Genetic and Proteomic Characterization of rpoB Mutations and Their Effect on Nematicidal Activity in Photorhabdus luminescens LN2

Rifampin resistant (RifR) mutants of the insect pathogenic bacterium Photorhabdus luminescens LN2 from entomopathogenic nematode Heterorhabditis indica LN2 were genetically and proteomically characterized. The RifR mutants showed typical phase one characters of Photorhabdus bacteria, and insecticidal activity against Galleria mellonella larvae, but surprisingly influenced their nematicidal activity against axenic infective juveniles (IJs) of H. bacteriophora H06, an incompatible nematode host. 13 out of 34 RifR mutants lost their nematicidal activity against H06 IJs but supported the reproduction of H06 nematodes. 7 nematicidal-producing and 7 non-nematicidal-producing RifR mutants were respectively selected for rpoB sequence analysis. rpoB mutations were found in all 14 RifR mutants. The rpoB (P564L) mutation was found in all 7 mutants which produced nematicidal activity against H06 nematodes, but not in the mutants which supported H06 nematode production. Allelic exchange assays confirmed that the Rif-resistance and the impact on nematicidal activity of LN2 bacteria were conferred by rpoB mutation(s). The non-nematicidal-producing RifR mutant was unable to colonize in the intestines of H06 IJs, but able to colonize in the intestines of its indigenous LN2 IJs. Proteomic analysis revealed different protein expression between wild-type strain and RifR mutants, or between nematicidal-producing and non nematicidal-producing mutants. At least 7 putative proteins including DsbA, HlpA, RhlE, RplC, NamB (a protein from T3SS), and 2 hypothetical proteins (similar to unknown protein YgdH and YggE of Escherichia coli respectively) were probably involved in the nematicidal activity of LN2 bacteria against H06 nematodes. This hypothesis was further confirmed by creating insertion-deletion mutants of three selected corresponding genes (the downregulated rhlE and namB, and upregualted dsbA). These results indicate that the rpoB mutations greatly influence the symbiotic association between the symbionts and their entomopathogenic nematode hosts.


Introduction
Rifampin (Rif), first introduced in 1972 as an antitubercular drug, was initially extremely effective against Mycobacterium tuberculosis, and other bacteria [1][2]. With its widespread and extended use, the number of bacterial isolates resistant to Rif has increased. Most Rif-resistance mutations in M. tuberculosis as well as in Escherichia coli and Staphylococcus aureus were conferred by a set of restrictive mutations in the rpoB gene, which encoded the bsubunit of RNA polymerase (RNAP) in bacteria [3][4]. DNAdependent RNAP, which contains an essential catalystic core enzyme (a 2 bb'v) and one of the sigma (d) factors, is the central enzyme for expression of genomic information in all organisms. Rif inhibits transcription initiation by blocking the rpoB of bacterial RNAP [5][6]. E. coli rpoB mutations that suppress the auxotrophy due to lack of stringent response were demonstrated to affect the transcription of stringently controlled genes by destabilizing the RNAP-stable RNA promoter complex [7]. The Rif resistant (Rif R ) M. tuberculosis mutations of the rpoB gene were found in nearly 95% of clinical isolates [4]. Most of the mutations were located from nucleotides 1276 to 1356 (codon 432-458 in M. tuberculosis rpoB gene and codon 507-533 in E. coli rpoB gene). An 81 bp core region was called the Rif resistance determining region (RRDR) of rpoB [8][9][10]. However, a significant number of Rif R mycobacteria with no mutations in the rpoB gene have been isolated from different clinical samples [11][12][13][14]. A 191A/C mutation in the Rv2629 gene was reported to be significantly associated with Rifresistance in M. tuberculosis [15]. Recently, it was reported that the K1 uptake regulator TrkA played an important role in intrinsic and acquired antibiotic resistance in mycobacteria [16]. Besides Rif-resistance, the rpoB mutation (A621E) conferred dual heteroresistance to daptomycin and vancomycin in Staphylococcus aureus [17], but most rpoB mutations were involved in reduced vancomycin susceptibility [18]. It suggested that different rpoB mutations may have different effect on bacteria.
Photorhabdus and Xenorhabdus bacteria belonging to the Enterobacteriaceae are symbiotically associated respectively with entomopathogenic Heterorhabditis and Steinernema nematodes, which are used as a commercial bioinsecticide for many economically important insect pests [19]. The association between the nematodes and their symbiotic bacteria plays an important role in the pathogenicity and production of these nematode-bacterium complexes. The infective juveniles (IJs) of these nematodes are a developmentally arrested non-feeding form, ensheathed in the second stage cuticle and harbor Photorhabdus or Xenorhabdus cells as symbionts in their intestines. The IJ nematodes properly maintain and carry the bacteria needed for killing insects and providing a suitable environment for the reproduction of new vectors [20][21][22]. Different Photorhabdus or Xenorhabdus bacterial isolates differ in their ability to support in vitro monoxenic cultures of non-host nematodes [23][24][25][26][27][28][29][30] and to retain the bacterial cells in the IJ intestines [22][23]26,28,31].
The trans-specific nematicidal activity of P. luminescens subsp. akhurstii LN2, a normal symbiont of H. indica LN2, against H. bacteriohora H06 nematodes was previously observed [40]. These bacteria secrete unidentified toxic factors lethal for H06 nematodes although the bacteria produce signals which trigger the recovery of H06 IJ nematodes [30]. A novel P. luminescens LN2 gene involved in the nematicidal activity against H. bacteriophora H06 IJs was identified [41].
Xenorhabdus and Photorhabdus bacterial isolates resistant to Rif were used in several references [42][43][44][45][46]. When different Rif R mutants of P. luminescens LN2 were monoxenically combined respectively with the axenic IJs of a Chinese isolate H. bacteriophora H06, the involvement of rpoB mutation in the nematicidal activity (incompatible symbiosis) was discovered.
To achieve an overall view of phenotypic, genetic and metabolic modifications associated with different Rif R mutants, the experiments were conducted to determine: (1) the effects of different Rif R mutants of P. luminescens LN2 on the growth of their corresponding incompatible nematode hosts, H. bacteriophora H06; (2) the phenotypic and biochemical characters of the Rif R mutants; (3) the rpoB mutations in the Rif R mutants; (4) the effects of rpoB mutations in the Rif R mutants on the nematode growth; (5) the mutualistic colonization of H06 IJs by the Rif R mutants; (6) the proteomic analysis of the mutants and wild-type bacterial strain; (7) the effects of differentially expressed proteins detected from proteomic analysis on nematicidal activity.

Nematode Species, Bacterial Strains, Plasmids and Culture Conditions
Bacterial strains and plasmids used in this study are listed in Table 1. P. luminescens subsp. akhurstii LN2 isolated from its host nematode H. indica LN2 was used for the isolation of spontaneous Rif R mutants. P. luminescens H06 or HNA were used for the mass production of H. bacteriophora H06. The bacterial strains were cultured in LB1 broth (1% tryptone, 0.5% yeast Extract, 0.5% NaCl) or on LB1 agar at 25uC. The primary form (phase one) of these bacteria was obtained by selecting green or blue-green colonies on NBTA or red colonies on MacConkey agar, and repeated subculturing [47]. Stock cultures were maintained in 15% glycerol (v/v) in LB1 at 280uC. E. coli strains were grown in LB2 broth (1% tryptone, 0.5% yeast Extract, 1% NaCl) or on LB2 agar at 37uC.
When required, antibiotics were added to the medium with the following concentrations: ampicillin (Amp), 100 mg/mL; kanamycin (Km), 50 mg/mL; rifampin (Rif), 50 mg/mL; and tetracycline (Tc), 25 mg/mL; chloramphenicol (Cm), 25 mg/mL. All the antibiotics used in this study were purchased from Sigma Chemical Company and all medium components from Oxoid Company, England.

Production of Axenic Heterorhabditis IJs
Axenic H. bacteriophora H06 IJs for the monoxenic nematodebacterium recombinations were obtained according to the method as previously described [30]. Briefly, IJs of H06 were grown monoxenically on nonspecific P. luminescens HNA on a sponge medium consisting of 1% yeast extract, 5% egg yolk, 15% soya flour, 5% corn oil, 8% polyether polyurethane sponge and 50% distilled water [30]. The IJs were collected by centrifugation and migration through a 30 mm nylon cloth sieve under sterile conditions, surface-sterilized in 0.5% streptomycin-sulfate (Merck, Germany) for 6 h and then rinsed three times in sterile distilled water. The axenicity of these surface-sterilized IJs was checked as previously described [30]. Because these IJs can be reared with the provided bacterial isolates, and are not able to contain the bacteria in their intestines, they are free of bacteria after surface sterilization.

Nematicidal Bioassay of the Rif R Mutants
The Rif R mutants from LB1 agar with Rif were screened for their nematicidal activities against H. bacteriophora H06 IJs according to the method as previously described [30]. Approximately 100 axenic H06 IJs were introduced to the 2-day old lawn of wild-type strain or Rif R mutants of P. luminescens LN2 grown on LB1 agar in 96-well tissue culture plate (Corning, New York, USA). Mortality and growth of the IJs were observed daily and recorded until 15 days. A lawn of wild-type P. luminescens H06 was used as a control. A mutant was considered positive for nematode growth if the tested nematodes were able to survive at least 7 days and produce the next generation of juveniles from the hermaphrodites. If the mutants were unable to support the survival of nematodes, the introduced IJs died after 7 days. The effect of the Rif R mutants on H06 IJs were verified by repeating the nematode survival and growth experiments three times, each with 12 replicates. Among the 34 tested Rif R mutants, 13 mutants were identified positive, and 21 mutants negative for the growth of H. bacteriophora H06. 7 positive (LN2-R2, LN2-R6, LN2-R12, LN2-R15, LN2-R28, LN2-R31, LN2-R33) and 7 negative mutants  (LN2-R3, LN2-R5, LN2-R7, LN2-R8, LN2-R11, LN2-R16,  LN2-R25) were selected for further study (Table 2).

Colonial Characterization of the Mutants and Wild-type Bacterial Strain
Colony pigmentation was determined on LB1 agar, NBTA, and MacConkey agar plates. pH-sensitive pigment production in LB1 was determined by addition of 1 M NaOH or 1 M HCl. Tests for the production of antibiotic substances were conducted as previously described [48], using Bacillus subtilis as test organism, and scored positive when a growth inhibition zone of .3 mm was measured around the P. luminescens colonies at 96 h after inoculation of the overlay culture. Bioluminescence was observed by dark-adapted eyes in a dark room. Cell morphology was observed microscopically. Catalase activity was tested by introducing 0.1 ml 3% hydrogen peroxide into the bacterial cultures and observing the release of oxygen. For all assays, both wide-type and mutant colonies were characterized on the same agar plate ad positive and negative controls. At least three plates for each medium were established.

DNA Manipulation
Plasmid DNA preparation, extraction of genomic DNA, restriction enzyme digestions, and ligations were carried out as previously described [49]. Restriction enzymes (Promega, USA) and T4 ligase (Novagen, Germany) were used according to the manufacturer's instructions. Plasmids were extracted from E. coli with QIAprep Spin Miniprep kit (Qiagen, Netherlands). When required, DNA fragments were extracted and purified from agarose gels using E.Z.N.A. TM Gel Extractio kit (Omega, USA). The genomic DNA was isolated from P. luminescens bacteria using E.Z.N.A. TM Bacterial DNA Kit (Omega).

Mutation Analysis of the rpoB Gene from Different Strains
To examine the rpoB sequence from the Rif R mutants and wildtype strain (LN2-W), together with a spontaneous Amp R mutant LN2-A and the namA mutant LN2-M1 [41], the gene was amplified from the genomic DNA of different strains, by PCR with PfuUltra TM II Fusion HS DNA Polymerase (Stratagene, Germany), using the primers rpoB-BamHI-F (59-GCTGGATC-CATGGTTTACTCCTATACCGAG-39) and rpoB-PstI-R (59-GCACTGCAGTTATTCGTCTTCCAGCTCGATG-39). The amplified gene was cloned into pCR4-TOPO vector (Invitrogen, USA), and transformed into E. coli TOP10 (Invitrogen). DNA sequencing was performed by Invitrogen Trading (Shanghai) Co. Ltd. All strains were sequenced at least twice. The sequence data of the rpoB gene were assembled and analyzed with DNAstar and CLUSTAL W program. The rpoB sequence data from wild-type strain of P. luminescens LN2 has been submitted to the GenBank database under accession number (JN177303).

Allelic Exchange Mutagenesis of the rpoB Gene
The plasmids of pCR4-TOPO-rpoB-LW and pCR4-TOPO-rpoB-LR31 containing corresponding rpoB genes from LN2-W and LN2-R31, were digested with BamHI (GGATCC) and PstI (CTGCAG), respectively. The resulting rpoB fragments were purified and ligated into the suicide vector pPHU281 [50] digested with BamHI and PstI to yield plasmids pPHU281-rpoB-LW and pPHU281-rpoB-LR31. The resulting plasmids were transferred into E. coli S17-1 (lpir) [51]. Strains LN2-WDrpoB-LR31 (LN2-W containing the mutant rpoB allele from LN2-R31) and LN2-R31DrpoB-LW (LN2-R31containing the wild type rpoB allele from LN2-W) were created by allelic exchange with pPHU281-rpoB-LR31 and pPHU281-rpoB-LW, respectively, using biparental mating method. Rif R .Amp R .Tc S exconjugants of LN2-WDrpoB-LR31 and Rif S .Amp R .Tc S exconjugants of LN2-R31DrpoB-LW were selected on LB1 agar plates with appropriate antibiotics. The exconjugants had undergone allelic exchange and lost the wild-type or mutated copy of rpoB and the plasmid vehicle. The mutants were verified for the presence of the appropriate rpoB allele by sequencing rpoB gene as described above. The resulting confirmed allelic exchange mutants were determined for the nematicidal activities against the IJs of H06 as described above.

Insecticidal Injection Assays
To check the insecticidal activity of the Rif R mutants, the wildtype strain and Rif R mutants of P. luminescens LN2 were grown overnight in LB1 broth without antibiotics, subcultured into fresh LB1 broth with 1% of bacterial culture, and incubated at 25uC for 24 h prior to injection. These cultures were washed and diluted to concentrations of 10, 100, 1000 CFU/mL in sterile phosphate buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM NaHPO 4 , 2 mM KH 2 PO 4, pH 7.4). Last instar larvae of greater wax moth Galleria mellonella were incubated on ice for approximately 5 min. 10 mL of the diluted cultures or sterile PBS were injected into the first proleg of each of 10 insect larvae using a 30gauge syringe (Hamilton, Reno, NV). Three replicates with 10 insect larvae per replicate were established. Insects were monitored every 6 h for 120 h post injection. Dead insects were observed to confirm the presence of red color and bioluminescence.

Colonizations of H06 IJs by the GFP-labelled Mutants
To observe the colonization of IJ nematodes by the Rif R mutant bacteria, the Rif R mutant LN2-R31 positive for the growth of H06 nematodes was labeled with GFP by transposon mutagenesis of a pMini-Tn5 [52] containing an expressed egfp gene (pMini-lac-egfp). The pMini-lac-egfp was constructed as follows. A fragment containing an egfp gene and ribosome binding site (GAAGGTTTA-GAC) was obtained from pUC19-egfp with primers egfp-SD2 (59-GAAGGTTTAGACATGGGCAAAGGAGA-39) and egfp-rev (59-TAGCGGCCGCTTATTTGTATAGTTCATC-39) (NotI). The amplified 750 bp PCR product was cloned into the pMD-18T vector (TaKaRa, Japan) and transformed into E. coli DH5a. Green clone on LB2 plates with ampicillin was selected to extract the plasmid pMDegfp. A NotI-NotI fragment containing the lac promoter and egfp gene from pMD-egfp was cloned into pMD-19T Simple vector (TaKaRa) to generate pMD-lac-egfp, with primers lac-F (59-AGCGGCCGC-GAGCGCAGCGAGTCAGTGAGC-3) (NotI) and egfp-rev (NotI ). After transformed into E. coli DH5a, clones (Amp R ) expressing GFP were detected using epifluorescence microscope (Nikon Eclipse 80i). To construct a transposon delivery vector pMini-lac-egfp, the NotI-NotI fragment carrying ribosome binding site, lacZ promoter and egfp gene from pMD-lac-egfp was inserted into the NotI site of pMini-Tn5. The ligation product was transformed into E. coli S17-1 (lpir). Clones Table 2. rpoB mutations of the Rif R mutants and their effect on H06 nematode growth.

Bacterial mutants
Nucleotide change Amino acid change Effect on the growth of H06 nematodes (Km R ) with green fluorescence were used to deliver the egfp gene into the chromosome of the Rif R mutant LN2-R31 by diparent conjugation. Conjugants (Km R .Rif R ) were selected on LB1 plates at 25uC. GFP-labeled LN2-R31 was observed to express stable green fluorescence, even in the absence of antibiotic selection.
To check the colonization of H06 IJs by the Rif R mutant LN2-R31, the nematodes were cultured on sponge medium [30] inoculated with GFP-labeled LN2-R31 as described above respectively. The IJs were extracted from the sponge and observed for GFP-labeled bacteria. The IJs were also homogenized with a sterile glass homogenizer after surface sterilization with 0.1% merthiolate and 5-time rinse with sterile distilled water. The presence of the GFP-labeled bacteria retained in the IJs intestines was determined by plating dilutions of surface-sterilized and homogenized nematodes on LB1 agar plates.

2-DE Analysis and Protein Identification by MALDI-TOF-MS
The 48 h old bacterial cells of the wild-type strain and Rif R mutants (one negative mutant LN2-R16 and three positive mutants LN2-R2, LN2-R31 and LN2-R33, for H06 growth) grown on LB1 plates at 25uC were harvested and washed three times with cold PBS by centrifugation (6000 g, 10 min, 4uC). The cell pellets were resuspended in lysis buffer (8 M urea, 0.2% w/v Bio-Lyte 3/10 Ampholyte (Bio-Rad, USA), 4% CHAPS, 65 mM DTT) containing Protease Inhibitor Cocktail (Calbiochem, Germany) and Benzonase (Novagen, Germany) and disrupted by liquid nitrogen freezing-thawing three times. Cell debris was removed by centrifugation (20000 g, 60 min, 4uC). The supernatant (total cell protein) was divided into aliquots and stored at 280uC until use. Protein concentrations were determined by the Bradford method using Modified Bradford Protein Assay Kit (Sangon, China).
The 2-DE was performed according to the methods described previously [53] and the manufacturer's instruction. The first dimension (isoelectric focusing) was conducted using the IPGphor IEF system (Bio-Rad) at 20uC. For analytical gels, 350 mg protein was solubilized in 400 mL rehydration solution (8M urea, 0.2% w/ v Bio-Lyte 3/10 Ampholyte, 4% CHAPS, 65 mM DTT, 0.001% w/v bromophenol blue), and loaded onto a 17 cm pH 3-10 NL IPG strip (Bio-Rad). Focusing was performed for 13 h at 50V, 1 h at 500 V, 1 h at 1000 V, and 5 h and 30 min at 8000 V (total = 45 kVh). The IPG strips were equilibrated as previously described [53]. The second dimension was performed with 12% (w/v) SDS-polyacrylamide gels using the Protean II xi 2D Multicell system (Bio-Rad). Proteins were stained with silver nitrate, and gels were digitized using Image ScannerII (Amersham Biosciences). Digitized 2-DE gel patterns were edited and matched using the PDQUEST software package (PDI, Humington Station). Triplicate experiments were run to confirm the reproducibility of results.
Spots of interest in gels staining with silver nitrate were cut out, washed, reduced, S-alkylated with iodoacetamide and in-gel digested at 37uC overnight with sequencing grade porcine trypsin (Promega, USA). After extraction in extractant of 50% ACN (Fisher) and 2.5% TFA (Sigma), peptide mixtures were analyzed using a saturated solution of 5 mg/mL a-cyano-4-hydroxycinnamic acid (CHCA, Sigma) in ACN containing 0.1% TFA (Sigma) (50/50 v/v) using a 4800 Proteomics Analyzer equipped with matrix assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) (Applied Biosystems, Framingham, MA, USA). For MS calibration, the trypsin autolysis peptides were used as internal calibrants. Monoisotopic peak masses were automatically determined within the mass range of 800-4000 Da, with a minimum S/N of 50. Five of the most intense ion signals were selected as precursors for MS/MS acquisition. Combined MS and MS/MS queries were performed with the MASCOT search engine (V2.1, Matrix Science, UK) embedded in GPS-Explorer Software (V3.6, Applied Biosystems), using the P. luminescens database (Gene DB). MASCOT protein scores (based on combined MS and MS/MS spectra) of greater than 61 were considered statistically significant (p#0.05). The individual MS/ MS spectrum with statistically significant (confidence interval .95%) best ion score (based on MS/MS spectra) were also accepted.

Insertion-deletion Mutations of the Corresponding Genes from Differentially Expressed Proteins
Compared with the nematicidal-producing mutant LN2-R16 and wild type strain, four proteins including RplC, RhlE, NamB (a putative transport and binding protein from type III secretion system), and a hypothetical protein similar to unknown protein YggE of E. coli were downregulated; three proteins including DsbA, HlpA, and a hypothetical protein highly similar to unknown protein YgdH of E. coli were upregulated in the nonnematicidal-producing mutants (LN2-R2, LN2-R31 and LN2-R33) ( Table 3-4, Figure S1, S2, S3, S4, S5). At least these 7 putative proteins were probably involved in the nematicidal activity of LN2 bacteria against H06 nematodes.
To confirm these results, the downregulated rhlE and namB, and upregualted dsbA (GenBank accession number JX274431, JX274430 and JX274432 respectively) from the differentially expressed proteins were selected for construction of insertiondeletion mutations to determine the effects of the knock-out genes on the nematicial activity. Three P. luminescens LN2 mutants termed as LN2DrhlE, LN2DnamB and LN2DdsbA were created. Insertion-deletion mutations in these three genes were constructed using fusion PCR strategy as previously described [54]. For each gene, three fragments F1 (the upstream of the target gene), camR (Chloramphenicol cassette) and F2 (the downstream of the target gene) were generated using primer pairs of P1 and P2, P3 and P4, and P5 and P6 (Table S1), respectively. The camR gene was amplified from the plasmid pSZ21 [55] and the F1 and F2 gene fragments were amplified from P. luminescens LN2 genomic DNA. Approximately equal amounts of the three purified fragments F1, camR and F2 were mixed, and used as a template to generate a new DNA fragment by a second PCR performed with the primers P1 and P6. Three resulting fragments, which corresponded to rhlE::Cm, namB::Cm and dsbA::Cm, respectively, were separately cloned into a pMD-19T Simple vector. Then the resulting plasmids of pMD-rhlE::Cm, pMD-namB::Cm and pMD-dsbA::Cm were separately ligated to the same enzyme digested suicide vector pKNG101 [56] to generate pKNG101-rhlE::Cm, pKNG101-namB::Cm and pKNG101-dsbA::Cm. P. luminescens LN2 mutants termed as LN2DrhlE, LN2DnamB and LN2DdsbA were created by allelic exchange with pKNG101-rhlE::Cm, pKNG101-namB::Cm and pKNG101-dsbA::Cm, respectively, as previously described [54]. The phenotypic characterization, rpoB sequence and effects on nematicidal activity of three resulting mutants were determined as described above.

Results
Isolation and Characterization of the Rif R Mutants of P. luminescens LN2 Several hundreds of the Rif R mutants of P. luminescens LN2 were isolated and 34 mutants were randomly selected for further study. The wild type strain and the selected mutants showed the typical Table 3. Total proteins with altered level of synthesis in the nematicidal-producing and non nematicidal-producing mutants.

Spot
No characteristics of phase one bacteria as described: uptake of dye from NBTA and MacConkey agar, production of pH-sensitive pigments, occurrence of inclusion bodies, antibiotic activity, and bioluminescence.
The Effects of the Rif R Mutants on the Growth of H06 Nematodes 13 of 34 Rif R P. luminescens LN2 mutants were able to support the growth of H06 IJs, with hermaphrodites containing living juveniles inside and outside after 12 days on the agar plates, while 21 of them were negative for the growth of H06 nematodes. On the bacterial lawns with those mutants or the wild-type, which did not support the nematode production, all the nematodes did not grow beyond adults and died after 7 days.
The Mutation Loci of rpoB Gene in the Rif R Mutants 7 positive mutants (LN2-R2, LN2-R6, LN2-R12, LN2-R15, LN2-R28, LN2-R31 and LN2-R33) and 7 negative mutants (LN2-R3, LN2-R5, LN2-R7, LN2-R8, LN2-R11, LN2-R16 and LN2-R25) for H06 nematode growth ( Table 2) were randomly selected for rpoB gene sequencing. The entire rpoB sequences of 14 selected Rif R mutants, wild type strain, LN2-A and LN2-M1 were sequenced at least twice. The rpoB gene from all colonies was 4029 bp in length, the same to that of P. luminescens subsp. laumondii TT01 [57]. The identity of rpoB genes between the wild-type strain of LN2 and TT01 was 96.13%. All of the 14 Rif R mutants carried mutations in the rpoB gene. 10 mutants showed a single nucleotide mutation resulting in an amino acid substitution, and 2 mutants presented two nucleotide mutations resulting in two amino acid substitutions, but only one mutant displayed three nucleotide mutations resulting in two amino acid substitutions ( Table 2). No mutation was observed in the rpoB gene of Amp R mutant LN2-A and namA mutant LN2-M1.
The rpoB (P564L) mutation was found in all 7 mutants which produced nematicidal activity against H06 nematodes, but not in the mutants which supported H06 nematode production. While the single mutations of V146F, S512F, Q513K and double mutations of A313D and R529C were detected respectively in the mutants which supported H06 nematode production ( Table 2). The single and double mutations resulted in loss of nematicidal activity against H06 nematodes and ability to supported H06 nematode production.

Allelic Exchange Assays
The recombinant LN2-WDrpoB-LR31 (LN2-W strain containing the mutant rpoB allele from LN2-R31) and the LN2-R31DrpoB-LW (LN2-R31 containing the wild type rpoB allele from LN2-W) were selected on Am R Rif R Tc s and Am R Rif S Tc s LB1 agar plates, respectively.
Successful homologous recombination of rpoB gene in the recipient strains was verified by randomly selecting three colonies from each recipient and checking their rpoB gene sequences by PCR and DNA sequencing. The sequences of rpoB gene from three colonies of each recipient were 100% identical.
The recombinant LN2-WDrpoB-LR31 showed Rif resistance and lost the nematicidal activity against H06 IJs, while LN2-R31DrpoB-LW was sensitive to Rif and restored the nematicidal activity. These results clearly indicated that rpoB mutation was responsible for the Rif-resistance and the absence of nematicidal activity of LN2-R31.

Insecticidal Activity
The Rif R mutants, including nematicidal-producing LN2-R16 and non nematicidal-producing LN2-R2, LN2-R31 and LN2-R33, together with wild-type of P. luminescens LN2 caused 100% mortality of G. mellonella at the concentrations of 1000 CFU/mL after 24 h, 100 CFU/mL after 30 h, and 10 CFU/mL after 36 h. No insect mortality was recorded in the control after 120 h. It appeared that the mutant bacteria also displayed insecticidal activity against G. mellonella larvae.

IJ Colonization of the GFP-labelled Mutants
No H06 IJs from the culture with GFP-labeled LN2-R31 mutant contained GFP-labeled bacteria in their intestines. No GFP-labeled bacteria were also observed from the mechanically disrupted H06 IJs. However, the IJs of H. indica LN2 from GFPlabeled LN2-R31 mutant contained GFP-labeled bacteria in their intestines. Bacterial colonization of the intestines of IJs is an important process in the nematode-bacterium symbiosis. The present result demonstrated that the mutation of rpoB gene restored the nutrient suitability of the LN2 bacteria for the reproduction of H06 nematodes by silencing the nematicidal activity of the bacteria, but did not establish the environment for bacterial colonization of the IJs.

The Proteomic Analysis of the Mutants and Wild-type
The effects of rpoB mutations on the nematicidal activity of LN2 bacteria were further investigated by identifying the differentially expressed proteins by 2-DE. The parental and selected rpoB mutant strains grown on LB1 agar plates were collected after 48 h. Cells were disrupted and whole cell proteins were separated on 2-DE gels spanning the pH 3-10, silver stained, and analyzed by MS. Protein levels were expressed as percentage volume, which corresponds to the percentage ratio between the volume of a single spot and the total volume of all spots present in a gel. The mean values of spot intensity were calculated using at least three gels. Spots showing more than 15% variation were not considered (Student's test, with 7 degrees of freedom, p,0.05). Little deviation was observed in the patterns on replica gels.
The spots with intensity changes by a factor of at least two were selected for MALDI-TOF-MS analysis (Table 3), using NCBI website and the PhotoList database (http://genolist.pasteur.fr/ PhotoList/).
Proteomic analysis revealed major difference between wild-type strain and Rif R mutants, and between nematicidal-producing and non nematicidal-producing mutants. In all the analyzed rpoB mutants, 15 putative proteins (YbdQ, Hcp, GlyA, UreG, ViuB, FabH, Ndk, Upp, Kst, MprA, DsbA, GpmA, AgaV, one bacteriophage protein and one unknown protein) were upregulated, and 11 (AhpC, CipA, cipB, HcaB, ISPPlu3Y, ISPlu10J, YqjD, TatA, DksA, FliC, and three hypothetical proteins) were downregulated. In particular, the following putative proteins were not detected from all the analyzed rpoB mutants: MacA (probable macrolide-specific ABC transporter, spot 13); DegQ (protease precursor, spot 37, 38); and KatE (catalase, spot 40, 41). Interestingly, an unknown function protein Brca2 (similar to breast cancer type 2 susceptibility protein, spot 21) was not present in the mutants, but present in the wild type strain. It appeared that the absence of these proteins was due to the rpoB mutation rather than the antibiotic pressure, because they were absent also from a namA disruption mutant of LN2 [41] without rpoB mutation in the culture without any rifampin (unpublished data).

Genetic Confirmation of Differentially Expressed Proteins
LN2DrhlE, LN2DnamB and LN2DdsbA mutants showed the typical characteristics of phase one bacteria as the wild type strain. No mutation in rpoB gene was observed in these mutants. LN2DrhlE and LN2DnamB mutants were able to support the growth of H06 IJs, with hermaphrodites containing living juveniles inside and outside after 12 days on the agar plates, while LN2DdsbA mutant was negative for the growth of H06 nematodes as the wild type strain. The results confirmed the involvement of these selected genes in the nematicidal activity against H06 nematodes.

Discussion
In this study, a similar mechanism determining Rif-resistance in E. coli and M. tuberculosis [3,11,58] was verified in Photorhabdus bacteria. Surprisingly, the Rif R mutants influenced the nematicidal activity of P. luminescens LN2 bacteria against a different nematode, H. bacteriophora H06. Furthermore, some but not all rpoB mutants of LN2 bacteria lost nematicidal activity against H06 IJs. The rpoB mutation was demonstrated to be responsible for the Rif-resistance and the effect on the nematicidal activity in the Rif R mutants of LN2. There are fundamental connections between rifampin resistance, RNA polymerase structure and function and global gene expression in the literatures. Rif mutations in E. coli affected a wide variety of phenotypes, including altered growth properties and stimulated secondary metabolism [59]. A novel rpoB mutation in B. subtilis showed a unique spectrum of effects on growth and various developmental events [60]. An rpoB mutation in Streptomyces lividans activated antibiotic production and reduced growth rate [61]. A spontaneous Rif R mutation isolated from Saccharopolyspora erythraea stimulated bacterial secondary metabolism and was severely impaired in erythromycin production [62]. To the best of our knowledge, this is the first report that rpoB mutations influenced the nematicidal activitity of a nematode symbiont on a non-cognate nematode partner. The symbiosis between the entomopathogenic nematodes and their associated bacteria will be also influenced by the rpoB mutations. However, how rpoB mutations affect this nematicidal activity needs to be further explored.
The mutants exhibited several phenotypes of phase one variant as previously described [47], e.g. absorption of the dye from NBTA and MacConkey agar, production of bioluminescence and occurrence of crystalline inclusion proteins in the cells. It was reported that the nematicidal activity occurred only in phase one of P. luminescens LN2 [31]. Apparently, the loss of nematicidal activity in the LN2 mutants against H06 nematodes was not the result of a typical phase variation. The physiological status of symbiotic Photorhabdus and Xenorhabdus bacteria (such as phase variation, mutants) may influence their fitness for nematode production. As rpoB mutations were associated with the nematode growth, screening of rpoB mutants of symbiotic bacteria of entomopathogenic nematodes may provide a way to select beneficial rpoB mutants by Rif for effective mass production of the nematodes.
One of the important characters in Photorhabdus bacteria is their insecticidal activities towards different insects [32][33]35]. The present result indicated that the rpoB mutations did not change the expression of the toxin genes, at least in the tested mutants, for the mutants also displayed the insecticidal activity against G. mellonella larvae.
Different rpoB mutations were associated with their ability to support H06 nematode production. However, the P564L mutation was not associated with the loss of nematicidal activity. The reasons why the mutations affect the physiology and metabolism of the bacterial mutants are not known. The RNA polymerase complex may contact every promoter in the genome, thus any change in critical portions of the enzyme can lead to global changes in gene transcription. Mutations within the Rif binding pocket of rpoB gene may alter the structure of RNA polymerase and hence its regulated interaction with specific promoters, and hence physiology and metabolism [63].
Proteomic analysis revealed at least 7 putative proteins including DsbA, RhlE, NamB (a protein from T3SS), HlpA, RplC and 2 hypothetical proteins YggE and YgdH might be involved in the nematicidal activity. All these proteins may play different roles in different organisms (64)(65)(66)(67)(68)(69)(70)(71). In the present study, it was hard to establish the functional relationship among these proteins in the nematicidal activity of LN2 bacteria. However, the insertion-deletion method confirmed the involvement of the selected corresponding genes (such as rhlE, namB and dsbA) from the differentially expressed proteins in the nematicidal activity against H06 nematodes. It seems that a big network system is involved in this nematicidal acitivity. Further work is needed to explore this system to understand the molecular mechanism on the trans-specific nematicidal activity of incompatible symbionts. Figure S1 2-DE map of total cell proteins from P. luminescens LN2 wild type strain. A representative gel shows the identified differentially expressed protein spots. 350 mg of total cell proteins was loaded onto a 17 cm pH 3-10 NL IPG strip, separated in the second dimension by SDS-polyacrylamide gel electrophoresis on a 12% gel and stained with silver nitrate. (TIF) Figure S2 2-DE map of total cell proteins from P. luminescens LN2 Rif R mutant LN2-R2. A representative gel shows the identified differentially expressed protein spots. 350 mg of total cell proteins was loaded onto a 17 cm pH 3-10 NL IPG strip, separated in the second dimension by SDS-polyacrylamide gel electrophoresis on a 12% gel and stained with silver nitrate. (TIF) Figure S3 2-DE map of total cell proteins from P. luminescens LN2 Rif R mutant LN2-R16. A representative gel shows the identified differentially expressed protein spots. 350 mg of total cell proteins was loaded onto a 17 cm pH 3-10 NL IPG strip, separated in the second dimension by SDS-polyacrylamide gel electrophoresis on a 12% gel and stained with silver nitrate. (TIF) Figure S4 2-DE map of total cell proteins from P. luminescens LN2 Rif R mutant LN2-R31. A representative gel shows the identified differentially expressed protein spots. 350 mg of total cell proteins was loaded onto a 17 cm pH 3-10 NL IPG strip, separated in the second dimension by SDS-polyacrylamide gel electrophoresis on a 12% gel and stained with silver nitrate. (TIF) Figure S5 2-DE map of total cell proteins from P. luminescens LN2 Rif R mutant LN2-R33. A representative gel shows the identified differentially expressed protein spots. 350 mg of total cell proteins was loaded onto a 17 cm pH 3-10 NL IPG strip, separated in the second dimension by SDS-polyacrylamide gel electrophoresis on a 12% gel and stained with silver nitrate. (TIF)