SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Gram-negative pathogen Pseudomonas aeruginosa invades epithelial cells in vivo and in vitro. We have examined the pathway(s) by which epithelial cells internalize P. aeruginosa strain PA103 using Madin-Darby canine kidney (MDCK) cells. We have recently demonstrated that P. aeruginosa internalization occurs by an actin-dependent Toxin B-inhibited pathway which becomes downregulated as epithelial cells become polarized, suggesting that one or more of the Rho family GTPases is involved in bacterial internalization. Here, we demonstrate that activation of the Rho family GTPases by cytotoxic necrotizing factor 1 (CNF-1) stimulates P. aeruginosa internalization. Examination of the roles of the individual Rho family GTPases in internalization shows that expression of a constitutively active allele of RhoA (RhoAV14), but not of constitutively active Rac1 (Rac1V12) or Cdc42 (Cdc42V12), is sufficient to increase uptake of PA103pscJ. This relative increase persists when bacterial infection is established at the basolateral surface of polarized cells, suggesting that the effect of RhoAV14 is not simply due to its known ability to disrupt tight junction integrity in polarized cells. RhoAV14-mediated stimulation of bacterial uptake is actin dependent as it is abrogated by exposure to latrunculin A. We also find that endogenous Rho GTP levels in epithelial cells are increased by infection with an internalized strain of P. aeruginosa; conversely, a poorly internalized isogenic strain expressing the bacterial anti-internalization protein ExoT causes decreased Rho GTP levels. Experimental inhibition of Rho, either by expressing dominant negative RhoAN19 or by inhibiting native Rho using a membrane permeable fusion construct of a Rho-specific inhibitor, C3 ADP-ribosyltransferase, does not inhibit PA103pscJ internalization in MDCK or HeLa cells. Models consistent with these data are presented.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Gram-negative pathogen Pseudomonas aeruginosa is one of many bacterial pathogens which can be internalized by non-professional phagocytes, such as epithelial cells (Fleiszig et al., 1994, 1995). Much is known about the general mechanisms of bacterial internalization by epithelial cells. Internalization is frequently the result of bacterial manipulation of the host cell actin cytoskeleton at the level of the small GTPases of the Rho family, proteins whose activity is associated with particular membrane and cytoskeleton rearrangements, such as lamellipodia (Rac), filopodia (Cdc42) and stress fibres (RhoA) (reviewed in Hall, 1998). Thus, Salmonella typhimurium promotes its internalization through a type III secreted bacterial protein SopE, which acts as a guanine-nucleotide exchange factor (GEF) for the Rho family GTPases Cdc42 and Rac1 (Hardt et al., 1998; Rudolph et al., 1999). In contrast, other bacteria modify the host cell via type III secreted effectors so as to block their own uptake. For example, Yersinia pseudotuberculosis inhibits its internalization via YopE (Mecsas et al., 1998), a protein with GTPase-activating protein (GAP) activity in vitro for RhoA, Cdc42 and Rac (Black and Bliska, 2000; von Pawel-Rammingen et al., 2000). The ability to resist internalization by host cells, in particular professional phagocytes of the immune system, may contribute significantly to the virulence of certain pathogens. However, as is postulated for Yersinia, internalization may nonetheless be required during certain steps of the disease process, such as breaching the mucosal barrier of the intestinal lumen by invading and crossing the M cells of the Peyer's patches (Finlay and Falkow, 1997).

P. aeruginosa has been found inside numerous epithelial cell types in vivo as well as in vitro (Fleiszig et al., 1994, 1995, 1996). The role of internalization in pathogenesis is not understood, although in certain disease models, such as corneal ulceration, bacterial internalization in the absence of other type III mediated virulence factors appears sufficient to cause disease (Fleiszig et al., 1994). Like Yersinia spp., P. aeruginosa appears to inhibit its internalization through a type III secreted effector protein, ExoT. PA103 mutants lacking this protein, or mutants defective in type III secretion, are thus internalized more efficiently than ExoT-expressing strains (Evans et al., 1998; Hauser et al., 1998; Cowell et al., 2000) (Garrity-Ryan et al., 2000). P. aeruginosa internalization is inhibited by actin depolymerizing agents such as cytochalasin D (Evans et al., 1998), suggesting that the actin cytoskeleton is involved in bacterial uptake. Moreover, toxin B, a Clostridium difficile protein that inactivates Rac, Rho and Cdc42 (Just et al., 1995), inhibits P. aeruginosa invasion into non-polarized cells, providing evidence that the Rho family GTPases play a role in P. aeruginosa invasion (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). The anti-internalization factor ExoT contains a GAP domain homologous to that found in YopE, suggesting that it may inhibit P. aeruginosa internalization by inactivating one or more of the Rho family GTPases. This hypothesis is supported by the finding that mutation of the conserved arginine within this domain (Arg149) significantly reduces the ability of ExoT to prevent bacterial internalization (Garrity-Ryan et al., 2000).

Although the pathway which governs P. aeruginosa internalization is incompletely understood, it is appreciated that host cell factors influence internalization. In epithelial cell monolayer cultures, for example, internalization is enhanced by manipulations such as mechanical wounding (Plotkowski et al., 1999), disruption of tight junctions by exposure to calcium chelators (Fleiszig et al., 1997; Pereira et al., 1997; Plotkowski et al., 1999), treatment of cells with hepatocyte growth factor/scatter factor (HGF) (Fleiszig et al., 1998) or tissue culture under conditions which do not promote the formation of polarized monolayers (Plotkowski et al., 1999). Conversely, internalization occurs to a lesser extent in cell lines which acquire a high transepithelial resistance (Fleiszig et al., 1997). These results are taken to show that P. aeruginosa internalization preferentially occurs following bacterial interaction with the basolateral domain of polarized epithelial cells; thus, disruption of tight junction integrity or wounding may allow apically infecting bacteria access to this membrane domain, whereas HGF exposure or growth under conditions not promoting polarization might result in the inappropriate appearance of basolateral determinants on the apical surface of cells. However, the identity of the basolateral determinant responsible for P. aeruginosa internalization has remained elusive.

Although spatial restriction of a P. aeruginosa receptor may contribute to the relative resistance of polarized epithelial cell monolayers to bacterial invasion, recent work from our laboratory suggests that incompletely polarized cells possess a pathway for P. aeruginosa internalization which is downregulated during the acquisition of polarity. This pathway is sensitive to actin depolymerizing agents, such as latrunculin A, and requires activity of a small GTPase as it can be inhibited by toxin B (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). In this work, we explicitly show that activation of RhoA, but not of Rac1 or Cdc42, is sufficient to increase P. aeruginosa internalization. The biological relevance of this finding is strengthened by our observation that infection with an internalized strain of P. aeruginosa, PA103ΔUΔT, is accompanied by increases in endogenous Rho GTP levels, whereas infection with an isogenic strain expressing the anti-internalization factor ExoT leads to significant decreases in Rho GTP levels.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Activation of Rho family GTPases by cytotoxic necrotizing factor 1 (CNF-1) treatment promotes P. aeruginosa internalization

We have shown that treatment of confluent although incompletely polarized Madin-Darby canine kidney (MDCK) cells with the C. difficile protein toxin B profoundly inhibits the internalization of P. aeruginosa PA103pscJ, suggesting that internalization is promoted by the activity of one or more of the toxin B-sensitive Rho family GTPases (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). We tested this hypothesis explicitly by treating MDCK II cells with cytotoxic necrotizing factor 1 (CNF-1), a bacterial toxin that activates Rho family GTPases and causes the accumulation of stress fibres, focal adhesions, microspikes and membrane ruffles in treated cells (Flatau et al., 1997; Schmidt et al., 1997; Lerm et al., 1999). Exposure of MDCK cells to CNF-1 (0.3 nM) for 15 h resulted in increased PA103pscJ internalization compared with untreated cells (Fig. 1A). The number of PA103pscJ bacteria bound to epithelial cells did not differ significantly between control and CNF-1-treated samples, as seen in Fig. 1A (P = 0.449, Student's two-tailed t-test).

image

Figure 1. CNF-1 treatment of MDCK cells causes increased P. aeruginosa internalization.

A. Confluent MDCK cell monolayers were treated with CNF-1 for 15 h before bacterial infection. Bacterial adhesion and internalization were measured as described in the Experimental procedures. Bars represent the mean number of adherent or internalized bacteria per Transwell (average of six wells) ± SEM in a representative experiment.

B. Uninfected MDCK cell monolayers treated with CNF-1 (0.3 nM) for 15 h compared with untreated cells. In all panels, F-actin staining by phalloidin appears red, and protein marker staining appears green. Cells were fixed, permeabilized and stained with Texas Red phalloidin and anti-gp135 (a), anti-ZO-1 (b) or anti-E-cadherin (c), followed by the appropriate Alexa 488-conjugated secondary antibody. Cells were visualized using confocal laser microscopy; the panels illustrate representative vertical sections. Untreated cells (left) show appropriately restricted localization of gp135 (apical), ZO-1 (tight junctions) and E-cadherin (cell–cell contacts, basolaterally). CNF-1-treated cells (right) show restriction of gp135 to apical membrane domains. ZO-1 localization is primarily cytoplasmic, whereas E-cadherin staining appears aberrantly in both apical and basolateral membranes.

Download figure to PowerPoint

The absence of increased bacterial binding is surprising as CNF-1 markedly altered the morphology of the MDCK cell monolayer in a manner which should improve bacterial access to the basolateral pole, where bacterial binding and internalization are thought to occur preferentially. CNF-1-treated cells continued to maintain cell–cell contacts, but showed much less uniform cell size and shape than control cells. Apical restriction of gp135 persisted in treated cells, but anti-zona occludens 1 (ZO-1) staining of tight junctions was disrupted and E-cadherin staining was no longer restricted to basolateral membranes (Fig. 1B). Nonetheless, overall bacterial binding to cells remained unchanged, suggesting that CNF-1 treatment increased PA103pscJ internalization by altering the efficiency with which bound bacteria were internalized by cells.

Expression of the CA RhoAV14 protein, but not of CA Rac1V12 or CA Cdc42V12, is sufficient to promote P. aeruginosa internalization by MDCK cells

As CNF-1 activation of Rho family GTPases increased P. aeruginosa uptake by epithelial cells, we next asked whether activation of any single GTPase was sufficient to cause increased bacterial internalization of PA103pscJ. Stably transfected MDCK cells expressing constitutively active (CA) mutant alleles of Rac1 (Rac1V12), RhoA (RhoAV14) and Cdc42 (Cdc42V12) under the control of a tetracycline repressible promoter system were used for these studies. This resulted in reproducible expression of the mutant proteins in all cells of the population under study (Jou and Nelson, 1998) and allowed us to measure changes in bacterial internalization by populations of cells. Expression of the mutant alleles was induced by growing the cells in the absence of doxycycline for 18–24 h before plating them to Transwell filters at high density (1.0–1.5 × 106 cells/cm2), resulting in the formation of an ‘instant’ monolayer. Cells were routinely assayed 36 or 60 h after plating to filters; there were no differences between data obtained from these time points. Cells induced to express the CA allele RhoAV14 internalized PA103pscJ≈ 10- to 75-fold more than uninduced cells (Fig. 2, P < 0.0001, Student's two-tailed t-test; see also Figs 3 and 4). This was similar to the increase seen in CNF-1-treated cells compared with controls (14- to 50-fold) (Fig. 1; data not shown). In contrast, expression of CA Rac1V12 led to a less than twofold increase in internalization which was not statistically significant (P = 0.106, Student's two-tailed t-test), whereas expression of CA Cdc42V12 resulted in a small decrease in internalization which also did not reach statistical significance (P = 0.219, Student's two-tailed t-test). Thus, only expression of the active allele RhoAV14 was sufficient to promote P. aeruginosa internalization.

image

Figure 2. Expression of a constitutively active allele of RhoA, but not of Rac1 or Cdc42, is sufficient to increase P. aeruginosa internalization. Mutant MDCK cell lines either uninduced (–) (maintained in the presence of doxycycline 20 ng ml−1) or induced (+) by withdrawal of doxycycline to express RhoAV14, Rac1V12 or Cdc42V12 were infected with PA103pscJ. The number of bacteria internalized per Transwell was determined as described in the Experimental procedures. Bars show the average of triplicate wells (± SD) of a representative experiment; the error bars on some points are too small to appear on this scale.

Download figure to PowerPoint

image

Figure 3. RhoAV14 expression increases P. aeruginosa and S. flexneri internalization by MDCK cells, but has no effect on S. typhimurium internalization. Mutant MDCK cell lines either uninduced (–) (maintained in the presence of doxycycline 20 ng ml−1) or induced (+) by withdrawal of doxycycline to express RhoAV14, Rac1V12 or Cdc42V12 were infected with PA103pscJ. Bars show the average of triplicate wells (± SD) of a representative experiment; the error bars on some points are too small to appear on this scale.

Download figure to PowerPoint

image

Figure 4. RhoAV14-induced increases in P. aeruginosa internalization are inhibited by latrunculin A and toxin B. Mutant MDCK cells were either maintained in the presence of doxycycline (20 ng ml−1, ‘uninduced’) or induced by doxycycline withdrawal to express the CA RhoAV14 protein. The number of PA103pscJ internalized per Transwell was determined as previously described. Bars represent the average of triplicate wells (± SD) of a representative experiment.

A. Cells were with latrunculin A (10 µg ml−1) or with medium plus DMSO alone.

B. Cells were treated with toxin B (100 ng ml−1) in medium or medium alone (control).

Download figure to PowerPoint

RhoAV14 expression does not cause non-specific increases in bacterial internalization

Rho family GTPases have been demonstrated to be required for the internalization of two bacterial pathogens, S. typhimurium and Shigella flexneri. Shigella internalization requires RhoA activity (Adam et al., 1996; Watarai et al., 1997), whereas Salmonella internalization requires Cdc42 and Rac1 activity and is independent of Rho (Chen et al., 1996). Consistent with these findings, the internalization of S. flexneri M90T was increased by CA RhoAV14 expression, whereas internalization of S. typhimurium SL1344 was not significantly altered by CA RhoAV14 expression (Fig. 3). These results demonstrate that activation of RhoA in this system does not result in non-specific uptake of all bacterial pathogens.

CA RhoAV14-mediated internalization of PA103pscJ is actin dependent and can be inhibited by latrunculin A or toxin B

The internalization of P. aeruginosa by non-polarized cells has been shown to require actin polymerization (Evans et al., 1998) (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). We tested whether the uptake process stimulated by CA RhoAV14 expression also required actin polymerization. Cells expressing RhoAV14 were pretreated for 30 min before infection with latrunculin A, an agent which binds monomeric G-actin and shifts the equilibrium between F-actin and G-actin in favour of actin disassembly (Lamaze et al., 1997). Latrunculin A was able to abrogate completely the RhoAV14-mediated increase in internalization, as seen in Fig. 4A. Latrunculin A did not inhibit PA103pscJ internalization by control cells, consistent with our previous report that polarized epithelial cells no longer show actin-dependent uptake of this strain (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). We confirmed that the decreased number of invasive bacteria recovered in the latrunculin A-treated samples was not due to preferential loss of the RhoAV14-expressing cells by comparing the number of filter-bound cells, as visualized by Texas Red phalloidin and Syto9 staining in uninduced vs. induced cells after treatment with latrunculin A (data not shown).

Toxin B is a C. difficile protein which monoglucosylates Rho, Rac and Cdc42 proteins, leading to their irreversible inactivation (Aktories, 1997). Treatment of RhoAV14-expressing MDCK cells with toxin B before bacterial infection abrogated the RhoAV14-mediated increase in P. aeruginosa internalization (Fig. 4B). This finding suggests that the activity of a Rho family GTPase is required for increased internalization to occur after RhoAV14 expression. We again confirmed that the effect of toxin B was not the result of preferential loss of RhoAV14-expressing cells (data not shown). Toxin B minimally decreased internalization into the uninduced control cells (P > 0.1, Student's two-tailed t-test); this observation is in agreement with our prior observations that toxin B has no inhibitory effect on P. aeruginosa internalization in polarized epithelial cells (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). It is not known whether toxin B modification of RhoAV14 results in the effective inhibition of RhoAV14 interactions with its downstream effectors. Thus, these results may signify that RhoA is directly involved in the pathway of Pseudomonas uptake in MDCK cells or that RhoA activates another toxin B-sensitive GTPase, which in turn increases P. aeruginosa uptake.

CA RhoAV14 stimulates PA103pscJ uptake in a basolateral model of infection

Rho activity regulates the function of multiple cellular structures, among them tight junctions (Nusrat et al., 1995; Jou et al., 1998). It has been shown that disruption of tight junctions by calcium chelators results in increased P. aeruginosa internalization by polarized epithelial cells, presumably by allowing bacteria improved access to the basolateral surfaces of these cells where internalization is postulated to occur preferentially (Fleiszig et al., 1997, 1998; Pereira et al., 1997; Plotkowski et al., 1999). We therefore developed a basolateral infection model to test explicitly whether CA RhoAV14 was increasing P. aeruginosa internalization by improving bacterial access to the basolateral domain of MDCK cells. We induced RhoAV14 expression, as described above, and cultured the control or induced cells on Transwell filters with 3-µm-diameter pores, as the pore size (0.4 µm) in the filters used in our usual model system prevented P. aeruginosa movement across the filter. Under these culture conditions, the cells acquired a transepithelial resistance characteristic of cells cultured on 0.4 µm filters, and appeared to segregate correctly apical and basolateral markers such as gp135 and E-cadherin. The only difference noted was that cell height was less for the 3 µm filter-grown cells than that seen with growth on 0.4 µm filters (data not shown). Figure 5 demonstrates that cells expressing CA RhoAV14 still showed 10-fold increases in bacterial internalization compared with uninduced control cells after basolateral infection. Thus, RhoAV14 expression stimulates PA103pscJ internalization at a step distal to the bacteria gaining access to discrete membrane surfaces of polarized epithelial cells.

image

Figure 5. RhoAV14 expression increases P. aeruginosa internalization in a basolateral model of infection. Mutant MDCK cells were either maintained in the presence of doxycycline (20 ng ml−1, ‘uninduced’) or induced by doxycyline withdrawal to express the CA RhoAV14 protein. Cells were plated to 3.0 µm filters and infected with PA103pscJ at their basolateral surfaces and the number of internalized bacteria per Transwell determined as before. Bars represent the average of triplicate wells (± SD) from a representative experiment.

Download figure to PowerPoint

DN RhoAN19 does not inhibit internalization of PA103pscJ

The experiments described thus far demonstrate that the activated allele of RhoA is sufficient to promote P. aeruginosa internalization. We now asked whether Rho activity was required for PA103pscJ internalization by testing whether expression of the dominant negative (DN) allele of RhoA, RhoAN19, inhibited bacterial internalization. As can be seen in Fig. 3, inducing RhoAN19 expression in our stably transfected MDCK system did not inhibit P. aeruginosa internalization, but, instead, resulted in small but reproducible increases in internalization (2.6-fold increase, P = 0.0269, Student's two-tailed t-test). Moreover, S. flexneri internalization was also increased, despite the fact that RhoA has been shown to be necessary for Shigella internalization in other experimental systems (Adam et al., 1996; Watarai et al., 1997). As reported previously (Jou et al., 1998), increased paracellular diffusion of the tracer molecule fluorescein isothiocyanate (FITC)-inulin was seen following RhoAN19 expression (data not shown). As expression of dominant negative alleles often results in incomplete inhibition of endogenous Rho family GTPases (Feig, 1999), we developed a second experimental approach to ask whether Rho activity is required for Pseudomonas internalization.

Inhibition of RhoA by C3 ADP-ribosyltransferase does not inhibit P. aeruginosa uptake by epithelial cells

The Clostridium botulinum toxin C3 ADP-ribosyltransferase (C3) is a specific inhibitor of RhoA, B and C (Aktories, 1997). Although some cell types have been treated directly with purified C3 (Morii and Narumiya, 1995; Watarai et al., 1997), the protein is very membrane impermeable and thus usually administered to cells by microinjection. Such an approach is impractical for assaying entire populations of cells. We therefore developed a membrane-permeable C3 fusion protein, using an 11-amino-acid sequence from the HIV TAT protein fused at the N-terminus of C3 to confer membrane permeability on the protein (Nagahara et al., 1998). Treatment of HeLa or MDCK cells grown on coverslips or MDCK cells grown on permeable filters with TAT–C3 for 18–20 h resulted in the complete loss of stress fibres (Fig. 6; data not shown). This cytoskeletal change is not seen after treatment with equal or higher concentrations of a TAT–GFP fusion and is consistent with inhibition of Rho (data not shown).

image

Figure 6. Treatment of MDCK cells with TAT–C3 ADP ribosyltransferase protein results in the loss of stress fibres. MDCK cells were plated on coverslips 12 h before being treated with TAT–C3 ADP-ribosyltransferase protein (300 nM) in MEM Earles plus 5% FBS for 20 h. Cells were fixed as described in the Experimental procedures and F-actin stained with Texas Red phalloidin. Samples were imaged using a Bio-Rad 1024 confocal microscope (60× objective); the image shown here is projection of a Z-series collected through the depth of the cell at 1.0 µm steps. No stress fibres could be visualized in the TAT–C3 ADP-ribosyltransferase treated cells.

Download figure to PowerPoint

Treatment of polarized MDCK cells (assayed 3 days after plating to filters) with TAT–C3 resulted in increased PA103pscJ internalization, as had been seen in cells expressing the DN RhoAN19 allele (Fig. 7). TAT–C3 treatment also increased paracellular diffusion of FITC-inulin across these monolayers in a dose-dependent fashion (data not shown). Confluent MDCK cell monolayers which were incompletely polarized (assayed 1 day after plating to filters) also showed no inhibition of PA103pscJ internalization (Fig. 7), nor did treated HeLa cells (data not shown). Both incompletely polarized MDCK cells and HeLa cells, however, show toxin B-sensitive PA103pscJ internalization (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). A broad range of TAT–C3 concentrations (50 nM to 12 µM) and treatment intervals (12–20 h) was used in these experiments, resulting in dose-dependent loss of stress fibres and cell rounding (data not shown); thus, the absence of inhibition by TAT–C3 is not likely to be due to inappropriate or inadequate exposure of cells to reagent.

image

Figure 7. TAT–C3 transferase treatment does not inhibit P. aeruginosa internalization. MDCK cells were plated to filters at high density in the presence of TAT–C3 or TAT–GFP at the indicated concentrations (‘day 1 MDCK′) or 2 days before treatment with TAT–C3 or TAT–GFP (‘day 3 MDCK′). Twenty hours after exposing the cells to fusion protein, PA103pscJ was added to the apical compartment of the Transwell and infection allowed to proceed for 2 h at 37°C; the cells were then washed, treated with amikacin-containing medium, lysed and processed for internalized bacteria as described in the Experimental procedures. Bars represent the average of triplicate wells (± SD) from a representative experiment.

Download figure to PowerPoint

P. aeruginosa internalization is accompanied by activation of endogenous RhoA

Ren et al. (1999) have recently developed a reagent, the Rhotekin Rho-binding domain (TRBD) fused to GST, which allows the specific affinity precipitation of RhoA–GTP from cell lysates. We used this reagent to assay whether bacterial internalization affects endogenous RhoA activity in epithelial cells. Two isogenic bacterial strains were compared. PA103ΔUΔT is an invasive strain which is type III secretion competent but no longer synthesizes the anti-internalization factor ExoT, whereas PA103ΔU produces ExoT and is poorly internalized (Garrity-Ryan et al., 2000). Infection of HeLa cells or incompletely polarized MDCK monolayers with PA103ΔUΔT resulted in marked increases in RhoA–GTP levels assayed 1–3 h after infection (Fig. 8; data not shown). In contrast, infection of HeLa cells with the non-internalized strain expressing ExoT, PA103ΔU, resulted in marked decreases in measurable Rho–GTP levels (Fig. 8). These results are consistent with the finding that ExoT has GAP activity for RhoA in vitro (Krall et al., 2000) and in vivo (B. Kazmierczak and J. Engel, manuscript in preparation). Moreover, they suggest that modification of RhoA activity is biologically relevant to P. aeruginosa internalization.

image

Figure 8. Rho–GTP levels increase after infection with an internalized strain of P. aeruginosa. HeLa cells were infected with PA103ΔU (non-invasive) or PA103ΔUΔT (invasive) at a MOI of 20–50. Cell lysates were prepared 1.5 or 3 h after infection and affinity precipitated with GST–TRBD before SDS–PAGE and Western blotting with anti-RhoA antibody (‘Rho–GTP′). Aliquots of lysates were electrophoresed and blotted with anti-RhoA in parallel to determine total RhoA levels (‘total Rho’). The bar graph shows calculated RhoA–GTP levels expressed as a percentage of total RhoA. This experiment was repeated four times; results shown are representative.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The general process of bacterial internalization involves many steps: bacterial binding at specific receptor sites, signalling to the host cell, modification of host signal transduction pathways, membrane and cytoskeletal rearrangements and eventual engulfment of pathogens. Alterations at any step in this pathway can influence the final outcome, namely the successful internalization of a bacterium. We have recently reported that P. aeruginosa internalization by non-professional phagocytes is inhibited by the clostridial protein toxin B, suggesting that the activity of one or more of the small Rho family GTPases is required for the internalization of this pathogen (B. I. Kazmierczak, K. Mostov, J. N. Engel, submitted). We now demonstrate that activation of Rho, either through expression of a constitutively active allele of this protein (RhoAV14) or through activation of endogenous Rho by CNF-1, is sufficient to increase PA103pscJ internalization. Furthermore, epithelial cells infected with internalized strains of P. aeruginosa show measurable increases in Rho–GTP levels over uninfected cells.

A biologically significant role for Rho in Pseudomonas internalization is further supported by the finding that the P. aeruginosa anti-internalization factor ExoT acts as a GAP for RhoA in vitro (Krall et al., 2000) and in vivo (B. Kazmierczak and J. Engel, manuscript in preparation). We have recently demonstrated that mutations of the conserved arginine (Arg149) within the GAP homology domain of ExoT abolish its in vivo GAP activity for RhoA (B. Kazmierczak and J. Engel, manuscript in preparation) and attenuate its ability to block Pseudomonas internalization (Garrity-Ryan et al., 2000). Thus, modulation of endogenous RhoA activity appears to play a role in bacterial regulation of internalization.

Rho activation also increases the internalization of another clinical isolate of P. aeruginosa, strain 6294 (B. Kazmierczak, unpublished results), and of S. flexneri, but does not affect internalization of S. typhimurium. As the latter organism is internalized through a Rho-independent mechanism, Rho activation does not appear to lead to indiscriminate increases in pathogen internalization. A recent report from Lee et al. (2000) has demonstrated a requirement for Rac1 and Cdc42 in the phagocytosis of another P. aeruginosa strain, PAO1, by murine macrophages. In our stably transfected MDCK cell system, however, we have found that the expression of constitutively active Rac1V12 or Cdc42V12 does not promote P. aeruginosa internalization, whereas expression of the dominant negative alleles of Rac1 (Rac1N17) and Cdc42 (Cdc42N17) stimulates PA103pscJ uptake (B. Kazmierczak, T. S. Jou, K. Mostov, J. Engel, unpublished results). The differences between our findings and those of Lee et al. (2000) are likely to reflect differences in the predominant uptake pathways utilized by professional phagocytes, such as macrophages, and by epithelial cells.

Although RhoA activation is sufficient to promote Pseudomonas internalization, we have not been able to demonstrate that it is absolutely required for this process. Inhibition of Rho either by expression of the DN allele RhoAN19 or by treatment of cells with the TAT–C3 fusion toxin does not block bacterial internalization. However, there are several limitations to these experiments. First, inhibition of RhoA profoundly alters tight junctions of polarized cells (Nusrat et al., 1995; Jou et al., 1998), which may itself be sufficient to increase Pseudomonas internalization by increasing bacterial access to basolateral domains of polarized cells. Indeed, disruption of tight junctions with chelators such as EDTA or EGTA is reported to increase P. aeruginosa internalization by 10-fold to > 100-fold (Fleiszig et al., 1997; Pereira et al., 1997; Plotkowski et al., 1999). Thus, by inhibiting RhoA and consequently disrupting tight junctions, we may supply Pseudomonas with a new route of uptake which may be independent of Rho activity. In such a case, the net internalization of Pseudomonas would actually increase. We also find that TAT–C3 treatment of polarized cells disrupts tight junctions in a dose-dependent manner and results in increased internalization. This increase is more pronounced in polarized (day 3) cells; this is a result we would expect as (i) these cells form a ‘tighter’ monolayer than day 1 cells by several criteria (e.g. transepithelial resistance, paracellular diffusion of tracer molecules) and (ii) tight junction disruption has been observed to result in greater relative increases in Pseudomonas internalization in more highly polarized monolayers (Fleiszig et al., 1997). Second, it remains possible that modification of RhoA by TAT–C3 does not inhibit all of its in vivo interactions with downstream effectors; indeed, in vitro studies have shown that a subset of RhoA interactions with downstream effectors that are blocked by toxin B-mediated glucosylation of RhoA are not blocked by C3-mediated modification of RhoA (Sehr et al., 1998).

Several models may explain our findings. We favour a model in which a Rho-dependent pathway mediates P. aeruginosa uptake in epithelial cells. As illustrated in Fig. 9A, bound bacteria may trigger RhoA activation directly, through a still unidentified effector molecule, or indirectly, through interactions with cell-surface receptors, resulting in actin cytoskeleton rearrangements that lead to bacterial internalization. Bacteria expressing the Rho GAP ExoT, however, can inhibit RhoA activation and downstream signalling and thereby remain extracellular. Of note, Rho inhibition mediated by the bacterially translocated protein may be spatially or temporally restricted; we do not, for example, see evidence of tight junction disruption in monolayers exposed to ExoT-expressing bacteria (L. K. Garrity and J. N. Engel, unpublished data). Cells may also possess a Rho-independent uptake pathway; however, this pathway may be basolaterally localized or expressed at low levels in polarized cells and therefore may contribute little to bacterial uptake. Under conditions where the monolayer is disrupted (e.g. by RhoA inhibition or treatment with chelators), bacteria may gain access to this Rho-independent pathway (Fig. 9B). In such an instance, net internalization may increase even if Rho-dependent uptake is blocked. Non-polarized cells lacking distinct apical and basolateral domains, such as HeLa cells, may present bacteria with both Rho-dependent and Rho-independent pathways at all times (Fig. 9C). Because internalization by HeLa cells can be profoundly inhibited by toxin B (B. I. Kazmierczak, K. Mostov and J. N. Engel, submitted), the Rho-independent pathway may itself be regulated by a small GTPase, which our data would suggest is neither Rac1 nor Cdc42. Lastly, it remains formally possible that RhoA does not play a role in bacterial internalization, and that this underlies our inability to block internalization by inhibiting Rho. However, the finding that RhoA activation is sufficient to promote P. aeruginosa internalization combined with the observation that endogenous Rho–GTP levels are modified in concert with bacterial internalization makes this last possibility unlikely.

image

Figure 9. A model for Rho activation in P. aeruginosa internalization.

A. Polarized epithelial cell monolayers internalize P. aeruginosa when RhoA is activated or GTP bound. Rho–GTP levels decrease when bacteria translocate ExoT, a Rho–GAP, into epithelial cells; this blocks bacterial internalization.

B. When monolayer integrity is disrupted by treatment of cells with TAT–C3 or by expressing DN RhoAN19, bacteria gain access to the basolateral pole of epithelial cells. Bacterial uptake may occur here by a RhoA-independent pathway, even though Rho-dependent uptake is blocked.

C. In non-polarized cells, RhoA-dependent and -independent pathways are accessible to bacteria; thus, specific inhibition of RhoA does not completely block bacterial internalization.

Download figure to PowerPoint

The mechanism by which Rho activation leads to increased P. aeruginosa internalization remains to be elucidated. RhoAV14-mediated increases in P. aeruginosa internalization are dependent on actin polymerization, suggesting that they are mediated through the actin cytoskeleton. Activation of Rho leads to the bundling of actin filaments and formation of stress fibres and focal adhesions in numerous cell types (Machesky and Hall, 1997). These phenotypes are thought to depend on the interaction of Rho–GTP with downstream effectors such as ROCK and Dia1 (Amano et al., 1997; Watanabe et al., 1999) and may contribute to pathogen internalization. As an example, the formation of a ‘pseudo-adhesion’ structure has been postulated to underlie the requirement for Rho activity in the complex process of Shigella flexneri internalization by epithelial cells (Mounier et al., 1999; Tran Van Nhieu and Sansonetti, 1999). The Rho-dependent increase in P. aeruginosa internalization that we have described, however, is unlikely to be mediated via a p160ROCK/ROKα-dependent pathway as we have not observed inhibition of internalization with the ROCK-specific inhibitor Y-27632 in RhoAV14-expressing cells (Uehata et al., 1997) (B. Kazmierczak, T.-S. Jou, S. Narumiya and J. Engel, unpublished data). Several other downstream effectors of Rho have been identified which modulate activities other than stress fibre and focal adhesion formation, including protein kinase N, phospholipase D and phosphotidylinositol 4-phospho-5-kinase (Kjoller and Hall, 1999). This last effector mediates the activation of the ezrin–radixin–moesin (ERM) family proteins (Matsui et al., 1999) and has been implicated in the process of S. flexneri internalization (Skoudy et al., 1999; Dumenil et al., 2000). Experiments are currently underway to ask whether any of these effectors may be responsible for the increased internalization we observe following Rho activation.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains

PA103pscJ was created by a Tn5 insertion into the pscJ gene; this strain lacks a functional type III secretion system (Kang et al., 1997; Hauser et al., 1998). PA103pscJ was grown by shaking overnight in Luria–Bertani (LB) broth at 37°C for invasion assays. PA103ΔU has an in frame deletion (aa 330–571) within exoU; the resulting strain is non-cytotoxic, non-invasive and type III secretion competent (L. Garrity, B. Kazmierczak, J. Commolli, A. Hauser and J. Engel, submitted). PA103ΔUΔT is PA103ΔU carrying an axyl/aacC1 cassette in place of aa 36–348 of exoT; the resulting strain is non-cytotoxic, invasive and type III secretion competent (Garrity-Ryan et al., 2000). S. typhimurium SL1344 was a generous gift of Stanley Falkow (Stanford University). This strain was grown standing overnight in high-salt LB (0.3 M NaCl), then diluted 1:50 into medium and grown to an OD600 of 0.1 for invasion assays. S. flexneri M90T was a generous gift of Arturo Zychlinsky (Skirball Institute, NYU). M90T was grown by shaking overnight in Soy Trypticase Broth (Difco), diluted 1:50 into fresh medium and grown to an OD600 of 0.1 for invasion assays. TAT fusion proteins were expressed in strain BL21(DES) pLysS (Stratagene). GST–TRBD was expressed in strain DH5α.

Cell lines and tissue culture

All tissue culture cells were maintained at 37°C in a humidified atmosphere containing 5% CO2. HeLa cells (ATCC CCL-2) were grown in DME H-21 (UCSF Tissue Culture Facility) plus 10% fetal bovine serum (FBS) (Gibco BRL). MDCK type II cells were grown in Eagle's minimal essential medium (MEM) with Earle's balanced salt solution (BSS) (UCSF Tissue Culture Facility) plus 5% FBS. MDCK T23 cells stably transfected with expression vectors for NH2-terminal myc-tagged RhoAN19, RhoAV14, Rac1N17, Rac1V12, Cdc42N17, Cdc42V12 or carrying the empty expression vector pUHD (Jou and Nelson, 1998) were grown in Eagle's MEM with Earle's BSS containing 5% FBS and 20 ng ml−1 doxycycline (Sigma). Transfected cells were induced to express mutant GTPase by replating the cells at low density (< 25% confluent) in the absence of doxycycline for 18–24 h (40 h for RhoAN19). Cells were then trypsinized, counted and plated to 12 mm Transwell filters (Corning) at a density of 1.0–1.5 × 106 cells cm−2. This protocol allowed maximal expression of the myc-tagged proteins, as detected by Western blotting and immunofluorescence (data not shown). Controls were carried out comparing bacterial internalization by ‘induced’ and ‘uninduced’ T23 cells carrying the empty expression vector pUHD; bacterial internalization differed by less than 10% between these two conditions, confirming that no aspect of the induction protocol itself contributes to changes in bacterial internalization (data not shown).

MEM lite contains Minimum Essential Medium Eagle (Sigma), 20 mM HEPES, pH 7.6 (UCSF Tissue Culture Facility) and 4 mM sodium bicarbonate (Sigma). MEM, etc. is composed of MEM lite supplemented with 0.6% (w/v) bovine serum albumin (USB).

Drugs and reagents

Latrunculin A (Molecular Probes) was made up as a 2 mg ml−1 stock in DMSO. Cells were pretreated for 30 min before bacterial infection, and latrunculin A was present throughout the internalization assay. Toxin B (TechLab) was supplied at 0.38 mg ml−1 in PBS. Cells were pretreated for 4 h before bacterial infection and toxin B was present during the first 2 h after bacterial inoculation. Drug was not replenished, however, during bacterial exposure to amikacin as the effects of toxin B have been shown to be irreversible after incubation periods as short as 15 min (Just et al., 1995). We confirmed that neither latrunculin A nor toxin B was toxic to P. aeruginosa at the concentrations used in these experiments by comparing titres of bacteria after growth for 5 h in the presence or absence of these agents; neither inhibited P. aeruginosa viability (data not shown). CNF-1 was the generous gift of Peter Flynn and David Stokoe (UCSF Cancer Center, San Francisco, CA, USA). Antibiotics were obtained from Sigma and prepared as sterile stocks in water or 50% ethanol (doxycycline).

Internalization assays

HeLa cells (5 × 105) were grown for 24–48 h in 24-well plates. PA103pscJ was grown shaking overnight in LB, diluted in MEM, etc. to OD600 0.1 and 200 µl added per well [approximate multiplicity of infection (MOI) = 20]. Following a 2 h incubation with bacteria, cells were washed with MEM, etc., then incubated with MEM plus 400 µg ml−1 amikacin (Sigma). After 2 h, cells were again washed to remove antibiotic, then lysed by incubating them with 1 ml Hanks' Ca2+ Mg2+-free BSS (UCSF Tissue Culture Facility) plus 0.25% Triton X-100 (TX-100) (Sigma) for 30 min. Internalized bacteria were enumerated by plating serial dilutions of cell lysates to LB plates and counting colony-forming units (cfu). All assays were carried out on triplicate wells, and results are reported as the average of three wells ± SD (standard deviation).

MDCK cells were plated into 12 mM Transwell filters (Corning) (0.4 µm pore size) at 1 × 106 cells cm−2 1–3 days before assaying, as indicated. Cells were changed into MEM, etc. before apical infection with bacteria at a MOI of 10–20. After 2 h incubation with bacteria, the cells were washed with MEM, etc., then incubated for a further 2 h with MEM, etc. supplemented with 400 µg ml−1 amikacin. The filters were then cut from their supports and incubated in 1 ml Hanks' Ca2+ Mg2+-free BSS + 0.25% TX-100 for 30 min; complete cell lysis was ensured by vortexing the filters with sterile glass beads twice for 10 s. Internalized bacteria were quantified by plating serial dilutions to LB plates and counting cfu. Results are reported for triplicate wells ± SD or for six wells ± SEM (standard error of the mean).

Basolateral infections of MDCK cells were carried out with the following modifications. MDCK cells were plated to 12 mM Transwell filters (3.0 µm pore size) at 1 × 106 cells cm−2 1–3 days before assay. (Bacteria were unable to cross the 0.4 µm pore filters efficiently.) Bacteria were diluted in MEM, etc. to an OD600 of ≈ 0.6. Filters were infected by placing them directly on top of a 40 µl drop of bacterial suspension (MOI ≈ 25–50); 200 µl of medium was added to the apical surface of the cells during the infection. After 2 h incubation with bacteria, the filters were washed, treated with antibiotic-containing medium and lysed as described above.

Binding assays

MDCK cells were plated into 12 mM Transwell filters (Corning) (0.4 µm pore size) at 1 × 106 cells cm−2 3 days before assaying, as indicated. Cells were changed into MEM, etc. before apical infection with bacteria at a MOI of 10–20. After 2 h incubation with bacteria, the cells were washed three times with ice-cold PBS (Ca2+ Mg2+ free). The filters were then cut from their supports and incubated in 1 ml Hanks' Ca2+ Mg2+-free BSS + 0.25% TX-100 for 30 min; complete cell lysis was ensured by vortexing the filters with sterile glass beads twice for 10 s. Bound plus internalized bacteria were quantified by plating serial dilutions to LB plates and counting cfu. Results are reported for triplicate wells ± SD or for six wells ± SEM (standard error of the mean).

FITC-inulin assays

FITC-inulin (Sigma) was prepared as a 5 mg ml−1 stock in PBS and stored as single-use aliquots at −20°C. To assay paracellular diffusion, FITC-inulin was added to the apical compartment of Transwells at a final concentration of 20 µg ml−1; the apical and basolateral compartments were sampled after 2 h incubation at 37°C. Fluorescence was quantified using a Cytofluor (480/530, 70% gain). Results are reported as the average of triplicate determinations (± SD).

Construction of TAT–HA fusion protein expression vectors pTAT–HA–C3 transferase and pTAT–HA–GFP

A SalI fragment encoding C3 transferase plus a six-amino-acid N-terminal linker (DLQACN) was digested from pKG-9E10.C3 (a generous gift from Art Alberts, UCSF Cancer Center), gel purified and subcloned into the XhoI site of pTAT–HA (Nagahara et al., 1998) (kindly provided by Steve Dowdy, Washington University). The in frame fusion was confirmed by sequencing across the cloning joint. GFP was amplified from GFPmut3 (kind gift of Brendan Cormack, Johns Hopkins University) (Cormack et al., 1996) using the primers FOR/GFP (5′-GGAGCCATGGGTAAAGGAGAAGAATTATTC-3′) and REV/GFP (5′-CCATGAAGCTTGCATGCCTG-3′) with Pfu Turbo polymerase (Stratagene). The PCR product was gel purified, Qiaex extracted (Qiagen), digested with NcoI (NEB) and SphI (NEB) and subcloned into the NcoI–SphI sites of pTAT–HA using standard molecular biology techniques. Positive clones were picked by virtue of fluorescence using 480/530 filters.

Purification of TAT fusion proteins

The pTAT–HA–C3 transferase and pTAT–HA–GFP plasmids were transformed into BL21(DES)pLysS (Stratagene); glycerol stocks were prepared and stored at −80°C. These were used to start 100 ml overnight cultures (LB plus 100 µg ml−1 ampicillin; Sigma) at 37°C. The following morning, 1 l cultures were inoculated from these overnight cultures, grown for 1–1.5 h at 37°C before addition of 0.1 mM IPTG (Gibco BRL), and induced for a further 4–4.5 h. Cultures were harvested by centrifugation at 4°C (5000 g for 10 min), washed once with 50 ml ice-cold PBS (Ca2+ Mg2+ free) and resuspended in 10 ml Buffer Z (8 M urea/100 mM NaCl/20 mM HEPES, pH 8.0). Samples were sonicated on ice three or four times for 15 s, then clarified by centrifugation at 20 000 g for 10 min at 4°C. Imidazole (Sigma) was added to the supernatant to a final concentration of 20 mM before the supernatant was loaded at room temperature (RT) onto a 7.5 ml Ni-NTA (Qiagen) column pre-equilibrated with Buffer Z plus 20 mM imidazole. The column was allowed to flow by gravity. The column was washed with 50 ml Buffer Z plus 50 mM imidazole, then eluted with 10 ml steps of Buffer Z plus 100, 250, 500 or 1000 mM imidazole. Fractions were analysed by SDS–PAGE, and those containing only fusion protein (commonly the 250 mM imidazole step) were desalted over a PD-10 column (Amersham Pharmacia Biotech) pre-equilibrated with MEM lite plus aprotinin (10 µg ml−1) (Roche BMB), pefabloc (2 mM) (Roche BMB) and leupeptin (20 µg ml−1) (Roche BMB). Protein concentration was determined using the Bradford assay (Bio-Rad) with bovine serum albumin as a standard. The protein was then added to tissue culture cells at the concentrations and times indicated in each experiment. FBS was supplemented as necessary for a final concentration of 5%.

TRBD affinity precipitation assays

pGEX-2T-TRBD, a vector encoding the fusion protein GST–TRBD, was generously provided by Xiang-Dong Ren and Martin Schwartz (Scripps Institute). GST–TRBD was purified as described previously (Ren et al., 1999), flash frozen and stored at −80°C as single-use aliquots. HeLa cells (1 × 106 per 10 cm dish) or MDCK cells (2 × 107 per 7 cm Transwell filter) were plated 24 h before infection with PA103ΔU or PA103ΔUΔT at a MOI of 20–50. After 1–3 h, cells were washed twice with ice-cold TBS, pH 7.2, and lysed according to Ren et al. (1999). Twenty microlitres of each cleared lysate was set aside for determination of total RhoA; 500 µl of each lysate was incubated with 40–50 µg of GST–TRBD immobilized on glutathione-Sepharose 4B (Pharmacia) for 45 min at 4°C. Samples were washed three times with lysis buffer, then resuspended in 2× sample buffer. The lysate aliquot and affinity precipitated samples were subjected to SDS–PAGE through 13% acrylamide gels; these were then transferred to PVDF (Millipore) using a semidry transfer technique. Blots were blocked overnight in 5% milk powder in TBST (4°C) before incubation with anti-RhoA (Santa Cruz) diluted 1:200 in 5% milk/TBST for 2 h. Blots were washed with 5% milk/TBST, then incubated with HRP-conjugated goat anti-mouse IgG (Bio-Rad) diluted 1:2000 in 5% milk/TBST for 1 h. Blots were developed with ECL-Plus (Amersham Pharmacia) according to the manufacturer's instructions and then exposed to film for 3–300 s. Bands were quantified by analysing scanned films with IPLab Gel H software.

Fixation and immunofluorescence

Samples were washed 3× with PBS, fixed for 30 min with 4% paraformaldehyde (Sigma) in PBS and quenched for 10 min with 75 mM NH4Cl/20 mM glycine in PBS. Samples were blocked and permeabilized in PBS/0.7% fish scale gelatine (Sigma)/0.025% saponin (Sigma) (PFS) plus RNAse (100 µg ml−1) for 15–30 min at 37°C before staining. Expression of the myc-tagged small GTPase mutant alleles was confirmed by staining the samples with 9E10 (1:250) (Santa Cruz Biotechnology) in PFS for 60 min at 37°C. Other monoclonal antibodies used in these studies include anti-gp135 (1:10 000) (gift of George Ojakian, SUNY Downstate Medical Center), anti-ZO-1 (1:200) (Chemicon) and anti-E-cadherin (RR-1, 1:1) (gift of Barry Gumbiner, Sloan Kettering). Samples were routinely washed 4× with PFS before staining with Alexa 488-conjugated anti-mouse or anti-rat IgG (1:500) (Molecular Probes) plus Texas Red phalloidin (1:200) (Molecular Probes) in PFS for 30–45 min at 37°C before washing and mounting with Prolong Antifade (Molecular Probes). Samples were imaged on a Bio-Rad 1024 confocal laser microscope using a 60× objective, pseudocoloured and assembled into figures using Adobe Photoshop 5.0.

Statistical analysis

Statistical analysis was performed using instat 1.12 software.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Eric Brown and members of the Engel laboratory for critical reading of this manuscript, Melissa Woodrow for helpful discussions regarding the TRBD affinity precipitation assay, Arturo Zychlinsky and Stanley Falkow for bacterial strains, Art Alberts, Brendan Cormack, Steve Dowdy, Peter Flynn, Xiang-Dong Ren, Martin Schwartz and David Stokoe for generous gifts of plasmids and reagents, and George Ojakian and Barry Gumbiner for generous gifts of antibodies. This work was supported by a Howard Hughes Medical Institute Physician Postdoctoral Fellowship (B.I.K.), the American Lung Association (J.N.E.) and NIH grants K08 AI01636 (B.I.K.), RO1 HL55980 (K.M.) and R01 AI42806 (J.N.E.). J.N.E. is an established investigator of the American Lung Association.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Adam, T., Giry, M., Boquet, P., Sansonetti, P. (1996) Rho-dependent membrane folding causes Shigella entry into epithelial cells. EMBO J 15: 33153321.
  • Aktories, K. (1997) Bacterial toxins that target Rho proteins. J Clin Invest 99: 827829.
  • Amano, M., Chihara, K., Kimura, K., Fukata, Y., Nakamura, N., Matsuura, Y., Kaibuchi, K. (1997) Formation of actin stress fibers and focal adhesions enhanced by Rho-kinase. Science 275: 13081311.DOI: 10.1126/science.275.5304.1308
  • Black, D.B. & Bliska, J.B. (2000) The RhoGAP activity of the Yersinia pseudotuberculosis cytotoxin YopE is required for antiphagocytic function and virulence. Mol Microbiol 37: 515527.DOI: 10.1046/j.1365-2958.2000.02021.x
  • Chen, L.-M., Hobbie, S., Galan, J.E. (1996) Requirement of CDC42 for Salmonella-induced cytoskeletal and nuclear responses. Science 274: 21152118.DOI: 10.1126/science.274.5295.2115
  • Cormack, B.P., Valdivia, R.H., Falkow, S. (1996) FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173: 3338.DOI: 10.1016/0378-1119(95)00685-0
  • Cowell, B.A., Chen, D.Y., Frank, D.W., Vallis, A.J., Fleiszig, S.M.J. (2000) ExoT of cytotoxic Pseudomonas aeruginosa prevents uptake by corneal epithelial cells. Infect Immun 68: 403406.
  • Dumenil, G., Sansonetti, P., Tran Van Nhieu, G. (2000) Src tyrosine kinase activity down-regulates Rho-dependent responses during Shigella entry into epithelial cells and stress fibre formation. J Cell Sci 113: 7180.
  • Evans, D.J., Frank, D.W., Finck-Barbançon, V., Wu, C., Fleiszig, S.M. (1998) Pseudomonas aeruginosa invasion and cytotoxicity are independent events, both of which involve protein tyrosine kinase activity. Infect Immun 66: 14531459.
  • Feig, L.A. (1999) Tools of the trade: use of dominant-inhibitory mutants of Ras-family GTPases. Nature Cell Biol 1: E25E27.
  • Finlay, B.B. & Falkow, S. (1997) Common themes in microbial pathogenicity revisited. Microbiol Mol Biol Rev 61: 136169.
  • Flatau, G., Lemichez, E., Gauthier, M., Chardin, P., Paris, S., Fiorentini, C., Boquet, P. (1997) Toxin-induced activation of the G protein p21 Rho by deamidation of glutamine. Nature 387: 729733.DOI: 10.1038/42743
  • Fleiszig, S., Vallas, V., Jun, C., Mok, L., Balkovetz, D., Roth, M., Mostov, K. (1998) Susceptibility of epithelial cells to Pseudomonas aeruginosa invasion and cytotoxicity is upregulated by hepatocyte growth factor. Infect Immun 66: 34433446.
  • Fleiszig, S.M., Zaidi, T.S., Fletcher, E.L., Preston, M.J., Pier, G.B. (1994) Pseudomonas aeruginosa invades corneal epithelial cells during experimental infection. Infect Immun 62: 34853493.
  • Fleiszig, S.M., Zaidi, T.S., Pier, G.B. (1995) Pseudomonas aeruginosa invasion of and multiplication within corneal epithelial cells in vitro. Infect Immun 63: 40724077.
  • Fleiszig, S.M., Evans, D.J., Do, N., Vallas, V., Shin, S., Mostov, K. (1997) Epithelial cell polarity affects susceptibility to Pseudomonas aeruginosa invasion and cytotoxicity. Infect Immun 65: 28612867.
  • Fleiszig, S.M.J., Zaidi, T.S., Preston, M.J., Grout, M., Evans, D.J., Pier, G.B. (1996) Relationship between cytotoxicity and corneal epithelial cell invasion by clinical isolates of Pseudomonas aeruginosa. Infect Immun 64: 22882294.
  • Garrity-Ryan, L., Kazmierczak, B., Kowal, R., Commolli, J., Hauser, A., Engel, J. (2000) The arginine finger domain of ExoT is required for actin cytoskeleton disruption and inhibition of internalization of Pseudomonas aeruginosa into epithelial cells and macrophages. Infect Immun in press.
  • Hall, A. (1998) Rho GTPases and the actin cytoskeleton. Science 279: 509514.DOI: 10.1126/science.279.5350.509
  • Hardt, W.-D., Chen, L.M., Schubel, K.E., Bustelo, X.R., Galan, J.E. (1998) S. typhimurium encodes an activator of Rho GTPases that induces membrane ruffling and nuclear responses in host cells. Cell 93: 815822.
  • Hauser, A.R., Kang, P.J., Fleiszig, S.J.M., Mostov, K., Engel, J. (1998) Defects in type III secretion correlate with internalization of Pseudomonas aeruginosa by epithelial cells. Infect Immun 66: 14131420.
  • Jou, T.-S. & Nelson, W.J. (1998) Effects of regulated expression of mutant RhoA and Rac1 small GTPases on the development of epithelial (MDCK) cell polarity. J Cell Biol 142: 85100.
  • Jou, T.-S., Schneeberger, E., Nelson, W.J. (1998) Structural and functional regulation of tight junctions by RhoA and Rac1 small GTPases. J Cell Biol 142: 101115.
  • Just, I., Selzer, J., Wilm, M., Von Eichel-Streiber, C., Mann, M., Aktories, K. (1995) Glucosylation of Rho proteins by Clostridium difficile toxin B. Nature 375: 500503.
  • Kang, P.J., Hauser, A.R., Apodaca, G., Fleiszig, S., Wiener-Kronish, J., Mostov, K., Engel, J.N. (1997) Identification of Pseudomonas aeruginosa genes required for epithelial cell injury. Mol Microbiol 24: 12491262.
  • Kjoller, L. & Hall, A. (1999) Signaling to Rho GTPases. Exp Cell Res 253: 166179.DOI: 10.1006/excr.1999.4674
  • Krall, R., Schmidt, G., Aktories, K., Barbieri, J.T. (2000) Pseudomonas aeruginosa ExoT is a Rho GTPase-activating Protein. Infect Immun 68: 60666068.
  • Lamaze, C., Fujimoto, L., Yin, H., Schmid, S. (1997) The actin cytoskeleton is required for receptor-mediated endocytosis in mammalian cells. J Biol Chem 272: 2033220335.
  • Lee, D.J., Cox, D., Li, J., Greenberg, S. (2000) Rac1 and Cdc42 are required for phagocytosis, but not NF-kappaB-dependent gene expression, in macrophages challenged with Pseudomonas aeruginosa. J Biol Chem 275: 141146.
  • Lerm, M., Selzer, J., Hoffmeyer, A., Rapp, U.R., Aktories, K., Schmidt, G. (1999) Deamidation of Cdc42 and Rac by Escherichia coli cytotoxic necrotizing factor 1: activation of c-Jun N-terminal kinase in HeLa cells. Infect Immun 67: 496503.
  • Machesky, L.M. & Hall, A. (1997) Role of actin polymerization and adhesion to extracellular matrix in Rac- and Rho-induced cytoskeletal reorganization. J Cell Biol 138: 913926.
  • Matsui, T., Yonemura, S., Tsukita, S., Tsukita, S. (1999) Activation of ERM proteins in vivo by Rho involves phosphatidylinositol 4-phosphate 5-kinase and not ROCK kinases. Curr Biol 9: 12591262.DOI: 10.1016/s0960-9822(99)80508-9
  • Mecsas, J., Raupach, B., Falkow, S. (1998) The Yersinia Yops inhibit invasion of Listeria, Shigella and Edwardsiella but not Salmonella into epithelial cells. Mol Microbiol 28: 12691281.
  • Morii, N. & Narumiya, S. (1995) Preparation of native and recombinant Clostridium botulinum C3 ADP-ribosyltransferase and identification of Rho proteins by ADP-ribosylation. Methods Enzymol 256: 196206.
  • Mounier, J., Laurent, V., Hall, A., Fort, P., Carlier, M.-F., Sansonetti, P.J., Egile, C. (1999) Rho family GTPases control entry of Shigella flexneri into epithelial cells but not intracellular motility. J Cell Sci 112: 20692080.
  • Nagahara, H., Vocero-Akbani, A.M., Snyder, E.L., Ho, A., Latham, D.G., Lissy, N.A., et al. (1998) Transduction of full-length TAT fusion proteins into mammalian cells: TAT-p27Kip1 induces cell migration. Nature Med 4: 14491452.
  • Nusrat, A., Giry, M., Turner, J.R., Colgan, S.P., Parkos, C.A., Carnes, D., et al. (1995) Rho protein regulates tight junctions and perijunctional actin organization in polarized epithelia. Proc Natl Acad Sci USA 92: 1062910633.
  • Von Pawel-Rammingen, U.M., Telpnev, V., Schmidt, G., Aktories, K., Wolf-Watz, H., Rosqvist, R. (2000) GAP activity of the Yersinia YopE cytotoxin specifically targets the rho pathway: a mechanism for disruption of actin microfilament structure. Mol Microbiol 36: 737748.DOI: 10.1046/j.1365-2958.2000.01898.x
  • Pereira, S.H.M., Cervante, M.P., De Bentzmann, S., Plotkowski, M.C. (1997) Pseudomonas aeruginosa entry into Caco-2 cells is enhanced in repairing wounded monolayers. Microb Pathog 23: 249255.DOI: 10.1006/mpat.1997.0153
  • Plotkowski, M.C., De Bentzmann, S., Pereira, S.H., Zahm, J.M., Bajolet-Laudinat, O., Roger, P., Puchelle, E. (1999) Pseudomonas aeruginosa internalization by human epithelial respiratory cells depends on cell differentiation, polarity, and junctional complex integrity. Am J Respir Cell Mol Biol 20: 880890.
  • Ren, X.-D., Kiosses, W.B., Schwartz, M.A. (1999) Regulation of the small GTP-binding protein Rho by cell adhesion and the cytoskeleton. EMBO J 18: 578585.DOI: 10.1093/emboj/18.3.578
  • Rudolph, M.G., Weise, C., Mirold, S., Hillenbrand, B., Bader, B., Wittinghofer, A., Hardt, W.-D. (1999) Biochemical analysis of SopE from Salmonella typhimurium, a highly efficient guanosine nucleotide exchange factor for RhoGTPases. J Biol Chem 274: 3050130509.
  • Schmidt, G., Sehr, P., Wilm, M., Selzer, J., Mann, M., Aktories, K. (1997) Gln 63 of Rho is deamidated by Escherichia coli cytotoxic necrotizing factor-1. Nature 387: 725729.DOI: 10.1038/42735
  • Sehr, P., Joseph, G., Genth, H., Just, I., Pick, E., Aktories, K. (1998) Glucosylation and ADP ribosylation of Rho proteins: effects on nucleotide binding, GTPase activity and effector coupling. Biochemistry 37: 52965304.
  • Skoudy, A., Tran Van Nhieu, G., Mantis, N., Arpin, M., Mounier, J., Gounon, P., Sansonetti, P. (1999) A functional role for ezrin during Shigella flexneri entry into epithelial cells. J Cell Sci 112: 20592068.
  • Tran Van Nhieu, G. & Sansonetti, P.J. (1999) Mechanism of Shigella entry into epithelial cells. Curr Opin Microbiol 2: 5155.DOI: 10.1016/s1369-5274(99)80009-5
  • Uehata, M., Ishizaki, T., Satoh, H., Ono, T., Kawahara, T., Morishita, T., et al. (1997) Calcium sensitization of smooth muscle mediated by a Rho-associated protein kinase in hypertension. Nature 389: 990994.
  • Watanabe, N., Kato, T., Fujita, A., Ishizaki, T., Narumiya, S. (1999) Cooperation between mDia1 and ROCK in Rho-induced actin reorganization. Nature Cell Biol 1: 136143.
  • Watarai, M., Kamata, Y., Kozaki, S., Sasakawa, C. (1997) rho, a small GTP-binding protein, is essential for Shigella invasion of epithelial cells. J Exp Med 185: 281292.