Figure 1.
Bacterial aggregates adhering to host cells are highly resistant to shear stress.
Bacterial aggregates on the cellular surface were submitted to shear stress and the ability of the aggregates to remain intact was compared (scale bars corresponds to 50 µm). (A) A monolayer of endothelial cells was infected with N. meningitidis expressing GFP for a period of 3 hours to allow microcolony formation as observed by fluorescence microscopy. Infected cells were then submitted to 10 dynes/cm2 in a laminar flow chamber as depicted on the diagram for a period of 5 min. (B) Aggregates remained unchanged before and after application of 10 dynes/cm2 (compare left and right panels). (C) The number of colony forming units before (black bars) and after (white bars) application of 10 dynes/cm2 was determined by a dilution plating and CFU determination. (D) The effect of 100 dynes/cm2 was determined.
Figure 2.
Bacterial aggregates in suspension are sensitive to shear stress.
Bacterial aggregates in the absence of host cells were submitted to shear stress and the ability of the aggregates to remain intact was compared (scale bars corresponds to 50 µm). (A) GFP-expressing N. meningitidis proliferating in suspension in cell culture medium were analyzed under a microscope to visualize aggregates of whose number and size were determined by microscopy and automated image analysis. Bacterial aggregates were submitted to 2.5–10 dynes/cm2 shear stress levels in a cone and plate device as depicted on the diagram and aggregates analyzed. (B) Aggregates in suspension were disrupted after application of 10 dynes/cm2 (compare left and right panels). (C) The effect of different shear stress levels on the number of bacterial aggregates was determined. (D) The effect of shear stress was determined for the pilT strain deficient for pilus retraction. (E) Bacteria were immobilized on a glass slide coated with a monoclonal antibody directed against type IV pili, allowed to proliferate and colonies were submitted to 10 dynes/cm2. (F) Higher magnification view of a colony immobilized on a glass slide before and after shear stress application.
Figure 3.
Host cell cholesterol depletion with cyclodextrin renders bacterial microcolonies sensitive to shear stress.
Implication of specific cellular functions was tested using inhibitors targeting actin cytoskeleton, microtubules or plasma membrane cholesterol. (A) The effect of increasing shear stress on GFP-expressing wild type microcolonies growing on cells treated with cytochalasin D, Nocodazole, or methyl ß cyclodextrin (MßCD) was analyzed by a plating assay, before (black bars) and after flow increase (10 dynes/cm2, white bars). The effect of cholesterol repletion is also indicated (MßCD+Chol). (B) Application of shear stress on infected cells treated with MßCD led to the appearance of numerous flat two-dimensional microcolonies (right panel) visible under the fluorescence microscope in contrast with the large three-dimensional microcolonies (left panel) observed on untreated cells (scale bar correspond to 10 µm). (C) Images of 2D and 3D microcolonies were taken with a confocal microscope and the Z-section is presented (scale bars correspond to 3 µm). (D) The frequency of 2D (white bars) and 3D (black bars) microcolonies was determined under the different conditions and expressed as the percentage of the total number of colonies.
Figure 4.
Lipid microdomain disruption by cholesterol depletion prevents bacteria-induced cellular response.
The effect of cholesterol depletion by methyl-ß-cyclodextrin (MßCD) on the interaction of N. meningitidis with host cells was tested. (A) Ability of N. meningitidis microcolonies to recruit cellular components determined by immunofluorescence using Ezrin as a marker. Bacteria and nuclei were stained with DAPI (DAPI); Ezrin was detected with anti-Ezrin polyclonal anti-serum (Ezrin); and images were merged (Merge). Scale bar corresponds to 10 µm. The top set of images are untreated cells and in the bottom set, cells were treated with MßCD. (B) Frequency of bacterial microcolonies efficiently recruiting ezrin (recruitment index) for non-treated cells (NT), in the presence of Cytochalasin D (CD), Nocodazole (Noco), and MßCD. (C) Dose response effect of MßCD with regard to the ability of bacterial microcolonies to reorganize the cellular surface on the surface of epithelial cells (full circles) and on endothelial cells (open circles). To control that the effect of MßCD was due to cholesterol, repletion experiments with added cholesterol were performed with both cell types (squares, open for endothelial cells and full for epithelial cells). (D) Cholesterol localization under bacterial microcolonies. GFP-expressing bacteria were used (Bacteria); Cholesterol was detected with Filipin (Cholesterol); and images were merged (Merge). The scale bar corresponds to 5 µm.
Figure 5.
Lipid microdomain disruption prevents the formation of bacteria-induced cellular projections.
(A) Human epithelial cells infected with N. meningitidis were visualized by electron microscopy to document the organization of bacterial microcolonies in relation with the cellular surface. Low magnification scanning electron microscopy shows bacteria growing in tight aggregates on the cellular surface (scale bars are 1 µm). Higher magnification shows the presence of numerous projections under and around individual bacteria in the aggregates. Transmission electron microscopy analysis of bacterial microcolonies showing the dense network of projections surrounding the bacteria is presented in the lower inset. (B) Scanning electron microscopy analysis of infected endothelial cells shows cellular projections (scale bar is 1 µm). Cells were treated with MßCD during infection and processed for scanning electron microscopy. The number and length of cellular projections is strongly reduced after MßCD (compare left and right panels).
Figure 6.
The minor pilin PilV is necessary to induce cellular surface reorganization.
Characterization of the ability of a mutant in the pilV gene to reorganize host cell plasma membrane of endothelial cells. (A) High resolution scanning electron micrographs showing direct contacts between pili and bacteria-induced cellular protrusions on epithelial cells (arrows, scale bar corresponds to 1 µm). (B) Immunofluorescence analysis of the cellular response to infection with the pilV mutant (pilV). Bacterial aggregates on the endothelial surface were visualized with DAPI staining (DAPI). Ezrin immunostaining was used as a marker for the recruitment of cytoskeletal components (Ezrin). Scale bar corresponds to 10 µm. (C) Quantification of the effect of the pilV mutation on the ability of N. meningitidis to recruit ezrin under microcolonies. The frequency of ezrin recruitment under individual microcolonies (recruitment index) is represented for the wild type strain (WT), the comP, pilT, and pilV mutants and the complemented pilV strain (pilVind) in the presence or absence of inducer (100 µM IPTG). (D) Total protein levels of PilV in the different indicated strains and with different doses of IPTG. (E) Scanning electron microscopy analysis of the cellular surface reorganization induced by wild type (WT) on endothelial cells and absent in the pilV mutant (pilV), (scale bars correspond to 1 µm).
Figure 7.
PilV is required to maintain integrity of bacterial microcolonies in the presence of shear stress.
The ability of microcolonies formed by a pilV mutant to resist shear stress was tested. (A) Microcolonies on a cellular monolayer formed by wild type or the pilV mutant expressing GFP were submitted to liquid flow generated force (10 dynes/cm2) in a laminar flow chamber. Images of fluorescent bacteria before and after flow increase are presented (scale bars correspond to 50 µm). (B) The morphology of microcolonies was determined before and after shear stress application and the frequency of 2D (white bars) and 3D (black bars) microcolonies determined. (C–D) The number of bacteria adhering to cells before (black bars) and after application of 10 dynes/cm2 shear stress (white bars) was determined by the plating assay for the indicated strains.
Figure 8.
Schematic representation of the link between the ability of microcolonies to resist mechanical force and bacteria induced cellular response.
Wild type bacteria trigger a massive reorganization of the plasma membrane and thus resist flow. The pilV mutant, in contrast, is able to adhere and form three-dimensional microcolonies but is not able to reorganize the cellular surface and renders the microcolony sensitive to shear stress. A similar effect occurs when cholesterol is depleted from the host cell with MCßD.