The Global Transcription Factor Lrp Is both Essential for and Inhibitory to Xenorhabdus nematophila Insecticidal Activity

ABSTRACT

In the entomopathogenic bacterium Xenorhabdus nematophila, cell-to-cell variation in the abundance of the Lrp transcription factor leads to virulence modulation; low Lrp levels are associated with a virulent phenotype and suppression of antimicrobial peptides (AMPs) in Manduca sexta insects, while cells that lack lrp or express high Lrp levels are virulence attenuated and elicit AMP expression. To better understand the basis of these phenotypes, we examined X. nematophila strains expressing fixed Lrp levels. Unlike the lrp-null mutant, the high-lrp strain is fully virulent in Drosophila melanogaster, suggesting that these two strains have distinct underlying causes of virulence attenuation in M. sexta. Indeed, the lrp-null mutant was defective in cytotoxicity against M. sexta hemocytes relative to that in the high-lrp and low-lrp strains. Further, supernatant derived from the lrp-null mutant but not from the high-lrp strain was defective in inhibiting weight gain when fed to 1st instar M. sexta. These data suggest that contributors to the lrp-null mutant virulence attenuation phenotype are the lack of Lrp-dependent cytotoxic and extracellular oral growth inhibitory activities, which may be particularly important for virulence in D. melanogaster. In contrast, the high-Lrp strain was sensitive to the antimicrobial peptide cecropin, had a transient survival defect in M. sexta, and had reduced extracellular levels of insecticidal activity, measured by injection of supernatant into 4th instar M. sexta. Thus, high-lrp strain virulence attenuation may be explained by its hypersensitivity to M. sexta host immunity and its inability to secrete one or more insecticidal factors.

IMPORTANCE Adaptation of a bacterial pathogen to host environments can be achieved through the coordinated regulation of virulence factors that can optimize success under prevailing conditions. In the insect pathogen Xenorhabdus nematophila, the global transcription factor Lrp is necessary for virulence when injected into Manduca sexta or Drosophila melanogaster insect hosts. However, high levels of Lrp, either naturally occurring or artificially induced, cause attenuation of X. nematophila virulence in M. sexta but not D. melanogaster. Here, we present evidence suggesting that the underlying cause of high-Lrp-dependent virulence attenuation in M. sexta is hypersensitivity to host immune responses and decreased insecticidal activity and that high-Lrp virulence phenotypes are insect host specific. This knowledge suggests that X. nematophila faces varied challenges depending on the type of insect host it infects and that its success in these environments depends on Lrp-dependent control of a multifactorial virulence repertoire.

INTRODUCTION

When initiating infection, pathogens alter gene expression profiles to adapt to the new host environment by overcoming host defenses through suppression of host immunity or avoiding detection by host immune surveillance mechanisms (1, 2). The bacterium Xenorhabdus nematophila is a pathogen of several insect groups and during infection overcomes insect immunity (3–9), which is comprised of two central tissues, hemocytes and fat body. Hemocytes include several types of blood cells that mediate cellular responses, such as phagocytosis and aggregation (nodulation) around a microbial invader (10). Fat body is the major site of synthesis for macromolecular immune factors (comprising humoral immunity), such as antimicrobial peptides (AMPs), small cationic peptides that bind to and disrupt microbial cell membranes (11).

Our group and others have shown that X. nematophila suppresses aspects of cellular and humoral insect immunity and that this process requires factors expressed by living X. nematophila cells. The latter finding is based on the facts that heat-killed X. nematophila is recognized by immune surveillance systems and induces AMP expression (4, 12, 13) and that when wild-type X. nematophila is coinjected with an immunogenic bacterial strain it prevents induction of AMP expression (4). X. nematophila can inhibit phagocytosis and nodulation by secreting phospholipase A₂ (PLA₂) inhibitors (5, 6, 8, 9, 14–16). PLA₂ catalyzes the first step in the biosynthesis of eicosanoids (17), which mediate several cellular activities, including immune signaling pathways (18). Coinjection of immunogenic heat-killed X. nematophila and benzylideneacetone (BZA; a PLA₂ inhibitor produced by X. nematophila) suppressed AMP expression in the lepidopteran insect Spodoptera exigua, indicating that eicosanoid pathway inhibition is sufficient to suppress AMP induction in Lepidoptera (12).

This idea is consistent with the reported link in the dipteran Drosophila melanogaster between eicosanoid biosynthesis and the immune deficiency (IMD) pathway signaling system, one of the signaling pathways responsible for induction of AMP expression (19). However, since X. nematophila inhibits the expression of genes encoding AMPs in lepidopterans (3, 4, 20) but not D. melanogaster insects (21, 22), it appears that X. nematophila targets a step in AMP induction that is present in Lepidoptera but not D. melanogaster. Alternatively, since X. nematophila kills D. melanogaster flies despite its inability to suppress immunity (21), it may be that X. nematophila toxicity is sufficiently powerful to negate the need for immune suppression in this insect.

X. nematophila immune suppression and virulence phenotypically vary in a process known as virulence modulation, in which a subgroup within an isogenic population is virulent and immunosuppressive toward insects while another subgroup is attenuated in both these phenotypes (4, 23). This phenomenon is controlled by the transcription factor Lrp (leucine responsive regulatory protein). Lrp is necessary for virulence since lrp mutants are attenuated for virulence and fail to suppress immunity in Manduca sexta insects (20). Virulence modulation is caused by variation in Lrp protein levels (23), such that cells expressing high levels of Lrp are attenuated in virulence and unable to prevent the induction of AMPs. Conversely, cells expressing low levels of Lrp are virulent and able to prevent AMP expression in insects (23).

We previously suggested that the virulence defect of the X. nematophila lrp mutant is due to a lack of Lrp-dependent activities, while the virulence defect of cells expressing high levels of Lrp may be a result of overstimulation of the insect immune system by increased expression of Lrp-dependent activities (23) (Fig. 1). In this model, overproduction of degradative or surface factors during initial stages of infection triggers a robust immune response in insects before the bacteria are able to produce Lrp-dependent immunosuppressive factors, and this, in turn, leads to bacterial clearance. Another theory is that high levels of Lrp inhibit (e.g., through transcriptional repression of certain loci) the production or secretion of specific virulence factors required to bring about insect death (Fig. 1). These two models are not mutually exclusive since during early infection, high levels of Lrp may cause (relative to that of the wild type) lower expression of virulence determinants and higher levels of immune elicitors. This would concomitantly reduce virulence/toxicity and trigger an immune response that clears the infection. In the work presented here, we used these models as a context to investigate the distinct underlying causes of the virulence attenuation phenotypes of an X. nematophila lrp mutant and cells expressing high levels of Lrp.

FIG 1.

Schematic representation of proposed interaction between X. nematophila fixed-Lrp strains and Manduca sexta larvae. A working model is proposed in which the attenuated virulence of the X. nematophila lrp mutant (top row) relative to cells expressing low levels of Lrp (middle row) when injected into Manduca sexta is caused by its lack of expression of Lrp-dependent virulence factors. In contrast, the attenuated virulence phenotype of cells expressing high levels of Lrp (bottom row) is proposed to be due to overexpression of surface and secreted Lrp-dependent activities that trigger an early insect immune response that successfully clears the X. nematophila infection, despite damage occurring to host tissues and cells. The relative levels of activity are indicated by plus and minus signs (+++, high level; ++, medium level; +, low level; −, insignificant levels).

RESULTS

In a previous report, we showed that high levels of Lrp are associated with an increased production of enzymes, such as lipase, protease, and hemolysin, that are thought to be involved in late stages of infection for insect tissue degradation (23–25). We hypothesized that inappropriate expression levels of these and other factors may directly elicit an immune response by virtue of their detection by the insect immune surveillance system or indirectly due to the tissue damage they cause (Fig. 1). Further, the high levels of Lrp may impact X. nematophila resistance to induced insect immune factors or its ability to express virulence factors. To test these hypotheses, we monitored phenotypes of X. nematophila strains engineered to express fixed levels of lrp: lrp-null mutant strains carrying a vector control or plasmids conferring either low or high levels of lrp expression (23).

Low levels of Lrp are sufficient for cytotoxicity.

Pathogen toxicity can damage host tissues, and this can lead to the activation of immune signaling pathways (26). Hemocytes are one of the two central tissues that comprise the insect immune system (27). To assess the effects of Lrp levels on X. nematophila toxicity, we exposed M. sexta hemocytes to supernatant from the X. nematophila fixed-Lrp strains. In agreement with a previous report (20), the lrp mutant had a defect in cytolytic activity. Surprisingly, low levels of Lrp were sufficient to damage over 80% of M. sexta hemocytes. No significant difference in cytolysis was observed between the low- and high-lrp strains (Fig. 2), suggesting low levels of Lrp are sufficient to induce production of factors that cause tissue damage in the insect host.

FIG 2.

Low levels of Lrp are sufficient for X. nematophila cytolytic activity. M. sexta hemocytes were harvested from naive larvae, washed twice, and resuspended in Grace's insect medium (GIM). Hemocytes were then incubated with supernatant from stationary-phase cultures of X. nematophila. GIM was used as a negative control for cytolytic activity. Bars represent the means of three independent experiments. Different letters denote significant difference (P < 0.05) between treatments via a one-way ANOVA with a Tukey's post hoc test.

High levels of Lrp increase sensitivity to the AMP cecropin A.

After injection into M. sexta 4th instar larvae, the X. nematophila lrp-null mutant and high-lrp strains are unable to prevent the induction of the insect AMP cecropin (4, 23). However, the impact of various Lrp levels on the sensitivity of X. nematophila to AMP induction has not been determined. To explore this question, we tested X. nematophila fixed-Lrp strains for growth in the presence of commercially available polymyxin B and cecropin A. None of the X. nematophila strains grew in the presence of polymyxin B (MIC, <0.05 μg ml⁻¹). However, in the presence of cecropin A, a low-lrp strain (with an MIC of 12 μg ml⁻¹) was more resistant than an lrp-null mutant and a high-lrp strain (both with MICs of 9 μg ml⁻¹) (Table 1). To further assess the impact of AMP exposure, we determined the growth rates of X. nematophila fixed-Lrp strains in LB supplemented with 3.6 μg ml⁻¹ cecropin A. While all of the strains showed similar specific growth rates in rich medium without cecropin, their specific growth rates were reduced by exposure to 3.6 μg ml⁻¹ cecropin. Sensitivity to cecropin was most pronounced for the high-lrp strain and least pronounced for the low-lrp strain (Table 1).

TABLE 1.

Impact of AMP exposure to X. nematophila fixed-Lrp strains

Strain	MIC (μg ml−1)		Specific growth ratea
	Polymyxin B	Cecropin A	LB	LB + 3.6 μg ml−1 cecropin A
lrp::kan vector	<0.05	9	0.28 A	0.088 AB
lrp::kan low lrp	<0.05	12	0.31 A	0.181 B
lrp::kan high lrp	<0.05	9	0.28 A	−0.036 A

^aDifferent letters indicate significant difference via a two-way ANOVA with a Bonferroni post hoc test.

Low-lrp cells are present at higher levels within insects than high-lrp and lrp mutant cells.

Since the lrp-null mutant and the high-lrp strains are incapable of suppressing cecropin AMP induction and are sensitive to this compound, we predicted that, relative to that of the low-lrp strain, they would exhibit impaired survival in M. sexta insects. To assess this, M. sexta 5th instar larvae were injected with an lrp-null mutant or low- or high-lrp strains and bacterial load was assessed in harvested hemolymph. A slightly lower bacterial load was observed for the high-lrp strain relative to that of the lrp mutant and the low-lrp strain at 24 h, but this difference was not significant. At 48 h postinjection, the lrp mutant and high-lrp bacterial load was significantly lower than that of the low-lrp strain (Fig. 3). While these data do demonstrate a transient higher bacterial load of the low-lrp strain relative to that of the lrp mutant and the high-lrp strains, the underlying cause of this difference awaits further investigation. Possibilities include differences in the timing of infection, bacterial localization, and bacterial viability.

FIG 3.

Impact of fixed Lrp levels on bacterial load in M. sexta insects. X. nematophila survival during insect infection. M. sexta 5th instar larvae were injected with ∼10⁴ stationary-phase CFU of X. nematophila fixed-Lrp strains (lrp::kan + vector, white bars; lrp::kan + low-lrp, gray bars; lrp::kan + high lrp, black bars). Three insects per sampling period (24, 48, or 72 h postinfection [hpi]) were bled for each strain, and CFU per milliliter hemolymph was determined by dilution plating. Data represent the mean and standard error of the mean of two independent experiments. Different letters indicate significant difference (P < 0.05) via a two-way ANOVA with a Bonferroni post hoc test.

Suppression of host immunity does not increase virulence of a high-lrp strain.

If the virulence defect of the high-lrp strain in M. sexta is due to early induction of or inability to suppress host immunity, we reasoned that suppression of immunity prior to infection should result in an increase in virulence. To test this hypothesis, we injected M. sexta larvae with BZA prior to injection with the high-lrp strain. Coinjection with BZA reduced AMP expression in M. sexta (Fig. 4A). However, no significant difference in virulence was observed between X. nematophila high-lrp bacteria injected alone or in combination with BZA (Fig. 4B). Therefore, the inability of the high-lrp strain to suppress immunity is not solely responsible for the attenuated virulence of this strain.

FIG 4.

Suppression of AMP expression does not increase virulence. M. sexta 4th instar larvae were injected with DMSO (control) or BZA followed by injection with 10⁴ CFU of an X. nematophila high-lrp strain. (A) Cecropin expression in fat body was measured 7 h postinjection through RT-qPCR (n = 3); (B) insect survival was monitored over time. Survival graph is a representative of three independent experiments. No significant difference (P < 0.05) was observed between survival curves of insects injected with DMSO or BZA via a log rank (Mantel-Cox) test performed on all three replicates.

Insect growth inhibition by bacterial supernatant is positively correlated with Lrp levels.

Since our data above indicate that high-lrp X. nematophila has attenuated virulence even in an immunocompromised host, we postulated that it lacks expression of one or more virulence factors necessary for host killing. Since virulence factors are often secreted into the extracellular milieu during growth in laboratory culture (28), we assessed this possibility by either feeding or injecting M. sexta larvae with supernatant from an lrp-null mutant or low- or high-lrp strains.

When fed to 1st instar larvae, all X. nematophila treatments, including the lrp::kan mutant supernatant, caused reduced weight gain in larvae after an incubation period of 72 h, indicating that X. nematophila can inhibit larval growth in an Lrp-independent fashion. However, Lrp-dependent effects were observed, such that insects fed supernatant from the high-lrp strain but not the low-lrp strain gained significantly less weight than those fed supernatant from the lrp::kan mutant (Fig. 5). These data indicate that the attenuated virulence of the high-lrp strain is not due to defects in secreted compounds that inhibit the growth of M. sexta larvae.

FIG 5.

Oral inhibitory activity is positively correlated with Lrp levels. M. sexta 1st instar larvae (n = 5) were fed with a wheat germ diet spiked with fresh LB (control) or supernatant from X. nematophila fixed-Lrp strains. The relative weight gain (RWG) of larvae was assessed 72 h post feeding. The graph is a representative of four independent experiments. Different letters denote significant difference (P < 0.05) via a one-way ANOVA with a Tukey's post hoc test.

High Lrp levels abolish insecticidal activity in X. nematophila supernatant.

When injected, supernatant from the lrp mutant and low-lrp strains showed the highest percentage of insect mortality, killing ∼80% of insects, while supernatant from the high-lrp strain killed fewer than 40% of insects (Fig. 6A). Furthermore, supernatant from the low-lrp strain rescued the insecticidal activity of the high-lrp supernatant (Fig. 6A). Given that the high-lrp strain does not have a growth defect in rich media relative to that of the low-lrp strain (Table 1), this finding may indicate that the lack of insecticidal activity in the high-lrp strain is due to the absence or low concentration of a specific bacterial factor, or factors, and not to neutralization of insecticidal activity as a result of high levels of Lrp. Indeed, we previously have shown differences in enzymatic activities among the fixed-Lrp strains (23), suggesting that concentrations of specific proteins vary for each cell type.

FIG 6.

High Lrp levels reduce insecticidal activity of X. nematophila supernatant. Cell-free supernatant (A) or concentrate and filtrate protein fractions (B) from an X. nematophila lrp mutant and low- or high-lrp strains were used for injection in M. sexta 4th instar larvae. Bars represent the mean and standard error of the means of three independent experiments. Different letters denote significant difference (P < 0.05) in the mean insect mortality via a one-way ANOVA with a Tukey's post hoc test or a two-way ANOVA with a Bonferroni post hoc test.

To obtain general characteristics of the putative factor missing from the high-lrp strain, a centrifugal filter unit was used to concentrate and separate the protein content in the supernatant of the X. nematophila fixed-Lrp strains into a fraction containing compounds with a molecular mass below 10 kDa (filtrate) and a fraction with compounds above 10 kDa (concentrate). Both fractions were used to inject M. sexta larvae. Insecticidal activity was only present in the concentrate; more than 90% of insects died when injected with the concentrate fraction of the lrp mutant and low-lrp strain (Fig. 6B). In addition, the concentrate of the high-lrp strain resulted in the mortality of ∼50% of insects, suggesting that, although an insecticidal factor is being produced by the high-lrp strain, its abundance prior to concentration is not sufficient to bring about insect death. Taken together, these data suggest that high lrp levels are detrimental for production or secretion of a bacterial factor or factors that have molecular mass over 10 kDa, with injectable insecticidal activity against the lepidopteran insect pest M. sexta.

High levels of Lrp are not detrimental for virulence in D. melanogaster.

In contrast to several species of lepidopteran larvae, in adults of the dipteran D. melanogaster, wild-type X. nematophila elicits AMP induction (21), suggesting the stage of immunity targeted by X. nematophila is absent in D. melanogaster but present in Lepidoptera. Based on this fact, we speculated that monitoring the influence of Lrp levels on X. nematophila phenotypes in D. melanogaster may reveal further insights into Lrp-dependent activities and insect-specific immune pathways. At a dose of 100 CFU/insect, the virulences of the low- and high-lrp strains in adult D. melanogaster were not significantly different from each other, and the two strains were significantly more virulent than the lrp mutant (Fig. 7A). A possible explanation for these results is that strains expressing any amount of Lrp produce factors, such as hemocyte cytotoxins (Fig. 1 and 2), sufficient to kill D. melanogaster before the induced immune factors can clear the bacterial infection. In this context, the attenuated virulence of the lrp mutant in D. melanogaster may be due to its relatively reduced cytotoxicity and virulence factor production, such that it is unable to kill D. melanogaster before being cleared itself (Fig. 1).

FIG 7.

High Lrp is not detrimental to virulence in D. melanogaster. Naive (A) and immunized (B) D. melanogaster adult flies were injected with ∼100 CFU of X. nematophila fixed-Lrp strains. To immunize flies, flies were injected with ∼300 CFU of E. coli 6 h prior to injection with X. nematophila strains. Survival was monitored over time. Graph shows the average of three independent experiments. Low-lrp and high-lrp survival curves are not significantly different from each other, but they are significantly different from the lrp mutant curve of naive flies (P < 0.05) via a log rank (Mantel-Cox) test.

To test this theory, we immunized flies by injecting them with Escherichia coli cells, which are immunogenic but not pathogenic to flies, 6 h prior to injection with X. nematophila. Consistent with previous findings (21), immunization resulted in protection of flies; about ∼60% of flies survived after injection with X. nematophila, regardless of the levels of Lrp being expressed (Fig. 7B).

DISCUSSION

For successful infection and host killing, X. nematophila must overcome host immunity, produce virulence factors, and persist. In a previous report, we showed that pathogenesis of X. nematophila requires low levels of expression of the transcriptional regulator Lrp while the absence or high levels of Lrp are detrimental to virulence (23). In the current manuscript, we examined the underlying causes of the virulence attenuation phenotypes observed in lrp mutant and high-lrp strains. Overall, relative to the virulent low-lrp strain, the high-lrp strain showed defects in virulence and immune suppression in M. sexta (23), resistance to the immune factor cecropin, survival or persistence in early stages of infection, and supernatant injectable toxicity. In contrast, this strain displayed similar cytotoxicity toward hemocytes and virulence in Drosophila, and its growth supernatant had increased inhibitory oral activity against M. sexta. Based on these phenotypes, we suggest that the attenuated virulence of the high-lrp in M. sexta results from aggregate consequences of misregulation: inappropriate elevated expression of exoenzymes and surface structures prematurely triggers immune responses that cannot be adequately suppressed. Increased sensitivity to immune effectors causes delayed establishment of the strain within the insect, and the lack of production of virulence factors prevents the strain from killing the insect rapidly.

The attenuation of the lrp mutant strain appears to be caused by a deficiency in Lrp-dependent toxic activities. First, unlike the high-lrp strain, the lrp mutant is attenuated for virulence in both M. sexta and D. melanogaster. The most dramatic differences between the lrp mutant and the virulent low-lrp strain are in cytotoxicity toward hemocytes and immune suppression in M. sexta. The inability of the lrp mutant to kill hemocytes that are likely key players in early immune induction (29), coupled with its elevated sensitivity to antimicrobials, may result in attenuated virulence in both M. sexta and D. melanogaster. In addition, in comparison to those of the low- and high-lrp strains, the supernatant of the lrp mutant showed reduced ability to inhibit the growth of M. sexta larvae when provided in food. This suggests that Lrp-dependent toxic activities, such as hemolysins (20, 30), are necessary for virulence in both M. sexta and D. melanogaster insects.

The outcome of pathogen-host interactions largely depends on the ability of the pathogen to overcome or persist in the face of host immunity. Coordinated programming for the expression of virulence or toxin-encoding genes is essential for successful pathogenesis. For example, early induction of high levels of virulence factors may be recognized by host immune surveillance systems or may cause tissue damage, both of which would trigger an immune response. X. nematophila is adept at escaping host immunity through the production of several immunosuppressive compounds (12, 16), in part through the expression of genes controlled by the transcription factor Lrp (4, 20). Nonetheless, constitutive overexpression of this transcription factor is detrimental to immune suppression and virulence in M. sexta (23).

For many pathogens, resistance to AMP activity, through such activities as surface modifications or efflux, is essential to overcoming host immunity and, therefore, for successful pathogenesis (2). However, X. nematophila sensitivity to AMPs does not necessarily result in bacterial clearance from the host. The high-lrp strain displayed higher sensitivity to cecropin than either the lrp mutant or the low-lrp strains (Table 1). The difference in AMP sensitivity between the high- and no-lrp strains may suggest that X. nematophila utilizes different mechanisms to overcome AMPs; for example, the sensitivity of the lrp mutant may be due to decreased expression of proteases, while the hypersensitivity of the high-lrp strain may be due to overproduction of surface components that attract AMPs (e.g., lipopolysaccharides). Regardless, the fact that the lrp mutant grows almost as well as the low-lrp strain in cecropin (Table 1) indicates that the virulence defect of the lrp mutant is unlikely to be due to AMP sensitivity. Similarly, we show here that chemical suppression of AMP induction in insects is not sufficient to rescue virulence of the high-lrp strain, suggesting that, although the combined inability to suppress immunity and sensitivity to its effectors may affect X. nematophila high-lrp strain survival within the insect host, this phenotype is not solely responsible for the virulence defect of this strain.

While the lrp mutant displayed attenuated virulence in both M. sexta and D. melanogaster, constitutively high levels of the Lrp transcription factor impeded killing only in M. sexta. We propose that this difference in high-lrp phenotype may indicate that X. nematophila uses different mechanisms to kill these different hosts. In D. melanogaster, suppression of AMPs does not appear to be necessary for virulence, and any overstimulation of the immune system due to overexpression of the lrp regulon early during infection does not mitigate pathogenesis. In contrast to both the low-lrp and the high-lrp strains, the lrp mutant was defective in cytolytic activity. Thus, the ability of X. nematophila to kill D. melanogaster insects may hinge upon hemocyte death. Conversely, the high-lrp strain had defects in secretion of insecticidal activity toward M. sexta insects, and this activity could be restored by supplementation with supernatant from the low-lrp strain. Therefore, the ability to kill M. sexta may require expression of a specific virulence factor that is expressed by low-lrp but not high-lrp strains. Future experiments are necessary to establish the underlying cause for the lack of insecticidal activity toward M. sexta and to identify the compound or compounds responsible for this activity.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

X. nematophila (Table 2) cultures were grown at 30°C in lysogeny broth (LB) stored in the dark or supplemented with 0.1% sodium pyruvate. Plasmids (Table 2) were maintained by adding 15 μg/ml of chloramphenicol to the media.

TABLE 2.

Strains and plasmids used in this study

Strain or plasmid	Relevant genotype	Reference
HGB1966	lrp::kan + vector (pKV69)	23
HGB1967	lrp::kan + low lrp (pEH54 vector)	23
HGB1968	lrp::kan + high lrp (pEH56 vector)	23
pKV69	Multicopy mobilizable vector; Camr, Tetr	34
pEH54	pKV69 with lrp in the opposite orientation relative to Plac	23
pEH56	pKV69 with lrp under the control of Plac	23

Antimicrobial peptide sensitivity assay.

To assess sensitivity to the antimicrobial peptide cecropin, bacterial strains were grown to an optical density at 600 nm (OD₆₀₀) of 1.0 and then diluted 1:100 in 99 μl LB. To determine antimicrobial MIC, LB was supplemented with various concentrations of cecropin A or polymyxin B obtained from Sigma (St. Louis, MO). To determine growth rate, LB was supplemented with 0 or 3.6 μg ml⁻¹ cecropin A. All cultures were kept at 30°C with shaking. The MICs were determined by visual observation after a 48-h incubation period. Specific growth rate was calculated as previously described (31). Briefly, optical density readings were substituted in the following equation: ln(X/X₀)/T, where X is A₆₀₀ at the measured time point, X₀ is A₆₀₀ at time zero, and T is time (in hours).

Insect infection and immunity assays.

Tobacco hornworm Manduca sexta larvae were raised from eggs (obtained from Carolina Biological Supply Company) on an artificial diet (Gypsy moth wheat germ diet; MP Biomedicals, Aurora, OH) with a photoperiod of 16 h.

To assess the cytolytic activity of the X. nematophila fixed-Lrp strains, hemolymph from M. sexta 5th instar larvae was collected into anticoagulant buffer (32). Hemocytes were pelleted, resuspended in the same volume of Grace's insect medium (GIM; Sigma-Aldrich, St. Louis, MO), and allowed to bind to sterile glass coverslips for 15 min at room temperature in a moist chamber. The medium was replaced with 20 μl of supernatant from stationary-phase cultures of X. nematophila fixed-Lrp strains. Samples were incubated for 1 h at 30°C in a moist chamber, and trypan blue staining was used to identify dead cells. Four microscope fields were observed per treatment per experiment. Percent cytolysis was calculated using the following formula: [1.00 – (number of blue cells, dead ÷ number of total cells)] × 100. This experiment was repeated at least three times.

For survival assays, X. nematophila fixed-Lrp strains were grown overnight in LB, subcultured 1:100 into fresh LB with chloramphenicol (15 μg/ml), and incubated at 30°C for 16 to 20 h. These cultures were washed and diluted to the desired concentration in sterile phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 1.8 mM KH₂PO₄, pH 7.4). Fifth instar insect larvae were incubated on ice for approximately 10 min prior to injection. Ten microliters of the diluted culture was injected behind the first set of prolegs of each of 12 insect larvae per experiment using a 30-gauge syringe (Hamilton, Reno, NV). Dilution plating of the inoculum confirmed that, for each treatment, an individual insect received 10⁴ CFU. Insect hemolymph was collected and pooled from three dead and/or live insects at 24, 48, and 72 h postinjection. Hemolymph samples were diluted in PBS and spread on LB agar plates supplemented with ampicillin (150 μg/ml), chloramphenicol (15 μg/ml), and tetracycline (15 μg/ml). Plates were incubated for 24 h at 30°C prior to determining the CFU per milliliter of hemolymph collected.

For BZA coinjections, bacterial cultures were prepared as stated previously. Fourth instar larvae were injected with 200 ng of BZA or dimethyl sulfoxide (DMSO). One hour post BZA or DMSO injection, 12 insects were injected with 10 μl of bacterial culture. Survival was monitored over time. Fat body tissue from two larvae was dissected 7 h postinjection. The tissue was homogenized and stored in TRIzol reagent (Invitrogen, Grand Island, NY). Five micrograms of total RNA was treated with RQ1 RNase-free DNase I (Promega, Madison, WI) and used as the template for reverse transcription using the Mg primer 5′-CGGGCAGTGAGCGCAACGTTTTTTTTTTTT-3′ (Integrated DNA Technologies, Coralville, IA) and AMV reverse transcriptase (Promega, Madison, WI). Reverse transcriptase quantitative PCR (RT-qPCR) was performed with iTaq universal SYBR green supermix (Bio-Rad, Hercules, CA) on a Bio-Rad CFX96 real-time PCR detection system using cDNA as the template. Cecropin-6 transcript levels were measured and normalized against rpS3 using the following primers: cecropin-6 F, 5′-GGTCAAAGGATTCGTGACGC-3′; cecropin-6 R, 5′-TTTGATTGTCCTTTGAAAATGGCG-3′; rpS3 F, 5′-ACTTCTCAGGCAAGGAGTGC-3′; and rpS3 R, 5′-GTCACCAGGATGTGGTCTGG-3′. Data were normalized using the formula 2^Cq(rpS3)/2^{Cq(cecropin-6)} and presented as a ratio of infected versus PBS-injected larvae (Cq, quantification cycle).

Supernatants of X. nematophila fixed-Lrp strains were assessed for their oral inhibitory activity and injectable toxicity toward M. sexta insects. For both, bacterial cultures were incubated in LB for 3 d at 30°C, and bacterial supernatant was isolated by spinning cells for 8 min at 1,500 × g and filtering through a 0.2-μm syringe filter. For oral inhibitory activity, 75 μl of supernatant was added to a 1-cm square of artificial diet. First instar M. sexta larvae (n = 5) were incubated with spiked diet or fresh LB media (control) for 72 h. The relative weight gain (RWG) of live larvae was determined as follows: 1+ [(average weight treatment – average weight control)/average weight control] as previously described (33). For injectable toxicity assays, the supernatant was separated into two protein fractions: filtrate (molecular mass, <10 kDa) and concentrate (molecular mass, >10 kDa), using an Amicon Ultra-15 centrifugal unit (Millipore, Cork, Ireland). Bacterial supernatant was concentrated 75 times for the concentrate fraction (15 ml of culture into 0.2 ml). Ten microliters of each supernatant fraction was injected into each 4th instar M. sexta larvae (n = 10). Insect mortality was determined by counting the number of dead insects 72 h postinjection. The experiment was repeated three times.

For virulence assays in D. melanogaster, adult (7 to 10 days old) flies (strain Oregon-R) were anesthetized with carbon dioxide and then injected in the thorax with 18.4 nl of either 1× sterile PBS or E. coli (100 to 300 CFU) using a Nanoject II apparatus (Drummond Scientific) equipped with glass capillaries prepared with a micropipette puller (Sutter Instruments). After 6 h, flies were reinjected with 18.4 nl of each of the X. nematophila fixed-Lrp strains (70 to 100 CFU). Following injections, flies were transferred to fresh vials with instant media at 25°C, and survival was scored at 6-h intervals up to 48 h postinjection. Survival experiments were conducted three times on different days, and each experiment involved 10 to 12 flies for each treatment.

Statistical analyses.

RT-qPCR analysis, cecropin A resistance, survival, insect mortality, cytolytic activity, and oral toxicity assays were analyzed using one-way analysis of variance (ANOVA) with a Tukey's post hoc test or a two-way ANOVA with a Bonferroni post hoc test. Insect survival curves were analyzed by a log rank (Mantel-Cox) test. All statistical analyses were performed with Prism version 5 for Macintosh (GraphPad Software, San Diego, CA).