Src kinase Lyn is crucial for Pseudomonas aeruginosa internalization into lung cells

Authors

  • Shibichakravarthy Kannan,

    1. Department of Biochemistry and Molecular Biology, University of North Dakota School of Medicine and Health Sciences, Grand Forks, USA
    Search for more papers by this author
  • Aaron Audet,

    1. Department of Biochemistry and Molecular Biology, University of North Dakota School of Medicine and Health Sciences, Grand Forks, USA
    Search for more papers by this author
  • Jessica Knittel,

    1. Department of Biochemistry and Molecular Biology, University of North Dakota School of Medicine and Health Sciences, Grand Forks, USA
    Search for more papers by this author
  • Saman Mullegama,

    1. Department of Biochemistry and Molecular Biology, University of North Dakota School of Medicine and Health Sciences, Grand Forks, USA
    Search for more papers by this author
  • George F. Gao,

    1. Center for Molecular Immunology, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author
  • Min Wu

    Corresponding author
    1. Department of Biochemistry and Molecular Biology, University of North Dakota School of Medicine and Health Sciences, Grand Forks, USA
    • Department of Biochemistry and Molecular Biology, University of North Dakota School of Medicine and Health Sciences, 501 N Columbia Road,PO Box 9037, Grand Forks, ND58203-9037, USA, Fax: +1-701-777-2382
    Search for more papers by this author

Abstract

Lyn is an important B cell signaling kinase of the Src tyrosine kinase family with a broad range of functions from cytoskeletal changes to induction of apoptosis. However, the role of Lyn in infectious diseases is not clear. Here, we demonstrate that Lyn activation by phosphorylation significantly impacted invasion of an alveolar epithelial cell line, primary lung cells, and rat lungs by Pseudomonas aeruginosa (PA), a common opportunistic lung pathogen affecting individuals with deficient lung immunity. Our results indicate that activation of Lyn and its interaction with rafts and TLR2, played an important role in the initial stages of PA interaction with host cells. The role of Lyn was further evaluated using the pharmacologic Src-specific inhibitor PP2, a dominant negative mutant, and finally confirmed with Lyn-deficient (Lyn–/–) bone marrow-derived mast cells. Inhibition of Lyn's function by above approaches prevented PA internalization. Moreover, blocking of Lyn also affected downstream events: induction of inflammatory cytokines and apoptosis. This report brings out a new role of Lyn in infectious diseases and indicates potential new targets for prevention and treatment of infections.

Abbreviations:
AEC II:

alveolar type II epithelial (cells)

BMMC:

bone marrow-derived mast cells

CFTR:

cystic fibrosis transmembrane conductance regulator

DN:

dominant negative

MβCD:

methyl-β-cyclodextrin

MTT:

methylthiazoletetrazolium

PA:

Pseudomonas aeruginosa

Introduction

Numerous studies have explored the pathogenesis mechanism of Pseudomonas aeruginosa (PA) in the airway epithelium as the chronic infection occurs mainly in the bronchial epithelium in cystic fibrosis patients 15. PA also causes acute lung injury and severe pneumonia mainly by attacking alveolar epithelial cells 6. The molecular mechanism of PA invading the alveolar cells is not well characterized, although a recent report indicates that infection of primary type I epithelium is dependent on caveolin 2 7. Alveolar type II epithelial (AEC II) cells are lung progenitor cells that respond to infection and injury. Recently, AEC II cells have been reported to have innate immune functions, such as secretion of chemokines/cytokines 8, recruitment of dendritic cells 9, and antigen presentation 10. Therefore, AEC II cells may play crucial roles in PA infection by contributing to host innate immunity.

Src family tyrosine kinases are the major signaling components in various cells. It has been shown that invasion of human epithelial cells by PA involves activation of p60Src and p59Fyn 11. However, the exact role of p60Src or p59Fyn was not clearly defined. Src family tyrosine kinases are thought to play a major role in reorganizing the cytoskeleton which might be associated with bacterial internalization 12. The activation of Src family kinases may be a host response contributing to eliminate the bacteria. Even though all members of this family have very homologous sequences, their substrates and functions differ. Lyn is an important kinase in this family and is involved in regulating broad cellular functions in a variety of cells, such as cytoskeletal changes 13, cell proliferation 14, and apoptosis 15. However, it is unclear whether Lyn plays a role in host response to pathogens. Here, we investigated the interaction between PA and AEC II cells. Our results demonstrate that PA infection activated Lyn through lipid raft-mediated mechanism, which may play a critical role in host response to PA invasion.

Results

PA infection activates lipid rafts in alveolar epithelial cells

PA internalization was previously reported to be dependent on raft activation in airway cells. In A549 cells, we also observed raft activation by infecting with GFP-PAO1 for 30 min (Fig. 1a). Internalized bacteria were co-localized with raft aggregates determined by Alexa Fluor 544 nm-labeled cholera toxin B chain that binds ganglioside GM1 (red color; Sigma). Throughout this study, all co-localization data were analyzed with the Zeiss software, and a selected area with co-localization correlation coefficient >0.85 was considered significant. Using rabbit antisera specifically to PA, we stained the cells on coverslips after co-incubation with PAO1 for 30 min to further confirm the internalization of the bacterium. The antibody staining was done after membrane permeabilization to detect intracellular bacteria. Results show that a significant number of PA were inside the cells positively stained with the antibodies (Fig. 1b). The internalized bacteria were confirmed by performing z series of confocal images. Surface-adherent bacteria can be distinguished from internalized bacteria by 3D reconstruction of the z series using the Zeiss LSM software (Supplementary materials Fig. 1).

Figure 1.

PA infection activating lipid rafts in vitro and in vivo. (a) Formation of lipid raft aggregates in A549 cells following GFP-PAO1 infection on live cells. GFP-PAO1 (green) was co-cultured with A549 cells for the indicated times, and internalized co-localizing with activated rafts (aggregate, arrows) identified by Alexa Fluor 544 nm-cholera toxin B chain that binds ganglioside GM1 (rafts marker). (b) PAO1 internalization identified by antibodies against this bacterium. Internalization was done by incubating cells with PAO1 for 0 and 30 min, washed to remove free PA, and stained with rabbit antibodies against PAO1. (c) Infection of primary AEC II cells by PA. Live PAO1 was labeled with DiI by the affinity to lipopolysaccharide. Dead PAO1 was obtained by heating at 60°C for 60 min. After co-incubation of the cells with PA for 30 min, live PAO1 (DiI red) was internalized co-localizing with activated raft aggregates (arrows) identified by FITC-cholera toxin B chain. (d) In vivo infection of AEC II cells following airway delivery of PAO1 to rat lungs (60 min). Cryosections show co-localization between PAO1 (labeled with DiI, arrows) and AEC II marker SP-C (green) detected by immunofluorescence with monoclonal antibodies (merged, orange color) (original magnification ×400). Three rats were used per group and the experiments repeated three times. All the data in this figure (a–d) are representative of three experiments.

Evidence for PA invasion was also obtained in primary AEC II cells (1 day post-isolation from Sprague-Dawley rats cultured on collagen-coated coverslips), and lipid rafts identified by FITC-labeled cholera toxin B chain (green) were found to be clearly reorganized into aggregates at bacterial contact sites on the cell membrane. Neither the control without PA nor heat-killed PA showed raft aggregates (Fig. 1c). AEC II infection was assessed within 2 days of isolation while the cells were still expressing the typical AEC II markers, such as SP-C, cytokeratin 8, and lamellar bodies (Supplementary materials Fig. 2).

Figure 2.

PA invasion associated with Lyn activation. (a) Activation of tyrosine kinases by PAO1 probed by phospho-tyrosine (PY20) antibodies in lung epithelial cells. A significant increase in phosphorylation of the proteins with molecular weights similar to Src p53/56 was observed by incubating the cells with PA for 15 min. ATCC27853 is another WT PA strain used for comparison. Similar protein loading was confirmed by reprobing GAPDH. (b) Co-localization of Lyn with PAO1 stained with DiI in A549 cells (bar = 10 µm). The co-localization was determined by co-culturing A549 cells for 30 min with PAO1, which showed orange or purple color (arrows). Lyn was also co-localized with lipid rafts under similar conditions. (c) Decrease in phosphorylation of Lyn Tyr508 following PAO1 infection (30 min co-culture with A549 cells) as identified in lipid raft fractions (3–5) using sucrose density gradient. Similar protein loading was confirmed by reprobing with total Lyn. (d) Activation of Lyn after in vitro tyrosine kinase assay with the exogenous substrate enolase. Tyr397 phosphorylation was confirmed by phospho-Src antibodies after IP with Lyn antibodies. A significant increase (50%) in phospho-enolase and phospho-Lyn was seen by PA infection at 15 min and 30 min. (e) Determination of activation site in Lyn by cyanogen bromide peptide digestion. This assay was done in IP samples by Lyn antibodies after PA infection of A549 cells (30 min). The peptide fragment containing active site (Tyr397) was significantly phosphorylated, while the peptide with Tyr508 was not activated. Lyn–/– BMMC infected with PA for 30 min showed no increase in Tyr397 phosphorylation compared to the uninfected control. Band densitometry was done using Lumi-imager software. Ratios of each treatment calculated against the control are shown on the bottom panel. Data were analyzed by Student's two-tailed t-test with 95% confidence interval (*p<0.05). Data are presented as means ± SD. All the data in this figure (a–e) are representative of three experiments.

Next, we investigated in vivo infection of PA to Sprague-Dawley rat lungs by direct instillation of DiI-labeled PAO1 through trachea (animal protocols have been approved by the IACUC committee of University of North Dakota). The lungs were treated with polymyxin 1 h after infection and lavaged to remove un-internalized bacteria. In vitro data show that polymyxin is very efficient in killing the bacteria that were not internalized inside of cells. Immunofluorescence shows that PAO1 (red DiI staining) was co-localized with SP-C in lung tissues (Fig. 1d), while the control without primary antibody did not show staining with SP-C. It has been shown that DiI does not leak out from stained pathogens 16. The data indicate that PA could be internalized into alveolar type II epithelium.

PA invasion is associated with Lyn activation

To identify the proteins participating in the underlining signaling process, we assessed the phosphorylation patterns following incubation with PA (WT PAO1 and ATCC27853 strains) for 30 min. Various protein tyrosine kinases were activated in lung cell lysates that were analyzed by Western blotting with antibodies against the phospho-tyrosine (PY20) (Fig. 2a), as quantitated by densitometry (p<0.01). Some of the activated proteins correspond to several Src family members including p56/53 members. The quantification was performed based on proteins of molecular masses of 56/53 kDa, with background subtraction.

We hypothesized that Lyn is involved in the process of PA infection since it is a multifunctional signaling molecule with key role in cell survival. To test this hypothesis, we transfected A549 cells with a Lyn-GFP construct containing only the N-terminal amino acids including two sites for lipid anchors for targeting to the membrane. This construct has previously been used as a marker for lipid rafts 17. This construct is used as phenotypical marker of Lyn as well as lipid rafts, but without functional indication. We observed an increased co-localization of Lyn-GFP with DiI-labeled bacteria (Fig. 2b, top panel). Similarly, we observed co-localization between Lyn and lipid rafts when the cells were infected with PA (Fig. 2b, bottom panel; non-stained PA can be seen in the differential interference contrast image). Thus, the data suggest that Lyn may be involved in PA infection in a raft-dependent manner.

In order to study phosphorylation events, we sought to investigate which tyrosine residues in Lyn become phosphorylated during PA infection. Lyn and other Src members have two predominant tyrosine residues – one in the middle (Tyr397) and one in the cytoplasmic tail (Tyr508). When the Tyr508 is phosphorylated, the cytoplasmic tail folds to cover the active site, resulting in an inactive form. Certain extracellular signals cleave this phosphate (Tyr508) and unfold the protein to expose the active-site tyrosine (Tyr397). The active-site tyrosine residue can be phosphorylated and interact with other proteins that contain an Src homology 2 domain (SH2).

Using sucrose density gradient centrifugation, we showed that Lyn phosphorylation in raft fractions was decreased following PAO1 infection for 30 min (Fig. 2c). The phospho-Lyn antibodies recognize the Tyr508 epitope, thus this result implies a decrease in inactive form of Lyn in the raft fractions, which may yield increased Lyn activity. The loading of immunoblots was checked using total Lyn staining, and the distribution of total Lyn has not been significantly changed upon PA infection. To evaluate whether Lyn's active site is phosphorylated, we immunoprecipitated Lyn and performed in vitro kinase assay using enolase as a substrate. The results showed increased activity in PA-infected samples compared to the control (Fig. 2d).

To confirm these data, we performed cyanogen bromide peptide digestion assay. We detected all phosphorylated peptide fragments of Lyn by immunoblotting with anti-phospho-tyrosine (PY20) antibody and identified the fragments based on their molecular weights. The band corresponding to the active-site fragment (containing Tyr397, 8 kD) with 30-min PA infection showed over fourfold increase in phosphorylation compared to uninfected control in A549 cells, while no phosphorylation increase was observed in the inactive-site fragment (containing Tyr508, 4 kD) (Fig. 2e, top panel). However, there was no phosphorylation increase in the active-site fragment in Lyn–/– bone marrow-derived mast cells (BMMC) obtained from M. Hibbs, Royal Hospital of Melbourne 18 (Fig. 2e, middle panel). Equal sample loading was further checked by Coomassie blue staining (Fig. 2e, bottom panel). Although the data from BMMC (phagocytic cells) were not meant to reflect the function of epithelial cells, the roles of Lyn in PA infection may be similar.

Disruption of rafts and Lyn function hampers PA internalization

To further identify roles of activated Src kinases in bacterial invasion, we used a blocking assay to show that cell survival was increased through pretreatment of the cells with PP2 (specific blocker for Src53/56 members) as compared to the control (Fig. 3a). The internalization was done by incubating with bacteria at 37ºC for 30 min and surface bacteria were killed by polymyxin treatment. The cells were further incubated for 4 h before the methylthiazoletetrazolium (MTT) assay to detect cell survival. This test examined that blocking certain Src kinases can result in less internalization of PA, indicating that Lyn may be associated with cell survival. It should be noted that the increase of survival was subtracted for the bacterial control to get net mammalian cell survival.

Figure 3.

Disruption of Lyn hampering PA internalization. (a) Increase in cell survival by Src kinase inhibitor PP2 determined by the MTT assay. PP2 (5 nM) is a synthetic blocker specific for p53/p56 Src kinases. This concentration had no toxicity to A549 cells or to PA. A549 cells were pretreated with PP2 or PBS for 30 min, and infected with PA for 30 min. The cells were then washed and treated with polymyxin for 60 min. PP2-treated cells had significantly increased cell survival compared to controls. The Student's two-tailed t-test was used to analyze the results with 95% confidence interval. Data are presented as means ± SD. (b) Inhibition of PA internalization by DN Lyn mutation. A549 cells were transfected with Lyn pcDNA3 constructs (WT; K275D: DN) using LipofectAMINE and selected by neomycin for 1 wk to kill non-transfected cells. Left panel: immunofluorescence; right panel: quantitative data of the left panel by counting at least 75 cells to get a percentage. The three bars represent 0–5 (open), 5–10 (hatched) and >10 (solid) bacterial rods counted per cell. Data are presented as means ± SD. Lyn WT transfected cells had significantly higher PA internalization compared to Lyn K275D transfectants. Data were analyzed by one-way ANOVA with the multiple comparison procedures (*p<0.05). (c) Blocking of PA infection in vivo in rat lungs by Src kinase inhibitor PP2 (green, Lyn; red, DiI-PAO1; original magnification ×400). PAO1 (1×108) was co-localized with Lyn (arrow). Pre-treatment of the lung with PP2 (5 nM) for 30 min was done before the instillation of DiI-labeled PAO1. Un-internalized PAO1 was killed by incubation with 100 µg/mL of polymyxin for 60 min and washed away before fixing tissues. Three rats were used per group and the experiments repeated three times. All the data in this figure (a–c) are representative of three experiments.

Next, we sought to determine the role of Lyn in the bacterial invasion. Transfection of a dominant negative (DN) construct (LynK275D) drastically blocked bacterial infection in A549 cells, while cells transfected with Lyn WT construct showed comparable internalization to non-transfected controls (Fig. 3b). The cells were incubated with 10:1 ratio of PA for 30 min. The samples were then treated with polymyxin for 30 min, washed and examined by confocal fluorescent microscopy. Furthermore, we demonstrated that PP2 drastically blocked the infection of PA in vivo in rat lungs (Fig. 3c), strongly suggesting that Src kinases play a role in PA invasion. Moreover, the invaded bacteria appear to be preferentially located at type II cells. To ascertain the data about Lyn's role in PA infection, we have examined PA infection in a genetic knockout model using Lyn–/– BMMC. Consistent with the above results, Lyn–/– BMMC demonstrated less PA internalization than the WT control BMMC cells as determined by confocal microscopy (Fig. 3d). Taken together, our data suggest that Lyn is a critical signaling protein involved in PA internalization.

TLR2 is involved in PA-induced Lyn activation

To further elucidate the signaling pathway regulating epithelial infection by PA, we sought to identify other proteins that are associated with Lyn's activation. TLR2 was previously reported to initiate signals in PA infection 19, 20, and recently associated with lipid rafts 19, 20. Since Lyn is a raft resident protein involved in PA infection, we hypothesized that Lyn may be associated with TLR2 activation. Using co-IP, our data show a PA-induced increase of TLR2 in IP samples pulled down by Lyn antibodies, suggesting an interaction between Lyn and TLR2 (Fig. 4a). This is further confirmed by demonstrating that Lyn's immunoblots were also positive in IP samples pulled down by TLR2 antibodies (Fig. 4a). To examine whether this association with TLR2 is specific, we used TLR5 as control and no significant association of TLR5 with Lyn was detected. However, DN transfection of LynK275D construct abolished this TLR2-Lyn interaction (Fig. 4a, right panel), but had no influence on TLR5. These results indicate that Lyn may bind TLR2, a pattern recognition molecule associated with rafts, and activates Lyn through immunoreceptor tyrosine-based activation motif in the cytoplasmic tail of TLR2.

Figure 4.

TLR2 involvement with PA infection and Lyn. (a) TLR2 interaction with Lyn during PA infection. PA infection was accompanied with Lyn/TLR2 association, demonstrating a strong Lyn signal in samples pulled down by TLR2 antibodies. There was no indication that TLR5 was involved in binding with Lyn during PA infection (control). Control, vector or LynK275D DN construct-transfected cells were treated with PAO1 for 30 min. Cell lysates were immunoprecipitated with either Lyn or TLR2 antibodies. The samples were then Western-blotted with TLR2 and Lyn antibodies, respectively. (b) Reduction in PAO1-induced invasion by transfection of TLR2 DN mutant (TLR2 P681H). Left panel, immunofluorescence; right panel, quantitation was made to demonstrate three different states of bacteria invasion: surfaces only (open), insides of the cells (<10 bacteria per cell, hatched) and insides of the cells (>10 bacteria per cell, solid). Data were analyzed by the one-way ANOVA and followed by the Holm–Sidak multiple comparison procedure. Data are presented as means ± SD. (c) Increase in cell survival by introducingTLR2 DN mutation determined by the MTT assay (Student's two-tailed t-test; *p<0.05, **p<0.01). Cells with TLR2 knockout were incubated with PAO1 for 30 min, washed twice with medium and incubated for 6 h before MTT assay. Data are presented as means ± SD. MβCD and PP2 were used as described above. TLR2 DN transfected cells had less internalization compared to the controls. (d) Significant reduction in PA internalization in Lyn–/– BMMC. Lyn–/– BMMC and Lyn WT BMMC (expressing Lyn) were incubated with DiI PAO1 (MOI = 0:1) for 30 min before taking confocal images. All the data (a–d) in this figure are representative of at least three different experiments.

To confirm the result, we transfected A549 cells with TLR2 DN mutation to examine the effect on PA invasion using LipofectAMINE. Disruption of TLR2 significantly blocked PAO1 invasion to A549 cells compared to the controls (Fig. 4b). By counting the status of internalizing bacteria, we show that majority of the bacteria were on cell surfaces with DN TLR2-transfected cells, while the majority of the bacteria had entered the control cells expressing TLR2 (p<0.01). Transfection of a TLR2 DN mutant resulted in an increase in cell survival determined by MTT assay (p<0.01) (Fig. 4c). In addition, cell survival was further increased by perturbing lipid rafts with methyl-β-cyclodextrin (MβCD) (p<0.05) and by perturbing Lyn with PP2 (p<0.05) (Fig. 4c). Similarly, PA internalization in Lyn–/– BMMC was reduced after 30 min incubation with DiI PAO1 compared to Lyn WT BMMC (expressing Lyn). Although BMMC are phagocytic cells that may not be a good model for studying epithelial cells, these results indicate that Lyn plays a clear role involving PA initial contact and internalization into a variety of cell types.

Disruption of rafts and Lyn function prevents PA-induced inflammatory damage

Finally, we sought to confirm roles of rafts in bacterial invasion by blocking rafts. The concentrations of MβCD or filipin (alone) for inhibiting raft activation did not show ill effects on PA or A549 cells. Filipin (disrupts caveolin) decreased PA internalization into A549 cells, thereby increasing cell survival as determined by the cell proliferation assay (Fig. 5a). We then examined whether PA infection of epithelial cells could induce apoptosis. The results indicated that significant apoptosis was induced by PA as determined by terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay (Fig. 5b). Apoptosis, however, is reduced by addition of either PP2 or cholesterol chelator MβCD as assessed by TUNEL assay (Fig. 5b), suggesting that some Src kinases together with rafts have an impact on PA infection-induced cell death. PA-induced apoptosis was confirmed by annexin V and propidium iodide staining as shown in Fig. 5c. MβCD reduced the bacterium-induced apoptosis in lung cells against the control. More importantly, we demonstrate that the Src kinase Lyn was involved in apoptosis by transfecting the DN Lyn construct and by staining mitochondria with the mitotracker (Fig. 5d). The results indicate significant reduction of mitochondria toxicity by bacterial invasion in Lyn DN-transfected cells as compared to the control. The results suggest that Lyn is crucial for bacterial internalization and subsequent cell apoptosis, since blocking Lyn's function causes significant reduction of bacterial internalization and apoptosis.

Figure 5.

Decrease in PA internalization and apoptosis by disruption of rafts and Lyn. (a) Increase in cell survival by disrupting rafts with decreasing PA internalization. Cells were pretreated with 10 mM of MβCD, 5 µM of filipin. Data are presented as means ± SD. Then infection was performed for 30 min with PAO1, and unbound bacteria were washed away. The cells were incubated for 18 h and were assayed with a MTT kit. Multiple comparisons (versus the control) were performed by a one-way ANOVA and followed by the Holm–Sidak method (*p<0.05). Each treatment was performed in triplicates with n=50 000 cells per well in a 96-well plate. (b) Significant decrease in PA-induced apoptosis by disrupting rafts and Lyn in A549 cells. TUNEL was used to detect DNA damage in the nucleus (arrow) following PAO1 infection for 1 h. Raft or Lyn blockers were used as above. Bacteria were stained with propidium iodide. (c) Reduction in PA-induced apoptosis determined by annexin V and propidium iodide staining. Perturbation of lipid rafts by MβCD decreased apoptosis of A549 cells (Student's two-tailed t-test). At least 100 cells were screened for positive annexin staining in each group and cells with propidium iodide staining were excluded to distinguish apoptosis from necrosis. Open bars: control; solid bars: PA added. (d) Decrease in apoptosis by introducing Lyn DN mutant. Transfected cells (see above) were infected with PAO1 for 30 min. Apoptosis was manifested by less mitochondrial damage compared to the control (GFP-PAO1, green; mitotracker: red; bar = 10 µm). Mitotracker exclusively stains only the mitochondria. A diffuse cytoplasmic staining is indicative of dye leakage from damaged mitochondria. PA treatment at 0 min did not cause apparent mitochondrial damage (not shown). (e) Decrease in secretion of inflammatory cytokine IL-1β. PAO1 incubation of A549 cells increased the IL-1β secretion to the medium after co-incubation for 2 h determined by standard ELISA. Blocking was performed by addition of an appropriate amount of raft or Lyn inhibitors as described above for 30 min before addition of PAO1. Multiple comparisons test (pairwise) was performed by a one-way ANOVA and followed by the Holm–Sidak method; *p<0.05, **p<0.01. Data are presented as means ± SD. All the data in this figure (a–e) are representative of three experiments.

Since one of the grim effects of PA infection is excessive release of inflammatory cytokines (i.e. IL-1β), we examined the production of this cytokine after infection. The data show that blocking of lipid rafts by MβCD, and TLR2 antibodies significantly reduced IL-1β secretion following PA infection using a standard ELISA assay (p<0.01; Fig. 5e). Similarly, we observed that blocking of Src kineases by PP2 reduced IL-1β secretion following PA infection (p<0.01; Fig. 5e). The reduction of the cytokine by the blockers is significant, reverting to the levels of the non-treated control. The data suggest that blocking of rafts and/or the blocking of certain Src kinases may reduce inflammatory response against the bacterium.

Discussion

We report that Lyn is constitutively expressed in alveolar epithelial cells and is activated by PA infection. Our results also demonstrate that Lyn activation was dependent on lipid raft aggregation since modulations of raft environments by MβCD potentially disrupt PA internalization. We also revealed that Lyn is activated at its key tyrosine site (Tyr397) in PA infection by biochemical analyses, which was confirmed by several blocking approaches, particularly by DN mutations. Since Lyn is not reported to be associated with TLR2, we propose that this is a novel innate immune response that is divergent from the classical TLR2 (MyD88) pathway 21.

Previous studies of PA infection focused on A549 cells, but so far there are no studies on primary type II cells. Recently, AEC II cells have been linked to anti-microorganism immunity through directly secreting chemokines 8, inflammatory cytokines 22, surfactant proteins, and through stimulating dendritic cell migration 9. Also, a recent report shows that AEC II cells can have an antigen-presenting role in microbial infection 10. For certain pathogens (e.g.Chlamydia pneumoniae), AEC II cells are major target cells in chronic infection 23. A549 cells do not express cystic fibrosis transmembrane conductance regulator (CFTR) 24. Fetal alveoli express CFTR, while isolated AEC II cells express low levels of CFTR in cell culture 25. Based mainly on TEM studies in A549 cells 26, 27 and CFU assay 28, most previous studies demonstrated a slow infectious process, which may be due to limitations in techniques, bacterial strains, and cell types. In addition to the CFU assay, we used multiple approaches (different dye staining including GFP-expressing PAO1) through powerful laser scanning confocal images and demonstrated that the infectious process was relatively efficient with PA WT strain, but it is significantly impeded with a pili-deficient mutant (Supplementary materials Fig. 3), which is consistent with a previous study 29.

PA infection involves initial contact to airway epithelial cells through interactions between certain bacterial outer membrane components and host cell receptors like TLR2 (peptidoglycan), TLR4 (binding to LPS), TLR5 (binding to flagella) 19, etc. However, a recent study indicates that TLR2 rather than TLR4 is particularly involved in PA invasion of the airway epithelium 20. Many of these putative receptors have been found to be associated with rafts. Disruption of lipid rafts by various treatments has previously been shown to inhibit PA adhesion 30. Our result suggests an association of TLR2 and Lyn within the raft environment during PA infection of the alveolar epithelium. Since Lyn is located inside the membrane, we propose that receptors like TLR2 may play a role in activating Lyn upon PA infection. Roles of TLR in PA infection appear to be complex, and a recent report indicates that TLR2- and TLR4-knockout mice have no defect in their response to PA 31. Also, TLR5 was reported to be involved in PA infection through recognition of flagellin 32. Thus there remains a possibility of unknown PA receptors stimulating Src family kinases. The discrepancy in TLR roles may be due, in part, to experiment conditions, and strains of PA.

We showed that Lyn activation is a crucial event for bacterial internalization signal. Lyn activation by PA infection was seen with a corresponding increase in phospho-enolase by in vitro tyrosine kinase assay. Further, cyanogen bromide phospho-peptide mapping revealed that Tyr397 of Lyn is predominantly phosphorylated by PA infection. In vivo studies showed that PA invaded AEC II cells. Similarly, a recent study demonstrated that PA can infect alveolar epithelium, but only discussed type I cells 7. Blocking rafts and Lyn by MβCD or Src inhibitor PP2 was shown to decrease PAO1 adhesion to the alveolar epithelial cells. These results are similar to experiments conducted in cultured alveolar epithelial cells.

To firmly establish the role of Lyn in PA infection, we used DN strategy in A549 cells and Lyn-deficient BMMC to show the important roles of Lyn in PA internalization. The unique in vitro and in vivo approaches, plus the use of Lyn-deficient cells, have demonstrated unequivocally that Lyn is indeed an important player in early PA infection of respiratory epithelial cells. Despite being phagocytic cells, the experiment in Lyn–/– BMMC indirectly confirmed that Lyn has critical roles involved in PA internalization regardless of cell types. We have also found that Lyn-mediated PA internalization is through actin polymerization in macrophages (manuscript in preparation). We have not tested this in Lyn–/– mice, but it should be noted that Lyn–/– mice may exhibit lupus-like symptoms and may not be a reliable tool for studying infection 33.

Lyn is highly expressed in airway or alveolar epithelial cells, suggesting an important role in these cells 34, 35. Lyn can induce apoptosis in cancer cell lines upon stimulation with certain drugs and respond to oxidative stress 36. In different cell types and various conditions, Lyn may play a role in regulating apoptosis by collaborating with other factors. A study indicated that Lyn may negatively regulate genotoxic apoptosis by interacting with GADD34 37. We hypothesized that Lyn may be used by the host to curb PA infection through a regulated immune response and proper cytokine production. Appropriate immune and/or inflammatory responses can benefit the body against pathogens, but an unabated immune response and excessive inflammation may hamper the body's immune mechanisms, resulting in disseminated septicemia. If apoptosis is controlled to a certain degree, it will help limit inflammatory responses. In contrast, excessive apoptosis by heavy loads of bacteria may damage organs, thereby hampering the host immune system. How apoptosis is regulated by Lyn during PA infection is unclear and is worth studying.

Apoptosis involving PA and its role in pathogenesis are controversial. We found that apoptosis occurred around 2 h, but some reports showed that apoptosis did not occur at much later times in airway epithelial cells 38. This may be due to the different measuring methods, different cell types, and the ratios of cells to bacteria. We used 1:10 ratio of cells to bacteria, while another study used 1:1 ratio 38. Another issue is that apoptosis may be initiated without bacterial internalization 38, as a recent report indicates that exotoxin A of PA can cause apoptosis 39. Our study indicates that blocking bacterial internalization significantly reduced apoptosis, indicating that internalization of the bacteria may be related to apoptotic mechanisms that require further study. We also speculate that apoptosis may depend on the time of infection. Apoptosis with early infection may need the bacterial internalization, as this was supported by other reports 20, 30.

Also, it is still unclear whether internalization is necessary for PA to induce inflammatory response, since previous reports demonstrated that adhesion may be directly related to production of inflammatory cytokines 40, 41. Our study indicates that internalization may be required for increased production in inflammatory cytokine, since blocking PA entry to cells did correlate with reduced cytokine release, consistent with some others’ reports 20, 30. Gulbins and colleagues 42 showed that adhesion is necessary for PA-induced activation of mitochondrial damage and stress-activated protein kinases.

In conclusion, we have demonstrated that the Src tyrosine kinase Lyn plays a major role in PA infection through raft-mediated mechanisms. The data are confirmed by suppressing the key proteins Lyn and TLR2 with DN transfection. We also revealed that Lyn is activated at its key tyrosine site (Tyr397) during PA infection. Our studies represent a new mechanism that the typical immuno-inhibitor Lyn is critically linked to PA infection of alveolar epithelial cells, which may imply a role in controlling PA infection and may have potential use for designing therapeutic strategies.

Materials and Methods

Cells

Human alveolar epithelial A549 cells were purchased from American Type Culture Collection (Manassas, VA). Rat AEC II cells were isolated from female Sprague-Dawley rats (Harlan, Indianapolis, IN) as previously described 4345. The cells were isolated using IgG panning method 43 and cultured on plates or chamber slides in DMEM (Cellgro Mediatech Inc., Herndon, VA) with 10% newborn bovine serum (Highclone, Logan, UT) 46. Phenotypes of epithelial cells were confirmed by Tannic staining, lamellar bodies (TEM), and staining with antibodies against pan-cytokeratin, and SP-C (Sigma-Aldrich, St. Lous, MO) prior to and during experiments 47.

Bacterial strains

PA strain PAO1 WT was a gift from Stephen Lory (Harvard Medical School, Boston). PA strain ATCC27853 was purchased from American Type Culture Collection. GFP-PAO1 and anti-PAO1 rabbit serum were kindly provided by G. Pier (Harvard Medical School) 48.

Infection experiments

The bacteria were grown overnight in LB broth at 37°C with vigorous shaking. The next day the bacteria were pelleted by centrifugation at 5000 × g, resuspended in 10 mL of fresh LB broth, and allowed to grow until the mid-logarithmic phase. OD 600 nm was measured and density was adjusted to ∼0.1 OD (1×108 cells/mL). Epithelial cells were washed once with PBS after fasting overnight in serum-free DMEM. Epithelial cells were infected with PA in multiplicity of infection (MOI) with a 1:10 ratio of cells to bacteria.

Immunocytochemistry

Epithelial cells were grown in plain or collagen (human type IV; Sigma-Aldrich)-coated Lab-Tek 8-well chamber slides (Nalge Nunc International, Rochester, NY). After infection or other treatments, cells were fixed in 2% paraformaldehyde, permeabilized with 0.1% NP-40 (Sigma-Aldrich) in PBS and blocked with blocking buffer (PBS containing 1% newborn bovine serum and 0.1% NP-40) for 30 min. Cells were incubated with primary antibodies at 1:500 dilution in blocking buffer for 1 h and washed three times. After incubation with appropriate fluorophore-conjugated secondary antibodies, images were captured by Carl-Zeiss LSM 510 Meta confocal microscope (Carl Zeiss MicroImaging Inc., Thornwood, NY). Internalization was quantitated by randomly selecting five to seven fields on each coverslip containing 15–20 cells, and the means calculated. Thus 100 cells were scored for each condition. Differential interference contrast pictures were taken simultaneously. Images were processed using the software provided by the manufacturer. Bacteria were stained with DiI (Molecular Probes Inc., Eugene, OR), a fluorescent chemical with emission spectra in red.

Bacterial adherence and internalization assays

Epithelial cells were grown to sub-confluent monolayer and serum-deprived overnight. Adherence was attained with MOI cell-to-bacteria ratio of 1:10 for 30 min (while internalization at 37°C for 30 min) and washed three times with DMEM to remove non-adherent bacteria. For internalization the cells were then incubated for 1 h with polymyxin B (100 µg/mL; Sigma-Aldrich) in serum-free DMEM to kill extracellular bacteria (adhesion control was performed without polymyxin B) as this antibiotic cannot cross the cell membrane. The cells were washed twice with PBS and lysed with 5 mg/mL saponin in LB broth (Sigma-Aldrich) for 10 min at 37°C. Intracellular bacteria survived this procedure and grew in colonies when plated on tryptic soy agar plates. Bacterial colonies were counted after incubation for 24 h at 37°C.

Cell proliferation assay (MTT)

Cell Titer 96 Non-Radioactive Cell Proliferation Assay kits (Promega, Madison, WI) were used for measuring cell survival after infection experiments. This assay measures the cellular conversion of a tetrazolium salt into a blue formazan product. An equal number of cells (5000 cells/well) was loaded in all wells for infection or other treatments and assayed according to the manufacturer's instructions.

Immunoprecipitation and immunoblotting for Lyn quantification

Epithelial cells were grown to sub-confluent monolayer in serum-free DMEM with PA for 0, 10, 30, and 60 min. Cells were then washed once with PBS and lysed with non-denaturing lysis buffer (25 mM MES, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM sodium fluoride, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM sodium orthovanadate and 1 mM DTT) and freshly prepared protease inhibitor cocktail (Calbiochem, San Diego, CA) for 15 min on ice, and homogenized by sonication at 40% amplitude three times for 10 s with intermittent cooling on ice. The lysate was then centrifuged at 3000 × g for 10 min at 4°C to remove nuclei. Supernatants of each sample were incubated with 1:500 dilution of Lyn (44) antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA) overnight at 4°C in an axial rotator. Then, 50 µL of 50% slurry of protein A/G agarose beads (Upstate, Charlottesville, VA) in PBS were added to 500 µL of post-nuclear lysate, incubated for 2 h at 4°C in the axial rotator and centrifuged briefly (200 × g) to pellet the beads. The beads were washed with 1 mL lysis buffer three times. Finally, Laemmli sample buffer was added and the samples were boiled to dissociate proteins from the beads, and resolved by SDS PAGE.

The proteins were transferred to an Immobilon PSQ membrane (Millipore Corp., Billerica, MA) by a wet transfer method and the transfer efficiency was checked using ponceau staining. Membranes were blocked 2 h with 5% BSA blocking buffer and incubated with phospho-Lyn antibody (Tyr508), phospho-Src antibody, or phospho-tyrosine (PY20) antibodies (Cell Signaling Technology, Beverly, MA) at 1:1000 dilution in 5% BSA blocking buffer overnight at 4°C shaking. The membranes were washed three times and then incubated with HRP-conjugated anti-rabbit or anti-mouse secondary antibodies at 1:2000 dilution, and developed with enhanced chemiluminescence reagents (Pierce, Rockford, IL) 49.

For co-IP studies, the same protocol was observed with minor changes. The original post-nuclear lysate was split into two portions to probe for complementary antibodies. To detect Lyn and TLR2 interactions, one portion of the precipitate was incubated with Lyn antibodies and the other with TLR2 antibodies. For gel loading consistency, total Lyn and TLR2 present in each IP fraction was also checked by immunoblotting the same membrane with respective antibodies.

In vitro kinase assay

In vitro kinase assay was done as described by Gould and Hunter 50. Briefly, anti-Lyn immunoprecipitates were washed twice with kinase assay buffer (100 mM HEPES, pH 7.4, 5 mM MnCl2, and 500 µM Na3VO4) and resuspended in 25 µL of kinase buffer containing 2.5 µg of acid-denatured enolase as an exogenous substrate. The reaction was initiated by the addition of [µ-32P]ATP (10 μCi), incubated for 30 min at 30°C and the reaction was terminated by the addition of 2× Laemmli SDS sample buffer. 32P-containing proteins were visualized by autoradiography.

Lyn phospho-peptide mapping

Lyn was immunopurified using 35 µL of anti-Lyn conjugated to protein A/G agarose beads (Santa Cruz Biotechnology) and 1×108–2×108 cell equivalents of mammalian cells lysed in the lysis buffer as above. Lyn was eluted from the anti-Lyn beads by incubation with 0.1 M glycine HCl, pH 2.5. Next, the sample was dried using vacuum centrifugation (Thermo Savant, Holbrook, NY) and exchanged to 70% formic acid containing 100 mg/mL cyanogen bromide (ICN Biomedicals Inc., Aurora, OH) to digest the protein overnight at room temperature in the dark. The following day, 500 µL of H2O was added to digests, and this mixture was then evaporated to dryness in the vacuum centrifuge. This wash step was repeated three times, and dried samples were solubilized in 50 µL of 1× sample buffer. The sample was run on a 16.5% acrylamide tricine gel and then transferred to an Immobilon PSQ membrane (Millipore Corp.). Membranes were subsequently blocked with 5% BSA in TBST and then probed with phospho-tyrosine (PY20) antibodies, washed, and detected by chemiluminescence as described above.

Lipid raft isolation

Lipid rafts were isolated by sucrose density gradient ultracentrifugation as previously described 20, 51, with modifications. The cells were harvested and washed twice with ice-cold PBS, and 5×106 cells were lysed with ice-cold 500 µL of 1% Triton X-100 lysis buffer as above. After 10 min of incubation on ice, the cells were homogenized. The lysate was mixed with 80% sucrose buffer to a final 40% concentration. A gradient was formed by successive addition of 40% sucrose-lysate mix to the bottom, followed by 30% and 5% sucrose buffer. Tubes were centrifuged at 35 000 rpm for 12 h at 4°C using an SW40 rotor (Beckman Instruments, Palo Alto, CA). Gradients were fractionated into 1.3-mL aliquots withdrawn from the top of the tube, and fractions 3 to 5 (15–30% sucrose) contained detergent-insoluble lipid rafts. Raft fractions were assessed by immunoblots as above.

Cholesterol depletion

Before extraction of cholesterol, the cells were washed twice with RPMI medium. The cells (2×107) were incubated with 1-10 mM MβCD in DMEM supplemented with 0.1% BSA (Sigma-Aldrich) for 30 min at 37°C. As controls, the cells were either exposed to medium alone or were treated with MβCD and subsequently reconstituted with 400 µg/mL cholesterol (Sigma-Aldrich) in DMEM for 1 h at 37°C. After incubation, the cells were washed twice with ice-cold DMEM before use.

Transfection

pcDNA Lyn constructs (WT, DN – LynK275D, constitutive – LynY508F) were kindly provided by Tomas Smithgall (University of Pittsburgh). Lyn-truncated construct GFP-Lyn24 was kindly provided by Anthony DeFranco (University of California, San Francisco). TLR2 DN construct (P681H) was kindly provided by J. D. Li (University of Rochester), and TLR2 WT construct was kindly provided by S. Hong (Indiana University Medical Center, Indianapolis). Transfection of these constructs was all done using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA) following the manufacturer's protocol.

Apoptosis assays

Apoptosis of cultured cells was determined 2 h after infection using FITC-labeled annexin V/propidium iodide staining as previously described. DNA fragmentation assay and TUNEL assays (Roche Applied Science, Indianapolis, IN) were done following the manufacturer's protocol.

Lung sections

Female Sprague-Dawley rats (Harlan) were used to study activation of Lyn by PA. All treatments and infections were done via a catheter through the trachea and washed with PBS. The lungs were then perfused and inflated with OCT. Cryosections were made using cryostat (Tissue-Tek, Elkhart, IN). Tissue sections were fixed with ice-cold acetone for 10 min. Immunofluorescence staining was performed as above. Three rats were used per group and the experiments repeated three times.

Statistical analysis

All experiments were performed in triplicates and repeated three times. Data were presented as percent changes compared to mean ± SD from the three independent experiments. All error bars represent SD. Group means were compared by Student's t- test or one-way ANOVA followed by post-hoc analysis, using Sigmastat software, and difference was accepted at p<0.05.

Acknowledgements

This work is supported by North Dakota Biomedical Research Infrastructure Network (NIH NCRR P20-164-71), American Heart Association Scientist Development Grant (National Office), NIEHS (ES1469) and North Dakota Experimental Program to Stimulate Competitive Research (NSF) to M.W. We thank Dr. E. Carlson, T. Casavan and D. Laturnus for help with imaging.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

Ancillary