In Vivo Rho GTPase-Activating Protein Activity of Pseudomonas aeruginosa Cytotoxin ExoS

Abstract

ExoS is a bifunctional type III cytotoxin secreted by Pseudomonas aeruginosa, which comprises a C-terminal ADP ribosyltransferase domain and an N-terminal Rho GTPase-activating protein (GAP) domain. In vitro, ExoS is a Rho GAP for Rho, Rac, and Cdc42; however, the in vivo modulation of Rho GTPases has not been addressed. Using a transient transfection system and delivery by P. aeruginosa, interactions were examined between the Rho GAP domain of ExoS and Rho GTPases in CHO cells. Rho GTPases were expressed as green fluorescent protein (GFP) fusion proteins to facilitate quantitation. GFP fusions of wild-type and dominant active Rho, Rac, and Cdc42 localized to discrete regions of CHO cells and appeared functional based upon their modulation of the actin cytoskeleton. Coexpression of the Rho GAP domain of ExoS changed the intracellular distribution of GFP-Rac and GFP-Cdc42 from a predominately membrane location to a cytosolic location. Coexpression of the Rho GAP domain of ExoS did not change the distribution of GFP-Rho, which was primarily in the cytosol. Coexpression of dominant active Rac (DARac) and DACdc42 inhibited actin reorganization by the Rho GAP domain but did not maintain the formation of actin stress fibers, which indicated that Rho had been inactivated. Similar results were observed when ExoS was delivered into CHO cells by P. aeruginosa. These data indicate that in vivo the Rho GAP activity of ExoS stimulates the reorganization of the actin cytoskeleton by inhibition of Rac and Cdc42 and stimulates actin stress fiber formation by inhibition of Rho.

Pseudomonas aeruginosa is a gram-negative opportunistic pathogen that causes life-threatening infections in cystic fibrosis patients, individuals with burn wounds, and the immunocompromised (13). P. aeruginosa pathogenicity involves cell-associated and secreted virulence factors as well as intrinsic resistance to antibiotics. P. aeruginosa produces four type III cytotoxins (ExoS, ExoT, ExoU, and ExoY), which are delivered directly into the eukaryotic cell (9). ExoS is a bifunctional cytotoxin (4). The N terminus of ExoS is a GTPase-activating protein (GAP) for Rho GTPases, whereas the C terminus of ExoS is a 14-3-3-dependent ADP ribosyltransferase. Transfection of the N terminus of ExoS into cultured cells induces actin reorganization, whereas transfection of the C terminus results in cell death. While cell death requires ADP ribosyltransferase activity, the mechanism responsible for eliciting cytotoxicity has not been defined. ExoS is a Rho GAP for Rho, Rac, and Cdc42 in vitro (14), but the in vivo specificity of Rho GAP activity has not been established.

Rho GTPases, Rho, Rac, and Cdc42, maintain the organization of the actin cytoskeleton (23). Activation of Cdc42 stimulates the formation of filopodia (microspikes) and contributes to cell polarity (22), and activation of Rac stimulates lamellipodia (membrane ruffling) and contributes to cell motility. Activation of Rho stimulates the reorganization of preexisting actin filaments into stress fibers (28). Rho GTPases are molecular switches that are controlled through nucleotide exchange (15) and are active in the GTP form and inactive in the GDP form. Nucleotide exchange is regulated by guanine nucleotide exchange factors (24), which activate Rho by stimulating the exchange of GDP for GTP, and by GAPs, which inactivate Rho by stimulating the hydrolysis of the γ-phosphate of GTP, yielding Rho-GDP (30). Rho GTPases are synthesized in the cytosol and isoprenylated (geranyl geranylated) at the C-terminal CAAX (17), which tethers the Rho GTPase to the cell membrane. Guanine nucleotide dissociation inhibitor (GDI), Rho-GDI, can extract Rho-GDP from the cell membrane and sequester the isoprenylated tail to prevent membrane attachment (31). The stimulation of the release of GDI from the Rho-GDI complex is unclear, but it may involve conformational changes upon the binding of PIP₂ and ERM proteins (20). Thus, Rho-GTP is localized to the cell membrane while Rho-GDP is present in the cytosol (7).

Rho GTPases are preferred targets for bacterial toxins, modulating the organization of the actin skeleton and allowing invasion into nonprofessional cells and phagocytosis into professional phagocytes. These toxins modify the activity of Rho GTPases through covalent modification or regulation of the nucleotide state. Toxins that covalently modify Rho include C3, which inactivates Rho by ADP ribosylation of Arg41, Clostridium difficile toxins A and B, which inactivate Rho, Rac, and Cdc42 by monoglycosylation at Thr37, and CNF, which activates Rho, Rac, and Cdc42 by deamidation of Q63/61 (1). Toxins that regulate the nucleotide state of Rho include ExoS (14), ExoT (18), SptP (10), and YopE (5, 33), which are GAPs for various Rho GTPases, and SopE, which is a guanine nucleotide exchange factor for Rac and Cdc42 (16). This study addresses the in vivo activity of the Rho GAP domain of ExoS on Rho GTPases.

MATERIALS AND METHODS

Plasmid construction.

cDNAs encoding RhoA, Rac1, and Cdc42 were subcloned into the BamHI and EcoRI restriction sites of pGEX4T-1 and subsequently cloned into the BglII and EcoRI restriction sites of pEGFPC1 to produce a fusion protein carrying enhanced green fluorescent protein (EGFP) followed in frame by the respective Rho GTPase. Eukaryotic expression vectors encoding ExoS(1-234) and ExoS(1-234)R146K have been described previously (25, 26).

Site-directed mutagenesis.

DARhoAQ63E and DARac1Q61E were engineered by using a PCR-based site-directed mutagenesis kit (QuickChange; Stratagene). DNA encoding Cdc42Q61E was the gift of Klaus Aktories. Primers for mutagenesis were confirmed by sequence analysis and include the RhoAQ63E positive (5′-GGGAACTGGTCCTTGCTGAAAACTATGAGCAGCATG-3′) and negative (5′-CATGCTTGCTCATAGTTTTCAGCAAGGACCAGTTCCC-3′) primers and the Rac1Q61E positive (5′-GGGATACAGCTGGAGAAGAAGATTATCACAGATTACGCC-3′) and negative (5′-GGCGTAATCTGTCATAATCTTCTTCTCCAGCTGTATCCC-3′) primers.

Cell culture and transfection protocols.

CHO cells were cultured in Ham F-12 medium supplemented with 10% newborn calf serum, 1.4% sodium bicarbonate, and 0.5% penicillin-streptomycin at 37°C in 5% CO₂. Subconfluent CHO cells were transfected with the indicated plasmid by using Lipofectamine Plus (GIBCO/BRL) as suggested by the manufacturer. Total transfected DNA was normalized with pCMV-luciferase.

(i) Expression and subcellular distribution of Rho GTPases.

The expression and subcellular distribution of Rho GTPases were determined on subconfluent lawns of CHO cells in 100-mm-diameter dishes. CHO cells were transfected with 1.0 μg of the indicated pEGFP-Rho GTPase and 0.1 μg of pEGFP alone or with 0.5 μg of pExoS(1-234) or pExoS(1-234)R146K. Cell pellets were suspended with homogenization buffer 1 (250 mM sucrose, 3 mM imidazole [pH 7.4]) and centrifuged at 5,000 rpm in a Beckman Microfuge at 4°C for 5 min. Cell pellets were suspended in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer for analysis of expression or homogenization buffer 2 (250 mM sucrose, 3 mM imidazole, 0.5 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride [pH 7.4]) for determination of subcellular distribution. Cells were broken by passage (14 times) through a 25-gauge 5/8-inch needle. Broken cells were centrifuged at 250 × g for 5 min at 4°C, and the postnuclear supernatant was then centrifuged at 100,000 × g for 30 min at 4°C. The supernatant (cytosolic fraction) was mixed with an equal volume of SDS sample buffer, and the insoluble material (membrane fraction) was dissolved in homogenization buffer 2 containing 1% Triton X-100 and mixed with an equal volume of SDS sample buffer. Fractions were boiled at 100°C for at least 5 min and subjected to SDS-PAGE and Western blot analysis.

(ii) Cellular location of Rho GTPases and measurement of the effects of dominant active Rho GTPases.

Fluorescence microscopy of subconfluent lawns of CHO cells in 12-well plates was used to determine the cellular location of Rho GTPases and to measure the ability of dominant active Rho GTPases to stimulate reorganization of the actin cytoskeleton through the Rho GAP domain. CHO cells were transfected with 0.1 or 0.2 μg of pEGFP Rho GTPase alone or with 0.1 μg of pExoS(1-234) or pExoS(1-234)R146K. Twenty-four hours posttransfection, the cells were washed in phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde, and examined by fluorescence microscopy. For detection of the actin cytoskeleton, cells were fixed for 10 min, treated with 0.1% Triton X-100 in PBS for 5 min, and stained with Alexa Fluor 546 phalloidin (0.2 μM; Molecular Probes) in PBS with 1% bovine serum albumin. After 20 min, the cells were washed and examined by fluorescence microscopy.

Infection of CHO cells with P. aeruginosa.

CHO cells were transfected with the indicated plasmid and then infected at a multiplicity of infection (MOI) of approximately 8:1 (bacterium/CHO cell ratio). The amount of P. aeruginosa was quantified by using the following absorbance calculation: optical density at 540 nm = 4 × 10⁸ bacteria. At confluent growth, the number CHO cells was estimated to be 4 × 10⁶ cells/85-mm-diameter dish. Cells were washed with 12 ml of PBS and incubated with 6 ml of serum-free Ham F-12 medium containing 0.1% NaHCO₃. P. aeruginosa PA103 ΔexoU exoT::Tc(pUCP) or P. aeruginosa PA103 ΔexoU exoT::Tc(pUCP-ExoS) was added to CHO cells under the conditions of 5% CO₂ at 37°C for 3 h. At this time, the medium was removed, and the CHO cells were washed with 10 ml of PBS and then fixed in 4% paraformaldehyde in PBS. Transfected cells were identified by fluorescence microscopy and examined for morphology or stained with Alexa-546 phalloidin for actin content as described above.

SDS-PAGE and Western blot analysis.

Protein fractions were subjected to SDS-13.5% PAGE followed by Western blot analysis with a monoclonal anti-green fluorescent protein (GFP) antibody for GFP detection or polyclonal anti-hemagglutinin (HA) antibody for HA detection. The enhanced chemiluminescence-conjugated secondary antibodies were goat anti-mouse immunoglobulin G to detect anti-GFP antibody and rabbit anti-mouse immunoglobulin G to detect anti-HA antibody. Secondary antibodies were detected by using ECL Super Signal (Pierce Biochemicals). Densitometry was performed on an alpha photo imager. The concentration of GFP-Rho GTPase in lysates was determined by coanalysis with authentic GFP (BAbCO). The total GFP-Rho GTPase in a cell lysate was calculated, divided by the number of transfected cells, and expressed in nanograms of Rho GTPase/10⁶ CHO cells.

RESULTS

Expression of the Rho GAP domain of ExoS in CHO cells.

While the N terminus of ExoS (residues 1 to 234) is a Rho GAP for Rho, Rac, and Cdc42 in vitro, the in vivo targets of this Rho GAP have not been determined. The in vivo modulations of Rho GTPases were analyzed following transient transfection of CHO cells with pExoS(1-234). This system was chosen to eliminate the complication of interpreting the effect of Rho GAP activity during a bacterial infection. Since ExoS is cytotoxic to cultured cells, the GAP domain was analyzed in these studies. Transfection of the GAP domain elicits a similar reorganization of the actin cytoskeleton as observed upon delivery of the GAP domain during an infection by P. aeruginosa (26), which supported the use of this delivery system. Transfection conditions were established to express the Rho GAP domain of ExoS [pExoS(1-234)] at levels comparable to expression during a bacterial infection (25). The modulations of Rho GTPases by ExoS delivered by P. aeruginosa were also examined to address issues pertaining to the physiological delivery of the toxin.

Expression of GFP-Rho GTPases in CHO cells.

To facilitate quantitative analysis of in vivo modulation of the Rho GTPases, N-terminal GFP fusion proteins were engineered for RhoA, Rac1, and Cdc42 (Fig. 1A). GFP-Rho GTPases have been shown to be functional in vivo (21). CHO cells transiently transfected with each GFP-Rho GTPase were subjected to expression analysis and direct fluorescence microscopy. GFP immunoreactive proteins were detected in the lysates of cells transfected with DNA encoding each Rho GTPase fusion protein at the apparent molecular weight consistent with expression of the full-length fusion protein (Fig. 1B). GFP-Rac1 and GFP-Cdc42 were expressed as a single immunoreactive protein, whereas the expression of GFP-RhoA was more complex. The major immunoreactive band for GFP-RhoA migrated at the apparent molecular weight consistent with the expression of the full-length fusion protein with two additional immunoreactive bands detected, one with a greater apparent molecular weight and one with a lower apparent molecular weight. The more rapidly migrating band appeared to be a degradation product, and while the origin of the higher molecular weight immunoreactive fusion product was not defined, it may represent a posttranslational modification. The addition of protease inhibitors did not influence the amount of degradation product expressed by pGFP-RhoA. To simplify the analysis, only the immunoreactive band that corresponded to full-length GFP-RhoA was quantified. The expression of transfected GFP-Rho GTPases in CHO cells was determined by comparative immunoblotting against authentic GFP (Table 1). Transfection conditions were established where GFP-Rho GTPases were expressed at levels about 1.5- to 2-fold above those of the endogenous Rho GTPases (21). Under these conditions, coexpression of ExoS(1-234) or ExoS(1-234)R146K had a limited effect on the expression of GFP-Rho GTPases (Table 1).

FIG. 1.

(A) Construction of eukaryotic expression vectors for GFP-Rho GTPases. cDNAs encoding RhoA, Rac1, and Cdc42 were subcloned into pEGFPC1 in frame with GFP. Site-directed mutagenesis generated dominant active forms of GFP-RhoAQ63E, GFP-Rac1Q61E, and Cdc42Q61E. SV40, simian virus 40; Neo/Km, neomycin and kanamycin. (B) Steady-state expression of GFP-Rho GTPases. pEGFPC1 (0.2 μg) alone or with the indicated pGFP-Rho GTPase (0.2 μg) was transiently transfected into CHO cells in six-well dishes. After 24 h, the cells were harvested in SDS sample buffer containing 1% protease inhibitor cocktail III (Calbiochem) and 0.5 mM EDTA and boiled at 100°C for 5 min. Cell lysates were subjected to SDS-13.5% PAGE and Western blot analysis with anti-GFP antibody. Purified GFP translated from the pEGFPN1 vector (R-GFP, leftmost lane) has a lower molecular weight and migrates faster than GFP translated from pEGFPC1 (GFPC1). Molecular weights are indicated to the left of the gel.

TABLE 1.

Expression of GFP-Rho GTPases in CHO cells^a

Rho GTPase	Concn (ng/106 cells)
Rac1	107 ± 9
Rac1 + Sb	91 ± 12
Rac1 + Rc	180 ± 50
DARac1	160 ± 38
DARac1 + S	129 ± 21
DARac1 + R	177 ± 44
Cdc42	100 ± 2
Cdc42 + S	100 ± 12
Cdc42 + R	102 ± 14
DACdc42	138 ± 48
DACdc42 + S	140 ± 35
DACdc42 + R	148 ± 42
RhoA	90 ± 9
RhoA + S	95 ± 15
RhoA + R	86 ± 26
DARhoA	80 ± 20
DARhoA + S	86 ± 16
DARhoA + R	83 ± 21

^aCHO cells were cotransfected with 1.0 μg of pGFP-RhoGTPase or pGFP-DARhoGTPase and 0.1 μg of pGFP alone or with 0.5 μg of ExoS(1–234) or ExoS(1–234)R146K. After 24 h, the cell lysates were subjected to SDS-13.5% PAGE along with a serial dilution of authentic GFP. Western blotting against anti-GFP antibody allowed quantitation of protein expression against authentic GFP. The student t test indicated that there was no significant difference (P > 0.05) in expression of each GTPase alone or relative to that coexpressed with ExoS(1–234) or ExoS(1-234)R146K.

^bS, ExoS(1-234).

^cR, ExoS(1-234)R146K.

Each GFP-Rho GTPase possessed unique subcellular distribution (Fig. 2A). GFP-RhoA was predominantly in the cytosol with some perinuclear accumulation. DAPI (4′,6′-diamidino-2-phenylindole) staining showed that GFP-RhoA did not localize in the nucleus (data not shown). CHO cells expressing GFP-RhoA possessed a more defined network of actin stress fibers than did GFP-transfected CHO cells, which indicated that GFP-RhoA was functional. GFP-Rac1 localized to the plasma membrane, nuclear region, and cytosol with transfected cells showing membrane ruffling. GFP-Cdc42 was expressed in CHO cells predominantly in the plasma membrane and had some cytosolic and nuclear localization. CHO cells transiently transfected with GFP-Cdc42 showed a unique morphology. DAPI staining showed that GFP-Cdc42 was not within the nucleus (Fig. 2A), and when overstained for F-actin, it possessed filopodia observable at higher magnifications (data not shown). CHO cells expressing GFP-RhoA or GFP-Rac1 also contained enhanced actin stress fibers. This indicated an activated state of the Rho GTPases in CHO cells and that Rac sequentially activated Rho.

FIG. 2.

Inhibition of stress fiber formation by ExoS. CHO cells were transiently transfected with pEGFP, wild-type (WT) (A), or dominant active (B) pGFP-Rho GTPases. Transfected cells were then treated with 4% paraformaldehyde and 0.1% Triton X-100 followed by staining of cellular actin with Alexa Fluor 546 phalloidin. Cells were examined by fluorescence microscopy (magnification, ×60), and images were obtained of representative cells on a charge-coupled device digital camera by using SPOT2. (C) CHO cells transfected with 0.2 μg of the indicated pGFP-DARho GTPase or pEGFP (GFP) were incubated alone or infected with P. aeruginosa PA103 ΔexoU exoT::Tc(pUCP-ExoS) (+ExoS). Three hours postinfection, the cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 in PBS, and stained with Alexa Fluor 546 phalloidin. Cells were visualized by fluorescence microscopy for GFP-Rho GTPases (green) or Alexa Fluor 546 phalloidin (red).

Direct measurement of GTP-bound form of Rho GTPases.

The initial characterization of the interaction between ExoS and the Rho GTPases measured the modulation of the nucleotide state of the Rho GTPase. These experiments used the GTPase effectors Rhotekin (Rho binding domain) and P21-activated kinase 1 (Pak binding domain), which specifically bind GTP-bound RhoA and GTP-bound Rac1 and Cdc42, respectively (6, 27). However, feasibility experiments showed that, in control cell lysates, the GST-Rho binding domain bound only 2.6% of total Rho and the GST-Pac binding domain bound only 6% of total Rac and 2.0% of total Cdc42. This low recovery of Rho-GTP limited the analysis of the nucleotide state of Rho GTPases, since the amount of Rho GTPase binding in control lysates would represent an upper limit of detection. In addition, it is not clear whether Rho GTPase association with these GTP-specific probes represents association with subsets of Rho GTPases or a representative intracellular population of Rho GTPase. Thus, while these systems appear useful to measure the stimulation of Rho GTPases to their activated states (3), these assays have limited value in measuring the reductions in GTP-bound Rho GTPases (32). Two assays were developed to measure in vivo interactions between the Rho GAP domain of ExoS and Rho GTPases, the influence of the Rho GAP domain on the intracellular distribution of Rho GTPases, and the ability of dominant active Rho GTPases to reverse actin reorganization elicited by the Rho GAP domain of ExoS.

Expression of the Rho GAP domain of ExoS redistributes Rac1 and Cdc42 but not RhoA.

The activation states of Rho GTPases are related to their subcellular localization with GTP-bound Rho GTPases localized to membranes and GDP-bound Rho GTPase present in the cytosol (7). Experiments determined if expression of the Rho GAP domain of ExoS, ExoS(1-234), modulated the distribution of Rho GTPases between the membrane and cytosol (Fig. 3). In control CHO cells, the majority of GFP-RhoA localized to the cytosol, and coexpression of ExoS(1-234) did not change this localization. In contrast, the majority of GFP-Rac1 and GFP-Cdc42 were localized in membranes, and coexpression of ExoS(1-234) shifted the distribution of GFP-Rac1 and GFP-Cdc42 to the cytosol. The ability of the Rho GAP domain to redistribute Rac and Cdc42 was specific for Rho GTPases, since coexpression of ExoS(1-234) did not stimulate redistribution of GFP-Ras (Fig. 3). In addition, redistribution of GFP-Rac1 and GFP-Cdc42 to the cytosol required the expression of Rho GAP activity, since coexpression of ExoS(1-234)R146K, a GAP-deficient protein, did not redistribute either GFP-Rac1 or GFP-Cdc42. Western blot analysis of cell lysates showed that both ExoS(1-234) and ExoS(1-234)R146K were expressed at similar levels. Other experiments showed that ExoS(1-234), but not ExoS(1-234)R146K, changes the steady-state localization of endogenous Cdc42 from the membranes to the cytosol (data not shown). These data indicate that the Rho GAP domain of ExoS modulated Rac and Cdc42 in vivo, but the modulation of Rho could not be resolved. This may be due to the observation that Rho is primarily a cytosolic protein in these epithelial cells.

FIG. 3.

Coexpression of the Rho GAP domain of ExoS redistributes Rac1 and Cdc42 but not RhoA. CHO cells were transfected with the indicated pGFP-Rho GTPase or pEGFPC1 (1 μg) alone or with pExoS(1-234) or pExoS(1-234)R146K (0.5 μg). After 24 h, cells were harvested and equivalent membrane and cytosol fractions were subjected to SDS-PAGE followed by Western blotting with anti-GFP antibody. The quantitation of data from two or more experiments is shown. An asterisk indicates a statistically significant difference (P < 0.05) relative to cells transfected with Rho GTPase in the absence of the Rho GAP domain of ExoS. Mem, membrane; Cyto, cytosol.

Expression of dominant active Rho GTPases in CHO cells.

The expression of dominant active GFP-Rho GTPases in CHO cells was similar to the expression of the wild-type forms of the GTPases (Table 1). Each dominant active Rho GTPase had a unique intracellular localization (Fig. 2B). GFP-DARhoA was predominantly in the cytosol with some perinuclear accumulation, GFP-DARac1 was located at the plasma membrane with some nuclear localization, and GFP-DACdc42 was located predominantly near the Golgi with some plasma membrane and cytosolic localization. F-actin staining showed that expression of each dominant active Rho GTPase stimulated the formation of actin stress fibers (Fig. 4). This suggests that DACdc42 and DARac stimulated Rho in cultured epithelial cells.

FIG. 4.

Expression of DARac and DACdc42 inhibit actin reorganization by the Rho GAP domain of ExoS. CHO cells transfected with 100 ng (gray bars) or 200 ng (black and striped bars) of the dominant active form of Rho GTPase(s) were infected (MOI, 8:1) with P. aeruginosa PA103 ΔexoU exoT::Tc(pUCP-ExoS) or cotransfected with 0.1 μg of pExoS(1-234). Three hours postinfection or 24 h posttransfection, the percentage of round cells in two fields (approximately 75 cells/field) was determined. The percentage of round cells observed with the expression of ExoS in the presence of dominant active Rho GTPase was subtracted from the percentage of round cells observed with the expression of ExoS alone to yield the percent inhibition of rounding. Error bars represent standard deviations from triplicate experiments. An asterisk indicates a statistically significant difference (P < 0.05) relative to control cells treated with ExoS in the presence of EGFP.

Coexpression of DARac1 and DACdc42 inhibits reorganization of the actin cytoskeleton by ExoS Rho GAP.

Coexpression was used to associate Rho GTPases in ExoS Rho GAP-mediated reorganization of the actin cytoskeleton. DNA encoding GFP fusions of the DARho GTPases (pRhoAQ63E, pRac1Q61E, or pCdc42Q61E) were cotransfected with pExoS(1-234) into CHO cells and scored for the organization of the actin cytoskeleton. Cotransfection with individual dominant active forms of Rho GTPases had limited ability to inhibit actin reorganization elicited by the ExoS Rho GAP domain (Fig. 4). This suggested that none of the Rho GTPases alone were sufficient to inhibit the reorganization of the actin cytoskeleton and presented the possibility that this reorganization involved multiple Rho signaling pathways. To test this possibility, CHO cells were cotransfected with combinations of dominant active Rho GTPases and the Rho GAP domain of ExoS. Coexpression of DARac1 and DACdc42 inhibited Rho GAP-mediated reorganization of the actin cytoskeleton to a greater extent than did coexpression of DARhoA and DARac1 or DARhoA and DACdc42. Coexpression of the three dominant active Rho GTPases did not enhance the inhibition elicited by coexpression of DARac1 and DACdc42. The ability of the Rho GTPases to inhibit actin reorganization by the Rho GAP domain of ExoS was dose dependent (Fig. 4). These data support a role for Rac and Cdc42 in the disruption of the actin cytoskeleton by the Rho GAP domain of ExoS. Transfected cells were also intoxicated with P. aeruginosa expressing full-length ExoS in order to analyze the effect of the physiological form of the toxin during an infection. While transfection of the individual dominant active Rho GTPases or cotransfection with DARhoA and DACdc42 or DARhoA and DARac did not interfere with the ability of bacterially delivered ExoS to reorganize the actin cytoskeleton (data not shown), coexpression of DARac and DACdc42 inhibited actin reorganization elicited by ExoS delivered by P. aeruginosa (Fig. 4). This degree of inhibition correlated well with that seen in transfection studies, supporting the identification of Rac1 and Cdc42 as intracellular targets.

ExoS inhibits the formation of actin stress fibers in CHO cells.

While coexpression of DARac and DACdc42 inhibited the reorganization of the actin cytoskeleton by ExoS, phalloidin staining showed that these cells lacked actin stress fibers (Fig. 2C). Expression of DARhoA stimulated the expression of stress fibers, which indicated that ExoS had inactivated stress fiber formation via a Rho-dependent pathway (Fig. 2C). The phenotype was observed when ExoS was expressed by transfection or when delivered by P. aeruginosa (Fig. 5). This indicated that ExoS inhibited Rho GTPases in vivo and also showed the presence of two independent pathways for Rho GTPase signaling in CHO cells, which involved actin reorganization (a Rac and Cdc42 function) and actin stress fiber formation (a Rho function).

FIG. 5.

Expression of ExoS(1-234) inhibits the stimulation of stress fibers by DARac and DACdc42. CHO cells transfected with 0.2 μg of the plasmid carrying the DARac and DACdc42 were infected (MOI, 8:1) with P. aeruginosa PA103 ΔexoU exoT::Tc(pUCP-ExoS) (striped bars) or cotransfected with 0.1 μg of pExoS(1-234) (black bars). Three hours postinfection or 24 h posttransfection, cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 in PBS with 1% bovine serum albumin, stained with Alexa Fluor 546 phalloidin, and examined by fluorescence microscopy. The percentage of nonround transfected cells expressing stress fibers in two fields (approximately 25 cells/field) was determined. Error bars represent standard deviations from duplicate experiments. Plus signs indicate statistically significant differences (P < 0.05) in the percentage of cells with stress fibers when comparing cells expressing DARac/DACdc42 relative to cells expressing DARac/DACdc42 that expressed the Rho GAP domain of ExoS (+) or were infected with P. aeruginosa expressing ExoS (+′). Asterisks indicate statistically significant differences (P < 0.05) in the percentage of cells with stress fibers when comparing cells expressing DARac/DACdc42 that express the GAP domain of ExoS to cells expressing DARho/DARac/DACdc42 that express the GAP domain of ExoS (*) or cells expressing DARac/DACdc42 that were infected with P. aeruginosa expressing ExoS to cells expressing DARho/DARac/DACdc42 that were infected with P. aeruginosa expressing ExoS (*′).

DISCUSSION

Type III-delivered Rho GAPs target specific Rho GTPases in vitro. SptP is a Rho GAP for Rac and Cdc42, and ExoS, ExoT, and YopE are Rho GAPs for Rho, Rac, and Cdc42. Since the in vivo and in vitro specificity of eukaryotic Rho GAPs differ (29), determination of the in vivo specificity of the type III Rho GAPs is of interest. Previous in vivo measurements of type III cytotoxins have utilized both bacterial and transfection-mediated delivery systems. Each delivery system possesses advantages and limitations. While bacterial delivery is the natural route of translocation of type III cytotoxins into eukaryotic cells, interpreting phenotypes is complicated due to the interaction of the bacteria with the host cells and the possibility that other type III products are delivered into the eukaryotic cell (8). While transfection-mediated delivery of type III effectors allows direct measurement of the consequence of individual type III cytotoxins on cell physiology, the physiological significance of these studies relies upon the similarity of the phenotype elicited by the type III cytotoxin upon delivery by transfection and the bacterium. Expression of the Rho GAP domain of ExoS in eukaryotic cells stimulates actin reorganization when delivered by transfection or by P. aeruginosa (25, 26), which suggests that transfection is a useful delivery mechanism to analyze the Rho GAP domain. In addition, as observed for full-length ExoS, the Rho GAP domain of ExoS is localized to the perinuclear region and processed in eukaryotic cells (25, 26), which suggested that intracellular expression of the Rho GAP domain followed a similar intracellular pathway and supported the characterization of this domain. Finally, utilization of ExoS(1-234)R146K, which lacks Rho GAP activity and actin reorganization, allowed the resolution of effects due to protein-protein interaction versus Rho GAP activity (14). The in vivo modulation of Rac, Cdc42, and Rho by the Rho GAP domain of ExoS, when delivered by either transfection or P. aeruginosa, implicates the inactivation of several physiological processes, including cellular motility, phagocytosis, and oxidative potential (15). Interference of these processes gives P. aeruginosa an advantage in direct host-parasite interactions.

Utilization of GFP fusion proteins of the Rho GTPases allowed evaluation by immunological and microscopic techniques. At the level of this investigation, the GFP fusion proteins possessed properties that were similar to the reported properties of endogenous Rho GTPases with respect to subcellular localization and effects on actin organization. A recent study showed that GFP fusion proteins of the Rho GTPases were functional within several eukaryotic cell lines (21). The amounts of GFP-Rho GTPase expressed in the present study were adjusted to be similar to the amounts of endogenous Rho GTPase in eukaryotic cells (21).

Two approaches evaluated the influence of the Rho GAP domain of ExoS with the Rho GTPases, physical distribution of the wild-type Rho GTPases and the ability of dominant active Rho GTPases to modulate Rho GAP activity. Coexpression of the Rho GAP domain stimulated the redistribution of Rac1 and Cdc42 from the membrane to the cytosol. Coexpression of DARac1 and DACdc42 reversed the reorganization of the actin cytoskeleton elicited by the Rho GAP domain of ExoS. These cells lacked actin stress fibers, which could be stimulated by the transfection of DARhoA. Together, these data indicated that Rac1, Cdc42, and Rho were in vivo targets of the Rho GAP domain of ExoS. The inactivation of Rac and Cdc42 was responsible for the actin reorganization elicited by ExoS while inactivation of Rho was responsible for the loss of actin stress fibers. Two other toxins stimulate the intracellular redistribution of Rho GTPases; YopT stimulates the movement of RhoA from the cell membrane to the cytosol by an undefined mechanism (34) and the clostridial glucosyltransferases block the recycling of Rho GTPases, resulting in the accumulation of the Rho GTPases in the cytosol (12). Other studies have shown that upon stimulation, Rho GTPases migrate from the cytosol to membranes within eukaryotic cells (7, 19).

Recent studies by Andor and coworkers (2) indicated that Rac was a preferred in vivo target of YopE, a Rho GAP that has the same Rho GTPase activity in vitro as ExoS. This conclusion was based upon the ability of YopE to inhibit Rho GTPase signaling upon stimulation by ligands that activate the Rho GTPase at the cell membrane. YopE inhibited the activation of Rac signal transduction by bradykinin, a direct stimulator of Cdc42, while YopE expression did not inhibit Rac activation upon stimulation of sphingosine-1 phosphate, which directly activates Rac. While preferred targeting of Rac by YopE is one explanation for this phenotype, another interpretation is that the activation state (GTP/GDP) influences the observed sensitivity of Rho GTPases to YopE, where direct ligand activation shifts a greater amount of the Rho GTPase to its GTP-bound form than does activation by intracellular signal transduction, in this case Rac by Cdc42. Thus, Rho GTPases activated by direct ligand activation would appear more resistant to the action of YopE than would Rho GTPases that are activated by upstream intracellular signaling pathways. Thus, the steady-state level of activated Rho GTPase, in addition to affinity, may contribute to the observed sensitivity of a Rho GTPase to a Rho GAP in vivo.

Unlike ExoS, expression of DARhoA reversed actin reorganization stimulated by YopE (5). There are several explanations for the different responses to DARhoA. First, in vivo actin reorganization elicited by YopE and ExoS may be due to the modulation of different classes of Rho GTPases. Second, the processing and localization of YopE and ExoS may differ in vivo, providing unique classes of Rho GTPases as targets for inactivation. Previous studies showed that ExoS is localized to an intracellular compartment of the cell (26), whereas the intracellular localization of YopE in vivo is not clear. Other studies have reported the inactivation of RhoA by the expression of ExoT (11). Reported inactivation was based upon affinity precipitation of Rho-GTP from lysates of cells infected with P. aeruginosa expressing ExoT, which was not observed in our experimental system because there was not sufficient Rho-GTP precipitated in the control cells to allow detection of an inactivation of the GTPase. Potential differences in the in vivo activities of ExoT and ExoS could be due to differences in their intracellular localization or expression in eukaryotic cells.