Use of Galleria mellonella as a Model Organism to Study Legionella pneumophila Infection

Legionella pneumophila, the causative agent of a severe pneumonia named Legionnaires' disease, is an important human pathogen that infects and replicates within alveolar macrophages. Its virulence depends on the Dot/Icm type IV secretion system (T4SS), which is essential to establish a replication permissive vacuole known as the Legionella containing vacuole (LCV). L. pneumophila infection can be modeled in mice however most mouse strains are not permissive, leading to the search for novel infection models. We have recently shown that the larvae of the wax moth Galleria mellonella are suitable for investigation of L. pneumophila infection. G. mellonella is increasingly used as an infection model for human pathogens and a good correlation exists between virulence of several bacterial species in the insect and in mammalian models. A key component of the larvae's immune defenses are hemocytes, professional phagocytes, which take up and destroy invaders. L. pneumophila is able to infect, form a LCV and replicate within these cells. Here we demonstrate protocols for analyzing L. pneumophila virulence in the G. mellonella model, including how to grow infectious L. pneumophila, pretreat the larvae with inhibitors, infect the larvae and how to extract infected cells for quantification and immunofluorescence microscopy. We also describe how to quantify bacterial replication and fitness in competition assays. These approaches allow for the rapid screening of mutants to determine factors important in L. pneumophila virulence, describing a new tool to aid our understanding of this complex pathogen.


Introduction
Animal models of infection have proved invaluable in the determination of bacterial virulence factors. However, invertebrate models have gained increased attention as a viable alternative to traditional mammalian models of infection. The larvae of the wax moth, Galleria mellonella is increasingly being used to study a number of important human pathogens, including Gram-positive 1 and Gram-negative bacteria 2,3 and several pathogenic fungi 4,5 . Using an insect model has a number of advantages over traditional mammalian models, as an invertebrate, G. mellonella is not subject to the ethical limitations of mammalian models. In addition, the larvae can be easily maintained, infected by injection without anesthesia, undergo pretreatment with chemical inhibitors 6 and sustain incubation at 37 °C 7 . Interestingly, a good correlation between the pathogenicity of several microorganisms in G. mellonella and mammalian models of infection has been established 2,8 . The increased understanding of the immune system of G. mellonella has also assisted in the characterization of this model organism. Although insects do not have an adaptive immune system as found in mammals, they do have sophisticated cellular and humeral defenses including the production of antimicrobial peptides 9 . Hemocytes are the major mediator of cellular defenses and are the most numerous cell type found in the hemolymph (or blood) of G. mellonella 10 , These cells are professional phagocytes and perform similar functions to human macrophages and neutrophils by both taking up and degrading bacteria in a phago-lysosomal compartment 10,11 and forming nodules around invading bacteria, physically restricting bacterial replication 12 .
Legionella pneumophila is a respiratory pathogen that causes severe pneumonia (Legionnaires' disease) in susceptible populations such as the elderly or immunocompromised 13 . Legionella is found ubiquitously in both environmental and man-made water sources, where it is a pathogen of various species of fresh water amoebae 14,15 . Legionella survives and replicates within these professional phagocytes by utilizing a multi-protein complex known as the Dot/Icm (defective in organelle trafficking/intracellular multiplication) type 4 secretion system (T4SS) to translocate over 275 effector proteins into the host cell [16][17][18][19][20] . These proteins serve to subvert the normal host cell phagocytic pathways, leading to the creation of the Legionella containing vacuole (LCV). The LCV avoids fusion with lysosomes and instead recruits endoplasmic reticulum (ER)-derived vesicles, resulting in a specialized compartment that resembles the rough ER 21

Quantification of Bacterial CFU
1. Add 100 μg/ml spectinomycin to CYE plates to avoid contamination by gut flora. L. pneumophila strain 130b is naturally resistant to spectinomycin 40 . 2. Before hemolymph extraction, weigh 1.5 ml centrifuge tubes. 3. Extract hemolymph as described in section 6 and place in weighed tubes, add 1 μl of 5 mg/ml digitonin, mix well and incubate for 5 min at RT to lyse hemocytes. 4. Reweigh the tube with hemolymph and determine the weight of hemolymph extracted. 5. Perform ten fold serial dilutions of the hemolymph in sterile AYE media. 6. Using a pen, divide the base of a CYE plate into six equal sectors and label. 7. Plate three drops of 25 μl of each dilution (starting with the most dilute) in each section of the plate. 8. Incubate the plates with the lids upmost overnight at 37 °C. 9. Once the drops have dried fully, turn the plate over and incubate at 37 °C for at least a further two days. 10. Quantify the bacteria extracted by counting the colonies at each dilution and normalize to the weight of hemolymph extracted.

Determination of the Competitive Index (CI)
1. Confirm that both strains grow equally well in broth culture and on CYE agar plates prior to attempting the competitive index. 2. Prepare WT or kanamycin resistant mutant bacterial suspensions as described in section 1 and mix in a 1:1 ratio.
1. Plate serial dilutions of the inoculum onto CYE spectinomycin (100 μg/ml) and CYE spectinomycin /kanamycin 3. Infect larvae and extract hemolymph at suitable time points as described above. 4. Determine viable counts by extracting hemolymph and plating serial dilutions onto CYE spectinomycin and CYE spectinomycin/kanamycin.

Representative Results
Here it is demonstrated that G. mellonella is an appropriate, easy to use model to study L. pneumophila infection. Previously it has been shown that L. pneumophila virulence in macrophages, amoebae and mammalian models is dependent on the presence of the Dot/Icm secretion system [41][42][43] . G. mellonella larvae were infected as described above and the virulence of the wild type (WT) and a Dot/Icm-deficient strain compared. Infection with 10 7 CFU of L. pneumophila strain 130b resulted in 100% mortality within 24 hr post infection (p.i.). However, the L. pneumophila ΔdotA strain, which does not have a functional Dot/Icm T4SS secretion system, was avirulent in this assay (Figure 1). This demonstrates that L. pneumophila virulence in G. mellonella depends on the translocation of Dot/Icm effectors, making this model suitable for characterization of the function of these proteins.
Recently, it was shown that inhibition of phagocytosis by cytochalasin treatment increased the susceptibly of the larvae to infection by the yeast Candida albicans 6 . As L. pneumophila is an intracellular pathogen, it was decided to determine if uptake of the bacteria is crucial in its pathogenesis in this model. Larvae were pretreated with 10 μl of 100 μM Cytochalasin D (CyD) for 4 hr at 37 °C, then infected with 10 7 CFU of WT L. pneumophila 130b and mortality monitored at 24 hr p.i. Treatment with the inhibitor alone did not affect larval survival. However, pretreated, infected larvae displayed significantly greater survival (P = 0.0066, unpaired T-test) compared to DMSO-treated, infected insects (Figure 2). The effect of CyD treatment was abolished by 48 hr p.i. (results not shown); this may be due to the half-life of the drug in G. mellonella. This demonstrates that uptake of L. pneumophila into G. mellonella hemocytes is a crucial aspect of bacterial virulence.
In order to validate expression and determine the subcellular localization of an effector protein in G. mellonella, hemocytes were extracted and processed for immunofluorescence microscopy. Larvae were infected with WT and ΔdotA L. pneumophila 130b expressing a fragment of the well-defined T4SS effector, SidC 41-918 , fused to 4 N-terminal HA tags. This effector was demonstrated to bind the LCV via a phosphoinositide-4-phosphate-binding domain 44 . Using anti-HA (red) and anti-Legionella (green) antibodies, 4HA-SidC 41-918 localized to the LCV in infected hemocytes (Figure 3). This localization has previously been shown in the amoebae Dictyostelium discoideum and in mammalian macrophages 44,45 confirming the comparability of this model.
The importance of proteins for virulence is usually determined by comparing the growth kinetics of wild type and mutant bacteria. In order to follow the bacterial replication kinetics over the course of the infection, three larvae were sacrificed at each time point (0, 5, 18, and 24 hr p.i.), the hemolymph collected and pooled and the CFU/0.1g of extracted hemolymph determined. After an initial dip at 5 hr p.i., the CFU of the WT bacteria increases up to 24 hr p.i. however, the ΔdotA strain undergoes no replication and is cleared at 18 hr p.i. (Figure 4).
The ability of L. pneumophila to cause lysis of macrophages in a T4SS-dependent manner has long been documented 46 , however no similar studies have been performed in vivo. The concentration of circulating hemocytes was determined at 5, 18, and 24 hr p.i. Larvae were infected with WT or ΔdotA L. pneumophila 130b, hemocytes extracted from infected insects and viable cells counted using the trypan blue exclusion method. At 5 hr p.i. no difference in hemocyte counts between the strains could be seen ( Figure 5). However, at 18 hr p.i. there was a significant drop in hemocyte concentration in WT, but not ΔdotA, infected larvae. This difference persisted at 24 hr p.i. The drop in hemocyte number,