Fig 1.
Mitochondrial localization of bacterialβ-barrel proteins. A, HeLa cells constitutively expressing GFP, FLAG-OmpA from Simkania negevensis (Sn), FLAG-OmpC from E. coli, or FLAG-OmpA from Ctr were fractionated, and mitochondrial and cytosolic fractions were subjected to anti-FLAG or anti-GFP Western blotting to analyze the subcellular localization of GFP, OmpC or OmpA. TOM20 and BAK were used as mitochondrial markers, α-tubulin as a marker of the cytosol. B, SnOmpA, E. coli OmpC and Ctr OmpA were ectopically expressed in HeLa cells. Cells were treated with ABT-737 (1 μM) and S63845 (500 nM) for 4h. Cells were fixed, permeabilized and stained for active caspase-3 to measure the number of apoptotic cells. Data are representative of three independent experiments. Error bars represent SEM and significance was tested using 2-way ANOVA (***, p < 0.001, ****, p < 0.0001).
Fig 2.
Interaction between OmpA and BAK on mitochondria.
A, HeLa EGFP expressing cells and HeLa OmpA expressing cells were treated with ABT-737 (1 µM) and S63845 (500 nM) for 3h in the presence of the caspase-inhibitor QVD-OPh (10 μM; to block cell death downstream of mitochondria). BAK was immunoprecipitated with an antibody recognizing active and inactive BAK. Proteins were run on SDS-PAGE, and BAK and OmpA were detected by immunoblotting. Aliquots of the input and IP-reactions were loaded separately. Right lane shows beads with no antibody. Data are representative of three independent experiments. B, control HeLa cells (upper panel) and Tet-OmpA HeLa cells (carrying a tetracycline/AHT-inducible OmpA [12], bottom panel) were seeded on cover slips. 48h post-stimulation with AHT, cells were treated with ABT-737 (1 μM) and S63845 (500 nM) for 4h in the presence of the caspase-inhibitor QVD-OPh (10 μM). Cells were fixed, permeabilized and processed for PLA with antibodies against BAK (aa23-38; active BAK) and OmpA (red). Mitochondria were labeled using antibodies directed against TOM22 (green), and DNA was stained with Hoechst dye (blue). Confocal microscopy was performed and the overlay shows the co-localization of the PLA-signal with the mitochondrial protein TOM22. Images were acquired under identical conditions and exposure times. Data are representative of three independent experiments. Scaling bar, 10 μm. A control experiment with BAK-deficient cells is shown in S2 Fig. C, same set-up as in B. HeLa cells carrying a tetracycline-inducible OmpA were incubated with 100 nM AHT for 48 h to induce OmpA expression. Cells were fixed, permeabilized and processed for PLA using a different antibody specific for BAK (Ab-1(TC-100; active BAK)) and OmpA (red). Mitochondria were labeled using antibodies directed against TOM20 (green), and DNA was stained with Hoechst dye (blue). Cells were imaged by confocal microscopy. Stacks were then processed with the deconvolution software AutoQuantX and 3D reconstruction analysis was performed using the Imaris software. Images were acquired under identical conditions and exposure times. Left panel represents a single Z-slice, right panels indicate 3D reconstructions of the indicated areas using multiple Z-slices. Data are representative of three independent experiments. Scaling bar, 10 μm.
Fig 3.
OmpA is detected at mitochondria during Ctr infection.
A, 24 h after Ctr-infection, HeLa cells on cover slips were fixed, permeabilized and stained for two-color 3D SMLM super-resolution images of mitochondria and Ctr using antibodies against TOM20 and OmpA (left panel). The panels in the middle and to the right show magnified views of the pink boxed region. The x, y, and z- dimensions of the magnified region are 830 nm x 1490 nm x 780 nm. Scale bar: 1 μm. Note the localization of OmpA (red) on mitochondria TOM20 (green) during Ctr-infection. Insets show different perspectives of the pink boxed region. Also see S1 and S2 Movies. B, immuno-gold electron microscopy identifies OmpA outside chlamydial inclusions and on mitochondria. HeLa cells were infected with Ctr and then fixed and stained for OmpA using immuno-gold particles and were imaged using electron microscopy as described in the methods. Yellow asterisks represent mitochondria, yellow arrows indicate cytoskeletal structures and blue arrows show gold particle labelling of OmpA on mitochondrial outer membranes. I (inclusion). Note the dense labeling of the bacteria as well as the labeling of mitochondrial membranes. Scale bar references are indicated. Data represents one experiment. C, BAX-deficient HeLa cells were seeded on cover slips. 24 h post-infection (MOI = 5), cells were treated with ABT-737 (1 μM) and S63845 (500 nM) for 4 h in the presence of the caspase inhibitor QVD-OPh (10 μM). Cells were fixed, permeabilized and processed for PLA using antibodies against BAK (Ab-1(TC-100); active BAK) and OmpA. The red signal (PLA) indicates close proximity of BAK and OmpA. Mitochondria were stained using antibodies directed against TOM20 (green), and DNA was stained with Hoechst dye (blue). Cells were imaged by confocal microscopy. Stacks were then processed with the deconvolution software. AutoQuantX and 3D analysis was performed using the Imaris software. Left panel represents a single Z-slice, right panels indicate 3D reconstructions of the indicated areas using multiple Z-slices. Data are representative of three independent experiments. Scale bar, 5 μm. D, BAX-deficient HeLa cells were either Ctr-infected (MOI = 5) or mock-infected. 24 h post-infection, cells were fractionated, and mitochondria were purified from heavy membrane fractions using magnetic beads labelled with anti-TOM20 antibodies. Purified mitochondria were subjected to sodium carbonate extraction (pH 11.5) to separate integral from attached membrane proteins. Membranes were pelleted, and the fractions were run on SDS-PAGE. Mitochondrial membranes and membrane-integrated proteins are found in the pellet fractions. Proteins were detected using the indicated antibodies. Chlamydial HSP60, BiP (ER), golgin-84 (golgi-apparatus), lamin B (nuclear envelope) and Rab7 (endosomes) were used as organelle markers. Release of SMAC shows extraction efficiency. Data are representative of three independent experiments.
Fig 4.
OmpA resides in membrane vesicles.
A, electron microscopy was performed on chlamydial derived membrane vesicles, which were isolated from Ctr-infected HeLa cells. B, CDVs were isolated from Ctr-infected (MOI = 5, 48h) HeLa cells. Uninfected cells were subjected to the same procedure. Isolated CDV-containing fractions were subjected to Western blotting. OmpA and IncA but not the chlamydial RNA polymerase B (RpoB, as a marker for bacteria) was seen in vesicles purified from Ctr-infected HeLa cells. VDAC was used to detect mitochondrial contamination. BiP was used to detect ER. GAPDH was used as a loading control. Data are representative of three independent experiments. C, Nanoflow cytometry (nFCM) analysis shows CDVs and control vesicles that were stained by immunofluorescence staining for specific surface markers such as OmpA, LPS and EV-derived host cell marker CD9. D, Histograms from Nanoflow cytometry (nFCM) analysis showing size distribution of CDVs and control vesicles stained by immunofluorescence staining for specific surface markers such as OmpA, LPS and EV-derived host cell marker CD9.
Fig 5.
CDVs colocalize with mitochondria.
A, HeLa cells were incubated with CDVs or control vesicles for 6h. Cells were fractionated, and mitochondria were purified from heavy membrane fractions using magnetic beads labelled with anti-TOM20 antibodies. Purified mitochondria and the fractions were run on SDS-PAGE. Proteins were detected using the indicated antibodies. Chlamydial OmpA, BiP (ER), golgin-84 (Golgi-apparatus), lamin B (nuclear envelope) and RAB7 (endosomes) were used as organelle markers. SMAC shows intact mitochondria. TOM20 (mitochondria). B, HeLa cells were seeded on cover slips. Cells were incubated with chlamydia derived vesicles or fractions isolated from uninfected control cells for 4h. Cells were fixed and permeabilized. Mitochondria were labeled using antibodies directed against TOM20 (red, outer mitochondrial membrane), OmpA (green), and DNA was stained with DAPI (blue). Confocal microscopy was performed with Zeiss LSM 880 and image analysis was performed using the Imaris software. 3D reconstructions of the indicated area using Z-slices were performed with Imaris software and the overlay shows the co-localization of the chlamydial vesicles with the mitochondrial protein TOM20. OmpA signal colored in pink indicates CDVs that are in contact or overlap with mitochondria. OmpA signal in yellow indicates no direct contact to mitochondria. Images were acquired under identical conditions and exposure times. Data are representative of three independent experiments. Scale bar, 8 μm. C, CDVs were labelled with rhodamine B chloride (R-18) (red) and were added to uninfected HeLa cells. Live cell imaging was performed for 50min on the confocal microscope Zeiss LSM 880. Mitotracker green was used to stain mitochondria. Graph displays fusion events of the R-18-labelled CDVs with host cell membrane over time. Scale bar, 10 μm. Data are means/SEM from three independent experiments. Control CDVs refers to fractions isolated from uninfected HeLa cells.
Fig 6.
CDVs deliver OmpA to BAK and inhibit apoptosis.
A, HeLa cells were seeded on cover slips. Cells were incubated with either chlamydial vesicles or preparations from uninfected HeLa cells (CDV control) for 1h. Cells were then treated with ABT-737 (1 μM) and S63845 (500 nM) for 3h in the presence of the caspase-inhibitor QVD-OPh (10 μM). Cells were fixed, permeabilized and processed for PLA using antibodies against active BAK (Ab-1(TC-100)) and OmpA (red) (Biozol Diagnostica, #LS-C79219). Mitochondria were labeled using antibodies directed against TOM20 (green) and DNA was stained with DAPI (blue). Confocal microscopy was performed and the overlay shows the co-localization of the PLA signal with the mitochondrial protein TOM20. Images were acquired under identical conditions and exposure times with the confocal microscope Zeiss LSM 880. BAK-deficient cells were used as a specificity control. Scale bar, 10 μm. The diagram shows the quantification from three independent experiments (columns are mean/SEM). Significance was calculated using the Kolmogorov–Smirnov test (**, p < 0.01). B, HeLa cells were treated with CDVs or control vesicles for 2h. Cytosolic and mitochondria-containing (heavy membrane) fractions were separated and Western blot analysis was performed. OmpA is found on the heavy membrane fraction on HeLa cells treated with CDVs. BAX retro-translocation to the cytosol is enhanced upon infection. Columns/error bars give means/SEM of four individual experiments (individual results are shown as symbols). GAPDH was used for normalization. Lines connect results of the experiments. Data are representative of four independent experiments. C. HeLa cells (control or BAK-deficient cells) were treated with CDVs or fractions isolated from uninfected cells (control vesicles) for 1h. Cells were then treated with ABT-737 (1 μM) and S63845 (500 nM) for 3h. Cells were fixed, permeabilized and active caspase-3 staining was performed to measure the number of apoptotic cells. Data are means/SEM of three individual experiments. Significance was tested using 2-way ANOVA (****, p < 0.0001).