Colonization with bacterial species from the Burkholderia cepacia complex (Bcc) is associated with fast health decline among individuals with cystic fibrosis. In order to investigate the virulence of the Bcc, several alternative infection models have been developed. To this end, the fruit fly is increasingly used as surrogate host, and its validity to enhance our understanding of host-pathogen relationships has been demonstrated with a variety of microorganisms. Moreover, its relevance as a suitable alternative to mammalian hosts has been confirmed with vertebrate organisms.
The aim of this study was to establish Drosophila melanogaster as a surrogate host for species from the Bcc. While the feeding method proved unsuccessful at killing the flies, the pricking technique did generate mortality within the populations. Results obtained with the fruit fly model are comparable with results obtained using mammalian infection models. Furthermore, validity of the Drosophila infection model was confirmed with B. cenocepacia K56-2 mutants known to be less virulent in murine hosts or in other alternative models. Competitive index (CI) analyses were also performed using the fruit fly as host. Results of CI experiments agree with those obtained with mammalian models.
Citation: Castonguay-Vanier J, Vial L, Tremblay J, Déziel E (2010) Drosophila melanogaster as a Model Host for the Burkholderia cepacia Complex. PLoS ONE 5(7): e11467. https://doi.org/10.1371/journal.pone.0011467
Editor: François Leulier, CNRS - Université Aix-Marseille, France
Received: March 9, 2010; Accepted: May 19, 2010; Published: July 12, 2010
Copyright: © 2010 Castonguay-Vanier et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by the Canadian Institutes of Health Research (CIHR) Operating grant MOP-77642 to ED. JCV received a master of science (M.Sc.) scholarship and LV a postdoctoral fellowship, both from the Fondation Armand-Frappier. JT was recipient of a doctor of philosophy (Ph.D.) scholarship from the Fonds quebecois de la recherche sur la nature et les technologies (FQRNT). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Members of the Burkholderia bacterial genus are well known for the versatility of their ecological niches. They were first isolated from the phytosphere where they were found to be pathogenic to plants . However, it is now known that many Burkholderia also have developed beneficial interactions with their plant hosts and have considerable ecological importance: several species of Burkholderia have proven to be very efficient biocontrol and bioremediation agents , . Burkholderia species are among the most antibiotic-resistant bacteria associated with human infections . Some species can in fact survive in antimicrobial agent solutions , and inside macrophages ,  and free-living amoebae . Within the genus, the Burkholderia cepacia complex (Bcc) has channelled a great part of the interest for all these reasons. Furthermore, species from the Bcc are responsible for chronic granulomatous disease  and are posing a considerable threat to immunocompromised individuals such as cystic fibrosis (CF) patients. The seriousness of a Bcc infection is highlighted by the fact that CF patients infected with Bcc strains suffer a faster health decline than when infected with Pseudomonas aeruginosa . The Bcc is composed of at least seventeen closely related genomic species or genomovars, all of which having been recovered from CF individuals , . Their prevalence is however not equal: B. multivorans and B. cenocepacia are the most encountered, with the latter also associated with the highest mortality rate within the CF community .
Much remains to be done to better understand the mechanisms behind the broad virulence of the Bcc, and development of animal models seems therefore inevitable. The traditional murine model has proven useful in the quest for understanding the virulence mechanisms of the Bcc , , , , but the search is on for more cost-effective alternatives and for somewhat less controversial widescreen models with faster generation time.
Over recent years, several alternative infection models have been developed for the Bcc, notably Galleria mellonella, Caenorhabditis elegans and alfalfa (Medicago sativa) ,,. The alfalfa seedlings model was the first proposed alternative to mice. This simple model to assess virulence revealed various patterns of infection between Bcc strains . The larvae of the greater wax moth G. mellonella has recently been proposed as an useful model for the testing of different strains of the Bcc , partly because it had previously shown good correlation between P. aeruginosa infection outcomes in mammals and in lower organisms . However, M. sativa and G. mellonella are not easily manipulated genetically. Hosts with which both reverse and forward genetics are readily possible represent additional advantages. Thus, the nematode C. elegans, easily genetically manipulated, has also been suggested , but produced mixed results regarding infection outcomes , , .
The genome sequence of the fruit fly has been unravelled several years ago  and mutants are readily available. Although Drosophila does not possess an acquired immune system, its innate counterpart is very similar to the mammalian one , . The fruit fly is capable of cellular as well as humoral responses when faced with invaders: the phagocytosis is done by its plasmatocytes and its fat body produces an array of antimicrobial peptides. The signalling cascades involved in the production of these molecules represent the milestone of the similarity between the innate immune system of vertebrates and of the rest of the animal kingdom . For these reasons, the fruit fly offers great potential to give insights on host-pathogen interactions. In fact, Drosophila melanogaster has already proven to be a great tool in the study of plant or fungal pathogens such as Erwinia carotovora  and Cryptococcus neoformans , but also of human opportunistic ones, such as P. aeruginosa .
In this work, we establish the use of the fruit fly as an effective model of infection for not only discriminating species and strains within the Bcc but also for the study of Bcc virulence factors. We also show that use of competitive index provides supplemental discriminating power for the characterization of virulence factors.
Results and Discussion
Burkholderia cepacia complex does not kill D. melanogaster when fed to the fly
Validating the fruit fly as an effective model in the study of the virulence of the Bcc species seemed promising because Drosophila was already used successfully with other pathogenic bacteria. Two different methods have been employed to infect Drosophila with bacteria: fly feeding, which involves feeding starved flies with bacteria, and nicking, which implies pricking flies in the thorax with a needle dipped into bacterial suspension. For instance, both feeding and pricking infections performed with P. aeruginosa are lethal to the flies , , .
Hence, several species from the Bcc were first tested for their capacity to kill fruit flies following ingestion. Interestingly, B. multivorans LMG16660, B. vietnamiensis LMG 18835, B. ambifaria HSJ1, B. pyrrocinia LMG21824, B. cenocepacia K56-2, B. cenocepacia LMG18830, B. dolosa LMG21819 and B. stabilis LMG18870 were all incapable of producing mortality during the trials (data not shown). Variables that could potentially have an influence on infection outcomes were then modified: flies were deprived of food and water for 9 h instead of 7 h, the initial temperature of 21°C was raised to 25°C, and bacterial concentration on which flies were left to feed was doubled. Still, the flies were not killed by Bcc bacteria using this method. For several feeding assays, the bacterial concentration inside the flies was recorded to verify that the flies had indeed ingested bacteria. For instance, up to 1.9×105 CFU per fly could be recovered for B. ambifaria HSJ1 twenty days after the beginning of the infection, without visible adverse effect on the flies.
Since ingested Bcc species are able to colonize the flies without harming them, it is interesting to speculate that these bacterial species could behave as endo-residents of Drosophila. To date, however, no finding of any Burkholderia species in laboratories flies ,  or in natural populations  supports this hypothesis.
Fly pricking is effective at generating a mortal infection
Given that feeding the flies with Bcc strains did not produce any mortality, assays with the nicking method were conducted.
Figure 1 shows survival curves for D. melanogaster when challenged with B. cenocepacia strain K56-2 on three different assays. Results demonstrate that the method can reveal fly killing by a Bcc strain with effectiveness. No statistical difference was observable among survival curves (Log-rank (Mantel-Cox) test), showing that this method is highly reproducible and accurate.
Pricking assays were performed in three independent replicates, each with a minimum of 30 flies. Statistical significance (Log-rank analysis (Mantel-Cox)) between survival curves is shown with *p<0.05 and ***p<0.0005.
Pricking experiments uncovered differences in terms of virulence between the Bcc species and strains, allowing discrimination between strains of one particular genomovar but also between Bcc genomovars. As a demonstration, flies were infected with five different wild-type strains of B. cenocepacia (genomovar III), revealing variability in infectious capacity between strains of the same Bcc genomovar (Figure 2). Among these strains, we tested K56-2 and J2315 two closely related strains belonging to the epidemic ET-12 lineage . We observed that J2315 is less virulent than K56-2 in the D. melanogaster model, which is in agreement with recent data from Uehlinger and colleagues who reported that J2315 is also less virulent than K56-2 in two other alternative models, G. mellonella and C. elegans. While strain LMG 18830 (CEP511) exhibits moderate virulence in the Drosophila, mouse and wax moth models, it produces very different infection outcomes in alfalfa (pathogenic) and C. elegans (non lethal) , .
Pricking assays were performed with a minimum of 30 flies for each strain. Statistical significance (Log-rank analysis (Mantel-Cox)) between survival curves is shown with *p<0.05 and ***p<0.0005.
A number of other Bcc genomovars were then tested and found to display large differences in virulence toward D. melanogaster (Figure 3). B. cepacia strains were from different sources (one was first isolated from onions while the other is a clinical strain) and yet, both are among the most virulent strains in the fly pricking model (Figure 3A and B). They kill 100% of the flies within 3 days. These results are similar to what has been obtained with other hosts, murine or alternative models alike. These data obtained with B. cepacia illustrate also the environmental isolates appear particularly virulent in alternative infection models, and sometimes more virulent than the clinical isolates as also observed with the C. elegans model .
Pricking assays were performed with a minimum of 30 flies for each strain. A: B. cepacia LMG1222, B: B. cepacia LMG18821, C: B. multivorans LMG16660, D: B. stabilis LMG18870, E: B. vietnamiensis LMG22486, F: B. vietnamiensis LMG18835, G: B. dolosa LMG21819, H: B. dolosa LMG21443, I: B. ambifaria AU0212, J: B. ambifaria CEP0996, K: B. pyrrocinia LMG21824, L: B. ubonensis LMG20358.
B. pyrrocinia LMG21824 and B. ubonensis LMG20358 kill flies in less than 75 hours, hence being among the most lethal to flies (Figure 3K and L). There are no data in mammalian hosts for these two genomovars; however, they are among the most virulent strains in the C. elegans model .
B. stabilis (Figure 3D), B. ambifaria (Figure 3I and 3J) and B. vietnamiensis (Figure 3E and 3F) can be classified as intermediate in their virulence towards the fly. These results are consistent with previous work on mammals, on alfalfa and on G. mellonella , ; the only exception being B. ambifaria CEP0996, scoring 3 on 3 in pathogenicity tests conducted with C. elegans .
At the other end of the pathogenicity spectrum, B. multivorans LMG16660 (Figure 3C) and B. dolosa LMG21819 (Figure 3G) are poorly virulent. The other tested B. dolosa strain, LMG21443 (Figure 3H), is slightly more virulent than LMG21819 but still takes more than 12 days to kill only 40% of the flies (Figure 3G and H). The latter B. dolosa strain was also avirulent in C. elegans  while strain LMG21443, more pathogenic to flies, had a LD50 of 40,000 CFU in G. mellonella  and a pathogenicity score of 2 on 3 with C. elegans .
The results obtained for B. multivorans LMG16660 (Figure 3C) are comparable to other works performed with alternative host models. For instance, no or very little mortality was observed with C. elegans  or alfalfa , even with several other strains of that species. B. multivorans strains C5393 and C1376 were tested in the rat agar bead model  and the animals once again showed very little signs of pathology. In this case, it was assumed to be because of a lower ability of the microorganism to grow in its host. However, poor growth or persistency of the bacteria cannot explain the lack of pathogenicity of B. multivorans towards flies. Indeed, 6.55×106 CFU per fly were recovered on day 8 following a septic injury. Experiments performed on Panagrellus redivivus, a nematode capable of surviving several days at 37°C revealed that B. multivorans strains were only able to kill the model at 37°C but not at 25°C suggesting that this genomovar carries virulence functions upregulated at 37°C .
Overall, there is a very good correlation between the results obtained in the fly pricking assay and those obtained with alternative or murine hosts.
Drosophila mortality correlates with bacterial growth and persistence in vivo
To verify the presence of bacteria in pricked flies throughout the infections, bacterial survival in vivo was measured for three Bcc strains. Figure 4 demonstrates that every strain tested was capable of colonizing the fly and able to replicate inside the host, although with different rates.
At the indicated time points, bacterial load was quantified from living fruit flies as described in Materials and Methods. A: B. cenocepacia LMG18830, B: B. cenocepacia K56-2, C: B. cepacia LMG18821.
The CFU/fly for the three strains measured one hour post infection were all equivalent: 4.55×104±2.73×104 CFU per fly were recovered for B. cenocepacia LMG18830 while it was 4.63×104±3.15×104 CFU per fly for B. cenocepacia K56-2 and 1.73×104±8.26×103 CFU per fly for B. cepacia LMG18821. However, while it only took 1 day for B. cepacia LMG18821 CFUs to reach 106 CFU/fly, approximately 6 days were needed for B. cenocepacia LMG18830 to reach the same CFU number in the flies. Accordingly, B. cepacia LMG18821 kills flies much faster than B. cenocepacia LMG18830 or K56-2; all LMG18821-infected flies died in less than 50 hours whereas approximately 90% of the flies for K56-2 and LMG18830 were still alive at that time during infection. These results support the hypothesis that strains displaying slower in vivo growth rates also kill flies more gradually.
Fly mortality when infected with Bcc mutants
An effective alternative infection model should allow discrimination between virulent and avirulent bacterial strains. Since the pricking assay revealed conclusive, mutants of B. cenocepacia strain K56-2 previously reported to display reduced virulence towards mammals were thereby examined in the fly (Table S1). The in vitro growth rates of the mutants had been determined to be the same than the wild-type (data not shown).
Two zinc-dependent metalloproteases, ZmpA and ZmpB, known to cleave several proteins important in host defence, are clearly involved in the virulence of B. cenocepacia K56-2. Indeed, a zmpA mutant is less virulent in a rat chronic respiratory infection model , as is also the zmpB and the double zmpA zmpB mutants . However, in C. elegans, G. mellonella and alfalfa, zmpA and zmpB mutants are as virulent as the wild-type K56-2 . Significantly, in our fly pricking model the zmpA mutant is significantly less pathogenic than the wild-type strain, while the zmpB mutant does not show a reduced pattern of mortality (Figure 5A). Moreover, infections performed with the zmpA-zmpB double mutant produce survival curves very similar to the ones obtained with the zmpA mutant. One possible explanation for the lack of contribution of ZmpB in the fruit fly is that this protease is only active when the temperature reaches 28°C . Indeed, our pricking experiments were conducted under a controlled temperature of 25°C. However, the report that a zmpB mutant was not attenuated in the G. mellonella and alfalfa models in experiments performed respectively at 30 and 37°C does not support this hypothesis .
The killing ability of wild-type B. cenocepacia K56-2 was compared to several mutants: A: zmpA-, zmpB- and zmpA-zmpB-, B: RSF12 and RSF13. C: bscN-, D: hldA-, E: cepI- and cepR-. Pricking assays were performed with a minimum of 30 flies for each strain. Statistical significance (Log-rank analysis (Mantel-Cox)) between survival curves is shown with ***p<0.0005 and ns = non-significant.
Another protease HtrA is required for survival to environmental stress . B. cenocepacia K56-2 strains RSF13 and RSF12 are deficient for the production of this protease (mutation in BCAL2829 gene encoding for the HtrA protease and mutation in a two-component regulatory system, respectively). Experiments performed in the rat agar bead model showed these two strains cannot survive in vivo during lung infections . However, we observed no difference for these two strains compared to the virulence of the wild-type strain in Drosophila survival experiments (Figure 5B). A result also obtained with C. elegans, G. mellonella and alfalfa (Table S1).
Bcc, as a wide range of bacterial pathogens, utilizes a type III secretion system to deliver virulence proteins directly into target host cells. BscN is a Type III secretion system ATP-binding protein that likely generates energy for the secretion of virulence proteins . In the mouse agar bead model, CFU recovered from the lungs and spleens were significantly lower for bscN mutants . Our trials conducted with the bscN mutant of strain K56-2 shows a partial difference in fly mortality pattern when compared to the parental strain (Figure 5C). This experiment, performed in triplicate, was repeated on two occasions and each time produced similar results: a subtle pathogenicity lag for the bscN mutant about mid-time post-infection. Interestingly, a bscN mutant of B. cenocepacia H111 also gave mixed infection outcomes when performed on C. elegans: difference between wild-type and bscN mutant could only be observed at particular time-points post-infection .
The role of LPS as a virulence factor of B. cenocepacia has been tested using strain SAL1 defective in the expression of hldA and hldD genes, which code for enzymes involved in the synthesis of complete LPS core oligosaccharides . Results in the rat agar bead model of chronic lung infection showed a reduced infectious capability for the SAL1 strain . The rats had in fact completely cleared the mutant two weeks after infection. We also found SAL1 to be attenuated in the D. melanogaster infection model (Figure 5D). Decreased virulence of this mutant was also observed in G. mellonella and C. elegans (Table S1). Mutations in LPS structure make the cells more sensitive to antimicrobial peptides, a key component of the host innate immune defense response, notably in Drosophila . Thus an increased susceptibility to antimicrobial peptides could explain the results obtained with strain SAL1.
The cepI and cepR genes are part of a quorum sensing system widely found in Bcc species . These genes affect the expression of various virulence factors, notably the zinc-dependent metalloproteases ZmpA and ZmpB , . Both cepI and cepR mutants have shown a decrease in mean percentage of lung inflammation in the rat agar bead infection model when compared to wild-type K56-2 . In our fly pricking model, the cepR and cepI mutants are not attenuated (Figure 5E). Accordingly, the cepI mutant of K56-2 is as virulent as the wild-type strain in two infection models: G. mellonella and alfalfa. However, the same mutant showed reduced killing of C. elegans . This divergence from the other infection models is probably due to AidA, a protein regulated by quorum sensing and one of the major virulence factors for C. elegans pathogenicity .
Competitive index assays
Usually, virulence with alternative models is evaluated in terms of lethality (time, or number of bacteria, LD50, required to kill the infected host). Some information may however be lost in the process. Also, if a gene mutation does not increase the mortality rate of the flies, it does not necessarily mean that the gene product does not play a role in the virulence of the bacterium. CI is a well-established, sensitive method to examine bacterial virulence in mammalian host such as the mouse or the rat , . These assays provide information on the capacity of a mutant strain to compete in vivo with the wild-type strain. For several genes investigated in our pricking assays (e.g. hldA or htrA), their role as virulence determinants had previously been demonstrated by CI analysis in the rat agar bead model , .
To provide enhanced sensitivity and discriminating power to the fly mortality tests, CI was adapted to be performed in Drosophila (see Materials and methods for details). The CI is defined as the ratio between the mutant strain and the wild-type in the output (96 h post-infection) divided by the ratio of the two strains in the input (inoculum).
In preliminary studies, we observed that the bacterial concentration used as inoculum or the choice of the time point for the output have no effect on the CI value (data not shown). Figure 6A shows three independent tests performed with the zmpA mutant that all produced very similar mean indices, confirming the reproducibility of the approach. For each experiment, the small variability between individual indices also highlights the precision of the results and further validates the method. In every case, mean CIs were all below 1, which indicates that the zmpA mutant is less competitive than the wild-type in vivo.
CI is defined as the ratio between the wild-type K56-2 and the mutant in the output (bacteria recovered from the fruit fly 96 h post infection) divided by their ratio in the input (inoculum). Each empty square represents the CI value obtained for one fly. A CI of less than 1 indicates a virulence defect. The mean of the CI is shown as a solid line. A. Three independent CI analyses performed with zmpA mutant. B. CI analyses of zmpA, zmpB and zmpA zmpB mutants. C. CI analyses of hldA, bscN, BCAL2831, cepI and cepR mutants. For htrA-, the CI was determined with strain RSF12 containing a mutation in the BCAL2831 gene.
During the pricking experiments, the zmpB mutant could not be differentiated from its parental strain (Figure 5A). However, the CI results revealed that this mutant is actually less competitive than the wild-type (Figure 6B). It was also hard to differentiate the Drosophila survival curves of zmpA from those of the double mutant zmpA zmpB. CI assays unravelled a difference between the strains: as expected, the double zmpA zmpB mutant proved less competitive than the single zmpA or zmpB mutants.
Similarly, CI assays were performed with the RSF12 strain and revealed that this mutant is less competitive than the wild-type (Figure 6C). This result is in concordance with those previously reported for rats by Flannagan et al. . In fact, in this latter study, conclusions regarding virulence were only based on CI where both RSF12 and RSF13 could not compete against the parental strain K56-2 and were completely cleared by the rats . Our data with CI in D. melanogaster confirm that HtrA is involved in bacterial survival in vivo.
Mortality assays performed with the bscN mutant showed only a partial difference with K56-2 wild-type (Figure 5C). In contrast, CI experiments clearly exposed the weakness of the bscN mutant (Figure 6C). The CI tests shed more light on the importance of type III secretion system in the pathogenesis of Bcc. CI performed with the mutant hldA (strain SAL1) produced the lowest CI observed, with a low mean result of 0.005 and a dramatic defect in in vivo survival. This result also highlights data obtained with the rat agar bead model of chronic lung infection .
Finally, the cepI mutant is less competitive than the wild-type, with a mean value of 0.42 for the CI. Interestingly, we obtained a value of 1.67 for the cepR mutant, which indicates that this mutant is more competitive than K56-2 wild-type in the fruit fly (Figure 6C). In B. cenocepacia, CepR functions both as a positive and negative regulator of virulence factors . One possibility is that CepR downregulates genes required for in vivo survival. CI in mammalian host has, to the best of our knowledge, never been performed with cepI or cepR mutants. Thus, a cepR mutant would be more competitive than the wild-type in mouse or rat models.
Taken together, not only CI experiments confirm results obtained with survival curves (e.g. for zmpA and hldA mutants) but also reveals new information not detectable by mortality tests (e.g. with zmpB, bscN or cepR mutants). We show for the first time that CI analysis can be used with invertebrate model hosts with results in accordance with those obtained with mammalian models.
Conclusion: potential of the Drosophila model
Several pathogens express virulence functions only above specific temperatures. Unlike G. mellonella or the murine model, both D. melanogaster and C. elegans cannot survive very long at 37°C . This could represent a drawback for the fruit fly as a host-model because Drosophila experiments are usually performed at temperatures of 21°C or 25°C. Studies with Salmonella enterica serovar Typhimurium have however been conducted at temperatures as high as 29°C ; and although the fly's life expectancy was somewhat shorten by the higher temperature, it still showed a difference between the mock and the true infection.
Results presented here clearly demonstrate the validity of the Drosophila pricking model in the study of Bcc virulence. As well as being a powerful tool for identifying Bcc virulence factors, D. melanogaster can enhance our understanding of host-pathogen interactions. The sequenced genome of the fly allows microarrays experiments to be performed , . Fluorescent proteins, such as GFP, have also been used before to monitor P. aeruginosa , , Escherichia coli , and Serratia marcescens  and S. Typhimurium  proliferation in the fly among others, but also to follow the expression of its immunity factors , . Such techniques could well be applied to the study of Bcc strains.
Surette and colleagues recently reported the use of D. melanogaster as a host for co-infections by P. aeruginosa concomitant with bacteria from the CF airways microflora . They discovered that some strains, non-pathogenic on their own, became infectious when in presence of P. aeruginosa. Knowing that essentially no environments are colonized by only one bacterial species, Drosophila infections opens an exciting door in the investigation of polymicrobial interactions and could most likely be used in the same fashion with the various Bcc strains .
Materials and Methods
Bacterial strains and culture conditions
Bcc strains used in this study are listed in Table 1. Escherichia coli SM10 λpir (thi-1 thr leu tonA lacY supE recA::RP4-2-Tc::Mu Kmr λpir) served as donor for conjugation experiments . Unless stated otherwise, Bcc strains were routinely grown in tryptic soy broth (TSB) (Difco) at 37°C with shaking (240 rpm) or on TSB agar plates. When required, antibiotics were added at the following final concentrations: 75 µg/ml trimethoprim and 200 µg/ml tetracycline.
Growth rates were verified with a Microbiology Bioscreen C Reader (Labsystems, Finland) in 100-well microplates using 200 µl of TSB. The optical density of the cultures was measured with the wideband 420–580 nm filter.
Construction of mutants
A 394-bp internal fragment of cepI and a 553-bp fragment of bcscN were amplified from B. cenocepacia K56-2 using the following primers (Table 2): cepIF with a KpnI site and cepIR with a XbaI site for the cepI gene, and primers bscNF and bscNR for the bscN gene with the same restriction sites. The PCR products were digested with KpnI and XbaI and ligated to the XbaI and KpnI sites of the suicide vector pKNOCK-Tet . The constructs were then electroporated into E. coli SM10 cells. The plasmids were then mobilized from SM10 into B. cenocepacia K56-2 cells by mating. Single-crossover insertion mutants were selected on TSB agar containing tetracycline. Plasmid insertion into the target gene was confirmed by PCR. The same procedure was used for the construction of the B. cenocepacia zmpA zmpB double mutant: a 547-bp internal fragment of zmpB was amplified from B. cenocepacia K56-2 using primers zmpBF with a KpnI site and zmpBF with a XbaI site. It was then cloned in the XbaI and KpnI sites of pKNOCK-Tet. B. cenocepacia K56-2 zmpA was used as the recipient strain. Single-crossover insertion mutants were selected on TSB agar containing trimethoprim and tetracycline.
D. melanogaster stock and maintenance
Wild-type Oregon R flies were used throughout this study. They were maintained on standard cornmeal sucrose medium and kept in a controlled environment of 25°C and 65% humidity with a 12-hours light cycle (Percival Scientific incubator). All experiments were conducted under these conditions unless stated otherwise.
Fly feeding assays
Oregon R wild-type flies, typically seven days old, were infected by a feeding assay modified from Chugani et al. . Plastic vials containing 5 ml of 5% sucrose and 1.5% agar were prepared. Whatman filter disks (2.3-cm diameter) were placed inside these vials on top of the agar surface. Bacterial cultures were grown in TSB to an OD600 of 4. Cells were then collected by centrifuging 1 ml of each culture at 7,500× g, washed once with 1 ml sterile PBS/5% sucrose solution and resuspended in 70 µl sterile PBS/5% sucrose. Bacterial suspensions were added to the surface of the filter paper in the plastic vials and let to dry for 15 min.
Flies starved for food and water for 7 h were anesthetised with CO2 and transferred to the vials per batches of 10 to 11 for a minimum of 30 flies per experiment. They were left to feed on the bacteria, and consequent death was recorded everyday. Control vials were inoculated with 70 µl sterile PBS/5% sucrose.
Fly pricking assays
Adult female flies of 8±2 days old were infected according to a protocol modified from Baldini et al.  and Tzou et al. . The flies were anesthetised with CO2 and pricked in the dorsal thorax with a 26S-gauge Hamilton needle previously dipped in the appropriate bacterial cell suspension. Bacterial cultures were grown in TSB to an OD600 of 2. Cells were then collected by centrifuging 1 ml of culture at 7,500× g, washed once with 1 ml of sterile 10 mM MgSO4 supplemented with 500 µg/ml ampicillin and resuspended in 1 ml of the same buffer. The addition of ampicillin to the buffer was done to prevent a possible infection with bacteria present on the surface of the fly. For every bacterial strain to be tested, a minimum of 30 flies were pricked and subsequently distributed by groups of 9 to 11 in a plastic vial containing 5 ml of 1.5% agar and 5% sucrose. Ten controls flies were pricked with a solution of 10 mM MgSO4 supplemented with ampicillin. The needle was washed between every replicate with 70%-grade ethanol and rinsed with the MgSO4 buffer. Fly survival was scored daily.
Bacterial growth in vivo
Between twenty and thirty flies were pricked with bacterial cell suspension prepared as described in the fly pricking assays section. To monitor bacterial loads of the flies during the course of an infection, the number of CFU per fly on specific days following infection were recorded as follow: flies were individually put in microfuge tubes containing 70%-grade ethanol and surface-sterilized by mixing by inversion for 1 min. They were then rinsed for 1 min. in sterile water and individually grinded in 200 µl of sterile PBS with a micropestle. The suspensions obtained were then serially diluted in 0.8% NaCl and plated on TSB agar containing 25 µg/ml gentamycin and 25 µg/ml polymyxin B. For the time point regarded as zero, flies were allowed to rest for 1 h after infection before anaesthesia and homogenization as described above. Five living flies were used for each time point in order to quantify bacterial loads.
Survival data was analyzed using Kaplan-Meier survival curves using the GraphPrism 5.0 software. Significance between survival curves was assessed using the Log-rank (Mantel-Cox) test.
Competitive index analyses
For these experiments, bacterial cultures were grown in TSB to an OD600 of 1. A 1∶1 ratio mix (approximately 106 CFU/ml of wild-type and mutant) was prepared in 10 mM MgSO4 with 500 µg/ml ampicillin. The CFUs for the two strains used in the input were counted by plating the serial dilutions of the bacterial solution using the appropriate antibiotics so to distinguish the strains. Flies were then injected with the bacterial mixture. Ninety-six hours following infection, 8 flies were sacrificed according to the method described for the measurement of in vivo bacterial growth. Strain discrimination was then performed by plating the bacterial suspension on TSB with the appropriate antibiotics. TSB agar contained 25 µg/ml gentamycin to determine total bacterial number; contained 25 µg/ml gentamycin and 75 µg/ml trimethoprim for mutants carrying a trimethoprim selection (zmpA, zmpB, zmpA zmpB, htrA, hldA mutants), or contained 25 µg/ml gentamycin and 200 µg/ml tetracycline for cepI, cepR and bscN mutants selection. The competitive index (CI) is defined as the CFU output ratio of the mutant strain when compared to the wild-type strain, divided by the CFU input ratio (inoculum) of the mutant over the wild-type.
The authors would like to thank Pamela Sokol (University of Calgary) and Miguel Valvano (University of Western Ontario) for providing bacterial strains and mutants. Oregon R flies were kindly provided by Yiorgos Apidianakis and Laurence Rahme (Massachusetts General Hospital, Harvard), and by Paul Lasko (McGill University). Special thanks to Marie Christine Groleau for technical assistance and helpful discussions.
Conceived and designed the experiments: JCV LV ED. Performed the experiments: JCV LV JT. Analyzed the data: JCV LV JT ED. Contributed reagents/materials/analysis tools: JCV LV JT. Wrote the paper: JCV LV JT ED.
- 1. Burkholder W (1950) Sour skin, a bacterial rot of onion bulbs. Phytopathology 64: 468–475.
- 2. Chiarini L, Bevivino A, Dalmastri C, Tabacchioni S, Visca P (2006) Burkholderia cepacia complex species: health hazards and biotechnological potential. Trends Microbiol 14: 277–286.
- 3. Vial L, Groleau MC, Dekimpe V, Déziel E (2007) Burkholderia diversity and versatility: an inventory of the extracellular products. J Microbiol Biotechnol 17: 1407–1429.
- 4. Aaron SD, Ferris W, Henry DA, Speert DP, Macdonald NE (2000) Multiple combination bactericidal antibiotic testing for patients with cystic fibrosis infected with Burkholderia cepacia. Am J Respir Crit Care Med 161: 1206–1212.
- 5. Weber DJ, Rutala WA, Sickbert-Bennett EE (2007) Outbreaks associated with contaminated antiseptics and disinfectants. Antimicrob Agents Chemother 51: 4217–4224.
- 6. Martin DW, Mohr CD (2000) Invasion and intracellular survival of Burkholderia cepacia. Infect Immun 68: 24–29.
- 7. Lamothe J, Huynh KK, Grinstein S, Valvano MA (2007) Intracellular survival of Burkholderia cenocepacia in macrophages is associated with a delay in the maturation of bacteria-containing vacuoles. Cell Microbiol 9: 40–53.
- 8. Marolda CL, Hauroder B, John MA, Michel R, Valvano MA (1999) Intracellular survival and saprophytic growth of isolates from the Burkholderia cepacia complex in free-living amoebae. Microbiology 145 (Pt7): 1509–1517.
- 9. Coenye T, Vandamme P, Govan JR, LiPuma JJ (2001) Taxonomy and identification of the Burkholderia cepacia complex. J Clin Microbiol 39: 3427–3436.
- 10. McClean S, Callaghan M (2009) Burkholderia cepacia complex: epithelial cell-pathogen confrontations and potential for therapeutic intervention. J Med Microbiol 58: 1–12.
- 11. Mahenthiralingam E, Baldwin A, Dowson CG (2008) Burkholderia cepacia complex bacteria: opportunistic pathogens with important natural biology. J Appl Microbiol 104: 1539–1551.
- 12. Vanlaere E, Lipuma JJ, Baldwin A, Henry D, De Brandt E, et al. (2008) Burkholderia latens sp. nov., Burkholderia diffusa sp. nov., Burkholderia arboris sp. nov., Burkholderia seminalis sp. nov. and Burkholderia metallica sp. nov., novel species within the Burkholderia cepacia complex. Int J Syst Evol Microbiol 58: 1580–1590.
- 13. Sajjan U, Thanassoulis G, Cherapanov V, Lu A, Sjolin C, et al. (2001) Enhanced susceptibility to pulmonary infection with Burkholderia cepacia in Cftr(−/−) mice. Infect Immun 69: 5138–5150.
- 14. Tomich M, Griffith A, Herfst CA, Burns JL, Mohr CD (2003) Attenuated virulence of a Burkholderia cepacia type III secretion mutant in a murine model of infection. Infect Immun 71: 1405–1415.
- 15. Sousa SA, Ulrich M, Bragonzi A, Burke M, Worlitzsch D, et al. (2007) Virulence of Burkholderia cepacia complex strains in gp91phox-/- mice. Cell Microbiol 9: 2817–2825.
- 16. Chu KK, Davidson DJ, Halsey TK, Chung JW, Speert DP (2002) Differential persistence among genomovars of the Burkholderia cepacia complex in a murine model of pulmonary infection. Infect Immun 70: 2715–2720.
- 17. Seed KD, Dennis JJ (2008) Development of Galleria mellonella as an alternative infection model for the Burkholderia cepacia complex. Infect Immun 76: 1267–1275.
- 18. Bernier SP, Silo-Suh L, Woods DE, Ohman DE, Sokol PA (2003) Comparative analysis of plant and animal models for characterization of Burkholderia cepacia virulence. Infect Immun 71: 5306–5313.
- 19. Cardona ST, Wopperer J, Eberl L, Valvano MA (2005) Diverse pathogenicity of Burkholderia cepacia complex strains in the Caenorhabditis elegans host model. FEMS Microbiol Lett 250: 97–104.
- 20. Jander G, Rahme LG, Ausubel FM (2000) Positive correlation between virulence of Pseudomonas aeruginosa mutants in mice and insects. J Bacteriol 182: 3843–3845.
- 21. Markey KM, Glendinning KJ, Morgan JA, Hart CA, Winstanley C (2006) Caenorhabditis elegans killing assay as an infection model to study the role of type III secretion in Burkholderia cenocepacia. J Med Microbiol 55: 967–969.
- 22. Adams MD, Celniker SE, Holt RA, Evans CA, Gocayne JD, et al. (2000) The genome sequence of Drosophila melanogaster. Science 287: 2185–2195.
- 23. Fallon PG, Allen RL, Rich T (2001) Primitive Toll signalling: bugs, flies, worms and man. Trends Immunol 22: 63–66.
- 24. Medzhitov R, Preston-Hurlburt P, Janeway CA Jr (1997) A human homologue of the Drosophila Toll protein signals activation of adaptive immunity. Nature 388: 394–397.
- 25. Hoffmann JA, Reichhart JM (2002) Drosophila innate immunity: an evolutionary perspective. Nat Immunol 3: 121–126.
- 26. Nicolas E, Reichhart JM, Hoffmann JA, Lemaitre B (1998) In vivo regulation of the IkappaB homologue cactus during the immune response of Drosophila. J Biol Chem 273: 10463–10469.
- 27. Apidianakis Y, Rahme LG, Heitman J, Ausubel FM, Calderwood SB, et al. (2004) Challenge of Drosophila melanogaster with Cryptococcus neoformans and role of the innate immune response. Eukaryot Cell 3: 413–419.
- 28. D'Argenio DA, Gallagher LA, Berg CA, Manoil C (2001) Drosophila as a model host for Pseudomonas aeruginosa infection. J Bacteriol 183: 1466–1471.
- 29. Chugani SA, Whiteley M, Lee KM, D'Argenio D, Manoil C, et al. (2001) QscR, a modulator of quorum-sensing signal synthesis and virulence in Pseudomonas aeruginosa. Proc Natl Acad Sci U S A 98: 2752–2757.
- 30. Lutter EI, Faria MM, Rabin HR, Storey DG (2008) Pseudomonas aeruginosa cystic fibrosis isolates from individual patients demonstrate a range of levels of lethality in two Drosophila melanogaster infection models. Infect Immun 76: 1877–1888.
- 31. Brummel T, Ching A, Seroude L, Simon AF, Benzer S (2004) Drosophila lifespan enhancement by exogenous bacteria. Proc Natl Acad Sci U S A 101: 12974–12979.
- 32. Roh SW, Nam YD, Chang HW, Kim KH, Kim MS, et al. (2008) Phylogenetic characterization of two novel commensal bacteria involved with innate immune homeostasis in Drosophila melanogaster. Appl Environ Microbiol 74: 6171–6177.
- 33. Cox CR, Gilmore MS (2007) Native microbial colonization of Drosophila melanogaster and its use as a model of Enterococcus faecalis pathogenesis. Infect Immun 75: 1565–1576.
- 34. Mahenthiralingam E, Vandamme P, Campbell ME, Henry DA, Gravelle AM, et al. (2001) Infection with Burkholderia cepacia complex genomovars in patients with cystic fibrosis: virulent transmissible strains of genomovar III can replace Burkholderia multivorans. Clin Infect Dis 33: 1469–1475.
- 35. Uehlinger S, Schwager S, Bernier SP, Riedel K, Nguyen DT, et al. (2009) Identification of Specific and Universal Virulence Factors in Burkholderia cenocepacia Strains by Using Multiple Infection Hosts. Infect Immun 77: 4102–4110.
- 36. Laws TR, Smith SA, Smith MP, Harding SV, Atkins TP, et al. (2005) The nematode Panagrellus redivivus is susceptible to killing by human pathogens at 37 degrees C. FEMS Microbiol Lett 250: 77–83.
- 37. Corbett CR, Burtnick MN, Kooi C, Woods DE, Sokol PA (2003) An extracellular zinc metalloprotease gene of Burkholderia cepacia. Microbiology 149: 2263–2271.
- 38. Kooi C, Subsin B, Chen R, Pohorelic B, Sokol PA (2006) Burkholderia cenocepacia ZmpB is a broad-specificity zinc metalloprotease involved in virulence. Infect Immun 74: 4083–4093.
- 39. Flannagan RS, Aubert D, Kooi C, Sokol PA, Valvano MA (2007) Burkholderia cenocepacia requires a periplasmic HtrA protease for growth under thermal and osmotic stress and for survival in vivo. Infect Immun 75: 1679–1689.
- 40. Loutet SA, Flannagan RS, Kooi C, Sokol PA, Valvano MA (2006) A complete lipopolysaccharide inner core oligosaccharide is required for resistance of Burkholderia cenocepacia to antimicrobial peptides and bacterial survival in vivo. J Bacteriol 188: 2073–2080.
- 41. Lewenza S, Conway B, Greenberg EP, Sokol PA (1999) Quorum sensing in Burkholderia cepacia: identification of the LuxRI homologs CepRI. J Bacteriol 181: 748–756.
- 42. Sokol PA, Sajjan U, Visser MB, Gingues S, Forstner J, et al. (2003) The CepIR quorum-sensing system contributes to the virulence of Burkholderia cenocepacia respiratory infections. Microbiology 149: 3649–3658.
- 43. Huber B, Feldmann F, Kothe M, Vandamme P, Wopperer J, et al. (2004) Identification of a novel virulence factor in Burkholderia cenocepacia H111 required for efficient slow killing of Caenorhabditis elegans. Infect Immun 72: 7220–7230.
- 44. Hunt TA, Kooi C, Sokol PA, Valvano MA (2004) Identification of Burkholderia cenocepacia genes required for bacterial survival in vivo. Infect Immun 72: 4010–4022.
- 45. Kukavica-Ibrulj I, Bragonzi A, Paroni M, Winstanley C, Sanschagrin F, et al. (2008) In vivo growth of Pseudomonas aeruginosa strains PAO1 and PA14 and the hypervirulent strain LESB58 in a rat model of chronic lung infection. J Bacteriol 190: 2804–2813.
- 46. Riedel K, Arevalo-Ferro C, Reil G, Gorg A, Lottspeich F, et al. (2003) Analysis of the quorum-sensing regulon of the opportunistic pathogen Burkholderia cepacia H111 by proteomics. Electrophoresis 24: 740–750.
- 47. Mylonakis E, Aballay A (2005) Worms and flies as genetically tractable animal models to study host-pathogen interactions. Infect Immun 73: 3833–3841.
- 48. Brandt SM, Dionne MS, Khush RS, Pham LN, Vigdal TJ, et al. (2004) Secreted Bacterial Effectors and Host-Produced Eiger/TNF Drive Death in Salmonella-Infected Fruit Fly. PLoS Biol 2: e418.
- 49. Vodovar N, Acosta C, Lemaitre B, Boccard F (2004) Drosophila: a polyvalent model to decipher host-pathogen interactions. Trends Microbiol 12: 235–242.
- 50. Apidianakis Y, Mindrinos MN, Xiao W, Lau GW, Baldini RL, et al. (2005) Profiling early infection responses: Pseudomonas aeruginosa eludes host defenses by suppressing antimicrobial peptide gene expression. Proc Natl Acad Sci U S A 102: 2573–2578.
- 51. Fauvarque MO, Bergeret E, Chabert J, Dacheux D, Satre M, et al. (2002) Role and activation of type III secretion system genes in Pseudomonas aeruginosa-induced Drosophila killing. Microb Pathog 32: 287–295.
- 52. Sibley CD, Duan KM, Fischer C, Parkins MD, Storey DG, et al. (2008) Discerning the Complexity of Community Interactions Using a Drosophila Model of Polymicrobial Infections. Plos Pathogens 4(10): e1000184.
- 53. Elrod-Erickson M, Mishra S, Schneider D (2000) Interactions between the cellular and humoral immune responses in Drosophila. Current Biology 10: 781–784.
- 54. Nehme NT, Liegeois S, Kele B, Giammarinaro P, Pradel E, et al. (2007) A model of bacterial intestinal infections in Drosophila melanogaster. PLoS Pathog 3: e173.
- 55. Dionne MS, Ghori N, Schneider DS (2003) Drosophila melanogaster is a genetically tractable model host for Mycobacterium marinum. Infect Immun 71: 3540–3550.
- 56. Tzou P, Meister M, Lemaitre B (2002) Methods for studying infection and immunity in drosophila. Methods in Microbiology. London: Academic Press. pp. 507–529.
- 57. Vial L, Déziel E (2008) The fruit fly as a meeting place for microbes. Cell Host Microbe 4: 505–507.
- 58. Simon R, Priefer U, Pühler A (1983) A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram negative bacteria. Bio/Technology 1: 784–791.
- 59. Alexeyev MF (1999) The pKNOCK series of broad-host-range mobilizable suicide vectors for gene knockout and targeted DNA insertion into the chromosome of gram-negative bacteria. Biotechniques 26: 824–828.
- 60. Baldini RL, Lau GW, Rahme LG (2002) Use of plant and insect hosts to model bacterial pathogenesis. Methods Enzymol 358: 3–13.
- 61. Mahenthiralingam E, Coenye T, Chung JW, Speert DP, Govan JR, et al. (2000) Diagnostically and experimentally useful panel of strains from the Burkholderia cepacia complex. J Clin Microbiol 38: 910–913.
- 62. Vandamme P, Holmes B, Coenye T, Goris J, Mahenthiralingam E, et al. (2003) Burkholderia cenocepacia sp. nov.-a new twist to an old story. Res Microbiol 154: 91–96.
- 63. Nelson MJ, Montgomery SO, Mahaffey WR, Pritchard PH (1987) Biodegradation of trichloroethylene and involvement of an aromatic biodegradative pathway. Appl Environ Microbiol 53: 949–954.
- 64. Coenye T, Vandamme P, LiPuma JJ, Govan JR, Mahenthiralingam E (2003) Updated version of the Burkholderia cepacia complex experimental strain panel. J Clin Microbiol 41: 2797–2798.
- 65. Vial L, Lépine F, Milot S, Groleau MC, Dekimpe V, et al. (2008) Burkholderia pseudomallei, B. thailandensis, and B. ambifaria produce 4-hydroxy-2-alkylquinoline analogues with a methyl group at the 3 position that is required for quorum-sensing regulation. J Bacteriol 190: 5339–5352.
- 66. Coenye T, Mahenthiralingam E, Henry D, LiPuma JJ, Laevens S, et al. (2001) Burkholderia ambifaria sp. nov., a novel member of the Burkholderia cepacia complex including biocontrol and cystic fibrosis-related isolates. Int J Syst Evol Microbiol 51: 1481–1490.