Figure 1.
An extracellular fragment of CAR is detected from conditioned media of human glioma U87-MG cells and mouse embryonic hippocampal neurons.
(A) A Western blot of conditioned media and cell lysates from U87 LNCX (control) and U87 CAR cells using the anti-CAR N-terminus antibody 2240. The shed extracellular fragment detected from conditioned media migrates at approximately 32 kDa, while full-length CAR detected from cell lysates migrates at approximately 50 kDa. (B) A Western blot of conditioned media from embryonic hippocampal neurons (3 days in vitro), using the anti-CAR N-terminus antibody 2240, revealed the presence of a 32 kDa fragment, similarly to the U87-MG CAR-expressing cell line. Also shown is a Western blot of full-length CAR detected from neuronal lysate using the anti-CAR C-terminus antibody RP291.
Figure 2.
CAR shedding is stimulated by the calcium ionophore ionomycin and the phorbol ester PMA.
(A) U87 CAR cells were treated with the calcium ionophore ionomycin at the indicated concentrations for 30 minutes. Ionomycin treatment stimulated the ECD shedding of CAR. (B) PMA treatment (25 ng/ml) of U87 CAR cells did not trigger CAR ECD shedding within one hour. At this concentration, 4 hours of PMA treatment led to robust CAR ECD shedding, but remained lower than constitutive shedding over 16 hours. Volumes of conditioned media loaded on SDS-PAGE were adjusted according to lysate protein concentrations. (C) The protein kinase C (PKC) inhibitor Gö 6983 decreased PMA-stimulated shedding of CAR ECD into conditioned media in a dose-dependent manner. PMA was used at a final concentration of 1 μM, and Gö 6983 at the following concentrations: + = 1 nM; ++ = 10 nM, and +++ = 100 nM. For the Western blots shown in these panels, the anti-CAR N-terminus antibodies 2239 or 2240 were used.
Figure 3.
CAR shedding is mediated by metalloproteases.
(A) U87 CAR cells plated on poly-L-lysine coated plates were pre-incubated for 45 minutes with a variety of protease inhibitors (10 µM pepstatin A, 10 µM leupeptin, 10 µM E64, 250 µM O-phenanthroline, 25 µM TAPI-1) followed by 3 hours of treatment with 1 µM of PMA. None of the treatments were toxic to the cells under these conditions and concentrations of inhibitors. CAR ECD released into conditioned media was detected via Western blot using anti-CAR N-term. antibody 2240. The broad-spectrum metalloprotease inhibitors TAPI-1 and O-phenathroline decreased PMA-stimulated CAR ECD shedding, while the aspartyl protease inhibitor pepstatin, the cysteine protease inhibitor E64, and the cysteine/serine protease inhibitor leupeptin had no effect. Also shown are Western blots of full-length CAR from the corresponding cell lysates (anti-CAR C-term. antibody RP291). (B) U87 CAR cells were treated with PMA (1 µM) or DMSO vehicle, in the presence of 25 µM of the broad spectrum metalloprotease inhibitor GM6001 or its negative control. GM6001, but not its negative control, inhibited PMA-stimulated shedding of CAR ECD. (C) U87 CAR cells were incubated for 3 hours with 1 µM PMA along with 10 µg/ml of TIMPs 1, 2 or 3. TIMP1 and TIMP3, but not TIMP2, decreased PMA-mediated ECD shedding of CAR, suggesting that ADAM10 may be a sheddase. For the Western blots shown in these panels, the anti-CAR N-terminus antibodies 2239 or 2240 were used.
Figure 4.
ADAM10 is involved in CAR ECD shedding.
(A) U251N cells stably expressing CAR were transfected with empty plasmid (mock) or ADAM10 plasmid. 24 hours after transfection, cells were washed and incubated in opti-MEM for 24 hours, and conditioned media and cell lysates were analyzed by Western blots using anti-CAR N-terminus antibody (2239). Overexpression of ADAM10 increased constitutive CAR ECD shedding. (B) U87 CAR cells were treated with 10 μM purified ADAM10 prodomain (versus an equivalent volume of buffer as a control), and conditioned media and cell lysates were collected as previously described. A Western blot for CAR extracellular domain (2240 antibody) shows that the prodomain of ADAM10, which inhibits ADAM10 activity, decreased CAR ECD shedding.
Figure 5.
shRNA knockdown of ADAM10 decreases constitutive, PMA-mediated and ionomycin-mediated CAR ECD shedding.
(A) A Western blot for ADAM10 using cell lysates of U87 CAR cells containing either control shRNA or ADAM10 (#6676) shRNA, in biological duplicates. Equal amounts of proteins were loaded on SDS-PAGE gel. Anti-GAPDH antibody was used as a loading control. Quantification of mean ADAM10 band intensities normalized over GAPDH revealed a decrease of approximately 60% with anti-ADAM10 shRNA compared to control shRNA. (B) A constitutive shedding experiment was performed using U87 CAR stable cell lines containing either control shRNA or anti-ADAM10 shRNA (#6676). Conditioned media and cell lysates were collected after 24 hours of incubation of cells in opti-MEM, and Western blotting was done using the anti-CAR N-terminus antibody 2239. With anti-ADAM10 shRNA, there was a significant decrease (40%) in the levels of shed CAR. Results from 4 independent experiments performed in duplicates were quantified (unpaired t-test; p=0.0004 (***)). (C) U87 CAR cells containing either control shRNA or ADAM10 shRNA (#6676) were treated with 1 µM PMA. Conditioned media and cell lysates were collected after 3 hours, and Western blots were performed using the anti-CAR N-terminus antibody 2239. With shRNA knockdown of ADAM10, there was a significant decrease of 41% in levels of shed CAR compared to control shRNA. Results from 3 independent experiments performed in duplicates were quantified (one-way ANOVA with Tukey’s multiple comparison test; *** = p < 0.001). (D) U87 CAR cells containing either control shRNA or ADAM10 shRNA (#6675) were treated with 1.5 µM ionomycin (vs. DMSO vehicle). Conditioned media and cell lysates were collected after 30 minutes of treatment, and Western blots were performed using the anti-CAR N-terminus antibody 2240. With shRNA knockdown of ADAM10, there was a significant decrease of 52% in levels of shed CAR with ionomycin treatment compared to control shRNA. Results from 3 independent experiments (n=3 per group) were quantified (one-way ANOVA with Bonferroni’s multiple comparison test; * = p < 0.05, ** = p < 0.01, *** = p < 0.001).
Figure 6.
Partial rescue of shRNA-mediated loss of CAR shedding using an shRNA-resistant mutant of ADAM10.
(A) shRNA stable cell lines of U251N CAR cells were transfected with either empty plasmid (mock) or an shRNA-resistant construct of ADAM10. 18-24 hours post-transfection, cells were washed and incubated in opti-MEM. Conditioned media and cell lysates were collected 24 hours later and analyzed by Western blot for CAR extracellular domain (anti-CAR antibody 2239). The shRNA-resistant ADAM10 mutant partially rescued CAR shedding in the ADAM10 (6676) shRNA cell line, compared to transfecting this cell line with empty plasmid (third and fourth Western blot bands from the left). (B) The band intensities of shed CAR detected by Western blot were quantified from 4 independent experiments (One-way ANOVA with Newman-Keuls multiple comparison test; * = p < 0.05, ** = p < 0.01, *** = p < 0.001).
Figure 7.
Confirmation of the area of ECD cleavage on CAR in HEK 293 cells.
(A) A schematic showing putative ADAM10 cleavage sites on CAR’s extracellular domain (arrows), as obtained from in vitro peptide digestion and mass spectrometry (Figure S4). (B) HEK 293 stable cell populations were generated to express wild-type CAR, the mutant MLRLAAAA or the mutant Δ221-232. The MLRLAAAA mutant did not shed into conditioned media of HEK 293 cells, while the Δ221-232 mutant shed its ECD similarly to wild-type CAR. Cell surface biotinylation experiments revealed that wild-type CAR and the two mutants are expressed similarly on the cell surface. Note that the low level of endogenous CAR in HEK 293 was detected after enrichment of cell surface biotinylated proteins (mock lane). Western blotting was performed using the anti-CAR N-terminus antibody 2240.
Figure 8.
CAR undergoes RIP by the γ-secretase complex.
(A) A Western blot for CAR’s intracellular domain (antibody RP291) shows lower molecular weight fragments in U87 CAR lysates at approximately 20 kDa and 14 kDa (CTF1 and CTF2, respectively). Full-length CAR (CAR FL) migrates at 50 kDa. (B) U87 CAR cells were treated with the γ-secretase inhibitors MG132 or DAPT for 16 hours. Equal amounts of proteins from cell lysates were analyzed by Western blot using the anti-CAR intracellular domain antibody RP291. Drug treatments resulted in accumulation of CAR CTF1 (20 kDa) and a decrease in CAR CTF2 (14 kDa) levels in a dose-dependent manner. (C) Overnight treatment of U87 CAR cells with Compound E, another inhibitor of the γ-secretase complex, at the indicated concentrations also diminished RIP of CAR. Equal quantities of proteins from cell lysates were analyzed by SDS-PAGE and Western blot using the anti-CAR intracellular domain antibody RP291. With increasing concentrations of Compound E, there was a corresponding accumulation of a 20 kDa fragment (CAR CTF1) and a decrease in a 14 kDa fragment (CAR CTF2). 3 different exposure times of the Western blot are shown. (D) A model of CAR proteolysis, with molecular weights of the resulting fragments indicated. Cleavage of CAR by ADAM10 (represented by the top pair of scissors) releases a 32 kDa fragment (CAR ECD) into the extracellular environment. The remaining 20 kDa fragment (CTF1) is processed by the γ-secretase complex (represented by the lower pair of scissors), generating a 14 kDa fragment (CTF2).
Figure 9.
Generation of CAR CTF1 precedes CTF2 production.
(A) Stable cell lines of U87-MG and U251N expressing either the V5 tag alone (mock) or CAR with a C-terminal V5 tag were generated. Equal amounts of cell lysates were analyzed by SDS-PAGE and Western blot using a mouse monoclonal antibody raised against the V5 tag. Full-length CAR, CTF1 and CTF2 were detected similarly to lysates of U87 CAR cells probed with anti-CAR C-term. antibody RP291 (Figure 8A). (B) U251N V5 and U251N CAR-V5 cells were treated overnight with MG132 (25 µM) or DMSO vehicle control. Equal amounts of proteins from cell lysates were used for anti-V5 Western blots. In the case of the MG132-treated cells, CTF1 levels accumulated while CTF2 nearly disappeared, similar to previous experiments with the U87 CAR cell line (Figure 8B). (C) MEF wild-type (MEF WT) or PS 1- and 2-knockout MEF cells (MEF PS1/2 KO) were infected with lentivirus to express full-length CAR with a C-terminal V5 tag. Cells were lysed 3 days post-infection and lysates were analyzed by Western blot using antibody raised against the V5 tag. MEF WT cells, but not MEF PS1/2 KO cells, contained CAR CTF2, indicating that presenilin is required for generation of the 14 kDa CTF2 fragment of CAR. (D) Verification of knockdown in ADAM10 expression in U87 CAR-V5 cells using ADAM10 shRNA (#6675); shown are anti-ADAM10 and anti-GAPDH Western blots. Lysates were also analyzed by Western blot using antibody raised against the V5 tag. The ADAM10 shRNA stable cell line had a decreased CAR CTF1 level, as expected. CAR CTF2 levels also decreased, indicating that shedding is a prerequisite for RIP of CAR.
Figure 10.
CAR’s intracellular domain (ICD) enters the nucleus.
Immunofluorescence and confocal microscopy images showing the presence of CAR ICD in nuclei of U87-MG cells. U87-MG cells were transiently transfected with empty pcDNA3.1 V5/His plasmid, full-length CAR-V5 plasmid or with CAR ICD-V5 plasmid. Immunofluorescence staining was performed 24-48 hours post-transfection using anti-V5 tag antibody and Alexa Fluor 555 secondary antibody (red). Nuclei were stained with DRAQ5 (blue). Images were acquired with a confocal microscope (63x oil objective). Images are representative of at least 3 independent experiments. Scale bars: 5 µm.