The pathogenesis of enteropathogenic Escherichia coli (EPEC) is characterized by the type III secretion system-dependent exploitation of target cells that results in attaching and effacing (A/E) lesions, actin rearrangements and pedestal formation. This pathology is mediated by effector proteins which are translocated by the type III secretion system into the host cell such as the translocated intimin receptor (Tir) and several E. coli secreted proteins (Esp). Secretion of virulence proteins of EPEC is tightly regulated. In response to Ca2+, Esp secretion is drastically reduced, whereas secretion of Tir is increased. Membrane insertion of Tir, secreted under low Ca2+ conditions, is therefore independent of Esp. Furthermore, espB and espD mutant strains of EPEC, unable to form the translocation pore, still translocate Tir into host cells membranes. This autointegrated Tir is functional, as it is able to complement a tir mutant strain in recruiting actin to bacterial contact sites. The uptake of Tir into the host cell appears to depend on the C-terminal part of the protein, as deletion of this part of Tir prevents autointegration. Together, our results demonstrate that under conditions of limited Ca2+ an alternative mechanism for Tir integration can trigger the induction of A/E lesions.
Enteropathogenic Escherichia coli (EPEC) are a leading cause of persistent infantile diarrhoea particularly in developing countries (Kaper et al., 2004). The hallmark of EPEC pathogenesis is the induction of attaching and effacing (A/E) lesions on epithelial cells that is characterized by loss of microvilli followed by further exploitation of signal transduction pathways of the host cell. This leads to rearrangement of the cytoskeleton, formation of actin-rich pedestal-like structures, induction of apoptosis and a breach in the gastrointestinal barrier (Nataro and Kaper, 1998; Goosney et al., 2001). The factors necessary for induction of the A/E histopathology are encoded within a chromosomal pathogenicity island of about 35 kb called the ‘locus of enterocyte effacement’ (LEE) (Elliott et al., 1998). The LEE of EPEC prototype strain E2348/69 encodes a type III secretion system (T3SS) for the contact-dependent translocation of bacterial effector proteins into host cells (Hueck, 1998; Knutton et al., 1998). Translocated effector proteins of EPEC, which are directly associated with pathogenicity, are the E. coli secreted proteins (Esp): EspA, EspB, EspD, EspF, EspG, Map and the translocated intimin receptor (Tir). EspA forms a filamentous translocation tube on the bacterial surface connecting the bacterial and the target cell membrane (Knutton et al., 1998). EspB and EspD are inserted into the eukaryotic cell membrane and are believed to form a pore structure allowing the passage of effectors into the host cell (Wolff et al., 1998; Goosney et al., 2000, Ide et al., 2001).
Tir is translocated into the host cell, where it is phosphorylated and facilitates intimate attachment by serving as a membrane receptor for the LEE-encoded adhesin intimin (Hartland et al., 1999; Kenny, 1999). Membrane inserted Tir forms a ‘hairpin’ structure exhibiting the intimin-binding domain in the extracellular loop. The cytosolic amino- and carboxyterminal domains of Tir recruit cytoskeletal components underneath the adhering bacteria (Kenny et al., 1997a; Deibel et al., 1998) triggering a cascade of signalling events in the host leading to pedestal formation (Deibel et al., 1998; Devinney et al., 2001). However, the exact mechanism of Tir insertion into host cell membranes remains elusive. Tir secretion and translocation depends on its chaperone CesT, as a cesT mutant strain expresses and secretes significantly reduced levels of Tir and is unable to induce pedestal formation in host cells (Elliott et al., 1999; Delahay et al., 2002). Mutants lacking Tir or intimin are unable to attach efficiently and are greatly reduced in virulence, emphasizing the central role of these two bacterial proteins in EPEC pathogenesis (Kenny et al., 1997a). Previously, we identified a role for Ca2+ in expression and secretion of LEE-encoded proteins (Ide et al., 2003). Interestingly, Ca2+ depletion leads to a reduced secretion of Esp but in turn considerably enhances Tir secretion and translocation. This suggested that integration of Tir into host cell membranes might proceed by an Esp-independent alternative mechanism.
In this study, we analysed Esp-independent integration of Tir into host cells and identified a role for the C-terminal domain of Tir in uptake and insertion into host membranes. Furthermore, we show that autointegrated Tir serves as a functional receptor for intimin, as it complements a tir mutant strain in recruiting actin and tyrosin-phosphorylated proteins underneath bacterial contact sites. These results suggest that under limited free Ca2+ the autointegration of Tir might trigger the first steps of EPEC pathogenesis.
Esp-independent autointegration of Tir into host cell membranes
According to the current model of EPEC infection, Tir is translocated through the T3SS into the host cell membrane, where it serves as a receptor for intimin (Jerse et al., 1990; Donnenberg and Kaper, 1991). Although the exact integration mechanism of Tir is not known, studies show that early during infection Tir can first be detected in cytoplasmic membranes and only at later time points in the cytosol (Gauthier et al., 2000). Recently we suggested, that Tir secreted under Ca2+-depleted conditions and also recombinant TirHis integrates into host cell membranes (Ide et al., 2003).
To further analyse contact-independent delivery of Tir into host cell membranes, HeLa cells were incubated with recombinant Tir and subsequently separated into cytosolic fractions (CF) and membrane fractions (MF). After 15 min of incubation the recombinant Tir protein was detected in the CF (TirC), and after 30 min Tir appeared in the MF (TirM) (Fig. 1A). As during a natural infection EPEC interact with polarized epithelial cells with intact microvilli, we employed fully differentiated Caco-2 cells as a model for human enterocytes. Caco-2 cells were incubated with recombinant Tir for 30 min and CF and MF were analysed for Tir integration. As with the HeLa cells, Tir can be detected in CF as well as MF of Caco-2 cells with anti-Tir as well as anti-His antibodies, indicating that the type III secretion needle complex is not required for the integration of Tir into differentiated enterocytes (Fig. 1B). The possibility that TirC represents a contamination with TirM due to the cell fractionation procedure was excluded by analysing fractions for cytosolic protein kinase C (PKC). PKC was only detected in the CF but not in MF (Fig. 1C). These results were also confirmed by electron microscopy, where gold-labelled recombinant TirHis appears to be associated with vesicles in the cytosol after 15 min of incubation (Fig. 1D). Autointegration of recombinant TirHis into host cell membranes demonstrates that no additional bacterial factors such as Esp proteins are involved in this process. Furthermore, Tir is first taken up into the host cytosol and only subsequently inserts into the host cell membrane.
Integration of Tir into host cell membranes by EPEC espB and espD mutant strains
Autointegration of TirHis suggests that formation of the EspB/D pore complex is not needed for Tir uptake during infection. To analyse the role of EspB/D in Tir uptake, we infected HeLa cells with EPECΔespB (UMD864) or EPECΔespD (UMD870) for 5 h, fractionated the cells into CF and MF and analysed MF for Tir. Indeed, even after infection of eukaryotic cells with EPEC strains lacking EspB or EspD, Tir could be detected in MF of HeLa cells (Fig. 2). These data confirm the assumption that Tir is able to autointegrate into cell membranes independently of a pore structure formed by EspB/D.
The C-terminal domain of Tir mediates uptake of Tir into host cells
Previous studies identified a signal for Tir secretion and translocation residing in the N-terminal part of Tir (Abe et al., 1999). To localize domains within Tir that mediate the Esp-independent uptake of the receptor into eukaryotic cells, different N- or C-terminally deleted versions of recombinant His-tagged Tir were constructed (Fig. 3A). HeLa cells were incubated with these recombinant TirHis proteins for 30 min and CF and MF of HeLa cells were analysed for the presence of Tir. As shown in Fig. 3B, TirHis missing the N-terminal 220 amino acids can be detected in the cell cytosol and in MF. In contrast, Tir deleted of at least the C-terminal 160 amino acids is neither taken up nor integrated into the host membrane (Fig. 3B). Together these data indicate a central role of the C-terminus of Tir in Esp-independent integration of the receptor into host cells and this seems to be independent of the transmembrane domains of Tir.
Autointegrated Tir is a functional receptor for EPEC
Following Tir delivery into host cells, binding to intimin is essential for the formation of A/E lesions (Deibel et al., 1998; Kenny, 1999). To analyse whether Tir integrated into host cell membranes by an Esp-independent mechanism is functional as a receptor for EPEC, we incubated HeLa cells with Tir-containing bacteria-free supernatant derived from EPEC grown in Ca2+-free Dulbecco's modified Eagle's medium (DMEM). This should result in increased Tir and strongly reduced EspB/D secretion. Subsequently, the cells were infected with the tir mutant strain SE896. Strain SE896 is neither able to form pedestals nor to recruit phosphorylated proteins to bacterial contact sites (Fig. 4D–F) (Elliott et al., 1999). When host cells were incubated with Tir-containing supernatant of EPEC prior to infection, the tir mutant strain was able to accumulate actin and to recruit tyrosine-phosphorylated proteins to bacterial attachment sites (Fig. 4G–I). To exclude that residual EspB or EspD in the supernatant might facilitate Tir insertion, we incubated HeLa cells with recombinant HisTir protein prior to infection (Fig. 5) and with the espB and espD mutant strains of EPEC (data not shown). Tir-containing supernatant as well as recombinant Tir was able to complement the Tir deficiency of strain SE896 as shown by the induction of actin accumulation and recruitment of tyrosine-phosphorylated proteins (Figs 4 and 5). Strong signals for Tir could be detected underneath strain SE896 (Figs 4J–L and 5F), showing autointegrated Tir underneath contact sites of bacteria. The merged images demonstrate the colocalization of accumulated actin and tyrosine-phosphorylated proteins or Tir underneath attached bacteria of strain SE896 (Fig. 4C, I, L). Additionally, the tir mutant strain regains the ability to form pedestals when complemented with Tir-containing supernatant derived from EPEC E2348/69 or UMD896 (ΔespB) (Fig. 6), emphasizing that autointegrated Tir serves as a functional receptor for EPEC.
According to the current model of EPEC infections, Tir is translocated into host cell membranes via the T3SS by an EspB/EspD-dependent mechanism (Kenny, 1999; Hecht, 2001). Previous studies show that secretion of proteins by the T3SS is influenced by the Ca2+ concentration in the environment and that depletion of Ca2+ promotes Tir secretion in EPEC, but in contrast reduces secretion of other Esp (Straley et al., 1993; Kenny et al., 1997b; Ide et al., 2003). EspB and EspD are forming a pore structure in the eukaryotic cell membrane, which is thought to be absolutely necessary for allowing Tir to get access to the eukaryotic host cell.
In this study we demonstrate that Tir secreted under low Ca2+ conditions is able to integrate functionally into host cell membranes by an Esp-independent mechanism. The involvement of additional factors and a pore structure formed by EspB and EspD can definitely be excluded, as even isolated and purified recombinant Tir inserts into host cell membranes (Ide et al., 2003). These data are confirmed by the fact that Tir is detected in host cell membranes after infection with an EPEC strain lacking EspB or EspD. In some experiments using supernatants of ΔespB and ΔespD mutants the amount of Tir inserted by incubation with the ΔespD mutant appeared to be higher; however, this can be attributed to variations associated with the trichloracetic acid (TCA) precipitations of the secreted proteins. The exact mechanism of Tir insertion even through the T3SS is currently unknown, but Gauthier et al. detected Tir at early time points exclusively in MF and only at later time points in CF. This might potentially be due to a ‘membrane over-flow’ with Tir (Gauthier et al., 2000). Data from other studies support a different model, in which Tir is translocated into the cytoplasm of host cells via the T3SS and subsequently integrates rapidly into the membrane, as no cytosolic intermediate(s) of Tir can be detected (Crawford and Kaper, 2002). In this study, we could show that Tir, when delivered Esp-independently, can first be detected in the cytoplasm before it inserts into the cytoplasmic membrane of target cells at later time points. This is further supported by the detection of gold-labelled recombinant Tir in vesicle-like structures by electron microscopy at early time points. By analysing the fractions for cytosolic PKC, a contamination of cellular fractions due to the isolation procedure could be excluded. Whether these results are in accordance with the model proposed by Crawford and Kaper or whether under distinct environmental conditions Tir is delivered to the membrane by a different mechanism is currently not known and will be investigated further.
The mechanism by which Tir is able to penetrate and subsequently insert into the host membrane in an Esp-independent fashion remains controversial. Inhibition of actin polymerization by cytochalasin D had no effect on Tir uptake and integration (data not shown). In addition, gold-labelled Tir could not be detected in coated pits, but appeared to be associated with vesicles. Inhibition of host tyrosine kinases and of PI3-kinase activity also had no effect on Tir uptake (data not shown). Tir–intimin interaction modulates host cell kinase activity (Kenny, 1999; Celli et al., 2001); however, this activity does not appear to be necessary for membrane integration of Tir. Additionally, an incubation temperature of 4°C had no effect on autointegration of Tir into host membranes, excluding many phagocytotic uptake processes for Tir. Although some receptor–ligand interactions can occur at 4°C, subsequent events leading to engulfment are energy-dependent and responsive to temperature (Mukherjee and Maxfield, 2000). Our data support a model, where Tir is first integrated into the membrane from the cytosolic side before signal transduction pathways involving host kinases are induced in a co-ordinated fashion. It is already known that translocators of the T3SS are secreted prior to effectors, so that effectors will be translocated directly into the host cell (Cornelis, 2002; Deng et al., 2005). In this regard, Tir may be the first secreted protein necessary for close adherence of EPEC to their host cells under certain environmental conditions such as limited free Ca2+. How these processes are controlled, however, remains elusive.
The uptake of extracellular Tir without the need for (bacterial) accessory factors is reminiscent of proteins like TAT of HIV-1 or Antennapedia of Drosophila. These proteins contain a protein transduction domain (PTD) that enables them to traverse biological membranes in a temperature-, receptor- and endosome-independent way (Lindgren et al., 2000). However, this view has been challenged recently, and an endocytotic uptake can no longer be excluded for PTD-containing proteins (Green et al., 2003). A potential involvement of a putative PTD in Tir uptake needs to be analysed. Previous studies identified signals for efficient delivery and translocation of Tir in N-terminal residues 1–200 (Abe et al., 1999; Crawford and Kaper, 2002). We could show that the deletion of residues 1–233 of Tir prevents neither uptake nor integration of the receptor into host cells. The amount of ΔN-HisTir in cellular fractions is smaller compared with full-length HisTir. This might be due to misfolding or a lower stability of the truncated protein. In contrast, truncation of different C-terminal parts of Tir completely inhibits autointegration, although we cannot exclude that C-terminally deleted Tir might not fold correctly. However, we propose that residues 390–550 of Tir are involved in the Esp-independent uptake of the protein into host cells. The detailed molecular mechanism of Tir autointegration is currently under investigation.
Host cell membrane-inserted Tir serves as a receptor for intimin, resulting in actin accumulation and the induction of signal transduction pathways, leading to pedestal formation. Mutants of EPEC, defective in either Tir or intimin, fail to form pedestals (Kenny et al., 1997a). We show that autointegrated Tir is able to complement a tir mutant strain as a functional receptor for EPEC, leading to actin accumulation and tyrosine phosphorylation of host cell proteins. These findings support an alternative model for membrane insertion of Tir, as the current model requires EspA, EspB and EspD for Tir delivery. Interestingly, Rabinowitz et al. (1996) reported that a mutant strain of EPEC, which does not secrete EspB, induces Tir-dependent A/E lesions in host cells. This is in agreement with the alternative model for Esp-independent Tir integration proposed in this study. Additionally, the finding that mutant strains of EPEC lacking EspB or EspD are able to translocate Tir into host cell membranes confirms a pore-independent Tir integration. Lai et al. (1997) analysed MF after infection with the EspD mutant strain (UMD870) only for the phosphorylated form of Tir (Hp90) and were not able to detect Hp90 in the membrane. However, the finding that Tir seems not to be phosphorylated after infection with these mutant strains indicates that EspB and EspD may be necessary for Tir phosphorylation. This hypothesis is supported by the present study, as recombinant as well as secreted Tir are also not phosphorylated after autointegration. Only after infection with the tir mutant strain (SE896), accumulated actin and tyrosine-phosphorylated proteins are recruited underneath adherent bacteria. The fact that we were not able to detect phosphorylation of Tir after autointegration by Western blotting analysis might also be due to the lower amount of Esp-independently integrated Tir. Autointegration of Tir is clearly less effective under the experimental conditions described than integration via the T3SS. However, by using recombinant HisTir and Tir-containing supernatant we were able to complement the tir mutant strain in recruiting actin and tyrosin-phosphorylated proteins, confirming functionality of autointegrated Tir as a receptor for EPEC. This provides an additional possibility to ascertain Tir-intimin mediated adhesion in response to different environmental conditions during the infectious process.
The results of this study prompt us to suggest an alternative initial step in EPEC pathogenesis under conditions of limited Ca2+. As a first step, EPEC attach to microvilli via BFP. Second, limited Ca2+ or other environmental conditions induce enhanced secretion and autointegration of Tir into the host cell membrane, which is followed by intimate attachment mediated by Tir–intimin interaction. In a third step, the needle complex is formed after host cell contact and effector proteins are injected. This is followed by signal transduction events leading among other pathological effects to cytoskeletal rearrangements and pedestal formation.
Bacterial strains and tissue culture cell lines
For bacterial strains and plasmids used in this study see Table 1. HeLa cells were maintained in DMEM supplemented with 10% fetal calf serum, 1 mM glutamine and antibiotics at 37°C in a 10% CO2 atmosphere. Caco-2 cells were maintained in DMEM supplemented with 10% fetal calf serum, 10 µg ml−1 human transferrin and antibiotics and cultivated for 17 days to differentiate into a polarized monolayer. Medium was changed every 2–3 days. Ca2+ limitation was achieved by the addition of 5 mM EGTA.
Table 1. Escherichia coli strains and plasmids used in this study.
pET24b(+) carrying nucleotides 660–1650 of EPEC tir
pET24b(+) carrying nucleotides 1–1170 of EPEC tir
pET24b(+) carrying nucleotides 1–699 of EPEC tir
Analysis of EPEC proteins
For protein analysis, EPEC strains were grown as static cultures in DMEM with or without 5 mM EGTA for 1–6 h at 37°C and 10% CO2. The bacteria were centrifuged at 3000 g at 4°C for 15 min and the resulting supernatant twice at 17 000 g, 4°C for 15 min. For infection experiments, the supernatant was filtered using a 0.45 µm sterile filter (Schleicher and Schuell). For concentration of secreted proteins, TCA was added to a final concentration of 10% for at least 1 h on ice, centrifuged for 15 min, 4°C at 17 000 g, washed with cold acetone followed by suspending the protein pellet in SDS sample buffer. For protein analysis of whole bacteria, cells were centrifuged and bacteria were resuspended in SDS sample buffer. Proteins were separated by 12.5% SDS-PAGE and transferred to nitrocellulose for immunoblot analysis. Protein amounts were normalized according to equivalent units of optical density.
Plasmid constructions and protein purification
The strains and plasmids used in this study are listed in Table 1. The cloning of pTir2348His and the expression and purification of recombinant proteins have been described (Ide et al., 2003). For cloning of ΔN-HisTir220−550 the tir fragment was amplified by using oligonucleotides delntir(+) and HisTir(–), containing restriction sites for NheI and XhoI respectively. ΔC-HisTir and ΔC2-HisTir were constructed by using oligonucleotides HisTir2348(+) and delctir(–) or HisTir2348(+) and delc2tir(–). The 5′ primer (+) contains a BamHI site, whereas the 3′ primer (–) for both deletion constructs contains a XhoI site. After restriction, the tir fragments were ligated into pET24b(+) to create Tir fusions with a C-terminal His-tag. All oligonucleotides used in this study are listed in Table 2.
Table 2. Oligonucleotides used in this study.
Fractionation of HeLa and Caco-2 cells
Confluent HeLa cells grown in 100 mm dishes were incubated with EPEC supernatant or with recombinant His-tagged Tir (approximately 15–25 µg ml−1 final concentration). After incubation (15–30 min), cells were washed twice with cold phosphate-buffered saline (PBS) and resuspended in 1 ml of sonication buffer (50 mM Tris-HCl (pH 7.6), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 30% glycerol, 0.4 mM NaVO4, 1 mM NaF) supplemented with Complete EDTA-free Protease Inhibitor Cocktail (Roche Biochemicals) and sonicated four times for 1 s to permeabilize HeLa cell membranes. After centrifugation (30 min, 108 000 g, 4°C) the supernatant contained soluble cytoplasmic proteins (CF). The insoluble pellet was resuspended in sonification buffer complemented with 1% Triton-X-100, incubated with rotary shaking for 30 min at 4°C and centrifuged again for 30 min, 108 000 g at 4°C. The resulting supernatant containing Triton-X-100-soluble membrane proteins (MF) and the CF were precipitated with TCA as described before and analysed by SDS-PAGE and Western blotting. The protein concentrations of the samples were equalized. Data are representative of at least three independent experiments. For the fractionation of Caco-2 cells, post-confluent fully differentiated cells grown in 75 cm2 tissue culture flasks were incubated with recombinant His-tagged Tir for 30 min and fractionated and analysed as described above.
Infection of HeLa cells and immunofluorescence microscopy
For the analysis of Esp-independent functional Tir insertion, HeLa cells were incubated for 1 h with filter-sterilized supernatants of EPEC strain E2348/69 grown in the presence of 5 mM EGTA for 6 h, or with recombinant HisTir protein. Prior to the addition to HeLa cells, supernatants were supplemented with 10 mM CaCl2 to avoid detachment of cells from the coverslips. Subsequently, cells were infected with EPEC E2348/69 or SE896 bacteria that were diluted 1:20 in DMEM 3 h prior to infection. A bacteria–cell ratio of 100:1 for EPEC E2348/69 and of 200:1 (for strain SE896) was used to obtain a comparable attachment rate of wild-type and mutant strain for 2 h at 37°C in a 10% CO2 atmosphere. Subsequently, cells were gently washed twice with PBS and fixed in 4% paraformaldehyd (PFA) in PBS containing Ca2+ and Mg2+ for 15 min at room temperature (RT). To quench free aldehydes, cells were treated with 0.2% Glycin in PBS for 10 min at RT and then permeabilized for 5 min with 4% PFA containing 0.1% Triton-X-100 in PBS at RT. After blocking with 3% BSA in PBS for 30 min at RT, cells were incubated with a mouse monoclonal anti-phosphotyrosine-antibody (4G10, Upstate Biotechnology) diluted 1:200 in 0.3% BSA-PBS or with anti-Tir-antibody [clone B51, directed against Tir of Shiga-toxin producing E. coli strain 413/89-1 (Deibel et al., 1998)] at RT for 1 h. The cells were washed three times with PBS and treated with Cy2-labelled goat anti-mouse secondary antibody diluted 1:200 in 0.3% BSA-PBS (Dianova). Filamentous actin was stained with texas-red-conjugated phalloidin (1:100, Sigma-Aldrich), which was added during incubation with the secondary antibody. Adherent bacteria were stained with 300 nM DAPI (4,6-diamidino-2-phenylindole dihydrochloride, Sigma-Aldrich). After washing, coverslips were mounted in mounting medium (DAKO) as antifade agent and cells were inspected using a DM RXA fluorescence microscope (Leica). Confocal images were obtained using a LSM 510 confocal laser scanning microscope (Zeiss). Data are representative of at least four independent experiments.
Electron microscopy procedures
Gold-labelled recombinant Tir was prepared as described before (el Baya et al., 1997). Cells were incubated with recombinant gold-labelled Tir for 15 or 30 min at 37°C. For the detection of pedestal-like structures, HeLa cells were incubated with Tir-containing supernatants from EPEC E2348/69 or UMD864 for 30 min and then infected with strain SE896 as described above. Subsequently, HeLa cells were fixed with 2% glutaraldehyde for 24 h at 4°C and post-fixed with 1% osmium tetroxide. Cells were dehydrated through a graded ethanol series and embedded in epoxy resin. Ultrathin sections were prepared, stained with uranyl acetate and Reynold's lead citrate, and were examined with a Philipps 410 electron microscope. Data are representative of at least three independent experiments.
We are indebted to James B. Kaper and Michael Donnenberg (Baltimore) for providing strains, and to Frank Ebel (München) for providing anti-Tir antibody. This work was supported by Grants SFB293 B5 from the Deutsche Forschungsgemeinschaft and by a grant from the Bundesministerium für Bildung und Forschung (BMBF) Project Network of Competence Pathogenomics Alliance ‘Functional genomic research on enterohemorraghic, enteropathogenic and enteroaggregative Escherichia coli’, Project Group Schmidt/Karch, Universitätsklinikum Münster to M.A.S., and by an Innovative Medical Research grant (IMF:HE120201) of the Medical School of the University of Münster to G.H. This study is part of the PhD. thesis of S. Michgehl.