|Home | About | Journals | Submit | Contact Us | Français|
Recent work has shown that a domain of YopE of Yersinia pseudotuberculosis ranging from amino acids 54 to 75 (R. Krall, Y. Zhang, and J. T. Barbieri, J. Biol. Chem. 279:2747-2753, 2004) is required for proper localization of YopE after ectopic expression in eukaryotic cells. This domain, called the membrane localization domain (MLD), has not been extensively studied in Yersinia. Therefore, an in cis MLD deletion mutant of YopE was created in Y. pseudotuberculosis. The mutant was found to secrete and translocate YopE at wild-type levels. However, the mutant was defective in the autoregulation of YopE expression after the infection of HeLa cells. Although the mutant translocated YopE at wild-type levels, it showed a delayed HeLa cell cytotoxicity. This delay was not caused by a change in GTPase activating protein (GAP) activity, since the mutant showed wild-type YopE GAP activity toward Rac1 and RhoA. The MLD mutant displayed a changed intracellular localization pattern of YopE in HeLa cells after infection, and the YopEΔMLD protein was found to be dispersed within the whole cell, including the nucleus. In contrast, wild-type YopE was found to localize to the perinuclear region of the cell and was not found in the nucleus. In addition, the yopEΔMLD mutant was avirulent. Our results suggest that YopE must target proteins other than RhoA and Rac1 and that the MLD is required for the proper targeting and hence virulence of YopE during infection. Our results raise the question whether YopE is a regulatory protein or a “true” virulence effector protein.
Yersinia pseudotuberculosis is a gram-negative bacterium that causes gastroenteritis in humans. In addition to Y. pseudotuberculosis, the genus of Yersinia consists of two other human pathogens, Y. enterocolitica and Y. pestis. In common for these pathogenic strains is the type III secretion system (T3SS). The T3SS is essential for the virulence of these pathogens and is used for the direct delivery of virulence-associated effector proteins, called Yops, from the bacterium into the eukaryotic host. Y. pseudotuberculosis expresses at least five different Yop effector proteins. They all target different host signaling pathways, resulting in the blockage of phagocytosis, as well as downregulation of the immune response, allowing for extracellular proliferation of the bacterium in lymphatic tissue (33).
One translocated Yop, YopE, is a GTPase activating protein (GAP) that targets small Rho GTPases within the target cell (34). YopE has been shown to target Rac1 and RhoA in vivo. However, interaction with Rac1 and RhoA is not a prerequisite for virulence in mice, indicating that there must be additional targets for YopE. This is in line with the finding that translocated YopE can inhibit the extent of effector translocation, as well as autoregulate Yop expression during infection of HeLa cells (4). The regulatory role of YopE has also been demonstrated by the group of Bliska, where the GAP activity of YopE has been shown to be required for regulation of pore formation (32). Further, the Rho proteins A, B, and C, together with active YopE, are required for effector translocation control (10, 21). YopE is not the only bacterial GAP with a regulatory role. Recently, this phenomenon was also described for ExoS in Pseudomonas aeruginosa, where ExoS can prevent further T3SS triggering and effector translocation (13).
Barbieri and coworkers have shown that YopE harbors a specific region defined as the membrane localization domain (MLD) that likely is involved in the virulence associated functions of YopE (20). However, this latter point remains elusive. A putative MLD is also present in the bacterial GAP proteins ExoS and ExoT from P. aeruginosa, as well as AexT and AexU of Aeromonas veronii (35). For ExoS it has been shown in transfection studies that the MLD region is both required and sufficient for membrane localization of ExoS within the host cell. For YopE, amino acids 54 to 75 were identified as the membrane localization domain. The MLDs of YopE and ExoS were demonstrated to be functionally interchangeable. The MLD of YopE localized ExoS to intracellular membranes, and the resulting hybrid protein was able to ADP-ribosylate eukaryotic proteins (20). Additional work from the same group showed that the membrane localization of ExoS was mediated by hydrophobic interaction via the multiple leucines and isoleucines present within the MLD (35). These researchers further showed that ExoS may interact with Rab5, -6, and -9 via the MLD (36), suggesting that also YopE may be targeted toward other proteins besides RhoA and Rac1.
We have here constructed an in cis deletion of the MLD region of YopE. The domain was found to be required for localization, as demonstrated both by immunofluorescence stainings and biochemical fractionation experiments. Importantly, the yopEΔMLD mutation was still able to inactivate Rac1 and RhoA to the same level as the corresponding wild-type strain. Thus, the deletion did not affect the GAP activity of YopEΔMLD. Finally, an in vivo imaging system (IVIS) bioluminescence experiment demonstrated that the MLD is biologically important since the yopEΔMLD mutant was attenuated in virulence. Our results suggest that the MLD is required for proper intracellular localization of YopE. Hence, the MLD is critical for positioning YopE in an environment where YopE will inactivate a target essential for the virulence of the pathogen.
Bacterial strains and plasmids used in the present study are listed in Table Table1.1. Bacteria are grown as indicated in Luria-Bertani (LB) broth or brain heart infusion (BHI) medium. The medium BHI+Ca2+ was supplemented with 2.5 mM CaCl2, whereas the medium BHI−Ca2+ was supplemented with 5 mM EGTA and 20 mM MgCl2. When appropriate, the following concentration of antibiotics was used: carbenicillin at 100 μg/ml, kanamycin at 50 μg/ml, and chloramphenicol at 25 μg/ml.
To generate yopEΔMLD (YPIII/pIB529), overlapping PCR was used to amplify a 750-bp DNA fragment lacking codons 50 to 74 of the yopE gene with the primers prΔ50-74.for (5′-CACTGAAAGCCAACGCATGTTCTCGG-3′) and prΔ50-74.rev (5′-ACATGCGTTGGCTTTCAGTGCGCCC-3′).
This fragment was first subcloned into the pCR4-TOPO TA cloning vector (Invitrogen) and sequenced to confirm deletion of amino acids 50 to 74. The fragment was cloned into the XbaI/SphI-digested suicide vector pDM4 (22), resulting in pEI1. pEI1 was introduced in the Y. pseudotuberculosis wild-type YPIII/pIB102 by conjugational mating using Escherichia coli S17-1λpir as the donor strain. For selection of appropriate homologues recombination events, established methods were used (22). To create the double mutant yopEΔMLD-(R144A) (YPIII/pIB529-562), the pDM4 vector pMA18 (1) carrying the R144A mutation was introduced into the YPIII/pIB529 strain.
For the IVIS experiments, YPIII/pIB1-XEN4 was purchased from Caliper Life Sciences (formerly Xenogen). YPIII/pIB1-XEN4 is derived from wild-type Y. pseudotuberculosis (11) with the luxCDABE operon integrated on the virulence plasmid. To generate the desired yopE mutants in cis, pDM4 plasmids pMA95, pMA18, and pEI1 was introduced by conjugational mating into YPIII/pIB102-XEN4 resulting in YPIII/pIB526-XEN4 (ΔyopE), YPIII/pIB529-XEN4 (yopEΔMLD), and YPIII/pIB562-XEN4 [yopE(R144A)].
All mutants were confirmed by sequencing.
Bacterial strains were grown in BHI medium with or without Ca2+ at 26°C. Overnight cultures were diluted 20-fold in fresh BHI medium with or without Ca2+, grown at 26°C for 30 min, and then shifted to 37°C for 1 h. Samples were centrifuged for 1 min at 13,000 rpm, and the supernatant was analyzed by Western blotting with polyclonal antibodies directed to total Yops or YopE (AgriSera Sweden).
Membranes were probed with horseradish peroxidase-labeled anti-rabbit antibody (Amersham Pharmacia Biotech), and proteins were detected with a chemiluminescence detection kit (Millipore).
Bacterial strains were grown in BHI−Ca2+ at 26°C. Overnight cultures were diluted 20-fold in fresh medium, grown at 26°C for 30 min, and then shifted to 37°C for 1 h. Whole cells, supernatant, and pellet were analyzed by Western blotting with total Yop or YopE polyclonal antibodies.
Overnight cultures were prepared in BHI+Ca2+, diluted 20-fold in fresh medium, followed by 30 min of incubation at 26°C and 1 h at 37°C. At this time point, chloramphenicol (50 μg/ml) was added. Samples were collected at 0, 5, 10, 15, 20, 30, 45, and 60 min after chloramphenicol addition. All samples were analyzed by Western blotting with a total Yop polyclonal antibody.
HeLa cells were routinely maintained as described earlier (29).
YopE expression in the presence or absence of HeLa cells was examined as described earlier (4). In short, 5 × 105 HeLa cells were seeded in four wells in a six-well plate. Next day, the monolayer was washed twice with phosphate-buffered saline (PBS) and antibiotic-free minimal essential medium (MEM) was added to all six wells. Bacterial strains were grown overnight in LB medium, diluted 150-fold in antibiotic-free MEM, and then incubated at 26°C for 30 min, followed by 1 h at 37°C. For infection, 500 μl of bacterial culture was added in triplicate wells per tested strain. To one set of HeLa cells, chloramphenicol was added at a concentration of 50 μg/ml to stop de novo protein synthesis. After 90 min of infection, total well content was recovered and analyzed by Western blotting with a polyclonal YopE antibody.
In a separate duplicate experiment, after 90 min of infection, the cell medium was removed and centrifuged. The pellet and supernatant was examined for YopE presence using a YopE antibody. HeLa cells were lysed using 1% digitonin, scraped off, and centrifuged for 10 min. The pellet and supernatant were analyzed by using a YopE antibody. Quantification of signals from four independent experiments was performed by using the Bio-Rad Imager ChemiDoc XRS system.
Translocation was analyzed by using the established proteinase protection-digitonin extraction assay (23). Briefly, 2 × 106 HeLa cells were seeded in 10-cm petri dishes the day before infection. Overnight bacterial cultures in LB were diluted in MEM without antibiotics and grown at 26°C for 30 min, followed by 1 h at 37°C. HeLa cell monolayers were infected in duplicates for 3 h and, after incubation, the dishes were washed twice with PBS. Proteinase K (500 μg/ml; Roche) was added to all plates for 1 min and then removed. The dishes were incubated at room temperature for 20 min. To stop the protease activity, the inhibitor phenylmethylsulfonyl fluoride (4 mM; Sigma) was added. One set of HeLa cells was lysed by the addition of 1% digitonin (Sigma) in PBS; to the duplicate plate, PBS alone was added. Cells were scraped off and centrifuged to remove the intact cells and cell debris. Supernatants were analyzed by Western blotting with a polyclonal YopE antibody and a monoclonal pan-Erk antibody (BD Transduction Laboratories).
The translocation assay was performed as described above. However, the resulting supernatant, the postnuclear supernatant (PNS), was ultracentrifuged at 112,000 × g for 30 min at 4°C (20). Pellet (membranes) and supernatant (cytosol) were analyzed by Western blotting with YopE and pan-Erk antibodies.
A total of 105 HeLa cells were seeded per coverslip the day before infection. Bacterial strains were grown in LB medium at 26°C overnight. The next day, bacterial cultures were diluted to an optical density at 600 nm of 0.1 in fresh medium, followed by incubation at 26°C for 30 min and at 37°C for 1 h. HeLa cells were washed twice in PBS, overlaid with antibiotic-free MEM, and then infected with bacteria at a multiplicity of infection (MOI) of 10. At 3 h postinfection, the HeLa cells were washed twice in PBS, fixed in 2% paraformaldehyde, and permeabilized in 0.5% Triton X-100 (Sigma). The HeLa cell membrane was labeled with wheat germ agglutinin (Molecular Probes/Invitrogen). Affinity-purified rabbit polyclonal YopE antibody was added, followed by a secondary donkey anti-rabbit Alexa 488 (Molecular Probes/Invitrogen). Coverslips were mounted and analyzed by confocal scanning microscopy (Leica).
A total of 105 HeLa cells were seeded per coverslip the day before infection. Bacterial strains were grown in LB overnight at 26°C. The next day, overnight cultures were diluted in fresh cell culture medium and grown for 30 min at 26°C, followed by 1 h at 37°C. HeLa cells were infected at an MOI of 10. After 45 min and 2 h, cell coverslips were fixed in 2% paraformaldehyde, and the cytotoxicity (altered morphology of cultured cells) was assessed as previously described (27, 29) by using a microscope (Zeiss Axioskop) at a magnification of ×20. Images were captured by a charge-coupled device camera (Hamamatsu).
For the pulldown of Rac1, GST-PAK-Rac-BD was used (where GST is glutathione S-transferase, PAK is p21-activated kinase, and BD is the binding domain). For RhoA pulldowns, GST-Rhotekin-Rho-BD was utilized. GST fusion proteins were purified and used as described previously (26, 30). HeLa cells were seeded in 15-cm petri dishes and grown to 70% confluence. Bacterial strains were pregrown in BHI medium, and HeLa cells were infected at an MOI of 10. After infection, as indicated in the figure legends, cells were washed twice in ice-cold PBS and lysed on ice with lysis buffer (for Rac1: 1% Triton X-100, 100 mM NaCl, 50 mM Tris-HCl [pH 7.5], 15 mM MgCl2, and 1 mM EDTA; and for RhoA: 1% Triton X-100, 500 mM NaCl, 50 mM Tris-HCl [pH 7.2], 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], and 10 mM MgCl2). Cell lysates were cleared by 10 min of centrifugation at 15,000 × g for 10 min at 4°C. Total protein concentration was determined and equal amount of cell lysates was incubated with GST fusion protein beads for 30 min (Rac1) or 45 min (RhoA) at 4°C. The beads were washed (for Rac1: 1% NP-40, 10 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, and 12 mM MgCl2; for RhoA: 1% Triton X-100, 50 mM Tris-HCl [pH 7.2], 150 mM NaCl, and 10 mM MgCl2), and bound GTPase was separated by SDS-polyacrylamide gel electrophoresis (PAGE), followed by immunoblotting with monoclonal antibodies to Rac1 and RhoA, respectively (Santa Cruz Biotechnology). Then, 5 μg of whole-cell lysate was analyzed on the same immunoblots for the presence of the GTPase. The GTPase-bound protein accounts for 1 to 5% of the total GTPase, depending on the cell conditions and different GST fusion protein batches. Pulldown experiments were performed at least three times for each GTPase.
Eight-week-old BALB/c female mice (Scanbur DK) were divided into four groups: seven animals for the wild-type (YPIII/pIB1-XEN4) and ΔyopE (YPIII/pIB526-XEN4) strains, and eight animals for the yopEΔMLD (YPIII/pIB529-XEN4) and yopE(R144A) (YPIII/pIB526-XEN4) strains. The mice were deprived of food and water 18 h prior to oral infection. Bacteria were grown overnight in LB medium at 26°C and resuspended in sterile tap water supplemented with 150 mM NaCl. The bacterial solutions were serial diluted, and the infection CFU counts were determined to 2.2 × 108 bacteria/ml for the wild-type and yopEΔMLD strains, 2.3 × 108 for the ΔyopE strain, and 1.9 × 108 for the yopE(R144A) strain.
Prior to imaging, the mice were anesthetized with 2.5% isoflurane gas and then placed in an IVIS imaging chamber (Caliper Life Sciences). During the imaging process, the animals were kept anesthetized. Total photon emissions of each animal were acquired for 10 or 60 s. Acquired images were analyzed by the software Living Image (Caliper Life Sciences), where a series of all strains was created for each imaging time point. Bioluminescence signals are shown using a pseudo-color scale, where red represents the most intense signals and blue represents the least intense. The images were compiled into a single figure using Adobe Photoshop and Adobe Illustrator without any manipulation of the original images.
C57BL/6 female mice (Scanbur BK), with three animals in each group, were injected intraperitoneally with increasing doses (104 to 107 CFU/ml) of the different strains. Bacteria were grown overnight in LB medium at 26°C, pelleted, and serially diluted in PBS prior to injection. The 50% infective dose (ID50) was calculated (25), rather than the traditional 50% lethal dose, since the animals were sacrificed immediately when they showed symptoms of severe disease. The animal experiments were approved by the Swedish National Board for Laboratory Animals.
Zhang and Barbieri have recently identified a putative MLD present in most bacterial GAP proteins (Fig. (Fig.1)1) (35). To explore a possible role for this domain in the YopE-dependent virulence of Y. pseudotuberculosis, an in cis deletion mutant of amino acids 50 to 75 of yopE was created in wild-type Y. pseudotuberculosis YPIII/pIB102, generating the strain YPIII/pIB529. The yopEΔMLD deletion was also introduced into the GAP-deficient yopE(R144A) mutant, where the arginine at position 144 is substituted with an alanine. This mutant was called YPIII/pIB529-562.
The yopEΔMLD deletion overlapped with the second half of the interaction domain between YopE and the cognate chaperone YerA (SycE), which binds to the YopE region amino acids 15 to 75 (9, 31). Therefore, it was important to first analyze the effect of the mutation on the expression and secretion of Yop proteins since a YerA-null mutant is affected in secretion (17). In line with earlier results (12), no effect of the mutation on either in vitro Yop expression or Yop secretion was found when the two MLD mutants were compared to the wild-type strain (data not shown).
We have earlier shown that in the presence of eukaryotic target cells the total levels of YopE protein is auto-regulated. This autoregulation is dependent on a functional GAP domain. The yopE(R144A) mutant displays increased levels of YopE in the presence of HeLa cells, whereas YopE from the wild-type strain remains at a lower level, regardless of cell presence (3, 4). We examined the levels of YopE protein in the presence or absence of eukaryotic cells to evaluate whether the MLD of YopE was important for this autoregulatory control or not. Bacteria were added to wells either containing semiconfluent HeLa cells or cell-free culture medium. At 90 min after the onset of infection at 37°C, SDS sample buffer was added to all wells to recover total protein content. Equal amounts were loaded onto a SDS-PAGE gel and immunoblotted using YopE antiserum. Wild-type Y. pseudotuberculosis displayed the same level of YopE expression regardless of HeLa cell presence (Fig. (Fig.2).2). As reported previously (3, 4), the GAP-deficient mutant yopE(R144A) showed elevated levels of YopE protein when the bacteria were incubated in the presence of HeLa cells, indicating that this mutant is defective in autoregulation of YopE. Interestingly, the yopEΔMLD mutant showed the same response as the yopE(R144A) mutant (Fig. (Fig.2),2), suggesting that the MLD was important for the autoregulatory control loop of YopE. Further, the phenotype of the double mutant yopEΔMLD(R144A) corresponded to that of the single mutants yopEΔMLD or yopE(R144A), respectively. Hence, combination of mutations did not have an additive effect on the autoregulatory control.
We proceeded to investigate the level of YopE translocation and intracellular localization within HeLa cells during infection, using a modified proteinase K treatment-digitonin extraction assay (2, 23). Semiconfluent HeLa cell monolayers were infected in duplicates with Yersinia strains for 3 h prior to treatment with proteinase K to remove all extracellular protein. A ΔyopD mutant (15), which is unable to translocate effectors, was included to ensure the effectiveness of the proteinase K treatment. To stop the proteinase activity, the inhibitor phenylmethylsulfonyl fluoride was added, followed by lysis of one set of cells using the detergent digitonin. The duplicate was treated with PBS alone. All samples were centrifuged to remove intact cells and nuclei. The supernatant from digitonin-treated and untreated cells were analyzed by Western blotting with antiserum to YopE and the eukaryotic protein Erk. Erk is a marker of cell lysis and also a control of equal sample loading (Fig. (Fig.3A).3A). The difference between digitonin treated (+) versus non-digitonin-treated (−) cells was considered to be translocated protein. The yopEΔMLD mutant was found to translocate YopEΔMLD protein at the same levels as the wild type, whereas the yopEΔMLD(R144A) double mutant showed an overtranslocation phenotype similar to that of the yopE(R144A) GAP mutation alone. Thus, in accordance with previous findings, it was found that translocation control requires an active YopE protein (3, 4, 10, 21, 32). It is also evident that the yopEΔMLD mutant translocates YopE at the same level as the wild-type strain, bearing in mind that the deleted region constitutes half the YerA chaperone binding domain.
However, intriguingly, in the presence of target cells, the yopEΔMLD mutant had lost control of YopE expression, yet it translocated YopE at wild-type levels. In order to determine the location of the extra YopE protein, a modified version of the auto-regulation assay was performed in parallel. It was clear that the overproduced YopEΔMLD protein was not leaking out into the cell medium supernatant, since no YopE or other Yops could be detected in the supernatant. As a positive control, a ΔyopN mutant was used, since it lacks polarized translocation and secretes Yops into the cell medium (28). The HeLa cells were lysed by using 1% digitonin and scraped off and centrifuged. The resulting pellet consisted of extracellular located YopE, whereas the supernatant consisted of translocated intracellular YopE. By quantification, we found that ca. 60% more YopEΔMLD than wild-type YopE was detected in the cell-associated pellet. The levels of translocation were also quantified and intracellular YopEΔMLD was found in the same amount as the wild-type YopE protein (data not shown). We conclude that the YopMLD protein is not translocated as efficiently as wild-type YopE and that the “extra” YopEΔMLD made is trapped between the bacterium and the target cell.
To examine the intracellular localization of translocated YopE within the HeLa cell, a modified version of the above-described translocation assay was used. After digitonin treatment, the lysed cells were centrifuged to remove intact cells and nuclei. The resulting supernatant, called the PNS, was subjected to ultracentrifugation that separated the cytosol from membranes (20). All fractions, including the PNS (denoted lysate in Fig. Fig.3B),3B), membrane pellet, and cytosolic supernatant, were analyzed by SDS-PAGE and immunoblotted with antisera to YopE and the eukaryotic protein Erk (a marker for the eukaryotic cell cytosol and a loading control) (Fig. (Fig.3B).3B). Nonlysed infected HeLa cells were analyzed on the same Western blots as a control of translocation. The nonlysed cells were negative for YopE/Erk protein, verifying the efficiency of proteinase K treatment and that protein detected within the lysed cells was translocated protein (data not shown).
Wild-type YopE localized only within the pellet (membrane) fraction (Fig. (Fig.3B)3B) in line with previous findings (20). The yopEΔMLD mutant, however, showed a different localization pattern. Compared to the wild type, only a small portion of translocated YopE was detected in the pellet fraction, whereas the major part was found in the supernatant fraction (cytosol) (Fig. (Fig.3B).3B). This difference in YopE localization indicates that the MLD is required for proper intracellular localization, corroborating previous data from the group of Barbieri and coworkers (20). When the localization of YopE(R144A) was investigated using the same assay (Fig. (Fig.3B),3B), the levels of translocated protein detected, was as expected, increased. Interestingly, roughly 50% of total translocated YopE(R144A) was detected in the membrane fraction and 50% in the cytosolic fraction. The appearance of YopE(R144A) in the cytosol is likely due to the increased translocation levels, resulting in saturation effect. For the double yopEΔMLD(R144A) mutant, all translocated YopE was found within the cytosolic fraction, further demonstrating the importance of the MLD for the proper localization of YopE.
Previous immunofluorescence experiments have shown that translocated YopE protein localizes to the intracellular membrane-rich perinuclear region of infected HeLa cells (4, 28). In addition, Barbieri and coworkers (20) demonstrated that a transfected green fluorescent protein-tagged YopE amino acids 54 to 75 construct localized to the perinuclear region of the HeLa cell, whereas a construct expressing only YopE amino acids 64 to 75 did not. Therefore, we wanted to determine the localization of YopEΔMLD using immunofluorescence stainings. However, HeLa cells infected with wild-type Y. pseudotuberculosis display severe cytotoxicity. This renders determination of localization of translocated YopE difficult, since the HeLa cell becomes severely affected in morphology and decreased in size. Therefore, we chose to use GAP-deficient mutants in this assay. The GAP-deficient mutants do not affect the HeLa cell morphology and also overtranslocate YopE. HeLa cells were infected for 3 h and thereafter fixed and permeabilized. The HeLa cell membrane was visualized with wheat germ agglutinin conjugated to rhodamine. YopE was visualized using an affinity-purified YopE rabbit antibody, followed by a secondary donkey anti-rabbit Alexa 488-conjugated antibody. Samples were analyzed by confocal scanning microscopy (Fig. (Fig.4).4). The yopE(R144A) strain displayed the expected perinuclear localization of translocated YopE protein. However, the yopEΔMLD(R144A) mutant protein did not localize to this region, instead the YopE staining was diffusely scattered within the whole HeLa cell, clearly showing the importance of the MLD for intracellular localization. An interesting observation was that cell-associated YopEΔMLD bacteria were stained green in contrast to wild-type bacteria that did not stain (28), corroborating the finding that YopEΔMLD is associated with the bacteria and trapped between the bacterium and the target cell.
YopE possesses a cytotoxic activity that disrupts the actin microfilament structure of infected HeLa cells (27). This activity is dependent on a functional GAP domain (1, 10, 34). HeLa cells were infected for 45 min and 2 h, respectively, and cytotoxicity was observed by microscopy (Fig. (Fig.55).
The Y. pseudotuberculosis wild-type strain induced a morphological change of the HeLa cells with cell rounding and visible actin retraction tails (Fig. (Fig.5).5). As expected, the two GAP-deficient mutants yopE(R144A) and yopEΔMLD(R144A) were unable to induce a cytotoxic effect (10, 34). The yopEΔMLD mutant induced a cytotoxic effect, albeit not to the same extent, and the effect seen was much delayed compared to the wild type. From the translocation experiments, we know that YopEΔMLD is translocated at the same level as wild-type YopE, indicating that the delay seen in cytotoxicity is not due to less translocated protein. This result could argue for the fact that the yopEΔMLD mutation affected the GAP activity of YopE per se.
To address this question, GAP activity assays were performed in the HeLa cell model system of infection (4). The results from previous work by Aili et al. have shown that YopE targets Rac1 and RhoA during infection of HeLa cells (4). We used the same pull-down technique as earlier (4), utilizing GST fusions of the binding domain of downstream targets of the active form of each GTPase to measure the activation state of endogenous GTPases in HeLa cells (26, 30). Cells were infected with wild-type, yopE(R144A), yopEΔMLD, and yopEΔMLD(R144A) strains and lysed after 30 min for Rac1 and after 1 h for RhoA. Cleared supernatants were incubated with GST-PAK beads or GST-Rhotekin, and bound GTPase was separated on SDS-PAGE and analyzed by immunoblotting. The results showed that the YopEΔMLD protein was as effective as wild-type YopE to inactivate Rac1, whereas the double yopEΔMLD(R144A) mutant had no effect on the activity of Rac1 (Fig. (Fig.6A).6A). YopE has been shown to induce downregulation of Rac1 as early as 5 min postinfection (4). Thus, the kinetics of Rac1 inactivation by the mutant was also studied. We observed that as with wild-type YopE (4), YopEΔMLD was able to downregulate Rac1 already 5 min postinfection (Fig. (Fig.6B).6B). These results demonstrated that the yopEΔMLD mutant was as efficient as the wild type in inactivating Rac1. When the effect on RhoA was investigated, it was clear that the yopEΔMLD mutant was as efficient as the wild type in inactivating RhoA (Fig. (Fig.6C).6C). This result is in perfect agreement with our results showing that YopEΔMLD is not defective in the translocation control, since recent work from Bliska and coworkers demonstrated that RhoA, -B, and -C were involved in the translocation control (21). This finding also indicated that YopE interacts with GTPases other than RhoA and Rac1 to induce cytotoxicity. Also, the delayed cytotoxicity seen with yopEΔMLD is not due to a defect in GAP activity per se but rather to an inability of this protein to find and interact with a eukaryotic target GTPase.
In order to verify the biological importance of the MLD, we took advantage of the new IVIS technology. The IVIS in vivo imaging system allows for real-time monitoring of bacterial infections. The different yopE mutations (Fig. (Fig.7)7) were introduced into the bioluminescent Y. pseudotuberculosis wild-type strain pIB1-XEN4. For infection, 30 female BALB/c mice were divided into four groups, one group per tested strain. Mice were infected orally with 1.9 × 108 to 2.3 × 108 bacteria/ml. Prior to bioluminescence imaging, all mice were anesthetized and placed in the IVIS chamber (Fig. (Fig.77).
On day 1 postinfection, all mice were infected at similar levels when total photons emitted per animal were compared. On day 3, the wild-type-infected mice started to show clear signs of disease, the most prominent being ruffled fur and diarrhea. This visual observation correlated to the bioluminescent signals where the wild type emitted 10 times more total photons than mice infected with the mutant strains. None of the three yopE mutant-infected animals showed any visual sign of disease, and there was no difference in total light intensity in comparison with each other. The condition of the wild-type-infected mice deteriorated on day 6, with diarrhea, hunched posture, and ruffled fur, and the mice were immobile. Two of them were considered moribund and were euthanized. Surprisingly, imaging of the whole animal resulted in a very low light signal, most likely due to the condition of the fur that caused the emitted light to scatter and that disease had spread into organs in deeper tissue. The dissected organs showed, however, clear symptoms of disease, indicating that the infection had become systemic. At 9 days postinfection, the remaining wild-type-infected mice still elicited high photon emission. The ΔyopE infection was cleared, and only weak signals were detected from the yopEΔMLD and yopE(R144A) mutants infected mice when the whole animal was imaged.
No light signals were captured from any mice infected with the three yopE mutant strains on day 13 postinfection, so the experiment was terminated. The one remaining wild-type-infected mouse had recuperated, judging by visual inspection; however, IVIS imaging still revealed a high bacterial load as determined by total photons detected.
The IVIS in vivo imaging system offers many significant advantages compared to traditional animal experiments. However, the IVIS technique is novel and has not been used for Yersinia infections previously. Therefore, to verify the findings above, infections of mice by intraperitoneal injection were also performed. Mice were infected intraperitoneally with wild-type and the yopEΔMLD mutant strains at different doses of bacteria. The wild-type showed an ID50 of 0.7 × 103 bacteria, whereas the yopEΔMLD mutant showed a 2-log (ID50 = 105) attenuation compared to the wild-type strain. The ΔyopE mutant, as well as the yopE(R144A) mutant, was found to be avirulent with ID50 values of ~107 bacteria. These latter results were in line with earlier findings showing that R144 of YopE is essential for Yersinia to cause systemic infection in mice (1). Hence, the intraperitoneal-injection experiment corroborated the data from the IVIS, showing the requirement of both the presence of the MLD, as well as a functional GAP domain in YopE for full virulence of the bacterium.
There is a family of closely related bacterial GAP proteins, and YopE from Yersinia spp. and ExoS of P. aeruginosa spp. are the most studied bacterial GAPs within this family of proteins (5, 7). YopE and ExoS have been shown to affect the activity state of Rho proteins of the target cell in vitro and to cause disruption of the actin cytoskeleton in vivo. YopE inactivates Rac1, RhoA, and Cdc42 in vitro, whereas it only targets Rac1 and RhoA during in vivo conditions (4, 6, 10, 34). Interestingly, the YopE-mediated GAP activity toward RhoA and Rac1 is not essential for virulence since specific point mutants of yopE that cannot inactivate Rac1 or RhoA are still virulent in mice (4), arguing for the presence of additional targets of YopE. This target(s) is likely another GTPase, since the GAP activity of YopE is essential for virulence (1, 10). In fact, it has been demonstrated that YopE is localized to the perinuclear region of infected eukaryotic cells (28), further supporting the idea that RhoA and Rac1 are not the sole targets of YopE.
Barbieri and coworkers (24) have shown that ExoS exhibits a specific sequence of amino acids that targets ExoS to the perinuclear region of the eukaryotic cell after ectopic expression in eukaryotic cells. This domain, called the MLD, is localized between amino acids 51 and 72 of ExoS, and the presence of MLD allows ExoS to associate with the membrane fraction of the cell. The N-terminal domain of ExoS shows high homology with YopE, and ExoS is in fact a bifunctional cytotoxin possessing an N-terminal Rho GTPase activity (YopE homology) and a C-terminal ADP-ribosyltransferase activity (8). Given the high homology between ExoS and YopE, it was not surprising that YopE also possesses an MLD similar to that of ExoS with conserved hydrophobic residues (20). To study the importance of the MLD of YopE in more detail, we constructed a yopEΔMLD mutant in cis of Y. pseudotuberculosis (YPIII/pIB529) by deleting codons corresponding to amino acids 50 to 74. Since this part of YopE constitutes half the chaperone binding site, it was first important to phenotypically characterize the yopEΔMLD mutant. The mutant was indistinguishable from the wild type with respect to in vitro Yop secretion. This was a somewhat surprising result, since it could be anticipated that interaction between YerA and YopE should be essential for YopE secretion given that a yerA mutant secrets lowered amounts of YopE in the culture supernatant (17). It is also interesting that the level of translocation of YopE and YopEΔMLD was the same after infection of HeLa cells. Thus, the yopEΔMLD can be used to study the role of MLD during in vivo conditions without any major concerns related to the secretion and translocation levels of the YopEΔMLD protein.
yopE mutants devoid of GAP activity are affected in their ability to regulate Yop expression after contact between the target cell and the bacterium has been established (3, 4). These mutants show elevated levels of Yop expression compared to the corresponding wild-type strain after infection of HeLa cells (4). We have defined this event as autoregulation. This phenotypic behavior has also been shown for ExoS, and an ExoS GAP-deficient mutant does not downregulate ExoS expression after eukaryotic cell contact (13). Interestingly, the yopEΔMLD mutant was also affected in this auto-regulatory control loop. This was surprising, since, in addition, all other mutants defective in autoregulation also display an overtranslocating phenotype, whereas the yopEΔMLD mutant did not. This implies that it is possible to dissect the regulatory role of YopE into two separate parts, one being translocation control and the other being autoregulation.
Bliska and coworkers showed convincingly the involvement of both RhoA and YopE GAP activity as important players in translocation control of Yops after infection of eukaryotic cells (10, 21, 32). The yopEΔMLD mutant retained translocational control, indicating that the mutant still can interact with and deactivate RhoA. This assumption was verified by a RhoA pulldown, where YopEΔMLD inactivated RhoA as wild-type YopE. We also assayed the effect of YopEΔMLD on Rac1, and it was found that YopEΔMLD inactivated Rac1 to the same extent as the wild-type YopE. Thus, we conclude that YopEΔMLD has retained its GAP function in vivo showing the same inactivation pattern as wild-type YopE on these two substrates. In addition, the yopEΔMLD mutant showed a delayed cytotoxicity after HeLa cell infection compared to the wild-type strain. This indicates that RhoA and Rac1 are not the main targets responsible for YopE-induced cytotoxicity. Thus, besides the Rho proteins RhoA and Rac1, YopE has at least one other target(s) that is responsible for the autoregulatory phenotype, as well as for the induction of the cytotoxic HeLa cell response.
Since YopE regulates the level of translocation, one may ask whether YopE functions as a regulatory protein for Yop translocation or whether YopE is a “true” virulence effector. We favor the former hypothesis. Hence, we suggest that YopE titrates the level of translocation to down tune unwanted host immune reactions induced by excess of translocated Yops. This idea is also supported by our earlier work showing that, although a ΔyopK mutant is much more aggressive compared to the wild type with respect to levels of translocation and cytotoxicity, the mutant is still avirulent (18, 19). Together, these results indicate that the pathogen has devoted part of its virulence arsenal to avoid the induction of innate immune defense mechanisms to remain in a stealth mode during infection.
YopE is localized to the perinuclear region of the cell after translocation but is not found in the nucleus (28). In contrast, the YopEΔMLD protein was found to be dispersed in the whole body of the target cell, including the nucleus with no distinct pattern of localization. In addition, presence of the MLD was also required to direct YopE to the intracellular membrane fraction of the eukaryotic cell (20). Thus, the MLD is important for targeting of YopE, and it is reasonable to believe that additional targets of YopE are to be found in the perinuclear compartment of the cell. ExoS is partially colocalized with Rab9, Rab6, and Rab5, whereas ExoSΔMLD does not colocalize with these proteins (36). In line with this is also the finding that Rab5 and Rab9 can be ADP ribosylated by ExoS (14, 16). Together, these results indicate that the MLD is important for directing ExoS to the perinuclear compartment involving early and late endosomes. We suggest that the MLD of YopE also is associated with vesicular trafficking, and it will be of considerable interest to define the role of YopE in this pathway since vesicle trafficking provides a novel and interesting mechanism for Yersinia-mediated virulence.
We found that a yopEΔMLD mutant was strongly attenuated for virulence both after oral and intraperitoneal infection. The yopEΔMLD mutation does not affect the level of secretion, translocation, or GAP activity of YopE, but the mutation affects the intracellular localization of YopE. This supports the importance of correct intracellular targeting for efficient function of the YopE GAP activity in mammalian cells. Further studies to understand the role of the MLD for correct targeting and identification of the target(s) of YopE should provide novel insights into bacterial virulence.
This study was supported by the Swedish Research Council. E.L.I. acknowledges the J. C. Kempes Memorial Fund for financial support.
Editor: J. B. Bliska
Published ahead of print on 17 August 2009.