OUP user menu

The Photorhabdus Pir toxins are similar to a developmentally regulated insect protein but show no juvenile hormone esterase activity

Nicholas Waterfield, Shizuo George Kamita, Bruce D. Hammock, Richard Ffrench-Constant
DOI: http://dx.doi.org/10.1016/j.femsle.2005.02.018 47-52 First published online: 1 April 2005


The genome of the insect pathogen Photorhabdus luminescens strain TT01 contains numerous genes predicting toxins and proteases. Within the P. luminescens TT01 genome, the products of two loci, plu 4093-plu 4092 and plu 4437-plu 4436, show oral insecticidal activity against both moth and mosquito larvae. The proteins encoded by these loci, here termed ‘Photorhabdus insect related’ (Pir) proteins A and B, show similarity both to δ-endotoxins from Bacillus thuringiensis (Bts) and a developmentally regulated protein from a beetle, Leptinotarsa decemlineata. The beetle protein has been inferred to possess juvenile hormone esterase (JHE) activity due to its developmentally regulated pattern of expression and the Photorhabdus proteins PirA and PirB have been proposed to be mimics of insect JHEs that can disrupt insect metamorphosis by metabolizing the insect growth regulator juvenile hormone (JH) [Nat. Biotechnol. 21 (2003) 1307–1313]. Here we confirm that, when injected together, PirA and PirB from two different Photorhabdus strains have insecticidal activity against caterpillars of the moth Galleria mellonella but show no oral activity against a second moth species Manduca sexta. Direct measurement of JHE activity, however, shows that the Pir proteins are not able to metabolise JH. These data show that the Pir proteins have no JHE activity, as suggested, but leave the mode of action of these interesting proteins uncertain.

  • Juvenile hormone esterase
  • Insecticidal toxin
  • Photorhabdus luminescens

1 Introduction

The genome of the insect pathogen Photorhabdus luminescens subsp. laumondii strain TT01 (termed TT01 for brevity) contains numerous genes predicting toxins, hemolysins and proteases, which may be important for insect pathogenicity [1]. Of the predicted toxins, only two classes of genes have been investigated in detail. The first, the toxin complex (tc) genes, encode high molecular weight Tc toxins with oral activity against a range of caterpillar pests [2]. Although the functional activity of different Tc protein sub-domains has recently been investigated via their expression in tissue culture cells [3], their precise mode of action remains obscure. The second class of toxins, encoded by homologs of the makes caterpillars floppy (mcf) gene, promote programmed cell death (apoptosis) within the hemocytes (insect phagocytes) and midgut epithelium of injected insects [4]. This destruction of the midgut reduces the caterpillar's body turgor and makes them ‘floppy’. Interestingly, the Mcf protein carries a Bcl-2 homology domain 3-like (BH3-like) domain and directly activates apoptosis in treated cells, suggesting that it may be mimicking a pro-apoptotic BH3 domain only protein [5].

Hints that a third class of insecticidal toxin genes were encoded in Photorhabdus genomes originally came from sample sequencing of the P. luminescens subsp. akhurstii strain W14 (termed W14) genome [6]. Two clones from the W14 sample sequence predicted proteins with limited similarity to insecticidal δ-endotoxins from Bacillus thuringiensis[7]. These same clones also predict similarity to a developmentally regulated protein from the Colorado potato beetle Leptinotarsa decemlineata[8]. This beetle protein was presumed to have juvenile hormone esterase (JHE) [9] activity based largely on its developmental profile in insects [8,1013], but in the absence of a direct demonstration of juvenile hormone (JH) hydrolysis via JHE enzyme activity. JHs and/or its metabolites are involved in a wide range of functions in larval and adult insects including roles in development, reproduction, diapause, and migration [6]. But JH is perhaps most commonly known as the hormone whose presence ensures a larval–larval moult rather than a larval–pupal moult [14]. As the absence of JH at the time of moulting leads to a larval–pupal moult, disruption of JH titre by JHE or a JH expoxide hydrolase has been proposed as a mechanism of insect control [9]. In the paper describing the sequencing of the complete P. luminescens TT01 genome, the insecticidal activity of the proteins produced by these δ-endotoxin and JHE-like genes was investigated [1]. These proteins are encoded at two distinct loci in the TT01 genome, plu 4093-plu 4092 and plu 4437-plu 4436, each with a pair of proteins encoded at each locus [1]. Recombinant Escherichia coli clones carrying either of these two pairs of TT01 genes showed oral activity against both mosquito larvae and the larvae of the moth Plutella xylostella[1]. As plu 4092 and plu 4436 encode bacterial proteins with similarity to a developmentally regulated protein from the Colorado potato beetle [1], and since this beetle protein is indirectly inferred to have esterase activity against JH [8,1013], we directly tested the gene products plu4093 and plu4092 of TT01, and their homologs from P. asymbiotica strain ATCC43949, for insecticidal and JHE activities.

Here, we show that these insecticidal proteins are not associated with any JHE activity. We have therefore term these proteins ‘Photorhabdus insect related’ (Pir) toxins, rather than JHE-like proteins, as they have no detectable JHE activity. We also show, via expression in E. coli, that injection of both proteins PirA and PirB is necessary for activity against caterpillars of the wax moth Galleria mellonella. Caterpillars injected with PirA and PirB together showed rapid morbidity and subsequent darkening of the cuticle and death. This study shows that Pir proteins are not associated with any JHE activity and also casts doubt on the presumed function of the developmentally regulated L. decemlineata protein. However, the mechanism whereby the Pir toxins induce melanization or insect death remains unclear.

2 Materials and methods

2.1 Cloning of the pir genes

The toxin encoding genes pirA and pirB were amplified from Photorhabdus genomic DNA using the rTth DNA polymerase system (Perkin Elmer) in the polymerase chain reaction (PCR). PCR products were digested and cloned into the KpnI and HindIII restriction sites of the arabinose-inducible expression plasmids pBAD30 using T4-DNA ligase (New England Biolabs) and EC100 electrocompetent cells (Epicentre). The pirA and pirB genes were cloned with and without their own promoters using the PCR primers shown in Table 1 resulting in the clones detailed in the diagram in Fig. 1.

View this table:
Table 1

Sequences of the PCR primers used to amplify pirA and pirB genes from P. asymbiotica ATCC43949 and P. luminescens TT01

pBAD30 derived cloneCloned PCR product (see Fig. 2)Promoters driving expressionPrimer sequence 5′?3′
Lowercase bases mismatched to create restriction sites
pPirABproThe [pirABpro] PCR product from strainPara+PpirJHEasyProF: CGTTTATTGGTaccGTAATGAAAGGCA
The pirAB operon including promoter regionATCC43949JHEasyR: ATACAAAgcTTGCCGACATCAAAAGA
pPirABThe [pirAB] PCR product from strainParaJHEasyF: GAATAggTAccTGTAAGTTGAGTAGGTT
The pirAB operon excluding the promoter regionATCC43949JHEasyR: ATACAAAgcTTGCCGACATCAAAAGA
pPirAThe [pirA] PCR product from strainParaJHEasyF: GAATAggTAccTGTAAGTTGAGTAGGTT
pPirBThe [pirB] PCR product from strainParaAJHEF:CGCGgTaCcACTAGAAATCTATTGGGT
pPirABproThe [pirABpro] PCR product from strainPara+PpirJHEtt01ProF: TGAAAGGTAccTGAATTGTATTCA
The pirAB operon including promoter regionTT01JHEtt01R: TAAAAGCTtATAACATTCTACGTACA
pPirABThe [pirAB] PCR product from strainParaJHEtt01F: AGTGGTAccACTGTGTTTTGAAAATAT
The pirAB operon excluding the promoter regionTT01JHEtt01R: TAAAAGCTtATAACATTCTACGTACA
Figure 1

Homology of predicted P. luminescens PirB to (A) Leptinotarsa decemlineata developmentally regulated protein (AF039135) and (B) the parasporal crystal protein Cry2A of Bacillus thuringiensis (CAA10670).

2.2 Insect bioassays

Insects were bioassayed with recombinant PirA and PirB proteins either via injection directly into the caterpillar hemocoel or, for oral bioassay, via application of bacterial preparations to artificial diet. For the insect bioassays, recombinant E. coli EC100 containing the expression plasmids pPirABaymb, pPIRABproaym, pPirAaymb, pPirBaymb, pPirABtt01, pPIRABprott01 and the pBAD30 vector alone were grown at 30 °C with aeration to an OD600 of 0.4, whereupon l-arabinose was added to a final concentration of 0.2% (w/v) to induce protein expression, and the cultures grown overnight. Harvested bacterial cells were then washed twice in an equal volume of sterile phosphate buffered saline (PBS) and re-suspended in an equal volume of PBS. For injection, last instar larvae of G. mellonella were cooled on ice for 20 min and then twenty animals per sample were injected with 10 μl of re-suspended bacteria.

For bioassays of oral activity caterpillars of a different moth, Manduca sexta, were used in a procedure described previously [2]. Briefly, 100 μl of whole overnight cultures were applied to 1 cm3 disks of artificial wheat germ diet. Treated food blocks were allowed to dry for 20 min and then three first instar M. sexta were placed on each of six food blocks per treatment. Treated blocks were held at 25 °C for 7 days and larvae were then weighed and scored as alive or dead. Final larval weights were expressed relative to the pBAD30 only control. Injectable toxicity against M. sexta was also tested. Here 50 μl of these same induced cultures were injected into a cohort of 10, fourth instar M. sexta, which were monitored over 7 days for any deleterious affects.

2.3 Juvenile hormone esterase assay

To assess the potential JHE activity of the recombinant Pir proteins of Photorhabdus, 0.2 ml of an overnight culture of each clone was added to 10 ml of Luria Broth (LB) medium supplemented with 100 μg ml−1 ampicillin. These cultures were grown with aeration to an OD600 of 0.3, at which point l-arabinose was added to a final concentration of 0.2% (w/v). Sterile filtered water was added in place of L-arabinose for the un-induced negative control. At 8 h post-induction, bacteria were removed by centrifugation (2000g for 10 min, at 5 °C) and supernatants were stored overnight at 5 °C. These supernatants were concentrated fivefold using a Centricon30 (Millipore) protein concentration column. The presence of toxin in these concentrated supernatants was confirmed by running 28 μl on a 12% SDS–PAGE gel. JHE activity was tested according to the protocol of Hammock and Sparks [15], using a mixture of 3H-labeled (New England Nuclear) and unlabeled (Sigma) JH-III as a substrate (final concentration of 5 × 10−6 M containing 700–800 counts min−1 of 3H-JH-III). Each assay which consisted of a 100 μl reaction in 10 mM Tris, pH 8.0, or LB supplemented with ampicillin was run for 15 min at 30 °C in duplicate. Purified recombinant JHE (rJHE) from M. sexta, expressed from recombinant baculovirus infected High 5 cells [16], was assayed under the same conditions as a positive control.

3 Results and Discussion

3.1 Pir sequence, expression and toxicity

The pir genes were first identified in a sample sequence of the genome of P. luminescens strain W14, as clones (numbers 01891 and 01973) predicting proteins with similarity to δ-endotoxin genes from B. thuringiensis[6]. In current database searches, whilst the predicted amino acid sequence of pirA shows little significant homology with known sequences, the predicted amino acid sequence of pirB shows significant BlastX matches to both a developmentally regulated protein from the beetle L. decemlineata and to a δ-endotoxin from B. thuringiensis (plu 4092 and plu 4436, E= 9e−66 and 5e−58 to the beetle protein and 4e−4 and 1e−3 to the endotoxin respectively). Specifically, PirB shows 20.5% identity and 41.5% similarity over 229 amino acids to the N-terminal domain of the Cry2A insecticidal toxin (Fig. 1(a)). We note that this region of the Cry2A protein is the pore-forming domain, suggesting that the Pir proteins may carry a similar motif. PirB also shows more extensive similarity (35.2% identity and 58.4% similarity over 401 amino acids), partly over the same region, [6] to a developmentally regulated protein from the beetle L. decemlineata[8] (Fig. 1 (b)). This developmentally regulated protein from L. decemlineata has been purified and its expression profile monitored throughout development in relation to the titre of JH [8,1013]. Although well characterized, to our knowledge no direct JHE activity has been measured from this beetle protein. We, therefore, questioned whether the similarity of the Photorhabdus Pir proteins to the beetle protein did indeed infer that Pir proteins are JHE-like or that they carry JHE activity.

To examine the insecticidal toxicity of Pir proteins and to test for JHE activity, we expressed recombinant PirA and PirB proteins from two different species of Photorhabdus. We tested the gene products of plu 4093 and plu 4092 from P. luminescens TT01, and their homologs from P. asymbiotica strain ATCC43949, for insecticidal and JHE activities. Induction of recombinant E. coli cultures carrying pPirAB plasmids from either P. luminescens TT01 or P. asymbiotica ATCC43949 revealed expression of proteins of the predicted sizes for both PirA (45 kDa) and PirB (14 kDa) via SDS–PAGE (Fig. 2). Injection of both PirA and PirB proteins together into Galleria caused the larvae to become moribund and die within 72 h. All the plasmid constructs tested, that carried both pirA and pirB genes together, from either species of Photorhabdus, showed 100% mortality within the injected larvae with 72 h (Table 2). We note, however, that combinations of PirA and PirB had neither injectable nor oral activity against a second caterpillar M. sexta (data not shown). Injection of either PirA or PirB alone into caterpillars of Galleria was not associated with any mortality. However, subsequent mixture of individual PirA and PirB preparations reconstituted full activity against this insect (Table 2).

Figure 2

Genomic organization of loci encoding PirA and PirB in two different species of Photorhabdus. Note that the genomic organization is conserved in P. luminescens strain TT01 and P. asymbiotica strain ATCC43949. PCR products used for cloning are shown (horizontal arrows). The low GC-content of the region and the location of ERIC sequences (vertical arrows) suggest that this region has been horizontally acquired. The black ORF has homology to fis-DNA invertase.

View this table:
Table 2

Bioassay and JHE activity data

Clone (see Table 1)InductionSpeciesGalleria mortality by 24 h (%)Galleria mortality by 48 h (%)Galleria mortality by 72 h (%)

Finally, it is interesting to note that uninduced pBAD30 plasmids still show 100% insecticidal activity (Table 2). We, therefore, suspect that the pBAD30 plasmids are ‘leaky’ and that sufficient Pir proteins are made within un-induced cultures to cause full mortality amongst injected larvae, despite the fact that the amount of recombinant protein is not detectable by SDS–PAGE (Fig. 3). Further, when similar plasmids carry the likely native promoters of pirA and pirB, this effect is reduced. Thus, it seems likely that these genes are tightly downregulated by the DNA that surrounds them and in its absence even a low level of protein expression is still sufficient for insect toxicity.

Figure 3

SDS–PAGE analysis of E. coli cell lysate showing expression of the pirAB genes under control of the Para-promoter with (+) and without (−) 0.2% arabinose induction. Note the appearance of proteins of the expected size upon arabinose induction. Note also that when the native promoter element was included in the PCR products (clones pPirABpro), protein expression was significantly reduced, suggesting that the promoter is under strong repression, by normal E. coli factors, even under arabinose induction.

These experiments support the concept that the PirA and PirB proteins are insecticidal when applied together. However, these data do not clarify the likely mode of secretion or release of the mature toxins from the host bacteria. As the PirA and PirB proteins can be detected in the cytosol of recombinant E. coli, it seems reasonable to expect them to be released via lysis of the recombinant E. coli cells following an insect immune response to their injection. Unfortunately, we cannot investigate the role of lysis directly using these recombinant bacteria, as lysis of the host E. coli cells alone leads to significant insect mortality, probably due to extensive release of E. coli LPS or peptidoglycan.

3.2 Pir proteins have no JHE activity

No JHE activity could be detected from recombinant PirA and PirB when expressed E. coli. Under the same conditions tested (15 min incubation with 0.5 nmoles of JH-III substrate, at 30 °C), control experiments indicated that approximately 18% of the substrate is hydrolyzed by 80 ng of purified rJHE from M. sexta. This corresponds to the “expression” of less than 1600 ng of “JHE” of Photorhabdus in the 10 ml of culture of induced bacteria. This data suggest that these recombinant Pir proteins do not have JHE activity. This lack of any demonstrable JHE activity in the Photorhabdus Pir proteins, but their clear similarity to a developmentally regulated insect protein, has two major implications. First, the concept that they might be JHE-like is based on the assumption that the developmentally regulated protein in L. decemlineata is indeed a JHE. Despite the extensive study of the hormonal regulation of this beetle protein, we still feel that JHE activity data is lacking for the beetle protein and that this protein itself cannot therefore be confirmed to have such an activity. Second, Photorhabdus does indeed make Pir toxins both injectably Fig. 4, and orally [1], active against different insects and these do show some similarity to both a protein involved in beetle development and to Bt Δ-endotoxins. Therefore, although they carry no JHE activity, the Photorhabdus Pir toxins could still potentially modulate insect development or be distant relatives of the Cry toxins of Bt. The mode of action of these interesting and novel bacterial toxins is therefore worthy of further study.

Figure 4

The effect of PirAB on Galleria mellonella. 10 μl of an overnight induced culture was injected per insect. (A) E. coli [pBAD30] negative control (B) E. coli [pPirABasymbiotica].


We thank Julian Parkhill and all in the Pathogen Sequencing group at the Sanger for sequencing P. asymbiotica. Supported by grants from the BBSRC initiative Exploiting Genomics and by a Royal Society Merit Award to R. ff-C.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
View Abstract