We first examined intracellular growth of L. pneumophila strain AA100jm in A549 and NCI-H292 cells by 72-h continuous cultures. The colony forming unit (CFU) per well of AA100jm growing in A549 and NCI-H292 cell cultures began to increase after 24 h and then increased time-dependently (Figure 1A). However, the CFU of the avirulent mutant strain with a knockout in dotO, encoding a protein essential for the type IV secretion system, did not increase during the 72-h period (Figure 1B). Furthermore, the multiplication of flaA mutant was slightly inhibited in A549 cells compared with the wild-type Corby (Figure 1C). This observation suggests that flagellin plays a role in uptake into non-phagocytic cells.

Monolayers of A549 and NCI-H292 cells were infected with L. pneumophila strain AA100jm for up to 24 h. Total cellular RNA was isolated from these cells at 1, 2, 4, 8, 16 and 24 h after the infection and IL-8 gene expression was analyzed by reverse transcription-polymerase chain reaction (RT-PCR). IL-8 mRNA expression increased immediately after the infection and its signal became gradually stronger in A549 and NCI-H292 cells until 24 h after infection (Figure 2). In another series of experiments, in which A549 cells were infected with AA100jm at different concentrations for 8 h (Figure 3A), AA100jm induced dose-dependent expression of IL-8 mRNA. Next, we examined the correlation between IL-8 expression level and the virulence of L. pneumophila. As shown in Figure 2, IL-8 mRNA expression was induced immediately after the infection and reached a peak level at 4 h, but became gradually weaker from 8 to 24 h after infection with the avirulent dotO mutant. Moreover, Legionella flagellin seemed to be involved in IL-8 expression, since a flaA knockout mutant, defective in flagellin production, failed to induce IL-8 mRNA immediately after the infection (Figure 2). However, its signal was induced by the flaA mutant at 16 h post-infection.

To determine the correlation between IL-8 expression levels and bacterial proteins of L. pneumophila, heat-killed AA100jm was used to infect A549 and NCI-H292 cells at a multiplicity of infection (MOI) of 100. IL-8 mRNA expression levels at 6 and 24 h in both cell lines infected with the heat-killed strain were significantly lower than those of cell lines infected with the live strain (Figure 3B). Furthermore, paraformaldehyde-fixed L. pneumophila showed no induction of epithelial IL-8 gene expression (data not shown).

We used enzyme-linked immuno-sorbent assay (ELISA) to determine IL-8 protein levels in culture supernatants of A549 and NCI-H292 cells at 24 h after infection with either the parental strain AA100jm or dotO mutant strain and either the parental strain Corby or flaA mutant strain at an MOI of 100. L. pneumophila did not alter the cell viability within an MOI of 100 and time frame tested (data not shown). We found IL-8 induction by AA100jm and Corby. On the other hand, the production levels of IL-8 by A549 and NCI-H292 cells infected with the dotO and flaA mutant strains were significantly less than those of cells infected with the wild-type strains (Figure 4).

To delineate the mechanism by which L. pneumophila induces IL-8 gene transcription, we identified L. pneumophila-responsive promoter elements in the IL-8 promoter. This was achieved by transfecting A549 cells with various plasmid constructs containing the luciferase reporter gene driven by the IL-8 promoter (Figure 5A). Twenty-four hours post-transfection, cells were infected with L. pneumophila strain AA100jm. The 5' region 1,481 bp full-length promoter was reproducibly activated by L. pneumophila infection in an MOI-dependent manner (Figure 5B). These results indicate that L. pneumophila induces IL-8 expression in A549 cells at transcriptional level. Next, we used a deletion analysis approach to identify essential promoter elements for transcriptional upregulation following a stimulus. High induction levels were observed with a reporter construct containing IL-8 5'-flanking sequence starting with position -1,481 to position -133. Deletion of sequences upstream of position -50 abolished inducibility by L. pneumophila infection (Figure 5C). The IL-8 gene fragment spanning positions -133 to -50 bp contains three prominent DNA-protein interaction sites for the transcription factors AP-1, NF-IL-6 and nuclear factor-κB (NF-κB) (Figure 5A). This maps the region from -133 to -50 bp as a L. pneumophila-responsive region, which is likely to contain individual L. pneumophila-responsive regulatory elements.

Previous studies have shown that the AP-1 and NF-IL-6 sites act cooperatively with NF-κB to induce IL-8 gene transcription 151617. To identify the cis-acting elements in the -133 to -50 bp region of the IL-8 promoter, which served as a L. pneumophila-responsive regulatory element, we prepared and tested site-directed mutant constructs (Figure 6A). The results shown in Figure 6B indicate that mutation in the NF-κB site (NF-κB mut-luc) resulted in a significant reduction of L. pneumophila-induced IL-8 production. Mutation of the AP-1 site (AP-1 mut-luc) decreased basal activity but had little effect on L. pneumophila-induced luciferase activity (generated a 11.8-fold stimulation). However, mutation of the NF-IL-6 site (NF-IL-6 mut-luc) did not result in reduction of either basal activity or activity induced by L. pneumophila infection. These results indicate that activation of the IL-8 promoter in A549 cells in response to L. pneumophila infection requires an intact binding site for the NF-κB element.

To further determine the exact site that is sufficient for L. pneumophila-induced IL-8 production, we used various constructs (Figure 6C). Each construct contained three tandemly repeated copies of one of the following elements: the NF-κB or NF-IL-6 site from the IL-8 gene or two copies of the AP-1 binding site from the IL-8 gene. Each was linked to the minimal IL-8 promoter spanning -50 to +44 bp and to the luciferase reporter gene. The construct × 3 NF-κB site-luc, which contains three tandem repeats of the NF-κB site from the IL-8 gene, generated a 4.2-fold stimulation by L. pneumophila infection. In contrast, constructs × 2 AP-1 site-luc and × 3 NF-IL-6 site-luc did not exhibit L. pneumophila-induced gene expression in A549 cells (Figure 6D). These results indicate that the NF-κB site of the IL-8 gene is important for activation of L. pneumophila-induced gene expression. In contrast, the AP-1 and NF-IL-6 sites, essential for IL-8 induction by various stimuli in other cell types, are not involved in this form of L. pneumophila-responsiveness.

To investigate the trans-acting factors involved in the activation of IL-8 by L. pneumophila infection, we performed electrophoretic mobility shift assay (EMSA) with a probe spanning positions -83 to -68 bp of the IL-8 promoter, which includes the NF-κB binding site at -80 to -71 bp (Figure 7A, upper panels). L. pneumophila infection induced a single predominant NF-κB binding complex. This binding was specifically competed by an excess of unlabeled NF-κB from the IL-8 gene (Figure 7B, lane 2) but not by an oligonucleotide containing AP-1 from the IL-8 gene (Figure 7B, lane 3). To further characterize the NF-κB complex, we performed supershift experiments using specific antisera. An antibody against NF-κB p50, p65 or c-Rel specifically shifted and diminished the formation of the complex (Figure 7B, lanes 4–6), while an antibody against p52 or RelB had no effect (Figure 7B, lanes 7 and 8). These data suggest that this DNA-protein complex represents a p50–p65 or p50-c-Rel heterodimer product. In contrast, AP-1 binding was constitutively present in nuclei from uninfected cultures, appearing as a single band on EMSA (Figure 7A, lower panels). This DNA-protein complex contained Fra-2 and JunD (data not shown). Nuclear extracts from L. pneumophila-infected A549 cells incubated with the AP-1 probe showed no increase in binding compared with uninfected extracts (Figure 7A, lower panels). Taken together, the results of cold competition and supershift assays indicate that L. pneumophila-induced proteins bind the NF-κB site but not the AP-1 site, which is in agreement with the functional data shown in Figure 6.

In resting cells, NF-κB proteins are predominantly sequestered in the cytoplasm by NF-κB inhibitory proteins IκBα and IκBβ. The activation of NF-κB requires phosphorylation of two conserved serine residues of IκBα (serines 32 and 36) and IκBβ (serines 19 and 23) within their N-terminal domain 18. Phosphorylation leads to ubiquitination and 26S proteasome-mediated degradation of IκBs, thereby releasing NF-κB from the complex followed by its translocation to the nucleus to activate various genes 18. To investigate whether NF-κB activation is mediated through alteration of phosphorylation of IκBα, A549 cells were infected with L. pneumophila and their protein extracts were checked for phosphorylated IκBα expression. L. pneumophila infection induced the phosphorylated IκBα. Kinetic analysis of L. pneumophila-induced degradation of IκBα in A549 cells revealed gradual replacement of IκBα levels (Figure 7C), suggesting that induction of phosphorylation of IκBα by L. pneumophila infection leads to degradation of IκBα protein.

Since activation of the IL-8 promoter by L. pneumophila infection required activation of NF-κB, we blocked NF-κB activation with Bay 11-7082, an inhibitor of IκBα phosphorylation 19, or N-acetyl-Leu-Leu-norleucinaldehyde (LLnL), a proteasome inhibitor 20. These studies served to demonstrate a link between NF-κB activation and upregulation of IL-8 expression in L. pneumophila-infected A549 cells. As shown in Figure 8A, Bay 11-7082 or LLnL markedly inhibited L. pneumophila-induced expression of IL-8 mRNA. We next examined whether L. pneumophila-mediated activation of IL-8 gene expression involves signal transduction components in NF-κB activation. A high-molecular-weight complex, IκB kinase (IKK) complex, which is composed of two catalytic subunits (IKKα and IKKβ) and a regulatory subunit (IKKγ), phosphorylates IκBs 18. Members of the mitogen-activated protein kinase kinase kinase protein kinase family mediate physiological activation of IKK 21. These kinases include NF-κB-inducing kinase (NIK) 22. We tested the ability of dominant interfering mutants of IκBα, IκBβ and IKKγ, and kinase-deficient mutants of IKKα, IKKβ and NIK to inhibit L. pneumophila-mediated activation of the IL-8-driven reporter gene. Expression of these various inhibitory mutants abolished the induction of IL-8 promoter by L. pneumophila infection (Figure 8B). These data show that signaling components involved in the activation of NF-κB are necessary for L. pneumophila activation of IL-8 promoter.

The 90 kDa heat shock protein, hsp90 is a major molecular chaperone of the cell, and appears to have particular significance to cellular regulatory processes. Recent studies have revealed that most of client proteins of hsp90 are protein kinases or transcription factors that play important roles in cellular carcinogenesis 2324. Currently, several groups documented that hsp90 plays a critical role in inflammatory response and its inhibitor resulted in a reduced immune response as indicated by a decrease of proinflammatory mediator production 252627. Furthermore, a recent study indicated that IKKα and IKKβ are clients of hsp90 28. hsp90 has been found to associate stoichiometrically with the IKK complex by binding to the IKKα and IKKβ kinase domains 29. As a possible mechanistic link between L. pneumophila infection and inflammation, we hypothesized the involvement of hsp90. To test this hypothesis, we investigated the effect of L. pneumophila infection on hsp90 and evaluated the effect of hsp90 inhibitor, 17-AAG, on L. pneumophila-induced IL-8 expression. A549 cells constitutively expressed hsp90 protein (Figure 9C), but L. pneumophila did not affect its expression (data not shown). Next, we analyzed whether hsp90 inhibitor could prevent L. pneumophila-induced IL-8 expression. For this purpose, A549 cells were pretreated with 17-allylamino-17-demethoxygeldanamycin (17-AAG) prior to L. pneumophila infection. The significant induction of IL-8 mRNA expression was completely inhibited by pretreatment of 17-AAG (Figure 9A). The finding suggests possible involvement of hsp90 in L. pneumophila-induced IL-8 signaling.

Because 17-AAG seemed to suppress L. pneumophila-induced IL-8 expression and NF-κB was identified as a critical regulator responsible for IL-8 transcription, we tested the direct influence of 17-AAG on L. pneumophila-induced transcriptional activity of NF-κB as a major transcriptional factor of IL-8 using by EMSA. Pretreatment with 17-AAG decreased the retardation of gel mobility through the inhibition of DNA binding activity of NF-κB complex, indicating repression of transcriptional activity of NF-κB (Figure 9B). Since a recent study indicated that IKKα and IKKβ are clients of hsp90 28, we examined the effect of 17-AAG on expression of IKKα and IKKβ. Treatment of A549 cells with 17-AAG reduced the amounts of IKKα and IKKβ proteins but not those of IKKγ and hsp90 (Figure 9C). These results suggest that 17-AAG-mediated hsp90 inhibition depletes IKKα and IKKβ, resulting in inactivation of NF-κB.

Several hsp90 client proteins are degraded by the proteasome following hsp90 inhibition. To examine whether proteasomal degradation was responsible for decreased levels of client proteins after 17-AAG treatment, A549 cells were cultured in a medium containing 17-AAG and the proteasomal inhibitor LLnL. 17-AAG-mediated degradation of IKKα and IKKβ proteins was blocked by LLnL (Figure 9C). In contrast, IKKγ and hsp90 were not destabilized by 17-AAG, and LLnL did not change the levels of IKKγ and hsp90. The reversal of 17-AAG-induced degradation of the IKKα and IKKβ proteins by LLnL suggests that these proteins are subjects to ubiquitin-dependent turnover.

L. pneumophila serogroup 1 strain AA100jm 44 is a spontaneous streptomycin-resistant mutant of strain 130b, which is virulent in guinea pigs, macrophages and amoebae. The avirulent dotO mutant was constructed by random transposon mutagenesis, as described previously 44. This mutation results in severe defects in intracellular growth and evasion of the endocytic pathway 45. The Corby flaA mutant derived from the wild-type Corby is defective in flagellin 46. L. pneumophila strains were grown at 35°C in a humidified incubator on either buffered charcoal-yeast extract-agar medium supplemented with α-ketoglutarate (BCYE-α) or in buffered yeast extract broth supplemented with α-ketoglutarate (BYE-α). Heat-killed bacteria were prepared by heating the bacterial suspension at 100°C for 1 h. Bacterial inactivation was achieved by paraformaldehyde (4%, 15 min followed by three washes in PBS) treatment. Both treated suspensions were confirmed to contain no viable bacteria by plating them on BCYE-α agar. Human A549, a type II alveolar epithelial cell line, and NCI-H292, a tracheal epithelial cell line, were maintained in RPMI 1640 containing 10% heat-inactivated FBS and were grown at 37°C in the presence of 5% CO₂.

LLnL and Bay 11-7082 were purchased from Sigma-Aldrich (St. Louis, MO, USA) and Calbiochem (La Jolla, CA, USA), respectively. 17-AAG was purchased from Alomone Labs (Jerusalem, Israel). Rabbit polyclonal antibodies to IκBα and NF-κB subunits p65, p50, c-Rel, p52 and RelB were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse monoclonal antibodies to hsp90 and actin were purchased from BD Transduction Laboratories (San Jose, CA, USA) and NeoMarkers (Fremont, CA, USA), respectively. Mouse monoclonal antibody to phospho-IκBα (serines 32 and 36) and rabbit polyclonal antibodies to IKKα and IKKβ were purchased from Cell Signaling Technology (Beverly, MA, USA). Rabbit polyclonal antibody to IKKγ was purchased from Sigma-Aldrich.

Unless otherwise indicated, cells were inoculated with either the parental strain AA100jm or the mutant strain dotO and either the parental strain Corby or the mutant strain flaA at an MOI of 100 for 2 h. The extracellular fluid and bacteria were removed by washing three times with PBS, and cells were further incubated with fresh medium for the indicated time periods. In some experiments, heat-killed or paraformaldehyde-fixed bacteria were inoculated in the same manner. The intracellular growth assay of L. pneumophila in lung epithelial cells was performed in 24-well plates for 72 h. Cells were inoculated with either AA100jm or dotO mutant and either Corby or flaA mutant at an MOI of 100, followed by three times washing with PBS and 2 h gentamycin treatment (100 μg/ml) to kill any extracellular bacteria. The cells were washed three times again with PBS and were further incubated with fresh medium. The infected cells and supernatant in each well were harvested at the indicated time intervals by washing the wells three times with sterilized distilled water. These bacterial suspensions were diluted in sterilized water and plated in known volume onto BCYE-α agar. The numbers of CFU in infected monolayers were enumerated at the indicated time points after infection.

The IL-8 content in the culture supernatants was measured by ELISA (BioSource International, Camarillo, CA, USA). A549 cells were cultured in RPMI 1640 supplemented with 10% FBS in 24-well plates. Subconfluent monolayers of A549 cells were infected with L. pneumophila for 24 h. The supernatants were then collected after centrifugation and stored at -80°C until assayed for IL-8 by ELISA. The concentration of IL-8 was determined using a standard curve obtained with recombinant IL-8.

Total cellular RNA was extracted with TRIZOL reagent (Invitrogen, Carlsbad, CA, USA) according to the protocol provided by the manufacturer, and the amount of total RNA was determined by measuring the absorbance at 260 nm. First-strand cDNA was synthesized from 1 μg total cellular RNA in a 10-μl reaction volume using an RNA PCR kit (Takara Bio Inc., Otsu, Japan) with random primers. Thereafter, cDNA was amplified for 30 and 28 cycles for IL-8 and β-actin, respectively. The oligonucleotide primers used were as follows: for IL-8, sense, 5'-ATGACTTCCAAGCTGGCCGTG-3' and antisense, 5'-TTATGAATTCTCAGCCCTCTTCAAAAACTTCTC-3'; and for β-actin, sense, 5'-GTGGGGCGCCCCAGGCACCA-3' and antisense, 5'-CTCCTTAATGTCACGCACGATTTC-3'. Product sizes were 302 bp for IL-8 and 548 bp for β-actin. Cycling conditions were as follows: denaturing at 94°C for 30 s, annealing at 60°C for 30 s and extension at 72°C for 90 s. The PCR products were fractionated on 2% agarose gels and visualized by ethidium bromide staining. The photographs of the gels were scanned, and densitometry measurements of the scanned bands were performed using AlphaEase^® FC software (Alpha Innotech, San Leandro, CA, USA). Data were normalized to β-actin controls.

The dominant negative mutants of IκBα, IκBαΔN, and IκBβ, IκBβΔN (provided by D. W. Ballard, Vanderbilt University School of Medicine, Nashville, TN, USA) are deletion mutants of IκBα and IκBβ lacking the N-terminal 36 amino acids and 23 amino acids, respectively 4748. The dominant negative mutants of IKKα, IKKα (K44M), and IKKβ, IKKβ (K44A), and dominant negative mutant of NIK, NIK (KK429/430AA) (provided by R. Geleziunas, Merk and Company Inc., West Point, PA, USA) and dominant negative mutant of IKKγ, IKKγ (1–305) (provided by K.-T. Jeang, National Institutes of Health, Bethesda, MD, USA), have been described previously 4950. Plasmids containing serial deletions of the 5'-flanking region of the IL-8 gene linked to luciferase expression vectors were constructed from a firefly luciferase expression vector 51. Site-directed mutagenesis of the IL-8 AP-1, NF-IL-6 and NF-κB sites in the -133-luc plasmid was introduced, converting the AP-1 site TGACTCA (-126 to -120 bp) to TatCTCA, the NF-IL-6 site CAGTTGCAAATCGT (-94 to -81 bp) to agcTTGCAAATCGT and the NF-κB site GGAATTTCCT (-80 to -71 bp) to taAcTTTCCT (lower case letters indicate location of base changes). These constructs were designated as AP-1 site-mutated, NF-IL-6 site-mutated and NF-κB site-mutated plasmids, respectively. Two copies of the IL-8 AP-1 binding site, three copies of the NF-IL-6 site or three copies of the IL-8 NF-κB site were inserted upstream of the IL-8 enhancerless core promoter (-50 to +44 bp) linked to the firefly luciferase gene (-50-luc plasmid).

A549 cells were transfected with 40 ng of appropriate reporter and 2 μg of effector plasmids along with 0.4 μg of phRL-TK, an internal control renilla luciferase expression vector (Promega, Madison, WI, USA), using Lipofectamine (Invitrogen). After 24 h, L. pneumophila was infected for 48 h. The cells were washed in PBS and lysed in reporter lysis buffer (Promega). Lysates were assayed for reporter gene activity with the Dual Luciferase Reporter Assay System (Promega). Relative luciferase amounts were normalized to equivalent renilla expression to control for transfection efficiency.

NF-κB and AP-1 binding activities to NF-κB and AP-1 elements were examined by EMSA, as described previously 52. In brief, 5 μg of nuclear extracts were preincubated in a binding buffer containing 1 μg poly(dI-dC)·poly(dI-dC) (Amersham Biosciences, Piscataway, NJ, USA), followed by addition of ³²P-labeled oligonucleotide probes containing NF-κB and AP-1 elements from the IL-8 promoter (approximately 50,000 cpm). These mixtures were incubated for 15 min at room temperature. The DNA protein complexes were separated on a 4% polyacrylamide gel and visualized by autoradiography. To examine the specificity of the NF-κB element probe, we preincubated unlabeled competitor oligonucleotides with nuclear extracts for 15 min before incubation with probes. The probes or competitors used were prepared by annealing the sense and antisense synthetic oligonucleotides as follows: NF-κB of the IL-8 gene, 5'-gatcCGT

GGAATTTCCT

TGACTCA

Cells were lysed in a buffer containing 62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 6% 2-mercaptoethanol and 0.01% bromophenol blue. Equal amounts of protein (20 μg) were subjected to electrophoresis on SDS-polyacrylamide gels followed by transfer to a polyvinylidene difluoride membrane and probing sequentially with the specific antibodies. The bands were visualized with the enhanced chemiluminescence kit (Amersham Biosciences).

Data were expressed as mean ± SD and analyzed by using the Student's t-test. A P value < 0.05 was considered to be significant.