Invasion of epithelial cells represents a potential pathogenic mechanism for Pseudomonas aeruginosa. We explored the role of mitogen-activated protein kinase kinases (MEK 1/2) and the extracellular signal-regulated kinases (ERK 1/2) in P. aeruginosa invasion. Treatment of corneal epithelial cells with MEK inhibitors, PD98059 (20 μM) or UO126 (100 μM), reduced P. aeruginosa invasion by ∼60% without affecting bacterial association with the cells (P=0.0001). UO124, a negative control for UO126, had no effect on bacterial internalization. Infection of cells with an internalization-defective flhA mutant of P. aeruginosa was associated with less ERK 1/2 tyrosine phosphorylation than infection with wild-type invasive P. aeruginosa. An ERK-2 inhibitor, 5-iodotubercidin (20 μM), reduced P. aeruginosa invasion by ∼40% (P=0.035). Together, these data suggest that P. aeruginosa internalization by epithelial cells involves a pathway(s) that includes MEK and ERK signaling proteins.
The opportunistic bacterial pathogen Pseudomonas aeruginosa is capable of causing serious infections if host immune defenses are compromised. Acute nosocomial pneumonia, chronic airway infection in cystic fibrosis patients, corneal ulcers, burn-wound infections, and bacteremia frequently involve this bacterium. The ability to invade and replicate within mammalian cells has been identified as a potential pathogenic mechanism for many clinical isolates of P. aeruginosa[2,,,5]. Invasion and intracellular replication requires bacterial lipopolysaccharide [5,,7], and internalization of P. aeruginosa into epithelial cells is significantly reduced by mutation of the pilin gene, pilA or the flagellar assembly gene, flhA. Host cell factors involved in P. aeruginosa invasion include: the actin cytoskeleton and protein tyrosine kinase activity; the cystic fibrosis transmembrane-conductance regulator; the surface glycosphingolipid asialo-GM1; Src-family tyrosine kinases [10,11]; intracellular calcium and calmodulin signaling; and RhoA. The manner in which these factors cooperate to allow cells to internalize P. aeruginosa, however, is yet to be determined.
Mitogen-activated protein (MAP) kinases represent a superfamily of cytoplasmic enzymes that are involved in mediating and coordinating cellular responses to a wide variety of extracellular stimuli [13,14]. At least three parallel MAP kinase signaling cascades (modules) have been identified, the activation of which results in phosphorylation of extracellular signal-regulated kinases (ERK 1/2) or the stress-activated kinases, JNK and p38, respectively. These pathways represent a point of convergence for multiple cell signaling events and are involved in the activation of cellular transcription factors, regulatory protein kinases, and cytoskeletal proteins.
Several recent studies have identified MAP kinase signaling proteins as being involved in, or associated with, the invasion of mammalian cells by different bacteria. These include Listeria monocytogenes, Staphylococcus aureus, Porphyromonas gingivalis, and Shigella flexneri. Pathway(s) of P. aeruginosa invasion have not yet been defined, but given the central role of MAP kinase proteins in cellular signal transduction processes and in regulation of actin cytoskeleton function, it is likely that one or more of these cascades is involved in the signaling events associated with P. aeruginosa invasion.
In this study, involvement of the MEK–ERK signaling module of MAP kinase signaling in P. aeruginosa invasion was examined. Within this signaling cascade, ERK 1/2 are activated (dual threonine and tyrosine phosphorylation) by MEK 1/2 which are themselves activated (dual serine phosphorylation) by Raf-1 kinase. This pathway is important in regulating cell growth and differentiation, and is activated by a number of extracellular signals including growth factors, and integrin binding to extracellular matrix proteins [19,20]. A role for MEK and ERK in P. aeruginosa invasion was hypothesized because of the association of these signaling proteins with other host cell factors involved in P. aeruginosa uptake, e.g. Src-family tyrosine kinases are involved in MEK–ERK activation [19,21], as is intracellular Ca2+. Furthermore, MEK and ERK are thought to be important in regulating the actin cytoskeleton [13,20].
The role of MEK–ERK signaling was explored by examining the effects of specific drug inhibition of MEK 1, MEK 2, and ERK 2 on P. aeruginosa invasion and by comparing the tyrosine phosphorylation levels of ERK 1 and 2 in corneal epithelial cells infected with wild-type invasive P. aeruginosa or an internalization-defective flhA mutant.
Materials and methods
Experiments were conducted using the wild-type invasive P. aeruginosa strains 6294 (serogroup O6), PA01 (serogroup O5), and an internalization-defective flhA mutant, PA01 flhA::Cmr (PA01 flhA−). Effects of restoring flhA were examined using a complemented strain PA01 flhA::Cmr+pPZ375flhA and its control PA01 flhA::Cmr+pP375. Bacteria were grown overnight on trypticase soy agar plates at 37°C which was supplemented with chloramphenicol (150 μg ml−1) for growth of PA01 flhA::Cmr, and with carbenicillin (300 μg ml−1) for selection of pP375 and pP375flhA.
Immortalized rabbit corneal epithelial cells were grown on 6- or 24-well tissue culture plates (Corning, New York, NY, USA) and fed on alternate days with modified SHEM containing bovine pituitary extract (5 μg ml−1) instead of cholera toxin. Experiments were performed using cells grown for 3–7 days after passaging, and results presented were obtained from cells grown between passages 6 and 20.
PD98059 and 5-iodotubercidin were obtained from Biomol Research Laboratories (Plymouth Meeting, PA, USA), and UO126 and UO124 were obtained from Calbiochem (San Diego, CA, USA). These drugs were dissolved in DMSO as 10 mM stock solutions which were aliquoted and stored at −20°C. Concentrations recommended by manufacturers were used. For the MEK inhibitors, UO126 and PD98059, concentrations used ranged from 10 to 100 μM, and for the ERK 2 inhibitor, 5-iodotubercidin, concentrations used ranged from 10 to 20 μM. Matching concentrations of DMSO were added to untreated controls. Bacterial and corneal epithelial cell viability in the presence of these drugs was monitored in control samples that were included in each experiment. None of the drugs affected bacterial viability or growth (assessed by viable counts), or corneal cell viability (assessed by trypan blue staining). Mouse monoclonal antibodies reactive with phospho-ERK 1 and 2, and goat polyclonal anti-ERK 1 and ERK 2 antibody that was used for the loading control were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). All other reagents were obtained from Sigma (St. Louis, MO, USA).
Gentamicin survival assays were used to quantify P. aeruginosa invasion of corneal epithelial cells grown on 24-well tissue culture plates as previously described. Bacteria were suspended into MEM (minimal essential Eagle's medium, with Earle's salts and l-glutamine [Cellgro™, Fisher Scientific, Pittsburgh, PA, USA] to which NaHCO3 [2.2 g l−1] had been added). Bacterial concentrations were determined by spectrophotometry (OD650), and confirmed by viable counting. Before the addition of bacteria, cells were washed once with MEM (500 μl) prior to subjecting them to inhibitor treatments (200 μl MEM+inhibitor at 10–100 μM concentrations) for 1 h at 37°C. A bacterial inoculum of 2×105 cfu in 200 μl of MEM was added per well of cells. Each well was calculated to contain ∼106 cells; thus the multiplicity of infection (MOI) was ∼0.2. Following a 3 h infection at 37°C in the continued presence of inhibitor, survivors of 1.5 h treatment with gentamicin (200 μg ml−1) were enumerated by viable counts. Prior to viable counting, cells were washed with MEM (500 μl) to remove residual antibiotic and were lysed by treatment with 200 μl Triton X-100 in MEM (0.25% v/v). Four to six wells were used for each sample, and experiments were repeated three times.
The total number of associated (adherent and invaded) bacteria over a 3-h period of incubation with cells was determined using the same method described for invasion assays, but without gentamicin treatment to kill extracellular bacteria. Instead, cells were washed three times with PBS before cell lysis with 1 ml Triton X-100 (0.25% v/v) in MEM. These assays were performed using cells grown on Transwell filters (12 mm, 0.45 μm pore) (Costar, Cambridge, MA, USA). Removal of the filter from its plastic holder prior to cell lysis allowed exclusion of bacteria that had adhered to the plastic insert.
Detection of ERK and ERK tyrosine phosphorylation
Cells were grown on 6-well tissue culture dishes. After one wash with MEM (2 ml), cells were exposed to ∼5×108 cfu bacteria in 2 ml of medium (MOI ∼6) for 15, 45, or 75 min. After removing the bacteria, cells were quickly washed once with ice-cold PBS (2 ml), and lysed with 1 ml of a boiling solution of sodium dodecyl sulfate (1% w/v), sodium orthovanadate (1 mM) and Tris–HCl (10 mM, pH 7.4). Cells were then scraped off the plastic, and the cell lysate transferred to a microcentrifuge tube (1.5 ml), and boiled for 5 min. Viscosity was reduced by passaging the lysate five times through a 27-gauge needle. The lysate was centrifuged (14 000×g, 4°C) for 5 min to remove cell debris and the supernatant collected. The total protein concentration was measured for each sample using a BCA assay kit (Sigma, St. Louis, MO, USA). Equal amounts of each sample (equivalent to ∼15 μg protein) were mixed with double strength SDS–PAGE sample buffer, boiled for 1 min, then subjected to SDS–PAGE (20 mA, 1.5 h) using precast Tris–HCl polyacrylamide gels 10% (Bio-Rad, Hercules, CA, USA).
Proteins were transferred from gels to nitrocellulose membranes by Western immunoblot (30 V, overnight, 4°C). Membranes were blocked using bovine serum albumin (1%, w/v) in TBS (Tris-buffered saline) with Tween 20 (0.1%, v/v) for 1 h at room temperature. Tyrosine phosphorylated ERK 1 and 2 was detected by exposing membranes to anti-phospho-ERK 1/2 monoclonal antibodies (diluted 1:1000 in blocking buffer) for 1 h at room temperature, then washing six times (5 min per wash) with TBS–Tween 20 (0.1%, v/v). Bound anti-phospho-ERK 1/2 antibody was detected using an horseradish peroxidase (HRP)-conjugated goat anti-mouse secondary antibody diluted 1:2000 in blocking buffer (1 h, room temperature). After six further TBS–Tween 20 (0.1%, v/v) washes, tyrosine phosphorylated ERK 1/2 was visualized using an enhanced chemiluminescence detection system (NEN, Belmont, MA, USA) and Kodak X-ray film.
Total ERK 1 and 2 was detected using the same protocol, except that goat anti-ERK 1/2 antibody was used to detect ERK, and HRP-conjugated donkey anti-goat secondary antibody was used.
Data are presented as the mean±S.D. Statistical assessment of the differences between means was performed using ANOVA for comparison of multiple groups, or Student's t-test for comparison of two groups. P values <0.05 were considered significant.
MEK inhibitors reduced P. aeruginosa invasion
Treatment of corneal epithelial cells with the MEK inhibitor PD98059 (20 μM) significantly reduced the invasion of P. aeruginosa (∼60% inhibition, P=0.0001, t-test) (Fig. 1A). Increasing the concentration of PD98059 to 40 μM or 100 μM did not significantly increase the amount of invasion inhibition achieved (data not shown). PD98059 (20 μM) had no significant effect on bacterial association with cells (P=0.202, t-test) (Fig. 1B), suggesting that the inhibitor blocked internalization of P. aeruginosa after they had attached.
The MEK inhibitor, UO126 (100 μM) also reduced P. aeruginosa corneal epithelial cell invasion (∼60% inhibition, P=0.0001, t-test) (Fig. 2A). Treatment of corneal cells with UO124 (100 μM), a negative control for UO126, had no significant effect on P. aeruginosa invasion when compared to controls. When directly compared to UO124, UO126 caused a 57% decrease in P. aeruginosa invasion (P=0.0002, t-test) (Fig. 2A). There were no significant differences in P. aeruginosa association between cells treated with UO126, UO124, or untreated controls (P>0.05, ANOVA) (Fig. 2B).
An internalization-defective flhA mutant of P. aeruginosa is associated with reduced tyrosine phosphorylation of ERK 1 and 2
Tyrosine phosphorylation of ERK 1 and 2 was detected in corneal epithelial cells after 15, 45, and 75 min of exposure to the wild-type P. aeruginosa strain PA01, an flhA mutant of PA01 that is defective in internalization, and a medium control (no bacteria). There was little difference in the tyrosine phosphorylation of ERK 1 or 2 at 15 min whether or not samples were incubated with bacteria (data not shown). After 45 min, however, there was significantly more ERK 1 and ERK 2 tyrosine phosphorylation in cells exposed to the wild-type invasive strain PA01 as compared to cells exposed to either the flhA mutant or no bacteria (Fig. 3A, upper panel). This difference was even more pronounced after another 30 min of bacterial exposure (75 min time point) with respect to ERK 1 phosphorylation. There continued to be less ERK 2 phosphorylation in cells exposed to the flhA mutant when compared to cells exposed to wild-type bacteria at 75 min. However, at this time point, the flhA mutant caused more phosphorylation of ERK 2 than of ERK 1, and also caused more ERK 2 phosphorylation when compared to controls (no bacteria). In the loading control (Fig. 3A, lower panel), levels of ERK 1 and ERK 2 proteins were found to be similar for all samples.
Corneal epithelial cells were infected with the PA01flhA mutant in which the flhA mutation was complemented in trans using the plasmid pPZ375flhA. As a control, other cells were exposed to the mutant complemented with the empty plasmid vector pPZ375. A 75 min incubation was chosen since this corresponded to the greatest difference in ERK 1/2 phosphorylation between the wild-type and flhA mutant bacteria in other experiments. The complementation with flhA resulted in more ERK 1/2 phosphorylation in corneal cells than complementation with the vector alone (Fig. 3B, upper panel). As seen previously, there was more ERK 2 phosphorylation in cells exposed to the flhA mutant than in no bacteria controls. In the loading control (Fig. 3B, lower panel), levels of ERK 1 and ERK 2 were similar across all samples.
Effect of an ERK 2 inhibitor on P. aeruginosa invasion of corneal epithelial cells
The changes in ERK 1/2 phosphorylation in response to an internalization-defective mutant of P. aeruginosa showed that phosphorylation of these proteins correlated with the ability to invade cells. To directly test whether or not ERK 2 played a role in invasion, the effect of 5-iodotubercidin, an inhibitor of ERK 2, was examined. Corneal epithelial cells were treated with this drug and the effect on P. aeruginosa invasion was assessed (Fig. 4). A 10 μM concentration of 5-iodotubercidin had no significant effect on P. aeruginosa invasion, while 20 μM of this ERK 2 inhibitor reduced invasion by ∼40% (P=0.035, t-test).
The data provide evidence that P. aeruginosa internalization by corneal epithelial cells involves both MEK and ERK signaling proteins. Although a pathway for P. aeruginosa invasion has not yet been defined, the established role of MEK 1 and 2 as upstream activators of ERK 1 and 2 suggests that P. aeruginosa internalization will involve the MEK–ERK module of MAP kinase signaling. This contrasts with Salmonella enterica (serovar typhimurium) invasion of epithelial cells and macrophages for which MEK inhibition did not reduce bacterial invasion.
MEK 1 activation, with subsequent involvement of ERK 2 MAP kinase, has been shown to be involved in the invasion of HeLa cells by L. monocytogenes. PD98059 inhibits both MEK 1 and MEK 2 without affecting related kinases [28,29]. PD98059 has an IC50 of 4 μM for MEK 1 inhibition, and an IC50 of 50 μM for inhibition of MEK 2. Thus, the dose of PD98059 that reduced internalization of P. aeruginosa in the present study (20 μM) would be more likely to have inhibited MEK 1 than MEK 2, implicating MEK 1 involvement in P. aeruginosa internalization.
PD98059 blocks the activation of MEK, while UO126 blocks the activity of phosphorylated MEK. Neither drug completely blocked P. aeruginosa internalization, suggesting that factors other than MEK are involved. MEK inhibition by PD98059 can be overcome by potent agonists at growth factor receptors. Although it is possible that P. aeruginosa receptor binding could partially override MEK inhibition, it is likely that P. aeruginosa can also utilize MEK-independent internalization pathway(s). Multiple invasion mechanisms have been described for other Gram-negative bacteria such as Yersinia spp., Salmonella spp., and Shigella spp..
The inhibition of P. aeruginosa invasion by 5-iodotubercidin suggested that ERK 2 is involved. The results of phosphorylation studies, however, suggest that ERK 1 may also play a role. Phosphorylation of both ERK 1 and 2 by flhA-competent P. aeruginosa was reduced by flhA mutation which also reduces P. aeruginosa internalization. This suggests that the P. aeruginosa invasion pathway involving MEK/ERK differs from that of L. monocytogenes, for which invasion was associated with the activation of ERK 2 but not of ERK 1. It may also differ from S. aureus invasion of osteoblasts, for which ERK 1 and 2 activation was associated with both bacterial attachment and invasion.
The bacterial–host cell signaling events upstream of MEK and ERK that help mediate P. aeruginosa internalization are yet to be defined. Bacterial lipopolysaccharide (LPS) is a well established activator of ERK 1/2 and other MAP kinases in mammalian cells [32,33], and MEK–ERK signaling is involved in P. aeruginosa LPS-mediated upregulation of mucin gene expression. Since P. aeruginosa LPS is also involved in bacterial attachment and invasion of corneal epithelial cells, the role of MEK and ERK in P. aeruginosa internalization could be linked to LPS-associated invasion events. However, both P. aeruginosa pili and flagella can also activate MEK–ERK signaling in mammalian cells via Ca2+-dependent pathways resulting in upregulation of IL-8 or mucin gene expression respectively. Moreover, pili- (PilA) and flagella- (FlhA/FliC) associated proteins are also involved in P. aeruginosa invasion. Most of the ERK 1 phosphorylation, and part of the ERK 2 phosphorylation, induced by P. aeruginosa in the present study was shown to involve flhA. However, residual ERK 2 phosphorylation induced by the flhA mutant, which lacks flagella, may involve pili- or LPS-mediated signaling, although it is not known if this phosphorylation is associated with P. aeruginosa invasion.
In conclusion, this study shows that both MEK and ERK signal transduction proteins are involved in the internalization of P. aeruginosa by corneal epithelial cells. This is most likely via a MEK–ERK module of MAP kinase signaling involving flhA-dependent P. aeruginosa invasion mechanisms. Delineation of the signaling pathway(s) used by P. aeruginosa to invade mammalian cells may allow the development of therapies that prevent invasion as a pathogenic mechanism.
This research was supported by a grant from the National Institutes of Health, Bethesda, MD, USA (EY10221), and a Faculty Research Grant from the University of California, Berkeley, CA, USA to S.M.J.F. We also wish to express our thanks to Alcon Corp., Bausch and Lomb, and CIBAVision for their generous research support.