Salmonella typhimurium SipA-induced neutrophil transepithelial migration: involvement of a PKC-α-dependent signal transduction pathway

Milton Silva, Cecilia Song, William J. Nadeau, Jeffrey B. Matthews, Beth A. McCormick

Abstract

Salmonella typhimurium elicits an intense proinflammatory response characterized by movement of polymorphonuclear neutrophils (PMN) across the epithelial barrier to the intestinal lumen. We previously showed that S. typhimurium, via the type III secretion system effector protein SipA, initiates an ADP-ribosylation factor-6- and phospholipase D-dependent lipid-signaling cascade that directs activation of protein kinase C (PKC) and subsequent transepithelial movement of PMN. Here we sought to determine the specific PKC isoforms that are induced by the S. typhimurium effector SipA in model intestinal epithelia and to link the functional consequences of these isoforms in the promotion of PMN transepithelial migration. In vitro kinase PKC activation assays performed on polarized monolayers of T84 cells revealed that S. typhimurium and recombinant SipA induced activation of PKC-α, -δ, and -ε. To elucidate which of these isoforms play a key role in mediating epithelial cell responses that lead to the observed PMN transepithelial migration, we used a variety of PKC inhibitors with different isoform selectivity profiles. Inhibitors selective for PKC-α (Gö-6976 and 2,2′,3,3′,4,4′-hexahydroxyl-1,1′-biphenyl-6,6′-dimethanoldimethyl ether) markedly reduced S. typhimurium- and recombinant SipA-induced PMN transepithelial migration, whereas inhibitors to PKC-δ (rottlerin) or PKC-ε (V1-2) failed to exhibit a significant decrease in transepithelial movement of PMN. These results were confirmed biochemically and by immunofluorescence coupled to confocal microscopy. Our results are the first to show that the S. typhimurium effector protein SipA can activate multiple PKC isoforms, but only PKC-α is involved in the signal transduction cascade leading to PMN transepithelial migration.

  • inflammation
  • intestine
  • bacterial pathogenesis

the pathophysiology of localized enteritis caused by nontyphoidal Salmonella is characterized by movement of electrolytes and water as well as polymorphonuclear neutrophils (PMN) into the intestinal mucosa and lumen from the underlying microvasculature (5, 24, 34). The recruitment of inflammatory cells is considered to be a key virulence determinant underlying the development of S. typhimurium-elicited enteritis (22). Such PMN recruitment is coordinated by the epithelial release of an array of proinflammatory cytokines, including two potent PMN chemoattractants: IL-8 and pathogen-elicited epithelial chemoattractant (PEEC). IL-8 is secreted basolaterally by epithelial cells in response to the bacterial product flagellin. This signaling pathway involves basolateral Toll-like receptor-5, which increases IL-8 gene expression via NF-κB (8, 9, 11). The basolateral secretion of IL-8 establishes a stable haptotactic gradient across the lamina propria that guides PMN to the basolateral aspect of enterocytes (20, 21). Subsequent PMN transit through the epithelial monolayer to the luminal surface (defined as PMN transepithelial migration) is directed by PEEC, which is secreted apically in response to the Salmonella effector protein SipA (23). The interaction of transmigrating PMN with the epithelium and the subsequent loss of intestinal epithelial barrier function are thought to be key events in mediating the clinical manifestations of S. typhimurium-induced enteritis.

Modeling of this final, rate-limiting step of PMN movement in vitro revealed that addition of the S. typhimurium SPI-1 effector protein SipA to the apical aspect of human polarized epithelial cells is sufficient to elicit PMN transmigration (17). The significance of these results has been confirmed by the recent finding that SipA plays an important role in eliciting host secretory and inflammatory responses during S. typhimurium infection of calves, a relevant in vivo model system used to study human enterocolitis (43). We know that SipA appears to activate a novel signal transduction pathway involving an ADP-ribosylation factor 6 (ARF-6)- and phospholipase D-dependent lipid-signaling cascade that, in turn, activates protein kinase C (PKC) and, subsequently, causes the apical secretion of PEEC (3).

The PKC family of isozymes is composed of ≥13 different but structurally related serine/threonine kinases. These can be grouped according to their biochemical requirements for activation. Conventional PKC (cPKC) isozymes (α, βI, βII, and γ) are dependent on phosphatidylserine, diacylglycerol (DAG), and Ca2+, and they can also be activated by phorbol esters. Novel PKC (nPKC) isozymes (δ, ε, η, μ, and θ) are similar to the conventional isoforms in sensitivity to activators, but they contain full catalytic activity in the absence of Ca2+ (25, 26, 33). A third, more recently identified group includes PKC-τ, -λ, and PKC-ζ. These isoforms are structurally similar to cPKC and nPKC isoforms, but because they require neither DAG nor Ca2+, they are considered to be atypical PKC isoforms (25, 26, 33).

The cPKC and nPKC isoforms are stimulated when DAG is released, usually by the action of phospholipases on inositol phospholipids to yield DAG and inositol phosphates (36). These phospholipids are regulated by many growth factors and hormones; therefore, it is widely accepted that PKC isozymes play an important role in regulating proliferation and differentiation, as well as short-term cellular responses, such as secretion and ion flux. Another possible outcome of PKC activation is stimulation of signaling through the MAPK pathway (36).

The identification of multiple members of the PKC family has led to the speculation that individual isozyme responses play different roles in regulating distinct cell functions (25, 26, 33, 36). This speculation is based on evidence that 1) expression profiles of individual family members are extremely heterogeneous, depending on cell type and stimulus, 2) overexpression of specific isoforms, as well as isoform-specific inhibitors, has distinct effects on cell function, and 3) the substrate specificities of each isoform are very different, determined in large part by differences in subcellular localization of activated kinases, directed by isozyme-specific interactions with receptors for activated C kinase. Therefore, the specificity of a given stimulus and a subsequent response lie in the particular PKC isoform that is activated and the distinct localization of that isoform within the cell.

Because cell-specific expression and subcellular localization of individual PKC isoforms can indicate important isoform-specific functions and may reveal downstream events, we sought to determine the specific isoforms that are induced by the S. typhimurium effector SipA in model intestinal epithelia and also to link the functional consequences of these isoforms to the promotion of PMN transepithelial migration. Our findings are the first to demonstrate that the S. typhimurium effector protein SipA can activate multiple PKC isoforms, but only PKC-α is involved in the signal transduction cascade leading to PMN transepithelial migration.

MATERIALS AND METHODS

Cell culture.

T84 intestinal epithelial cells (passages 45–65) were grown in a 1:1 mixture of Dulbecco-Vogt modified Eagle's medium and Ham's F-12 medium supplemented with 15 mM HEPES buffer (pH 7.5), 14 mM NaHCO3, 40 mg/l penicillin, 8 mg/l ampicillin, 90 mg/l streptomycin, and 5% newborn calf serum. Polarized monolayers of T84 cells were formed and maintained on 0.33-cm2 ring-supported collagen-coated polycarbonate filters (Costar, Cambridge, MA) as previously described with recently detailed modifications (18). T84 cell monolayers reached a steady-state resistance 4–6 days after plating with some variability largely related to cell passage number. For clarity, we refer to this polycarbonate filter with the attached monolayer of T84 cells and matrix as “cell culture inserts.” Cell culture inserts were utilized for bacterial invasion and PMN transmigration assays 5–14 days after plating, as described previously (18). Cell culture inserts of inverted monolayers, used to study transmigration of PMN in the physiological basolateral-to-apical direction, were constructed as previously described (18, 27, 29).

Bacterial strains.

SL1344 is an invasive mouse-virulent S. typhimurium strain (40). Construction of mutant SL1344 derivatives EE633 (sipA::lacZY4) and VV341 (hilA::kan-339) has been previously described (13).

Growth of bacteria for assays using cell culture inserts.

Nonagitated microaerophilic bacterial cultures were prepared by inoculating 10 ml of Luria-Bertani broth (32) with 0.01 ml of a stationary-phase culture and incubating the culture overnight (∼18 h) at 37°C. Bacteria from such cultures were in the late logarithmic phase of growth and correlated with 5–7 × 108 colony-forming units/ml. Routinely, colony-forming units were determined by diluting and plating onto MacConkey agar medium (Difco) or L agar, as previously described in detail (20, 22). Ampicillin (50 μg/ml; Sigma) was added to bacterial culture media when necessary.

Purification of SipA-hemagglutinin fusion protein.

Crude lysate from Escherichia coli DH5α expressing the SipA-hemagglutinin (HA) recombinant fusion protein was precleared by passage through a 1-ml Sepharose column (Amersham Pharmacia) to remove any proteins that bind nonspecifically to Sepharose, as previously described (17). Briefly, the HA-affinity matrix (HA.11 affinity matrix; Convance, Berkeley, CA) was equilibrated at 4°C with buffer C (200 mM MES-HEPES, pH 6.2, 0.1 mM MgCl2, and 0.1 mM EDTA containing 0.05% Tween 20 and 0.5 M NaCl) before the addition of the precleared E. coli lysate. The HA affinity matrix was mixed with the E. coli lysate for 18 h at 4°C with constant end-over-end shaking. The unbound lysate material was washed free from the column with 40 ml of buffer C. To elute the bound protein, the column resin was mixed with 1 mg of HA peptide dissolved in 2.5 ml of column buffer, and the resin-peptide mixture was warmed to 30°C for 20 min. The collected fraction was buffer exchanged with Hanks' balanced saline solution and analyzed by SDS-PAGE and Western blotting to verify protein purity.

In vitro kinase assays.

Confluent T84 monolayers grown on 4.7-cm2 permeable supports were treated with SipA (40 μg/ml) and washed twice with cold PBS. Proteins were extracted by 30 min of incubation on ice with 500 μl of apical lysis buffer containing 50 mM Tris·HCl, pH 7.5, 150 mM NaCl, 1% (vol/vol) Triton X-100, 2 mM EDTA, 1 mM EGTA, 30 mM sodium pyrophosphate, 50 mM NaF, 100 μM Na3VO4, and complete protease inhibitor cocktail tablets (Boehringer-Mannheim) (37). The protein concentration of each sample was measured and adjusted to contain 500 μg in 400 μl of apical lysis buffer. Polyclonal antibodies (Biomol, Plymouth Meeting, PA) against cPKC (2 μg) or nPKC (4 μg) were added to each sample for overnight rotation at 4°C. After incubation, the immune complexes were precipitated using protein A-agarose beads, washed, resuspended in 20 μl of kinase buffer consisting of 35 mM Tris·HCl, pH 7.5, 10 mM MgCl2, 0.5 mM EGTA, 10 μCi [γ-32P]ATP (NEN), 60 μM cold ATP, and 1 mM Na3VO4, and incubated with 10 μg of myelin basic protein as a substrate at 30°C for 30 min. After incubation, the reaction was terminated by the addition of 5× Laemmli sample buffer, and the samples were boiled for 5 min. The supernatants were subjected to SDS-PAGE (12% gels), and the gel was dried and subjected to autoradiography.

PKC translocation assay.

Movement of PKC to a detergent-insoluble membrane fraction was assayed as described by Ferro et al. (6). All supernatants were concentrated over a 10-kDa molecular mass cutoff filter (Centricon-10; Amicon). In subsequent experiments, the intermediate (buffer B) step of the protocol of Ferro et al. was omitted, thereby yielding cytosolic and membrane subcellular fractions. Fifty micrograms of each fraction were separated on 10% polyacrylamide gels, transferred to nitrocellulose, and immunoblotted for PKC isoforms (Upstate Biotechnology) and then for horseradish peroxidase-conjugated goat anti-rabbit (AP Biotech). The PKC inhibitors Gö-6976, 2,2′,3,3′,4,4′-hexahydroxy-1,1′-biphenyl-6,6′-dimethanoldimethyl ether (HBDDE), rottlerin, V1-2, and C2-4 were obtained from Biomol. T84 cell monolayers were pretreated for 1 h at 37°C before the addition of rSipA, wild-type S. typhimurium, or the buffer control. Treatment of the polarized monolayer with PKC inhibitors did not affect the relative distribution of protein between cytosolic and membrane fractions. The antibody for the phosphorylated form of PKC-α was purchased from Upstate Biotechnology.

Because of the day-to-day differences, basal values of membrane-associated PKC were found to vary. The results are representative immunoblots from at least three independent experiments. The relative amount of PKC translocated in each instance is compared with the uninfected control treatment condition.

PMN transepithelial migration assay.

Cell culture inserts of inverted T84 monolayers were used for the physiologically directed (basolateral-to-apical) PMN transepithelial migration assay as previously described (20). Human PMN were isolated from normal volunteers as described elsewhere (12). Transmigration was quantified by assaying for the PMN azurophilic granule marker myeloperoxidase as described previously (29, 30). After each transmigration assay, nonadherent PMN were extensively washed from the surface of the cell culture inserts, and PMN cell equivalents, estimated from a standard curve, were assessed as the number of PMN associated with the cell culture inserts and the number that had completely traversed the cell culture insert (i.e., into the basolateral reservoir). In a subset of experiments, rSipA and S. typhimurium induction of PMN transepithelial migration was performed in the presence of PKC inhibitors. For these studies, T84 cell monolayers were pretreated with the inhibitor for 1 h at 37°C before the addition of rSipA, wild-type S. typhimurium, or the buffer control.

Confocal microscopy.

Monolayers grown on 0.33-cm2 permeable supports were treated with rSipA (40 μg/ml) and washed three times with cold PBS. Cells were then fixed in 4% paraformaldehyde for 1 h at room temperature, washed twice with PBS, permeabilized with 0.1% (vol/vol) Triton X-100 in PBS for 7 min, and rinsed twice with PBS (37). Filter membranes were cut out in a rectangular shape from the Transwell support, placed between 50 μl of blocking buffer (1% normal goat serum and 3% BSA in PBS) at the top and bottom of the monolayers, and incubated for 30 min at room temperature. Polyclonal antibodies against PKC-α were diluted to 10 μg/ml in blocking buffer containing 0.1% Triton X-100, and 50 μl of the antibody were placed at the top and bottom of the monolayers. After overnight incubation in a moist chamber at 4°C, monolayers were washed three times in PBS for 10 min and incubated in rhodamine-conjugated goat anti-rabbit polyclonal IgG (1:100 dilution) for 1 h at room temperature along with FITC-phalloidin for F-actin staining. Monolayers were then washed three times in PBS and mounted on a microscope slide with Vectashield mounting medium (Vector Laboratories). Confocal images were acquired using a Zeiss inverted microscope equipped with MRC-1024 and Lasersharp software (Bio-Rad). Secondary antibodies conjugated to different fluorescent dyes were obtained from Jackson Laboratories.

Data presentation.

Because of day-to-day and passage-to-passage variability in transepithelial resistance between groups of monolayers (baseline resistance = 650–1,500 Ω·cm2) and in PMN obtained from different donors, individual experiments were performed using large numbers of monolayers and PMN from single blood donors on individual days. PMN isolation was restricted to 10 different donors (repetitive donations) over the course of these studies. S. typhimurium invasion and myeloperoxidase assay data were compared by Student's t-test. PMN transmigration results are represented as PMN cell equivalents derived from a daily standard PMN dilution curve. PMN that completely traversed the monolayer are represented as the number of PMN (cell equivalents/ml in a total volume of 1 ml). Values are means ± SD of an individual experiment done in triplicate and repeated at least three times.

RESULTS

SipA regulates translocation of PKC to a membrane fraction.

Our previous findings revealed that S. typhimurium, via the effector SipA, activates a PKC-dependent signal transduction pathway that orchestrates transepithelial PMN movement (3, 17). Because most PKC isoforms are phosphorylated and shift from host cell cytosol to the membrane on activation, we initially examined whether subcellular distribution of PKC to the membrane occurred during apical invasion of S. typhimurium into polarized T84 cell monolayers. We found that wild-type S. typhimurium promoted a significant increase in PKC in the Triton X-100-insoluble membrane preparation compared with the uninfected buffer control (Fig. 1). Furthermore, direct addition of purified rSipA to the apical surface of T84 cell monolayers elicited a marked rise of PKC in the Triton X-100-insoluble membrane fraction, with potency similar to that of wild-type S. typhimurium. Such PKC activation appeared to be dependent on SipA, because the SipA mutant strain EE633 failed to induce an increase in PKC in the subcellular membrane fraction (Fig. 1). This SipA mutant strain adhered to and entered the apical surface of T84 cell polarized monolayers as efficiently as its wild-type parent (data not shown) (17). To control for the increase in PKC movement into the Triton X-100-insoluble membrane fraction, we stimulated the T84 cell monolayers with the phorbol ester PMA, a direct nonphysiological activator of PKC. These results suggest that purified SipA is able to recapitulate signaling pathways that connect wild-type S. typhimurium to PKC.

Fig. 1.

Triton X-100 membrane-insoluble fraction exhibits an increase in PKC in response to apical addition of Salmonella typhimurium or recombinant SipA (rSipA). T84 cells were infected apically with wild-type SL1344 (WT), rSipA (40 μg/ml), or the SipA mutant strain EE633 for 1 h and then lysed and enriched for Triton X-100-insoluble membranes. T84 cells incubated with only Hanks' balanced saline solution served as the negative control (buffer); stimulation of T84 cell monolayers with PMA (1 μg/ml) represented the positive control. Protein (50 μg from each fraction) was separated on 10% acrylamide gels and immunoblotted for conventional PKC (cPKC) isoforms using a pan-cPKC antibody diluted 1:1,000 (Upstate Biotechnology). Data are from a single experiment and are representative of ≥3 experiments that showed an identical pattern of results.

We next sought to elucidate the PKC isoform involved in regulation of the signal cascade that mediates PMN transepithelial migration. For these studies, we used rSipA, rather than wild-type S. typhimurium, because this allowed us to determine the specific PKC activation patterns induced by SipA without interference from other effector proteins delivered by Salmonella type III secretion systems or other virulence determinants. Therefore, using an in vitro PKC kinase assay, we first examined what PKC isoforms became activated on rSipA exposure to the apical surface of T84 cell polarized monolayers. Of the 13 known PKC isoforms we examined by Western blot, only 5 (α, βII, δ, ε, and ζ) are shown to be expressed in T84 cells. Because PKC-βII is constitutively active in T84 cells, only the three major PKC isoforms (α, δ, and ε), which were responsive to the phorbol ester PKC agonist PMA, were examined on the addition of rSipA. We found that rSipA specifically induced the kinase activity of the cPKC isoform PKC-α, in addition to the nPKC isoforms PKC-δ and PKC-ε. Furthermore, such activation occurred as early as 15 min after rSipA exposure (Fig. 2A) and was found to correlate with the translocation of PKC to the membrane after apical exposure to S. typhimurium or rSipA (data not shown). Additionally, densitometric analysis revealed that PKC-α was the isoform most strongly activated (Fig. 2B) early after SipA stimulation (15 min), whereas PKC-δ and -ε activation was maximal 30 min after stimulation. Also, PKC-α activation in response to rSipA stimulation was biphasic, because there was a statistically significant transient decrease in activation after 30 min. Taken together, these data suggest that apical exposure of rSipA increases the kinase activity of PKC-α, -δ, and -ε, with PKC-α being the most potently activated early after stimulation.

Fig. 2.

Induction of PKC isoforms by rSipA. A: monolayers were subjected to an in vitro kinase assay to determine which PKC isoforms are activated in T84 cells in response to rSipA. Monolayers were stimulated with rSipA for 15, 30, or 45 min and immunoprecipitated with isoform-specific antibodies to assess their kinase activities on myelin basic protein blots. B: densitometric analysis of the kinase activation experiment in A. *P < 0.005; †P < 0.05 (Student's t-test) compared with time 0. Data are representative of ≥3 experiments.

PKC-α is required for S. typhimurium-induced PMN transepithelial migration.

To elucidate which of the activated PKC isoforms (α, δ, or ε) was directly involved in the regulation of S. typhimurium SipA-induced PMN transepithelial migration, we first examined the effect of structurally unrelated PKC inhibitors on the ability of rSipA to induce PMN transepithelial migration. We initially assessed whether the PKC-α inhibitors Gö-6976 and HBDDE adversely influenced the ability of rSipA to induce PMN transepithelial migration. Gö-6976 has been shown to inhibit conventional PKC isoforms exclusively [IC50 = 2 nM against PKC-α in vitro (19)], with no demonstrable in vitro inhibitory activity against Ca2+-independent atypical PKC isoforms, even at high micromolar concentrations. This inhibitor also has no effect on δ-, ε-, or ζ-isoforms (19). Furthermore, HBDDE inhibits PKC-α (IC50 = 43 μM) without inhibiting PKC-δ or PKC-βI and -βII (15). Using these inhibitors at optimal inhibitory concentrations, we found that pretreatment of T84 cell monolayers for 1 h with 5 μM Gö-6976 or 5 μM HBDDE significantly reduced the ability of S. typhimurium to induce PMN transepithelial migration by ∼80 and 85%, respectively (Fig. 3). We next examined the PKC-δ inhibitor rottlerin for its ability to inhibit rSipA-induced PMN transepithelial migration. Although rottlerin has been shown to be specific for the nPKC δ-isoform at 10 μM (IC50 = 3–6 nM), it is weakly active against the cPKC isoforms (IC50 = 30 μM) and is inactive against PKC-ε (IC50 = 30 μM) (70). Pretreatment of T84 cell monolayers with rottlerin at 10 μM, the concentration known to inhibit PKC-δ activity, had no effect on the ability of rSipA to elicit PMN transepithelial migration (Fig. 3). Similarly, the PKC-ε inhibitor V1-2, which blocks the function of PKC-ε, but not PKC-α, -βI, or -δ (10), also failed to inhibit PMN transepithelial migration induced by rSipA (Fig. 3). None of the PKC isoform inhibitors at the concentrations used for these studies affected transepithelial resistance or paracellular permeability (data not shown).

Fig. 3.

PKC-α inhibitors block ability of rSipA to induce polymorphonuclear neutrophil (PMN) transepithelial migration. Monolayers were pretreated with PKC isoform-selective inhibitors (at optimal inhibitory concentrations) for 1 h and rSipA for an additional hour. −, Control conditions, i.e., monolayers incubated only in the presence of HBSS. HBDDE, 2,2′,3,3′,4,4′-hexahydroxy-1,1′-biphenyl-6,6′-dimethanoldimethyl ether. *P < 0.001 (Student's t-test) compared with positive control. Data are from a single experiment and are representative of 5 experiments that showed an identical pattern of results.

As a nonspecific inhibitor, we used the PKC-βII inhibitor C2-4, because this cPKC isoform was not activated on rSipA stimulation. As expected, we found that this inhibitor did not adversely affect rSipA-induced PMN migration across T84 cell monolayers (Fig. 3). Furthermore, PMN transmigration to imposed gradients of N-formylmethionyl-leucyl-phenylalanine [which mediates PMN transepithelial migration by an SipA-independent process (20)] was not influenced by any of the PKC inhibitors examined (data not shown), indicating that these inhibitors do not block all responses to PMN movement across polarized cell monolayers. Because cPKC-α, -ε, and -δ are the isoforms activated in response to rSipA (within the sensitivity of the antibodies we used), results from these inhibitor studies imply that although PKC-α, -δ, and -ε are activated in response to rSipA exposure, PKC-α may be the key isoform responsible for the regulation of PMN transepithelial migration, at least for this portion of the signaling pathway.

Because rSipA is capable of reproducing the signaling events that link S. typhimurium to PKC, we next examined the PKC inhibitor sensitivity profiles with respect to wild-type S. typhimurium. Specifically, the PKC-α inhibitors Gö-6976 and HBDDE reduced the ability of wild-type S. typhimurium to induce PMN transepithelial migration by 60 and 70%, respectively, and this inhibitory response was dose dependent (Fig. 4, A and B). Conversely, the PKC-δ-, -ε-, and -βII-specific inhibitors failed to inhibit PMN transepithelial migration induced by wild-type S. typhimurium (Fig. 4, C and D). As described above, the inhibitory effect with the PKC-α inhibitor was not a global phenomenon, because PMN migration to imposed gradients of N-formylmethionyl-leucyl-phenylalanine was not influenced by any of the PKC inhibitors examined (data not shown). In addition, T84 monolayers incubated with the S. typhimurium strain VV341, an avirulent isogenic derivative of SL1344 that is rendered entry deficient by deletion of the hilA locus and does induce PMN transepithelial migration, were used as the negative background control. These results demonstrate that, even in the presence of other effector proteins and virulence factors, wild-type S. typhimurium retained a PKC inhibitor sensitivity profile similar to that of rSipA.

Fig. 4.

PKC-α inhibitors block ability of wild-type S. typhimurium to induce PMN transepithelial migration. Monolayers were pretreated with PKC isoform-selective inhibitors for 1 h and wild-type S. typhimurium for an additional hour. *P < 0.025; **P < 0.05 (Student's t-test). Data are from a single experiment and are representative of 5 experiments that showed an identical pattern of results.

Selectivity of PKC-α inhibitor activation in in vitro activation assays.

Having confirmed the functional effect of PKC-α inhibitors on the ability to inhibit PMN transepithelial migration under conditions induced by rSipA or wild-type S. typhimurium, we next sought to verify these observations biochemically. For these studies, we chose to examine only the cellular distribution of PKC-α to the membrane by Western blot analysis, because these results precisely correlated with the in vitro kinase activation assay. For these experiments, T84 cell monolayers were treated for 1 h in the presence and absence of the PKC-α inhibitor Gö-6976, the PKC-ε inhibitor V1-2, or the PKC-δ inhibitor rottlerin before the addition of rSipA to the apical surface. The specificity and selectivity for each of these inhibitors have been confirmed by dose-response inhibition experiments and by in vitro kinase assays in T84 cells (37). It is also important to note that the mechanistic action of these inhibitors functions to prevent the phosphorylation of the PKC isoforms, rather than to block their translocation to the membrane (38).

Thus, on the basis of the observation that PKC-α may be the key isoform in the signal transduction pathway leading to PEEC secretion and, subsequently, PMN transepithelial migration, we initially determined whether the PKC-α inhibitor Gö-6976 was able to block PKC-α phosphorylation after SipA stimulation of T84 cell monolayers. As expected, we found that rSipA initiated the phosphorylation of PKC-α (Fig. 5) and that such activation was inhibited in the presence of the PKC-α isoform inhibitor. As a control, the Western blot was probed using a nonphosphorylated PKC-α antibody; there was no appreciable difference in the intensity of the PKC-α band whether the T84 cells were treated in the absence or presence of the inhibitor. We next assessed the cellular distribution of PKC-α to the membrane by Western blot analysis. We found that, similar to the PKC-δ and -ε inhibitors (rottlerin and V-12, respectively), the PKC-α inhibitor (Gö-6976) did not considerably prevent the movement of PKC to the membrane (data not shown). Thus, not only do these results validate previous studies utilizing these inhibitors (37), but, more significantly, they also validate the functional consequence of the PKC-α inhibitor, because treatment of Gö-6976 blocked the activation of PKC-α but not the movement to the membrane. Moreover, because we previously showed that S. typhimurium via SipA initiates an ARF-6-dependent lipid-signaling cascade that directs the activation of PKC and release of PEEC, we examined whether blocking PKC-α activity also corresponds with a reduction in PEEC release. Indeed, we found that treatment of T84 cell monolayers with the PKC-α inhibitor Gö-6976 (5 μM) before infection with wild-type S. typhimurium completely ablated the ability of model intestinal epithelia to elicit PEEC release (57.3 and <1.0 pmol, i.e., less than the limit of detection, in the absence and presence of the inhibitor, respectively). Control uninfected monolayers were also evaluated for PEEC levels, which were lower than the limit of detection (<1.0 pmol).

Fig. 5.

PKC-α inhibitors block PKC activation but do not prevent the increase of PKC to the Triton X-100-insoluble membrane fraction. A: PKC-α inhibitor Gö-6976 blocked PKC-α activation. Control monolayers were incubated in the presence or absence of 5 μM Gö-6976 for 1 h without rSipA stimulation. Then the monolayers were stimulated for 1 h with rSipA (40 μg/ml) at 37°C, and Triton X-100-insoluble membrane fraction membrane was collected and prepared. Protein (50 μg) was separated on a 10% acrylamide gel and immunoblotted with the phosphorylated form of PKC-α antibody (Cell Signaling) to determine whether PKC was activated. The nonphosphorylated PKC-α antibody (Upstate Biotechnology) was used as the control. All the specific PKC isoform antibodies used in this study were obtained from Cell Signaling. Data are from a single experiment and are representative of ≥3 experiments that showed an identical pattern of results.

Subcellular localization of PKC-α after SipA induction.

PKC isoforms are distributed to varying degrees in mammalian tissue- and cell-specific patterns. Moreover, because PKC isoforms exhibit distinct subcellular localizations within individual cell types, we next examined the subcellular localization of PKC-α after apical exposure to rSipA using immunofluorescence in association with confocal microscopy. Consistent with our biochemical analysis, we found that, under control conditions, PKC-α was initially localized to the basal cytoplasm (Fig. 6). However, after apical exposure to rSipA, PKC-α began to redistribute toward the apical pole in many cells, and as early as 15 min, PKC-α was clearly localized to the apical membrane and subapical cytoplasmic domain (Fig. 6).

Fig. 6.

rSipA transloctaes PKC-α to the apical membrane. Monolayers were incubated with rSipA (40 μg/ml) for 15 min, fixed, permeabilized, and fluorescently labeled with anti-PKC-α (red) and FITC-phalloidin (green). Confocal images of the vertical (x-z) sections of each monolayer are shown. Control monolayers show immunolocalization of PKC-α at the basolateral cytoplasm. Under conditions of rSipA (40 μg/ml) stimulation at the apical membrane surface, PKC-α redistributes to the apical region. Data are from a single experiment and are representative of ≥3 experiments that showed an identical pattern of results.

DISCUSSION

Recruitment of inflammatory cells is considered to be a key virulence mechanism underlying Salmonella-elicited enteritis (22, 28). We are beginning to identify the molecular basis of this requirement and have established that SipA, a secreted substrate of the S. typhimurium SPI-1 type III secretion system, is not only necessary but is sufficient to activate epithelial signaling pathways and promote PMN transepithelial migration. Recent insight into this signal transduction pathway revealed that S. typhimurium, via the effector SipA, initiates an ARF-6- and phospholipase D-dependent lipid-signaling cascade that, in turn, directs activation of PKC and subsequent transepithelial PMN movement (7, 17). In the present study, we have used biochemical and cell biological methods to establish the pivotal role of PKC-α in mediation of the epithelial cell responses that lead to PMN transepithelial migration.

We have provided several lines of evidence linking activation of PKC-α induced by the S. typhimurium effector SipA to the epithelial signaling pathways that lead to PMN transepithelial migration. Our initial results from the in vitro kinase PKC activation assay revealed that, after rSipA exposure to the apical surface of T84 cell monolayers, this protein induced the activation of one cPKC isoform (α) and two nPKC isoforms (δ and ε) within the host cell. We also determined that only the PKC-α inhibitors (Gö-6976 and HBDDE), but not the PKC-δ-specific inhibitor (rottlerin) or the PKC-ε inhibitor (V1-2) significantly reduced S. typhimurium- and rSipA-induced PMN transepithelial migration across model intestinal epithelia. These results, combined with the PKC kinase activation studies, suggest that PKC-α plays an important role in mediating the signal transduction pathway that culminates in PMN transepithelial migration. The results of these inhibitor studies were also confirmed biochemically, in that we determined PKC translocation by subcellular fractionation coupled to Western blotting. Although we were not able to effectively localize PKC to the cytosolic fraction of T84 cells, we were able to proficiently localize PKC to the cytosolic fraction using the same protocol when other epithelial cells, e.g., Madin-Darby canine kidney cells (3), and A549 cells (unpublished observations) were examined. Interestingly, most of the PKC in the T84 cell monolayers partitioned to a fraction abundant in acidic glycolipids (i.e., phosphoinositols), which is consistent with the observation that the apical surface of T84 cells is rich in acidic glycophospholipids (unpublished observations). Unfortunately, despite numerous attempts, we were unable to modify the protocol to successfully enrich for PKC in the cytosolic fraction. Nevertheless, these results are consistent with our previous findings that cPKC isoforms participate in epithelial cell signaling to PMN (3). Finally, immunolocalization studies showed that, in response to rSipA, PKC-α relocalizes to the apical membrane domain.

Our observations indicate that although SipA can induce epithelial cell activation of PKC-α, -δ, and -ε, only PKC-α plays an important role in mediating the epithelial cell's responses that lead to PMN transepithelial migration. At this point, it is not clear how SipA might induce PEEC production via activation of host PKC-α or why there is a transient decrease in the activation of this isoform 30 min after stimulation. Traditionally, PKC is a known initiator of MAPK cascades in many cell types, where activation occurs downstream of purinergic receptor activation (14, 28, 41). PKC activation can also lead to the phosphorylation and degradation of IκB, which allows NF-κB to migrate to the nucleus and act as a transcriptional regulator. However, inhibitors of the NF-κB pathway have no effect on S. typhimurium-induced PMN migration. Another target of PKC is the regulatory light chain of myosin II, which in epithelial cells is predominantly found in the actin bundles that attach to junctional complexes (39). Although phosphorylation of myosin light chain by PKC-α has been found to stimulate the contractile activity of the motor and promote an overall increase in epithelial cell permeability, such a mechanism would not likely account for PEEC synthesis and release. Nonetheless, an effect of PKC on epithelial barrier function, which would assist in pathogen-induced PMN movement, is possible, inasmuch as several enteric pathogens produced alterations in tight junction permeability via PKC-dependent pathways (31).

In the gastrointestinal tract, several PKC isoforms, including PKC-α, -δ, and -ε, have been detected in rat and human colonic mucosa (4). There is now increasing awareness that PKC contributes to the pathogenesis of a number of inflammatory disease states. For instance, activation of PKC activity has been shown to lead to the expression and release of several proinflammatory mediators, including products of arachidonic acid metabolism and cytokines (2). Moreover, PKC activation has been shown to induce a severe inflammatory response in mouse skin, whereas selective inhibitors of PKC can act as an anti-inflammatory agent in some animal models (16, 42). On the basis of these observations, activation of PKC-α by S. typhimurium SipA likely plays an important role in mediating proinflammatory responses, rather than altering tight junctional components. Indeed, our finding that rSipA induced the activation of PKC-α, -δ, and -ε in intestinal epithelial cells but that only PKC-α appears to be involved in transepithelial signaling to PMN is consistent with recent studies that investigated the role of PKC isoforms in trinitrobenzene sulfonic acid-induced models of colitis in rats (1). Analogous to our findings, experimentally induced colitis in the rat resulted in elevated levels and activation of PKC-α, -δ, and -ε, and, in addition, these studies report that early expression and activation of PKC-α may play an important role in promoting colitis in this acute model of inflammation. Further substantiating the important contribution of PKC in the pathogenesis of inflammatory bowel diseases, biopsies from inflamed colonic mucosa of ulcerative colitis patients have been found to have an increased PKC activity in the cellular particulate fraction, suggesting an activated state of this enzyme (35).

Little information is available on the distribution and roles of individual PKC isoforms in the events leading to colonic inflammation. In the present study, we used a physiologically relevant epithelial model to provide evidence that closely links the ability of the S. typhimurium effector protein SipA to mediate transepithelial migration with the activation of PKC-α. Thus this work identifies PKC-α as an important mediator of the immune inflammatory response that results in PMN influx. Although PKC-δ and -ε were also found to be activated on stimulation with SipA, they appear to be unimportant in mediating the events leading to PMN transmigration; consequently, roles for these proteins in S. typhimurium pathogenesis await further investigation. We anticipate that subsequent identification of proteins phosphorylated by PKC-α during S. typhimurium-host cell interactions should provide a connection between PKC-α activation and apical release of the potent PMN chemoattractant PEEC. Such studies may lead to a better understanding of pathways by which chronic states of intestinal inflammation are aberrantly activated. Furthermore, on the basis of these findings, it may also be possible to selectively target the specific PKC isoform in the context of anti-inflammatory drug development.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-56754 and DK-33506 (to B. A. McCormick) and DK-48010 and DK-51630 (to J. B. Matthews).

Acknowledgments

We thank Dr. Bryan Hurley for numerous productive discussions regarding this investigation, Drs. Randy Mrsny and Dario Siccardi for quantifying PEEC, and Michael Pazos for superb technical assistance.

Footnotes

  • 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.

REFERENCES

View Abstract