INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Along with Streptococcus pneumoniae and Haemophilus influenzae, Legionella species are some of the most common etiologic agents of bacterial pneumonia (10). Legionella pneumophila and Legionella micdadei are the two most common Legionella species that are associated with the majority of outbreaks of Legionnaires' disease (29). Upon inhalation, both of these Legionella species are phagocytosed by alveolar cells (30). After internalization, both Legionella species replicate within pulmonary cells, and this intracellular replication is believed to be the hallmark of Legionnaires' disease (15, 22). While L. pneumophila is targeted into a rough endoplasmic reticulum (RER)-surrounded phagosome that does not fuse with lysosomes, L. micdadei is targeted into a RER-free phagosome that is thought to fuse to lysosomes in human monocytes (16). Interestingly, within the RER-surrounded phagosome, L. pneumophila undergoes dramatic alterations in gene expression, which are thought to be required for bacterial adaptation to this intracellular niche (2-4, 6, 7). Gene expression by intracellular L. micdadei has never been examined.

Outbreaks of Legionnaires' disease occur through droplet transmission of the bacteria from an environmental water source, predominantly air conditioning cooling towers and showerheads (16). In the environment, Legionella species are ubiquitous and are parasites of at least 13 species of amoebae and ciliated protozoans (16). Invasion by and intracellular replication of L. pneumophila in protozoans play major roles in the infectivity and transmission of the Legionnaires' disease bacteria to humans (16). We have recently shown that L. pneumophila possesses genetic loci that are required for survival and replication within both human macrophages and protozoans, and these loci have been designated pmi (protozoan and macrophage infectivity) loci (19, 20). However, there are other distinct bacterial loci that are uniquely required for intracellular replication within macrophages but not within protozoans, and these loci have been designated mil (macrophage-specific infectivity) loci (19, 21).

Uptake of L. pneumophila by monocytes occurs by microfilament-dependent coiling and conventional phagocytosis mechanisms, while uptake of L. micdadei occurs only by conventional phagocytosis (30, 35). In contrast to uptake of L. pneumophila by macrophages, entry of the bacterium into the protozoan host Hartmannella vermiformis occurs by different mechanisms and is not inhibited by the microfilament disrupting agent cytochalasin D (5, 20, 21, 24, 27). Recently, we demonstrated the presence of a 170-kDa galactose N-acetylgalactosamine (Gal/GalNAc)-inhibitable lectin in H. vermiformis (24, 34). We have recently shown that type IV pili, which are filamentous appendages expressed on the bacterial surface, are involved in bacterial adherence to mammalian and protozoan cells (32). Whether the type IV pili are the ligands for the Gal/GalNAc lectin is not known.

Most studies have used L. pneumophila as a model pathogen for studying Legionnaires' disease, and the mechanism of entry of L. micdadei into H. vermiformis and the intracellular environment in which L. micdadei resides within the host cell are not well-defined. Hence, we sought to address these issues, which may be important for understanding the intracellular life cycle and pathogenesis of L. micdadei in its protozoan host, H. vermiformis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strain and culture. Virulent strain AA100 of L. pneumophila is a virulent clinical isolate that has been described previously (3). The virulent strain L. micdadei Rivera was obtained from Barry Fields and Janet Purckler, Centers for Disease Control and Prevention (Atlanta, Ga.). The strains were grown on buffered charcoal yeast extract (BCYE) agar plates.

Protozoan culture. H. vermiformis CDC-19 (= ATCC 50237) has been cloned and grown in axenic culture and has been used as a model organism to study the pathogenesis of L. pneumophila (1, 17). This strain was isolated from a water source during an outbreak of nosocomial Legionnaires' disease in a hospital in South Dakota (11, 17). The amoebae were maintained in American Type Culture Collection axenic culture medium 1034 supplemented with 10% fetal calf serum at 37°C (17).

Infection protocol. H. vermiformis was harvested and infected with L. micdadei as previously described (24, 34). Briefly, amoebae were incubated overnight in culture flasks in serum-free axenic medium. Then amoebae were harvested by centrifugation and resuspended in fresh serum-free axenic medium. Aliquots containing 2 × 10⁷ amoebae/ml were infected with 10⁹ L. micdadei cells. After several intervals of coincubation at 37°C, amoebal cell lysates were prepared for immunoblot analysis as described below.

Transmission electron microscopy. Monolayers were infected with L. micdadei or L. pneumophila by using a multiplicity of infection of 100 for 1 h, followed by extensive washing of extracellular bacteria with American Type Culture Collection axenic medium 1034 supplemented with 10% fetal calf serum. Ultrathin sections were prepared as described previously (21). Briefly, infected amoebae were fixed with 3.5% gluteraldehyde and then with 1% OsO₄, dehydrated with ethanol, and embedded in Eponate 12 resin (Ted Pella, Redding, Calif.). Ultrathin sections were stained with uranyl acetate and then with lead citrate and were examined with a model H-7000/STEM electron microscope (Hitachi Inc., Tokyo, Japan) at 75 kV.

Preparation of cell lysates and Western blotting. After incubation of H. vermiformis with L. micdadei, infection was stopped by using cold stop buffer containing 1× phosphate-buffered saline (pH 7.2) and the phosphatase inhibitors NaF (5 mM) and Na₃VO₄ (1 mM) (Sigma Chemical Co., St. Louis, Mo.) (34). To examine the effects of methylamine on tyrosine phosphorylation, amoebae were pretreated with methylamine, and cells were also infected in the presence of this inhibitor, which blocks entry of the bacteria (see below). The cells were washed three times with stop buffer and pelleted by low-speed centrifugation at 735 × g for 2 min. The supernatant containing bacteria was discarded, and the amoebae were lysed with cold 1% Triton X-100 lysis buffer (20 mM Tris-HCl [pH 7.6], 150 mM NaCl, 10 mM NaF, 1 mM Na₃VO₄, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 2 µg of aprotinin per ml, 2 µg of leupeptin per ml) for 30 min on ice (34). The detergent-soluble and -insoluble fractions were separated by centrifugation at 16,000 × g for 30 min at 4°C in a microcentrifuge. Proteins from detergent-soluble fractions were resolved on sodium dodecyl sulfate-10% polyacrylamide gel electrophoresis gels under reducing conditions. The blots were transferred onto Immobilon-P membranes (Millipore, Bedford, Mass.) in a transfer cell (Bio-Rad, Hercules, Calif.) for 1.5 h with 0.2 M Tris-0.025 M glycine buffer containing 20% (vol/vol) methanol. After the proteins were transferred, the membranes were incubated for 30 min in a blocking buffer containing 1.5% bovine serum albumin. The membranes were then probed with a 1:5,000 dilution of horseradish peroxidase-conjugated recombinant anti-phosphotyrosine antibody RC20H (Transduction Laboratories, Lexington, Ky.) for 30 min at room temperature. After extensive washing, the blots were developed with an enhanced chemiluminescence kit (DuPont NEN, Boston, Mass.) according to the manufacturer's instructions.

Inhibition of L. pneumophila uptake by sugars. H. vermiformis was infected with L. micdadei as described below (23). To analyze the effects of different sugars on invasion of H. vermiformis by L. micdadei, infection experiments were performed in triplicate in the presence of the following sugars: galactose (Gal), N-acetyl-D-galactosamine (GalNAc), glucose, mannose, and lactose (Sigma). Solutions of these sugars were prepared in assay medium and stored at 4°C (1).

Invasion of amoebae in the presence of methylamine. Uptake of Legionella strains by amoebae was studied by using gentamicin protection invasion assays (24, 31). Briefly, H. vermiformis was resuspended in axenic medium at a concentration of 10⁷ cells/ml. The amoebae were incubated for 1 h in the presence of 100 mM methylamine (Sigma), an inhibitor of receptor-mediated endocytosis, prior to infection (13, 27). Following infection with 10⁹ L. micdadei cells for 1 h, the extracellular bacteria were killed with 50 µg of gentamicin per ml, and the intracellular bacteria were released following lysis of the amoebae with 0.04% Triton X-100. Lysis of the amoebae was monitored microscopically and was complete within 1 min. Different dilutions of bacteria were plated onto BCYE plates for colony enumeration (34).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Tyrosine dephosphorylation of H. vermiformis proteins after infection by L. micdadei. In this study we examined the changes in host protein tyrosine phosphorylation status during attachment to and invasion of the protozoan host H. vermiformis by L. micdadei. Amoebae were coincubated with L. micdadei for different periods of time, and cell lysates were immunoblotted with anti-phosphotyrosine antibody (Fig. 1). Infection of H. vermiformis with L. micdadei resulted in prominent time-dependent tyrosine dephosphorylation of several host cell proteins, including 200- to 205-, 170-, 150-, 69-, and 55-kDa proteins (Fig. 1, arrowheads).

View larger version (83K): [in this window] [in a new window]   FIG. 1.   Attachment to and invasion of H. vermiformis by L. micdadei induces reversible tyrosine dephosphorylation of several host cell proteins. H. vermiformis cell extracts were prepared from uninfected cells (lane 1) or cells infected for 1, 5, 15, and 30 min (lanes 2 through 5), subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and probed with anti-phosphotyrosine antibody. Lane 6 contained cell extracts from uninfected cells incubated at 37°C for 30 min. Lane 7 contained extracts from H. vermiformis infected for 30 min; then the extracellular bacteria were washed off and the amoebae were incubated for 15 min. Arrowheads indicate the positions of proteins that were markedly tyrosine dephosphorylated. The second arrowhead from the top indicates the position of the 170-kDa Gal/GalNAc lectin (34). The results are representative of the results of at least two independent experiments. KD, kilodaltons.

View larger version (102K): [in this window] [in a new window]   FIG. 2.   Tyrosine dephosphorylation of H. vermiformis proteins induced by L. micdadei is specific and requires attachment of the bacteria to the Gal/GalNAc lectin. H. vermiformis cell extracts were prepared from uninfected cells (lane 1) or cells infected for 30 min with live (lane 2) or dead (lane 3) bacteria. Amoebal extracts were also prepared after stimulation with latex beads (lane 4), bacterial culture supernatants (lane 5), or supernatant collected from an infection (lane 6) as additional controls. Lanes 7 through 9 contained host cell extracts prepared from infections in the presence of Gal, GalNAc, and mannose, respectively. The arrowhead indicates the position of the 170-kDa Gal/GalNAc lectin (34). The results are representative of the results of at least two independent experiments. KD, kilodaltons.

L. micdadei-induced protein tyrosine dephosphorylation in H. vermiformis is reversible. Next, we examined whether attachment to and invasion of H. vermiformis by L. micdadei resulted in a permanent change in the tyrosine phosphorylation status of cellular proteins or whether the alteration in host cell processes was reversible. H. vermiformis was coincubated with L. micdadei for 30 min, and extracellular bacteria were washed away after this incubation period. H. vermiformis cells were subsequently incubated for an additional 15 min, and the patterns of tyrosine-phosphorylated proteins were determined in immunoblots probed with anti-phosphotyrosine antibody. The levels of all of the tyrosine-dephosphorylated proteins returned to the normal levels observed with uninfected cells (data not shown). Overall, these observations indicated that attachment and invasion by L. micdadei resulted in tyrosine dephosphorylation of multiple proteins in H. vermiformis. This bacterium-induced tyrosine dephosphorylation in H. vermiformis was reversible and dependent on the continuous presence of bacteria.

Invasion of H. vermiformis by L. micdadei can be blocked by Gal and GalNAc. Previous work in our laboratory showed that a 170-kDa 2 integrin-like Gal/GalNAc lectin is present in H. vermiformis (Fig. 2, arrowhead) (34). This lectin mediates attachment of L. pneumophila and undergoes tyrosine dephosphorylation during this process (34). Invasion and intracellular replication of L. pneumophila in H. vermiformis were effectively blocked by the Gal and Gal/GalNAc monomers (34). To examine the functional significance of the 170-kDa lectin of H. vermiformis in the invasion of L. micdadei, blocking experiments were performed in the presence of Gal and Gal/GalNAc. Amoebae were preincubated with Gal or Gal/GalNAc for 15 min prior to infection with L. micdadei (Fig. 3). Mannose, lactose, and glucose were used as negative controls to verify the specificity of the sugars (data not shown). The presence of GalNAc had a dramatic effect on invasion by L. micdadei. Gal had a less dramatic effect but still sufficiently blocked the invasion of bacteria. Mannose, glucose, and lactose had no effect on bacterial invasion, and the number of intracellular bacteria at the end of the infection period in the presence of these sugars was very similar to the number in the untreated control (data not shown). The inhibition of bacterial invasion by Gal and Gal/GalNAc was dose dependent (data not shown), and these sugars were not toxic to H. vermiformis (34). These results indicated that the 170-kDa Gal/GalNAc lectin was involved in attachment to and invasion of H. vermiformis by L. micdadei.

View larger version (12K): [in this window] [in a new window]   FIG. 3.   Inhibition of invasion of H. vermiformis by L. micdadei in the presence of different sugar monomers. The growth kinetics of L. pneumophila in cocultures with H. vermiformis were determined in the absence or presence of sugars at a concentration of 100 mM. At several times after infection, the numbers of bacteria in the cocultures were determined following growth on agar plates. The bacteria did not replicate extracellularly in the cocultures, and thus the increase in the number of bacteria was due to intracellular replication.

Attachment of L. micdadei to H. vermiformis is required to induce protein tyrosine dephosphorylation of host cell proteins. We examined whether bacterial attachment to the Gal/GalNAc lectin was required to cause changes in the pattern of protein tyrosine phosphorylation in H. vermiformis. In these experiments, H. vermiformis cells were incubated with L. micdadei in the presence of Gal, Gal/GalNAc, or mannose (Fig. 2, lanes 7 through 9). Incubation of H. vermiformis with these sugars does not alter the profile of tyrosine phosphorylated proteins in resting amoebae (34). H. vermiformis cell lysates were prepared after 20 min of infection and were analyzed by using immunoblots probed with the anti-phosphotyrosine antibody. The presence of Gal (Fig. 2, lane 7) or Gal/GalNAc (Fig. 2, lane 8) completely blocked bacterium-induced tyrosine dephosphorylation of the 170- and 69-kDa proteins (Fig. 2, lanes 7 and 8). These proteins have been identified as the Gal/GalNAc lectin (170 kDa) (34) and paxillin (69 kDa) (unpublished data). As expected, mannose failed to block tyrosine dephosphorylation of proteins in H. vermiformis (Fig. 2, lane 9). These results indicated that contact of L. micdadei with the Gal/GalNAc lectin on H. vermiformis was required to mediate early changes in the tyrosine phosphorylation of host cell proteins.

Invasion of H. vermiformis by L. micdadei is not required to induce protein tyrosine dephosphorylation. Since bacterial attachment was essential for induction of tyrosine dephosphorylation of amoebal proteins (Fig. 2), we examined whether entry of L. micdadei into H. vermiformis was also required to mediate these effects. Methylamine has been shown to block the entry of L. pneumophila into H. vermiformis (27). We examined whether methylamine blocks invasion by L. micdadei by using the gentamicin protection assay (see above). In this assay, following a period of infection, extracellular bacteria were killed with the antibiotic gentamicin, while intracellular bacteria were protected and subsequently enumerated (34). The data showed that like inhibition of invasion of H. vermiformis by L. pneumophila, invasion by L. micdadei was inhibited 99% in the presence of 100 mM methylamine (data not shown).

View larger version (K): [in this window] [in a new window]   FIG. 4.   Tyrosine dephosphorylation of H. vermiformis proteins induced by L. micdadei does not require entry of the bacteria. H. vermiformis cell extracts were prepared from uninfected cells in the absence (lane 1) or presence (lane 2) of methylamine. Lane 3 contained cell lysates prepared after 30 min of infection of H. vermiformis with L. micdadei in the presence of methylamine. The arrowheads indicate the positions of the proteins that were tyrosine dephosphorylated. The second arrowhead from the top indicates the position of the 170-kDa Gal/GalNAc lectin (34). The results are representative of the results of at least two independent experiments. KD, kilodaltons.

L. micdadei is surrounded by a smooth phagosome in H. vermiformis. In H. vermiformis, L. pneumophila is targeted into an RER-surrounded phagosome (1). Transmission electron microscopy showed that after invasion of H. vermiformis by L. micdadei, the bacteria were targeted into RER-free phagosomes (Fig. 5). None of the approximately 100 phagosomes examined that contained L. micdadei was surrounded by RER; in contrast, 83% of the phagosomes containing L. pneumophila were surrounded by RER (1, 20). By 18 h postinfection, the amoebae were heavily infected with both Legionella species. These data showed that despite the fact that L. pneumophila and L. micdadei had similar mechanisms for utilizing the Gal/GalNAc lectin in bacterial attachment and invasion, these two species were targeted into phagosomes that were morphologically distinct, at least at the ultrastructural level.

View larger version (162K): [in this window] [in a new window]   FIG. 5.   L. micdadei and L. pneumophila are targeted into morphologically distinct phagosomes in H. vermiformis: transmission electron micrographs of H. vermiformis infected with L. micdadei (A and C) or L. pneumophila (B and D). The micrographs were obtained 4 h postinfection (A and B) or 18 h postinfection (C and D). The arrows in panel B indicate the RER-surrounded phagosome. b, bacteria. (A and B) Bars = 0.5 µm. (C and D) Bars = 1 µm.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study focused on the initial signaling events that are activated upon attachment to and invasion of H. vermiformis by the Legionnaires' disease bacterium L. micdadei. We recently identified the Gal/GalNAc lectin of H. vermiformis as a potential receptor utilized by L. pneumophila for attachment to and invasion of protozoans (34). In this study we examined the role of the Gal/GalNAc lectin in the attachment to and invasion of H. vermiformis by L. micdadei. While entry of several pathogenic bacteria, including Salmonella typhimurium and Shigella flexneri, induces host tyrosine phosphorylation events (18), uptake of L. micdadei by H. vermiformis is associated with rapid and reversible tyrosine dephosphorylation of host cell proteins. These effects of L. micdadei on H. vermiformis are similar to the effects induced by Yersinia pseudotuberculosis during invasion of macrophages, which are mediated by a bacterial tyrosine phosphatase (28). Dephosphorylation of cellular proteins prevents uptake of Yersinia cells (18). In contrast, tyrosine dephosphorylation of H. vermiformis proteins appears to be associated with attachment and uptake of L. micdadei. These observations indicate that there is a novel mechanism for uptake of L. micdadei by H. vermiformis.

The Gal/GalNAc lectin of Entamoeba histolytica is antigenically similar to the human 2 integrin and is involved in cytoskeleton-mediated capping and shedding of antibody and complement from the protozoan surface in order to evade the immune system of the human host (8, 9). Utilization of the Gal/GalNAc lectin on the protozoan host and its tyrosine dephosphorylation upon bacterial attachment and invasion may provide a selective advantage for L. micdadei in the regulation of its uptake process. In contrast to the induction of tyrosine phosphorylation of the receptor and other proteins upon engagement of receptors with their ligands (12), attachment of L. micdadei to the Gal/GalNAc lectin was associated with tyrosine dephosphorylation of the lectin and other host proteins. We offer three hypotheses to explain this observation. First, bacterium-induced tyrosine dephosphorylation of the lectin may disrupt communication of the lectin with the underlying cytoskeleton in order to avoid shedding of bacteria from the protozoan surface. This hypothesis is supported by our recent observations that several lectin-associated proteins are dissociated upon attachment of L. pneumophila to H. vermiformis (unpublished data). Some of these dissociated proteins may represent important cytoskeletal proteins involved in cell shape, motility, and chemotaxis in amoebae. In addition, four protozoan cytoskeletal proteins, including paxillin, vinculin, and pp125^FAK, undergo tyrosine dephosphorylation upon attachment to L. pneumophila, indicating that disruption of the cytoskeleton occurs (33). Second, dephosphorylation of H. vermiformis proteins upon attachment to L. micdadei may downregulate host signals in order to prevent activation of phagocytic signals and microbicidal mechanisms that result in degradation of engulfed bacteria. Subversion of the bactericidal protozoan host cell processes by L. micdadei may be supported by observations made with neutrophils (14). Suppression of host cell functions, like superoxide anion production, chemotaxis, and bactericidal activity, by L. micdadei was observed when human neutrophils were allowed to attach to and ingest bacteria (14). Third, dephosphorylation of H. vermiformis proteins, including the Gal/GalNAc lectin, by L. micdadei may increase the efficiency of bacterial uptake. Overall, attachment to and invasion of H. vermiformis by L. micdadei mediated by the Gal/GalNAc lectin may be required to accomplish more than one of the strategies proposed above to ensure efficient uptake and subsequent targeting of the bacteria into a safe phagosome in amoebae.

The Gal/GalNAc lectin of H. vermiformis is important for invasion by L. micdadei, as determined by blocking experiments performed with Gal and Gal/GalNAc. Recent studies in our laboratory have demonstrated the importance of the 170-kDa Gal lectin in the invasion of H. vermiformis by L. pneumophila (24, 34). The sugars Gal and Gal/GalNAc block invasion of H. vermiformis by L. pneumophila (24, 34). It is interesting to note that the two most frequent causative agents of Legionnaires' disease, L. micdadei and L. pneumophila, utilize a common host cell receptor to invade H. vermiformis.

Intracellular pathogens like L. pneumophila and Mycobacterium tuberculosis utilize the CR3 receptor to enter human macrophages, which prevents activation of oxidative bursts and other microbicidal mechanisms (36, 37). Uptake of many intracellular pathogens through the Fc receptor (instead of CR3) targets the internalized bacteria into lysosomes instead of a replicative niche (25, 26). Although L. pneumophila and L. micdadei utilize the Gal/GalNAc lectin for attachment and invasion of their protozoan host, H. vermiformis, the intracellular fates of the two bacterial species are different, at least at the ultrastructural level. Accordingly, a phagosome that contains L. pneumophila but not L. micdadei is surrounded by RER. These observations are similar to observations made with a phagosome containing both of these Legionella species in macrophages and epithelial cells (unpublished data). These results suggest that additional mechanisms (other than utilizing the Gal/GalNAc lectin) may be involved in the trafficking L. micdadei to the RER-free replicative phagosome. It is possible that L. micdadei resides in a relatively nascent phagosome, while L. pneumophila exists in a more mature phagosome that is surrounded by RER.

In summary, this study and our earlier report (34) revealed novel but similar signaling mechanisms that are triggered in the protozoan host H. vermiformis following attachment and invasion by the two Legionella species that have been most commonly implicated in outbreaks of Legionnaires' disease. Despite the similarities in what appears to be receptor-mediated signaling events, the two species are targeted into different compartments in the protozoan host. Additional studies should help explain the exact strategies utilized by Legionella spp. to enter and invade two evolutionarily distant hosts, human cells and environmental free-living amoebae.