The Salmonella effector protein SigD is an inositol phosphate phosphatase that inhibits phosphatidylinositol 3-kinase-dependent signaling. Because epidermal growth factor (EGF) inhibits chloride secretion via phosphatidylinositol 3-kinase, we explored whether Salmonella infection might modify the inhibitory effect of EGF. As expected, EGF inhibited chloride secretion induced by carbachol in T84 epithelial cells. Infection with wild-type (WT) but not sigD− mutant S. typhimurium SL1344 decreased CCh-stimulated chloride secretion. Moreover, WT but not sigD− Salmonella reduced the inhibitory effect of EGF on carbachol-stimulated chloride secretion. Complementation of sigD restored the ability of mutant Salmonella to reverse the inhibitory effect of EGF. EGF-induced EGF receptor phosphorylation was similar in cells infected with either WT or mutant Salmonella, and neither WT nor sigD− Salmonella altered recruitment of the p85 subunit of phosphatidylinositol 3-kinase to EGF receptor, implying that SigD acts downstream of these signaling events. Furthermore, transepithelial resistance fell more rapidly in cells infected with WT vs. sigD− Salmonella, indicating an early role for SigD in reducing barrier function, perhaps via activation of protein kinase C. We conclude that the Salmonella bacterial effector protein SigD may play critical roles in the pathogenesis of disease caused by this microorganism.
- chloride secretion
- Salmonella typhimurium
- epidermal growth factor
the intestinal epithelium lines the entire gastrointestinal tract and has recently been established as not simply a physical barrier against microbial pathogens but rather as an active participant in the host response to infection via expression of a defined set of genes. Cross talk between the intestinal epithelium and bacterial pathogens is also presumed to contribute to diarrheal symptoms evoked by such organisms. However, detailed insights into the underlying mechanisms of diarrheal illness produced by non-enterotoxin-producing pathogens have remained elusive. Overall, the interactions that occur between pathogenic microorganisms and their host cells to produce such disease are complex and intimate.
Salmonella, a gram-negative bacterium that can cause gastroenteritis and enteric fever, is a significant cause of diarrheal disease worldwide. All Salmonella serotypes share the ability to invade the host by inducing their own uptake into intestinal epithelial cells. The uptake into these nonphagocytic cells is facilitated by virulence proteins delivered into the host cell cytoplasm by specialized type III protein secretion systems (TTSS) (11). The genes coding for subunits of TTSS are present in a number of distantly related gram-negative pathogens such as Yersinia, Shigella, Salmonella, and Pseudomonas. In Salmonella, two separate TTSS are encoded in genetic loci called Salmonella pathogenicity islands (SPIs). Whereas the SPI-2 TTSS is essential for the systemic part of the infection, the SPI-1 TTSS contains the genes required for invasion, manipulating cellular signal transduction pathways and the subsequent triggering of the epithelial inflammatory response, which culminate in diarrhea (25). However, effectors secreted by the SPI-1 TTSS are not necessarily encoded within the borders of SPI-1, and several are found elsewhere on the chromosome.
SigD (also called SopB), which is encoded in SPI-1, is one such effector protein secreted from Salmonella and has been characterized as an inositol phosphate phosphatase. It is capable of hydrolyzing the intracellular messenger inositol 1,3,4,5,6-P5 to inositol 1,4,5,6-P4 (21a). Previous studies (13) have shown that infection of T84 human intestinal epithelial cells with Salmonella, but not other invasive bacteria, leads to an increase in intracellular inositol 1,4,5,6-tetrakisphosphate (IP4). Furthermore, this inositol 1,4,5,6-P4 isoform of IP4 indirectly increases Cl− secretion by counteracting the inhibitory effect of epidermal growth factor (EGF) on basolateral K+ efflux (3), thus supporting increased Cl− secretion. The effect of EGF on secretion is mediated via the EGF receptor (EGFr) and consequent activation of the downstream effector phosphatidylinositol 3-kinase (PI3K). In turn, this enzyme activates phosphorylation of phosphatidylinositol 3,4-bisphosphate (PIP2) to phosphatidylinositol 3,4,5-trisphosphate (PIP3) (30). In addition to its effect on IP4, SigD has been shown to countermand PIP3 signaling by hydrolyzing it to PIP2 (19).
Salmonella mutants lacking sigD retain a normal ability to invade epithelial cells in vitro (30) but are attenuated in their ability to induce fluid secretion and neutrophil accumulation in a bovine ligated ileal loop model infected with Salmonella dublin (21a). We hypothesized that the attenuated diarrheal phenotype of sigD− mutants might result from their inability to countermand signals that normally limit Cl− secretion. Thus, by acting as an inositol phosphate phosphatase, we explored the possibility that SigD might reverse the inhibitory effect of EGF on Ca2+-dependent Cl− secretion. In the present study, we therefore sought to define cell physiological correlates of the known biochemical properties of this particular bacterial effector protein.
MATERIALS AND METHODS
Bisindolylmaleimide I (EMD Biosciences, La Jolla, CA); blocking reagent for Western blotting (skim milk), Rainbow colored protein molecular weight markers, ECL Plus (Amersham Pharmacia Biotech, Piscataway, NJ); BM chemiluminescence ELISA reagent (Boehringer Mannheim, Mannheim, Germany) carbachol, Phalloidin-FITC (Sigma-Aldrich, St. Louis, MO); anti-rat PI3K antibody (rabbit whole antiserum), anti-human EGFr antibody (mouse monoclonal IgG1) (Upstate Biotechnology, Lake Placid, NY); anti-zonula occludens (ZO)-1 (rabbit polyclonal IgG), Cy5 goat anti-rabbit IgG (Zymed Laboratories, San Francisco, CA); and Cy3 rabbit anti-goat IgG (Chemicon International, Temecula, CA) were obtained from the sources indicated. All other chemicals were of at least reagent grade and were obtained commercially.
Cell culture and infection.
The colonic epithelial cell line, T84, was cultured routinely in DMEM/Ham's F-12 medium (1:1) (JRH, Lenexa, KS) supplemented with 5% newborn calf serum (vol/vol) (HyClone, Logan, UT). For Ussing chamber/voltage clamp studies, ∼5 × 105 cells were seeded onto 12-mm Millicell-HA Transwells (Millipore, Bedford, MA). For experiments involving immunoprecipitation and Western blotting, ∼106 cells were seeded onto 30-mm Millicell-HA Transwells. Cells seeded onto Millicell filters were cultured for 10–15 days before use. Under these conditions, T84 cells develop a polarized phenotype characteristic of epithelial cells in vivo (12). The cell monolayers were infected apically for the times indicated with either wild-type Salmonella typhimurium SL1344, an isogenic mutant strain containing a deletion in sigD, or the sigD− mutant complemented with sigD provided on a plasmid (19). Bacteria were maintained using LB broth or LB agar plates. Before infection, bacteria were grown overnight without shaking in LB broth supplemented with 300 mM NaCl. After an optical density (at 600 nm) of ∼0.4 was reached, bacteria were washed twice in PBS and resuspended in Ringer solution before infection.
Colony-forming units and IL-8 determination.
T84 monolayers were grown to confluence and infected apically with bacteria at a multiplicity of infection (MOI) of 25 for 1 h to allow bacterial entry to occur. Extracellular bacteria were then removed by washing the monolayers, and gentamicin (50 μg/ml) was added to kill any remaining extracellular bacteria. Numbers of intracellular bacteria were measured by counting colony-forming units after lysing epithelial cells in 50% PBS and 0.1% Triton X-100 and plating serial dilutions of the lysate on LB agar plates for overnight culture.
IL-8 secretion was determined using T84 intestinal epithelial cells grown to confluence on permeable supports and infected with bacteria for 1 h as described above, followed by incubation of the cells for 8 h. Supernatants were harvested, and IL-8 was measured by ELISA as described previously (14).
Immunoprecipitation and Western blotting.
T84 cell monolayers were washed three times in Ringer solution and allowed to equilibrate for 30 min at 37°C. Cells were then infected (±inhibitors) as appropriate. Incubations were terminated by washing twice with ice-cold PBS. Ice-cold lysis buffer was added (consisting of 1% Triton X-100, 1 mM NaVO4, 1 μg/ml leupeptin, 1 μg/ml pepstatin, 1 μg/ml antipain, 1 mM NaF, 1 mM EDTA, and 100 μg/ml phenylmethylsulfonyl fluoride in PBS), and the cells were incubated at 4°C for 30 min. Cells were then scraped into microcentrifuge tubes and centrifuged at 10,000 rpm for 10 min, and the pellet was discarded. An aliquot was removed from each sample to determine protein content, and samples were adjusted so that they contained equal amounts of protein. For immunoprecipitation studies, monoclonal EGFr antibody (5 μg) was added to each sample and allowed to incubate, with shaking, for 60 min at 4°C, followed by addition of 40 μl protein A sepharose, and an additional incubation for 60 min at 4°C. After centrifugation, pellets were washed once with ice-cold lysis buffer, twice with ice-cold PBS, and then resuspended in gel loading buffer (50 mM Tris, pH 6.8, 2% SDS, 100 mM dithiothreitol, 0.2% bromophenol blue, 20% glycerol). The samples were boiled for 5 min, and proteins were separated by SDS-PAGE. Separated proteins were transferred onto polyvinylidene difluoride membrane (DuPont NEN) overnight at 4°C. The membrane was washed in 1% blocking buffer for 30 min, followed by incubation of the membrane with an appropriate dilution of primary antibody in 1% blocking buffer for 60 min. This was followed by washing three times in TBS with 1% Tween. After washes, a horseradish peroxidase-conjugated secondary antibody was added to the membrane in 1% blocking buffer and allowed to incubate for an additional 30 min. This was followed by a further washing three times in TBS with 1% Tween. Immunoreactive proteins were detected by an enhanced chemiluminescense detection kit and exposure of the membrane to X-ray film. Quantitation of protein phosphorylation was performed by densitometry using National Institutes of Health Image software.
Measurement of ion transport across T84 cells.
T84 cells were grown as monolayers on permeable filter supports as described above. After 10–15 days in culture, monolayers were infected (or treated appropriately as controls), mounted in Ussing chambers (aperture = 0.6 cm2), voltage-clamped to zero potential difference, and monitored for changes in short-circuit current, which in this cell line are wholly reflective of active chloride secretion (12). Short-circuit current measurements were carried out in Ringer solution containing (in mM) 140 Na+, 5.2 K+, 1.2 Ca2+, 0.8 Mg2+, 119.8 Cl−, 25 HCO3−, 2.4 H2PO4−, and 10 glucose. Results were normalized and expressed as μA/cm2.
Measurement of transepithelial resistance across T84 epithelial monolayers.
Monolayers of T84 cells were grown on permeable filter supports, as described above, and infected or treated appropriately as controls. Transepithelial resistance (TER) was monitored at various times thereafter with a voltohmeter.
Immunohistochemical analysis of T84 epithelial monolayers.
For immunohistochemical analysis, infected and control epithelial monolayers were fixed in 5% neutral buffered formalin and blocked (1% BSA, 0.2% blocking reagent, 0.3% Triton X-100). Sections were incubated respectively with antibodies, subsequently washed three times in PBS, and incubated with rabbit anti-goat IgG Cy3-, goat anti-rabbit IgG Cy5-conjugated secondary antibody, or phalloidin-FITC. Images were captured with a DeltaVision Restoration microscope system (Applied Precision, Issaquah, WA) attached to an inverted, wide-field fluorescent microscope (Nikon TE-200). Optical sections were acquired in 0.2-μm steps in the z-axis, and the resolution of ×40 images was enhanced using DeltaVision's deconvolution process and the algorithm by Agard et al. (1). Images were saved, processed, and analyzed using the DeltaVision software package softWoRx (version 2.50). Quantification of fluorescence was accomplished using only the linear range of the digital camera.
All data are expressed as means ± SE for a series of n experiments. Student's t-tests were used to compare paired data. One-way analysis of variance with a Student-Neuman-Keul posttest was used when three or more groups of data were compared. P values <0.05 were considered to be statistically significant.
Colony-forming units and IL-8 secretion are similar in cells infected with either wild-type or sigD-deficient S. typhimurium.
To confirm that a SigD-deficient mutant is able to infect and proliferate inside T84 intestinal epithelial cells to the same extent as wild-type bacteria, T84 cell monolayers were infected with wild-type or sigD− S. typhimurium for 1 h, washed and treated with gentamicin (50 μg/ml), and then incubated further for 3, 10, and 24 h. Determination of intracellular colony-forming units, shown in Fig. 1A, revealed that wild-type and sigD− Salmonella had a similar ability to invade and proliferate inside T84 cells. Because functional SigD has also been suggested to be involved in the accumulation of neutrophils in infected intestinal loops, we determined the secretion of the chemokine IL-8 as a representative of the host response. As shown in Fig. 1B, IL-8 secretion evoked by wild-type and sigD− mutant Salmonella was also similar.
Salmonella infection activates EGFr.
We next examined intracellular epithelial signal transduction pathways induced by Salmonella. Both strains, wild-type and sigD− mutant Salmonella, led to a rapid and transient tyrosine phosphorylation and therefore activation of the EGFr. Maximal activation was seen within 30 s to 1 min after initial infection (Fig. 2). These data suggested that early Salmonella-induced activation of EGFr is independent of the presence of SigD. Therefore, these data provided the basis to investigate the effect of SigD on signal transduction pathways downstream of the activation of EGFr itself. Western blotting experiments also showed that a 40-min period of Salmonella infection followed by a 15-min incubation with EGF did not modify EGF-induced EGFr activation, whether either wild-type or the sigD− mutant Salmonella were used (Fig. 3A). Finally, recruitment of the p85 subunit of PI3K to the EGFr was investigated under the same conditions. At this time point, neither EGF alone nor the addition of EGF to cells previously infected with wild-type or mutant Salmonella induced recruitment of p85 to the EGFr (Fig. 3B). Overall, the epithelial signaling events investigated here did not distinguish between the wild-type and sigD− Salmonella. This, therefore, allowed us to analyze possible effects of SigD on epithelial ion transport.
Effect of wild-type or sigD− S. typhimurium infection on chloride secretory responses.
Infection with wild-type S. typhimurium SL1344, but not the isogenic mutant strain containing a deletion in sigD, decreased CCh-stimulated Cl− secretion compared with the response in uninfected monolayers (Fig. 4A). Moreover, previous studies from our laboratory have shown that pretreatment of T84 intestinal epithelial cells with EGF inhibits Cl− secretion (29). Control experiments verified that EGF (100 ng/ml) inhibited Cl− secretion in response to the Ca2+-dependent agonist CCh (100 μM) by >50% (Fig. 4B). However, in cells infected with wild-type S. typhimurium, the inhibitory effect of EGF on CCh-stimulated secretion was reduced, whereas this reversal was not seen in cells infected with sigD− mutant S. typhimurium (Fig. 4B). In these experiments, cells were infected for 40 min, mounted in Ussing chambers, and then pretreated with EGF for 15 min before CCh stimulation. These time points were selected because peak IP4 production after Salmonella infection reportedly occurs between 40 and 60 min (13).
To verify that the failure of the sigD− mutant to abrogate the inhibitory effect of EGF on chloride secretion was indeed related to the loss of SigD, complementation experiments were carried out. By providing sigD on a plasmid, we were able to confer an ability to inhibit CCh-induced chloride secretion on the sigD-deficient S. typhimurium mutant strain (compare Fig. 5A with Fig. 4A). Moreover, as shown in Fig. 5B, complementation of the deleted sigD restored the ability of the mutant S. typhimurium to abrogate the inhibitory effect of EGF on CCh-stimulated chloride secretion to an extent similar to that of wild-type Salmonella.
Different pattern of changes in TER in cells infected with wild-type vs. sigD− S. typhimurium.
The pathogenesis of Salmonella-induced diarrhea likely also involves effects of the bacteria on epithelial barrier function. To explore a possible role for SigD in this response, T84 cells were infected with wild-type or mutant Salmonella, and effects on TER were noted. As shown in Fig. 6A, T84 cells infected with either wild-type S. typhimurium or the sigD− strain show different epithelial resistance responses. Although both wild-type and mutant Salmonella eventually lowered TER to a similar extent, the fall in resistance was significantly delayed in cells infected with Salmonella lacking SigD. Similarly, conductance measured in Ussing chambers after 60 min of infection was significantly higher in wild-type Salmonella-infected monolayers than in either uninfected monolayers or monolayers infected with sigD− mutant bacteria (Fig. 6B). These data imply a permissive role for SigD in the early changes in TER evoked by Salmonella infection.
Reduction in TER is induced through PKC and can be restored in sigD− mutants complemented with pACDE.
To verify that the reduction in TER was due to SigD, we compared the effects of wild-type bacteria, sigD− mutants, and the sigD− mutants complemented with SigD encoded on pACDE. Monolayers of T84 cells were again infected apically, and TER was measured. We focused on TER in infected cells at an early time point (40 min) and expressed TER as a percentage of starting values. Figure 7A illustrates the data obtained in these studies. At 40 min of infection, both wild-type and pACDE-complemented mutants reduced TER significantly compared with both control cells and cells infected with the sigD− mutants, verifying a role for SigD in the early reduction of TER.
Phosphatase activity can lead to PKC activation (22), which has been shown to play a role in the control of tight junctional proteins (8, 23, 27). We therefore investigated a role for PKC in the early SigD-dependent decrease in TER. Interestingly, as can be seen in Fig. 7B, when cell monolayers were pretreated with the PKC inhibitor bisindolylmaleimide (50 nM) before infection, the early decrease in TER induced by wild-type Salmonella infection was inhibited, suggesting a role for PKC in SigD-induced TER reduction. On the other hand, pretreatment of monolayers with bisindolylmaleimide did not alter the final nadir in TER values in cells infected with wild-type or sigD− Salmonella (data not shown), indicating a role for PKC in the SigD-induced, early phase of TER reduction but not in the later, more delayed effects of infection in these cells.
Immunolabeling of tight junctional proteins shows a more distinct temporary loss of actin structure and loss of colocalization of ZO-1 and occludin in monolayers infected with wild-type vs. sigD− mutant S. typhimurium.
To explore how the early differences in the effects of wild-type and sigD− infection on TER occur, a role for SigD in inducing changes in tight junctional protein structures was investigated using protein immunolabeling followed by deconvolution microscopy. Costaining for F-actin (green) and ZO-1 (red) (Fig. 8) showed that, in monolayers infected with wild-type bacteria for 20 min (Fig. 8A), the apical actin structure was temporarily lost in a large percentage of the cells with locally condensed accumulation, whereas the loss was more limited after infection with sigD− Salmonella (arrows). Interestingly, in both wild-type and sigD− mutant-infected monolayers, an elongation/distortion of cell shape was observed at this time point of infection (arrows). Furthermore, after 40 min of infection, the distortion of the cells declined, and the actin structure appeared to reorganize. Actin, however, was still locally accumulated mainly in cells infected with wild-type bacteria where this recovery also appeared only partial.
In similar experiments with immunohistological staining for ZO-1 (red) and occludin (green), another tight junctional protein difference was observed in monolayers in the early phase of infection with either wild-type or sigD− S. typhimurium. In control monolayers, there was colocalization of ZO-1 and occludin at the level of the tight junctions, whereas in cells infected with wild-type bacteria, occludin, but not ZO-1, was selectively lost from the junction after 40 min of infection. In contrast, in cells infected with the sigD− mutant bacteria, the loss of occludin from the junction was delayed (Fig. 9). Cells infected with wild-type bacteria also appeared to have a markedly irregular contour after 40 min of infection, an effect that was less marked in monolayers infected with the sigD− mutant. The differential effects of the wild-type and mutant bacteria on the cytoskeleton, as well as tight junction structure and biochemistry, may accordingly underlie the differential kinetics of their effects on barrier function.
Elucidation of the functional activities of a number of bacterial effector proteins has underscored the specificity of host-bacteria interactions. Translocation of bacterial proteins into the host cell that are able to interfere with the host cell's signal transduction pathways illustrate how bacterial pathogens manipulate normal cellular signaling responses, presumably to their own benefit. Our goal in the present work was to examine how secreted Salmonella effectors, and in particular SigD, might interdict to modify host cell signaling important in the maintenance of epithelial transport and barrier function.
In this study, deletion of SigD had no influence on the ability of Salmonella to invade T84 cells, as also found previously (30). SigD has been shown to participate in the bacteria-induced host membrane ruffling thought to be utilized in Salmonella invasion via its role as a phosphatase (28). However, as shown by Zhou et al. (30), the role of SigD in bacterial entry may be redundant with the roles of two other effector proteins, SopE and SopE2. Hence, if only one of these effector proteins is missing, the bacteria can still enter host cells with high efficiency. Our findings on invasion support this interpretation. Similarly, we also showed that the induction of IL-8 by wild-type and sigD− mutant Salmonella was similar, suggesting that other early cellular anti-infectious responses to both wild-type and mutant Salmonella might be comparable as well.
Wild-type and sigD− S. typhimurium initiated similar initial signaling responses in T84 cells, notably EGFr activation. Investigators at one of our laboratories have recently shown that certain effects of Salmonella on epithelial signaling require only bacterial attachment rather than invasion (5). Thus activation of EGFr induced by both wild-type and sigD− Salmonella may occur before any actual invasion. Moreover, although activation of the EGFr probably does not participate in Salmonella invasion (20), prompt phosphorylation of this basolaterally located receptor (24) after apical infection of a polarized monolayer is striking and deserving of further investigation (5). Studies on epithelial signaling in response to Salmonella infection, especially relating to the role of specific effector proteins, are still in relative infancy, but this rapidly moving field of investigation should ultimately shed light on pathophysiological mechanisms.
Although Salmonella induces activation of EGFr, the pattern and downstream effects differ from the mode of EGFr activation by EGF itself. This together with the SigD-induced actions on ion transport pathways induced by EGF is intriguing. Although we do not yet have a definite answer, our laboratory’s prior work with signaling through the EGFr leads us to believe that Salmonella may induce divergent dimerization of the EGFr than EGF itself. The EGFr (ErbB1) is part of the ErbB family, which consists of, thus far, four receptors (ErbB1–4) (6, 7, 26). On activation, and dependent on agonist, these receptors form dimers. Data suggest that EGF will induce dimerization of ErbB1 and ErbB2 (18), whereas, for instance, CCh will induce homorization between two ErbB1s (18), and the downstream effects of these activations differ accordingly. Without certainty, we speculate that Salmonella will form dimerization between ErbB receptors too. The actual EGFr activation does not, however, seem to interfere with the phosphatase activity induced by Salmonella through SigD. Thus, although Salmonella does activate the EGFr, this activation may induce different actions than what is observed with EGF.
Although wild-type and sigD− mutant Salmonella had similar properties with respect to invasion, initial host cell signaling, and IL-8 production, their effects on epithelial secretory responses were divergent. Thus wild-type Salmonella modified secretory responses and associated signaling, whereas sigD− mutant Salmonella did not. First, calcium-dependent Cl− secretion was decreased in wild-type but not sigD− Salmonella-infected cell monolayers, indicating a role for SigD in controlling secretory responses. Indeed, several inositol phosphates and phosphatidylinositol phosphates have been shown to interfere with cellular secretory pathways (3, 13, 17). This suggests that the ability of SigD to function as an inositol phosphate phosphatase renders the protein as a key mediator of Salmonella-induced alterations in epithelial secretory responses. Second, we observed that, although wild-type Salmonella with intact phosphatase activity was capable of abrogating the inhibitory effect of EGF on chloride secretion, the sigD− mutant was not. This finding strongly supports our initial hypothesis, which held that because SigD acts as inositol phosphate phosphatase and converts Ins(1,3,4,5,6)P5 to Ins(3,4,5,6)P4 (21a), it can thereby antagonize inhibitory actions of PIP3 on potassium efflux (3, 9, 17). Likewise, the ability of SigD to convert PIP3 to PIP2 (19, 21a) would also be predicted to abrogate the inhibitory effect of EGF on chloride secretion, which is known to depend on the enzyme responsible for PIP3 production, PI3K (29). Thus SigD induces diverse effects on chloride secretory responses, including inhibition of agonist-induced Cl− secretion, while at the same time antagonizing normal cellular inhibitory pathways that would otherwise limit secretion by themselves, perhaps via IP4 production and/or PIP3 metabolism (Fig. 10A). On the other hand, it is unlikely that wild-type Salmonella abrogates responses to EGF by virtue of prior activation of the receptor for this ligand, because this effect was induced equally by the wild-type bacteria and the sigD− mutant.
The ability of the wild-type bacteria to activate EGFr phosphorylation, although antagonizing effects of EGF-induced receptor activation, is on the surface somewhat paradoxical. However, we take these data to indicate that the specific effects of SigD in the latter regard occur downstream of the receptor and its immediate ability to recruit signaling intermediates. Furthermore, other ongoing studies from our laboratory imply that the effect of Salmonella on EGFr phosphorylation do not require bacterial invasion. Finally, our laboratory and others (4, 16, 18) have shown that the outcome of EGFr-mediated signaling may differ considerably depending on whether the receptor is activated by its ligand or secondary to intracellular signals, such as are likely to be used by the bacteria studied here. Any or all of these features may therefore resolve the seeming paradox. Our data from this study also elaborated a possible role for SigD in the early phase of changes in TER associated with Salmonella infection. Thus wild-type Salmonella, as well as the sigD pACDE mutant, reduced resistance in T84 epithelial monolayers more rapidly than the sigD− mutant Salmonella, even though the change in resistance was eventually similar. Correspondingly, epithelial conductance, as measured in Ussing chambers, differed between monolayers infected with wild-type Salmonella vs. sigD− mutant-infected monolayers or uninfected controls. Also, studies at the same time points showed that the SigD-dependent reduction in TER is likely mediated via PKC, although it does not appear to influence TER at later time points. This distinguishes the overall effect of Salmonella on epithelial barrier function from that evoked by another pathogen, enterohemorrhagic Escherichia coli (23).
SigD has been shown to mediate bacterial entry by stimulating membrane ruffling and actin cytoskeleton rearrangements (15). Although a loss in actin structure was also seen in sigD− mutant-infected monolayers, it was less extensive in these early time points, and this may contribute to differing effects of the bacterial strains on the initial drop in TER. Other studies have shown that SigD induces effects on the actin-organizing small GTP proteins, Cdc42, whereas another effector protein, SopE, induces effects on both Cdc42 and Rac (15). Thus a possible explanation for the delayed reduction in TER and the reduced effect on actin structure in epithelial monolayers infected with the sigD− mutant could be the lack of an early SigD-induced effect on Cdc42, although eventually this can be compensated for by SopE. This, together with the PKC-dependent early TER reduction we observed, may link the bacterial effector protein-induced changes in actin structure and the pathophysiological response of reduced TER. SigD is also known to play a role in bacterial-induced membrane ruffling (28), perhaps contributing to epithelial dysfunction even before invasion. Furthermore, the reduction in colocalization of occludin and ZO-1 was considerably less remarkable in sigD− mutant- than in wild type-infected monolayers, and a relation between PKC activation, altered occludin phosphorylation, and increased permeability have been suggested (10). Thus, although these data do not prove conclusively that SigD directly impairs colocalization of occludin and ZO-1 rather than influencing this association downstream from an effect on the actin cytoskeleton, the faster reduction in TER observed in wild-type vs. sigD− mutant-infected monolayers may be attributable to the deranged localization of these tight junctional proteins. Therefore, the early reduction in TER produced by SigD is dependent on 1) a possible direct action on actin filaments, 2) a disruption of the colocalization of occludin and ZO-1, and 3) possible mechanical rupture of tight junctions as reflected by the altered contour of individual cells in infected monolayers. On the other hand, the ability of wild-type Salmonella to activate EGFr is unlikely to be involved in reducing TER, because this effect was evoked in an equivalent manner by the SigD− mutant. We have also shown that an EGFr kinase inhibitor fails to prevent Salmonella-induced resistance changes (5).
In summary, we have presented evidence that specific cell physiological parameters of epithelial monolayers infected with S. typhimurium are affected by the bacterial effector protein SigD. The effects of SigD extend to both the regulation of Cl− secretory responses as well as the time course of changes in epithelial resistance. SigD may interfere with intestinal epithelial secretion via several mechanisms. First, its function as an inositol phosphate phosphatase may antagonize inhibitory pathways for epithelial secretion (Fig. 10A). Second, it apparently is required, at least initially, through a PKC-dependent mechanism, to alter the integrity of epithelial tight junctions through its actions on actin structure and ocludin/ZO-1 colocalization (Fig. 10B). The reversal of EGF-induced inhibition of Cl− secretion by the wild-type but not the sigD-deficient mutant also supports the hypothesis that Salmonella likely increases Cl− secretion in vivo by counteracting negative regulatory epithelial signaling pathways. Attenuation of signaling pathways that normally limit Cl− secretion may account, at least in part, for the diarrhea that so often accompanies Salmonella infection.
This work was supported by National Institutes of Health Grants DK-35108 Unit 5 and AT-01180 (to K. E. Barrett).
We thank Glenda Wheeler for assistance with manuscript submission.
This work was presented, in part, at the 2001 Annual Meeting of Experimental Biology and has been published in abstract form (FASEB J 15: 656.5, 2001).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society