Involvement of Vitamin B₆ Biosynthesis Pathways in the Insecticidal Activity of Photorhabdus luminescens

ABSTRACT

Photorhabdus luminescens is a Gram-negative entomopathogenic bacterium which symbiotically associates with the entomopathogenic nematode Heterorhabditis bacteriophora. P. luminescens is highly virulent to many insects and nonsymbiotic nematodes, including Caenorhabditis elegans. To understand the virulence mechanisms of P. luminescens, we obtained virulence-deficient and -attenuated mutants against C. elegans through a transposon-mutagenized library. From the genetic screening, we identified the pdxB gene, encoding erythronate-4-phosphate dehydrogenase, as required for de novo vitamin B₆ biosynthesis. Mutation in pdxB caused growth deficiency of P. luminescens in nutrient-poor medium, which was restored under nutrient-rich conditions or by supplementation with pyridoxal 5′-phosphate (PLP), an active form of vitamin B₆. Supplementation with three other B₆ vitamers (pyridoxal, pyridoxine, and pyridoxamine) also restored the growth of the pdxB mutant, suggesting the existence of a salvage pathway for vitamin B₆ biosynthesis in P. luminescens. Moreover, supplementation with PLP restored the virulence-deficient phenotype against C. elegans. Combining these results with the fact that pdxB mutation also caused attenuation of insecticidal activity, we concluded that the production of appropriate amounts of vitamin B₆ is critical for P. luminescens pathogenicity.

IMPORTANCE The Gram-negative entomopathogenic bacterium Photorhabdus luminescens symbiotically associates with the entomopathogenic nematode Heterorhabditis bacteriophora. P. luminescens is highly virulent to many insects and nonsymbiotic nematodes, including Caenorhabditis elegans. We have obtained several virulence-deficient and -attenuated P. luminescens mutants against C. elegans through genetic screening. From the genetic analysis, we present the vitamin B₆ biosynthetic pathways in P. luminescens that are important for its insecticidal activity. Mutation in pdxB, encoding erythronate-4-phosphate dehydrogenase and required for the de novo vitamin B₆ biosynthesis pathway, caused virulence deficiency against C. elegans and growth deficiency of P. luminescens in nutrient-poor medium. Because such phenotypes were restored under nutrient-rich conditions or by supplementation with B₆ vitamers, we showed the presence of the two vitamin B₆ synthetic pathways (de novo and salvage) in P. luminescens and also showed that the ability to produce an appropriate amount of vitamin B₆ is critical for P. luminescens pathogenicity.

INTRODUCTION

The Gram-negative bacterium Photorhabdus luminescens (Enterobacteriaceae), symbiotically associated with the entomopathogenic nematode (EPN) Heterorhabditis bacteriophora, is highly pathogenic toward many insects (1). P. luminescens is carried by the development-arrested stage of H. bacteriophora, called infective juvenile (IJ), in the intestinal lumen. IJs enter the insect host from natural openings or penetrate through the exoskeleton and release the symbiotic bacteria into the hemocoel. Released bacteria produce several toxins to kill insects while suppressing the host immune response (2–5). After the insect dies, recovered IJs from developmental arrest begin to propagate with feeding on symbiotic bacteria within the insect cadaver. During EPN propagation, the insect cadaver is protected from other microorganisms and scavengers with metabolites produced by the symbiotic bacteria (6–8). When the insect cadaver is filled with propagating nematodes, IJs reassociate with the symbiotic bacteria and escape from the cadaver.

Because H. bacteriophora is resistant to the pathogenicity of P. luminescens, it can develop and proliferate by feeding on the symbiotic bacteria. However, other nematodes, including the genetic model organism Caenorhabditis elegans, are susceptible to the toxicity of P. luminescens (9, 10). For example, P. luminescens inhibits the growth of C. elegans larvae and kills >90% of adult C. elegans within 5 days (10). C. elegans molecular genetics is a powerful tool for analysis of bacterial pathogenicity mechanisms, for example, broadly conserved innate immunity (10–12), toxin mode of action (13–15), and screening of new virulence-related genes (16, 17).

To understand the virulence mechanisms of P. luminescens, we constructed a transposon-mutagenized library and screened virulence-deficient or -attenuated mutants using the model nematode C. elegans. We isolated several mutants and identified one of the genes encoding erythronate-4-phosphate dehydrogenase (pdxB) that functions in the de novo vitamin B₆ biosynthesis pathway. Here, we reveal that vitamin B₆ biosynthesis is involved in the virulence of P. luminescens and that the pdxB gene plays an important role in insecticidal activity.

MATERIALS AND METHODS

Nematodes, bacterial strains, and culture conditions.

Culturing and handling of C. elegans were performed as described previously (18). The wild-type Bristol (N2) strain was provided by the Caenorhabditis Genetics Center (University of Minnesota). P. luminescens TT01 was isolated from Galleria mellonella infected with H. bacteriophora TT01 (kindly provided by Ann Burnell, National University of Ireland-Maynooth) as described previously (10). In this study, we used the spontaneous rifampin-resistant mutant of P. luminescens TT01 as a wild-type strain. P. luminescens was grown in Luria-Bertani (LB) broth, nematode growth medium (NGM), and Vogel-Bonner minimal salts (medium E) (19) plus 0.4% (wt/vol) glucose (VBG) medium at 30°C with 200 rpm shaking, unless otherwise indicated.

Construction of a P. luminescens mutant library.

The Tn5 or Tn10 transposon-inserted P. luminescens mutants were obtained by conjugation with Escherichia coli S17-1 (λpir) carrying pUT mini-Tn5 Km (Biomedal, Seville, Spain) as a donor of Tn5 transposon or pBSL180 (provided by National BioResource Project-E. coli, National Institute of Genetics [NIG], Japan) as a donor of the Tn10 transposon. Overnight cultures of E. coli and P. luminescens were diluted with LB broth (100 μg/ml ampicillin and 25 μg/ml kanamycin were added only for E. coli) to obtain an optical density at 600 nm (OD₆₀₀) of 0.05 and incubated until mid-exponential phase (OD₆₀₀ of 0.6 to 0.7) (measured by BioSpec mini system; Shimadzu, Japan). The cultures of E. coli and P. luminescens were washed with LB broth twice, resuspended in a minimal volume of LB broth, and then mixed and dropped onto the LB agar plate (the mixture ratio was 4:1 [P. luminescens/E. coli]) and incubated overnight (about 15 h) at 30°C. Bacterial cells were collected with LB broth, suspended thoroughly and spread onto LB agar supplemented with 50 μg/ml rifampin and 25 μg/ml kanamycin, and incubated for 2 days at 30°C. Colonies of exconjugants were picked and incubated in LB broth in 96-well microtiter plates with appropriate antibiotics to confirm resistance.

Screening of virulence-deficient mutants.

Transposon-inserted P. luminescens strains were incubated in 200 μl of LB broth supplemented with 50 μg/ml rifampin and 25 μg/ml kanamycin in 96-well microtiter plates for 24 h at 30°C without shaking. Then 50 μl of each culture was spotted onto NGM agar contained in 24-well microtiter plates and incubated for 24 h at 30°C. Synchronized L1-stage C. elegans nematodes were prepared by treating egg-containing adults with sodium hypochlorite (20). Approximately 200 L1-stage nematodes suspended in 10 μl of M9 buffer were transferred into each well and incubated at 20°C. Because C. elegans larvae cannot grow on wild-type P. luminescens, we selected virulence-deficient candidate mutants by the existence of adult nematodes after several days. The strains selected were subjected to a rescreening and confirmed to lack virulence against C. elegans.

Identification of the transposon-inserted gene.

To identify the disrupted gene, the transposon insertion site was determined using an LA PCR in vitro cloning kit (TaKaRa Bio, Shiga, Japan) according to the manufacturer's instructions. Two Tn5 specific primers, Tn5Rev_S1 (5′-GCA CTA AGC ACA TAA TTG CTC ACA GCC AAA CTA TC-3′) and Tn5Rev_S2 (5′-TTA TCG CAA TAG TTG GGC GAA GTA ATC GCA ACA TC-3′), were used for nested PCR in combination with cassette primers C1 and C2 supplied in the kit. The amplified DNA fragment was purified and sequenced using the ABI Prism BigDye Terminator cycle sequencing kit and analyzed on an ABI 3130xl sequencer (ABI, USA). The sequence data were submitted to a BLASTN search (National Center for Biotechnology Information [NCBI], http://www.ncbi.nlm.nih.gov/) to identify the transposon-inserted site.

Plasmid construction for complementation testing.

For complementation testing, we constructed a pdxB expression plasmid and transformed it into the P. luminescens mutant. Primers used in this experiment are listed in Table S1 in the supplemental material. The coding sequence of the pdxB gene with a putative promoter region (Fig. 1A) was first amplified by PCR from genomic DNA of wild-type P. luminescens TT01 using primers pdxB_up_F and usg_R and then reamplified from the amplicon using Pnat_pdxB_Inf_F and pdxB_Inf_R to add cloning adaptors. Linearized pBK-miniTn7-gfp2 (kindly provided by Søren Molin, Danmarks Tekniske Universitet) was made by PCR from pBK-miniTn7-T0 using primers pBK-miniTn7-Pnat_For and pBK-miniTn7-Pnat_Rev. To attach the lambda T0 transcriptional terminator, adaptor-added pdxB gene was ligated into the linearized pBK-miniTn7-T0 with an In-Fusion HD cloning kit (Clontech, CA, USA) to construct pBK-pdxB-T0. The fragment of pdxB with the putative promoter and lambda T0 transcriptional terminator was amplified from pBK-pdxB-T0 using primers pUC19_Pnat_pdxB_InF_F and pUC19_T0_InF_R and ligated into the linearized pUC19 plasmid (linearized by PCR with primers pUC19_linearize_F and pUC19_linearize_R) (Epicentre, WI, USA) by In-Fusion to construct pUC19-pdxB. The plasmid pUC19 (blank, without any insertion) or pUC19-pdxB was introduced into the mutant cells by electroporation using EasyjecT Prima (Equibio Ltd., United Kingdom), following the instruction manual for E. coli transformation. Detailed methods for P. luminescens transformation by electroporation are as follows. Overnight cultures of P. luminescens were diluted with LB broth to obtain an OD₆₀₀ of 0.05 and recultured until the OD₆₀₀ reached 0.6 to 0.7 (mid-exponential growth stage; measured with Synergy H1 [Biotek, VT, USA]). Thirty milliliters of the P. luminescens culture was chilled on ice for 20 min and centrifuged for 20 min at 3,000 rpm at 4°C (this condition was used for all centrifugations in this transformation experiment). After removal of the supernatant, the cells were washed with same volume (30 ml) of chilled Milli-Q quality water and centrifuged. After removal of the supernatant (Milli-Q quality water), the cells were washed three times with half volumes of Milli-Q quality water (15 ml). After the third wash, the cells were resuspended in 4.5 ml of chilled 10% (wt/vol) glycerol and centrifuged. After removal of the supernatant (10% glycerol), the cells were resuspended in 160 μl of 10% glycerol to prepare electrocompetent cells. The competent cells were used for electroporation immediately. One microliter of 20 to 25 ng/μl plasmid was mixed with 40 μl of the competent cells and incubated on ice for 1 min. The mixture was transferred into a chilled cuvette (Gene Pulser/Micropulser with 0.2 cm of gap width; Bio-Rad, CA, USA) and given a pulse at 2.5 kV as the output voltage. The cells were immediately resuspended in 960 μl of SOC medium and incubated for 2 h at 30°C with shaking at 200 rpm. The cells were then spread on LB agar supplemented with 50 μg/ml ampicillin and incubated at 30°C to obtain transformants. The transformants obtained were named pdxB::Tn5/pUC19 and pdxB::Tn5/pUC19-pdxB. The plasmids in the transformants were stable when P. luminescens was cultured in medium supplied with 50 μg/ml ampicillin (data not shown).

FIG 1.

Physical map of Tn5 insertion site and vitamin B₆ biosynthesis pathway. (A) The transposon Tn5 insertion site in the P. luminescens TT01 genome is shown with an arrowhead. The fragment used for complementation testing is shown as a striped bar. The sequence information was obtained from GenBank (accession no. BX571869). (B) Overview of the vitamin B₆ biosynthesis pathways in E. coli. In the de novo pathway, pyridoxal 5′-phosphate (PLP) is synthesized from d-erythrose-4-phosphate (E4P) by serial enzymatic reactions. The dashed arrows indicate steps processed by several reaction. PdxB catalyzes the dehydrogenation of 4-phospho-d-erythronate (4PE) in this pathway. In the salvage pathway, PLP is synthesized from B₆ vitamers pyridoxine (PN), pyridoxal (PL), and pyridoxamine (PM) by phosphorylation and/or oxidation. 4PHT, 4-phosphohydroxy-l-threonine; DXP, deoxyxylose 5′-phosphate; PNP, pyridoxine 5′-phosphate; PMP, pyridoxamine 5′-phosphate.

Pathogenicity of pdxB mutant toward C. elegans.

We cultured P. luminescens strains (wild-type TT01 and mutant pdxB::Tn5) overnight in LB broth. Then 200 μl of each bacterial culture was transferred onto 6-cm NGM plates, spread to cover all the plate surface, and incubated for 24 h at 30°C. Synchronized L1-stage C. elegans nematodes were transferred onto NGM plates seeded with each P. luminescens strain and incubated at 20°C. After 24, 48, and 72 h of incubation, nematodes were collected with M9 buffer, transferred onto a 5% agar pad, and covered with a coverslip. Pictures of nematodes were taken with differential interference contrast (DIC) optics (Eclipse E600; Nikon, Japan) on a microscope equipped with a cooled charge-coupled device (CCD) camera (VTCH1.4ICE; Visualix, Japan). Nematode body lengths were measured with ImageJ (NIH; http://imagej.nih.gov/ij/). We confirmed reproducibility in triplicate experiments.

Complementation test.

We cultured each P. luminescens strain (wild-type TT01 and mutant pdxB::Tn5 and two transformants, pdxB::Tn5/pUC19 and pdxB::Tn5/pUC19-pdxB) overnight in LB broth. The pathogenicity of each P. luminescens strain against C. elegans was determined mainly as described above, except that 50 μg/ml ampicillin was added in NGM when we cultured the two transformants. After 48 h of incubation, nematodes were collected with M9 buffer, transferred onto a 5% agar pad, covered with a coverslip, and measured. The differences in nematode body lengths on each of the four P. luminescens strains were statistically analyzed by Kruskal-Wallis plus Wilcoxon rank sum tests followed by Bonferroni correction. We confirmed reproducibility with triplicate experiments.

Growth test in three different media.

Each P. luminescens strain (wild-type TT01 and mutant pdxB::Tn5 and two transformants, pdxB::Tn5/pUC19 and pdxB::Tn5/pUC19-pdxB) cultured overnight in LB broth was washed and diluted to an OD₆₀₀ of 0.05 with LB broth, NGM, or VBG medium, and then incubated at 30°C with shaking at 200 rpm. The OD₆₀₀ values were measured every hour for 12 h and every 12 h (in LB and NGM) or every 24 h (VBG) until saturation with an absorption spectrophotometer (Synergy H1). Ampicillin at 50 μg/ml was added as a supplement for maintenance of plasmids when the P. luminescens transformants were cultured.

Vitamer auxotrophy test.

For the B₆ vitamer auxotrophy test, we used pyridoxal 5′-phosphate (PLP) (Sigma-Aldrich Co., MO, USA), pyridoxal (PL) (Wako Pure Chemical Industries, Ltd., Osaka, Japan), pyridoxine (PN) (Wako), and pyridoxamine (PM) (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan). P. luminescens wild-type TT01 and mutant pdxB::Tn5 strains were cultured overnight in LB broth, washed and diluted to an OD₆₀₀ of 0.05 with VBG medium supplemented with each B₆ vitamer at concentrations of 0, 1, 10, and 100 μM, and incubated at 30°C with shaking at 200 rpm. The OD₆₀₀ values were measured after 48 h with an absorption spectrophotometer (Synergy H1) and compared by two-tailed t tests between each strain.

Complementation of P. luminescens mutant pathogenicity by PLP supplement.

We cultured wild-type TT01 and mutant pdxB::Tn5 strains overnight in LB broth. Then 200 μl of each bacterial culture was spread onto 6-cm NGM plates supplemented with or without 10 μM PLP to cover the plate surface and incubated for 24 h at 30°C. The body lengths of the nematodes grown on each plate were measured as described above. We evaluated the effects of bacterial strain and PLP supplementation on nematode body lengths by generalized linear model (GLM) analysis using R software (version 3.2.3). Gamma distribution and log link function were employed in the GLM analysis. Body length was the response variable; strain (TT01 or pdxB::Tn5), PLP supplementation (0 or 10 μM), and interaction (strain × PLP supplementation) were the explanatory variables. The significance of the interaction effect was investigated by a likelihood ratio test using chi-squared approximation. We confirmed reproducibility with triplicate experiments.

Test of pathogenicity toward insects.

The pathogenicities of two P. luminescens strains (wild-type TT01 and mutant pdxB::Tn5) toward the insect host were evaluated using the superworm Zophobas morio (purchased from Mito Koorogi, Aichi, Japan). The mid-exponential phase of each P. luminescens strain cultured in LB broth at 30°C was washed with sterilized phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 8 mM Na₂HPO₄, 1.46 mM KH₂PO₄) and resuspended with sterilized PBS at the appropriate cell density (7 × 10⁸ cells/ml). Five microliters of cell suspension (equivalent to 3.5 × 10⁴ cells) was injected into Z. morio larvae (the average weight was 650 ± 50 mg) with an 1702RN Neuros syringe (Hamilton Co., NV, USA). Insects were incubated individually or in three pairs in the plastic case (4.5 cm in diameter by 4.5 cm in height), supplied with wheat bran at 25°C, and checked every 12 h for viability. The survival curves were analyzed by the Kaplan-Meier procedure, and significant differences between the survival curves were calculated by the log rank test with the statistical software Excel Tokei 2006 (SSRI, Tokyo, Japan).

RESULTS

Screening of virulence-deficient mutants using C. elegans as a host.

We screened a total of 700 independent P. luminescens mutants by using C. elegans as a host organism and identified three virulence-deficient or -attenuated mutants (two mutants were from Tn5 and one was from a Tn10 insertion library). We selected one virulence-deficient mutant and identified the location of the Tn5 insertion between nucleotides 81 and 82 of the pdxB gene (Fig. 1A). pdxB was predicted to encode erythronate-4-phosphate dehydrogenase and to function in de novo vitamin B₆ biosynthesis; this gene and pathway are also known in other bacteria, including E. coli (Fig. 1B). We confirmed the mutant pdxB::Tn5 phenotype by comparing body lengths of C. elegans grown on the wild-type TT01 and mutant pdxB::Tn5 from the L1 stage. L1-stage nematodes never grew to the adult stage when transferred onto the wild-type TT01 plate. In contrast, nematodes grew to the adult stage 3 days after being transferred onto the mutant pdxB::Tn5 plate (Fig. 2A; see also Table S2 in the supplemental material).

FIG 2.

Development of C. elegans larvae on P. luminescens strains. (A) Body lengths of C. elegans grown on wild-type TT01 and mutant pdxB::Tn5 (green and red box plots, respectively) from L1 stage at 20°C. Representative data from three independent trials are shown. See Table S2 in the supplemental material for details on body lengths in each trial. (B) Body lengths of C. elegans grown on four different strains, wild-type TT01, mutant pdxB::Tn5, transformed mutant with blank plasmid pdxB::Tn5/pUC19, or transformed mutant with rescue plasmid pdxB::Tn5/pUC19-pdxB, from the L1 stage. Representative data from three independent trials are shown. See Table S3 in the supplemental material for details on body lengths in each trial. Differences among the four strains were compared, and similar results were obtained in triplicate experiments (*, P < 0.01; Wilcoxon rank sum test followed by Bonferroni correction). NS, not significant.

Virulence deficiency is caused by a pdxB gene mutation.

In order to confirm that the phenotype was caused by the pdxB gene mutation, we amplified the coding sequence and putative promoter region of the wild-type pdxB gene (Fig. 1A), fused with the lambda T0 transcriptional terminator and ligated into pUC19 to construct the rescue plasmid pUC19-pdxB. The mutant pdxB::Tn5 was transformed with the plasmid pUC19-pdxB, and its virulence was tested against C. elegans (Fig. 2B; see also Table S3 in the supplemental material). The body lengths of C. elegans grown on the transformed mutant pdxB::Tn5/pUC19-pdxB were as short as those grown on wild-type TT01. The body lengths of C. elegans grown on the transformed mutant with blank plasmid pdxB::Tn5/pUC19 were similar to those grown on the mutant pdxB::Tn5 (Fig. 2B; see also Table S3), suggesting that virulence deficiency was complemented by the expression of the pdxB gene in trans. Therefore, we concluded that the pdxB gene is necessary for P. luminescens pathogenicity against C. elegans.

Phenotype of growth and pigmentation deficiency is complemented by nutrient-rich condition.

Next, we characterized the effect of pdxB mutation on P. luminescens growth in three different liquid media: LB (nutrient-rich medium), NGM (nematode growth medium), and VBG (minimal medium) (Fig. 3). When P. luminescens was cultured in nutrient-rich LB medium, the growth curve of the mutant pdxB::Tn5 was identical to that of the wild-type TT01 (Fig. 3A). No effects of pUC19-pdxB and pUC19 transformation into the mutant pdxB::Tn5 were observed (Fig. 3A). When P. luminescens was cultured in NGM, the stationary growth phase of the mutant pdxB::Tn5 was around an OD₆₀₀ of 0.25, which was lower than the wild-type TT01 OD₆₀₀ of 0.45 (Fig. 3B). This growth deficiency of the mutant pdxB::Tn5 in NGM was complemented by pUC19-pdxB but not by pUC19 (Fig. 3B). When P. luminescens was cultured in minimal medium (VBG), wild-type TT01 reached a stationary growth phase (OD₆₀₀ of 1.0) at around 72 h. In contrast, the mutant pdxB::Tn5 grew to a stationary growth phase at OD₆₀₀ of around 0.35 in VBG medium (Fig. 3C). This growth deficiency of the mutant in VBG medium was also complemented by pUC19-pdxB but not by pUC19 (Fig. 3C). Production of reddish pigment was observed in VBG medium from wild-type TT01 and the complemented mutant (pdxB::Tn5/pUC19-pdxB) but not from the other two strains (pdxB::Tn5 and pdxB::Tn5/pUC19) (Fig. 3D). Mutation in pdxB leads to growth- and pigmentation-deficiency phenotypes under nutrient-poor conditions, but such phenotypes can be aided by the nutrient-rich condition.

FIG 3.

Growth and pigmentation of P. luminescens in different nutrient media. Growth curves correspond to the four strains, TT01, pdxB mutant, pdxB::Tn5/pUC19, and pdxB::Tn5/pUC19-pdxB, in LB (A), NGM (B), and VBG medium (C). OD₆₀₀ values were measured every hour for 12 h and every 12 h (LB and NGM) or every 24 h (VBG) until stationary phase. The data are the means ± standard errors of the means (SEM) from at least three biological replications. (D) Pigments produced by four P. luminescens strains in VBG medium, grown for 5 days at 30°C.

Supplementation of B₆ vitamers restores growth deficiency.

Because PdxB is involved in de novo vitamin B₆ biosynthesis (Fig. 1B), we expected that supplementation of PLP, the active form of vitamin B₆, might restore growth. First, we supplemented PLP (0, 1, 10, and 100 μM in final concentration) into VBG medium and measured the growth of wild-type TT01 and mutant pdxB::Tn5. Supplementation of PLP had no effect on wild-type TT01 at any concentrations; however, it effectively restored growth of the mutant pdxB::Tn5 in a dose-dependent manner (P < 0.01; two-tailed t test) (Fig. 4A). These results suggest that the mutation in pdxB results in loss of de novo PLP biosynthesis that attenuates P. luminescens growth under nutrient-poor conditions.

FIG 4.

Growth of P. luminescens in VBG medium with B₆ vitamers. OD₆₀₀ values are shown for two P. luminescens strains, TT01 and mutant pdxB::Tn5, in VBG medium supplemented with PLP (A), PL (B), PN (C), and PM (D) (final concentrations were 0, 1, 10, and 100 μM), grown at 30°C for 48 h. The data are the means ± SEM from three biological replicates. Differences between strains at each concentration of supplement were statistically examined (*, P < 0.01; **, P < 0.001; two-tailed t test).

In addition to the de novo pathway, bacteria are known to have a salvage pathway to synthesize the active form of vitamin B₆ (Fig. 1B); in that pathway, PLP is biosynthesized from three different vitamin B₆ forms (PN, PM, and PL) through phosphorylation by PdxY and PdxK kinases. By in silico analysis, we found pdxY but not pdxK in the P. luminescens genome using the Kyoto Encyclopedia of Genes and Genome (KEGG) (http://www.genome.jp/kegg/). Therefore, we expected that supplementation of the three vitamers could restore the growth deficiency of the mutant pdxB::Tn5 by the salvage pathway. We supplemented three vitamers (at concentrations of 0, 1, 10, and 100 μM) into VBG medium and measured the growth of wild-type TT01 and mutant pdxB::Tn5. As seen for the PLP supplementation, the three vitamers could restore growth in pdxB::Tn5 (Fig. 4B to D). From these results, we reveal the existence of a salvage pathway that can compensate for the main vitamin B₆ biosynthesis pathway in P. luminescens. We also confirmed that the growth deficiency of the mutant pdxB::Tn5 is due to the lack of vitamin B₆.

Vitamin B₆ is an essential compound for P. luminescens pathogenicity.

We tested the effect of PLP supplementation on mutant pdxB::Tn5 virulence against C. elegans. C. elegans L1 larvae on the P. luminescens mutant pdxB::Tn5 could grow to a body size of around 700 μm after 48 h at 20°C. Supplementation of PLP in NGM (final concentration was 10 μM) fully restored the pathogenicity of the mutant pdxB::Tn5; the body size of C. elegans on the mutant pdxB::Tn5 with PLP supplementation was as short as that on the wild-type TT01 (Fig. 5; see also Table S4 in the supplemental material). Data for this observation were investigated by GLM analysis. The effects of strain (wild-type TT01 or mutant pdxB::Tn5), PLP supplementation (with or without supplementation), and interaction (strain × PLP supplementation) were determined by using coefficients estimated by the GLM (see Table S5A in the supplemental material). Values for Akaike's information criterion (AIC) were compared among the models using the package MuMIn (https://cran.r-project.org/web/packages/MuMIn/index.html) for model selection (see Table S5B), and model 5 was always selected as the best model in each of the three trials. We performed likelihood ratio tests using chi-squared approximation, where model 4 was used as a null model and model 5 was used as an alternative model (see Table S5B). The difference between the deviances for the two models was used for a test statistic. The result suggested that model 5 is more suitable in significance statistically (df = 1, P < 0.001); namely, the effect of interaction is significant. Therefore, we concluded that vitamin B₆ is an essential compound for P. luminescens pathogenicity against C. elegans. We then evaluated the virulence of the wild-type TT01 and mutant pdxB::Tn5 against the insect Z. morio. The pathogenicity of the mutant pdxB::Tn5 was clearly lower than that of the wild-type TT01 (P < 0.05; log rank test) (Fig. 6).

FIG 5.

Effect of PLP supplementation on P. luminescens virulence against C. elegans. Data represent body lengths of C. elegans grown on wild-type TT01 and mutant pdxB::Tn5 with or without PLP supplementation (10 μM) from the L1 stage. Representative data from three independent trials are shown. See Table S4 in the supplemental material for details on body lengths in each trial. The effect of the interaction between strain and PLP supplementation was significant for each trial (df = 1, P < 0.001; likelihood ratio test).

FIG 6.

P. luminescens pathogenicity against Z. morio. Survival curves correspond to Z. morio injected with wild-type TT01 or mutant pdxB::Tn5, based on three trials (total n = 36 in TT01 and n = 35 in pdxB::Tn5). The virulence of the P. luminescens mutant against Z. morio is significantly lower than that of the wild type (P < 0.05; log-rank test).

DISCUSSION

Vitamin B₆ (or B₆ vitamer) is one of the essential vitamin compounds involved in many aspects of metabolism in all living organisms (21, 22). Vitamin B₆ is the generic name encompassing six biologically interconvertible compounds: pyridoxal (PL), pyridoxine (PN), pyridoxamine (PM), and the phosphorylated forms pyridoxal 5′-phosphate (PLP), pyridoxine 5′-phosphate (PNP), and pyridoxamine 5′-phosphate (PMP) (http://www.chem.qmul.ac.uk/iupac/misc/B6.html). PLP is the active form of vitamin B₆ that plays an important role as a cofactor for more than 140 enzymes, which include those involved in amino acid, lipid, and carbohydrate metabolism (21, 22). Because several reports show evidence of its antioxidative properties, vitamin B₆ is also sold as a food supplement for human health (23–25). Two distinct pathways are known for de novo PLP biosynthesis: a small group of Gammaproteobacteria, including E. coli, have the deoxyxylose 5′-phosphate (DXP)-dependent pathway; and archaea, fungi, plants, and other bacteria, such as Bacillus subtilis, have the DXP-independent pathway (22, 26, 27). In the DXP-dependent pathway, mainly studied in E. coli, PLP is synthesized from d-erythrose-4-phosphate (E4P) by several enzymes. 4-Phosphohydroxy-l-threonine (4PHT) and DXP are used as the substrates to form PNP by PdxA (EC 1.1.1.262) and PdxJ (EC 2.6.99.2) and then oxidized by PdxH (EC 1.4.3.5) to form PLP (Fig. 1B).

In addition to the de novo pathway, bacteria have the salvage pathway to synthesize PLP (28, 29). In E. coli, PN and PM are phosphorylated and then oxidized by PdxK (EC 2.7.1.35) and PdxH (EC 1.4.3.5) to form PLP, and PL is phosphorylated by PdxK or PdxY (EC 2.7.1.35) to form PLP. We confirmed that supplementation of PL, PN, and PM restored the growth deficiency of the pdxB mutant in minimal medium. Since there is no PdxK-coding gene in the P. luminescens genome, we supposed that PdxY might act as a universal kinase for all three vitamers in the P. luminescens salvage pathway.

PLP is known as a cofactor for more than 140 enzymes (21, 22) and has been recently considered a virulence factor (30–33). Two Gram-positive human-pathogenic bacteria, Mycobacterium tuberculosis and Streptococcus pneumoniae, are known to have a DXP-independent de novo vitamin B₆ pathway, which is required for growth in vitamin B₆-free medium and host colonization in the mammalian model (31, 32). Two Gram-negative pathogenic bacteria, Helicobacter pylori and Campylobacter jejuni, are known to have a DXP-dependent de novo B6 pathway, and the role of PLP as a cofactor for a virulence-related enzyme is partially understood (30, 33). In H. pylori, PLP is a cofactor that aids PseC's aminotransferase activity. PseC is necessary for glycosylation of flagella, which are essential for motility and colonization during host infection (30, 34). pdxA gene mutation also causes defects in flagellar glycosylation, resulting in attenuation of motility and virulence in H. pylori (30). Because P. luminescens does not have the pseC gene, PLP might not contribute to the pathogenicity in P. luminescens via flagellar modification as observed in H. pylori and C. jejuni.

Poor growth caused by the lack of PLP biosynthesis could result in virulence deficiency. Mutations in some genes related to de novo PLP biosynthesis are also reported to cause growth deficiencies in other pathogens (30–32). In the present study, the pdxB gene mutant lacks de novo vitamin B₆ synthesis, causing growth- and pigmentation-deficient phenotypes under nutrient-poor conditions that are suppressed by nutrient-rich conditions and PLP supplementation. P. luminescens produces some secondary metabolites such as toxins and pigments during post-log-phase to stationary-phase growth (35, 36). In addition, recent research revealed that the group-coordinated behavior of Photorhabdus, which includes pathogenicity, is controlled by density-dependent cell-cell communication (37–39). Therefore, poor growth might cause inadequate production of secondary metabolites that results in the virulence deficiencies.

The P. luminescens pdxB mutant exerted pathogenicity against the insect host Z. morio but more slowly than the wild-type TT01 (Fig. 6). As PLP is a fundamental compound for all organisms (21) and Z. morio is reported to contain 3.2 mg of PN per 1 kg (40), the pdxB mutant might exert weak pathogenicity despite lacking the de novo vitamin B₆ synthetic pathway by scavenging it from the host.

P. luminescens has the ability to change between two phenotypic forms (pathogenic [P] form and mutualistic [M] form), depending on the environmental conditions (41). The highly pathogenic P form is known to produce more secondary metabolites, including essential nutrients for H. bacteriophora development. On the other hand, the M form constructs specific fimbriae on the surface of the cell that are essential for adhesion to maternal intestinal cells of Heterorhabditis nematodes. Switching between such forms is controlled by the inversion of a single promoter that coordinates bacterial pathogenic and mutualistic life cycles (41, 42). Opposite expression patterns between the pdxA and pdxJ genes were observed in the two different formations with expression ratios of the M form to the P form of 2.1 in pdxA and 0.49 in pdxJ (41). Therefore, the vitamin B₆ pathway might also play an important role in mutualistic life cycles in P. luminescens.

The screening method established here might reveal new virulence mechanisms of P. luminescens against insects and nonsymbiotic nematodes. Further investigation with this method is being conducted in order to understand how the EPN created a symbiotic relationship with such a highly pathogenic bacterium during their evolution and to understand the pathogenic mechanism of bacteria.