (A) Cells were lysed, and the extent of JAM-C knockdown and rescue was determined by western blot relative to a tubulin loading control. The expressed WT and Quad-K constructs were similar to endogenous levels. (B–D) The monolayer was scratched, and the cells were allowed to recover over a 0- to 60-h time period whilst being imaged on a brightfield Olympus (Tokyo, Japan) microscope. (B) A representative experiment showing the percentage recovery of cells over a 16-h time period. Mock treated cells expressing the GFP construct recover relatively quickly. JAM-C siRNA-treated cells fail to recover when rescued by transduced Quad-K JAM-C–GFPout or GFP construct. However, almost complete recovery is apparent using transduced WT JAM-C–GFPout construct after 16 h. Error bars represent SD. (C) The percentage recovery of siRNA-treated GFP, WT JAM-C–GFP, or Quad-K JAM-C–GFP transduced cells at 16 h is shown normalised to the mock/GFP rescued condition. Data shown from n = 4 experiments, error bars represent SEM (*p ≤ 0.05, **p ≤ 0.01; unpaired t test). (D) Representative images of recovery of scratch wounds after 60 h. Scale bar, 200 μm. Underlying data are found in S6 Data. GFP, green fluorescent protein; HUVEC, human umbilical vein endothelial cell; JAM-C, junctional adhesion molecule-C; kd, knockdown; luc, luciferase; Quad-K, 4 lysine residues mutated to arginine residues; SEM, standard error of the mean; siRNA, small interfering RNA; WT, wild type.
Fig 9.
JAM-C is internalised from junctions near the front of migrating cells.
(A) JAM-C–GFPout was expressed in HUVECs, and the cells were allowed to form a confluent monolayer. The cell layer was scratched, and 45 min later, live-cell imaging was carried out with images acquired every 30 s for 45 min. The dotted line represents the edge of the scratch wound, asterisks highlight vesicles, arrows show holes that form between cells, and the boxed region is shown magnified. Scale bar, 20 μm. (B) Model of the role of JAM-C trafficking in endothelial cell migration. WT: (1) JAM-C is endocytosed from the junction either side of the leading edge and is localised to endosomes. (2) As JAM-C enters endosomes, the E3 ligase CBL ubiquitylates the cytoplasmic tail on lysine residues, allowing recruitment of the ESCRT complex and the formation of intraluminal vesicles. (3) JAM-C is degraded in the lysosomes, preventing recycling and regulation of integrins and/or abrogating signalling. (4) The remaining JAM-C that is not ubiquitylated is likely returned to the cell surface to reform new junctions after the cell has moved. Inhibition of JAM-C ubiquitylation: (5) JAM-C is endocytosed from the junction either side of the leading edge and is localised to endosomes. (6) Mutation of JAM-C or absence of CBL prevents ubiquitylation and recruitment to the intraluminal vesicles of late endosomes. (7) Failure to degrade JAM-C results in a failure of cell migration potentially due to mistargeting of JAM-C to the appropriate region of the cell junction, misregulation of integrin molecules, and/or abrogation of signalling. CBL, Casitas B-lineage lymphoma; ESCRT, endosomal sorting complex required for transport; GFP, green fluorescent protein; HUVEC, human umbilical vein endothelial cell; JAM-C, junctional adhesion molecule-C; WT, wild type.