Introduction Legionella pneumophila is a Gram-negative bacterium found ubiquitously in environmental water reservoirs where it replicates in free-living protozoa (38). Following inhalation of contaminated aerosols, L. pneumophila is capable of infecting human alveolar macrophages and causing disease ranging from mild flu-like symptoms to Legionnaires’ disease, a severe, life-threatening pneumonia (16). L. pneumophila thrives in professional phagocytes by avoiding killing by the phago-lysosomal pathway (21). Instead it establishes a specialized Legionella containing vacuole (LCV), which shows characteristics of the rough endoplasmic reticulum (ER) (48). L. pneumophila employs several specialized protein secretion systems, e.g. the twin-arginine translocation (Tat) pathway and a type II secretion system (T2SS), to secrete virulence factors, some of which have been shown to contribute to Legionella’s intracellular survival and pathogenicity (11, 12). However, the essential virulence determinant of L. pneumophila is the Dot (defective in organelle trafcking)/Icm (intracellular multiplication) type IV secretion system (T4SS), which is indispensable for intracellular survival and establishment of the replication-permissive LCV in both amoebae and macrophages (4, 45). The Dot/Icm T4SS is a multi-protein complex able to translocate at least 275 effector proteins directly into host cells (35, 53). Although it has been demonstrated that several T4SS effectors manipulate host cell vesicular trafficking, inhibit apoptosis and immune signaling, the function of the majority of T4SS effectors during infection is still unknown (5). Free-living freshwater amoebae such as Acanthamoeba castellanii or Hartmannella vermiformis routinely serve as model hosts to study molecular aspects of Legionella pathogenesis (1, 20). As natural hosts, these professional phagocytes are believed to have exerted evolutionary pressure for the selection of Legionella’s virulence factors that enable the bacteria to overcome the antimicrobial activities of human macrophages (32). In addition, Dictyostelium discoideum has become a prevalent protozoan model organism as it can readily be genetically modified (47). Although, protozoan Legionella infection models have proven successful, they do not fully reflect the infection of macrophages as amoeba employ less complex antimicrobial mechanisms than mammalian cells. The nematode Caenorhabditis elegans possesses an innate immune system and is a wellestablished model for several bacterial pathogens including Legionella spp. (5). However, one caveat to the use of C. elegans is that bacteria replicate in the intestinal lumen and do not invade intestinal epithelial cells, limiting the usefulness of this model to study virulence determinants required for Legionella’s intracellular lifestyle. Typically, human Legionella infection is modeled using mammalian hosts (3, 6). Disease progression in the guinea pig resembles Legionellosis in humans and pathology includes lymphocyte infiltration, goblet cell metaplasia, mild fibrosis and emphysema (3). In contrast, the majority of mouse strains are resistant to Legionella infection (52) with the exception of the inbred albino A/J mouse, which develops a self-limiting infection (6). Due to the high cost and ethical considerations associated with the use of mammalian hosts, the search for alternative models is ongoing. Insect model organisms, in particular Drosophila melanogaster, have been introduced to study bacterial pathogenesis (44). L. pneumophila replicates in D. melanogaster and kills the flies in a Dot/Icm T4SS-dependent manner (27). The human and insect innate immune systems demonstrate many similarities (24, 29) with most insect species containing specialized cells known as haemocytes that phagocytose pathogens and form aggregates which encapsulate and neutralize foreign microorganisms (30). Moreover, activated haemocytes can trigger a phenoloxidase (PO) melanisation cascade leading to physical restriction of intruders and the production of antimicrobial compounds (8). Haemocyte-mediated responses are complemented by the production and secretion of anti-microbial peptides by the insect fat body, an organ similar to the mammalian liver (29, 31). Besides D. melanogaster, the larva of the greater wax moth Galleria mellonella has become a widely adopted insect model to study a wide range of human pathogens including Listeria spp. (23), Streptococcus pyogenes (36), Campylobacter jejuni (10), Yersinia pseudotuberculosis (9) and several pathogenic fungi (17, 33). G. mellonella larvae can be easily maintained and infected by injection without anesthesia and sustain incubation at 37C (33). A good correlation between the pathogenicity of several microorganisms in G. mellonella and other mammalian models of infection has been established (22, 23). The aim of this study was to determine if G. mellonella could be used as a model to study L. pneumophila pathogenesis. Material and Methods Bacterial strains and G. mellonella larvae. L. pneumophila serogroup 1 strain 130b is a spectinomycin-resistant clinical isolate from the Wadsworth Veterans Administration Hospital, Los Angeles, CA (14). The L. pneumophila ∆DotA strain is a dotA insertion mutant (kanamycin resistance) of L. pneumophila strain 130b (41). L. pneumophila strain JR32 is a salt sensitive streptomycin-resistant L. pneumophila strain Philadelphia-1 isolate (39) and the ∆IcmT strain is an icmT isogenic mutant in the JR32 strain (46). L. pneumophila strain Lp02 is a thymine auxotroph streptomycin-resistant derivative of the Philadelphia-1 strain (4). L. pneumophila strain Paris is a worldwide epidemic strain (7) G. mellonella larvae were obtained from Livefoods, UK and stored at room temperature in the dark. Infection of G. mellonella. L. pneumophila strains were cultured on charcoal-yeast extract (CYE) plates for four days then inoculated into ACES yeast extract (AYE) as described previously (43). For the Lp02 strain, thymidine (100 μg/ml) was added. After 21 h of growth, bacteria were diluted in Dulbecco’s phosphate buffered saline (PBS) to an OD600 of 1 which corresponds to 10 CFU/ml unless otherwise indicated. Gene expression in strains containing the p4HA plasmid was additionally induced during infection with 1mM isopropyl -d-1thiogalactopyranoside (IPTG). Ten G. mellonella larvae were injected with 10 μl of bacterial suspension as previously described (37) and were incubated at 37C in the dark. As a control ten larvae were injected with PBS alone and ten untreated insects were included with every experiment. Larvae were individually examined for pigmentation and time of death was recorded. Assays were only allowed to proceed for 3 days as pupa formation could occasionally be seen by day 4. At least three independent replicates of each experiment were performed. Intracellular growth assay. At 0, 2, 5, 18 and 24 h post infection (p. i.) haemolymph was extracted from three infected larvae and pooled as previously described (23). Cells were lysed by incubation of the haemolymph with 1 μl of 5 mg/ml digitonin for 5 min at room temperature. Extracted haemolymph was serially diluted in AYE media and plated onto CYE plates. To prevent contamination, the extracted haemolymph was plated on CYE plates supplemented with spectinomycin (50 μg/ml) for the L. pneumophila strain 130b or streptomycin (100 μg/ml) for the Philadelphia-1-derived strains. Plates were incubated at 37 C for three days, viable bacteria were enumerated and the number of CFU was normalized to the weight of haemolymph extracted. Plasmids. A fragment of the SidC homologue from L. pneumophila 130b containing the phosphatidylinositol-4 phosphate binding domain (amino acids 41 to 918) was cloned into the Xbal and BamHI sites of the p4HA plasmid (13) to yield the IPTG-inducible 4HA-SidC41918 expression plasmid pICC562 using the forward primer 5’cgtattctagataacacctgccaaacagcagttgag-3’ and the reverse primer 5’ggctaggatccctatttctttataactcccgtgtac-3’ and standard molecular biology techniques. Indirect immunofluorescence on extracted haemocytes. Haemolymph from infected G. mellonella was extracted at 5 and 24 h post infection. The extracted haemolymph was dispensed onto poly-L-lysine coated glass coverslips and centrifuged at 500 x g for 10 min to allow sedimentation and attachment of haemocytes. Coverslips were washed twice with PBS and fixed using 4% paraformaldehyde for 20 min followed by quenching with 50 mM ammonium chloride. Extracellular L. pneumophila were stained with a mouse anti-L. pneumophila LPS antibody (ViroStat) and a donkey anti-mouse Rhodamine Red-Xconjugated antibody (Jackson ImmunoResearch Laboratories, Inc.). After permeabilization of the cells with 0.1% Triton in PBS and blocking with 2% (w/v) bovine serum albumin (BSA) in PBS, total bacteria were stained with a rabbit anti-L. pneumophila antibody (Affinity BioReagents) and a donkey anti-rabbit Alexa Fluor 488-conjugated antibody (Jackson ImmunoResearch). To visualize 4HA-SidC41-918 in haemocytes, fixed cells were permeabilised and blocked for 1 h in PBS containing 2% (w/v) BSA. Samples were stained with rabbit anti-L. pneumophila antibody (Affinity BioReagents), donkey anti-rabbit Alexa Fluor 488-conjugated antibody (Jackson ImmunoResearch), mouse anti-HA conjugated to Tetramethyl Rhodamine IsoThiocyanate (TRITC) (Sigma) and 5 μg ml 1 of 4 ,6-diamidino-2-phenylindole (DAPI) to visualize DNA. Samples were analyzed using an Axio M1 Imager microscope and images processed with the AxioVision software (Carl Zeiss). Staining of formalin fixed sections of G. mellonella. G. mellonella were fixed in formalin for one week at room temperature, paraffin embedded, sectioned and stained either with haematoxylin and eosin (H&E) or by indirect immunofluorescence as described previously (18). L. pneumophila was stained with rabbit anti-L. pneumophila antibody (Affinity BioReagents) and donkey anti-rabbit Alexa Fluor 488-conjugated antibody (Jackson ImmunoResearch). Cellular and bacterial DNA was stained with DAPI and the shape of the tissues was visualized using Rhodamine Phalloidin (Invitrogen). Transmission electron microscopy. Haemolymph was extracted from ten infected G. mellonella per condition and time point. Cells were spun down onto 6 well plates, washed once with PBS and fixed in 2% glutaraldehyde. Samples were processed as described previously (26) and examined using a Tecnai12 (FEI) electron microscope. Images were taken with a CCD camera (TVIPS, Gauting, Germany). Haemocyte quantification and viability assay. Infected haemolymph was extracted at 5 and 18 h p.i., Trypan blue (0.02% (v/v) in PBS) was added to cells and incubated at room temperature for 10 min. Viable cells were enumerated using a haemocytometer and each sample was analysed in triplicate. The average of three independent experiments was plotted graphically. Phenoloxidase (PO) activity assay. At 5 and 18 h p.i. haemolymph from three infected insects per condition was extracted and pooled. Cells and debris were removed by centrifugation at 20000 x g for 10 min at 4 C. The phenoloxidase activity in the plasma was quantified using a microplate enzyme assay as described previously (15). The change in absorbance at 490 nm was read for 1 h at room temperature with a reading taken every minute using a Fluostar Optima plate reader (BMG labtech, Germany). The experiment was performed in triplicate and independently repeated at least three times. Phenoloxidase activity was expressed as the mean OD490/minute. RNA extraction and RT-PCR. At indicated time points fat bodies from three larvae were collected and stored in RNAlater (Qiagen) at 4 C until processing. Tissue was homogenized by a gentleMACS homogeniser (Miltenyi Biotech) using M tubes and the 90 s RNA setting. RNA was extracted using a RNAeasy kit (Qiagen) and contaminating DNA was digested using Turbo DNA-free kit (Ambion) following the manufacturer's instructions. Two-step RTPCR was performed using Superscript reverse transcriptase (Invitrogen) using 2 μg of RNA as a template and random hexamers (Invitrogen). Genes were amplified using RedTaq readymix (Sigma) and 0.6 pM of gene specific primers (Table 1) as described previously (23). DNA was analyzed on a 1% agarose gel with SYBRSafe (Invitrogen) and quantified using ImageJ software (NIH).