Fig 1.
T3SS-dependent targeting of TfnR to the apically infected plasma membrane.
(A) Basolaterally internalized Tfn and Rab11a partially co-localize at apical infection sites. Polarized MDCK-PTR9 cells were infected with EPEC-escV, or EPEC-wt and exposed to basolateral Tfn-AF647 [Tfn (Basal)]. Cells were then fixed and immunolabeled with anti-Rab11a antibodies. Representative confocal images are shown. Boxed regions highlight areas of Tfn and Rab11a co-residence at apical infection sites. Arrows point towards additional regions where Rab11a, or Tfn, are visualized at infection sites. Bar = 5 μm. (B) Rab11a and Tfn recruitment at apical infection sites depends on functional Myo5b motors. Polarized MDCK-PTR9 cells expressing the GFP- Myo5b-FL or GFP-Myo5b-tail mutant were infected with EPEC-wt. These cells were exposed to basolateral Tfn-AF647 [Tfn (Basal)], fixed, immunolabeled with anti-Rab11a antibodies, and stained with DAPI. Representative confocal images and recruitment of Rab11a, Myo5b and Tfn at infection sites are shown. Yellow arrows point towards infecting microcolonies. Red arrows point towards Myo5b-tail labeled structures. Bar = 5 μm. (C) T3SS-dependent increase in Tfn transcytosis. Polarized MDCK-PTR9 cells were infected with EPEC-escV, EPEC-wt, or left uninfected. Tfn basolateral-to-apical transcytosis and recycling were measured. Results are mean ± SE of n≥12 samples analyzed in four independent experiments. (D) T3SS-dependent increase in the abundance of s-TfnR on the apical surface. Polarized MDCK-PTR9 (top panel) and caco2-BBe (bottom panel) cells were infected with EPEC-escV, or wt, fixed and TfnR located at the apical (Ap) and basolateral (Bl) surfaces (s-TfnR) were immunostained with the B3/25 antibodies. Cells were then permeabilized, stained with TR-phalloidin (F-actin) and imaged by confocal microscopy. x-z representative images and the apical/basal distribution of s-TfnR (Z-plots) are shown. Results are mean ± SE of 12 images analyzed in three independent experiments. Bar = 5 μm.
Fig 2.
T3SS elements mediate an increase in apical endocytic turnover.
(A-D) EPEC-wt promotes an increase in Tfn apical endocytosis. (A) Biochemical analysis: Polarized MDCK-PTR9 cells were infected with EPEC-escV, wt, or left uninfected. Cells were exposed to human holo-Tfn (S2 Table) administered at either their apical [Tfn (apical)] or basolateral [Tfn (Basal)] surface. In another experiment, the dynamin inhibitor, Dynasore, was supplemented with the apical Tfn to the cells. Cells were lysed and cell-associated Tfn was detected by Western blotting, using anti-human Tfn antibodies (S4 Table). Representative Western blots are shown. For the quantitative analysis, cell-associated Tfn was normalized to α-tubulin, and results are presented as “% change of uninfected” cells. Results are mean ± SE of three independent experiments. (B) FACS analysis: Polarized MDCK-PTR9 cells were infected and exposed to Tfn-AF647, as above. The amount of cell-associated ligand was analyzed by the ‘endocytosis assay’ using flow cytometry. Results are mean ± SE; n≥12 were analyzed in three independent experiments. (C) EPEC-wt specifically affects Tfn endocytosis. Polarized MDCK-GFP-TfnR cells were infected and exposed to apical Tfn-AF647 [Tfn (apical)], as above. One set of cells was supplemented with a 50x molar excess of unlabeled Tfn administered to the apical surface while another set of cells was exposed to a 50x molar excess of unlabeled Tfn administered to the basal surface. Cell-associated ligand was assessed by the ‘endocytosis assay’ using flow cytometry. Data are presented as "% change of uninfected" cells. Results are mean ± SE; n≥6 were analyzed in two independent experiments. (D) T3SS-dependent increase in apical endocytosis of Tfn in Caco2-BBe cells. Polarized Caco2-BBe cells were infected with EPEC-escV, wt, or left uninfected. Cells were then exposed to apical Tfn-AF647, and cell-associated ligand was assessed by flow cytometry. Results are mean ± SE; n≥12 were analyzed in three independent experiments. (E) T3SS-dependent increase in apical recycling of Tfn. Polarized MDCK-PTR9 or MDCK-GFP-TfnR cells were infected with EPEC-escV, wt, or left uninfected. Cells were then exposed to apical Tfn-AF647 and the levels of released Tfn were determined by the ‘recycling assay’, using flow cytometry. Results are mean ± SE; n = 12 were analyzed in three independent experiments.
Fig 3.
T3SS secreted elements redistribute endosomes from perinuclear to peripheral host sites.
(A). T3SS elements promote a reduction in the Tfn-positive perinuclear recycling endosome area and an increase in endosomal abundance at peripheral infection sites. HeLa cells were infected with EPEC-escV, or EPEC-wt, or remained uninfected. Cells were exposed to Tfn-AF647 concomitant to infection and the subcellular localization of endocytosed Tfn-AF647 was determined by whole cell projection of confocal images. The determination of Tfn-positive perinuclear recycling endosomal (RE) punctum area was determined. Representative confocal images and processed (Otsu-threshold) images whose cell edges have been manually marked, are shown. Results presented in the box plot are mean ± SE of n≥150 infected cells analyzed in three independent experiments. Red arrows point towards the Tfn-positive perinuclear endosomal punctum, while yellow (confocal images) and green (Otsu) arrows point towards infecting microcolonies. Bar = 5 μm. (B-C) T3SS-dependent increase in Tfn-loading of TfnR and EEA1-positive peripheral endosomes. HeLa cells were infected with EPEC-escV, or EPEC-wt, or remained uninfected. Cells were exposed to Tfn-AF647 concomitant to infection, fixed, immunostained with anti-TfnR, or anti-EEA1 antibodies, and imaged by confocal microscopy. The intensity level of Tfn-AF647, TfnR and EEA1 fluorescence was determined with the ‘Image Particle Analysis’ tool. Representative confocal images (except for uninfected cells which looked very similar to the EPEC-escV infected cells) and results representing fluorescence intensity of TfnR, Tfn and their ratio (B) and EEA1, Tfn and their ratio (C) in individual endosomes (n≥600 single particles) are shown in scatter plots. Arrows point towards infecting microcolonies. Bar = 5 μm. (D) Electron micrograph showing enrichment of endosomes containing Tfn at peripheral EPEC-wt infection sites. HeLa cells were infected with EPEC-escV or EPEC-wt and concomitantly exposed to Tfn-HRP. Cells were then processed for Tfn-HRP visualization by electron microscopy. Representative images are shown. Blue arrowheads point towards infecting bacteria and red arrows point towards internalized Tfn-labeled endosomes. Bar = 1 μm.
Fig 4.
Shuttling of Rab11a and Tfn from perinuclear recycling endosomes to peripheral infection sites.
HeLa cells transiently expressing tdEos-Rab11a were exposed to Tfn-AF647 and then infected with EPEC-escV, or EPEC-wt. Immediately upon bacterial attachment to the host cell, tdEos-Rab11a and Tfn labeled central puncta were subjected to photo-conversion and images were acquired by live-cell confocal microscopy (S7 and S8 Movies). Representative images and quantitative analysis of the photo-converted tdEos-Rab11a recruitment at infection sites are shown. Results are mean ± SE; n≥20 bacterial microcolonies (yellow arrow) were analyzed in three independent experiments. Red arrow points towards a tdEos-Rab11a/Tfn positive perinuclear endosomal punctum subjected to photoconversion. Bar = 5 μm.
Fig 5.
Myo5b is essential for T3SS-dependent recruitment of Rab11a/Tfn-positive endosomes to infection sites and increased endocytic activity.
(A) Effects of GFP-Myo5b expression on Tfn recruitment at infection sites. HeLa cells transiently expressing the indicated GFP-Myo5b constructs were infected with EPEC-wt and exposed to Tfn-AF647. Cells were fixed and stained with DAPI and subjected to confocal imaging. Representative images and quantitative analysis of the recruitment of the indicated markers at infection sites are shown. Results are mean ± SE; n≥30 bacterial microcolonies were analyzed in at least three independent experiments. Bacterial microcolonies are pointed out by arrows. Bar = 5 μm. (B) Effects of GFP-Myo5b expression on F-actin and Rab11a recruitment at infection sites. HeLa cells transiently expressing the indicated GFP-Myo5b constructs were subjected to EPEC-wt infection, fixed, immunostained with anti-Rab11a antibodies and then stained with TR-phalloidin (F-actin) and DAPI (DNA). Representative images and quantitative analysis of marker recruitment at infection sites are shown. Results are mean ± SE; n≥30 microcolonies were analyzed in at least three independent experiments. Arrows point towards infecting bacterial microcolonies. Bar = 5 μm. (C) Effects of GFP-Myo5b expression on Tfn endocytosis. HeLa cells expressing the indicated GFP-Myo5b constructs were infected with EPEC-wt concomitant to Tfn-AF647 exposure, or left uninfected. The levels of internalized ligand in GFP-expressing cells, and in cells that showed no detectable GFP expression ('non-expressing cells'), were measured by FACS. Untransfected cells served as control. Results are mean ± SE; n≥12 were analyzed in at least three independent experiments.
Fig 6.
Rab11a and Rab11b are essential for Tfn-positive endosome recruitment to infection sites and increased endocytic activity.
(A) Rab11a and Rab11b silencing by siRNA. HeLa cells were treated with scrambled siRNA, siRab11a, siRab11b or a mixture of siRab11a and siRab11b (a+b). Cells were subjected to Western blotting analysis using anti-Rab11a, anti-Rab11b, and anti-α-tubulin antibodies (S4 Table). A representative Western blot and quantitative analysis of the Rab11 levels are shown. Results are mean ± SE of three experiments. (B) Expression of Rab11a and Rab11b is required for Tfn recruitment at EPEC-wt infection sites. HeLa cells transfected with the indicated siRNAs were infected with EPEC-wt and exposed to Tfn-AF488. Cells were stained with TR-phalloidin and DAPI, and imaged by confocal microscopy. Representative images and quantitative analysis of markers’ recruitment at infection sites are shown. Results are mean ± SE; n≥30 infecting microcolonies were analyzed in three independent experiments. Arrows point towards infecting microcolonies. Bar = 5 μm. (C) Rab11a and Rab11b are required for EPEC- mediated increase in Tfn endocytosis. HeLa cells treated with the indicated siRNAs were infected with EPEC-wt and exposed to Tfn-AF488. The levels of internalized ligand were measured by flow cytometry. Results are mean ± SE; n≥12 were analyzed in three independent experiments.
Fig 7.
EspF and Map mediate the recruitment of Tfn/TfnR and Myo5b/Rab11a at infection sites.
(A) Screening for type III secreted protein effectors mediating Tfn recruitment to infection sites. HeLa cells were infected with EPEC-wt, or with the indicated mutant strains (S1 Table), or remained uninfected. Cells were exposed to Tfn-AF647 during the infection, fixed, stained with TR-phalloidin (F-actin) and DAPI and analyzed by confocal microscopy. Quantitative analysis of F-actin and Tfn recruitment at infection sites are shown. Results are mean ± SE; n≥30 bacterial microcolonies. (B) EspF and Map are essential for Tfn/TfnR recruitment at infection sites. HeLa cells were infected with the EPEC-espF or map mutant strains and their corresponding espF+EspF, or map+Map complemented strains. Effector protein expression was induced by IPTG. Cells were exposed to Tfn-AF647 during infection, fixed, immunolabeled with anti-TfnR antibodies, stained with TR-phalloidin and DAPI and analyzed by confocal microscopy. Quantitative analysis of F-actin, TfnR and Tfn recruitment at infection sites are shown. Results are mean ± SE of n≥30 bacterial microcolonies analyzed in three independent experiments. Arrows point towards infecting microcolonies. Bar = 5 μm. (C) EspF and Map are essential for Myo5b and Rab11a recruitment at infection sites. HeLa cells were infected with the indicated EPEC strains, immunolabeled with anti-Myo5b and anti-Rab11a antibodies, stained with TR-phalloidin and DAPI and visualized by confocal microscopy. Quantitative analysis of F-actin, Myo5b and Rab11a recruitment at infection sites are shown. (D) Simultaneous mutation of espF and map markedly reduces the recruitment of Tfn, TfnR, Myo5b and Rab11a at infection sites. HeLa cells were infected with the EPEC-espF/map mutant strain and the level of the recruitment of the indicated protein markers at infection sites was evaluated as in panels B&C. Results are mean ± SE; n≥30 infecting microcolonies analyzed in three independent experiments. Representative confocal images are shown in S12 Fig.
Fig 8.
EspF and Map mediate alterations in endosomal distribution and stimulate an increase of endocytic turnover.
(A-B) Translocated EspF and Map prompt a reduction in Tfn-positive perinuclear endosome area and an increase in endosomal abundance at peripheral infection sites. HeLa cells were infected with the indicated EPEC strains and exposed to Tfn-AF647 during infection. The area of Tfn-positive perinuclear recycling endosome (RE) punctum was determined. Results are mean ± SE; n≥75 infected cells were analyzed in three independent experiments. Red and yellow arrows point towards the Tfn positive perinuclear endosomal puncta and infecting bacterial microcolonies, respectively. Bar = 5 μm. (C-D) EspF and Map-dependent increase in Tfn-loading of TfnR and EEA1-positive peripheral endosomes. HeLa cells were infected with the indicated EPEC strains and exposed to Tfn-AF647 during the infection. Quantitative analyses of the fluorescence intensities of Tfn, TfnR and EEA1 in individual endosomes were determined. Results present fluorescence intensity of Tfn normalized to the fluorescence intensity of TfnR or to the fluorescence intensity of EEA1, from n≥500 single particles (endosomes). (E-F) EspF and Map increased the Tfn endocytic turnover. HeLa cells were infected with the indicated EPEC strains and exposed to Tfn-AF647 during infection. Tfn-AF647 endocytosis, recycling and surface-bound ligand were determined by flow cytometry. Results are mean ± SE of n≥6 samples from three independent experiments. (G-H) The interactions of EspF and Map with host proteins are essential for eliciting Tfn endocytosis. HeLa cells were infected with the indicated EPEC strains (see S3 Table) and exposed to Tfn-AF647 during the infection. The capacity of cells to endocytose Tfn was determined by flow cytometry. Results are mean ± SE; n≥6, analyzed in three independent experiments.
Fig 9.
EspF, but not Map, mediates the recruitment of TfnR and Rab11a at apical infection sites in polarized epithelial cells.
(A) Infection with an EPEC-espF, but not EPEC-map, mutant results in diminished recruitment of Rab11a at infection sites. Polarized MDCK cells transiently co-expressing mRFP-LifeAct and GFP-Rab11a were infected with the indicated EPEC strains (S1 Table) and imaged by live-cell confocal microscopy. Results are mean ± SE of n≥20 bacterial microcolonies analyzed in three independent experiments. (B) EspF-dependent recruitment of Rab11a, TfnR and Tfn at infection sites. Polarized MDCK cells transiently co-expressing mRFP-LifeAct and GFP-Rab11a, or mRFP-LifeAct and GFP-TfnR, were infected with EPEC-espF or EPEC-espF+EspF. The latter cells were exposed to basolateral Tfn-DL649 during the infection. Cells were then imaged by live-cell confocal microscopy, and the degree of recruitment of the expressed host proteins and internalized Tfn at apical infection sites were quantitatively determined. Arrows point to infection sites. Results are mean ± SE; n≥20 bacterial microcolonies (arrows) were analyzed in three independent experiments. (C) EspF-dependent increase in endocytic turnover. The apical surface of polarized MDCK-GFP-TfnR cells was infected with the indicated EPEC strains or remained uninfected. Cells were exposed to apical Tfn-DL649 and cell-associated Tfn was determined by FACS analysis. Results are mean ± SE of six independent experiments. (D) EspF-dependent increase in Tfn transcytosis. Polarized MDCK-PTR9 cells were infected with the indicated EPEC strains, or remained uninfected. Tfn transcytosis was analyzed as in Fig 1C. Results are mean ± SE of five independent experiments.
Fig 10.
Translocated EspF interacts with SNX9, N-WASP, and the novel SNX18, SNX33, WIPF1 proteins.
(A) Major host cell proteins co-immunoprecipitated with EspF. HeLa cells were infected with EPEC-espF or espF+EspF. The Flag-tagged EspF was immunoprecipitated and co-immunoprecipitated proteins were identified, as described in Methods. 232 proteins were identified in at least two of four replicates (S6 Table). Student’s t-test was used to assess which proteins were significantly more abundant in the EspF pull-down compared to the control sample. Six proteins enriched by at least 80-fold with p-value < 0.05 were retrieved. One is EspF itself (not shown), three are sorting nexin proteins (SNX9, SNX18, and SNX33) and two are actin polymerization regulators, WIPF1 (WASP/WASL-interacting protein) and N-WASP (also named WASL). (B) STRING protein interaction analysis. To identify potential interactions within the EspF binders, we used the STRING protein interaction database [85], updated in September 2018, under the highest confidence and experimentally validated interaction setting. Cluster analysis shows that WIPF1, WASL, SNX9 and SNX18 form a highly connected interaction network with experimentally validated interactions (based on the STRING annotation), suggesting that these proteins are pulled-down as one interaction unit. Previous studies have identified only SNX9 and N-WASP/WASL as EspF binding partners [22,23]. The EspF interactions were added using the Cytoscape program. We, as a result, identified here three additional interactors, SNX33, SNX18 and WIPF1. (C) SNX9, SNX18 and SNX33 co-immunoprecipitation with EspF confirmed by Western blotting. HeLa cells infected with EPEC-espF or EPEC-espF+EspF were subjected to EspF immunoprecipitation (IP), using anti-FLAG antibodies. The immunoprecipitated EspF and associated proteins were subjected to SDS-PAGE followed by Western blotting analyses. EspF (~26 and 15 kDa) was detected by anti-FLAG antibodies, and co-immunoprecipitated SNX9, SNX18 and SNX33 (~75 kDa) were detected by the appropriate antibodies (S4 Table). The results show that the SNX proteins are specifically co-immunoprecipitated with translocated EspF (D) Co-immunoprecipitation efficiency of SNX9, SNX18 and SNX33 with EspF. Experiments were performed as in panel C, except that cells were infected with the indicated EspF mod strains. Results are mean ± SE of three independent experiments. The results show reduced levels of co-immunoprecipitated SNX with EspF mod-RD compared to EspF mod-wt.
Fig 11.
Schematic model of EPEC-induced recruitment of TfnR-Rab11-positive endosomes at infection sites of non-polarized (A) and polarized (B) epithelial cells.
In non-polarized cells, EPEC, through its effectors EspF and Map, prompts Myo5b-dependent shuttling of Rab11-positive recycling endosomes to the peripheral infected plasma membrane, where some are clustered at bacterial infection sites. These endosomes exert local endocytic and recycling activities, thereby stimulating the endocytic turnover of plasma membrane proteins at the infected plasma membrane. This activity enriches the plasma membrane infection sites with the missorted endosomes and their cargo proteins. In polarized cells, the Rab11-positive apical recycling endosomes are located beneath the infected (apical) plasma membrane, promoting apical recycling of apical plasma membrane protein. They also utilize their apical recycling capacity to promote transcytosis of basolaterally recycled cargo that has been missorted from the common endosomes to the apical recycling endosomes [25]. Upon infection, translocated EspF promotes Myo5b-dependent recruitment of the Rab11-positive apical recycling endosomes to infection sites, where they locally foster endocytosis and recycling (i.e. endocytic turnover). We hypothesize that EspF, and possibly additional protein effectors, interfere with the basolateral recycling sorting machinery. The consequence of this event is missorting of internalized basolateral plasma membrane recycling proteins (e.g. TfnR) from the common endosomes (which mediate basolateral recycling) to the subapical Rab11 recycling endosomes hijacked by EPEC (dashed line arrows). These endosomes then use their apical recycling capacity to translocate the basolateral plasma membrane proteins into apical plasma membrane infection sites, where they undergo continuous endocytosis and apical recycling. This activity enriches the apical plasma membrane infection sites with plasma membrane proteins originally derived from the basolateral surface of the host.