Complete Killing of Caenorhabditis elegans by Burkholderia pseudomallei Is Dependent on Prolonged Direct Association with the Viable Pathogen

Background Burkholderia pseudomallei is the causative agent of melioidosis, a disease of significant morbidity and mortality in both human and animals in endemic areas. Much remains to be known about the contributions of genotypic variations within the bacteria and the host, and environmental factors that lead to the manifestation of the clinical symptoms of melioidosis. Methodology/Principal Findings In this study, we showed that different isolates of B. pseudomallei have divergent ability to kill the soil nematode Caenorhabditis elegans. The rate of nematode killing was also dependent on growth media: B. pseudomallei grown on peptone-glucose media killed C. elegans more rapidly than bacteria grown on the nematode growth media. Filter and bacteria cell-free culture filtrate assays demonstrated that the extent of killing observed is significantly less than that observed in the direct killing assay. Additionally, we showed that B. pseudomallei does not persistently accumulate within the C. elegans gut as brief exposure to B. pseudomallei is not sufficient for C. elegans infection. Conclusions/Significance A combination of genetic and environmental factors affects virulence. In addition, we have also demonstrated that a Burkholderia-specific mechanism mediating the pathogenic effect in C. elegans requires proliferating B. pseudomallei to continuously produce toxins to mediate complete killing.


Introduction
Caenorhabditis elegans is well established as a facile model to decipher the genetic factors mediating pathogen-host interactions. To date, a wide and still expanding range of pathogens have been reported to infect C. elegans including the Gram-negative bacteria Pseudomonas aeruginosa [1], Salmonella typhimurium [2,3], Serratia marcescens [4] and Burkholderia species [5,6], the Gram-positive bacteria Staphylococcus aureus and Enterococcus faecalis [7] as well as fungi such as Drechmeria coniospora [8] and Yersinia pseudotuberculosis [9]. In the case of P. aeruginosa and B. cepacia, nematode killing is growth medium-dependent and killing is associated with the proliferation of pathogenic bacteria within the digestive tract and the production of toxins. S. typhimurium, S. marcescens and E. faecalis are capable of stably colonizing and establishing persistent infection in the nematode intestinal tract. D. coniospora adheres to the surface in the region of the mouth and vulva whilst Y. pseudotuberculosis attaches to the cuticle in the head region and forms a biofilm which covers the mouth of C. elegans, thus preventing the worm from feeding.
Burkholderia pseudomallei is the etiologic agent for melioidosis, a disease endemic to south and east Asian countries as well as Australia [10]. Outbreaks of melioidosis in animals, including sheep, pigs, goats, cattle and dolphins have also been documented in endemic and non-endemic areas [11]. In areas where this bacterium is endemic, infection by B. pseudomallei has been estimated to be responsible for 20% to 30% of mortality due to septicaemia and 40% of sepsis-related mortality [12]. Human melioidosis exhibits a diverse clinical picture ranging from an asymptomatic state, to benign pneumonitis, to acute or chronic pneumonia, or to overwhelming septicaemia. The incubation period from defined inoculating events to onset of melioidosis was previously ascertained as 1-21 days [13] but the latent period has been documented to be as long as 62 years after exposure [14]. Unfortunately, treatment of B. pseudomallei infection is difficult as the bacterium is intrinsically resistant to many antibiotics [15].
Previous studies assessing killing kinetics of C. elegans by B. pseudomallei implicated the involvement of unidentified diffusible bacterial toxin(s) [5,16]. They also provided support that different environmental factors could affect C. elegans susceptibility and/or bacterial virulence [16]. In this study, we extended the investigations by determining the virulence of B. pseudomallei isolated from different sources in Malaysia on C. elegans, and by assessing the contribution of environmental factors in nematode killing. We further ascertained, by visualization of green fluorescent protein (GFP)-expressing B. pseudomallei, whether killing of C. elegans is mediated primarily by active infection of the nematode gut or by secreted toxin(s) into the media.

Results
Differential susceptibility of C. elegans to B. pseudomallei isolated from different sources In the initial experiments conducted, we observed that the worms were laden with eggs, which in some cases hatched internally, a phenotype called ''bagging of worms'' [1,2]. As the killing assay extended over the C. elegans generation time, progeny production may also interfere with the enumeration of surviving nematodes. Therefore, to fully eliminate any possible anomaly as a result of the bagging phenotype and laying of new progeny, we used sterile germ line proliferation-deficient (Glp) animals generated by knocking down the cdc-25.1 gene expression [17]. The cdc-25.1 knock-down did not affect the ability of C. elegans to survive infection as the TD mean of Glp animals (25.6160.93 hours) was not significantly different from wild type (Bristol N2) animals (23.6260.92 hours) after infection by B. pseudomallei strain Human R15 (Logrank (Mantel-Cox) test, p = 0.06). All B. pseudomallei isolates tested (Table 1) were virulent on C. elegans: all isolates grown on NG medium caused a significant reduction in adult life span with a mean time to death (TD mean ) of 18 hours to 56 hours compared to the worms exposed to the laboratory food source Escherichia coli OP50 ( Figure 1A). In mortality assays on Glp animals, Sheep 4523 isolate was the most virulent, with a TD mean of 18 (Table 2). Since all B. pseudomallei strains were grown under identical conditions, the observed variability is likely due to intrinsic differences in genetic determinants among strains. The strains utilized in this study were previously described by Lee et al. [18]. All isolates (with the exception of Human PMC2000) were used to infect mice and the lethal dose 50% (LD 50 ) assay was performed to determine virulence levels. To assess for possible concordance in B. pseudomallei virulence between the mouse and C. elegans models, we re-analyzed our previous data on infected mice with the Kaplan-meier analysis programme to obtain the TD mean for mice infected with different isolates (Table 2).
Interestingly, the bacterial virulence of a particular strain varies in different hosts, for example, the Sheep 4523 isolate, which was highly lethal to C. elegans had a low virulence in BALB/c mice. In contrast, the Ostrich 9166 isolate, which was less lethal to C. elegans, had a high virulence in BALB/c mice, suggesting that these isolates possess a distinct set of virulence factors that could potentiate different host immune responses to limit subsequent infection and further supporting that C. elegans may recognize a different component of pathogen-associated molecular patterns (PAMPs) than those recognized by the mammalian system. Nevertheless, virulence of the Human R15, Human D286, Human H10 and Rabbit 2514 isolates on C. elegans appeared concordant to their virulence in mice.
C. elegans susceptibility to B. pseudomallei infection is environmental factor-dependent As the expression of bacterial virulence factors is tightly regulated by various environmental signals [19], we determined the virulence of these isolates towards C. elegans by growing the pathogen on plates containing peptone-glucose (PG) medium which has increased concentration of media components [20]. In contrast to the observations on NG medium ( Figure 1A), worms exposed to every bacterial strain with the exception of Human H10, died significantly faster when the bacteria were grown on PG medium ( Figure 1B). The time taken for nematode killing by B. pseudomallei grown on PG medium ranges from 5 to 35 hours. None of the worms exposed to E. coli OP50 died over the course of the experiment, suggesting that the difference in worm mortality is not due to intrinsic toxicity of the PG medium.  Table 3). The data obtained above suggest that the specific dynamics of a C. elegans-B. pseudomallei interaction could be appreciably influenced by distinct environmental cues. As Mahajan-Miklos et al. [20] had previously demonstrated that expression of P. aeruginosa toxins involved in killing C. elegans is induced by high osmolarity, we repeated the experiments on Sheep 4523 and Human PMC2000 by replacing PG medium with peptone-glucose-sorbitol (PGS) medium possessing higher osmolarity. Similar killing rates were observed when B. pseudomallei Sheep 4523 and Human PMC2000 isolates grown on either PG or PGS medium were exposed to C. elegans. Sheep 4523 killed all nematodes after 12 hours post-infection on PGS medium with a TD mean of 5.6260.21 hours whereas Human PMC2000 killed all the nematodes after 20 hours post-infection on PGS medium with a TD mean of 9.7360.47 hours (Table 3). The role of diffusible toxin(s) in nematode killing To determine if diffusible secreted factors contribute to B. pseudomallei-mediated killing of C. elegans, we assayed the survival of C. elegans exposed to culture filtrates alone. This was achieved by growing B. pseudomallei isolates on a 0.22 mm nitrocellulose filter placed over the media surface. As the filter pore size permits passage of secreted molecules but not the bacteria, any observed killing of the nematode population upon removal of the filter can be ascribed to the presence of diffusible toxin(s). When the filter assay was carried out on NG medium, with the exception of Human PMC2000, killing by the other isolates was negligible ( Figure 2A). To validate the effect of B. pseudomallei secreted factors on worm mortality, we used bacterial cell-free culture filtrate and repeated the killing assay. Filtrates from all the isolates had negligible effects on worm mortality up to 76 hours of exposure ( Figure 2B). Data obtained from both the filter and bacteria cellfree culture filtrate assays above demonstrated that, for all the isolates tested, secreted molecules alone were insufficient to mediate full killing. In addition, heat-killed B. pseudomallei also failed to cause worm mortality (data not shown). We therefore conclude that the lethal effects of B. pseudomallei require direct and prolonged contact between the viable bacteria and C. elegans.

B. pseudomallei does not persistently accumulate within the C. elegans gut
To determine whether B. pseudomallei can stably colonize the C. elegans intestine, we performed the shifting assay with the most virulent clinical isolate (Human R15). One-day old adult worms were exposed to Human R15 for 4 hours, following which they were shifted to E. coli OP50 grown on NG medium. Unlike the constant exposure assay ( Figure 1A), only 20% of the worms died at the 72 hour time point. The proportion of dead worms increased to 30% and 40% when worms were fed with Human R15 for 12 hours and 20 hours, respectively, prior to shifting onto E. coli OP50 ( Figure 3A). The results indicate that brief exposure to B. pseudomallei was insufficient to cause a similar rate of killing to  that observed when C. elegans was continuously in contact with Human R15 over the entire assay. Perhaps C. elegans is capable of overcoming a brief exposure to B. pseudomallei. To monitor the fate of B. pseudomallei within the tract of the C. elegans intestine upon ingestion, we repeated the experiment using GFP-expressing B. pseudomallei (R15-GFP). We noted that R15-GFP was unable to fully colonize the wild-type worm intestinal lumen despite continuous exposure for 4 hours up to 24 hours ( Figure 3B-C). Next, we asked if R15-GFP would be able to establish an infection following initial colonization of the intestinal lumen. We increased the efficiency of initial colonization of R15-GFP bacteria in the intestinal lumen of C. elegans by using tnt-3(aj3) mutant animals, which have a defective grinder and thus allow intact bacteria to enter the intestinal lumen, but otherwise are indistinguishable from wild-type worms for immune function [21]. Using tnt-3(aj3), we noted that over 50% of worms had lumens full of bacteria, as assayed by GFP fluorescence after 6 hours of exposure (data not shown). The other worms were in various degrees of partial colonization. When the partially colonized tnt-3(aj3) animals were shifted to plates containing only E. coli, we failed to detect GFP fluorescence in the gut as early as 2 hours post-shifting to OP50 ( Figure 3D-E). Among the worms that were fully colonized with R15-GFP ( Figure 3F), less than 10% of these animals retained luminal GFP at 4-hours post-shifting. By 20 hours after the shift to OP50, GFP fluorescence, and by extension R15-GFP, could not be detected in the intestine of any of the worms ( Figure 3G). Together, these observations indicate that although B. pseudomallei is not able to accumulate in the C. elegans intestine at the gross level, nevertheless, small numbers of this bacteria are sufficient to cause a deadly effect on C. elegans.

Discussion
The virulence of B. pseudomallei natural isolates in the C. elegans host model was assessed by measuring the survival of worms fed on pure cultures of these isolates. Our studies have shown differences in virulence between natural isolates of B. pseudomallei. These results indicated that the killing efficiency could be a result of strain specific virulence factors that directly mediate pathogenicity, as previously reported for other human pathogens such as P. aeruginosa [1], S. typhimurium [2], B. cepacia [5], E. faecalis [7] and S. aureus [22]. Generally, the strain-dependent pathogenicity could be explained by different genotypic and phenotypic characteristics of the B. pseudomallei isolates. For genetic characterization, a preliminary comparative genomics analysis of Sheep 4523, Human R15, Human PMC2000, Human D286 and Human H10 has indicated genomic differences among these five isolates predominantly within the 16 genomic islands previously reported for the B. pseudomallei K96243 isolate [23] (Lye and Nathan, in preparation). Variation in the pathogenic phenotype of B. pseudomallei natural isolates could reflect either the loss or acquisition of accessory genetic material through horizontal gene transfer events, which may confer functional diversity among    [24,25].
A wide range of animal models of B. pseudomallei infection have been developed and the host species vary widely in their pathogenesis of disease as well as their susceptibility to infection [26]. Therefore, it is not surprising that in our study some of the B. pseudomallei strains did not exhibit similar levels of virulence in both C. elegans and mice. These differences may be a consequence of the allelic constellation of the host and the microbe and their interactions with physical environments [27]. Even though we only tested a limited number of strains in this study, we do not hesitate to propose the use of C. elegans as a surrogate model to investigate virulent determinants of B. pseudomallei isolates. This is clearly reflected by the isolates that do indeed demonstrate a correlation between virulence of different B. pseudomallei isolates in both C. elegans and murine models. The impact of the complex interaction between host genetics and environmental factors also contributes to the difficulties in analyzing complex diseases in humans. In the context of B. pseudomallei, the symptoms of melioidosis vary widely in different host species. Therefore, it may be necessary to use different animal species to model different aspects of disease in humans [26].
As alluded to above, in addition to genetic factors, the influence of specific environmental cues on the various bacterial species is well documented. For example, E. coli OP50 is non-pathogenic to nematodes when grown on NG medium, but is as pathogenic as E. faecalis when grown on BHI agar [7]. P. aeruginosa grown on different media exhibited distinct modes of killing termed slow and fast killing. Slow killing resembles an infection like process over an extended period of time and correlates with the accumulation of bacteria in the nematode gut whilst fast killing is mediated by the action of diffusible toxins such as pyocyanins [1]. The high salt or hyperosmolarity conditions used in the fast killing assay are akin to those in the lung environment of cystic fibrosis patients. It is possible that the expression of particular toxins that kill C. elegans is induced by such conditions [19]. Our results demonstrated that B. pseudomallei isolates cultured on PG medium killed nematodes significantly more rapidly than on NG medium in concordance with that reported by Gan et al. [16]. Surprisingly, increasing the osmolarity by adding sorbitol into PG medium (PGS) did not significantly affect the dynamics of C. elegans killing by B. pseudomallei, indicating that the observed differential effects are less likely a result of change in osmolarity alone. The obtained results postulate that addition of glucose into the PG medium does contribute an additional effect towards virulence, which might explain why diabetes is a major risk factor for B. pseudomallei infected individuals. We have recently demonstrated that addition of different carbon sources such as glucose, sucrose and arabinose into the media did not favor an accelerated growth rate for B. pseudomallei (Abdul Aziz and Nathan, unpublished data). This rules out the possibility that the accelerated killing rate is associated with enhanced bacterial growth in the presence of glucose. The existence of a link between regulation of utilizable carbon sources (fructose, glucose) and expression of virulence genes has previously been proposed [28].
Previous studies have demonstrated that proliferation and accumulation of pathogens such as P. aeruginosa, S. typhimurium and E. faecalis in the intestine of C. elegans may constitute the primary cause of the earlier death of the worms. In marked contrast, GFP-expressing B. pseudomallei failed to accumulate in the worm gut to a significant extent even up to 24 hours after feeding on R15-GFP, indicating that killing of C. elegans by B. pseudomallei is not associated with the establishment and proliferation of pathogenic bacteria within the nematode gut. In addition, when we tested for persistence of B. pseudomallei within the C. elegans intestine in the transfer experiments, unlike S. typhimurium [2], B. pseudomallei fails to establish a long-lasting infection when worms are exposed to the bacteria for short periods of time. Ruling out the infection-mediated mechanism, the diffusible toxins could be the primary cause of worm killing. In contrast, the extent of killing observed via the filter assay is still significantly less than that observed in the direct killing assay (p,0.0001) implying that secreted toxin is incapable of independently causing worm mortality. Moreover, negligible nematode lethality following exposure to bacterial cell-free culture filtrates was in agreement with that reported by O'Quinn et al. [5] and Balaji et al. [29] where they demonstrated that bacteria-free culture failed to show any deleterious effect on C. elegans. Our findings differ from Gan et al. [16] who showed that in the presence of a filter, a substantial proportion (about 50%) of nematodes on NG medium were killed by B. pseudomallei strain KHW. This can be attributed to strain differences as previously reported [30]. Both these observations propose that neither the infection-mediated nor toxin-mediated mechanism is the primary cause of worm killing by B. pseudomallei. Therefore, permanent direct interaction between the host and the pathogen is crucial for a complete lethal effect i.e. the killing of worms requires living or proliferating B. pseudomallei to continuously produce toxins in order to mediate the full killing effect. Analysis of the B. pseudomallei infected worm transcriptome will reveal why the host fails to overcome infection by this pathogen and this analysis is currently on-going.
The complex interaction between a microbial pathogen and its host is the underlying basis of infectious disease. The nematode C. elegans has been used to provide important insights into how animals perceive and defend themselves against infection. The ability of B. pseudomallei to infect a wide range of animal species indicates the potential to develop a wide range of infection models. For example, in mice, the disease can range from a chronic or apparently latent infection to an acute fulminant disease depending on the route of infection, dose and mouse strain. Nevertheless, C. elegans is obviously limited to the study of acute infections and at the present time cannot be used to model chronic infections. This limitation underscores the necessity of a variety of complementary host-pathogen model systems in order to thoroughly understand the full complexity of B. pseudomallei -host interactions.

Materials and Methods
Bacterial isolates and nematode strains B. pseudomallei isolates used are listed in Table 1. Stock cultures were stored at 280uC and routinely cultured on Ashdown media [31] at 37uC. All B. pseudomallei isolates were previously characterized based on biochemical tests as well as by 16S rRNA sequencing [18]. All the experiments involving B. pseudomallei were approved by the Universiti Kebangsaan Malaysia Animal Ethics Committee (UKMAEC) and performed in a BSL2+ level laboratory. The wild type C. elegans N2 and tnt-3(aj3) mutant strains used in this study were obtained from the Tan Laboratory at Stanford University. The nematode was propagated on nematode growth medium (NG medium) and fed on the normal food source, E. coli OP50 [32], at 16uC.

Production of Glp worms
The cdc-25.1 RNAi clone was cultured overnight in Luria-Bertani (LB) broth supplemented with 100 mg mL 21 ampicilin at 37uC. One hundred mL of a 25-fold -concentration liquid overnight culture was spotted onto NG medium supplemented with 1 mM Isopropyl b-D-1-thiogalactopyranoside (IPTG) plates and incubated at room temperature for 24 hours. Wild type N2 or tnt-3(aj3) mutant strains were made sterile by an RNAi feeding method utilizing the cdc-25.1 RNAi clone as previously described [33]. The gene cdc-25.1 encodes a CDC25 phosphatase homolog which affects embryonic viability and is necessary for cell proliferation in the germ line. In brief, gravid worms were laid on cdc-25.1 RNAi plates for 4 hours and then transferred to similar plates for an additional 4 hours of egg laying. After that, gravids were removed and eggs were left to hatch and grow in the presence of cdc-25.1 RNAi to produce sterile germ line proliferation-deficient (Glp) worms. One-day old adult hermaphrodite Glp worms reared on the cdc-25.1 RNAi clone lawn were used in the experiments unless otherwise indicated.
C. elegans survival assays C. elegans survival assays were performed as previously described [1] with minor modifications. All B. pseudomallei isolates and E. coli OP50 were grown overnight in 1 mL Brain Heart Infusion (BHI) broth and LB broth respectively at 37uC. Ten mL of an overnight culture was spread over a small area on 3.5-cm NG medium or Peptone-Glucose (PG) (1% Bacto-Peptone/1% NaCl/1% glucose/1.7% Bacto-Agar) medium plates and incubated at 37uC for 24 hours. Plates were then allowed to equilibrate to room temperature for 12-24 hours before use. Thirty age-matched Glp worms were transferred to either NG medium or PG medium plates seeded with individual B. pseudomallei isolates and incubated at 25uC. The number of live and dead worms was scored at 4-6 hour intervals. As a negative control, B. pseudomallei was replaced with E. coli OP50. Three independent experiments were performed for each isolate.

Filter assay
One typical colony of an individual B. pseudomallei isolate was inoculated into 1 mL BHI broth and incubated at 37uC overnight. Ten mL of each B. pseudomallei isolate overnight culture was spread onto a 0.22 mm pore size sterilized nitrocellulose filter (Whatmann) placed on 6.0-cm NG medium plates. Following incubation for 24 hours at 37uC, the filter together with the entire bacterial lawn was removed and thirty age-matched Glp worms were seeded onto the conditioned agar. Heat-killed E. coli OP50 was supplied to prevent starvation. Mortality of the worms was monitored at 4hour intervals. E. coli OP50 was used in place of B. pseudomallei as the negative control. Three independent experiments were performed for each isolate.
Bacterial cell-free culture filtrate assay One typical colony of an individual B. pseudomallei isolate was inoculated into 1 mL BHI broth and incubated overnight at 37uC. The bacterial culture was centrifuged at 4000 g for 15 minutes at 4uC and the supernatant collected and filtered through a 0.22 mm pore size syringe membrane filter unit (Millipore). The supernatant was supplemented with 1 mL S-basal medium [34]. The assay was performed in six-well microtiter plates containing 2 mL of the supernatant and S-basal mixture. Thirty age-matched Glp worms were transferred into the conditioned liquid medium. Heat-killed E. coli OP50 was supplied to prevent starvation. Mortality of the worms was monitored at 4-hour intervals. E. coli OP50 was used in place of B. pseudomallei as the negative control. Three independent experiments were performed for each isolate.

Shift assay
Approximately 100 age-matched Glp N2 worms were seeded on a B. pseudomallei lawn and allowed to feed. After 4, 12 and 20 hours, the worms were apportioned to three plates containing E. coli OP50 with 200 mg mL 21 tetracycline to ensure no further transfer of non-ingested B. pseudomallei. Worm mortality was scored every 12 hours. E. coli OP50 was used in place of B. pseudomallei as the negative control. For the shift assay involving GFP-expressing B. pseudomallei (Ooi and Nathan, in preparation), age matched Glp tnt-3(aj3) mutant worms fed on GFP-expressing B. pseudomallei (R15-GFP) for 6 hours were washed off from the bacterial lawns with M9 buffer and seeded on plate containing E. coli OP50 with 200 mg mL 21 tetracycline. Approximately 40 worms were placed in 100 mM levamisole on a 2% agarose pad at 0, 2, 8 and 20 hours post-shifting for microscopic examination under 400x total magnification using an upright fluorescent microscope equipped with an I3 long-pass GFP filter (Leica Microsystems, Wetzlar, Germany). Fluorescence micrographs were collected at 100x or 400x magnification using a ProgRes C10 Plus digital microscope camera (Jenoptic Laser, Jena, Germany). Each image series was captured on identical settings. Merging of microscopic photographs was done using Adobe Photoshop 7.0.

Survival analysis
For all experiments, three replicates per trial were carried out (with a total of 90 worms) for statistical purposes. Nematodes were classified as dead when they failed to respond to touch and pharyngeal pumping was no longer observed. Worms that died as a result of being stuck to the wall were censored from the analysis. The resulting data were analyzed with StatView (R) 5.0.1 software (SAS Institute, Inc) and plotted using the Kaplan-Meier Cumulative Survival Plot for Time (nonparametric survival analysis). All experiments were replicated in a comparable fashion and found to produce findings similar to the data presented.