Identification of a Gene That Affects the Efficiency of Host Cell Infection by Legionella pneumophila in a Temperature-Dependent Fashion

Strains and plasmids used in this study are described in Table 1. The streptomycin-resistant variant of L. pneumophila strain AA100 (17), which has been shown to be virulent in both in vitro and in vivo models of infection (30), was used. L. pneumophila strains were passaged no more than twice in the laboratory in order to prevent loss of virulence, and the strains were grown on buffered charcoal yeast extract (BCYE) agar (15) for 3 days at either 30 or 37°C in 5% CO₂. Escherichia coli strains were grown in Luria-Bertani medium (Difco, Detroit, Mich.) at 37°C. When appropriate, kanamycin (Sigma) was added at a concentration of 25 μg/ml. The E. coli strain ψec47 has been described previously (10) and was used to propagate plasmids containing an R6K origin of replication. DNA manipulations were done essentially as described previously (37). XL-1 Blue and ψec47 electroporation-competent cells were prepared and frozen as described previously (10, 14).

HEp-2 cells (ATCC CCL23), established from a human epidermoid carcinoma, were grown in RPMI 1640 supplemented with 5% heat-inactivated fetal bovine serum (Gibco) and 2 mM l-glutamine at 37°C in 5% CO₂. HL-60 (ATCC CCL240) monocyte/macrophage cells were grown at 37°C in 5% CO₂ in RPMI 1640 supplemented with 10% fetal bovine serum and 2 mM l-glutamine.

An L. pneumophila total genomic DNA library was constructed in the vector pWKS130 (43). For this library, L. pneumophila genomic DNA was isolated as described previously (10), partially digested with Sau3AI to obtain fragments ranging in length from 0.5 to 2.5 kbp, and ligated into the BamHI site of pWKS130. Plasmid DNA from pools of greater than 5,000 E. coli colonies was then separately transformed into L. pneumophila by electroporation as described previously (10, 14). Six pools of greater than 5,000 L. pneumophila clones (total of 30,000 clones) were enriched for enhanced host cell infection variants using entry assays.

The ability of L. pneumophila to infect host cells was evaluated using standard entry assays, which were carried out as described previously (10). Briefly, 10⁶ HL-60 cells were differentiated in 24-well tissue culture dishes with 100 ng of phorbol 12-myristate 13-acetate (PMA; Sigma)/ml for 48 h prior to use. Nonadherent cells were removed by washing with phosphate-buffered saline (PBS), and 1 ml of culture medium was added. HEp-2 cells were seeded in 24-well dishes at a concentration of 1.5 × 10⁵ cells/ml 18 to 24 h before use. A ratio of 10 bacteria per cell was added to each well and incubated for 30 min with monocytic cells or for 90 min with epithelial cells. All infections were carried out at 37°C. Cells were washed two times with PBS, and fresh medium containing 100 μg of gentamicin/ml was added to kill extracellular bacteria. After 2 h at 37°C, the cells were washed with PBS and lysed with sterile water, and dilutions were plated for CFU determinations. Entry levels were determined by calculating the percentage of the inoculum that became gentamicin resistant over the course of the assay [i.e., percent entry = 100 × (CFU gentamicin resistant/CFU of inoculum)]. To correct for variations in the levels of uptake between experiments, entry is reported relative to that of the wild-type strain (i.e., relative entry = percent entry of test strain/percent entry of wild type).

Pools of clones from the expression library were enriched for enhanced host cell infection variants in modified entry assays. Entry assays for enrichment of enhanced host cell infection variants were carried out in HEp-2 cells in the same manner as standard entry assays, except that bacteria were coincubated with cells for 5 min prior to the addition of gentamicin. Individual colonies that resulted from the subsequent plating for CFU were then screened individually for an enhanced entry phenotype relative to wild type in standard entry assays.

Adherence assays were carried out by two methods, immediate assays (7-9) and formaldehyde assays (9, 16, 21), both of which have been described in detail previously. For immediate assays, HEp-2 and HL-60 cells were seeded and prepared as described for entry assays. After adding the bacteria at a ratio of 10 bacteria per cell, the medium was gently mixed by rocking back and forth, immediately washed five times with PBS to remove nonadherent bacteria, and then lysed by incubation for 10 min in 1 ml of water followed by vigorous pipetting. After lysis, the number of cell-associated bacteria was determined by plating for CFU on BCYE. Adherence assays on formaldehyde-fixed cells were carried out in the same manner except that the cells were fixed in 3.7% formaldehyde for 10 min, washed three times with PBS, and suspended in RPMI prior to addition of the bacteria. The bacteria were coincubated with the cells for 30 min (HL-60) or 90 min (HEp-2). Adherence levels were determined by calculating the percentage of the inoculum that became cell associated over the course of the assay [i.e., percent adherence = 100 × (CFU cell associated/CFU in inoculum)]. In order to correct for variation in levels of uptake between experiments, adherence is reported relative to that of wild-type strain AA100 at 37°C (i.e., relative adherence = percent adherence of test strain/percent adherence of AA100 at 37°C).

The growth kinetics of 30°C- and 37°C-grown L. pneumophila in PMA-differentiated HL-60 cells at 37°C was determined as described previously (7, 8). Briefly, bacteria were added to a monolayer of 10⁶ cells/well of HL-60 cells in 24-well tissue culture dishes at a multiplicity of infection of 0.1 and incubated for 30 min at 37°C. Extracellular bacteria were killed by gentamicin treatment for 2 h, the cells were washed twice, and one set of triplicate wells was lysed with water and another set of wells was incubated in fresh medium at 37°C for 72 h before lysis in water. Dilutions of the lysate were plated to determine the CFU present. Survival is expressed as the portion of CFU present at 72 h compared to time zero (T₀ = 2.5 h), i.e., mean (CFU T₇₂/ T₀) × 100.

Adherence and entry levels were confirmed through examination of stained coverslips by fluorescence microscopy. We differentially stained intracellular and extracellular bacteria to allow discrimination between adherence and uptake. Intracellular and extracellular bacteria were differentially stained by first prestaining the bacteria prior to infection of cells followed by antibody detection of extracellular bacteria (without permeabilizing the cells) with polyclonal anti-L. pneumophila antibody. The rabbit polyclonal anti-L. pneumophila antibody was prepared and tested as described previously (8). This procedure allows intracellular bacteria to appear green and extracellular bacteria to stain both green and red by fluorescence microscopy. Basically, L. pneumophila cells were prestained with Oregon green (OG) prior to infection in a similar manner to that previously described using fluorescein (18, 47). The bacteria were suspended in PBS plus 5 μg of OG/ml, incubated at 30 or 37°C (depending on the prior growth temperature) for 30 min, and washed five times to remove unbound OG. Stained bacteria were then used to infect cells in the same manner as for standard entry assays, except that no gentamicin incubation was used. After washing, the monolayers were fixed with 2% paraformaldehyde in PBS for 1 h at 4°C followed by staining with primary and secondary antibody as described previously (8), except that the secondary antibody was conjugated with R-phycoerythrin (Becton Dickinson). Dual images of multiple fields were captured using an Optronics charge-coupled device video camera. The OG staining procedure did not affect bacterial viability (data not shown) or the ability of L. pneumophila to adhere to or enter into HL-60 or HEp-2 cells (data not shown). Viability of the labeled bacteria was confirmed using the LIVE-DEAD assay (Molecular Probes, Eugene, Oreg.) and by plating dilutions for CFU on BCYE agar (Difco).

DNA sequencing was performed as described previously (10) using a BigDye Terminator (Applied Biosystems) cycle sequencing apparatus and subsequent analysis on an ABI 310 automated sequencing apparatus (Applied Biosystems). Sequence analysis and assembly were carried out using Gene Construction Kit 2 (Textco), searches for homologous sequences were done using BLAST (1), and protein alignments were made using MegAlign (DNASTAR).

An in-frame deletion in the coding region of lvhB2 was constructed by overlapping PCR as described previously (10). In this construct, the entire LvhB2 coding region was deleted with the exception of the ATG and TGA codons. The resulting PCR product has approximately 1 kbp of DNA flanking the lvhB2 start and stop codons and was cloned into the NotI site of pJDC15 (10). The resulting plasmid was transformed into L. pneumophila, and allelic exchange selected for sucrose and kanamycin as negative and positive selectable markers, respectively, as described previously (10). The presence of the appropriate deletion in the resulting lvhB2 mutant strain (ΔlvhB2) was confirmed by Southern and PCR analyses (data not shown). The complementing plasmid for lvhB2, pJDC38, was constructed using restriction digestion of L34 to delete all potential coding regions other than lvhB2. The resulting plasmid was then transformed into ΔlvhB2 and tested for complementation of the entry phenotype.

All in vitro experiments were carried out in triplicate and repeated at least three times. The significance of the results was analyzed using analysis of variance. P values of <0.05 were considered significant.

The GenBank accession number for the nucleotide sequence of lvhB2 and the surrounding chromosomal region from L. pneumophila strain AA100 is AF410854 .

Our previous studies found that a number of proteins are differentially regulated under growth conditions that enhance entry by L. pneumophila (8). We wished to develop a strategy that would allow identification of L. pneumophila genes specifically involved in interactions with host cells rather than other physiological processes that are expressed differently because of growth conditions. Since several genes involved in the efficiency of host cell infection appear to be repressed under standard laboratory conditions, regulatory mutations that increase their expression could result in an enhanced host cell infection phenotype. It is possible to enrich for enhanced host cell infection mutants from a population of bacteria by using entry assays that are based on gentamicin protection, where only those bacteria that enter host cells survive. Our investigators recently used a similar strategy to identify three L. pneumophila loci that are involved in entry (10). In the present study, we constructed a total L. pneumophila genomic DNA library in the low-copy-number vector pWKS130. This vector allows constitutive expression of L. pneumophila genomic DNA fragments from the trp-lac promoter (P_tac). This library was then enriched for L. pneumophila clones that display an enhanced host cell infection phenotype by using entry assays. We optimized these assays for time of coincubation, number of host cells used, host cell type, and multiplicity of infection to result in an uptake of 1 to 10 wild-type bacteria/well (data not shown).

We examined the phenotype of individual clones from the L. pneumophila expression library before and after enrichment. Prior to enrichment, 40 random clones were screened in standard entry assays and compared to wild-type L. pneumophila. None of the preenrichment clones displayed higher levels of entry into host cells than wild type (data not shown). After enrichment, two clones (5%), designated L1 and L34, out of 35 displayed at least twofold-enhanced entry into HEp-2 cells (data not shown) compared to that into the wild type (P < 0.01). In order to characterize the positive clones identified with this approach, plasmids were isolated from them and retransformed into wild-type L. pneumophila cells and the host cell infection phenotype of the resulting transformants was examined. All of the resulting retransformed clones displayed the enhanced host cell infection phenotype of the original clone (data not shown). These data suggest that the plasmids carrying L. pneumophila chromosomal fragments, rather than secondary mutations elsewhere in the chromosome, are responsible for the observed phenotype.

Restriction analysis of the clones that display an enhanced host cell infection phenotype, L1 and L34, with PstI, KpnI, and NheI demonstrated that they have identical physical maps (data not shown). A series of deletions were constructed in L34 to dissect the region responsible for the enhanced host cell infection phenotype (Fig. 1). We found that an internal PstI-NheI fragment is needed for the enhanced host cell infection phenotype conferred by the L34 plasmid (Fig. 2). In addition, neither portion of the L34 plasmid insert to the left (pJDC53) or right (pJDC52) of the KpnI restriction site conferred enhanced host cell infection alone, suggesting that the KpnI restriction site interrupts an open reading frame (ORF) that is necessary for the phenotype conferred by this plasmid. The L. pneumophila chromosomal DNA insert in the L34 plasmid was sequenced and found to contain an ORF encoding a putative protein of 96 amino acids in length that is interrupted by digestion with KpnI. This ORF is in the proper orientation to allow expression from the pWKS130 P_tac promoter. We initially designated this gene enhD to indicate that it has the ability to confer enhanced host cell infection. This gene has recently been sequenced in another strain of L. pneumophila, where it was designated lvhB2 and part of a large putative operon involved in type IV secretion (39). Although the lvh operon appears to play a role in conjugation, its role in virulence has not been investigated.

The protein encoded by lvhB2 is similar to the T-pilin subunit protein encoded by virB2 of Agrobacterium. This gene is part of a growing family of pilin genes that are involved in conjugation and the secretion of virulence proteins (26). Alignment of the similar proteins revealed that they contain several conserved domains (Fig. 3) and a signal peptidase cleavage site (26). Interestingly, LvhB2 has the greatest similarity to the Bordetella pertussis VirB2 homologue (25% identity, 70% similarity over 96 amino acids), which is involved in secretion of virulence proteins (31, 46). The next most similar proteins are from Brucella and Bartonella, both of which have approximately 20% identity with LvhB2. These data suggesting that lvhB2 may play a role in the efficiency of host cell infection are the first to infer a function other than conjugation for lvhB2 and indicate that further investigation of its role in pathogenesis is warranted.

In order to investigate whether lvhB2 plays an important role in the ability of L. pneumophila to infect host cells, we compared the interactions of an lvhB2 mutant and complemented strain with that of the wild type in the human monocyte/macrophage cell line HL-60. Since the lvhB2 gene is part of a potentially large operon, we constructed an in-frame deletion (ΔlvhB2) in this gene to prevent polar effects on downstream genes (Fig. 1). This construct carries a complete in-frame deletion in the LvhB2 coding region, producing an ORF that encodes only the translational start and stop codons for this protein. An L. pneumophila strain carrying this mutation was constructed by allelic exchange using sucrose as a counterselectable marker and confirmed by Southern and PCR analyses (data not shown). We found that the lvhB2 mutant does not display a significant defect in the ability to infect human monocytic or epithelial cell lines when grown at 37°C (Fig. 4A and B). This observation suggests that either this gene plays no role in host cell infection or it is not expressed well at higher temperatures under standard laboratory growth conditions unless it is expressed from an exogenous promoter. This is the case for the L34 construct, where the gene is downstream of the P_tac promoter.

T-pilus production and T-DNA transfer in Agrobacterium are optimal at lower temperatures (19, 20). Thus, we examined the effects of bacterial growth temperature on lvhB2 function. When the bacteria were grown at 30°C, the L. pneumophila lvhB2 mutant was ∼100-fold reduced in entry compared to wild type (Fig. 4C and D). This defect in the ability to infect host cells is complemented by the presence of vector containing only the L. pneumophila lvhB2 gene (pJDC38). Since all host cell infections were carried out at 37°C, this difference is likely due to changes in expression of lvhB2 at lower temperatures.

In order to confirm the difference in the ability of the lvhB2 mutant to infect host cells compared to wild-type L. pneumophila, we examined these infections by fluorescence microscopy. We used differential staining with fluorescent markers to determine whether the bacteria were intracellular or extracellular. All bacteria were first prestained with OG, and then only extracellular bacteria were labeled with a secondary antibody conjugated to R-phycoerythrin. This technique allows determination of the total number of cell-associated bacteria and whether they are adherent or intracellular and does not affect bacterial viability or their ability to enter into or adhere to host cells (data not shown). Similar to gentamicin protection assays, there is no apparent difference between the wild type and the ΔlvhB2 L. pneumophila mutant when the bacteria are grown at 37°C, but when grown at 30°C the ΔlvhB2 mutant displays obviously lower (P < 0.01) levels of cell association (Fig. 5). Since the staining technique used for microscopy should not be affected by bacterial viability, as would the CFU-based gentamicin protection assays, rapid killing by host cells following uptake is not solely responsible for the differences observed. The higher percentage of intracellular bacteria observed at 30°C with the monocytic cell line compared to the epithelial cell line is most likely the result of the constitutively phagocytic nature of monocytic cells. Based both on standard gentamicin protection assays for entry and microscopy, we found a significant defect in the ability of the L. pneumophila lvhB2 mutant to infect monocytic and epithelial cell lines.

It is possible that the role of lvhB2 in entry is primarily the result of a defect in the ability to adhere to host cells rather than trigger uptake. In order to explore this further, we examined the adherence of ΔlvhB2 and wild-type L. pneumophila to epithelial and monocytic cell lines after growth of the bacteria at both 37 and 30°C (Fig. 6). To completely rule out the possibility that entry and/or intracellular killing were solely responsible for the data obtained, both standard immediate adherence assays and formaldehyde adherence assays, where internalization cannot occur, were used. Both assay methods have been previously shown to allow accurate measurement of adherence to host cells by L. pneumophila (9). Both the wild type and ΔlvhB2 mutant adhere better to epithelial and monocytic cell lines at 30°C than at 37°C (P < 0.01). Furthermore, the ΔlvhB2 mutant displays a defect in adherence at 30°C (P < 0.01) but adheres at the same levels as wild type at 37°C.

Our laboratory's previous studies have found that the initial interactions with host cells, such as adherence and entry, can correlate with the subsequent ability of L. pneumophila to survive and replicate (7, 9). Since the L. pneumophila lvhB2 gene affects the efficiency of host cell infection, it may also impact the resulting intracellular events. We examined replication of the lvhB2 mutant in HL-60 cells (Fig. 7) and found that although there is a moderate effect on intracellular replication (P < 0.01 after 72 h) when the bacteria are grown at 37°C prior to infection, there is a much greater impact when the bacteria are grown at 30°C (P < 0.001). The generation time in HL-60 cells of wild-type L. pneumophila grown at both 30 and 37°C is between 5 and 6 h, whereas the generation time of the lvhB2 mutant grown at 30°C is between 8 and 9 h and slightly greater than 6 h when grown at 37°C. These observations suggest that lvhB2 is important for intracellular survival and replication when the bacteria are grown at 30°C and less so when they are grown at 37°C.