|Home | About | Journals | Submit | Contact Us | Français|
Invertebrate animal models are experimentally tractable and have immunity and disease symptoms that mirror those of vertebrates. Therefore they are of particular utility in understanding fundamental aspects of pathogenesis. Indeed, artificial models using human pathogens and invertebrate hosts have revealed conserved and novel molecular mechanisms of bacterial infection and host immune responses. Additional insights may be gained from investigating interactions between invertebrates and pathogens they encounter in their natural environments. For example, enteric bacteria in the genera Photorhabdus and Xenorhabdus are pathogens of insects that also mutualistically associate with nematodes in the genera Heterorhabditis and Steinernema, respectively. These bacteria serve as models to understand naturally occurring symbiotic associations that result in disease in or benefit for animals. X. nematophila is the best studied species of its genus with regard to the molecular mechanisms of its symbiotic association. In this review, we summarize recent advances in understanding X. nematophila-host interactions. We emphasize regulatory cascades involved in coordinating transitions between various stages of the X. nematophila life cycle: infection, reproduction, and transmission.
One approach to investigating animal-microbe communication is the development of model systems based on naturally occurring symbioses between experimentally tractable organisms (Moran, 2006; Graf et al., 2006). This approach is relevant given that immunity (Müller et al., 2008) and responses to pathogens (Aballay & Ausubel, 2002; Mylonakis et al., 2007) are conserved processes among animals. As such, invertebrates in particular have become useful model hosts for human bacterial pathogens (Vallet-Gely et al., 2008; Kurz & Ewbank, 2007; Dorer & Isberg, 2006). The experimental track record of model invertebrates such as Caenorhabditis elegans and Drosophila melanogaster make them an obvious choice for use in such studies. The relative simplicity of these artificial models compared to their natural human counterparts, combined with an array of powerful host genetic tools, have provided insight into bacterial factors required for virulence, conserved features of host innate immunity and pathogen recognition, and conserved links between aging and immunity (Vallet-Gely et al., 2008; Dorer & Isberg, 2006).
Although these models are revealing, laboratory associations between invertebrates and human pathogens are artificial, and cannot fully represent the complexities of naturally occurring associations that have evolved through natural selection. Enteric bacteria of the genera Xenorhabdus and Photorhabdus are natural pathogens of a broad range of insects (Peters, 1996), including the laboratory model Manduca sexta. Therefore, studying their virulence mechanisms can provide a window into complicated, multi-dimensional aspects of a naturally occurring disease.
The complementary use of both artificial and natural pathogen-invertebrate models are leading to a better understanding of conserved mechanisms of pathogenesis, as well as distinct strategies used by some pathogens in adapting to the host environment. For example, toxin complexes (Tcs) were discovered in Photorhabdus based on their insecticidal activity but are also found in Yersinia spp. that are not known to infect insects (Bowen & Ensign, 1998; Waterfield et al., 2007; ffrench-Constant & Waterfield, 2006). The Yersinia Tc complexes are active against mammalian cells (Hares et al., 2008). Thus, a virulence determinant identified in an invertebrate pathogenic association paved the way toward identifying homologous determinants with distinct specificities in human pathogens.
Similarly, invertebrate models can reveal factors with conserved function regardless of pathogen or host identity. For example, the two-component regulatory system PhoPQ is utilized by Salmonella for survival in macrophages and infection of C. elegans and by Photorhabdus for infection of insects (Alegado et al., 2003; Miller et al., 1989; Derzelle et al., 2004), highlighting the idea that host environments share a common “signature” that is sensed by bacteria through conserved signal transduction cascades (Prost & Miller, 2008). PhoPQ is critical for establishment of infection in both Salmonella and Photorhabdus and for resistance to antimicrobial peptides (Bennett & Clarke, 2005; Derzelle et al., 2004; Gunn et al., 1998; Guo et al., 1997). In contrast, Xenorhabdus nematophila requires PhoPQ for resistance to antimicrobial peptides (AMPs), but not for successful infection of insects (C. Lipke, 2006, M.S. thesis, UW-Madison). Instead, this bacterium prevents insect antimicrobial peptide induction (Park et al., 2007) apparently reducing the need for PhoPQ-dependent resistance.
As in S. enterica, the Y. pestis PhoPQ system positively regulates virulence and survival during infection of mammalian cells (Grabenstein et al., 2006). Concomitantly, PhoPQ negatively regulates hms genes necessary for biofilm formation (Itoh et al., 2005) critical for colonization of the flea vector but not mammalian cells (Hinnebusch et al., 1996; Sun et al., 2009). Genetic analysis of bacterial and host factors involved in biofilm formation was facilitated by the fact that Y. pestis also forms hms-dependent biofilms on C. elegans (Jarrett et al., 2004; Darby, 2008). The insect pathogen X. nematophila also forms hms-dependent biofilms on the heads of C. elegans, but apparently does not require this activity for association with its natural hosts (Drace & Darby, 2008). These findings leave open the question of what roles PhoPQ regulation and biofilm formation play in the life cycle of X. nematophila, and highlight the value of comparing pathogen activities in multiple host models
Invertebrate models also are providing insights into conserved mechanisms of mutually beneficial host-microbe interactions. The Xenorhabdus and Photorhabdus models provide a rare opportunity to explore both pathogenic and mutualistic interfaces, since these insect pathogens also have a second, beneficial symbiosis with nematodes. Here we emphasize X. nematophila regulatory hierarchies that control pathogenesis and immune modulation, as well as the transition to the mutualistic state. The reader is referred to recent reviews for information on Photorhabdus biology and a comparison of Photorhabdus and Xenorhabdus symbiosis (Clarke, 2008; Goodrich-Blair & Clarke, 2007)
Among Xenorhabdus spp. X. nematophila is the best studied with regard to molecular mechanisms of symbiosis (Goodrich-Blair, 2007; Herbert & Goodrich-Blair, 2007). Approximately 30–200 X. nematophila cells occupy a specific region of the intestinal lumen of a soil-dwelling, infective juvenile form of its nematode host, Steinernema carpocapsae. The nematode transports its bacterial symbiont into the insect blood and releases it by defecation (Herbert & Goodrich-Blair, 2007) (Fig. 1A). There, X. nematophila begins to reproduce and launches an assault on the immune system and tissues of the insect that leads to insect death and degradation. Indeed, X. nematophila is virulent in the absence of its nematode host when experimentally injected into insects (Forst et al., 1997; Herbert & Goodrich-Blair, 2007). Degraded products provide a nutrient source for X. nematophila and nematodes (Fig. 1B), and when these nutrients become limiting the bacteria and nematodes re-associate and exit the cadaver to search for their next host (Forst et al., 1997) (Fig. 1C). The X. nematophila life cycle can thus be divided into three phases: insect infection, bacterial and nematode reproduction, and transmission (i.e. re-association) (Fig. 1).
Three global regulators that each play a role in multiple phases of the X. nematophila life cycle are the leucine-responsive regulatory protein Lrp, the two-component system CpxRA, and the LysR-type regulator LrhA (Fig. 1). lrp mutants have defects in all three phases of the X. nematophila life cycle (Cowles et al., 2007), lrhA mutants display defects in virulence (Richards et al., 2008) and support of nematode reproduction (G. Richards and H. Goodrich-Blair, in prep.), and cpxR mutants are defective in infection and transmission (Herbert et al., 2007). This pleiotropy suggests these regulators control gene expression during transitions between X. nematophila life stages and adaptation to the shifting host environment.
lrp, cpxR, and lrhA mutants each have a defect in killing M. sexta insects. Consistent with this, Lrp, CpxR, and LrhA are part of a regulatory hierarchy that controls many of the activities and behaviors implicated in X. nematophila virulence (Fig. 1). Lrp controls motility, antibiotic production, protease activity, the hemolysins encoded by xhlA and xaxAB, the lipase encoded by xlpA, and the transcription factor encoded by lrhA (Cowles et al., 2007). Like Lrp, LrhA is necessary for flagellar motility and lipase activity and has a positive effect on transcription of the flagellar regulator FlhDC (Richards et al., 2008), which in turn leads to synthesis of the flagellar apparatus (Givaudan & Lanois, 2000). In addition, LrhA controls expression of the XlpA lipase at the levels of both transcription and secretion through the flagellar apparatus (Richards et al., 2008). CpxR is the response regulator of a two-component system and, like LrhA and Lrp, positively affects motility, XlpA, and lrhA expression (Herbert et al., 2007). However, unlike LrhA and Lrp, CpxR has a negative effect on hemolysin activity, as well as protease and antibiotic activities, through an as yet unknown mechanism (Herbert et al., 2007). The phenotypes described above are depicted in the simplified regulatory hierarchy model shown Fig. 1.
lrhA mutants have a severe virulence defect (Cowles et al., 2007; Herbert et al., 2007; Richards et al., 2008), suggesting LrhA plays a central role in controlling virulence gene expression. This idea is supported by the fact that virulence attenuation of a cpxR mutant is rescued by constitutive expression of lrhA (Tran and Goodrich-Blair, in revision). Only one LrhA-dependent gene, xptD1, which is predicted to encode a subunit of a Tc toxin, is known thus far to play a role in insect killing (Richards et al., 2008). Xenorhabdus Tc toxins have oral insecticidal activity (Sergeant et al., 2006), and xptD1 mutants have attenuated virulence when injected into M. sexta insects, although not as severe as the virulence defect of an lrhA mutant (Richards et al., 2008). Furthermore, the oral toxicity of X. nematophila xptD1, lrhA, and lrp mutants is lower than wild type cultures (Richards and Goodrich-Blair, in prep.), indicating a Tc toxin regulated by both Lrp and LrhA is involved in X. nematophila infection.
Insects possess immunity including humoral AMP induction and cellular responses consisting of phagocytosis and nodule formation around invading organisms (Kanost et al., 2004). X. nematophila inhibits transcription of AMPs and also can prevent the formation of nodules around the invading bacterial cells (Ji & Kim, 2004; Park et al., 2003; Park et al., 2004; Park et al., 2007). The latter activity is mediated by a soluble inhibitor of insect phospholipase A (Kim et al., 2005). However, the mechanism by which AMP gene expression is suppressed remains obscure. Both cpxR (Herbert and Goodrich-Blair, in revision) and lrp (Cowles et al., 2007) mutants have defects in suppressing AMP induction. lrp mutants are also defective in suppression of nodule formation (Cowles et al., 2007), but this ability has not been measured for cpxR mutants. Interestingly, lrhA mutants are able to suppress both cecropin transcription and nodulation activities (Richards and Goodrich-Blair, in prep.), indicating Lrp-dependent immune suppression factors lie outside the LrhA regulon. As yet, X. nematophila genes necessary for immune suppression have not been identified. The finding that such elements are Lrp-dependent but LrhA-independent may facilitate their identification through microarray analysis.
Within the insect cadaver X. nematophila promotes its own growth and that of its nematode host (Sicard et al., 2003; Mitani et al., 2004). Nematode reproduction may be influenced by the capacities of X. nematophila to provide nutrients and inhibit competing microorganisms. The in vivo nutritional requirements of the nematode host are largely unknown but include lipids (Wouts, 1981; Qiu et al., 2000). It has been presumed that activities within the Lrp, CpxR, LrhA and FlhD regulons (e.g. lipases, proteases, and hemolysins) contribute to insect tissue degradation for use by the nematodes. Consistent with this idea, X. nematophila lrp (Cowles et al., 2007) and lrhA (Richards and Goodrich-Blair, in prep) mutants support the production of fewer nematode progeny than wild type. Furthermore, the X. nematophila Lrp-, CpxR-, LrhA- and FlhD-dependent lipase XlpA recently was shown to be required for nematode progeny production in the insect host (Richards and Goodrich-Blair, in prep.). This finding indicates XlpA is needed for degradation of insect lipids that contribute to nematode reproduction.
The results described above suggest a model in which CpxR, Lrp, and LrhA regulate activities necessary for fitness of X. nematophila and its nematode host both during and after successful establishment of an infection. The model presented in Fig. 1 invokes a temporal regulatory progression in which subsets of regulon members are expressed at distinct times by X. nematophila within the insect. For example, based on its known contribution to virulence (Cowles & Goodrich-Blair, 2005), xlhA may be expressed during infection, while xlpA expression may occur later, during nematode reproduction. However, these ideas remain to be experimentally tested by monitoring X. nematophila gene expression at distinct life cycle phases. In addition, it will be informative to examine the contribution to nematode reproduction of certain activities (e.g. hemolysin and motility) that lack apparent roles in virulence.
Nematode fitness can be reduced by the presence of non-symbiont bacteria, including non-native Xenorhabdus species (Sicard et al., 2004). Thus, X. nematophila support of nematode reproduction may be partially due to its out-competing of other microbes. X. nematophila competition with other Xenorhabdus is thought to be mediated through antimicrobial bacteriocins (Boemare et al., 1992). One such bacteriocin, xenocin, is induced in response to mitomycin C exposure, iron limitation, and nutrient depletion (Singh & Banerjee, 2008). However, it is not known if the global regulators discussed here influence production of xenocin, nor has the role of xenocin in the X. nematophila life cycle been examined.
Xenorhabdus species also produce antibiotics effective against other bacterial genera (Jarosz, 1996; Li et al., 1997; McInerney et al., 1991; Paul et al., 1981; Sundar & Chang, 1993). As indicated above, X. nematophila antibiotic activity against Gram-positive bacteria is positively regulated by Lrp, negatively regulated by CpxR, and independent of the flagellar pathway. However, like the flagellar pathway, antibiotic activity is repressed by the two-component regulator OmpR (Park & Forst, 2006). While an analysis of antibiotic-defective mutants is needed to clarify the function of these factors in X. nematophila symbiosis, a likely function is suppression of microorganisms that may be released from the insect gut and other surfaces. Understanding how differential regulation by CpxR, Lrp, and OmpR contributes to the timing of antibiotic expression also will yield insights into antibiotic contributions to X. nematophila fitness.
High nematode population and low nutrient densities are thought to trigger initiation of the next phase of the X. nematophila life cycle: colonization of the infective juvenile stage of the nematode that will transmit it to the next insect host (Popiel et al., 1989) (Fig. 1). CpxR and Lrp appear to mediate this transition, since cpxR and lrp mutants show attenuated colonization of the infective juvenile nematode (Cowles et al., 2007; Herbert et al., 2007). X. nematophila nilBC genes encode colonization factors (Cowles & Goodrich-Blair, 2004; Cowles & Goodrich-Blair, 2008; Heungens et al., 2002) whose transcription is positively influenced by CpxR (Herbert et al., 2007) and repressed by Lrp in conjunction with a small transcription factor, NilR (Cowles & Goodrich-Blair, 2004; Cowles & Goodrich-Blair, 2006). Thus, nil gene expression may be repressed by Lrp at early stages of infection and growth when they are not needed, and induced by CpxR at late stages of the reproductive cycle, when the host environment is spent and the transmission to a new host is beneficial.
In addition to nilBC, NilR recently was shown to negatively regulate expression of the putative lipase-encoding gene estA (Richards and Goodrich-Blair, in prep.). The biological function of EstA is currently unknown; it is not required for virulence against M. sexta or colonization of S. carpocapsae, and unlike XlpA lipase, it is not required to support nematode progeny production (Richards and Goodrich-Blair, in prep.). However, co-regulation of estA with nil genes suggests that this lipase may be important at later stages of reproduction, perhaps for nutrient acquisition in a nutrient-deficient environment in which nematode progeny infective juveniles are forming. Thus individual X. nematophila lipases may have distinct biological functions that requires their expression at different times in the life cycle: XlpA lipase during nematode reproduction and EstA during entry into the transmission phase.
In the model shown in Fig. 1 the X. nematophila global regulators CpxR, LrhA, and Lrp contribute to expression of distinct genes in each of the multiple phases of the X. nematophila life cycle. Signals that modulate regulator activity to alter gene expression patterns may be nutritional cues that reflect the changing environmental conditions within the insect. In general, the metabolic profile of a host environment is a likely source of information regarding the status of infection, as it undergoes dramatic transformations during pathogenesis and upon host death. For example, mice infected with the parasite Plasmodium berghei show changes in their metabolic profile as the infection progresses (Li et al., 2008). Therefore the metabolic profile can serve as an inherent indicator of the temporal progression of infection that could be used by pathogens, including X. nematophila, as a signal for switching from expression of virulence determinants to catabolic activities (Schaible & Kaufmann, 2005). Nutrient consumption will result in yet another metabolic profile shift, which may serve as a signal to express genes necessary for transmission.
Several lines of evidence support the idea that Lrp, CpxR, and LrhA sense and respond to nutritional conditions. First, their homologs in other systems respond to specific metabolites and general nutrient conditions. For example, Escherichia coli Cpx senses flux through the Pta-AckA pathway, and is influenced by changes in the external environment such as the presence of certain metals (Wolfe et al., 2008; Yamamoto & Ishihama, 2006). E. coli Lrp controls various aspects of amino acid metabolism (Yokoyama et al., 2006) and can respond to leucine (Calvo & Matthews, 1994). Lrp recently was shown to mediate enterohemorrhagic E. coli responses to butyrate, a product of microbial fermentation in the distal ileum (Nakanishi et al., 2009). E. coli LrhA negatively regulates genes encoding aspartate and galactose chemoreceptors (Lehnen et al., 2002), and the Yersinia LrhA homolog, RovM, is upregulated during growth on minimal medium (Heroven & Dersch, 2006), implying a metabolic connection. However, although many LysR-type regulators are known to require metabolic co-inducers for activation (Schell, 1993), no LrhA homolog co-inducer has yet been found.
Direct evidence that X. nematophila Lrp, CpxR, and LrhA sense the nutritional composition of the host environment comes from the fact that X. nematophila mutants lacking each of these regulators display metabolic defects. X. nematophila cpxR mutants have a longer lag phase than wild-type during growth in LB and hemolymph (Herbert et al., 2007) and delayed systemic infection in vivo relative to wild type (Tran and Goodrich-Blair, in revision). Although lrp and lrhA mutants have wild type growth rates in LB, lrp mutants exhibit delayed growth upon transition from rich to poor nutrient sources (Cowles et al., 2007; Richards et al., 2008) and an lrhA mutant requires certain amino acids for growth on minimal media (Richards and Goodrich-Blair, in prep.). Thus, the primary contribution of X. nematophila CpxR, Lrp, and LrhA to X. nematophila symbiosis may be to link nutritional cues within the host environment to appropriate expression of virulence determinants, degradative activities, or nematode colonization factors.
X. nematophila nutritional requirements as well as the metabolic profiles of its host environments may give clues regarding the signals sensed by its global regulators. M. sexta hemolymph contains sugars and amino acids (Phalaraksh et al., 1999) and P. luminescens upregulates genes encoding histidine, ethanolamine, and tagatose catabolism upon infection, suggesting these nutrients are available within the insect host (Munch et al., 2008). Nucleosides in the insect environment are also utilized by X. nematophila, based on the fact that nucleoside-scavenging functions confer a competitive advantage to X. nematophila during infection (Orchard & Goodrich-Blair, 2005). X. nematophila yigL, which is predicted to encode a sugar phosphatase, is required for virulence and expression of XlpA lipase activity, implying the importance of metabolic activities during infection and support of nematode production (Richards and Goodrich-Blair, submitted).
Although the metabolic activity of X. nematophila during the transition from reproductive to transmission states has not been investigated, some insights have been gleaned regarding nutrients available once X. nematophila has colonized the infective juvenile nematode. The nematode environment has nucleosides, as well as sufficient levels of some amino acids and vitamins to support growth of auxotrophs, but apparently lacks sufficient quantities of methionine or threonine (Martens et al., 2005; Orchard & Goodrich-Blair, 2005). Further characterization of Lrp, CpxR, and LrhA regulons and their changes in expression during X. nematophila life stages should yield additional hints of metabolic pathways critical for X. nematophila adaptation and the environmental signals to which X. nematophila regulators respond.
Modulation of X. nematophila global regulators in response to signal variation has not been experimentally demonstrated. One likely mechanism is direct modulation by ligand binding (e.g. specific metabolites) or protein modification (e.g. phosphorylation). In addition, other transcription factors can alter the regulation of a subset of global regulon members. For example, as described above, NilR functions synergistically with Lrp to repress expression of nil colonization factors (Cowles & Goodrich-Blair, 2006). Therefore, signals that prevent NilR activity are expected to cause de-repression of nil gene expression during the transmission phase of the X. nematophila life cycle, even in the presence of Lrp. As another example, Lrp-dependent transcriptional changes during the transition from early infection to nutrient acquisition may be mediated by FliZ (Lanois et al., 2008). In a FlhDC-dependent manner fliZ is co-transcribed with fliA, which encodes the flagellar sigma factor (Park & Forst, 2006). FliA and FliZ affect expression of different flagellar genes and extracellular enzyme activities: FliA is necessary for expression of flagellar structural genes and xlpA (Park & Forst, 2006), while FliZ positively controls expression of flhDC (forming a positive feedback loop for its own expression) and the hemolysin-encoding genes xhlBA and xaxAB (Lanois et al., 2008). Both Lrp and FliZ bind directly to the xhlBA promoter, indicating the two may interact at this locus to regulate expression. Furthermore, like lrp mutants, fliZ and xhlBA mutants have virulence defects (Cowles et al., 2007; Lanois et al., 2008). Therefore, upon successful infection, FliZ inactivation would shift Lrp-dependent expression from genes encoding virulence factors (xhlA) to those encoding degradative activities (xlpA).
An additional regulator that modulates the Lrp- and LrhA-dependent influence on the flagellar pathway is the two-component system EnvZ/OmpR (Park & Forst, 2006). OmpR negatively affects the flagellar pathway, including motility, lipase, protease, hemolysin and antibiotic activity. Although the signals to which EnvZ responds are not known, it has been proposed that OmpR repression of the flagellar regulon may be alleviated after the insect is dead (Park & Forst, 2006). This idea supports the notion that degradative enzymes in the flagellar regulon are expressed at this point in order to support nutrient acquisition and nematode production.
X. nematophila research has established regulatory and structural factors required for its natural symbiotic interactions. The global regulators Lrp, CpxR, and LrhA coordinate activities necessary for pathogenesis and mutualism, and we argue here that they sense the nutritional status of the insect host environment. However, much remains to be learned regarding the specific metabolic cues that trigger regulatory transitions. Furthermore, the regulatory model presented here highlights the need to discern temporal changes in symbiotic factor expression and function that allow X. nematophila to move seamlessly through the multiple stages of its associations with invertebrates.