Fig 1.
Two independent functional genomics screens for regulators of ECT uncover AQP3.
(A) CCF4 ECT assay schematic. (B) Representative flow cytometry plot of the sorting of WT BMDCs based on ECT+ or ECT- populations. (C) AQP3 was significantly enriched in the ECT+ population (false discovery rate adjusted p-value<0.001) according to RNA-Seq differential gene expression analysis. (D) GAL4-TA ECT assay schematic. (E) Representative histograms of the GAL4-UAS reporter cell line demonstrating ECT after incubation with 10 μg/ml GAL4-TA overnight. (F) BMDCs expressing the GAL4-UAS reporter cell line were treated with various shRNAs followed by incubation with GAL4-TA overnight. Mean fluorescence intensity is shown. 2 independent experiments were performed. **P<0.01, two-tailed t-test.
Fig 2.
AQP3 is uniquely positioned among aquaporin family members to regulate ECT.
(A) HEK293 cells were transfected with various aquaporin overexpression constructs, flow cytometry-sorted for equivalent expression, and ECT efficiency (described in Methods) was evaluated with 100 μg/ml β-lactamase. Mean ± SEM is shown. (B) HEK293 cells were co-transfected with fluorescently-tagged AQP3 and AQP9 expression constructs and 3 days later, fixed and imaged by confocal microscopy. Scale bar: 2 μm. (C) HEK293 cells stably expressing fluorescently-tagged AQP3 were incubated with 1 μm latex beads for 2 hours prior to fixation and analysis by immuno-electron microscopy. 3 representative images are shown. Arrows point to phagosomal membrane. E: Endosome. P: Phagosome/latex bead. M: Mitochondria. Scale bar for top left image: 500 nm. Scale bar for top right and bottom left images: 200 nm. (D) On day 6 of culture, WT or AQP3-/- BMDCS were fixed in 4% PFA and permeabilized with 0.05% saponin, followed by detection with anti-AQP3. AQP3 can be seen on or near the plasma membrane and dotting the interior of the cell. Scale bar: 2 μm. (E) HEK293 cells were transfected with fluorescently-tagged AQP3 or AQP3 2xLLmut. 3 days later, cells were incubated with fluorescently-labeled OVA-coated 1 μm beads for 1 hour prior to cell homogenization and analysis by flow cytometry. (F) AQP3 2xLLmut does not localize to phagosomes. Stably-transfected HEK293s were incubated with 100:1 (beads:cells) ratio of fluorescently-labeled OVA-coated 1 um tosylactivated Dynabeads and live imaging was initiated immediately. Scale bar: 1 μm. 2–4 independent experiments were performed.
Fig 3.
Cytosolic ROS is important for ECT.
(A) ECT was evaluated in WT BMDCs with 100 μg/ml β-lactamase in the presence or absence of 100 μM H2O2. (B) TRIF-/- or WT BMDCs were loaded with the ROS sensor CM-H2DCFDA followed by incubation with 250 μg/ml dextran, 100 μg/ml each PGN and LPS, or 250 μg/ml β-lactamase, in the presence or absence of 1 μM TTFA. (C) MitoSOX Red-loaded BMDCs were incubated with 250 μg/ml β-lactamase. MedFI is displayed. (D) ECT was evaluated in WT BMDCs with 250 μg/ml β-lactamase in the presence of 1 μM TTFA or equivalent volume DMSO. (E) ECT was assessed in TRIF-/- or WT BMDCs with 100 μg/ml β-lactamase. (F) HEK293 cells were transfected with the indicated constructs and phagosomal H2O2 was evaluated with rHyPer in the presence or absence of 100 μM H2O2. Mean ± SEM is shown. *P<0.05, two-tailed t-test. 2–4 independent experiments were performed.
Fig 4.
AQP3 regulates phagosomal lipid peroxidation.
(A) AQP3-/- or WT BMDCs were co-incubated with fluorescently-labeled OVA-coated 1 μm beads and a C11-bodipy lipid peroxidation indicator for 1.5 hours, homogenized, and phagosomes were analyzed by flow cytometry. (B) Annotated volcano plot of ECT functional genomics screen highlights the connection between ECT and ROS generation and transport, which could potentiate endosomal lipid peroxidation and antigen release. (C) ECT was evaluated in WT BMDCs with 0.5 mg/ml β-lactamase in media that did or did not contain ISCOMATRIX™ adjuvant. (D) WT BMDCs were incubated with 100 mg/ml OVA in media that did or did not contain ISCOMATRIX™ adjuvant, in the presence or absence of 1 μM epoxomicin, followed by co-incubation with CFSE-labeled OT-I CD8+ T cells for 64 hours. The number of CD8+ T cells that had undergone division as a measure of cross presentation was evaluated by flow cytometry. (E) DHFR activity of recombinant β-lactamase-DHFR fusion protein was measured after incubation with 500 nM MTX or equivalent volume PBS. (F) The ability of increasing amounts of recombinant β-lactamase-DHFR fusion protein to undergo ECT was evaluated in WT BMDCs following treatment with 500 nM MTX or equivalent volume PBS. *P<0.05, *P<0.01, two-tailed t-test. 2–3 independent experiments were performed.
Fig 5.
AQP3 is involved in cross presentation efficiency in vitro and in vivo.
(A) ECT was evaluated in AQP3-/- or WT BMDCs with 50 μg/ml β-lactamase. (B-C) Splenic CD11c+ DCs from AQP3-/- or WT mice were isolated and ECT was evaluated in the CD11c+CD8+ population (B) or the CD11c+XCR1+ population (C) with 0.5 mg/ml β-lactamase. (D) BMDCs overexpressing AQP3 by viral transduction were incubated with increasing amounts of OVA or SIINFEKL peptide followed by co-incubation with CFSE-labeled OT-I CD8+ T cells. (E) BMDCs overexpressing AQP3 by viral transduction were incubated with 1 μg/ml αDEC205-OVA followed by co-incubation with CFSE-labeled OT-I CD8+ T cells. (F) BMDCs were incubated with 250 μg/ml OVA followed by co-incubation with CFSE-labeled OT-II CD4+ T cells. (G) ECT was evaluated in BMDCs expressing an AQP3 shRNA knockdown construct by viral transduction with 100 μg/ml β-lactamase. AQP3 shRNA knockdown was calculated to be 93% effective as measured by real-time PCR. (H) AQP3 KD cells were incubated with 1 μg/ml αDEC205-OVA followed by co-incubation with CFSE-labeled OT-I CD8+ T cells. (I,J) AQP3-/- and WT control mice were infected with LCMV Armstrong. 8 days later, viral titer in kidney was assessed (I) and splenocytes were stimulated with LCMV NP peptide to determine the generation of LCMV-specific CD8+ T cell clones (J). Each dot represents an individual mouse. (K) AQP3-/- and WT control mice were injected with rat HER-2/neu extended peptide and adjuvant. 1 week later, splenocytes were harvested and the generation of antigen-specific CD8+ T cell clones was evaluated using a HER-2/neu tetramer. Mean ± SEM is shown. *P<0.05, **P<0.01, two-tailed t-test or Mann-Whitney U test. 2–3 independent experiments were performed.