Biological Effects of Pseudomonas aeruginosa Type III-Secreted Proteins on CHO Cells

Abstract

A strain of Pseudomonas aeruginosa that fails to express known type III-secreted effector proteins was constructed as an expression host. Individual effectors were expressed in trans, and their biological effects on CHO cells were assessed in an acute cellular infection model. Intoxication with ExoS, ExoT, or ExoY resulted in alterations in cell morphology. As shown in previous genetic studies, ExoU expression was linked to acute cytotoxicity.

The exoenzyme S regulon, which encodes the type III secretion system of Pseudomonas aeruginosa, consists of coordinately regulated secretion, translocation, regulatory, and effector genes (6). To date, four type III-secreted effector proteins have been identified; ExoS, ExoT, ExoU, and ExoY (3, 10, 21–23). ExoS and ExoT are members of the family of bacterial ADP-ribosyltransferases (10, 21). Despite having 75% amino acid identity, ExoT possesses only 0.2% of the enzymatic activity of ExoS (21). ExoU functions as an acute cytotoxin in vitro and is associated with lung injury in vivo (3). The mechanism of ExoU-mediated toxicity remains unknown. ExoY is a recently discovered adenylate cyclase, which is activated by a eukaryotic protein that is distinct from calmodulin (23). Because P. aeruginosa produces multiple effector proteins in a strain-specific manner, it is difficult to determine the role of the individual products in pathogenesis. The goal of this study was to construct a strain of P. aeruginosa which fails to express the known effector proteins for use as an expression host. By constructing a host without effectors but possessing a functional secretion and delivery apparatus, the cellular effects of individual virulence determinants could be assessed.

Construction of a P. aeruginosa type III-secreted effector mutant.

P. aeruginosa PA103 was chosen as the parental strain (Table 1). PA103 produces significant amounts of ExoU and ExoT (3) but fails to express ExoY (23) and does not possess exoS (5). In addition, strain PA103 is easy to genetically manipulate and displays virulence in both tissue culture and acute lung infection models of P. aeruginosa pathogenesis (1, 3, 5, 11). Although strain PA103 produces large amounts of exotoxin A in vitro, this toxin appears to play no role in the tissue culture and acute infection models developed to measure the contribution of the type III-secreted products (1, 11). In previous studies, individual mutations in exoU (PA103ΔexoU) and exoT (PA103exoT::Tc) were constructed (3, 4). In this study, we constructed the double mutant PA103ΔexoUexoT::Tc and compared its properties with those of the parental (PA103) and individual mutant (PA103ΔexoU and PA103exoT::Tc) strains in a Chinese hamster ovary (CHO) cell model of infection.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Known type III effector protein(s) expressed or relevant characteristic(s)	Reference(s) or source
P. aeruginosa strains
PA103	ExoT, ExoU	B. H. Iglewski
PA103ΔexoU	ExoT	4
PA103exoT::Tc	ExoU	3
PA103ΔexoUexoT::Tc	None	This study
PA103toxA::Ω	ExoT, ExoU	B. H. Iglewski
PA103exsA::Ω	None	7
Plasmids
pMOBexoT::Tc	Allelic replacement vector encoding tetracycline-interrupted exoT	3, 18
pUCP18	pUC-derived cloning vector able to replicate in P. aeruginosa	17
pUCPexoS	Encodes the wild-type ADP-ribosyltransferase ExoS (100% activity)	12
pUCPexoSE381A	Encodes the noncatalytic mutant ExoSE381A (0.02% activity)	This study and 13
pUCPexoT	Encodes the ADP-ribosyltransferase ExoT (0.2% activity)	21
pUCPexoY	Encodes the adenylate cyclase ExoY	23
pUCPexoYK81M	Encodes the noncatalytic mutant ExoYK81M	23

To construct the double exoU-exoT mutation, pMOBexoT::Tc (3) was conjugated into strain PA103ΔexoU (4). Tetracycline-resistant merodiploids were selected and passaged on Vogel-Bonner minimal medium (20) with 100 μg of tetracycline per ml. Plasmid sequences and the wild-type exoT allele were resolved from the chromosome by selecting for strains resistant to 5% sucrose and tetracycline (7, 18). Isolates exhibiting the correct phenotype were grown under inducing conditions for the exoenzyme S regulon, and their extracellular protein profiles were determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie blue staining. Chromosomal DNA was isolated (7) from strains defective for the extracellular production of ExoU and ExoT and subjected to Southern blot analysis (14). We selected a single isolate, PA103ΔexoUexoT::Tc, which failed to hybridize to the exoU probe, exhibited tetracycline resistance, and failed to express extracellular ExoU and ExoT.

Single (PA103exoT::Tc and PA103ΔexoU) and double (PA103ΔexoUexoT::Tc) mutant strains were analyzed for their extracellular protein profiles by SDS-PAGE and Western blot analysis. In the parental PA103 strain, all of the tested extracellular proteins of the regulon (ExoU, ExoT, PcrV, and PopD) were induced by the inclusion of the chelator nitrilotriacetic acid (NTA) in the growth medium (Fig. 1A and B). Induction and secretion of type III proteins also occurred in each mutant strain. These results indicated that type III-mediated regulation and secretion were unaffected by the introduction of mutant alleles. Introduction of either single or double mutant alleles resulted in the absence of only the respective protein products (Fig. 1A and B).

FIG. 1.

Extracellular protein profiles, Western blot analysis, and infection of CHO cells with parental (PA103) and mutant (PA103ΔexoU, PA103exoT::Tc, and PA103ΔexoUexoT::Tc) strains of P. aeruginosa. (A) Coomassie blue-stained polyacrylamide gel (10%) of concentrated culture supernatants from strains grown in the absence (lanes 1, 3, 5, and 7) or presence (lanes 2, 4, 6, and 8) of 10 mM NTA, a chelator that induces the expression of the exoenzyme S regulon. Supernatant fractions were collected and concentrated 20-fold by the addition of a saturated ammonium sulfate solution to 55% from strains PA103 (lanes 1 and 2), PA103ΔexoU (lanes 3 and 4), PA103exoT::Tc (lanes 5 and 6), and PA103ΔexoUexoT::Tc (lanes 7 and 8). Molecular mass markers (MWM; in kilodaltons) and the relative mobilities of ExoU (72 kDa), ExoT (53 kDa), PcrV (32.2 kDa), and PopD (31 kDa) are indicated. (B) Western blot of a duplicate gel as shown in panel A. A mixture of specific antisera reactive to ExoU, ExoS-ExoT, PcrV, and PopD was used as the primary antibody. Bound antibodies were visualized with a peroxidase-labeled secondary antibody and 4-chloro-1-naphthol and peroxide as substrate. (C) Phase-contrast microscopy (40× objective) of CHO cell morphology following infection with parental or mutant strains of P. aeruginosa. The results of the trypan blue staining for uninfected cells, PA103, and PA103exoT::Tc (10× objective) are shown in the insets. Only infections with bacterial strains expressing ExoU resulted in trypan blue staining 3 to 4 h after bacterial infection.

Parental and mutant P. aeruginosa strains were transferred from Vogel-Bonner minimal medium to serum-free tissue culture medium and used to infect CHO cells at a multiplicity of infection of ≈5:1. Following either a 3-h (strains PA103 and PA103exoT::Tc) or a 4-h (uninfected; strains PA103ΔexoU and PA103ΔexoUexoT::Tc) infection at 37°C in 5% CO₂, duplicate wells were washed with phosphate-buffered saline and either fixed in 2% paraformaldehyde or stained for 5 min with trypan blue and photographed. An additional quantitative 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assay for cell viability was performed for each strain 4 h postinfection (2). CHO cells infected with parental strain PA103 differed in morphology, were permeable to trypan blue, and showed significant differences in cell viability compared to uninfected cells (Fig. 1C and Table 2). CHO cells infected with PA103ΔexoU (expressing ExoT) appeared rounded, were not permeable to trypan blue, and retained viability (Fig. 1C and Table 2). Cells infected with a strain expressing only ExoU (PA103exoT::Tc) possessed a phenotype similar to that of cells infected with the parental strain. Finally, when a strain of P. aeruginosa which fails to express any of the known type III-secreted effector proteins, PA103ΔexoUexoT::Tc, was used, the infected cells were indistinguishable from the uninfected control cells (Fig. 1C and Table 2). We interpreted these results as suggesting that part of the ExoU-mediated cytotoxic response (3) may involve cellular morphology changes and/or membrane damage. On the other hand, ExoT appears to cause cell rounding in the absence of membrane damage or changes in cell viability at this time point. Elimination of ExoT and ExoU appears to result in an avirulent strain in this tissue culture model.

TABLE 2.

Biological effects of the various bacterial strains on CHO cells

Strain	Effector(s)	% Altered morphologya	% Trypan blue staineda	Decreased viabilityb
None (uninfected)		6 ± 2	0	NAc
PA103	ExoT, ExoU	53 ± 2	53 ± 2	+
PA103exoT::Tc	ExoU	46 ± 8	46 ± 8	+
PA103toxA::Ωd	ExoT, ExoU	NDe	ND	+
PA103ΔexoU	ExoT	98 ± 2	0	−
PA103ΔexoUexoT::Tc	None	6 ± 2	0	−
PA103exsA::Ω	None	NCf	0	−
PA103ΔexoUexoT::Tc pUCP	None	5	0	−
PA103ΔexoUexoT::Tc pUCPexoS	ExoS	90 ± 3	0	−
PA103ΔexoUexoT::Tc pUCPexoSE381A	ExoSE381A	99 ± 1	0	−
PA103ΔexoUexoT::Tc pUCPexoT	ExoT	99	0	−
PA103ΔexoUexoT::Tc pUCPexoY	ExoY	95 ± 6	0	−
PA103ΔexoUexoT::Tc pUCPexoYK81M	ExoYK81M	7 ± 4	0	−

^aPercentage of CHO cells displaying an altered morphology or trypan blue staining via phase-contrast microscopy; results are averaged from photomicrographs taken of three random fields of infected monolayers. 

^bPresence or absence of a significant reduction in cell viability compared to uninfected control cells by an MTT viability assay. 

^cNA, not applicable; uninfected control cells were used as the basis of comparison for the MTT assay. 

^dControl for the MTT assay; a strain used to show that exotoxin A does not mediate the acute cytotoxic response in this tissue culture model. 

^eND, not done. 

^fNC, no change in cellular morphology relative to the uninfected control. 

Expression of individual effector proteins in PA103ΔexoUexoT::Tc.

In the acute in vitro infection model, cocultivation of strain PA103ΔexoUexoT::Tc with CHO cells resulted in no observable effect. This result suggested that this might be an ideal host strain from which to assess the biological effects of the individual type III-secreted effector proteins of P. aeruginosa. Strain PA103ΔexoUexoT::Tc was transformed with a vector control (pUCP18); pUCPexoS, a plasmid encoding a noncatalytic derivative of ExoS (pUCPexoSE381A); pUCPexoT; pUCPexoY; or a plasmid expressing a noncatalytic adenylate cyclase (pUCPexoYK81M). Expression plasmid pUCPexoSE381A was constructed by replacing the NsiI-BamHI fragment of pUCP18exoS (12) with that from pET16bΔRIexoSE381A (13). The strains were induced for expression of the exoenzyme S regulon, and the extracellular protein profile was analyzed by SDS-PAGE and Western blot analysis (Fig. 2A and B). This analysis indicated that expression of each product was variable. ExoS and the noncatalytic mutant ExoSE381A appeared to be equally expressed and secreted, as has been shown in previous studies (Fig. 2A and B, compare lanes 2 and 3) (22). When exoT was provided in trans, the protein was made and secreted in relatively large quantities (Fig. 2A and B, lanes 5). Both forms of ExoY were expressed and secreted in much smaller quantities than either ExoS or ExoT.

FIG. 2.

Extracellular protein profiles and Western blot analysis of PA103ΔexoUexoT::Tc expressing various effector proteins in trans and their effects on CHO cells. (A) Coomassie blue-stained polyacrylamide gel (11%) of concentrated culture supernatants from strains grown under inducing conditions (growth in the presence of 10 mM NTA). Supernatants are from strains PA103ΔexoUexoT::Tc pUCP18 (lane 1), PA103ΔexoUexoT::Tc pUCPexoS (lane 2), PA103ΔexoUexoT::Tc pUCPexoSE381A (lane 3), PA103ΔexoUexoT::Tc pUCPexoT (lane 4), PA103ΔexoUexoT::Tc pUCPexoY (lane 5), and PA103ΔexoUexoT::Tc pUCPexoYK81M (lane 6). Molecular mass markers (MWM; in kilodaltons) and the relative mobilities of ExoT (53 kDa), ExoS (49 kDa), and ExoY (42 kDa) are indicated. (B) Western blot of a duplicate gel as shown in panel A. A mixture of specific antisera to ExoS-ExoT and ExoY was used, and the bound antibodies were visualized with a peroxidase-labeled secondary antibody. (C) Cellular morphology of CHO cells infected with PA103ΔexoUexoT::Tc expressing various effector proteins in trans. The name of the expression plasmid in each strain is given above the appropriate picture.

CHO cells were infected for 4 h and subsequently stained with trypan blue or fixed in paraformaldehyde and observed by phase-contrast microscopy or subjected to the MTT assay. At this point in infection, permeability to trypan blue or changes in cell viability were not observed with strains expressing ExoS, ExoSE381A, ExoT, ExoY, or ExoYK81M (Table 2). Compared to the vector control, however, expression of ExoS, ExoSE381A, ExoT, and ExoY, but not of ExoYK81M, resulted in an altered cellular morphology (Fig. 2C and Table 2). Cells appeared rounded and eventually detached from the surface of the well. Our data confirm that the ADP-ribosyltransferase activity of either ExoS or ExoT is not required to cause a rounding of CHO cells, supporting earlier observations of HeLa cell morphology changes when ExoS and ExoSE381A were delivered by the Yersinia type III apparatus (8). In addition, we confirm that ExoY is capable of causing a similar morphological effect on CHO cells which is dependent on adenylate cyclase activity (23). As positive controls for cell morphological changes, CHO cells were intoxicated with Escherichia coli heat-labile enterotoxin or pertussis toxin. Heat-labile enterotoxin mediated CHO cell elongation while pertussis toxin mediated CHO cell clustering, indicating that the cell line we are using responds to changes in cyclic AMP levels as previously reported (data not shown) (9, 23).

Measurement of effector translocation into CHO cells.

In previous studies, we have used ExoS ADP-ribosyltransferase activity to measure type III-mediated translocation from P. aeruginosa 388 (expresses ExoS and ExoT) into CHO cells (19). To determine if PA103ΔexoUexoT::Tc could be used to study translocation of individual components, we pretreated CHO cells with cytochalasin D to inhibit the uptake of bacteria. Treated cells were infected with an inoculum of PA103ΔexoUexoT::Tc pUCPexoS in serum-free medium for 4 h at a multiplicity of infection of ≈5:1. The supernatant was removed, the number of viable bacteria was measured from a small aliquot, and the remaining sample was subjected to centrifugation at 14,000 × g, 4°C. A portion of the soluble fraction was retained for ADP-ribosyltransferase activity assays (supernatant-associated activity). The CHO cell monolayer was washed and treated for 2 h with 100 μg of ciprofloxacin per ml and 200 μg of gentamicin per ml to kill the extracellular bacteria. Infected CHO cells were lysed with 150 μl of distilled water. The lysate was subjected to centrifugation at 14,000 × g (4°C), and a portion of the soluble fraction was retained to perform ADP-ribosyltransferase activity assays (lysate-associated activity) as described previously (19). A duplicate well, not treated with antibiotics, was used to perform viable counts. Supernatant and lysate fractions from uninfected CHO cells or cells infected with strains containing the vector control plasmid (pUCP18), pUCPexoSE381A, or pUCPexoT were included as negative controls. Under these conditions, ExoS ADP-ribosyltransferase activity was predominantly associated with the CHO cell lysate, rather than the supernatant, indicating that PA103ΔexoUexoT::Tc is able to translocate ExoS (Fig. 3). Similar results were obtained when CHO cells were infected with PA103ΔexoUexoT::Tc pUCPexoY and cyclic AMP accumulation was measured (23). Activities that were slightly above background levels were measured from cells infected with strains expressing either ExoSE381A or ExoT. This amount of activity may represent the residual ADP-ribosyltransferase activity of the mutant proteins.

FIG. 3.

Cell- and supernatant-associated ADP-ribosyltransferase activity of PA103ΔexoUexoT::Tc expressing ExoS, ExoSE381A, or ExoT in trans. ADP-ribosyltransferase activity assays were performed on supernatant- and cell-associated samples collected from uninfected CHO cells or CHO cells infected with PA103ΔexoUexoT::Tc pUCP18, PA103ΔexoUexoT::Tc pUCPexoS, PA103ΔexoUexoT::Tc pUCPexoSE381A, or PA103exoUexoT::Tc pUCPexoT. Activity is normalized to CFU and expressed as 10⁻⁴ femtomoles of ADPRT (femtomoles of ADP-ribose transferred to soybean trypsin inhibitor).

Concluding remarks.

We constructed a strain of P. aeruginosa, PA103ΔexoUexoT::Tc, which fails to express any of the known P. aeruginosa type III-secreted effector proteins. Strain PA103ΔexoUexoT::Tc was used as an expression host, and the effects of individual translocated proteins were assessed in a cellular acute infection model. CHO cell viability was measured by using trypan blue staining and an MTT assay. Our results indicate that PA103ΔexoUexoT::Tc expresses and secretes type III effectors and that translocation into CHO cells is measurable by using either activity assays or changes in cell morphology or viability. Our analysis confirmed that ExoS, ExoSE381A, and ExoY alter CHO cell morphology but do not result in an acute cytotoxic response. ExoS, however, has been shown to mediate cytotoxic responses after longer infection (15) or transfection (16) periods. We demonstrated that delivery of ExoT also results in morphology changes and confirmed that ExoU is responsible for acute cytotoxicity. The biological effects of each of the type III-secreted effectors are summarized in Table 3.

TABLE 3.

Biological effects of P. aeruginosa type III-secreted proteins on CHO cells

Protein	Enzymatic activity	Effect on CHO cells
ExoS	ADP-ribosyltransferase	Morphological alterations
ExoT	ADP-ribosyltransferase	Morphological alterations
ExoU	Unknown	Acute cytotoxicity: membrane permeability, loss of viability
		Morphological alterations
ExoY	Adenylate cyclase	Morphological alterations