Examination of Xenorhabdus nematophila Lipases in Pathogenic and Mutualistic Host Interactions Reveals a Role for xlpA in Nematode Progeny Production

Abstract

Xenorhabdus nematophila is a gammaproteobacterium and broad-host-range insect pathogen. It is also a symbiont of Steinernema carpocapsae, the nematode vector that transports the bacterium between insect hosts. X. nematophila produces several secreted enzymes, including hemolysins, lipases, and proteases, which are thought to contribute to virulence or nutrient acquisition for the bacterium and its nematode host in vivo. X. nematophila has two lipase activities with distinct in vitro specificities for Tween and lecithin. The gene encoding the Tween-specific lipase, xlpA, has been identified and is not required for X. nematophila virulence in one insect host, the tobacco hornworm Manduca sexta. However, the gene encoding the lecithin-specific lipase activity is not currently known. Here, we identify X. nematophila estA, a gene encoding a putative lecithinase, and show that an estA mutant lacks in vitro lipase activity against lecithin but has wild-type virulence in Manduca sexta. X. nematophila secondary-form phenotypic variants have higher in vitro lecithinase activity and estA transcript levels than do primary-form variants, and estA transcription is negatively regulated by NilR, a repressor of nematode colonization factors. We establish a role for xlpA, but not estA, in supporting production of nematode progeny during growth in Galleria mellonella insects. Future research is aimed at characterizing the biological roles of estA and xlpA in other insect hosts.

Xenorhabdus nematophila is a gammaproteobacterium that maintains a mutually beneficial symbiosis with the nematode Steinernema carpocapsae (19, 28, 56), with which it causes disease and death of a broad range of insects, including lepidopterans (butterflies and moths), coleopterans (beetles and weevils), dipterans (flies), and orthopterans (grasshoppers, crickets, and locusts) (19, 20, 28). Many of these insect hosts, such as the tobacco hornworm, Manduca sexta, and the greater wax moth, Galleria mellonella (a parasite of beehives), are agricultural pests, and X. nematophila is used as a biocontrol agent (15, 17, 35, 47, 55).

The free-living form of the nematode, called the infective juvenile (IJ), serves as a vector to transport the bacterium between insect hosts (19, 28, 56). Once inside the insect, X. nematophila manufactures an arsenal of toxins and extracellular degradative enzymes, including hemolysins, lipases, and proteases, which are thought to contribute to its broad host range and may aid in virulence or decomposition of host tissues (7-9, 13, 19, 46, 53). The insect cadaver serves as a nutrient source for the growing bacteria and nematodes, which reproduce within the insect (19, 28). When nematode and bacterial populations exhaust these nutrients, the bacteria reassociate with the nematodes, which enter the IJ form of their life cycle and leave the cadaver in search of a new insect host (19, 28).

Although many of the specific in vivo nutritional requirements of both organisms are unclear, successful nematode reproduction requires lipids (22, 41, 58) and the presence of X. nematophila; S. carpocapsae nematodes that are not colonized by their bacterial symbiont produce significantly fewer nematodes than those that are colonized (22, 34, 48). The X. nematophila nutrients, signals, or other activities necessary for nematode reproduction and IJ development are unknown (31) but may include its many secreted enzymes. X. nematophila produces at least three distinct hemolysins, two lipases, and two proteases (7-9, 13, 18, 19, 46, 53). These activities can be distinguished in vitro based on substrate specificity. For example, one lipase has a primary activity against Tween, while the second is active against lecithin (7, 19, 50, 53). Some but not all of these activities have been genetically characterized. The Tween-specific lipase of X. nematophila is encoded by xlpA (39, 42), while the genetic locus encoding the lecithinase is unknown. However, the lecithinase activity of the F1 strain of X. nematophila has been biochemically characterized and shown to preferentially act on the substrate phosphatidylcholine (53). In addition, a role in virulence against particular insect hosts has been assigned for some enzymes but not others. The hemolysin XhlA contributes to virulence against M. sexta (13) and the cotton leafworm Spodoptera littoralis (30), while the Tween-specific lipase XlpA does not contribute to virulence in M. sexta (39, 42).

The expression of many of the secreted enzymes discussed above is affected by X. nematophila phenotypic variation. Wild-type X. nematophila exhibits two distinct colony forms that can be distinguished in most strains by binding (primary) or not binding (secondary) to bromothymol blue dye (19). In addition, the secondary form of X. nematophila strain ATCC 19061 lacks hemolytic activity but exhibits an increase in lipase activity against Tween compared to that exhibited by the primary form (11, 16, 57). These phenotypic forms do not have distinguishing effects on X. nematophila-host interactions; both forms are able to kill insects and colonize nematodes (11, 49, 57). However, in competition assays, the primary form is isolated more readily from nematodes and the secondary form more readily from insect cadavers, indicating that phenotypic variation may reflect adaptations to each host environment (12).

Although many X. nematophila-secreted enzymes have been proposed to contribute to the degradation and utilization of insect host tissues, in most cases a biological role has not been definitively determined. For example, while xhlA and xlpA mutants have been tested for insect virulence and colonization of nematodes after in vitro cocultivation, it is unknown if they have a defect compared to wild-type X. nematophila in supporting nematode progeny production during insect infection (13, 39, 42). In some other gram-negative pathogens, lipases have been shown to play a role in nutrient acquisition (38, 43). To assess the role of lipolytic activities in the biology of X. nematophila, we identified, mutated, and characterized the regulation of estA, encoding lecithinase activity, and analyzed the role of estA and xlpA in supporting nematode production within G. mellonella insects.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Bacterial strains and plasmids utilized in these experiments are listed in Table 1. Luria-Bertani (LB) broth (33) was used to culture bacteria at 30°C. Media used to grow X. nematophila strains were either supplemented with 0.1% sodium pyruvate or stored in the dark (59). Unless stated otherwise, plasmids were introduced into X. nematophila strains through conjugation with Escherichia coli S17-1 (λpir) as described previously (4, 56). Antibiotic resistance markers of plasmids and strains were selectively maintained at the following concentrations: ampicillin (Amp), 150 μg ml⁻¹; chloramphenicol (Cm), 30 μg ml⁻¹; kanamycin (Km), 50 μg ml⁻¹; and streptomycin (Sm), 25 μg ml⁻¹.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)	Source or reference
Strains
S17-1 (λpir)	E. coli donor strain for conjugations	51
DH5α (λpir)	E. coli general cloning strain	45
SM10 (λpir)	E. coli general cloning strain	51
HGB007	X. nematophila wild-type ATCC 19061 (acquired in 1996)	ATCC
HGB009	Bacillus subtilis AD623	A. Driks
HGB081	X. nematophila wild-type AN6/1 Rifr	S. Forst
HGB800	X. nematophila wild-type ATCC 19061 (acquired in 2003)	ATCC
HGB1300	HGB800 ΔxlpA2::Km	42
HGB1301	HGB800 ΔestA1::Km	This study
HGB1302	HGB1301 ΔestA1::Km ΔxlpA3::Sm	This study
HGB1320	HGB800 ΔlrhA2	This study
HGB1059	HGB800 lrp-2::Km	10
HGB1061	HGB800 secondary form	10
HGB1102	ΔnilR16::Sm	11
HGB1137	ΔnilR16::Sm (Tn7-nilQR)	11
HGB1230	ΔcpxR1::Sm	23
HGB151	HGB007 ΔrpoS1::Km	56
HGB597	HGB081 flhD4::Tn10; Km	42
Plasmids
pBluescript II SK+	Ampr; general cloning vector	Stratagene (La Jolla, CA)
pBluescript II KS+	Ampr; general cloning vector	Stratagene (La Jolla, CA)
pKNJ102	Source of aad (Strr) cassette (see Table 2 for primers)	12
pEV2	pBluescript KS+ with Km cassette (1.7 kb) in BamHI site	E. I. Vivas
pBlueStr	pBluescript SK+ with Sm cassette from pKNJ102 (1.17 kb) in BamHI site	This study
pBlueEstUp	pBluescript SK+ with EstUp (1,256-bp insert)	This study
pBlueEstUpDn	pBlueEstUp with EstDn (1,161-bp insert)	This study
pBlueEstUpKmDn	pBlueEstUpDn with Km from pEV2	This study
pBlueXlpUpDn	pBluescript SK with XlpUp (1,096-bp insert) and XlpDn (1,086 bp)	42
pBlueXlpUpSmDn	pBlueXlpUpDn with Sm from pBlueStr	This study
pKR100	Cmr; oriR6K suicide vector	K. Visick, Loyola University
pKREstKm	pKR100 with EstUpKmDn (3.9-kb insert)	This study
pKRXlpSm	pKR100 with XlpUpKmDn (5.2-kb insert)	This study
pKNG101	Smr; oriR6K suicide vector	27
pKNGLrhAUpDn	pKNG101 with LrhAUp (1,212-bp insert) and LrhADn (1,142 bp)	This study

Molecular biological methods.

This research was performed using standard molecular biological methods (45). DNA was PCR amplified using either ExTaq (Takara, Otsu, Shiga, Japan) or Platinum Pfx (Invitrogen, Carlsbad, CA) according to the manufacturers' instructions. To verify correct sequence, inserts of all constructs were sequenced at the University of Wisconsin Biotechnology Center by the use of BigDye version 3.1 (Applied Biosystems, Foster City, CA). PCR purification, plasmid preparation, and gel extraction kits (Qiagen, Valencia, CA) were used according to the manufacturers' directions, as were restriction enzymes (Promega, Madison, WI). The primers used in this work (Integrated DNA Technologies, Coralville, IA; University of Wisconsin Biotechnology Center, Madison, WI) are presented in Table 2.

TABLE 2.

Primers used in this study

Primer	Sequence (5′ to 3′)a	Use
EstAApaUpF	NNNNGGGCCCCCGATTGTTGTTTCTCCAGAC	Mutant construction
EstABamUpR	NNNNGGATCCGGCACTCTCCTTAAAAGTCGT	Mutant construction
EstABamDnF	NNNNGGATCCGCAGCGCTCAAGATCCTAACT	Mutant construction
EstAXbaDnR	NNNNTCTAGACCAGAATTTCGTCCGGTTCGA	Mutant construction
LrhAApaUpF	NNNNGGGCCCCCAGCATTTCGTCACCTGTAT	Mutant construction
LrhAKpnUpR	NNNNGGTACCGCAGATCGAGATCGAGGTTTA	Mutant construction
LrhAKpnDnF	NNNNGGTACCCCTCTGCCGAAAATATAGAC	Mutant construction
LrhABamDnR	NNNNGGATCCCGCTAGGTTTACTGACTTGA	Mutant construction
StrepBamF	NNNNGGATCCCCAGGACAGAAATGCCTCGAC	pBlueStr construction
StrepBamR	NNNNGGATCCGCGTCGGCTTGAACGAATTGT	pBlueStr construction
RecAminFor	TGTCCGTTTGGATATCCGCC	qPCR
RecAminRev	CCCAGAGTATTAATACCTTCCCCAT	qPCR
EstAintF	GCCACCACAGGTAAACAGTG	qPCR
EstAintF	CCCAAAGTGCATAGAGGAGCT	qPCR

^aEngineered restriction enzyme sites are underlined; N represents G, A, T, or C.

Construction of the estA1 and estA1 xlpA3 mutants.

The estA and xlpA genes were found during a BLAST (2) search of the X. nematophila genome (https://www.genoscope.cns.fr/agc/mage/) for homologs of genes encoding other bacterial extracellular lipases (see Table S1 in the supplemental material). For estA mutant construction, primers with engineered restriction sites (Table 2) were used to PCR amplify, from HGB800 chromosomal DNA, fragments located upstream (primers EstAApaUpF and EstABamUpR; 1,256-nucleotide [nt] product) and downstream (primers EstABamDnF and EstAXbaDnR; 1,161-nt product) of the 1,953-nt region to be deleted (including the entire coding region, except the last 56 nt). PCR amplifications were conducted using Platinum Pfx (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions, using annealing temperatures of 48°C and 52°C, respectively. These fragments were subsequently cloned, using the engineered restriction sites (Table 2), into pBluescript II SK+, and the Km^r cassette (BamHI digested from pEV2) was cloned into the unique engineered BamHI site between them. The ΔestA1::Km construct was then cloned, using the KpnI and XbaI restriction sites, into the suicide vector pKR100 to create pKREstKm. This construct was then conjugated from E. coli S17-1 (λpir) into HGB800 to create HGB1301, and allelic replacement in Km^r Cm^s exconjugants was verified by PCR amplification. For estA xlpA mutant construction, the Sm^r cassette (BamHI digested from pBlueStr) was cloned into the unique BamHI site of pBlueXlpUpDn (42), and the resulting ΔxlpA3::Sm construct was then cloned, using the KpnI and XbaI restriction sites, into the suicide vector pKR100 to create pKRXlpSm. To create the 1,052-nt deletion of xlpA, this construct was then conjugated from E. coli S17-1 (λpir) into HGB1301 (estA1::Km) to create HGB1302, and allelic replacement in Km^r Sm^r Cm^s exconjugants was verified by PCR amplification.

Construction of markerless lrhA deletion mutant.

The markerless ΔlrhA 922-nt deletion (of all but the first 45 nt and last 25 nt of the lrhA coding region) mutant was constructed using the method of Kaniga et al. (27). Primers with engineered restriction sites (Table 2) were used to PCR amplify, from HGB800 chromosomal DNA, fragments located upstream (primers LrhAApaUpF and LrhAKpnUpR; 1,212-nt product) and downstream (primers LrhAKpnDnF and LrhABamDnR; 1,142-nt product) of the region to be deleted, using Platinum Pfx (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions and using annealing temperatures of 49°C and 45°C, respectively. These fragments were subsequently digested with KpnI and ligated together, and the resulting fragment was PCR amplified using the LrhAApaUpF and LrhABamDnR primers. The product was cloned into the suicide vector pKNG101 by using the ApaI and BamHI restriction sites to create pKNGLrhAUpDn. This construct was then conjugated from E. coli SM-10 (λpir) into HGB800. HGB800 Sm^r exconjugants sensitive to 5% sucrose were then grown on LB agar overnight and subsequently grown on LB agar plus sucrose to select for sucrose-resistant exconjugants that had excised the vector. The Sm^s phenotype was verified, and deletion of the lrhA fragment was confirmed by PCR amplification.

In vitro phenotypic assays.

In vitro phenotypic plate analyses were performed as described previously to assay motility (56), lipase activity (50), lecithinase activity (6, 53), protease activity (6), and hemolytic activity (24, 44) on agar containing 5% defibrinated sheep blood (Colorado Serum Company, Denver, CO) and antibiotic activity (32, 56) against Bacillus subtilis. In vitro phenotypic experiments were conducted a minimum of two times, with a minimum of two replicates per experiment.

Insect virulence assays.

Tobacco hornworm M. sexta eggs (North Carolina State University) were reared to fourth-instar larval stage on an artificial Gypsy moth wheat germ diet (MP Biomedicals, Aurora, OH) as described previously (56). For virulence assays, three independent experimental replicates were performed. Stationary-phase cultures of X. nematophila strains were assessed as explained previously (13, 37). For each experimental replicate, overnight cultures grown at 30°C in LB broth were subcultured at 1:100 into fresh LB and grown for 18 to 24 h. The strains were then washed in PBS, diluted, and plated onto LB agar for calculation of CFU. Each treatment was injected into 10 insect larvae at a level of approximately 10⁴ CFU per insect by the use of a 30-gauge syringe (Hamilton, Reno, NV), and mortality was monitored until at least 72 h postinjection. Logarithmic-phase X. nematophila cultures were assayed with the following modifications, as described previously (42); when subcultured 1:100 in fresh LB broth, cultures were incubated at 30°C until they reached an optical density at 600 nm of 0.8. Cultures were then injected at a level of approximately 10² CFU.

Nematode colonization and progeny production assays.

Nematode cocultivations were executed as previously explained, with three independent experimental replicates (24). Briefly, for each experimental replicate, each X. nematophila strain was grown on three separate lipid agar plates to which sterile S. carpocapsae nematode eggs were added. For colonization assays, IJ nematodes reared on these plates were collected in White traps (29), surface sterilized, and sonicated. The sonicated nematode solution was serially diluted and plated for colonies, and this information was used to calculate the average number of CFU per nematode.

For nematode production assays, S. carpocapsae IJ nematodes were colonized with each X. nematophila strain as described above and subjected to assays similar to those of Mitani et al. (34), with the specified modifications. Uncolonized nematodes were reared on X. nematophila HGB151 rpoS1::Km mutants as previously described (56). Nematodes colonized by each strain were injected at a level of approximately 50 IJs into each of 12 G. mellonella larvae (Grubco, Hamilton, OH) using a 30-gauge syringe (Hamilton, Reno, NV) and subsequently placed in White traps. After injection, traps were monitored daily for IJ emergence, and the first 10 larvae per treatment to show IJ emergence were used to calculate average time postinjection to first emergence. To determine cumulative numbers of IJ progeny after the first day of emergence, IJs were collected from each trap and counted daily for the first 10 days, every 2 days from day 12 to day 20, and finally on days 21 and 28. Colonized nematodes carrying each X. nematophila strain were assayed for colonization preinjection and postemergence to verify the identities of strains and confirm wild-type levels of colonization. Any G. mellonella replicates showing signs of fungal contamination were excluded from the analysis.

Quantitative PCR detection of transcript levels.

Measurement of transcript levels by using quantitative PCR (qPCR) was performed as described previously (13). By the use of the TRIzol extraction procedure (Invitrogen, Carlsbad, CA), complete cellular RNA was isolated from X. nematophila logarithmic-phase cultures grown in LB to an optical density at 600 nm of 0.9. Residual DNA was removed from the cellular RNA samples with DNase I (Roche Diagnostics, Mannheim, Germany). cDNA was synthesized using random hexamer primers (Integrated DNA Technologies, Coralville, IA) and AMV reverse transcriptase (Promega, Madison, WI). The cDNA samples were next subjected to qPCR in duplicate 25-μl reaction mixtures containing iQ SYBR green supermix (Bio-Rad, Hercules, CA) and the relevant primers that are listed in Table 2. Water was added in lieu of cDNA template as a negative control. qPCR reactions were performed on a Bio-Rad iCycler machine, with the resulting data analyzed using Bio-Rad iCycler iQ software. Transcript levels of recA, detected with primers RecAminFor and RecAminRev, were used to normalize cycle threshold results between strain cDNA samples. The conclusion that recA levels were suitable for normalization in these experiments was based on several facts. First, recA cycle numbers do not vary dramatically between reactions of primary- and secondary-form DNA template (data not shown). Second, an lrp mutant that displays secondary-form characteristics (11) has similar levels of recA transcript according to whole-genome microarray data (X. Lu and H. Goodrich-Blair, unpublished data). Finally, our previously published data (11) show that most but not all genes, when normalized with recA, are expressed at lower levels in the secondary form than in the primary form, while the data presented here show that estA is expressed at higher levels in the secondary form. If a change in recA levels were occurring in the secondary form relative to those in the primary form, we would expect this change to have a consistent effect on our measurements of RNA levels regardless of the genes being tested.

Statistical analysis was performed on normalized cycle numbers, and data are presented accounting for the twofold change in the amount of product per cycle.

Statistical analysis.

Transcript level cycle numbers, antibiotic production, nematode colonization, time to IJ progeny emergence, cumulative IJ progeny numbers, virulence percent mortality, and LT₅₀ (time required to kill 50% of injected insects) data were analyzed with either an unpaired t test or one-way analysis of variance with Tukey's posttest at a 95% confidence interval by using GraphPad Prism version 3.0a for Macintosh (GraphPad Software, San Diego, CA).

RESULTS

Identification of estA, a gene predicted to encode the lecithinase activity of X. nematophila.

During laboratory growth, X. nematophila expresses two lipase activities with in vitro specificities against Tween and lecithin (6, 7, 53). It has been established that the gene xlpA (Fig. 1B) is necessary for the Tween-specific activity (39, 42). However, to date, the gene encoding X. nematophila lecithinase has not been reported. Using BLAST (2), we searched the X. nematophila genome (https://www.genoscope.cns.fr/agc/mage/) for additional genes predicted to encode homologs of other bacterial extracellular lipases implicated in virulence (see Table S1 in the supplemental material). This search revealed an open reading frame (XNC1_2104) that we named estA after a homolog encoded by the plant pathogen Serratia liquefaciens (Fig. 1A). In S. liquefaciens, EstA is a lipase required for the utilization of certain lipids as sole carbon sources (43). X. nematophila estA is predicted to encode a 668-amino-acid protein that is 38% identical and 56% similar to the S. liquefaciens homolog and also has similarity to the lip-1 gene of the related entomopathogenic bacterium Photorhabdus luminescens (50% identity, 66% similarity) (2). The X. nematophila estA genetic locus is syntenic with that of P. luminescens lip-1 but not with that of S. liquefaciens estA, which encodes the swr quorum-sensing system downstream. In both X. nematophila and P. luminescens, the estA/lip-1 homologs are flanked upstream by the tandemly oriented prsA (predicted to encode a phosphoribosylpyrophosphate synthetase) and downstream by the divergently oriented ychH (predicted to encode a conserved, hypothetical protein) (Fig. 1A). The X. nematophila EstA protein has a putative N-terminal phospholipase/lecithinase/hemolysin domain (amino acids 1 to 407) and a C-terminal autotransporter domain (amino acids 426 to 668) predicted to contain integral membrane β barrels involved in autosecretion (Fig. 1A) (2). In addition, EstA is predicted to contain the serine-aspartate-histidine catalytic triad (S34, D350, H353), which is characteristic of lipases (3, 25, 26, 36).

FIG. 1.

The estA (A) and xlpA (B) loci of X. nematophila. Arrows indicate genes and their directions of transcription. The regions of estA predicted to encode a phospholipase/lecithinase/hemolysin domain and an autotransporter domain are designated by white and gray shadings, respectively. Genes were named based on their similarity to those of E. coli, with the exceptions of estA and xlpA, which were named based on their homologs in S. liquefaciens and Yersinia enterocolitica, respectively.

X. nematophila mutants deficient in distinct lipase activities.

To determine if X. nematophila estA encodes the in vitro lecithinase activity, we created an estA deletion mutant, estA1::Km, by replacing the gene with a kanamycin-resistant cassette. Unlike the wild-type parent, the estA1::Km mutant did not have in vitro lecithinase activity against egg yolk (6, 53) but retained activity against Tweens 20, 40, and 60 (6, 50), indicating that estA likely encodes the lecithinase activity of X. nematophila (Table 3). xlpA lipase mutants exhibit defective in vitro lipase activity against Tweens 20, 40, and 60, as previously reported (39, 42), but have a wild-type level of lecithinase activity (Table 3). As expected, a double mutant lacking both estA and xlpA (estA1::Km xlpA3::Sm) lacked activity against Tween or lecithin (Table 3). Thus, the two X. nematophila in vitro lipase activities have now been attributed to distinct genetic loci.

TABLE 3.

Selected phenotypes of X. nematophila lipase mutants

Strain	Activity ofa:						LT50g
	Lipaseb	Lecithinasec	Proteased	Hemolysind	Motilitye	Antibiotic productionf	Logh	Stati
Wild type	+	+	+	+	+	+	22.4 ± 0.7	23.4 ± 2.7
estA1::Km mutant	+	−	+	+	+	+	24.2 ± 0.9	24.8 ± 0.9
xlpA2::Km mutant	−	+	+	+	+	+	25.5 ± 1.3	25.6 ± 0.3
estA1::Km xlpA3::Sm mutant	−	−	+	+	+	+	24.7 ± 2.1	26.7 ± 1.7

^a+, activity was indistinguishable from wild type; −, activity was not detected.

^bQualitative evaluation of halo surrounding the bacterial colony 3 days after inoculation on plates containing Tween 20, 40, or 60.

^cQualitative evaluation of halo surrounding the bacterial colony 4 days after inoculation on plates containing egg yolk.

^dQualitative evaluation of halo surrounding the bacterial colony 3 days after inoculation on milk or blood agar plates.

^eZone of growth 24 h after inoculation on 0.25% agar plates.

^fSize of halo within lawn of indicator strain (Bacillus subtilis) surrounding the Xenorhabdus colony after 24 h of incubation.

^gData are mean times (h) to 50% mortality ± standard error for M. sexta injected with the respective mutants (n = 3).

^hLogarithmic-phase cultures were injected.

ⁱStationary-phase cultures were injected.

To verify that the deletion of the estA and xlpA genes did not result in additional unexpected defects, estA1::Km, xlpA2::Km (42), and the estA1::Km xlpA3::Sm double mutant each were tested for several in vitro phenotypes, including hemolysin and protease activity, motility, and antibiotic production. Each of the three lipase mutants displayed wild-type phenotypes for each of these activities, as expected (Table 3).

X. nematophila lipase mutants exhibit wild-type nematode colonization and virulence against M. sexta insects.

To examine the potential roles of lipases in the host interactions of X. nematophila, estA1::Km, xlpA2::Km, and estA1::Km xlpA3::Sm mutants were tested for virulence in M. sexta insects and colonization of S. carpocapsae nematodes (Fig. 2) (42). Previously we reported that the X. nematophila xlpA2::Km mutant is as virulent as its wild-type parent when grown to logarithmic phase and injected into M. sexta larvae (42). Similarly, logarithmic-phase cells of estA1::Km and estA1::Km xlpA3::Sm mutants exhibited wild-type virulence, killing 90 to 100% of insects (Fig. 2A). Also, stationary-phase cells of each lipase mutant (the estA mutant, the xlpA mutant, and the double mutant) killed as many insects as did the wild-type control (stationary-phase cells typically kill fewer insects at higher doses than do logarithmic-phase cultures) (Fig. 2B). None of the mutants were delayed in the time it took them to kill M. sexta, as evidenced by their wild-type LT₅₀s (Table 3). Thus, the estA and xlpA lipase genes do not play a direct role in the ability of X. nematophila to cause disease in M. sexta insects. In addition, in vitro cultivation assays showed that each of the lipase mutants colonized the IJ stage of S. carpocapsae at levels no different from those of their wild-type parent (Fig. 2C), demonstrating that the XlpA and EstA lipases are not required by X. nematophila for colonization of the nematode host.

FIG. 2.

Host interactions of X. nematophila lipase mutants. Ability of logarithmic-phase (A) or stationary-phase (B) X. nematophila cultures to kill M. sexta insects (note that virulence of logarithmic-phase xlpA2::Km cultures are from a study by Richards et al. [42] and are included for comparison). Cultures were injected, and percent mortality at 72 h is shown. (C) X. nematophila colonization of S. carpocapsae nematodes cultivated on lawns of each strain. The wild type (black bars; wt) and estA1::Km (white bars), xlpA2::Km (cross-hatched bars), and estA1::Km xlpA3::Sm (diagonal lines) mutants are shown. Separate wild-type results are presented because independent experiments were performed for each mutant. Error bars represent standard errors (n = 3). No significant differences were found (P > 0.05).

Colonization of S. carpocapsae by xlpA lipase mutants results in delayed and decreased nematode progeny production during G. mellonella insect infection.

X. nematophila supports the development of S. carpocapsae nematodes within insects (22, 28, 34, 48), and Steinernema nematodes require host-derived lipids (1) for reproduction (14). However, the role of Xenorhabdus lipase activities in providing lipid compounds to their nematode hosts has not been directly tested. If X. nematophila lipase mutants are unable to provide specific lipid derivatives as a nutrient source to S. carpocapsae in insecta, it could be detrimental to nematode reproduction and lead to a delay in progeny IJ emergence and/or a decrease in the total number of progeny produced. This idea was tested by monitoring the ability of the lipase mutants to support nematode productivity in G. mellonella insects. In addition, we tested an X. nematophila lrhA mutant lacking the LysR-type regulator LrhA, since an lrhA mutant has a defect in Tween lipase activity and decreased levels of xlpA transcript (42). Nematodes colonized by wild-type X. nematophila or the estA1::Km, xlpA2::Km, estA1::Km xlpA3::Sm, or lrhA2 mutant were injected into G. mellonella larvae. Uncolonized nematodes, which do not produce IJs in this insect host (22, 34), were injected as a negative control, and the time postinjection to first progeny IJ emergence, as well as the total number of progeny IJs produced over time, was recorded. All G. mellonella larvae died within a few days of injection. This was expected, as S. carpocapsae is known to kill G. mellonella in the absence of X. nematophila (22, 34).

IJ progeny from nematodes colonized by wild-type X. nematophila took an average of 10.9 days postinjection to begin to emerge, and the times to emergence for estA1::Km (9.7 days) and lrhA2 (10.2 days) mutants were not significantly different from that of the wild type (P > 0.05) (Fig. 3A). However, both the xlpA2::Km mutant and the estA1::Km xlpA3::Sm double mutant showed a significant increase in the time to emergence (15.9 and 13.9 days, respectively; P < 0.05). No progeny IJs emerged from G. mellonella injected with uncolonized IJs, as expected. These data indicate that xlpA, but not estA, plays a role in the timing of nematode emergence from G. mellonella cadavers.

FIG. 3.

Progeny production of nematodes colonized by X. nematophila lipase mutants. (A) IJ nematodes colonized by X. nematophila wild type (black bars; wt) or lrhA2 (vertical lines) estA1::Km (white bars), xlpA2::Km (cross-hatched bars), or estA1::Km xlpA3::Sm (diagonal lines) mutants were injected into G. mellonella larvae, and the time to first emergence of progeny IJs was recorded. (B) Larvae were then monitored for total progeny IJ emergence at the indicated days. Nematodes with wild-type (black squares, solid lines) or lrhA2 (white circles), estA1::Km (white diamonds), xlpA2::Km (white squares), or estA1::Km xlpA3::Str (white triangles) mutant treatment are shown. Different letters indicate significantly different values (P < 0.05; n = 10). In the case of panel B, significant differences refer to cumulative numbers at day 28.

The xlpA2::Km and estA1::Km xlpA3::Sm mutants also produced significantly fewer progeny IJs than did the wild type over time. On day 5 of progeny IJ emergence, xlpA2::Km and estA1::Km xlpA3::Sm mutants produced, on average, fewer than half the number of IJs (61,808 and 60,916, respectively) as wild-type X. nematophila (145,186) (Fig. 3B). The average number of total progeny IJs emerging from xlpA2::Km and estA1::Km xlpA3::Sm mutant-infected insects remained significantly lower than that from the wild type throughout the experiment (P < 0.05) (Fig. 3B). One exception was that on day 14, the number of progeny from xlpA2::Km-infected insects, though fewer than those from wild-type-infected insects, was not significantly different (P > 0.05) (Fig. 3B). By 28 days, insects infected with the xlpA2::Km and estA1::Km xlpA3::Sm mutants produced, respectively, 74% (174,337) and 68% (161,124) the total number of progeny IJs produced by wild-type-infected insects (236,404) (Fig. 3B).

The number of cumulative progeny from insects infected with lrhA2 mutant-colonized nematodes was fewer than that from the wild-type treatment throughout the experiment, although this difference was only statistically significant on days 21 and 28 (P < 0.05) (Fig. 3B). In contrast, insects infected with estA1::Km mutant-colonized nematodes generated levels of progeny IJs similar to those for the wild type throughout the experiment. From these results, we conclude that xlpA, but not estA, aids in production of nematode progeny within the insect host G. mellonella. As expected, the number of progeny IJs from uncolonized nematodes was below the level of detection (data not shown), and the G. mellonella larvae injected with these nematodes developed fungal contamination, consistent with the observations of Mitani et al. (34).

X. nematophila secondary form variant has increased in vitro lecithinase activity that correlates with an increase in estA expression.

In addition to regulation by LrhA, it is known that xlpA expression is positively regulated by the global regulator Lrp (leucine-responsive regulator protein) and the master flagellar regulator FlhDC through the action of the flagellar sigma factor FliA (11, 39, 42). In addition, secondary-form X. nematophila cells exhibit more Tween lipase activity in vitro than do primary-form cells, even though xlpA is expressed at lower levels in the secondary form (11). In contrast, nothing is known about the regulation of the estA-encoded lecithinase, although it has been observed that the secondary form exhibits more in vitro lecithinase activity than does the primary form (7). Because the regulation of estA may provide insights into the biological function of lecithinase activity, we examined the in vitro lecithinase activities of the primary and secondary forms of wild-type X. nematophila on egg yolk agar. The secondary form displayed higher levels of lecithinase activity than did the primary form (data not shown), as previously observed (7). To determine if this difference in activity between the two forms resulted from a difference in estA transcription, we monitored transcript levels of estA in the primary and secondary forms of X. nematophila. The secondary form had over four times the amount of estA transcript as the primary (Fig. 4A), thereby demonstrating that the increase in in vitro lecithinase activity for the secondary form correlates with an increase in the expression of estA. The estA1::Km mutant, which served as a negative control, had no detectable expression of estA transcript, as expected.

FIG. 4.

Transcript levels of the X. nematophila estA gene. Total cellular RNA was extracted from logarithmic-phase X. nematophila cultures of primary-form wild type (black bar), secondary-form wild type (white bar), and the estA1::Km mutant (vertical lines, not visible) (A) or wild-type (black bar), nilR16::Sm (cross-hatched bar), and nilR16::Sm carrying a wild-type copy of nilR (diagonal lines) (B). cDNA was analyzed by qPCR. Levels of transcript are reported as percentages of primary-form wild type. Bars with different letters are significantly different from each other (P < 0.05; n ≥ 3).

NilR, a regulator of X. nematophila colonization factors, mediates repression of estA expression.

To further examine the regulation of estA expression, we tested a variety of X. nematophila transcriptional regulator mutants for differences in in vitro lecithinase activity compared to that of wild-type primary form. These included strains with mutations in lrhA (42), lrp (11), flhD (21, 42), cpxR (encodes the response regulator of a two-component system involved in both mutualism and pathogenesis of X. nematophila) (23), rpoS (encodes the stationary-phase sigma factor σ^S, which is required for nematode colonization) (56), and nilR (encodes a repressor of factors required for nematode colonization) (10). Each of these mutants had a level of lecithinase activity qualitatively similar to that of the wild-type primary form, with the exception of the nilR16::Sm mutant, which, like the wild-type secondary form, had a noticeable increase in lecithinase activity compared to that of the wild-type primary form. qPCR revealed that this higher in vitro activity corresponded with an estA transcript level approximately twice that of primary-form wild-type X. nematophila (Fig. 4B). Introduction of a wild-type copy of nilR into the nilR16::Sm strain (10) restored estA transcript levels to those of the primary-form wild type, confirming that nilR was responsible for this difference in transcription (Fig. 4B).

Although the in vitro lecithinase activity of the nilR16::Sm mutant was qualitatively indistinguishable from that of the secondary-form wild type, it appears that estA expression may be higher in the wild-type secondary form than in the primary-form nilR16::Sm mutant (Fig. 4). Thus, in an independent experiment, we directly compared estA levels between these two strains. The secondary-form wild type had higher estA transcript levels than the nilR16::Sm mutant (estA level in the secondary-form wild type was 276% that of the primary-form wild type, versus 174% for the nilR16::Sm mutant; n = 3), although this difference was not statistically significant. Nevertheless, the apparent difference in estA expression between the nilR16::Sm mutant and the secondary-form wild type implies that one or more additional regulators may contribute to estA expression in the secondary form.

DISCUSSION

estA is necessary for X. nematophila lecithinase activity.

The gene xlpA, encoding a Tween-specific lipase activity of X. nematophila, was characterized previously (39, 42), but the gene encoding a lecithinase-specific lipase activity remained unknown until this study. Here, we show that deletion of the estA gene, predicted to encode a lecithinase/phospholipase C (Fig. 1A), abolishes the in vitro lecithinase activity against egg yolk. The estA mutant did not display any differences from the wild type in other enzymatic activities, including protease, hemolysin, and lipase activities against Tween (Table 3). These data support the conclusion that X. nematophila estA encodes a lecithinase and, coupled with the previous characterization of the xlpA-encoded Tween lipase, establish genes responsible for two phenotypically distinct lipolytic activities produced by X. nematophila.

xlpA, but not estA, contributes to nematode productivity during infection of G. mellonella.

The lipase and lecithinase activities of X. nematophila are required for neither colonization of S. carpocapsae nematodes nor virulence against M. sexta insects (39, 42) (Fig. 2). The data presented here demonstrate a role for xlpA in supporting S. carpocapsae progeny IJ production during infection of G. mellonella insects (Fig. 3). Nematodes colonized by the X. nematophila xlpA2::Km or estA1::Km xlpA3::Sm mutants exhibited delayed emergence of progeny IJs (Fig. 3A) and produced significantly fewer total progeny IJs over time than did the nematodes colonized with wild-type X. nematophila (Fig. 3B). Based on these data, we conclude that xlpA contributes to, but is not essential for, the efficiency of nematode production. We predict that the contribution of xlpA to nematode production likely would impact the fitness of S. carpocapsae nematodes in nature, where scavengers and competitors threaten the persistence of the infected insect cadaver (for example, see reference 60). Furthermore, it is possible that the IJ progeny derived from insects infected with the xlpA mutants are less fit than those cultivated from wild-type-infected insects. Future analysis of other parameters important to the success of S. carpocapsae in nature, including size, longevity, infectivity, and lipid content, could reveal additional biological effects of xlpA mutants on the nematode life cycle.

Although not as severe at early time points, the X. nematophila lrhA2 mutant had an overall defect, similar to that of the xlpA mutants, supporting the production of fewer progeny IJs than the wild type (Fig. 3B). Like the lrhA2 mutant, an lrp mutant has been shown to exhibit a defect in production of nematode progeny compared to the wild type, although the lrp defect was measured during in vitro cultivations of nematodes (11). This is consistent with Lrp as a positive regulator of lrhA and LrhA, in turn, positively influencing the expression and secretion of XlpA (11, 42). However, transcription of xlpA is not abolished in an lrhA mutant (42), and it is likely that some XlpA production in both lrhA and lrp mutants accounts for these mutants' ability to support the production of higher numbers of IJ progeny than the xlpA mutant, either early on (lrhA) or overall (lrp).

In contrast to the xlpA2::Km lipase mutant, the estA1::Km lecithinase mutant did not have a defect in supporting nematode productivity in G. mellonella larvae (Fig. 3), nor did the estA1::Km mutant have defects in nematode colonization or virulence against M. sexta insects, leaving the biological function of EstA lecithinase activity unclear. It is possible that estA does contribute to nematode productivity but that its influence is too subtle to be detected in our assay. Alternatively, the G. mellonella and M. sexta insects used in these studies may not have substantial levels of a lipid substrate recognized by X. nematophila lecithinase (5, 53). Although detailed biochemical characterization has not yet been conducted, estA is predicted to encode a phospholipase C activity (cleavage of the phosphoryl group from a phospholipid), while xlpA is predicted to encode a phospholipase A (cleaving carboxy-ester bonds) with a potentially broader substrate range (3, 25, 26). Since insect lipid content, including the relative concentrations and types of phospholipids, can vary depending on age, diet, and species (5), it is plausible that the role of X. nematophila estA would be more apparent in assays conducted in wild insects or in another insect species.

The regulation of X. nematophila estA expression may provide insights into its function.

Transcript levels of estA are significantly higher in the secondary-form phenotypic variant of wild-type X. nematophila than they are in the primary form (Fig. 4A), which is consistent with the qualitatively higher level of in vitro lecithinase activity in secondary-form cells. While phenotypic variation of several activities is known to occur at the level of transcription (11), to our knowledge, estA is the first gene demonstrated to have elevated transcript levels in the secondary form. (Although secondary form cells also have increased in vitro lipase activity against Tween relative to primary form cells, xlpA transcript levels unexpectedly were lower in the secondary form than in the primary form [11].)

The lower level of estA transcript in primary form relative to that of secondary-form cells appears to be due to negative regulation mediated through a putative transcription factor, NilR, since the primary-form nilR mutant expresses estA at levels similar to those of wild-type secondary-form cells (Fig. 4B). NilR was discovered based on its role in repression of the nematode colonization factors nilA, -B, and -C (10). Although it is not yet known whether NilR regulates estA directly or if other regulators act on estA, the common regulation of estA with genes necessary for nematode colonization initiation may suggest that the lecithinase activity is important at late stages of nematode development, when population levels are high, nutrients are limiting, and IJs are forming. This is consistent with the fact that estA is not required for virulence during early stages of insect infection (Fig. 2). Although estA is not required for nematode productivity (Fig. 3), perhaps it functions to supply alternative or additional nutrients for the bacteria or nematodes in a nutrient-depleted environment. Indeed, S. liquefaciens estA mutants are defective in in vitro growth on lipids as a sole carbon source (43), and the role of phospholipases in phosphate acquisition has been posited for some bacteria (38, 52, 54). In vivo expression studies of nilR, estA, and nilA, -B, and -C during the course of insect infection could help shed light on when NilR and lecithinase activity are important.

The significant role of Xenorhabdus bacteria in supporting Steinernema nematode reproduction has long been recognized (40) and routinely has been presumed to be due to bacterial enzymatic activities. The identification of genetic elements responsible for X. nematophila lipase activity and the demonstration of a role for one of these, xlpA, in nematode productivity represent a first step toward understanding the molecular basis of this aspect of the symbiosis between X. nematophila and S. carpocapsae.