SEARCH

SEARCH BY CITATION

Summary

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

The human intestinal pathogen, enteropathogenic Escherichia coli (EPEC), causes diarrhoeal disease by a mechanism that is dependent on the injection of effector proteins into the host cell. One effector, EspF, is reported to be required for EPEC to disrupt tight junction integrity of intestinal cells and increase the paracellular movement of molecules, which is likely to  contribute  to  diarrhoea.  Here,  we  show  that  not one but three EPEC-encoded factors play important roles in this process. Thus, the Map (Mitochondria-associated protein) effector is shown to: (i) be as essential as EspF for disrupting intestinal barrier function, (ii) be able to function independently of EspF, (iii) alter tight junction structure and (iv) mediate these effects in the absence of mitochondrial targeting. Additionally, the outer membrane protein Intimin is shown to be crucial for EspF and Map to disrupt the intestinal barrier function. This function of Intimin is completely independent of its interaction with its known receptor Tir, revealing a physiologically relevant requirement for Intimin interaction with alternative receptor(s). This work demonstrates that EPEC uses multiple multifunctional proteins to elicit specific responses in intestinal cells and that EPEC can control the activity of its injected effector molecules from its extracellular location.


Introduction

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

Enteropathogenic Escherichia coli (EPEC) is a human intestinal pathogen that is a major cause of diarrhoea, particularly among infants in developing countries (Nataro and Kaper, 1998). EPEC attaches to gut epithelial cells where it induces an attaching and effacing (A/E) lesion resulting from the localized loss of microvilli (effacement) and formation of actin-rich pedestal-like structures beneath the adherent bacteria. EPEC and the closely related pathogen enterohaemorrhagic E. coli (EHEC) possess a homologous chromosomal DNA region called the LEE (Locus of Enterocyte Effacement), a 35 kb pathogenicity island that contains all the genes required for A/E lesion formation. Although the mechanism by which EPEC causes disease remains elusive, numerous factors involved in EPEC pathogenesis have now been identified and characterized (for recent reviews, see Kenny, 2002; Clarke et al., 2003). These include a type III secretion system that enables the bacterium to inject effector proteins such as Map (Kenny and Jepson, 2000), Tir (Kenny et al., 1997), EspF (McNamara and Donnenberg, 1998), EspG (Elliott et al., 2001) and EspH (Tu et al., 2003) directly into the host cell where they interfere with different aspects of host cell physiology.

Map (Mitochondrial-associated protein) carries a mitochondrial targeting sequence which can direct this protein to mitochondria where it interferes with their function (Kenny and Jepson, 2000). In addition, Map has been shown to promote filopodia formation in a Cdc42-dependent, but mitochondrial-independent manner (Kenny et al., 2002), and to mediate EPEC uptake into non-phagocytic cells (Jepson et al., 2003), revealing the multifunctional nature of this protein. In contrast, Tir (Translocated Intimin receptor) is inserted into the host plasma membrane where it interacts with the EPEC outer membrane protein Intimin (Kenny et al., 1997). Tir has also been shown to possess multiple functions including: (i) triggering cellular responses leading to actin nucleation resulting in pedestal formation, (ii) phosphorylation of the host protein, PLC-γ1, (iii) downregulation of Map-induced filopodia formation and (iv) a role in mediating uptake into non-phagocytic cells with most of these functions shown to be dependent on Tir interaction with Intimin (Kenny, 2002).

After EPEC entry into the gastrointestinal tract, the bacteria target epithelial cells within the small intestine. Epithelial cells possess apically located tight junctions that facilitate tight cell–cell contact and thus create an effective barrier to the paracellular movement of molecules. Tight junctions are dynamic structures composed of many different proteins, which enable regulation of this intestinal ‘barrier function’ (for a review, see Lapierre, 2000). Disruption of tight junction integrity is a mechanism believed to contribute to diarrhoea and, accordingly, is a common target of various enteric pathogens (Balkovetz and Katz, 2003). Intestinal epithelial cell lines grown on porous membranes provide a model to study epithelial properties such as barrier function. One method is to measure changes in transepithelial electrical resistance (TER), which has been shown to be altered after EPEC infection (Canil et al., 1993; Spitz et al., 1995; Philpott et al., 1996; Collington et al., 1998). Such EPEC-mediated losses in TER are accompanied by an altered distribution of the tight junction protein, occludin (Philpott et al., 1996) and an increase in the paracellular movement of specific markers (Spitz et al., 1995). This EPEC-mediated phenotype has been shown to be dependent on the type III delivery of the EspF effector protein as deletion of espF dramatically diminished the ability of EPEC to reduce TER (McNamara et al., 2001). However, it was noted that at high infection levels, the espF mutant induced a partial drop in TER, suggesting that other EPEC effectors may play a minor role (McNamara et al., 2001).

The goal of this article was to determine whether the Map and/or Tir effector molecules function along with EspF to cause maximal disruption of intestinal barrier function using polarized Caco-2 cells as a model intestinal epithelium. While a role for Tir was not observed for this EPEC-mediated phenotype, it was found that Map was as important as, if not more so than, EspF and this Map-mediated activity correlated with alterations in tight junction structure. Unexpectedly, Map and EspF's ability to disrupt intestinal barrier function, but not delivery into host cells, was found to be dependent on the expression of the bacterial outer membrane protein Intimin. Thus, this work revealed not only that the activity of three EPEC proteins is required to disrupt intestinal barrier function but that EPEC can control (directly or indirectly) the activity of effector molecules from its extracellular location through its outer membrane protein Intimin and independent of Intimin's primary receptor, Tir.

Results

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

EPEC-induced loss of TER is dependent on both Map and EspF

Given the multifunctional and redundant nature of effector molecules (Zhou, 2001; Kenny et al., 2002), we investigated whether Map and/or Tir act with EspF to cause maximal disruption of intestinal barrier function by monitoring changes in the electric resistance (TER) of polarized Caco-2 cells after infection with various EPEC strains. As previously reported with T84 cells (McNamara et al., 2001), the espF mutant induced only a relatively minor decrease in TER compared with wild-type EPEC (Fig. 1A). Thus, after a 6 h infection with wild-type EPEC, the TER decreased rapidly to 80.4% ± 10.7 (mean ± SEM) of uninfected cells. In contrast, the espF mutant triggered a more gradual and smaller decrease in TER resulting in a 29% ± 11.8 decrease at 6 h (mean ± SEM) compared with uninfected cells (Fig. 1A). Although a similar infection with the tir mutant resulted in an ≈ 80% decrease in TER (data not shown, but see Fig. 5A), ruling out a significant role for this effector molecule in this process, the map mutant, like espF, was strongly deficient in its ability to reduce TER. Thus, after 6 h of infection with map, the TER decreased by only 18% ± 3.2 (means ± SEM; Fig. 1A) with respect to uninfected cells.

image

Figure 1. EPEC infection of intestinal Caco-2 cells causes a loss in transepithelial electrical resistance (TER) that is dependent on both Map and EspF. Caco-2 cells were infected with wild-type (WT) EPEC; map, espF or the mapespF double mutant; map mutant with a plasmid carrying the map gene (map/pmap); espF mutant carrying T7-espF gene on a plasmid (espF/pespF  ). A. The effect on TER by the different strains is indicated in descending order in the legend. By 1 h post infection, there was a significant difference between infections with the map or espF mutants and WT (P < 0.04 for both mutants, Student's t-test), with this difference increasing over the 6 h infection process. In contrast, no significant difference is observed between uninfected and mapespF-infected cells (P = 0.345 at 6 h, Student's t-test). Complementation of map and espF mutants with plasmids expressing the missing gene product restored the ability to induce loss of TER to WT levels (P = 0.18 and 0.32, respectively, Student's t-test, 6 h post infection). Points represent means ± SEM from four independent experiments. B. Representative Western blot from three independent experiments of Triton X-100 soluble (membrane) and insoluble (containing adherent bacteria) fractions, isolated from Caco-2 cells infected for 4 h with the various EPEC strains and probed for EspF. The second slower migrating band evident only in the insoluble fraction presumably results from proteolytic breakdown. Note that the presence of the T7 tag alters the migration of the plasmid-expressed EspF protein. The presence of EspF in the insoluble, but not the membrane, fractions of cells infected with the type III secretion-defective strain cfm-14 demonstrates specific EspF delivery.

Download figure to PowerPoint

image

Figure 5. EPEC-mediated disruption of TER is independent of Tir, but not Intimin. Caco-2 cells were infected with wild-type (WT) EPEC, tir, intimin (eae) and the intimin mutant expressing Map from a plasmid (eae/pmap). A. TER measurements 6 h post infection revealed no significant difference between tir and WT EPEC-infected cells (P = 0.27, Student's t-test) in contrast to the intimin mutant, with or without the Map-expressing plasmid, which both induced increases in TER relative to uninfected cells. Points represent means ± SEM from three separate experiments. B. Representative Western blot of membrane fractions isolated from Caco-2 cells infected with the various strains and probed for EspF and Map. This reveals that Map and EspF are delivered into the host cell at comparable levels as WT-infected cells, with higher Map levels delivered after its expression from a plasmid in the intimin mutant (eae/pmap).

Download figure to PowerPoint

To exclude the possibility that the observed defects did not result from unwanted secondary mutations during the construction of the map and espF mutants, plasmids carrying the respective gene were introduced into the strains and used to infect Caco-2 cells. Figure 1A shows that plasmid expression of Map and EspF (as a T7-EspF fusion protein) in map and espF backgrounds, respectively, restored the ability of the strains to induce loss of TER to wild-type levels (Fig. 1A). Thus, these data confirm that the observed effects on TER result from the map and espF gene products alone and also showed that the T7-EspF fusion protein functionally substitutes for EspF.

As it was conceivable that deletion of map could affect EspF translocation into host cells, Western blot analysis was carried out on Triton X-100 soluble membrane and insoluble fractions derived from Caco-2 infected cells. This confirmed that EspF is still translocated into host cells by the map mutant and at levels comparable to that mediated by wild-type EPEC (Fig. 1B). It should be noted that plasmid expression of T7-EspF in the espF mutant increases the level of this protein delivered into the host cell (Fig. 1B). Specific delivery is demonstrated by the absence of EspF in the membrane, but not insoluble, fractions of cells infected with the cfm-14 secretion-defective mutant that is unable to deliver effector molecules into the host cell (Fig. 1B), and the absence of bacterial cytoplasmic protein, DnaK, in the membrane fractions (data not shown).

To investigate whether the deletion of both Map and EspF was sufficient to prevent EPEC's barrier-disrupting activities, a mapespF double mutant was generated and examined for its ability to alter TER. As shown in Fig. 1A, the double mutant failed to have any significant impact on TER compared with uninfected cells (P = 0.345, Student's t-test, 6 h post infection). It should be noted that wild-type EPEC, map, espF and mapespF strains grew at similar rates before infection and attached to the Caco-2 cells at comparable levels (results not shown). Therefore, our data reveal that Map and EspF are both essential factors required for EPEC to maximally perturb barrier function. The finding that both the map and espF mutants induce a small drop in TER (compared with wild type) is strongly indicative of synergism between these two effector molecules, as the TER value after wild-type infection is not due simply to the additive TER values of the single map and espF mutant infections (see Fig. 1A). Although these molecules appear to act synergistically, they retain some barrier disrupting activity in the absence of each other and thus display partial redundancy.

Map enables EPEC to cause maximal loss of TER independent of EspF and mitochondrial targeting

We next investigated whether increasing the expression levels of Map or EspF within EPEC in the absence of the other effector could restore the strain's ability to disrupt TER to levels observed with the wild-type strain. Thus, plasmids expressing Map or EspF were introduced into map, espF and the double (mapespF) mutant, generating map/pespF, espF/pmap, mapespF/pmap and mapespF/pespF. Infection of Caco-2 cells with these strains over a 5 h period revealed that plasmid expression of Map in either espF or mapespF backgrounds induced losses in TER similar to that mediated by wild-type EPEC (Fig. 2A). In contrast, plasmid expression of EspF in the map or mapespF mutants had little impact on TER (Fig. 2A) and even prevented the relatively minor decrease of TER associated with map infections (see Fig. 1A). Western blot analysis confirmed the presence of EspF at levels at least equivalent to that delivered by wild-type EPEC in host cells infected with strains expressing this protein, although it should be noted that plasmid expression of T7-EspF results in higher delivery levels of this protein which appears to hinder native EspF delivery (Fig. 2B, double arrow). Thus, these data suggest that Map, when expressed from a plasmid, can function independently of EspF to trigger maximal loss of TER whereas EspF requires the presence of Map for EPEC to disrupt TER.

image

Figure 2. Map can enable EPEC to induce dramatic loss in TER, in the absence of EspF. Caco-2 cells were infected with wild-type (WT) EPEC; single mutants deleted for map and espF carrying plasmids containing the T7-espF (map/pespF) or map (espF/pmap) genes, respectively; double mapespF mutant carrying a plasmid containing the T7-espF (mapespF/pespF) or map (mapespF/pmap) gene. A. Measurement of TER revealed no significant difference between mutants carrying the map plasmid (espF/pmap and mapespF/pmap) and WT EPEC (P = 0.48 and 0.4, respectively, 5 h post infection, Student's t-test). In contrast, a highly significant difference was detected between mutants carrying the espF plasmid (map/pespF and mapespF/pespF) and WT EPEC (P = 0.0001 for both strains). Points represent means ± SEM from four independent experiments. B. Representative Western blot from three independent experiments of Triton X-100 soluble ‘membrane’ fractions isolated from Caco-2 cells infected for 4 h with the various EPEC strains and probed for EspF. Arrowheads indicate the position of the delivered native EspF and T7-EspF fusion proteins in map/pespF-infected cells.

Download figure to PowerPoint

Map has been shown to possess at least two distinct functions in host cells that include induction of filopodia from a cytoplasmic location (Kenny et al., 2002) and targeting of mitochondria where it interferes with their function (Kenny and Jepson, 2000). To elucidate whether Map-mediated disruption of TER was dependent on mitochondrial targeting, a Tir_Map fusion protein (directed into host cells by the Tir domain) that fails to target mitochondria was expressed in the tirmap double mutant (preventing competition with Tir for host cell delivery) and this strain (tirmap/ptir_map) was used to infect Caco-2 cells. As shown in Fig. 3, this strain was as capable as wild-type EPEC in disrupting barrier function, although the kinetics of the TER decrease were slightly different. The apparent lag in TER decrease by tirmap/ptir_map probably results from the reduced growth rate of this strain compared with wild-type EPEC (data not shown) which is probably a consequence of the expression and delivery of this unnatural fusion protein into host cells.

image

Figure 3. Map targeting to mitochondria is not required for Map-mediated loss of TER. Caco-2 cells were infected with wild-type (WT) EPEC or a tirmap double mutant expressing a Tir_Map fusion protein (tirmap/ptir_map), which directs Map into host cytoplasm and abolishes mitochondrial targeting (Kenny et al., 2002). The tirmap mutant is used to avoid competition for translocation with the tir gene product, a protein shown to be unimportant for EPEC-mediated disruption of barrier function (see text and Fig. 5A). Although the decrease in TER at the early stages of infection caused by the tirmap/ptir_map strain was not as pronounced as that observed with WT EPEC, by 3 h post infection there was no significant difference between the TER values of WT and tirmap/ptir_map infections. It was noted that the growth rate of tirmap/ptir_map was notably slower than WT, probably resulting from the requirement to deliver an unnatural fusion protein, which may explain the difference between the two strains at early time points. Points represent the means ± SEM from four independent experiments.

Download figure to PowerPoint

Both Map and EspF contribute to EPEC-induced disruption of tight junctions

EPEC-mediated disruption of barrier function is linked to changes in tight junction structure. Given this, we infected Caco-2 cells with the various EPEC strains and examined the location of the tight junction protein, occludin. As previously reported, occludin forms a continuous perijunctional band in the apical region of uninfected cells (Simonovic et al., 2000) (Fig. 4A) and this distribution remained unaltered after a 5 h infection with the cfm-14 mutant (Fig. 4A), which is unable to deliver effector molecules into host cells. In contrast, cells infected with wild-type EPEC exhibited a fragmented peripheral pattern indicative of the loss of this protein from the tight junction. It should be noted that no observable accumulation of occludin in the cytoplasmic area was apparent in contrast to that reported with EPEC-infected T84 cells (Simonovic et al., 2000). Infection with the espF mutant did not induce any significant loss of occludin from the perijunctional region, as previously reported with T84 cells (McNamara et al., 2001), with identical patterns observed with the map and mapespF mutants (Fig. 4A). Plasmid expression of Map in the map, espF or mapespF mutants restored the ability to fragment occludin in a manner indistinguishable to wild-type infected cells (data not shown). This implies that Map, like EspF, participates in EPEC-mediated barrier dysfunction through disruption of tight junction structure.

image

Figure 4. Map and EspF function to enable EPEC to disrupt the perijunctional location of occludin in Caco-2 cells. Caco-2 cells were infected for 5 h with various EPEC strains and then either (A) fixed and stained for occludin or (B) cell membrane fractions were isolated and probed for occludin. A. While both uninfected cells and those infected with the type III secretion mutant cfm-14 displayed a complete peripheral pattern of occludin, infection with wild-type (WT) EPEC caused fragmentation of the occludin pattern. In contrast, infection with the map, espF and mapespF mutants, like the cfm-14 mutant, had no observable impact on the occludin staining pattern. It should be noted that EPEC infection leads to a type III secretion-dependent undulation of the cell surface, resulting in the need to show a composite of serial optical sections to reveal the occludin staining profile giving the image a ‘jagged’ effect. In contrast, the surface remains unaffected in uninfected or cfm-14 (secretion-defective strain) infected cells allowing the occludin pattern to be recorded in a single plane generating a sharper image.

Download figure to PowerPoint

Given the inability to detect EPEC-triggered relocalization of occludin to the cytoplasmic fraction using confocal microscopy, we used the alternative method of Western blot analysis to monitor such changes. This supported the microscopy data in that there was a strong correlation between loss of TER by the various EPEC strains and membrane levels of occludin (Fig. 4B). Thus, whereas similar levels of occludin were detected in the membrane fractions of uninfected cells and those cells infected with the map, espF, mapespF (Fig. 4B) and cfm-14 mutants (data not shown), dramatically reduced levels were apparent in the corresponding fractions of cells infected with wild-type EPEC, map/pmap (Fig. 4B), espF/pespF or mapespF/pmap mutants (data not shown). In contrast, no detectable differences in the level of occludin were observed in any of the insoluble or cytoplasmic fractions (data not shown).

Intimin-mediated signalling is essential for EPEC to induce loss of TER in Caco-2 cells

Although our findings revealed that the EPEC-mediated loss of TER does not require Tir (see above and Fig. 5A), it was surprising that an intimin mutant, which does not express the Intimin outer membrane protein (serves as a ligand for Tir), is documented to be unable to induce loss of TER (Canil et al., 1993; Philpott et al., 1996; Collington et al., 1998) or redistribute occludin (Simonovic et al., 2000). We first confirmed using Caco-2 cells that, as previously reported, the intimin (eae) mutant actually causes an increase in TER in contrast to wild type- or indeed tir-infected cells (Fig. 5A), and that plasmid expression of Intimin restored the barrier disrupting activity of the intimin mutant (data not shown). This finding implies that the intimin mutant either (i) cannot deliver Map and EspF into polarized cells or (ii) delivers these molecules but their barrier-disrupting activity is dependent on Intimin. To distinguish between these possibilities, membrane fractions from intimin-infected Caco-2 cells were probed for EspF and Map (Fig. 5B), revealing that both were translocated into host cells, at comparable levels to wild-type EPEC. Indeed, the cfm-14 mutant, which does not have a functional type III secretion apparatus, was unable to deliver EspF or Map into the membrane fraction and thus reveals specific effector delivery by the other strains. Thus, these results reveal that Intimin does not prevent Map or EspF delivery into the host cell and suggests therefore that Intimin is essential for Map and EspF to function within host cells to disrupt tight junctions, and that this function is independent of its interaction with Tir.

Intimin is essential for Map's barrier-disrupting activity but not its ability to induce filopodia

To support the Western blot data indicating that the intimin mutant can deliver effector molecules into host cells, we took advantage of the observation that Map delivery into HeLa cells leads to the transient expression of actin-rich filopodia whereas plasmid expression of Map within EPEC stimulates this process. The ability of EPEC to also induce filopodia formation on polarized Caco-2 cells was first assessed by infecting with map/pmap and indeed this resulted in the induction of prominent actin-rich filopodia (data not shown). Thus, Caco-2 cells were then infected with the intimin (eae) mutant, with or without the Map-expressing plasmid and TER was measured after a 6 h infection (Fig. 5A). In addition, duplicate infected cells were stained for polymerized actin at an earlier time point (when filopodia are evident, Fig. 6). Whereas plasmid expression of Map within the eae mutant had no effect on the ability of the mutant to decrease electrical resistance (Fig. 5A), it enabled the strain to trigger prominent filopodia at the site of infection (Fig. 6). In contrast, filopodia-like structures were almost completely absent from uninfected cells or cells infected with the map mutant (Fig. 6) or type III secretion-defective mutant cfm-14 (data not shown). Quantification of the number of filopodia evident on map- and eae/pmap-infected cells revealed a marked difference (eae/pmap = 10.8 ± 3.2; map = 0.5 ± 0.2; mean ± SE filopodia from 16 fields of view in four independent experiments), demonstrating that Map is delivered by eae/pmap and therefore Intimin controls the barrier-disrupting activity of Map (directly or indirectly), but not its filopodia-inducing activity. Thus, taken together, these results clearly reveal that Intimin plays a very important role in EPEC disruption of tight junctions and that this is independent of Map/EspF effector delivery or its interaction with Tir.

image

Figure 6. Intimin is required for Map to exhibit its barrier-disrupting activity, but not its filopodia-inducing function. Caco-2 cells were left uninfected or infected for 2 h with the intimin mutant carrying a Map-expressing plasmid (eae/pmap) before the cells were fixed and stained for polymerized actin (Red) and DNA (Blue). Infection with eae/pmap (middle) resulted in extensive filopodia (B) on the surface of the Caco-2 cells surrounding the bacterial microcolonies as evident from an aerial view (middle right) and in cross-sections of infected Caco-2 cells (middle left). In contrast, filopodia were not observed on uninfected cells (top) or cells infected with the map mutant (bottom). (A) Level of membrane support; (B) filopodia; (C) bacterial microcolonies on the surface of Caco-2 cells; (D) Caco-2 cell nuclei; (E) occludin staining.

Download figure to PowerPoint

Discussion

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

Under normal physiological conditions, the intestinal epithelium acts as a regulatory barrier to the movement of molecules. This intestinal ‘barrier function’ is maintained through the formation of intercellular tight junctions, which if disturbed can affect intestinal permeability and presumably contributes to diarrhoeal disease. It is well established that the diarrhoeal pathogen EPEC strongly diminishes the barrier function of in vitro grown epithelial cell lines (Canil et al., 1993; Spitz et al., 1995; Philpott et al., 1996; Collington et al., 1998) by a mechanism that is dependent on the delivery of the effector protein EspF into host cells via a type III secretion system (McNamara et al., 2001). The delivery of EspF into polarized T84 cells correlates with a decrease in TER and redistribution of the tight junction protein, occludin (McNamara et al., 2001).

In this article, we demonstrate that loss of barrier function is not as simple as previously inferred but in fact involves the coordinated action of at least three EPEC proteins. This finding stemmed from a search for the effector molecule(s) proposed to play an additional but minor role with EspF to enable EPEC to maximally disrupt barrier function (McNamara et al., 2001). Our assessment revealed that the Map effector protein is as equally important as EspF for this EPEC-mediated phenotype. Both the map and espF mutants induced only relatively small decreases in TER after a 6 h post-infection period (≈ 18% and 30% respectively) in contrast to wild-type EPEC which induced an ≈ 80% decrease. The fact that the map mutant consistently disrupted barrier function to a less degree than espF (although not to a statistically significant degree) suggests that it may play a more dominant role than EspF in this process, whereas the residual disrupting activity indicates that both effectors retain some function in the absence of the other. The validity of our data is supported by the findings that: (i) the map and espF defects could be restored by introducing plasmids expressing the missing gene product, (ii) the strains grew and adhered to host cells at comparable rates and (iii) a mapespF double mutant failed to induce any significant decreases in TER compared with uninfected cells.

Map's barrier-disrupting activity, like EspF, was shown to disrupt tight junction structure as espF, map and mapespF mutants all failed to alter the distribution of the tight junction protein occludin, unlike cells infected with wild-type EPEC or single mutants expressing the missing gene product from plasmids. Interestingly, our data strongly suggest that Map and EspF act synergistically to disrupt tight junctions, as in the absence of one another, each effector can only induce a minor loss in TER. In contrast, when expressed together (as in wild-type EPEC) their effect on both TER and tight junction alteration increases to a far greater extent than that caused by the individual single map and espF mutants.

The supposition that Map plays a more dominant role to EspF in mediating barrier dysfunction was supported by the finding that plasmid expression of Map in espF or mapespF backgrounds restored their ability to reduce TER to wild-type levels, unlike map or mapespF mutants expressing EspF from a plasmid which failed to disrupt barrier function. As plasmid expression increases the level of these molecules delivered into host cells, this suggests that higher cellular levels of Map either (i) alleviates its dependence on EspF function or less likely (ii) triggers alternative signalling pathway(s) that also happen to disrupt tight junction structure and barrier function. In contrast, EspF appears to require Map delivery into host cells to allow EPEC to effectively disrupt barrier function. However, rather puzzling, plasmid expression of EspF in a map mutant prevented it from triggering the small decrease in TER that is evident in the absence of the plasmid, although higher levels of EspF were detected in the infected host cells. This may suggest that the EspF-triggered cellular process(es) that mediate this low level barrier disruption are somehow sensitive to the intracellular levels of this effector molecule in the absence of Map. Alternatively, the ability of plasmid expression of one effector to interfere with the delivery of others (Figs 1 and 3) may reduce the level of the LEE-encoded Map protein to levels insufficient to trigger barrier dysfunction.

Map's critical role in barrier dysfunction extends the multifunctional repertoire of this effector molecule which has been shown to: (i) trigger Cdc42-dependent filopodia formation (Kenny et al., 2002), (ii) participate in Cdc42-dependent uptake of EPEC into non-phagocytic cells (Jepson et al., 2003) and (iii) target mitochondria to interfere with their function (Kenny and Jepson, 2000). While interference of mitochondrial function can affect ATP production to alter tight junction permeability (Mandel et al., 1993; Madsen et al., 1995; Gopalakrishnan et al., 2002), this mechanism is apparently not utilized by EPEC, as a Map molecule unable to target mitochondria restored the ability of the map mutant to disrupt barrier function. Although this molecule also enables the map mutant to trigger filopodia formation, this activity alone is unlikely to contribute directly to barrier dysfunction, as these two processes could be uncoupled (see below). Thus, Map appears to possess another activity, which works in a synergistic manner with EspF to disrupt barrier function.

Our finding that Tir plays no major role in barrier dysfunction was very important given the hitherto unexplained finding that an EPEC mutant defective in the expression of its outer membrane protein Intimin cannot disrupt barrier function (Canil et al., 1993; Philpott et al., 1996; Collington et al., 1998) or cause a redistribution of occludin (Simonovic et al., 2000). Our finding is highly significant, as the primary documented function of Intimin is to interact with the plasma membrane-located Tir molecule to mediate intimate adherence to the host cell and trigger various host cell signalling activities (Kenny, 2002). Rather surprisingly, the inability of the intimin mutant to disrupt barrier function has been overlooked in relation to EspF's essential role in this process. Importantly, our study shows that this defect is not at the level of effector delivery but rather that Intimin is essential for Map and EspF to exhibit their barrier-disrupting activities. Studies are currently underway to understand the mechanism by which Intimin influences the tight junction disrupting activities of EPEC. Interestingly, co-infection experiments with (i) map and espF mutants, (ii) intimin and cfm-14 mutants (secretion defective but expresses Intimin) or (iii) intimin and mapespF mutants all failed to significantly reduce the TER below that evident with the single mutants. This suggests that tight junction disruption by EPEC is likely to be a tightly coupled and temporally regulated process.

The ability of Intimin to apparently control the barrier-disrupting activities of Map and EspF independently of Tir clearly demonstrates that Intimin must mediate this activity through interaction(s) with other receptor(s). Indeed, if Intimin truly exists only as a bacterial surface protein, this implies that it must influence these Map and EspF functions via non-Tir receptors. Such an interaction would not be unprecedented, as Intimin interaction with non-Tir receptors has been indicated, although a physiological relevance for such interactions, until now, has not been established. Thus, previous work has identified cellular responses that are dependent on Intimin but not Tir (Frankel et al., 1996; Higgins et al., 1999; Phillips et al., 2000; Muza-Moons et al., 2003), whereas more recently Intimin from the closely related pathogen EHEC (serotype O157:H7) has been shown to bind the host receptor, nucleolin (Sinclair and O’Brien, 2002). Further studies are underway to determine whether nucleolin plays a role in the Intimin-dependent disruption of barrier function.

Recently, Muza-Moons et al. (2003) reported that EPEC-mediated disruption of barrier function in T84 cells was dependent on Tir, which contrasts with our results. However, this discrepancy may reflect the use of different cell lines and/or infection procedures, although our investigations into type III-dependent disruption of barrier function in T84 cells reveals that this process is dependent on the Map, EspF and Intimin, but not Tir, proteins (data not shown). It must also be noted that Muza-Moons et al. (2003) use a poorly characterized tir mutant and fail to demonstrate that the observed inability of this mutant to disrupt barrier function can be complemented by plasmid expression of Tir.

Several studies have associated EPEC's ability to disrupt barrier function with the phosphorylation of the 20 kDa myosin light chain (MLC20) (Yuhan et al., 1995; Savkovic et al., 2001; Zolotarevsky et al., 2002) although no work has linked EspF with this process. One explanation for this, as revealed by this study, may be the unforeseen complexity of EPEC signalling underlying this event, in addition to the partially redundant nature of Map and EspF in the process. However, having now unearthed the interdependent roles of these three EPEC factors, it should now be possible not only to characterize a link between MLC20 phosphorylation and Intimin/Map/EspF function but also to determine how Map, EspF and Intimin function together to alter barrier function. Preliminary confocal microscopy confirmed Map association with mitochondria and reveals partial colocalization of Map and EspF, although EspF is also evident in other, as yet, uncharacterized compartments. However, localization to such compartments does not appear to play an important role, as both Map (Fig. 4) and EspF (data not shown) retain their ability to disrupt barrier function when restricted to a cytoplasmic location.

A final important finding from this study is that EPEC can modulate, either directly or indirectly, a specific activity of a multifunctional effector molecule. Thus, Map's barrier-disrupting function but not its filopodia-inducing activity was shown to require the presence of Intimin, but not its known primary receptor Tir. This supports the proposition that EPEC can exert a controlling influence on its effector molecules from its extracellular location through the surface-located Intimin molecule (Kenny et al., 2002), and indeed extends the concept by showing that Intimin can influence effector function by both Tir-dependent (Kenny et al., 2002) and -independent mechanisms. Thus, this study reveals a complex interplay that exists between three EPEC LEE-encoded proteins to induce alterations in tight junction structure. Our results provide the opportunity to further decipher the molecular mechanisms that underlie EPEC pathogenesis and host cellular signalling pathways involved in tight junction physiology that are subverted after EPEC infection.

Experimental procedures

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

Bacterial strains

EPEC strains used in this study include wild-type EPEC E2348/69 (Levine et al., 1985), map (Kenny and Jepson, 2000), tir (Kenny et al., 1997), espF (Warawa et al., 1999), mapespF (this study; see below), cfm-14 (Donnenberg et al., 1990) and the intimin (eae) mutant CVD206 (Donnenberg and Kaper, 1991). All strains were streaked directly from frozen 10% glycerol stocks onto Luria–Bertani (LB) agar plates. Overnight cultures of E. coli were inoculated from single colonies into LB broth and left at 37°C without shaking. Strains containing plasmids derived from pACYC184 or pBluescript (pSK; Stratagene) were selected with chloramphenicol (25 µg ml−1 final concentration) and carbenicilin (100 µg ml−1 final concentration) respectively.

Generation of plasmid and strains

The generation of plasmids encoding Map (pACYC-map) and Tir_Map (pSK-tirΔMTSmap) have been described previously (Kenny and Jepson, 2000; Kenny et al., 2002). To generate pACYC-T7espF, the espF gene was PCR amplified using oligonucleotides that also introduced unique BamHI and SalI restriction sites (just upstream of the initiation codon and 884 bp downstream of the espF stop codon respectively), enabling its insertion into pbluescript (Stratagene) generating pSK-espF. This BamHI–SalI fragment was then used to replace a BamHI–SalI fragment carrying the espB gene in pACYC-T7espB (Markus Stein Thesis, University of Tubingen, Germany) to create an in frame T7-espF gene fusion.

The mapespF double mutant was generated using an espF mutant (Warawa et al., 1999) as a recipient for conjugation with SM10-λpir carrying the suicide vector pCVD422-Δmap (full) that carries the region upstream and downstream of map but has no map gene (Kenny and Jepson, 2000). Exchange of the map gene within the espF mutant for a fully deleted version was carried out as previously described for the generation of the map mutant (Kenny and Jepson, 2000).

Caco-2 cell culture and EPEC infection

Caco-2 cells (ATCC No. HTB-37) were cultured in tissue culture flasks at 37°C in a humidified atmosphere of 5% CO2. Cells were grown in Dulbecco's minimal Eagles Medium (DMEM) supplemented with 10% (v/v) heat-inactivated fetal calf serum, 2 mM l-glutamine and 1% non-essential amino acids.  Before  confluence,  Caco-2  cells  were  trypsinized  for ≈ 20 min and the resulting suspension was diluted accordingly (see below) in culture medium. This was added to plastic inserts containing a 1.1 cm2 membrane filter (0.4 µm pore size; Costar, Corning) to give ≈ 5 × 105 cells per filter. Monolayers were given fresh medium every 2 days and used in infection assays at 10–15 days post seeding. The TER at this time was typically 150–200 Ω·cm2, with the resistance of a blank filter taken as zero.

For infection of Caco-2 monolayers, the OD600 of overnight bacterial cultures was measured and the suspensions were diluted in DMEM (without supplements) to an OD600 of ≈ 0.03. Bacteria were grown in this medium (at 37°C in 5% CO2) for 3 h before being added to the apical surface of Caco-2 monolayers growing on the membrane filters (≈ 5 × 107 bacteria per well). Two hours before bacterial addition, the Caco-2 monolayers were washed in pre-warmed DMEM (without supplements) and left under normal culture conditions, to allow the TER to stabilize. TER was measured at 37°C over a 5–6 h post-infection period using an EVOM epithelial voltohmeter (World Precision Instruments). Bacterial adherence was determined post infection by washing infected cells three times in phosphate-buffered saline (PBS) to remove non-adherent bacteria, followed by lysis of cells with 1% Triton in PBS before plating serial dilutions on LB agar plates to enumerate the number of colony-forming units.

Immunofluorescence microscopy

Caco-2 monolayers grown on membrane filters were infected as  described  above.  At  various  time  points  post  infection, the cells were washed three times in PBS to remove non-adherent bacteria and then fixed for 20 min in 2.5% (w/v) paraformaldehyde at room temperature. Cells were then washed twice in PBS, permeabilized in 0.2% (v/v) Triton X-100 and incubated for 60 min in a staining solution containing 0.1 µM tetramethylrhodamine isothiocyanate (TRITC)-conjugated  phalloidin  (to  stain  polymerized  actin)  and an anti-occludin monoclonal antibody (Zymed Laboratories). The bound anti-occludin antibody was detected using a fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Jackson Laboratories). After staining, membrane filters were excised, mounted in Vectashield (Vector Laboratories), which contained DAPI to stain DNA, and sealed beneath a coverslip with nail varnish. Monolayers were viewed using a Leica TCS SP2 confocal microscope.

Isolation of cellular fractions

Caco-2 monolayers grown on membrane filters were infected with various EPEC strains and then fractionated as follows. Monolayers were chilled on ice and washed three times with cold PBS before permeabilization with 0.2% (v/v) saponin (Calbiochem) in PBS containing 1 mM phenylmethylsulphonyl fluoride. After a 5 min incubation at 4°C the cells were removed with a scraper, centrifuged (12 000 g; 5 min; 4°C) and the supernatant, containing the soluble ‘cytoplasmic’ components, was removed and kept on ice. The cell pellet was rinsed in PBS, resuspended in the saponin buffer containing 1% (v/v) Triton X-100, centrifuged (12 000 g; 2 min; 4°C) and the Triton X-100 soluble ‘membrane’ fraction was removed and kept on ice. The insoluble pellet was resuspended vigorously in PBS. The protein concentration of all samples was determined using Bradford's reagent (Sigma) and the samples were mixed with Laemmlli sample buffer (Laemmlli, 1970) and boiled for 5 min before Western analysis.

Western blot analysis

Twenty micrograms of each protein sample were loaded and resolved by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), before transfer to nitrocellulose and blocking in a 5% solution of dried skimmed milk. Membranes were then probed with antibodies directed against either Map (rabbit polyclonal, 1:400), EspF (rabbit polyclonal, 1:2000), occludin (mouse monoclonal, 1:2000; Jackson Laboratories) or DnaK (mouse monoclonal, Company, 1:2000). Primary antibodies were probed with peroxidase-conjugated (Jackson Laboratories) or alkaline phosphatase-conjugated (Zymed Laboratories) secondary antibodies and detected according to the manufacturers’ instructions by chemiluminescence substrates (Pierce) or alkaline phosphatase substrate (Promega) respectively.

Acknowledgements

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

This work was funded by a grant from the Wellcome Trust to B.K. as a Senior Fellow in Basic Biomedical Sciences. We would like to thank Sabine Quitard for technical assistance, Dr Roger James (Department of Surgery, Leicester University) for producing anti-Map monoclonal, anti-EspF and anti-Tir rabbit polyclonal antibodies. We also thank Dr Mark Jepson and Alan Leard (Cell Imaging Facility, Department of Biochemistry) for assistance with confocal microscopy and the MRC (UK) for supporting the School of Medical Science Cell Imaging Facility. Finally, we are grateful to Drs Marc Maresca, David Banbury (Department of Pathology and Microbiology) and Mark Jepson for providing constructive comments on the draft manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Balkovetz, D.F., and Katz, J. (2003) Bacterial invasion by a paracellular route: divide and conquer. Microbes Infect 5: 613619.
  • Canil, C., Rosenshine, I., Ruschkowski, S., Donnenberg, M.S., Kaper, J.B., and Finlay, B.B. (1993) Enteropathogenic Escherichia coli decreases the transepithelial electrical resistance of polarized epithelial monolayers. Infect Immun 61: 27552762.
  • Clarke, S.C., Haigh, R.D., Freestone, P.P.E., and Williams, P.H. (2003) Virulence of enteropathogenic Escherichia coli, a global pathogen. Clin Microbiol Rev 16: 365378.
  • Collington, G.K., Booth, I.W., Donnenberg, M.S., Kaper, J.B., and Knutton, S. (1998) Enteropathogenic Escherichia coli virulence genes encoding secreted signalling proteins are essential for modulation of Caco-2 cell electrolyte transport. Infect Immun 66: 60496053.
  • Donnenberg, M.S., and Kaper, J.B. (1991) Construction of an eae deletion mutant of enteropathogenic Escherichia coli by using a positive selection suicide vector. Infect Immun 59: 43104317.
  • Donnenberg, M.S., Calderwood, S.B., Donohuerolfe, A., Keusch, G.T., and Kaper, J.B. (1990) Construction and analysis of TnphoA mutants of enteropathogenic Escherichia coli unable to invade HEp-2 cells. Infect Immun 58: 15611571.
  • Elliott, S.J., Krejany, E.O., Mellies, J.L., Robins-Browne, R.M., Sasakawa, C., and Kaper, J.B. (2001) EspG, a novel type III system-secreted protein from enteropathogenic Escherichia coli with similarities to VirA of Shigella flexneri. Infect Immun 69: 40274033.
  • Frankel, G., Lider, O., Hershkoviz, R., Mould, A.P., Kachalsky, S.G., Candy, D.C.A., et al. (1996) The cell binding domain of intimin from enteropathogenic Escherichia coli binds to beta(1) integrins. J Biol Chem 271: 2035920364.
  • Gopalakrishnan, S., Dunn, K.W., and Marrs, J.A. (2002) Rac1, but not RhoA, signaling protects epithelial adherens junction assembly during ATP depletion. Am J Physiol Cell Physiol 283: C261C272.
  • Higgins, L.M., Frankel, G., Connerton, I., Goncalves, N.S., Dougan, G., and MacDonald, T.T. (1999) Role of bacterial intimin in colonic hyperplasia and inflammation. Science 285: 588591.
  • Jepson, M.A., Pellegrin, S., Peto, L., Banbury, D.N., Leard, A.D., Mellor, H., and Kenny, B. (2003) Synergistic roles for the  Map  and  Tir  effector  molecules  in  mediating  uptake of enteropathogenic Escherichia coli (EPEC) into non-phagocytic cells. Cell Microbiol 5: 773783.
  • Kenny, B. (2002) Enteropathogenic Escherichia coli (EPEC) – a crafty subversive little bug. Microbiology 148: 19671978.
  • Kenny, B., and Jepson, M. (2000) Targeting of an enteropathogenic Escherichia coli (EPEC) effector protein to host mitochondria. Cell Microbiol 2: 579590.
  • Kenny, B., DeVinney, R., Stein, M., Reinscheid, D.J., Frey, E.A., and Finlay, B.B. (1997) Enteropathogenic E. coli (EPEC) transfers its receptor for intimate adherence into mammalian cells. Cell 91: 511520.
  • Kenny, B., Ellis, S., Leard, A.D., Warawa, J., Mellor, H., and Jepson, M.A. (2002) Co-ordinate regulation of distinct host cell signalling pathways by multifunctional enteropathogenic Escherichia coli effector molecules. Mol Microbiol 44: 10951107.
  • Laemmlli, U. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680685.
  • Lapierre, L.A. (2000) The molecular structure of the tight junction. Adv Drug Deliv Rev 41: 255264.
  • Levine, M.M., Nataro, J.P., Karch, H., Baldini, M.M., Kaper, J.B., and Black, K.L. (1985) The diarrheal response of humans to some classical serotypes of enteropathogenic Escherichia coli is dependent on a plasmid encoding an enteroadhesiveness factor. J Infect Dis 152: 550559.
  • McNamara, B.P., and Donnenberg, M.S. (1998) A novel proline-rich protein, EspF, is secreted from enteropathogenic Escherichia coli via the type III export pathway. FEMS Microbiol Lett 166: 7178.
  • McNamara, B.P., Koutsouris, A., O'Connell, C.B., Nougayrede, J.P., Donnenberg, M.S., and Hecht, G. (2001) Translocated EspF protein from enteropathogenic Escherichia coli disrupts host intestinal barrier function. J Clin Invest 107: 621629.
  • Madsen, K.L., Yanchar, N.L., Sigalet, D.L., Reigel, T., and Fedorak, R.N. (1995) FK506 increases permeability in rat intestine by inhibiting mitochondrial function. Gastroenterology 109: 107114.
  • Mandel, L.J., Bacallo, R., and Zampighi, G. (1993) Uncoupling of the molecular fence and paracellular gate functions in epithelial tight junctions. Nature 361: 552555.
  • Muza-Moons, M.M., Koutsouris, A., and Hecht, G. (2003) Disruption of cell polarity by enteropathogenic Escherichia coli enables basolateral membrane proteins to migrate apically and to potentiate physiological consequences. Infect Immun 71: 70697078.
  • Nataro, J.P., and Kaper, J.B. (1998) Diarrheagenic Escherichia coli. Clin Microbiol Rev 11: 142201.
  • Phillips, A.D., Giron, J., Hicks, S., Dougan, G., and Frankel, G. (2000) Intimin from enteropathogenic Escherichia coli mediates remodelling of the eukaryotic cell surface. Microbiology 146: 13331344.
  • Philpott, D.J., McKay, D.M., Sherman, P.M., and Perdue, M.H. (1996) Infection of T84 cells with enteropathogenic Escherichia coli alters barrier and transport functions. Am J Physiol 33: G634G645.
  • Savkovic, S.D., Ramaswamy, A., Koutsouris, A., and Hecht, G. (2001) EPEC-activated ERK1/2 participate in inflammatory response but not tight junction barrier disruption. Am J Physiol 281: G890G898.
  • Simonovic, I., Rosenberg, J., Koutsouris, A., and Hecht, G. (2000) Enteropathogenic Escherichia coli dephosphorylates and dissociates occludin from intestinal epithelial tight junctions. Cell Microbiol 2: 305315.
  • Sinclair, J.F., and O'Brien, A.D. (2002) Cell surface-localized nucleolin is a eukaryotic receptor for the adhesin intimin-gamma of enterohemorrhagic Escherichia coli O157:H7. J Biol Chem 277: 28762885.
  • Spitz, J., Yuhan, R., Koutsouris, A., Blatt, C., Alverdy, J., and Hecht, G. (1995) Enteropathogenic Escherichia coli adherence to intestinal epithelial monolayers diminishes barrier function. Am J Physiol 31: G374G379.
  • Tu, X.L., Nisan, I., Yona, C., Hanski, E., and Rosenshine, I. (2003) EspH, a new cytoskeleton-modulating effector of enterohaemorrhagic and enteropathogenic Escherichia coli. Mol Microbiol 47: 595606.
  • Warawa, J., Finlay, B.B., and Kenny, B. (1999) Type III secretion-dependent hemolytic activity of enteropathogenic Escherichia coli. Infect Immun 67: 55385540.
  • Yuhan, R., Koutsouris, A., and Hecht, G. (1995) Enteropathogenic E. coli induced phosphorylation of myosin light chain increases intestinal epithelial paracellular permeability. Gastroenterology 108: A948A948.
  • Zhou, D.G. (2001) Collective efforts to modulate the host actin cytoskeleton by Salmonella type III-secreted effector proteins. Trends Microbiol 9: 567569.
  • Zolotarevsky, Y., Hecht, G., Koutsouris, A., Gonzalez, D.E., Quan, C., Tom, J., et al. (2002) A membrane-permeant peptide that inhibits MLC kinase restores barrier function in in vitro models of intestinal disease. Gastroenterology 123: 163172.