Bacterial strains, plasmids, primers and growth conditions

The strains and plasmids used in this study are listed in Table 1. All primers used in this study (Eurogentec) are listed in Supplemental Table 1. Photorhabdus and Escherichia coli strains were routinely grown at 28°C and 37°C, respectively, in Luria-Bertani (LB) broth or on LB solid medium. Photorhabdus was grown in M9 liquid medium with concentrations of MgSO₄ (10 µM and 10 mM) and supplemented with 0.1% casamino acids, 0.41 mM nicotinic acid, 9.1 mM sodium pyruvate, 0.2% glycerol and 0.1 mM CaCl₂. When required, antibiotics were used at the following final concentrations: kanamycin (Km) 20 µg/ml, gentamicin (Gm) 30 µg/ml, ampicillin (Ap) 100 µg/ml.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Strains and plasmids used in this study.

https://doi.org/10.1371/journal.pone.0110060.t001

Inference of the evolutionary relationships of Ail, PagC and OmpX-related proteins

Ail, PagC or OmpX annotated proteins in Photorhabdus luminescens TT01 (plu1967, plu2480, plu2481) and Photorhabdus asymbiotica ATCC43949 (PAU_02047 and PAU_02601) were identified and retrieved using the protein family analysis tool PipeAlign [38]. Outputs were pooled and resulting dataset were curated for protein multiple occurrences. The sequences were aligned using ClustalW [38] followed by manual curation. The sequence alignment was generated by Gblocks [39] and unambiguously aligned amino acid sites were retained for phylogeny inference using the maximum likelihood method implemented in PhyML [40]. Analyses were engendred under the LG model of amino acid replacement [41] with a gamma distribution of evolutionary rates across sites [42]. Internal branch supports were evaluated using the approximate Likelihood Ratio Test [43].

Molecular techniques and RNA preparation

DNA manipulations were carried out as previously described [44]. Plasmids were introduced into E. coli WM3064 (Table 1) by transformation and transferred to P. luminescens TT01 by filter mating [45]. All constructs were sequenced by Eurofins MWG Operon (Ebersberg, Germany). Total RNA was extracted with TRIzol reagent according to manufacturer's instructions (Invitrogen) and purified using RNeasy miniprep kit (Qiagen), including a DNase I treatment step. For each RNA preparation, we assessed DNA contamination by carrying out a control PCR. The quantity and quality of RNA were assessed with a NanoDrop 2000 spectrophotometer (Thermo Scientific) and an Agilent 2100 Bioanalyzer with the RNA 6000 Nano LabChip kit (Agilent), respectively. Material for real-time quantitative polymerase chain reaction (RT-qPCR) analysis was prepared by extracting total RNA from the P. luminescens wild-type strain and the phoP mutant grown in Luria broth (OD₅₄₀ = 0.5–0.8). The gene expression level was evaluated during the growth phase.

Real-time quantitative polymerase chain reaction analysis

RT-qPCR was performed in two steps. First, the cDNA was synthesized from 500 ng of total RNA, with Super Script II Reverse Transcriptase (Invitrogen) and random hexamers (100 ng/µl) (Applied Biosystems). We then carried out qPCR in triplicate with the LightCycler 480 SYBR Green I Master kit from Roche Diagnostics, with 1 µl of cDNA synthesis mixture (diluted 1∶100) and 1 µM of specific primers for the genes studied (Supplemental Table 1). The enzyme was activated by heating for 10 min at 95°C. All qPCRs were performed in three technical replicates, with 45 cycles of 95°C for 5 seconds, 60°C for 5 seconds and 72°C for 10 seconds, and were monitored with the Light Cycler 480 system (Roche). Melting curves were analyzed for each reaction and each curve contained a single peak. The data for each sample were expressed relatively to the expression level of gyr, using REST software 2009 [46] as previously described [47]. This method provided a relative quantification of the target gene expression with respect to a reference gene (gyrB).

Overexpression and purification of PhoP recombinant protein

The entire coding region of phoP gene from TT01 strain was amplified by PCR and digested by NdeI and BamHI. The obtained PCR product was ligated into the same site of the expression vector, pETPhos [48] inserting a His-tag in the N-terminal part of PhoP thus generating the plasmid P_T7PhoP-His. The recombinant plasmid encoding a PhoP-His fusion protein was transformed into E. coli BL21 (DE3) pLysS cells. At an OD between 0.5–0.8, the expression of PhoP-His was induced by adding Isopropyl-beta-D-thiogalactoside at 0.5 mM. An overnight induction was then performed at 18°C. Bacterial culture was centrifuged at 7,000×g for 15 min at 4°C and washed twice in resuspension buffer (Tris 5 mM pH 7.5, NaCl 300 mM, glycerol 10%, Imidazole 10 mM). The pellet was frozen at −80°C for 30 min, then suspended in 5 ml resuspension buffer and lysed by sonication during 10 min at 4°C. Lysis products were centrifuged at 10,000×g during 30 min at 4°C. Five hundred µl of pre-equilibrated beads of Ni-NTA agarose (Qiagen) in wash buffer (Tris 5 mM pH 7.5, NaCl 300 mM, glycerol 10%, Imidazole 15 mM) were added to the supernatant fraction and incubated during 45 min with shaking at 4°C. The fraction was centrifuged at 500×g during 2 min at 4°C and wash 5 times with wash buffer. Protein was eluted twice in 1 ml elution buffer (Tris 5 mM pH 7.5, NaCl 300 mM, glycerol 10%, Imidazole 200 mM). Concentration of recombinant protein was assessed by Bradford assay and controlled by SDS-PAGE gel. Recombinant proteins were conserved at −80°C until use.

Electrophoretic mobility-shift assays

The promoter of ail1_Pl was PCR-amplified from the genomic DNA of TT01 strain using specific primers (Supplemental Table 1) and purified using the High Pure PCR Product Purification kit (ROCHE). The 5′ ends of DNA fragment were labeled using [γ-³²P] ATP and T4 polynucleotide kinase (Promega). Radioactive DNA probe (2000 cpm/ml), 200 ng of poly(dI-dC)-poly(dI-dC) (SIGMA) and different amounts of PhoP-His were mixed with binding buffer (50 mM tris-HCl pH 8, 50 mM KCl, 50 µg/ml BSA) in a total 20 µl volume and incubated for 20 min at room temperature. The mixture was then loaded onto a native 6% (w/v) polyacrylamide TBE precast Gel (Invitrogen) and electrophoresed in 1% TBE (Tris-Borate-EDTA) buffer for 1 h at 100 V. Radioactive species were detected by autoradiography. PhoP-His was activated by in vitro phosphorylation with acetyl phosphate as previously described [49].

Construction of plasmids expressing gfp[AAV] under the control of ail1_Pl or ail2_Pl promoters

Plasmids expressing the reporter gene gfp[AAV] under the control of ail1_Pl, ail2_Pl or lac gene promoters were constructed using a previously described method [50]. The construction of P_lac-gfp[AAV] has been described elsewhere [51]. Briefly, DNA fragments located upstream from ail1_Pl (360 bp) and ail2_Pl (358 bp) were amplified by PCR from P. luminescens TT01 genomic DNA with primers containing the EcoRI and BamHI restriction site (Supplemental Table 1). The PCR products were EcoRI- and BamHI-hydrolyzed and inserted into the corresponding sites of pPROBE-gfp[AAV]. Finally, P_lac-gfp[AAV], P_ail1Pl-gfp[AAV], P_ail2Pl-gfp[AAV] were transferred by filter bacterial mating [45] in P. luminescens TT01 wild type and phoP strains.

Quantification of ail1_Pl and ail2_Pl expression in bacterial populations grown in different media

Wild-type strains carrying either P_ail1Pl-gfp[AAV], P_ail2Pl-gfp[AAV] or P_lac-gfp[AAV] constructs were cultured in black-sided, clear-bottomed 96-well plates (Greiner). For each well, a 1∶20 dilution of an overnight culture was added to the M9 minimal medium supplemented with kanamycin and different concentrations of MgSO₄ or to LB medium supplemented with kanamycin. Then, the plates were incubated at 28°C for 45 h with shaking on an orbital shaker, in an Infinite M200 microplate reader (Tecan). Absorbance at 600 nm and GFP fluorescence intensity (excitation at 485±4.5 nm; emission at 520±10 nm) were measured every 30 min. Specific fluorescence was obtained by dividing fluorescence units (at the maximum level of expression) by the absorbance value.

Serum-killing assay

The serum-killing assay was performed as described previously [5] with overnight culture of Escherichia coli strain XL1Blue in LB with ampicillin 100 µg/ml at 37°C. For each bacterial strain, three independent assays were performed with human serum from human male AB sterile plasma (Sigma-Aldrich, reference: H4522) or with serum that was heat-treated to inactivate complement (56°C, 30 min). The number of viable bacteria after incubation with serum at 37°C for 60 min was calculated by serial dilution, plating on LB agar with ampicillin 100 µg/ml and counting the colony forming units (CFU). The degree of killing was calculated as follow: log kill  =  (log₁₀ CFU per milliliter of initially added bacteria) - (log₁₀ CFU per milliliter of bacteria surviving the incubation). The resistance was expressed as the difference in log kill between XL1-Blue harboring pUC19 incubated in 50% human serum and XL1Blue harboring the recombinant plasmid incubated either in 50% human serum or heat-inactivated serum.

Cell association assays

The cell invasion and association assays were performed as described previously [52]. Escherichia coli strain XL1Blue were grown at 37°C in LB broth supplemented with ampicillin 100 µg/ml for 2.5 hours (10⁸ CFU/ml; optical density at 540 nm of 0.5; xponential-growth phase). Chinese hamster ovary cells (CHO) were grown and maintained at 37°C in RPMI Medium 1460-glutamax I (Gibco) complemented with 10% foetal bovine serum (Lonza), and 1% PenStrep 5000 U/ml (Gibco). Spodoptera littoralis cells derived from hemocytes (Sl2b) were grown and maintained at 28°C in G3 Medium pH 6,2 (TC100 2% from Gibco modified with 0.037% α-ketoglutaric acid, 0.04% β-fructose, 0.005% fumaric acid, 0.067% malic acid, 0.006% succinic acid, 0.26% sucrose, 0.02% choline chloride, 0.02% β-alanine, 0.035% sodium bicarbonate, 0.33% lactalbumin hydrolysate, and complemented with 5% foetal bovine serum, 0.016% penicillin, 0.006% streptomycin).

Susceptibility to antimicrobial peptides

In vitro susceptibility to polymyxin B sulfate (Sigma), colistin methanesulfate (Sigma), cecropin A (Sigma), Spodoptera frugiperda cecropin B was evaluated by determination of minimal inhibition concentration as previously described [53].

Phylogeny of Ail, PagC and OmpX-related proteins among Bacteria

Ail, PagC and OmpX homologs were searched in genome databases (see Material and Methods). Phylogenetic analysis (Fig. 1A) clearly separated the 89 proteins from 53 bacterial species into three well-supported groups. The Ail group recovered the y1324 canonical Ail protein described in Y. pestis KIM [4]. The PagC group contained the STM3031 PagC protein from Salmonella Typhimurum [54]. Finally, the OmpX group was characterized by several proteins annotated OmpX, such as Ent638_1301 from Enterobacter sp. 638. Interestingly, the protein y1682 of Y. pestis KIM usually named Ail clusters with the OmpX group [4]. Our phylogenetic analysis clearly indicates that all the Photorhabdus proteins clustered within the Ail group (plu2480 and plu2481 from P. luminescens and PAU_02047 from P. asymbiotica) and the PagC group (plu1967 from P. luminescens and PAU_02601 from P. asymbiotica).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. The Photorhabdus genus harbors ail and pagC genes.

A. Evolutionary relationships of Ail, PagC and OmpX-related proteins. Stars indicate branch supports higher than 0.85 (used as significance threshold). The scale bar corresponds to the number of substitutions per amino acid residue site. B. Conserved genomic context of the ail (up) and pagC (bottom) genes in Photorhabdus luminescens TT01 (Pl) and Photorhabdus asymbiotica ATCC43949 (Pa) genomes. The boxes above and below the axis represent ORFs in the forward and reverse orientations, respectively. The names of some genes are indicated. The names of genomic islands (GI) previously described in the P. luminescens and P. asymbiotica genomes [55] are given.

https://doi.org/10.1371/journal.pone.0110060.g001

Genomic organization of ail and pagC genes in Photorhabdus genomes

Analysis of the P. asymbiotica ATCC43949 and P. luminescens TT01 genome sequences [36], [37] revealed that ail and pagC genes were located in conserved regions in the two Photorhabdus species, previously described as genomic islands in P. luminescens TT01 (Fig. 1B) [55]. Both insertion points and contents in the coding sequences were similar. The genomic islands from P. luminescens genomes were composed of additional genes encoding proteins potentially involved in infectious process. The toxBCD operon in the vicinity of the ail genes (GI_59) is involved in toxoflavin biosynthesis of Burkholderia glumae [56]. The phaxA1B1 genes in the vicinity of the pagC (GI_47) encodes the XaxAB-like binary toxin with insecticidal and cytotoxic activity [57], [58].

The two ail genes from P. luminescens, referred to as ail1_Pl and ail2_Pl hereafter, shared 71% of nucleotidic identity and 66% of aminoacids identity and were separated by 517 nucleotides. This adjacent position (Fig. 1B) together with their closed clustering inside the Ail group (Fig. 1A) suggests a recent tandem duplication. By contrast, the ail gene from P. asymbiotica (ail_Pa), the pagC genes of P. luminescens and P. asymbiotica (pagC_Pl and pagC_Pa, respectively) were present in single copy.

PhoP is directly regulating the expression of ail1_Pl

Data on expression of genes encoding outer membrane proteins of the Ail/OmpX/PagC family remains limited except for PagC. Indeed, pagC gene in S. Typhimurium is a PhoP-activated gene [7]. A phoP mutant has been previously constructed and described for P. luminescens TT01 [21]. We therefore performed RT-qPCR to measure levels of pagC_Pl, ail1_Pl and ail2_Pl mRNA and to calculate their ratio of values in phoP and wild type backgrounds of P. luminescens TT01 (Fig. 2A). Only the level of ail1_Pl transcript was lower in the phoP mutant than in the wild type strain, indicating that the ail1_Pl gene expression requires PhoP. Contrarily, ail2_Pl and pagC_Pl expression was not PhoP-dependent. The differential regulation of ail paralogs were independently confirmed by measuring expression of P_ail1Pl-gfp[AAV] and P_ail2Pl-gfp[AAV] fusions in wild-type and phoP backgrounds (Fig. 2B). Next, the possible direct binding of PhoP protein upstream of ail1_Pl was investigated. Electrophoretic mobility shift assays (EMSAs) were carried out to compare the interaction profiles of different amounts of PhoP protein on the 360-bp ail1_Pl promoter region (Fig. 2C). A recombinant N-terminal His-tag PhoP protein (PhoP-His) was first produced from P_T7PhoP-His vector (Table 1). The PhoP-His protein was purified and phosphorylated in vitro by incubation with acetyl phosphate. Phosphorylation efficiency and dimer formation were assessed by migration on precast Phos-tag gel (Fig. S1). Then, different amounts of phosphorylated and unphosphorylated PhoP-His were mixed with radiolabeled ail1_Pl promoter. A gel shift pattern was observed when 1.5 µM of phosphorylated PhoP-His was added (Fig. 2C). No shifted bands were observed upon incubation with unphosphorylated PhoP-His. Therefore, PhoP-His protein can specifically bind to the promoter region of ail1_Pl and the active form of PhoP corresponds to the phosphorylated isoform.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Only ail1_Pl is directly regulated by PhoP.

A. RT-qPCR: ail1_Pl expression is PhoP-dependent. Total RNA from phoP mutant or TT01 wild-type strain of Photorhabdus luminescens was used for RT-qPCR analysis with internal primers specific for the indicated genes. mRNA levels were normalized against those of a reference gene (gyrB). Data are presented as a ratio of values for phoP mutant and TT01 wild-type strain. A ratio of 1 indicates no difference in expression level between both strains. The bars indicate standard errors calculated using Taylor's series. Significant differences (p-value <0.05) are indicated by asterisks (*). The relative quantification results were obtained from three independent experiments with the REST 2009 program. B. Gene transcription monitored by GFP quantification: ail1_Pl promoter region is positively regulated by PhoP. The dynamic expression of ail1_Pl and ail2_Pl promoter in TT01 and phoP backgrounds was monitored over time after growth in LB medium. Each histogram represents the specific fluorescence at the peak of expression for each condition. One experiment representative of more than three independent experiments is shown. Standard deviations represent technical replicates. C. PhoP-His binds the promoter region of ail1_Pl. Electrophoretic mobility shift assay was carried out to test the binding of PhoP-His protein activated in vitro with ACP 10 mM (P-PhoP-His) or non activated PhoP-His (PhoP-His) on ail1_Pl promoter. The PhoP-His concentrations indicated are in micromolar. To ensure that the fixation is specific, we used BSA proteins and poly(dI-dC) in the binding buffer. D. ail1_Pl expression is higher at low MgSO₄ concentrations. We evaluated the impact of low and high MgSO₄ concentrations on ail1_Pl and ail2_Pl expression. Cultures diluted at 1/200 were grown in M9 minimal medium supplemented with 10 µM or 10 mM MgSO₄. Each histogram represents specific fluorescence at the peak of expression for each condition. Experiments were realized at least three times.

https://doi.org/10.1371/journal.pone.0110060.g002

ail1_Pl expression is high at low MgSO₄ concentrations

It has been shown that low concentrations of Mg²⁺ activate the expression of PhoP-dependent genes in Salmonella whereas high Mg²⁺ concentrations repress the system (for review see [59], [60]. In order to evaluate the role of Mg²⁺ concentration on ail1_Pl or ail2_Pl expression, wild-type strain containing P_ail1Pl-gfp[AAV], P_ail2Pl-gfp[AAV] or P_lac-gfp[AAV] were grown in minimal medium M9 supplemented with 10 µM (activating concentration) or 10 mM (repressing concentration) of MgSO₄ and gene expression was monitored by recording GFP fluorescence (Fig. 2D). We observed a 5-fold decrease of ail1_Pl expression at 10 mM MgSO₄, whereas ail2_Pl expression was not dependent on the concentration of MgSO₄. As observed for PhoP-activated genes in S. Typhimurium [19], low MgSO4 concentration increases ail1_Pl expression in Photorhabdus.

ail2_Pl and pagC_Pl genes confer human serum resistance but no eukaryotic cell association phenotype

When introduced into Escherichia coli, individual members of the Ail/PagC/OmpX-related protein family have the property to confer resistance to human serum complement and to associate with eukaryotic cell [5], [6], [11], [13]. In order to test if the ail and pagC genes from P. luminescens and P. asymbiotica have similar phenotypes, we introduced ail1_Pl, ail2_Pl, pagC_Pl, ail_Pa and pagC_Pa under the control of the constitutive promoter P_lac into the high copy number pUC19 plasmid and transformed these pUC19 derivatives into an E. coli XL1blue strain. Active transcription of corresponding cloned genes in E. coli XL1Blue was controlled by real-time PCR (data not shown). We first evaluated the human serum resistance of the E. coli XL1Blue strains harboring the different derivative plasmids (Fig. 3). The positive control, E. coli XL1Blue expressing the ail gene from Y. pestis strain KIM, y1324 (ail_Yp), showed an elevated serum resistance, similar to the one observed in all strains when the complement in the serum was heated-inactivated. Moreover, we observed intermediate serum resistance with the ail2_Pl and pagC_Pl genes whereas the resistance with the ail1_Pl, ail_Pa and pagC_Pa genes was weak.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Human serum resistance of Escherichia coli XL1Blue strains carrying the plasmid pUC19 and its derivatives harboring ail or pagC genes.

Overnight grown bacteria were tested for viability at 37°C in 50% serum (black histograms) or heat-inactivated serum (hatched histograms). The resistance was expressed as the difference in log kill between XL1-Blue harboring pUC19 incubated in 50% human serum and XL1Blue harboring the recombinant plasmid incubated either in 50% human serum or heat-inactivated serum. Means and standard errors of results from triplicate experiments are shown.

https://doi.org/10.1371/journal.pone.0110060.g003

In order to assess if the association of E. coli XL1Blue to eukaryotic cells was affected by expression of Photorhabdus ail or pagC genes, association assays of mammals cells (CHO) and insect cells (Sl2b) were performed. E. coli XL1Blue expressing ail_Yp displayed association with CHO cells (% of association of 30.43±17.06) in regards to E. coli XL1Blue harboring the pUC19 plasmid (% of association of 0.57±0.05), which is consistent with previously published data [52]. None of the Photorhabdus over-expressed genes, ail1_Pl, ail2_Pl, pagC_Pl, ail_Pa and pagC_Pa, was able to confer to E. coli XL1Blue strain an improved association with Sl2b cells to E. coli XL1Blue strain (average % of association of 10.60±6.80).

ail1_Pl, ail2_Pl and pagC_Pl do not confer antimicrobial peptide resistance

Since P. luminescens TT01 and P. asymbiotica ATCC43949 are pathogenic towards insect, the role of ail and pagC genes was studied by cultivating recombinant E. coli in insect hemolymph, which is analogous to the mammalian blood regarding the immune system. The growth of the E. coli XL1Blue derivatives was not affected when cultivated in sterile hemolymph of Spodoptera littoralis fifth-instar larvae (data not shown). Then, we compared the susceptibility to different antimicrobial peptides of the negative control E. coli XL1Blue/pUC19 and the E. coli XL1Blue harboring the different derivative plasmids. Similar minimal inhibitory concentrations were observed towards colistin (1.5–3 µg/ml), polymyxin B (0.1–0.2 µg/ml), cecropin A (6–12.5 µg/ml) or cecropin B from the Lepidoptera Spodoptera frugiperda (1.5–3 µg/ml).

Photorhabdus genomes only harbor ail and pagC gene homologs

Our analysis of 89 proteins from 53 bacterial species belonging to the Ail/PagC/OmpX family clearly distinguished three sub-families. In the two whole-assembled genomes of Photorhabdus, both Ail and PagC proteins are encoded. Interestingly, in P. luminescens TT01, we identified two intra-genome homologs ail1 and ail2. Such homologs can arise through duplication, where both gene copies are named paralogs, or by acquiring similar genes from outside sources through horizontal gene transfer, where both gene copies are named xenologs [61]. The adjacent position of ail1_Pl and ail2_Pl genes suggests recent tandem duplication. Duplication is an important hallmark of the genome plasticity of P. luminescens both in short-term adaptation and long-term evolution. Several dozen of duplicated genes have already been described in P. luminescens TT01 [36]. Interestingly, another gene encoding an outer membrane protein, the ompF-like gene, probably underwent a tandem duplication [62]. Characteristic features of an ancient whole-genome duplication were also detected in the P. luminescens TT01 genome [63]. In addition, a 275-kilobase single block duplication, with cryptic phenotypic consequences, was observed in phenotypic variants of P. luminescens [64].

What are the factors regulating ail/pagC gene expression in Photorhabdus?

No data about transcription factors involved in the control of ail/pagC expression in the Photorhabdus genus are currently available. P. asymbiotica is considered an emerging human pathogen [34]. In an attempt to find host-interacting proteins that are relevant to either human or insect infections is interesting, Wilkinson et al. identified a thermoregulation of the secretion for its Ail-like homolog in P. asymbiotica ATCC43949 (Ail_Pa in our study) as Ail_Pa is secreted at 37°C but not at 30°C [37]. In our study, we decided to use the available phoP mutant of P. luminescens TT01 [21] to compare relative expression of ail1_Pl, ail2_Pl and pagC_Pl in wild-type and phoP backgrounds. Only P. luminescens ail1 expression is PhoP-dependent by contrast with ail2_Pl and pagC_Pl. The pagC gene from S. Typhimurium is regulated by PhoP in a indirect way via SlyA [18]. Surprisingly, our electromobility shift assays showed that the His-tagged PhoP protein from P. luminescens directly binds to the promoter region of ail1. This is the first time that ail has been described to rely on the direct control of PhoP in Gram-negative bacteria. In addition, we monitored the kinetic of ail1 expression in different culture conditions and determined that ail1 expression was reduced in the presence of high concentrations of MgSO₄. The environmental deprivation of Mg²⁺ as a signal activating the PhoP/PhoQ signal transduction cascade was first described in Salmonella [19]. In Photorhabdus, Derzelle et al. [21] showed that PhoP-dependent expression of the first gene of the pbgPE operon, involved in LPS modifications, relies on the Mg²⁺ concentrations in the culture medium. Like pbgPE, expression ail1_Pl is higher at low Mg²⁺ concentrations than at high concentrations. In Salmonella and P. luminescens, both the phoP gene and the pbgPE operon are involved in virulence in the mouse and the insect models, respectively [7], [21], [65], [66]. In P. luminescens, it is likely that the deficience of PhoP-dependent pbgPE expression is responsible for the avirulence of phoP mutant rather than the one of ail1_Pl.

What are the functions of Ail and PagC proteins in the life cycle of Photorhabdus?

To answer this question, we used a well-established assay by expressing in E. coli ail and pagC homologs from the Photorhabdus strains. Such strategy was successfully used for characterizing the role of proteins from the Ail/PagC/OmpX family in Salmonella and Yersinia in human serum resistance and invasion/adherence to eukaryotic culture cells [4]. Recombinant Escherichia coli clones, expressing cosmids from Photorhabdus, were also used in assays to study gain of toxicity against insects, nematodes, amoeba, and mammalian macrophages [67] and to attribute biological function to several Photorhabdus potential virulence loci [68]–[70]. With this heterologous assay, we showed that none of the over-expressed genes displayed a role in adherence to the tested mammal (CHO) or insect (Sl2b) cells. Photorhabdus life cycle is mainly extracellular except a transient invasive stage during the symbiont transmission to the new generation of IJ before they exit the insect cadaver [30]. This transmission stage is dependent on the production of the bacterial Mad pili [71]. By contrast with Ail and PagC from Yersinia and Salmonella species [5]–[8], our results do not suggest a role of Ail1, Ail2 or PagC in cell invasion. While P. asymbiotica ATCC43949 is a clinical isolate from human wounds resistant to human serum at 30°C and 37°C [37], it is surprising that neither ail1_Pa nor pagC_Pa participated in the human serum resistance. It is tempting to speculate that other P. asymbiotica proteins could be involved in serum resistance. Finally, when expressed in E. coli, only ail2_Pl and pagC_Pl appear to play a role in human serum resistance whereas P. luminescens TT01 was not described as a clinical isolate. One hypothesis could be that Ail2_Pl and PagC_Pl could play a role in resistance toward components of the insect blood, the hemolymph. Thus, we tested the resistance of recombinant E. coli bacteria toward different AMPs, a key component of the insect humoral insect immunity, though without any success. In pathogenic Yersinia species, Ail proteins bind substrates such as the host cell extracellular matrix proteins, fibronectin and laminin, as well as the complement regulatory proteins C4bp and factor H [72]. Therefore, in insects, Ail2_Pl and PagC_Pl may interact with similar hemolymph components yet to be identified.

What is the evolutionary significance of the two Ail proteins in P. luminescens TT01, Ail1_Pl and Ail2_Pl?

In Y. pestis KIM, four genes encoding the proteins from the Ail/PagC/OmpX family were identified including one OmpX protein (y1682) and three Ail proteins (y2446, y2034 and y1324). These three Ail proteins are phylogenetically distant and their corresponding genes are not adjacent on the Y. pestis KIM genome [73]. In P. luminescens TT01, we propose that the two ail genes result from a recent tandem duplication (see above). The genomic redundancy in prokaryotes can be explained as a consequence of three selective processes, (i) elevated protein dosage (identical and duplicated genes), (ii) protein diversification (divergent paralogs) and (iii) adaptation to environmental variations (ecoparalogs of intermediate divergence) [74]. The intermediate aminoacids identity between Ail1_Pl and Ail2_Pl suggests a case of ecoparalogs. Three clues also argue in favour of this hypothesis. First, as already described with ecoparalogs predicted to be on the outer membrane or in the periplasmic space where the environment influence is important for protein function and stability [74], Ail1_Pl and Ail2_Pl differ by their isoelectric point values (respectively, 9.0 and 7.0). Second, a usual role identified in the Ail/PagC/OmpX family, the resistance to human serum, was only conserved for one protein, Ail2_Pl. Finally, regulation of the expression of ail1_Pl and ail2_Pl genes is obviously different and in the case of ail1_Pl, we showed that the influence of external environment fluctuation by the way of magnesium concentration is important. All together, the existence of these two ecoparalogs is highly suggestive of an adaptation to multiple niches (at least, insect and nematode) in response to external fluctuation.