Figure 1.
PorB and BH3-only protein-induced signaling pathways cooperate to activate caspases.
(A) FLAG-tagged porins from N. gonorrhoeae (PorBNgo-FLAG) and commensal N. mucosa (PorBNmu-FLAG) were expressed in HeLa cells. Shown are the overlays of FLAG-tag (green), and cytochrome c (blue). The mitochondria with membrane potential were detected using MitoTracker staining (red). 25 µM BH3-mimetic compound BH3I-2 (Calbiochem) was added 24 h post transfection where indicated. (B) BH3I-2- or TNFα/cycloheximide (TNFa/CHX)-treated cells were subjected to immunoprecipitation with conformation specific antibodies recognizing activated Bak or Bax. Graph depicts quantities of precipitated Bak, Bax and actin (loading control), which were detected by western blot. Data are mean±SD of three independent experiments. (C) Caspase-3 activation was detected by western blot analysis upon treatment with BH3I-2 as in A. Cells were treated with either protonophore CCCP (Sigma) or the complex III inhibitor antimycin A (Sigma) at 1 µM as positive controls. Blot is representative of three independent experiments.
Figure 2.
PorB-induced rearrangement of the inner mitochondrial membrane.
(A) HeLa cells infected in the presence or absence of zVAD were analyzed by EM to detect the structural changes in the mitochondria. Representative sections of control, infected (Ngo) and infected, zVAD treated cells (Ngo+zVAD) are shown (upper panel). The mitochondria were viewed at higher magnification to analyze alterations in the cristae architecture (lower panel). (B) A total of 500 mitochondria were counted from cells infected and pretreated with zVAD and analyzed for condensation and loss of cristae. The graph shows the relative number of mitochondria with the indicated phenotypes of all analyzed mitochondria. Data are mean±SD of three independent experiments. (C) Transfection of porB leads to loss of cristae and matrix condensation. FLAG-tagged PorB was detected by immunogold labeling with a primary FLAG antibody and a secondary antibody linked to 12 nm gold particles (filled arrow head). Endogenous protein of the OMM (Tom20) was co-stained with the respective antisera and a secondary antibody linked to 6 nm gold particles (open arrow head). (D) Normal mitochondria or mitochondria with condensed matrix and disrupted cristae were quantified in PorB-expressing and control cells and expressed as % of total mitochondria. Data are mean±SD of three different experiments.
Figure 3.
Import of PorB into isolated mitochondria.
(A) Mitochondria were pretreated with 20 µg/ml of trypsin in the presence (−Trypsin) or absence (+Trypsin) of 600 µg/ml Soybean trypsin inhibitor (SBTI). In the latter (+Trypsin) trypsin was subsequently inactivated by addition of 600 µg/ml SBTI. Treated mitochondria were incubated with a radiolabelled PorB precursor for indicated times followed by carbonate extraction at pH 11.5. The graph depicts mean±SD of three independent import experiments. Maximum PorB import time into non- Trypsin treated mitochondria was set at 100%. (B,C,D) Mitochondria were isolated from tom40kd-2, sam50kd-2 and mtx2kd-2 cell lines grown in the absence (−Dox) or the presence (+Dox) of doxycycline for 7 days, except for tom40kd-2 cells which were treated with Dox for 5 days. After the import of radiolabelled PorB for indicated times, mitochondria were subjected to carbonate extraction at pH 11.5 and SDS-PAGE. Graph depicts mean±SD of three independent experiments. Maximum PorB import time into mitochondria in cells not treated with Dox was set at 100%. Immunoblots demonstrate the levels of import factors in the mitochondria used for import assays and are representative of three independent experiments. (E,F) Import of 35S-labelled N. gonorrhoeae PorB into mitochondria isolated from the yeast mutant strain sam50-1, Δsam37 strain and from the corresponding wildtype strains. The amount of imported PorB was normalized against the total amount of PorB in the import reaction. Graph depicts mean±SD of four independent experiments.
Figure 4.
Association of PorB with the mitochondrial inner membrane.
(A) Sucrose density gradient. 35S-labelled PorB, together with the marker proteins VDAC (outer membrane) and F1γ (the ATP-synthase γ-subunit, inner membrane), were synthesized in reticulocyte lysate and imported into isolated yeast mitochondria. Following treatment with proteinase K, membrane vesicles were prepared by sonication and separated on a linear sucrose gradient (0.85–1.6 M sucrose). The fractions were analyzed by SDS-PAGE and a phosphorimager. The highest value for each protein was set at 100% (control). (B) Localization of PorB at the IMM in HeLa cells. HeLa cells were transfected with an expression construct of FLAG-tagged PorB. PorB was detected in porB-transfected cells by immunogold labelling using an antibody directed against the FLAG-tag and a secondary antibody linked to 12 nm gold particles. Endogenous proteins of the OMM (Tom22) and IMM (Tim23) were co-stained with the respective antisera and a secondary antibody linked to 6 nm gold particles. Arrows represent different features of mitochondria (outer membrane, inner membrane and matrix). Arrowheads show 12 nm (closed arrowheads) or 6 nm (open arrowheads) gold particles on electronmicrographies. (C) Quantification of (B). Gold particles were counted to determine the intramitochondrial localization of PorB, Tom22 and Tim23. The actual number of counted gold particles is indicated above the individual bars.
Figure 5.
Electrophysiological characterization of PorB.
(A) Typical (n≥5 independent bilayers) current recordings of a bilayer containing PorB at indicated applied voltages. Buffer conditions were symmetrical with 250 mM KCl, 10 mM Mops/Tris pH 7.0 in cis and trans. (B) PorB is regulated by ATP. Current-voltage relationship of PorB in the absence or presence of ATP added at the indicated side of the membrane. Buffer conditions were as in (A) (n≥3). (C) Voltage-dependent open probability of PorB in the absence or presence of ATP added to the indicated side of the membrane. Quantification was performed by comparing the mean current determined over a time range of 2 min with the maximum current at a constant holding potential (n≥3). (D) Current recordings of a bilayer containing PorB in the presence of ATP at the indicated constant holding potential. Buffer conditions were as in (A) (n≥5). (E) HeLa cells transfected with empty vector and porB were loaded with calcein-AM (green) and CoCl2 as mentioned in Protocol S1. The cells were then incubated with MitoTracker orange (red) to check for membrane potential loss, and the expression of PorB (blue) was checked by costaining. The respective overlays of all the three channels are shown. (F) The loss of calcein-AM staining upon Ngo infection was monitored by FACS analysis. Data represent mean±SD of three independent experiments.
Figure 6.
PorBK98Q channels are not modulated by ATP.
(A) PorBK98Q is impaired in ATP binding. Neisserial strain N920 (PorBNgo), N886 (PorBK98Q) and N923 (PorBNmu) were subjected to ATP binding assay as described in Materials and Methods. Bacterial lysates were separated by SDS-PAGE, and labelled PorB was visualized by exposing the dried gel to X-Ray films. The lower panel represents PorB as visualized by coomassie staining. For quantification purposes, the amount of ATP bound by the strain N920 was taken to be 100%. (B,C) Typical (n≥5 independent bilayers) current recordings of a bilayer containing PorBK98Q at 120 mV and −120 mV. Buffer conditions were symmetric with 250 mM KCl, 10 mM Mops/Tris pH 7.0 in cis and trans. 1 mM ATP was added in cis and trans (C). (D,E) Voltage-dependent open probability of PorBK98Q in the absence (D) and presence (E) of ATP added to both sides of the membrane (n≥3). Data are mean±SD.
Figure 7.
ATP binding is crucial for the PorB-induced ΔΨm loss and apoptosis.
(A) PorBK98Q is impaired in inducing ΔΨm loss. HeLa cells were infected with the indicated strains N920 (containing wild type PorBNgo) and N886 (containing ATP-binding mutant PorBK98Q), and the loss of ΔΨm was measured using TMRE staining and FACS analysis. The data are expressed as the relative number of cells with low TMRE staining from total cells. Data are mean±SD of 3 independent experiments. (B) HeLa cells were infected with the indicated strains and ΔΨm was monitored by MitoTracker staining. Phase contrast images demonstrate the strong infection of the HeLa cells. Arrows in the section of magnified phase contrast images point to bacteria. (C) HeLa cells were infected with the indicated strains, fixed and stained with 1 µg/ml Hoechst 33342 (Invitrogen) for 10 min followed by intense washing with PBS. A minimum of 200 cells from 5 randomly chosen fields per slide was analyzed for chromatin condensation using a Zeiss immunofluorescence microscope. Data are mean±SD of at least three independent experiments. (D) FLAG-tagged porin from N. gonorrhoeae (PorBNgo-FLAG) was expressed in HeLa cells for 24 h prior to infection with strain N886. PorBNgo-FLAG expression (green), cytochrome c (blue) and MitoTracker staining (red) were monitored. Note that infection with N886 triggered the release of cytochrome c only in cells with PorBNgo-FLAG expression. Images are representative of three independent experiments.
Figure 8.
Model of PorB's role in infection induced apoptosis.
PorB integrates into the inner membrane and causes an ATP-dependent loss of ΔΨm. Cristae remodeling by PorB sensitizes mitochondria for the BH3-protein induced opening of the OMM by Bak.