SEARCH

SEARCH BY CITATION

Summary

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

Type III secretion is a widespread method whereby Gram-negative bacteria introduce toxins into eukaryotic cells. These toxins mimic or subvert a normal cellular process by interacting with a specific target, although how toxins reach their site of action is unclear. We set out to investigate the intracellular localization of a type III toxin of Pseudomonas aeruginosa called ExoU, which has phospholipase activity and requires a eukaryotic factor for activity. We found that ExoU is localized to the plasma membrane and undergoes modification within the cell by addition of two ubiquitin molecules at lysine-178. A region of five amino acids at position 679–683 near the C-terminus of the ExoU protein controls both membrane localization and ubiquitinylation. Site-directed mutagenesis identified a tryptophan at position 681 as crucial for these effects. We found that the same region at position 679–683 was also required for cell toxicity produced by ExoU as well as in vitro phospholipase activity. Localization of the phospholipase ExoU to the plasma membrane is thus required for activation and allows efficient utilization of adjacent substrate phospholipids.


Introduction

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

Type III secretion of proteins by Gram-negative bacteria is a widespread phenomenon whereby toxins are directly introduced into the cytoplasm of an infected eukaryotic cell (Cornelis and Van Gijsegem, 2000; Cornelis, 2002; Mota and Cornelis, 2005). A specialized multisubunit apparatus, the injectisome, is elaborated by the bacterium and serves as a conduit for the injected toxins (Coombes and Finlay, 2005). Once within the infected cell, the secreted toxins exert their effects, often by subverting normal cellular processes (Aktories et al., 2000; Cornelis, 2002; Aktories and Barbieri, 2005).

Many type III effectors have sequence, structural or functional similarity to eukaryotic proteins, allowing them to interfere with many different cellular processes, such as cytoskeletal organization (Francis et al., 1992; Barbieri et al., 2002), inflammatory signalling (Orth et al., 2000; Orth, 2002; Espinosa and Alfano, 2004), intracellular trafficking (Beuzon et al., 2000; Vazquez Torres et al., 2000), and progression through the cell cycle (Olson et al., 1999). These sophisticated actions allow the toxins to alter normal cellular functions for the benefit of the invading pathogen. Many of the targets of these type III secreted toxins are localized to specific intracellular organelles. Therefore, a general problem faced by type III toxins injected directly into the cell cytoplasm is how to achieve delivery to their site of action without becoming degraded within the cell. Some type III secreted toxins localize to very specific organellar locations such as the Golgi apparatus: for example, enterohaemorrhagic Escherichia coli toxin NleA (Gruenheid et al., 2004), Pseudomonas aeruginosa ExoS (Pederson et al., 2000) and the Salmonella typhimurium toxin SseG [at least in transfected cells (Salcedo and Holden, 2003)]. Others interact with actin, such as the Salmonella effectors SipA and SipC (Hayward and Koronakis, 1999; Zhou et al., 1999). The molecular mechanisms underlying these specific localizations and their relation to the pathogenicity of the microbe that secretes them are still poorly understood. However, the interactions between toxins and eukaryotic proteins ultimately will determine their delivery to specific cellular locations.

The type III secreted toxins produced by the human pathogen P. aeruginosa exemplify many typical features of type III toxins (Frank, 1997). This bacterium produces four type III secreted toxins: ExoS, T, U and Y. ExoS and ExoT are bifunctional toxins, with both ADPribosyltransferase and GTPase activating activities (Barbieri, 2000; Krall et al., 2000), ExoU is a phospholipase (Sato et al., 2003; Sato and Frank, 2004), and ExoY is an adenylate cyclase (Yahr et al., 1998). ExoS and T produce alterations in the cell cytoskeleton, which results in cell rounding and inhibition of phagocytosis (Barbieri and Sun, 2004). ExoS ultimately causes cell death whereas ExoT does not. ExoY alters cell morphology but does not produce cell death (Yahr et al., 1998; Vallis et al., 1999).

ExoU is the principal cytotoxin secreted by P. aeruginosa, producing rapid cell death (Finck-Barbancon et al., 1997; Hauser et al., 1998; Sato et al., 2003). It has recently been identified as a phospholipase, with at least phospholipase A2 activity, as well as possible lysophospholipase actions (Phillips et al., 2003; Sato et al., 2003; 2005; Tamura et al., 2004). It is an important virulence determinant in animal models of infection with this organism (Finck-Barbancon et al., 1997; Hauser et al., 1998; Shaver and Hauser, 2004). Although the phospholipase activity of ExoU is clear, there are no regions within the molecule homologous to known membrane interacting domains (Sato and Frank, 2004). How ExoU targets cellular phospholipids is therefore unclear.

Unusually, the activities of the P. aeruginosa type III secreted toxins are only enabled in the presence of eukaryotic cofactors (Coburn et al., 1991; Yahr et al., 1998; Sato et al., 2003). For ExoS and T, these have been identified as proteins of the 14-3-3 family (Fu et al., 1993); cofactors for the other secreted toxins are not known. Thus, type III secreted toxins from this organism have to interact with two eukaryotic targets: their activating cofactor and additionally the particular site within the cell where their biochemical activity can produce its effects. These may, of course, be at the same site. Activation by a eukaryotic cofactor prevents toxicity to the prokaryotic originator of the toxins.

We set out to investigate possible subcellular localization of the type III secreted toxins of P. aeruginosa and relationship, if any, to toxicity. We show here that one of these toxins, the phospholipase ExoU, is localized to the plasma membrane, and undergoes modification to a higher-molecular-weight form, found exclusively in a particulate cell fraction. The modification was identified by mass spectrometry as the addition of two ubiquitin molecules per molecule of ExoU, at lysine 178. Membrane localization and ubiquitinylation were dependent on a five-amino-acid region near the C-terminus of the ExoU molecule, which also controlled toxicity and phospholipase activity. Ubiquitinylation did not alter toxicity or activity of ExoU but did produce a modest increase in its turnover within the cell. This is the first demonstration that the localization of a type III toxin within a eukaryotic cell can dictate its enzymatic activity, efficiently targeting ExoU directly to adjacent substrate phospholipids.

Results

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

ExoU is modified to a higher-molecular-weight form within eukaryotic cells

In order to study the localization of ExoU within cells, we constructed a vector that directed the expression of ExoU fused to a FLAG epitope tag. Using the native functional ExoU sequence we were unable to detect expression of the ExoU protein, in common with other studies of this toxin (data not shown). This is due to its potent toxic effect, killing cells at very low levels and not allowing enough of the protein to accumulate for detection (Sato et al., 2003, and data not shown). We thus mutated the active site serine at position 142 to alanine, removing the phospholipase activity of ExoU (Sato et al., 2003). Following transfection of this construct (ExoU-S142A), appreciable levels of ExoU-S142A protein could be detected by immunoblotting, shown in Fig. 1A as a specific band of 85 kDa, not present in cells transfected with empty vector (data not shown). This is consistent with the molecular weight of the ExoU toxin with the FLAG epitope tag. In addition to this band, we consistently observed a higher molecular species of about 105 kDa that appeared to increase in abundance with time following transfection (Fig. 1A, marked with asterisk). This band was specific to cells transfected with the ExoU construct and not seen in cells transfected with empty vector.

image

Figure 1. Modification of ExoU to a higher-molecular-weight form.A. Immunoblot of HeLa cell lysates for ExoU at various times after cells were transfected with pCMV-Tag2-ExoU_S142A, an ExoU expression vector. Time in hours after transfection is shown above each lane. The main ExoU protein band is indicated to the right of the blot; the modified form is shown as ExoU*. Molecular weight markers in kDa are shown to the left of the gel.B. Immunoblot of HeLa cell lysates for ExoU 3 h after infection with either PA103ΔUΔT or PA103ΔUΔT transformed with the ExoU expression vector pUCP19-exoU_S142A-spcU. The main ExoU protein band is indicated to the right of the blot; the modified form is shown as ExoU*. Molecular weight markers in kDa are shown to the left of the gel.

Download figure to PowerPoint

To ensure that this additional band was not an artefact of translation of ExoU within a eukaryotic cell, we analysed ExoU within HeLa cells following infection with P. aeruginosa engineered to express the non-functional ExoU-S142A mutant. As shown in Fig. 1B, this higher-molecular-weight form of ExoU (ExoU*) was also detected in cells following infection.

The C-terminal region of ExoU is required for modification

In order to determine the region of ExoU required to produce this higher-molecular-weight form, we constructed vectors producing a number of deletion mutants of ExoU-S142A and transfected these into HeLa cells (Fig. 2A, C1–C7). Deletion of the C-terminus from amino acid 343 to the terminal amino acid at position 687 prevented formation of a modified ExoU (Fig. 2B, C3). Additional deletion mutants showed that a mutant lacking the terminal nine amino acids of ExoU was unmodified (Fig. 2B, C6), but that deletion of the terminal four amino acids did not alter modification (Fig. 2B, C7). Thus, the five amino acids between positions 679–683 were essential for modification of ExoU to the higher-molecular-weight form.

image

Figure 2. Determination of the region of ExoU required to produce modification.A. Diagram showing the different deletion constructs used and their relation to conserved residues important in the catalytic site of ExoU; these are shown as GXGXXG (111–116), S142 and D344 above the diagram. The different constructs are shown to the left with the amino acids deleted indicated. WT is wild type. The FLAG tag is shown as darker shading.B. Immunoblot of HeLa cell lysates for ExoU 16 h after transfection with the indicated expression vector constructs. All constructs contain the S142A mutation. Molecular weight markers in kDa are shown to the left.C. Diagram showing the last nine amino acids of ExoU.D. Immunoblot of HeLa cell lysates for ExoU 16 after transfection with expression constructs with either no changes (WT) or amino acid substitutions for alanine as shown. All constructs have the S142A mutation. Molecular weight markers in kDa are shown to the left.

Download figure to PowerPoint

The protein sequence of this region of ExoU is shown in Fig. 2C. In order to define which amino acids within this region were important for the modification, we individually mutated them to alanine residues and introduced these ExoU-S142A-expressing constructs into HeLa cells. Immunoblotting of extracts following transfection revealed that the tryptophan at position 681 was critical for the higher-molecular-weight modification (Fig. 2D, W681A). Individual mutation of the other amino acid residues within this C-terminal region to alanine produced no significant effect on modification of ExoU (Fig. 2D). Longer exposure of the immunoblot in Fig. 2D showed that there was a very low level of modified ExoU in the W681A mutant (not shown). This suggested that W681 was not itself the site of any modification, but was a critical amino acid residue in production of the higher-molecular-weight form of ExoU.

ExoU is ubiquitinylated at lysine 178

The increase of about 20 kDa in the apparent molecular weight of ExoU in HeLa cells as seen in the denaturing sodium dodecyl sulphate (SDS) polyacrylamide gels shown above suggested that ExoU was covalently modified by addition of a small peptide of eukaryotic origin. To establish the molecular identity of this modification, we immunoprecipitated ExoU-S142A from transfected HeLa cells using the FLAG epitope tag and cut the modified ExoU protein band from an SDS polyacrylamide gel. This material was digested with trypsin and the resulting peptides analysed by tandem mass spectrometry. Searching of the mass spectral data against a prokaryotic database confirmed the protein to be ExoU from P. aeruginosa. Using a human protein database, we identified numerous peptides from ubiquitin within this sequence (Fig. 3A, bold type). Detailed analysis of the peptide mass spectral data demonstrated that this ubiquitin was conjugated to the lysine at residue 178 (Fig. S1). Similar analysis of the unmodified ExoU protein showed that it was not modified by ubiquitin.

image

Figure 3. ExoU is ubiquitinylated at lysine 178.A. Peptides of ubiquitin identified by tandem mass spectrometry from trypsin digested modified ExoU are shown in bold type.B. HeLa cells were transfected with the indicated plasmid expression vectors and lysates made 16 h later. FLAG containing proteins were immunoprecipitated and analysed by immunoblotting for ubiquitin (left panel) or FLAG (right panel).

Download figure to PowerPoint

To confirm this site of ubiquitin conjugation, we mutated lysine 178 to arginine (K178R), a residue that is not ubiquitinylated. Cells were transfected with the ExoU-S142 construct either with lysine 178 intact or with the K178R mutation and the ExoU proteins immunoprecipitated with anti-FLAG. Immunoblotting of lysates from these cells with an antibody to ubiquitin demonstrated a 105 kDa band only in lysates from cells transfected with wild-type ExoU-S142A, not the K178R mutated form (Fig. 3B). This blot was stripped and reprobed with an anti-FLAG antibody. This demonstrated that the modified 105 kDa ExoU band was absent when the lysine at position 178 is mutated to arginine. Taken together, these data prove that ExoU is ubiquitinylated at the lysine 178 residue.

We determined the molecular weights of the two forms of immunoprecipitated ExoU by mass spectrometry. The molecular weight of the unmodified ExoU was measured at 75 493.0 Da, which is in good agreement with the predicted value of 75 489.0 Da, although the 4 Da discrepancy is slightly higher than the estimated error of ± 1–2 Da. The measured molecular weight of the modified ExoU was 17 094.4 Da higher than that of the unmodified ExoU, which could only be accounted for by the addition of two ubiquitin residues per molecule of ExoU (theoretical increase 17 093.6 Da). Further analysis of the mass spectral data obtained from the tryptic digest of modified ExoU showed that two different ubiquitin linkages could be identified, involving lysines at positions 48 and 63. However, the Lys-63-linked form was present at much higher abundance than the Lys-48-linked form.

Effect of ubiquitinylation on turnover of ExoU

As a start to understanding the functional consequences of ExoU ubiquitinylation, we measured the turnover of unmodified and modified ExoU (Fig. 4). Following inhibition of protein synthesis with cycloheximide, we followed the decay of the two ExoU forms by immunoblotting (Fig. 4A). Unmodified ExoU-S142A decayed with a half-life of between 7 and 8 h in three separate experiments. The ubiquitinylated ExoU-S142A consistently turned over rather faster, with a half-life between 5 and 6 h (Fig. 4B). This suggests that the addition of ubiquitin does not produce a dramatic change in ExoU breakdown, although it does make turnover of ExoU a little more rapid.

image

Figure 4. Influence of ubiquitinylation on the degradation of ExoU.A. Immunoblot of ExoU-S142A and ubiquitinylated ExoU-S142A (ExoU*) expressed within HeLa cells at various times in hours after the addition of 10 µM cycloheximide. Molecular weight markers in kDa are shown to the left.B. Graph of densitometric quantification of ExoU-S142A (solid line) and ubiquitinylated ExoU-S142A (dotted line, ExoU*) at various times after addition of 10 µM cycloheximide. Each point is the mean of three determinations, error bars are ± 1 SD.

Download figure to PowerPoint

Localization of ExoU within cells

Next, we analysed the subcellular localization of ExoU-S142A within cells using deconvolution immunofluorescence microscopy. In both cells transfected with the ExoU-S142A construct (Fig. 5A) and infected with P. aeruginosa that translocate the ExoU-S142A protein (Fig. 5B), we observed striking localization of ExoU to the plasma membrane. This often appeared to be rather uneven in distribution, with a punctate appearance (Fig. 5). This has previously been described for syringe-loaded recombinant ExoU within Chinese Hamster Ovary cells (Phillips et al., 2003). To determine the region of the ExoU molecule required for this localization, we analysed the distribution of the C-terminal truncation constructs that allowed us to define the region required for ubiquitinylation. In both transfected and infected cells, we found that deletion of the C-terminal amino acids from 679 to 687 (Fig. 5, C6) removed the plasma membrane localization and showed a diffuse cytoplasmic distribution. Deletion of the amino acids from 684 to 687 (Fig. 5, C7) did not alter the plasma membrane localization. Thus, exactly the same region of the C terminus was required to localize the ExoU protein to the plasma membrane as was needed to produce ubiquitinylation (Fig. 5), i.e. residues 679–683.

image

Figure 5. Distribution of ExoU within HeLa cells. A. HeLa cells were transfected with the indicated ExoU expression constructs and fixed and stained for ExoU (green) after 16 h. All constructs contain the S142A mutation. Deconvolved images are shown; nuclei are counterstained blue with DAPI.B. HeLa cells were infected with PA103ΔUΔT expressing the indicated ExoU construct. Cells were fixed 3 h after infection and stained for ExoU (green). Deconvolved images are shown; nuclei are counterstained blue with DAPI. Images are representative of at least three separate experiments. Scale bar = 10 µm.

Download figure to PowerPoint

We analysed the subcellular localization of ExoU-S142A with mutations of these C-terminal residues to alanine. The only change in distribution we observed was when the tryptophan at position 681 was mutated to alanine (Fig. 5A, W681A). This mutant ExoU was found in the plasma membrane, but this localization was much less marked than observed for the ExoU-S142A protein expressed in transfected cells. However, the distribution of the W681A mutant in infected cells was virtually indistinguishable from ExoU-S142A (Fig. 5B, W681A). This may reflect the different trafficking between the two methods of protein delivery.

This concordance between the amino acids necessary for ubiquitinylation and plasma membrane localization suggested that ubiquitinylation could be either the cause of the localization or an effect of the protein residing at this site in the cell. To differentiate between these possibilities, we analysed the localization of the K178R mutant ExoU that cannot be ubiquitinylated. This showed a clear localization to the plasma membrane that was indistinguishable from that seen with the wild-type protein (Fig. 5, K178R). Thus, the addition of the ubiquitin residues to ExoU is not the cause of plasma membrane localization, but rather seems to be the result of the protein trafficking to this location.

We have demonstrated that the C-terminal residues 679–683 were necessary for plasma membrane localization. To determine if this area of the molecule was sufficient to produce localization at this site, we made a fusion construct between green fluorescent protein (GFP) and the last 33 amino acids of ExoU. Such a fusion was found widely distributed within the cytoplasm, with no particular membrane localization. This distribution was the same as seen with wild-type GFP (Fig. S2). Thus, the C-terminal 33 amino acids of ExoU are necessary but not sufficient for plasma membrane localization.

Biochemical analysis of ExoU localization

HeLa cells transfected with ExoU-S142A were fractionated into soluble and particulate portions. Immunoblotting of these fractions showed a distribution of unmodified ExoU in both cell compartments (Fig. 6). However, the ubiquitinylated ExoU was always found exclusively within the particulate fraction (Fig. 6). Reprobing the blot with markers for cytoplasm (GAPDH) and membrane (calnexin) confirmed good separation of these two fractions (Fig. 6). These data thus support the hypothesis that ubiquitinylation of ExoU can only occur within a membrane compartment within the cell. The significant proportion of unmodified ExoU in the soluble fraction appears out of keeping with the distribution of ExoU as seen by immunofluorescence (Fig. 5). However, there is staining above background throughout the cytoplasm that accounts for the significant amount of ExoU detected in the soluble compartment by the biochemical assay (Fig. 6).

image

Figure 6. Distribution of ExoU and ubiquitinylated ExoU in cell fractions. HeLa cells expressing ExoU-S142A were fractionated into soluble and particulate fractions and analysed for ExoU by immunoblotting with anti-FLAG antibodies. ExoU and ubiquitinylated ExoU (ExoU*) are indicated to the left of the panel. Blots were stripped and reprobed with antibodies for calnexin (membrane protein) and GAPDH (cytoplasmic protein) to gauge purity of fractions.

Download figure to PowerPoint

Toxicity of ExoU depends on the C-terminal residues 679–683

To explore the relationship between membrane localization, ubiquitinylation and the toxicity of ExoU in eukaryotic cells, we measured the ability of various ExoU mutants to produce cell death in both transfected cells and cells infected with P. aeruginosa. First, we transfected HeLa cells with a number of expression vector constructs for ExoU with the mutations described above that allowed us to identify the region of the molecule essential for membrane localization and ubiquitinylation. Toxicity in this model was assayed by cotransfection of a constitutive expression vector for luciferase and assaying for this protein. Vectors expressing wild-type ExoU produced about 100-fold less luciferase compared with controls transfected with empty vector (Fig. 7A). Mutation of the active site serine in ExoU abrogated this toxic effect (Fig. 7A, ExoU-S142A), with luciferase levels the same as those seen in cells transfected with empty vector. Deletion of the C-terminal 24 amino acids of ExoU rendered the molecule completely non-toxic in this model (Fig. 7A, C4). Deletion of the C-terminal 17 or nine amino acids (Fig. 7A, C5 and C6) were also lacking in full toxicity, but were significantly slightly more toxic in this assay than the C4 or S142A mutants (Fig. 7A). Deletion of the terminal four amino acids (Fig. 7A, C7) produced a protein that was not significantly different in toxicity from the wild-type ExoU. Thus, the same region of five amino acids between positions 679–683 that localize the protein to the plasma membrane and direct ubiquitinylation are also required for full toxicity of the ExoU protein. A small additional contribution is made by amino acids 663–678. Substitution of alanine for the individual residues between 679 and 683 did not produce any significant alteration in toxicity of ExoU (Fig. 7A), including the W681A mutant. In addition, transfection of the expression construct for the K178R mutant that cannot be ubiquitinylated produced as much toxicity as wild-type ExoU (Fig. 7A, K178R). Thus, ubiquitinylation by itself does not alter ExoU toxicity.

image

Figure 7. Role of C-terminal amino acids on cytotoxicity and phospholipase activity of ExoU.A. Cytotoxicity of various ExoU constructs in the expression vector pCMV-Tag2 transfected into HeLa cells together with the luciferase expression vector, pGL2. Results are the means of triplicate determinations of luciferase activity (relative light units, RLU) assayed 16 h after transfection. Error bars are ± 1 SD.B. Cytotoxicity caused by infection of HeLa cells with PA103ΔUΔT expressing the indicated ExoU constructs. Results are the means of three determinations of percentage cytotoxicity assayed by LDH release; error bars are ± 1 SD.C. Phospholipase activity of indicated recombinant ExoU proteins alone or where indicated in the presence of a HeLa cell extract. Results are the means of three determinations; error bars are ± 1 SD.

Download figure to PowerPoint

We repeated these measures of ExoU toxicity using infection of cells with P. aeruginosa engineered to deliver the different ExoU proteins by the type III secretion system. In this assay, percentage cell death was determined by lactate dehydrogenase (LDH) release (Fig. 7B). Again, using the C-terminal deletion mutants, we found that the C6 mutant was virtually completely non-toxic, but that the C7 mutant was as toxic as the wild-type ExoU (Fig. 7B). Thus, in natural infection, as with transfected ExoU, amino acids 679–683 are critical for toxicity of ExoU towards eukaryotic cells. We also found that the W681A mutant was as toxic as the wild-type molecule, as was the K178R mutant that cannot be ubiquitinylated (Fig. 7B). These were the same results as found for transfected ExoU.

Phospholipase activity of ExoU depends on the C-terminal residues 679–683

The results described above have allowed us to identify a region of the ExoU molecule that directs it to the plasma membrane, is required for full toxicity and directs ubiquitinylation. In order to understand the contribution of the 679–683 amino-acid region towards the toxicity of ExoU, we examined the effect of these residues on the phospholipase activity of the toxin. We produced recombinant ExoU in a bacterial expression system and assayed its phospholipase A2 activity in vitro. In keeping with other studies, recombinant ExoU on its own showed no phospholipase activity [Fig. 7C, rExoU (Sato et al., 2003)]. To produce enzymatically active ExoU, we had to add a eukaryotic cofactor found in HeLa cell extract (Fig. 7C, rExoU + lysate). Mutation of the active site serine of ExoU abolished phospholipase activity as expected (Fig. 7C, S142A + lysate). Next, we tested the effect of the various C-terminal deletions on the phospholipase activity of ExoU. We found that removal of the C-terminal nine amino acids abolished enzymatic activity of the molecule (Fig. 7C, C6). Deletion of the terminal four amino acids did not affect enzymatic activity (Fig. 7C, C7). Thus, the five-amino-acid residues between positions 679–683 were critical for phospholipase activity of ExoU. These are the same residues essential for plasma membrane localization and ubiquitinylation. We also tested the effects of the W681A mutation on ExoU enzyme activity. In contrast to the lack of effect of this mutation on toxicity of ExoU within cells, this mutation did reduce ExoU phospholipase activity to about 20% of the levels produced by the wild-type toxin (Fig. 7C). Recombinant protein that had the K178R mutation and thus cannot be ubiquitinylated had a small reduction in phospholipase A activity (Fig. 7C).

The observed defects in ExoU localization, ubiquitinylation and enzymatic activity are dependent on the five amino acids at positions 679–683. In order to exclude an effect of this region on folding or gross conformation of the ExoU protein that could account for these differences, we compared circular dichroism (CD) spectra and thermal denaturation curves for recombinant wild-type ExoU and the C6 mutant protein, lacking the last nine-amino-acid residues (679–687) (Fig. 8). The far UV CD spectra show that both wild-type and truncated ExoU have broadly similar secondary structures. The spectral differences probably reflect contributions from tryptophan 668, which is absent from the C6 mutant-truncated ExoU. Significant contributions from tryptophan side-chains in the far UV region have been demonstrated in a number of studies (Freskgard et al., 1994; Grishina and Woody, 1994). The thermal unfolding profiles (Fig. 8, inset panel) show that the stabilities of the wild type and C6 truncation mutant of ExoU are comparable. Taken together, these data demonstrate that the overall conformation of the truncated ExoU has not been compromised by the C-terminal deletion. Thus, the effects of this C terminal region between residues 679–683 on the ubiquitinylation and enzymatic activity of ExoU most likely reflect the inability of this mutant protein to be delivered to its correct location in the plasma membrane, rather than improper folding or conformation.

image

Figure 8. Circular dichroism spectra and thermal denaturation of wild type and C6 truncation mutant of ExoU. Far UV CD spectra of wild type (solid line) and truncated (mutant C6 lacking amino acids 679–687, dashed line) ExoU. The inset shows the thermal unfolding profiles of both wild-type (squares) and truncated (circles) ExoU.

Download figure to PowerPoint

Localization of the eukaryotic cofactor for ExoU

Previous studies have shown the importance of the C-terminal region of ExoU for its toxicity and phospholipase activity in mammalian and yeast cells (Hauser et al., 1998; Phillips et al., 2003; Rabin and Hauser, 2005). We have demonstrated that five amino acids at positions 679–683 within the C-terminus of ExoU regulate its phospholipase activity and toxicity. These are well away from the regions of the enzyme known to be critical for its activity, the serine–aspartate dyad at positions 142 and 344, and the glycine-rich motif at positions 111–116 (Sato et al., 2003, Fig. 2A). The importance of the 679–683 region in membrane localization, as well as phospholipase activity, suggests that the required cofactor for enzyme activity may be membrane localized. To test this, we fractionated HeLa cell extracts into soluble and particulate fractions as before and assayed the ability of each fraction to produce activation of recombinant ExoU in an in vitro phospholipid assay. As shown in Fig. 9, nearly all the activating eukaryotic cofactor was found in the particulate fraction (HeLa P).

image

Figure 9. Cofactor for ExoU activation resides in a particulate fraction. Phospholipase activity of recombinant rExoU was determined alone or in the presence of unfractionated HeLa cell extract (lysate) or fractionated into a particulate (lysate P) or soluble (lysate S) fractions.

Download figure to PowerPoint

Discussion

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

We have demonstrated here that a type III secreted bacterial toxin, ExoU, is trafficked to the plasma membrane. This process is essential for the toxin to express its enzymatic activity and produce cytotoxicity. It also results in the addition of a diubiquitin moiety to the toxin that produces a modest increase in protein turnover but no alteration in enzymatic activity or toxicity. The area that produces this effect is distinct from the catalytic site responsible for the toxin’s phospholipase action and has no homology to any known eukaryotic membrane-localizing domains. These novel findings are the first demonstration of a specific cellular localization determining the ability of a type III secreted bacterial toxin to cause cytotoxicity. The capacity of a bacterial protein to be able to interact with eukaryotic membranes and to be ubiquitinylated is a remarkable adaptation to life within a cell from a different kingdom, and produces a highly efficient toxin that does not affect its prokaryotic originator.

The localization of ExoU to the plasma membrane requires five amino acids at the C-terminus of the protein between positions 679–683. Deletion of this region does not produce major alterations in protein folding or conformation as determined by CD spectroscopy and thermal denaturation (Fig. 8). Although essential for plasma membrane interaction, the C-terminal 33 amino acids are not sufficient to direct membrane localization, as shown in Fig. S2. Thus, additional sequences within ExoU are required for plasma membrane localization. We do not know what dictates the interaction of ExoU with the plasma membrane, but it could either be an interaction with a lipid or with a membrane protein. No homologies to known lipid interacting domains are present within this region (Hurley and Misra, 2000). Preliminary binding studies to immobilized lipids suggests that ExoU can bind to a number of phospholipid species, but this binding is not dependent on the C-terminal region (data not shown). As a phospholipase, ExoU will have a reasonable affinity to its substrates, and this may be sufficient to produce some membrane binding. This could then be strengthened by binding to a membrane protein. The eukaryotic cofactor required for ExoU action is only present in a membrane-rich, particulate cell fraction (Fig. 9). This could be a membrane protein or even a lipid, although previous work suggests the factor is proteinaceous (Sato et al., 2005) and we failed to find any stimulation of ExoU activity by the addition of a number of purified phospholipids (data not shown).

The same C-terminal amino acids required for plasma membrane localization are also important in producing the cytotoxic effects of ExoU and determining its phospholipase activity (Fig. 7). In these assays, mutation of tryptophan-681 to alanine significantly reduced phospholipase activity of ExoU but did not influence its cytotoxicity. This probably reflects the potency of ExoU as a phospholipase, such that even a reduction of its enzymatic activity to 20% of wild-type levels, as produced by the W681A mutation, is still sufficient to produce cell death (Sato et al., 2003).

How does the C-terminus affect the phospholipase activity of ExoU? We propose the following two models (Fig. 10). One possibility is that the C-terminal amino acids that we have identified as required for plasma membrane localization are involved in binding to a membrane protein that acts as a cofactor for ExoU activation (Fig. 10A). An alternative explanation is that the C-terminus is required to localize the ExoU to a lipid interface that is also required for enzyme activation (Fig. 10B). Such interfacial activation of a phospholipase has been described for mammalian cytosolic phospholipase A2 (Nalefski et al., 1994). Additional activation by a eukaryotic cofactor would still, however, be required, as on its own rExoU has no phospholipase activity (Fig. 7). Artificial targeting of the C6 truncation mutant of ExoU to the plasma membrane by incorporation of prenylation signals may allow a distinction between these two possibilities.

image

Figure 10. Model of ExoU interaction with the plasma membrane. ExoU (blue) interacting by its C-terminal amino acids (C-term) either directly (A) with a membrane-localized eukaryotic cofactor (CF in red) or with membrane lipids (B). Membrane localization leads to the addition of a two-ubiquitin molecule (Ub, green).

Download figure to PowerPoint

The plasma membrane localization of ExoU brings it into close contact with its major substrate, phospholipids. The enzymatic activity of ExoU shows it to have a fairly broad specificity, able to hydrolyse phosphatidylcholine and phosphatidylethanolamine (Sato et al., 2005), as well as arachidonic acid containing lipids (Rabin and Hauser, 2005) and lysophospholipids (Tamura et al., 2004). The net result of this activity is rapid necrotic cell death (Finck-Barbancon et al., 1997). Bacterial lipids are also attacked by ExoU, so that the activation of the enzyme by a eukaryotic factor is crucial in limiting the toxic effects of ExoU synthesized within the bacterial cell (Sato et al., 2005). Mammalian phospholipases that are homologous to ExoU such as cytosolic phospholipase A2 (cPLA2) and calcium-independent PLA2 (iPLA2) have similar substrate specificities. The effects of these mammalian enzymes within cells tend to be more important in initiating inflammation rather than cell death, although they can contribute to cell lysis (Cummings et al., 2000). ExoU does play a role in inflammation (McMorran et al., 2003), although its greater ability to produce cell death remains unexplained. cPLA2 associates with the plasma membrane when free calcium levels are elevated, so this localization cannot be the reason behind the different cellular effects of these phospholipases.

The ubiquitinylation of ExoU was unexpected. Two other examples of this modification of type III secreted toxins have been reported, SopB and SopE in Salmonella (Marcus et al., 2002; Kubori and Galan, 2003). ExoU only has two molecules of ubiquitin added, too few to allow diversion to the proteosomal degradation pathway. The majority of the added ubiquitin was conjugated via the Lys-63, a link that is not associated with targeting to the proteosome. In agreement with this, we did not find that ubiquitinylation produced much effect on ExoU turnover (Fig. 4). The addition of only two ubiquitins to a protein is unusual, with most proteins showing conjugation either to one ubiquitin or to a polyubiquitin chain (Pickart and Fushman, 2004). It is not clear what effect the addition of ubiquitin has on ExoU, as it does not alter its distribution and preventing ubiquitin addition has only a very small effect on enzymatic activity (Fig. 7C), and no effect on cytotoxicity (Fig. 7B). It is interesting that the ubiquitinylated form is only found in the particulate membrane-rich cell fraction (Fig. 6). Other proteins that are modified by a short-ubiquitin chain do show effects on localization, usually by altering their trafficking within the cell (Pickart and Fushman, 2004). Further work will be required to investigate this for ExoU. The tryptophan-681 residue plays a very important role in the ubiquitin addition. The data presented here suggest that ubiquitinylation is a result rather than a cause of membrane localization (Fig. 5). In other phospholipases, tryptophan has been shown to be important in binding to the interfacial region of zwitterionic phospholipid bilayers (Han et al., 1999; Beers et al., 2003). This may account for the effect of Trp-681 on phospholipase activity. Another possibility is that Trp-681 is involved in a direct interaction with a membrane-bound ubiquitin ligase complex.

In conclusion, we have shown here that a type III secreted bacterial toxin, ExoU, is specifically localized to the plasma membrane. We have defined a region of five amino acids within the C-terminus that determined this localization, as well as cytotoxicity and phospholipase activity of the toxin, and addition of ubiquitin. These novel findings demonstrate that localization of type III secreted toxins can influence their activity and provide further evidence of the remarkable adaptation of these toxins to a eukaryotic cell environment.

Experimental procedures

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

Plasmids

The plasmids used in this study are listed in Table 1.

Table 1.  Plasmids used in this study.
PlasmidDescriptionSource or reference
pCMV-Tag2N-terminal FLAG tag fusion mammalian expression vector, KanrStratagene
pCMV-Tag2-ExoUpCMV-Tag2 encoding ExoUThis study
pCMV-Tag2-ExoU_S142ApCMV-Tag2 encoding ExoU containing the S142A mutationThis study
pCMV-Tag2-ExoU_K178RpCMV-Tag2 encoding ExoU containing the K178R mutationThis study
pCMV-Tag2-ExoU_K679ApCMV-Tag2 encoding ExoU containing the K679A mutationThis study
pCMV-Tag2-ExoU_W681ApCMV-Tag2 encoding ExoU containing the W681A mutationThis study
pCMV-Tag2-ExoU_R682ApCMV-Tag2 encoding ExoU containing the R682A mutationThis study
pCMV-Tag2-ExoU_N683ApCMV-Tag2 encoding ExoU containing the N683A mutationThis study
pCMV-Tag2-ExoU_C4pCMV-Tag2 encoding ExoUΔ663–687This study
pCMV-Tag2-ExoU_C5pCMV-Tag2 encoding ExoUΔ671–687This study
pCMV-Tag2-ExoU_C6pCMV-Tag2 encoding ExoUΔ679–687This study
pCMV-Tag2-ExoU_C7pCMV-Tag2 encoding ExoUΔ684–687This study
pCMV-Tag2-ExoU_S142A_K178RpCMV-Tag2 encoding ExoU containing the S142A and K178R mutationThis study
pCMV-Tag2-ExoU_S142A_K679ApCMV-Tag2 encoding ExoU containing the S142A and K679A mutationThis study
pCMV-Tag2-ExoU_S142A_W681ApCMV-Tag2 encoding ExoU containing the S142A and W681A mutationThis study
pCMV-Tag2-ExoU_S142A_R682ApCMV-Tag2 encoding ExoU containing the S142A and R682A mutationThis study
pCMV-Tag2-ExoU_S142A_N683ApCMV-Tag2 encoding ExoU containing the S142A and N683A mutationThis study
pCMV-Tag2-ExoU_S142A_K684ApCMV-Tag2 encoding ExoU containing the S142A and K684A mutationThis study
pCMV-Tag2-ExoU_S142A_E685ApCMV-Tag2 encoding ExoU containing the S142A and E685A mutationThis study
pCMV-Tag2-ExoU_S142A_F686ApCMV-Tag2 encoding ExoU containing the S142A and F686A mutationThis study
pCMV-Tag2-ExoU_S142A_T687ApCMV-Tag2 encoding ExoU containing the S142A and T687A mutationThis study
pCMV-Tag2-ExoU_S142A_C1pCMV-Tag2 encoding ExoUΔ654–687 containing the S142A mutationThis study
pCMV-Tag2-ExoU_S142A_C2pCMV-Tag2 encoding ExoUΔ413–687 containing the S142A mutationThis study
pCMV-Tag2-ExoU_S142A_C3pCMV-Tag2 encoding ExoUΔ343–687 containing the S142A mutationThis study
pCMV-Tag2-ExoU_S142A_C4pCMV-Tag2 encoding ExoUΔ663–687 containing the S142A mutationThis study
pCMV-Tag2-ExoU_S142A_C5pCMV-Tag2 encoding ExoUΔ671–687 containing the S142A mutationThis study
pCMV-Tag2-ExoU_S142A_C6pCMV-Tag2 encoding ExoUΔ679–687 containing the S142A mutationThis study
pCMV-Tag2-ExoU_S142A_C7pCMV-Tag2 encoding ExoUΔ684–687 containing the S142A mutationThis study
pUCP19-ExoU-SpcUpUCP19 containing exoUspcU as a 2.8 kb EcoRV fragmentDW Frank
pUCP19-ExoU_S142A-SpcUpUCP19 encoding ExoU containing the S142A mutation and SpcUThis study
pUCP19-ExoU_K178R-SpcUpUCP19 encoding ExoU containing the K178R mutation and SpcUThis study
pUCP19-ExoU_W681A-SpcUpUCP19 encoding ExoU containing the W681A mutation and SpcUThis study
pUCP19-ExoU_C6-SpcUpUCP19 encoding ExoUΔ679-687 and SpcUThis study
pUCP19-ExoU_C7-SpcUpUCP19 encoding ExoUΔ684-687 and SpcUThis study
pUCP19-ExoU_S142A_K178R-SpcUpUCP19 encoding ExoU containing the S142A and K178R mutations and SpcUThis study
pUCP19-ExoU_S142A_W681A-SpcUpUCP19 encoding ExoU containing the S142A and W681A mutations and SpcUThis study
pUCP19-ExoU_S142A_C6-SpcUpUCP19 encoding ExoUΔ679–687 containing the S142A mutation and SpcUThis study
pUCP19-ExoU_S142A_C7-SpcUpUCP19 encoding ExoUΔ684–687 containing the S142A mutation and SpcUThis study
pEGFP-C3N-terminal EGFP tag fusion mammalian expression vector, KanrBD Biosciences Clontech
pEGFP-ExoU653-687pEGFP-C3 encoding ExoU653–687This study
PET-100/D-TOPON-terminal Xpress and His tag fusion E. coli expression vector, AmprInvitrogen
pET-100/D-ExoUpET-100/D encoding ExoUThis study
pET-100/D-ExoU_S142ApET-100/D encoding ExoU containing the S142A mutationThis study
pET-100/D-ExoU_K178RpET-100/D encoding ExoU containing the K178R mutationThis study
pET-100/D-ExoU_W681ApET-100/D encoding ExoU containing the W681A mutationThis study
pET-100/D-ExoU_C6pET-100/D encoding ExoUΔ679–687This study
pET-100/D-ExoU_C7pET-100/D encoding ExoUΔ684–687This study

exoU was amplified from a clinical strain of P. aeruginosa (PA4) genomic DNA using the forward primer CGGGATCCATG CATATCCAATCGTTGGGG and the reverse primer TCGGGC CCTCATGTGAACTCCTTATTCCGCCA and cloned into the BamHI and ApaI sites of pCMV-Tag2B.

pUCP19-ExoU-SpcU was used to express ExoU and its chaperone (kindly supplied by Prof D. Frank, Medical College of Wisconsin, USA).

The exoU gene was amplified from PA4 genomic DNA using the forward primer CACCATGCATATCCAATCGTTGGGG and the reverse primer TCATGTGAACTCCTTATTCCGCCA and cloned into pET-100/D-TOPO directional cloning vector.

Various mutations were introduced into the exoU gene in pCMV-Tag2-ExoU, pUCP19-ExoU-SpcU and pET-100D-ExoU using the QuikChange Site-Directed Mutagenesis Kit (Stratagene) according to manufacturer’s instructions. The primers used for the site-directed mutagenesis are listed in Table 2. Each site-directed mutagenesis reaction also introduced an extra restriction site that enabled rapid screening for successful mutants.

Table 2.  Primers used for site-directed mutagenesis.
PrimerSequenceRestriction sites
S142A-FwGTCCGGTTCGGCCGCTGGCGGCAEagI
S142A-RevTGCCGCCAGCGGCCGAACCGGACEagI
K178R-FwCTCGATAGCTCGAACAGGAAGCTTAAGCTGTTCCAACACAHindIII
K178R-RevTGTGTTGGAACAGCTTAAGCTTCCTGTTCGAGCTATCGAGHindIII
K679A-FwCTACCGTTGAGATGGCTGCAGCTTGGCGGAATAAGGAGTTPstI
K679A-RevAACTCCTTATTCCGCCAAGCTGCAGCCATCTCAACGGTAGPstI
W681A-FwCCGTTGAGATGGCCAAGGCTGCGCGGAATAAGGAGTTCMscI
W681A-RevGAACTCCTTATTCCGCGCAGCCTTGGCCATCTCAACGGMscI
R682A-FwTTGAGATGGCTAAGGCCTGGGCGAATAAGGAGTTCACATGAGStuI
R682A-RevCTCATGTGAACTCCTTATTCGCCCAGGCCTTAGCCATCTCAAStuI
N683A-FwATGGCTAAGGCTTGGCGCGCTAAGGAGTTCACATGAGCBssHII
N683A-RevGCTCATGTGAACTCCTTAGCGCGCCAAGCCTTAGCCATBssHII
K684A-FwCTAAGGCTTGGCGGAATGCGGAATTCACATGAGCGGCCEcoRI
K684A-RevGGCCGCTCATGTGAATTCCGCATTCCGCCAAGCCTTAGEcoRI
E685A-FwGCTTGGCGGAATAAGGCCTTCACATGAGCGGCStuI
E685A-RevGCCGCTCATGTGAAGGCCTTATTCCGCCAAGCStuI
F686A-FwGCTTGGCGGAATAAGGAGGCTACGTAAGCGGCCGCTCGAGTCSnaBI
F686A-RevGACTCGAGCGGCCGCTTACGTAGCCTCCTTATTCCGCCAAGCSnaBI
T687A-FwGGCGGAATAAGGAATTCGCATGAGCGGCCGCTEcoRI
T687A-RevAGCGGCCGCTCATGCGAATTCCTTATTCCGCCEcoRI
C4-FwACAACTACTCGGCACGAGGTTAACTGCGTTTCGGCAAACCHpaI
C4-RevGGTTTGCCGAAACGCAGTTAACCTCGTGCCGAGTAGTTGTHpaI
C5-FwCTTCCTGCGTTTCGGCAAACCCCTTTAAAGCACTACCGTTGAGATGGCTADraI
C5-RevTAGCCATCTCAACGGTAGTGCTTTAAAGGGGTTTGCCGAAACGCAGGAAGDraI
C6-FwCTACCGTTGAGATGGCTTAAGCTTGGCGGAATAAGGAGTTHindIII
C6-RevAACTCCTTATTCCGCCAAGCTTAAGCCATCTCAACGGTAGHindIII
C7-FwATGGCTAAGGCTTGGCGGAATTGAATTCTCACATGAGCGGCCGCTCEcoRI
C7-RevGAGCGGCCGCTCATGTGAGAATTCAATTCCGCCAAGCCTTAGCCATEcoRI

Pseudomonas aeruginosa strains and electroporation

The P. aeruginosa strain used in this study was PA103ΔexoUΔexoT::Tc, referred to hereafter as PA103ΔTΔU (kindly supplied by Prof D. Frank, Medical College of Wisconsin, USA).

Electrocompetent PA103ΔTΔU were prepared and transformed with approximately 0.5 µg of pUCP19-based vector DNA as previously described (Diver et al., 1990). The transformants were selected on Luria–Bertani (LB) agar containing 300 µg ml−1 carbenicillin.

Cell culture and transfection

HeLa cells were grown at 37°C with 5% CO2 in a humid atmosphere in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, 10 U ml−1 penicillin and 10 µg ml−1 streptomycin (all supplied by Invitrogen). HeLa cells were transfected with plasmid DNA once they had reached 90–95% confluence using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions.

Infection of HeLa cells with P. aeruginosa

For infection, PA103ΔTΔU was grown in LB medium (supplemented with 300 µg ml−1 carbenicillin as required) at 37°C with shaking until OD600 = 0.4. The bacteria were collected by centrifugation, washed twice, and resuspended in sterile PBS to obtain a suspension containing approximately 5 × 105 cfu µl−1. Cells were infected at a multiplicity of infection of 10 and analysed 3–4 h later.

Immunoblotting

Transfected HeLa cells were lysed directly into 2× SDS gel-loading buffer and analysed by immunoblotting. Immunoblots were incubated with the primary antibody for 1 h at the following dilutions: mouse anti-FLAG monoclonal antibody [1:1000 dilution (Sigma-Aldrich)], rabbit anti-ExoU polyclonal antibody (1/500 dilution), mouse anti-ubiquitin monoclonal antibody [1/1000 dilution (Sigma-Aldrich)], rabbit anti-GAPDH polyclonal antibody [1/1000 dilution (Abcam)], or rabbit anti-calnexin polyclonal antibody [1/2000 dilution (Stressgen)]. Bound antibody was detected with an appropriate biotinylated secondary antibody (Vector Laboratories), horseradish peroxidase (HRP)-conjugated streptavidin (Biosource) and ECL Plus Western Blotting Detection Reagents (Amersham Biosciences).

Immunoprecipitation of ExoU

The HeLa cells were washed twice with ice-cold PBS and lysed by incubating with lysis buffer (50 mM Tris HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100 containing 10 µl protease inhibitor cocktail per ml) on ice with agitation for 30 min. A post-nuclear supernatant was prepared by centrifuging the lysed cells at 16 100 g for 10 min at 4°C.

The Anti-FLAG M2 Affinity Gel (Sigma-Aldrich) was prepared by centrifuging the gel suspension at 16 100 g for 1 min at 4°C and washing the packed gel twice in ice-cold TBS. Immunoprecipitation of FLAG-tagged ExoU was achieved by incubating the HeLa cell lysate with the packed gel for 3 h at 4°C with rotation. The gel was then washed three times with ice-cold TBS to remove unbound proteins. For elution, 1.5 µl of 3× FLAG peptide (Sigma-Aldrich) stock solution (5 µg µl−1 3× FLAG peptide in 100 mM Tris HCl, pH 7.5, 200 mM NaCl) was added to 50 µl TBS (150 ng µl−1 final concentration) and this was incubated with the packed gel for 30 min at 4°C with rotation. The gel was then centrifuged at 16 100 g for 1 min at 4°C and the supernatant containing the eluted proteins was transferred to a fresh tube.

Mass spectrometric analysis

A 75 cm2 flask of HeLa cells was transfected with pCMV-Tag2-ExoU_S142A and 20 h after transfection, the HeLa cell lysate was immunoprecipitated with Anti-FLAG M2 Affinity Gel (Sigma-Aldrich) as described above. The proteins were separated by 4–12% SDS-PAGE and the proteins were fixed and stained with 0.25% Coomassie brilliant blue R250 (Sigma-Aldrich).

The Coomassie-stained bands were destained, reduced, carbamidomethylated and digested overnight with 10 ng µl−1 modified trypsin (Promega) in 25 mM ammonium bicarbonate at 30°C. The resulting peptide mixtures were separated by reverse-phase liquid chromatography (column: 0.1 × 100 mm, Vydac C18, 5 µm particle size), with an acetonitrile gradient (0–30% over 30 min) containing 0.1% formic acid, at a flow rate of 500 nl min−1. The column was coupled to a nanospray ion source (Protana Engineering) fitted to a quadrupole-TOF mass spectrometer (Qstar Pulsar I; Applied Biosystems/MDS Sciex). The instrument was operated in information-dependent acquisition mode, with an acquisition cycle consisting of a 0.5 s TOF scan over the m/z range 350–1500 followed by 2 s MS/MS scans (triggered by 2+ or 3+ ions), recorded over the m/z range 100–1700. Proteins were identified by database searching of the mass spectral data using Mascot software (Matrix Science). Mascot was also used to determine ubiquitinylation sites using the Lysine GlyGly ‘variable modification’ parameter. All identified ubiquitinylation sites were verified by manual interpretation of the corresponding MS/MS spectra. All the unassigned MS/MS spectra of significant intensity were also manually interpreted.

For the intact mass measurements, excess FLAG peptide was removed from immunoprecipitated ExoU by ultrafiltration through a 30 kDa membrane (Pall NanoSep 30K). A portion of the retentate was loaded onto a C4 reversed-phase capillary column (0.1 × 10 mm, Brownlee Aquapore Butyl) and eluted with a gradient of 10–50% acetonitrile containing 0.1% formic acid into the mass spectrometer as described above. TOF spectra were acquired over the m/z range 500–2400 and deconvoluted using the BioAnalystQS Bayesian Protein Reconstruct function.

Stability of ExoU and ubiquitinylated ExoU

HeLa cells were transfected with pCMV-Tag2-ExoU_S142A. Twelve hours after transfection, protein synthesis was inhibited by incubating the HeLa cells with 25 µg ml−1 cycloheximide (Sigma-Aldrich). At various times after addition of cycloheximide, the HeLa cells were washed twice in PBS and ExoU assayed by immunoblotting with the anti-FLAG antibody as described above. Densitometric analysis was performed using NIH Image (National Institute of Health).

Immunofluorescence staining

HeLa cells transfected with a pCMV-Tag2-based plasmid or infected with PA103ΔTΔU containing a pUCP19-based plasmid were immunostained with an anti-FLAG or a polyclonal anti-ExoU antibody respectively. Bound antibody was visualized using Alexafluor conjugated secondary reagents; nuclei were counterstained with DAPI. Cells were viewed using a Zeiss Axiovert S100 microscope. For each view, five images were taken at 0.5 micron intervals and Openlab software (Improvision) was used to deconvolve the images.

Cell fractionation

HeLa cells were transfected with pCMV-Tag2-ExoU_S142A. After 16 h of transfection, the cells were washed in ice-cold PBS and lysed in 0.5 ml ice-cold lysis buffer (20 mM Tris-Cl, 10 mM EDTA) supplemented with 5 µl ml−1 protease inhibitor cocktail (Sigma). The lysates were sonicated before being centrifuged at 1500 g for 5 min at 4°C to remove the nuclei. The supernatant was transferred to a pre-chilled 13 mm × 51 mm polyallomer centrifuge tube (Beckman Instruments) and centrifuged at 100 000 g for 1 h at 4°C. The supernatant was removed and stored while the pellet was washed by resuspending in 500 µl lysis buffer and centrifuging at 100 000 g for 1 h. The washed pellet was resuspended in SDS gel-loading buffer by vigorous vortexing and heating at 95°C for 5 min. For use in phospholipase assays, the pellet was solubilized in the original cell lysis buffer by sonication.

Luciferase assay

HeLa cells in a 24 well plate were transfected with the pGL2-control plasmid (Promega), which expresses firefly luciferase in mammalian cells, and a pCMV-Tag2-ExoU or mutant ExoU plasmid. Twenty-four hours after transfection, the HeLa cell media were replaced with fresh media and the cells were incubated for a further 24 h. Luciferase was assayed with a Luciferase Assay Kit (Stratagene) as a measure of viable cells.

Lactate dehydrogenase assay

HeLa cells were grown to 90% confluence in a 24 well plate. Two hours before infection, the normal HeLa media were replaced with DMEM minus phenol red supplemented with 1% FCS and 2 mM l-glutamine to reduce background absorbance. The HeLa cells were infected with P. aeruginosa as described above. After 4 h of infection, the HeLa cells in the 24 well plates were centrifuged at 250 g for 4 min. The HeLa cell culture supernatant were diluted 1 in 10 in DMEM minus phenol red supplemented with 1% FCS and 2 mM l-glutamine and 50 µl aliquots were transferred to a 96 well plate. The amount of LDH released by HeLa cell lysis after P. aeruginosa infection was measured using the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega) according to the manufacturer’s instructions.

Production of recombinant ExoU

BL21 Star (DE3) One Shot Chemically Competent Cells (Invitrogen) were transformed with 1 µl pET-100/D-ExoU or ExoU mutant according to the manufacturer’s instructions. The entire transformation mix was added to 10 ml of LB broth supplemented with 100 µg ml−1 ampicillin and grown overnight at 37°C with shaking. Five millilitres of the overnight culture was used to inoculate 500 ml of LB broth supplemented with 100 µg ml−1 ampicillin and the culture was incubated at 30°C with shaking until the OD600 = 0.5–0.8. ExoU expression was induced by adding 1 mM IPTG (Invitrogen) and incubating for a further 2 h at 30°C with shaking. The culture was centrifuged at 9950 g for 10 min at 4°C. The bacterial pellet was resuspended in 5 ml 1× Ni-NTA Bind Buffer (Novagen) by vigorous vortexing. Lysozyme was added to the resuspended bacteria at a final concentration of 0.5 mg ml−1 and the mixture was incubated at 30°C for 10 min. The lysate was sonicated for 1 min in 10 s bursts on ice, centrifuged at 9950 g for 20 min at 4°C and the supernatants were removed and stored on ice. Ni-NTA HisBind Resin (Novagen) was used to purify the His-tagged proteins as described by the manufacturer, with all the steps being carried out at 4°C. The eluted fractions were analysed for purity by SDS-PAGE and the amount of protein was quantified using the Bio-Rad Protein Assay (Bio-Rad) according to the manufacturer’s instructions.

Phospholipase activity assay

The cPLA2 Assay Kit (Cayman) was used to determine the phospholipase activity of the recombinant ExoU and mutant ExoU proteins. The HeLa cell lysate for the phospholipase activity assay was prepared by lysing a confluent 75 cm2 flask of HeLa cells with 0.5 ml ice-cold lysis buffer (50 mM Hepes pH 7.4, 1 mM EDTA). The cells were sonicated and centrifuged at 1000 g for 15 min at 4°C and the supernatant collected.

Circular dichroism and thermal denaturation

Circular dichroism spectra were obtained using a JASCO J-810 spectropolarimeter. Far UV CD spectra were recorded in a 0.02 cm path-length quartz cylindrical cell (Helma UK Ltd) using protein concentrations of 0.5 mg ml−1.

Acknowledgements

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

This work was supported by the Wellcome Trust, Action Research UK and the BBSRC. We are grateful to Dara Frank for the kind gifts of pUCP19exoUspcU and PA103ΔUΔT, and to Karen McCluskey for her help in protein structural determinations.

References

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