The indigenous Pseudomonas plasmid pQBR103 encodes plant-inducible genes, including three putative helicases

Abstract

Plasmid pQBR103 (∼400 kb) is representative of many self-transmissible, mercury resistant plasmids observed in the Pseudomonas community colonising the phytosphere of sugar beet. A promoter trapping strategy (IVET) was employed to identify pQBR103 genes showing elevated levels of expression on plant surfaces. Thirty-seven different plant-inducible gene fusions were isolated that were silent in laboratory media, but active in the plant environment. Three of the fusions were to DNA sequences whose protein products show significant homology to DNA-unwinding helicases. The three helicase-like genes, designated helA, helB and helC, are restricted to a defined group of related Pseudomonas plasmids. They are induced in both the root and shoot environments of sugar beet seedlings. Sequence analysis of the three plasmid-encoded helicase-like genes shows that they are phylogenetically distinct and likely to have independent evolutionary histories. The helA gene is predicted to encode a protein of 1121 amino acids, containing conserved domains found in the ultraviolet (UV) resistance helicase, UvrD. A helA knockout mutant was constructed and no phenotypic changes were found with plasmid-conferred UV resistance or plasmid conjugation. The other 34 fusions are unique with no homologues in the public gene databases, including the Pseudomonas genomes. These data demonstrate the presence of plant responsive genes in plasmid DNA comprising a component of the genomes of plant-associated bacteria.

Introduction

Plasmids are commonly isolated from phytosphere-colonising bacteria, but functional studies are limited [1–4]. Most analyses have been confined to genes that encode clear phenotypes, for example, genes involved in bioremediation, symbiosis or virulence, and genes encoding resistance to antibiotics, heavy metal ions or UV radiation[4]. In order to increase understanding of plasmid-encoded traits, it is necessary to create ways of studying genes that do not express a strong phenotype in the laboratory environment, but nevertheless may contribute toward ecological performance in the wild.

One way to progress is to identify genes of interest based on their level of expression in the wild compared to their level of expression in the laboratory environment: genes showing higher expression in the wild, being “niche specific genes”, are more likely to contribute toward ecological performance in nature than genes expressed equally across a range of environments – including the laboratory. Such a predication was recently confirmed for a locus encoding an acetylated cellulose polymer [5,6].

In two previous studies [5,7]Pseudomonas fluorescens genes, showing elevated levels of expression in the rhizosphere of sugar beet, were identified by a simple promoter trapping strategy (IVET, in vivo expression technology) [8,9]. The genes were identified on the basis of their ability to drive the expression of either a promoterless copy of panB, a gene encoding ketopantoate hydroxymethyltransferase (required for the biosynthesis of pantothenate)[7], or a promoterless copy of dapB [5], a gene encoding 2,3-dihydrodipicolinate reductase (required for the biosynthesis of diaminopimelate (DAP) and lysine): critically, both pantothenate and DAP are severely limiting in the plant environment.

Here we describe the application of a DAP-based IVET strategy to the functional analysis of a large indigenous plasmid, pQBR103. Plasmid pQBR103 is typical of a group of plasmids found naturally in a Pseudomonas population of field growing sugar beet[10]. It was exogenously isolated by its transfer proficiency and its ability to confer resistance to mercury chloride on the host bacteria, but otherwise it is cryptic. It has been shown in field release and plant pot experiments that carriage of pQBR103 initially reduces the fitness of P. fluorescens SBW25 in the early stages of plant growth, but enhances the competitive advantage of the host bacterium as the plant matures [11,12]. Consistently, isolation of natural plasmid transconjugants from the phytosphere of sugar beet in the field was limited to a mid- to late-season period[13]. These data suggested that the plasmid carries fitness-associated genes responding to plant-derived signals and/or seasonal factors. Here, using the IVET strategy, we demonstrate the presence of plant-inducible genes on plasmid pQBR103 and report the analysis of three predicted helicases.

Materials and methods

Bacterial strains and growth conditions

A summary of bacterial strains and plasmids used in this study is provided in Table 1. The ancestral strain of P. fluorescens SBW25 is a plasmid-free, non-pathogenic, rRNA group 1 fluorescent Pseudomonas, isolated from field-grown sugar beets at the University of Oxford farm, Wytham, Oxford, in 1989 [14,15]. Plasmid pQBR103 was acquired by genetically marked P. fluorescens SBW25EeZY6KX after it was released to sugar beet grown at the same field site[13]. Mercury chloride (0.1 mM) was used to select P. fluorescens strains carrying the environmental plasmids, which are prefixed with pQBR and listed in Table 2.

Escherichia coli strains were grown in Luria–Bertani (LB) medium at 37 °C. P. fluorescens was grown in LB or minimal (M9) medium[16] at 28 °C. Lysine (80 g ml⁻¹) and DAP (diaminopimelate, 800 μg ml⁻¹) were supplemented as required to cultivate P. fluorescens SBW25ΔdapB. Expression of the lacZY genes was observed in agar plates containing 45 μg ml⁻¹ of 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal). Antibiotics were used at the following concentrations (μg ml⁻¹): tetracycline (Tc), 10; ampicillin (Ap), 100; kanamycin (Km), 50; streptomycin (Sm), 100. CFC supplement (Cetrimide, Fucidin and Cephalosporin) from Oxoid (UK) was used at half strength to select P. fluorescens recovered from the plant.

DNA manipulations

Recombinant DNA techniques were performed using standard protocols[16]. DNA was transformed into E. coli by electroporation, using a Bio-Rad GenePulser according to the manufacturer's protocol (Bio-Rad Laboratories, Hertford, UK). Restriction enzymes were obtained from Gibco-BRL (UK) and New England Biolabs Inc. (NEB; Hitchin, UK). PCR amplifications were performed using Qiagen Taq DNA polymerase with an annealing temperature of 56 °C. Purified genomic DNA or whole cells were used as template DNA. Oligonucleotide primers were obtained from MWG Biotech (Ebersberg, Germany) and sequences are listed in Table 1. DNA sequencing, using Big-Dye Terminators (Applied Biosystems, Warrington, UK), was performed following the manufacturer's recommendation and separated on an automated DNA sequencer, model 310 (Perkin–Elmer, Warrington, UK). Plasmid conjugation was carried out using the filter plate mating method described by Lilley et al.[10], modified to provide a growth temperature of 28 °C.

Construction of the IVET libraries in pIVETD

The plasmid DNA of pQBR103 was prepared and purified from 50 ml of an overnight culture of P. fluorescens SBW25 (pQBR103) using the sodium dodecyl sulphate-lysis and sucrose gradient method of Wheatcroft and Williams[17]. Plasmid DNA was directly digested by Bgl II/Bam HI (New England Biolabs) and ligated to Bgl II-digested and alkaline phosphatase-treated pIVETD. The ligation mix was transformed into E. coli DH5αλpir. Transformants were selected on LB plates containing tetracycline and about 600 colonies were mixed as a non-random IVET library of pQBR103. In order to construct a random IVET library, 1–3 kb DNA fragments of pQBR103 were generated using a general sonication treatment[16] and ligated to the Sma I site of pUC18 after blunting the ends by standard Klenow Fragment treatment. The ligation mix was transformed into DH5α and the Lac⁻ transformants were grouped in pools of ∼200 colonies. Plasmid DNA was prepared from each pool (five in total) and then digested by Bam HI completely. The digestion mix was ligated to Bgl II-digested and alkaline phosphatase-treated pIVETD. After electroporation, about 600 E. coli colonies were obtained for each group of DNA (totally five) and they were mixed as IVET fusion plasmid libraries. Contaminant pUC18 vector DNA was removed in the following steps of homogenous recombination into P. fluorescens SBW25ΔdapB (pQBR103) and plant selection.

The genomic libraries of P. fluorescens SBW25 (pQBR103) were constructed as described in detail in Rainey[7] and Gal et al.[5], except that the DNA was obtained from an isolate of SBW25 carrying pQBR103. Briefly, genomic DNA was partially digested with Sau 3AI and separated by agarose gel electrophoresis. Fragments of 3–5 kb in size were extracted using the Qiagen Gel Extraction Kit (Qiagen, UK) and cloned into pIVETD. The fusion library was stored as 10 individual pools with a mixture of ∼600 E. coli transformant colonies in each.

The IVET library pools in E. coli were transferred into P. fluorescens SBW25ΔdapB (pQBR103) by conjugation with the help of the mobilizing plasmid pRK2013 (Tra⁺). Transformants containing integrations in pQBR103 were selected on M9 minimal agar containing tetracycline, lysine, DAP, HgCl₂ and X-Gal.

Identification of plant inducible genes

Detailed description of the use and application of the promoter traps, including the DAP-based strategy, have been published elsewhere [5,7,18]. Briefly, gene fusions showing elevated levels of expression were isolated by inoculating sugar beet seeds (Beta vulgaris var. Amethyst) with ∼10³ bacterial cells of each pool (10 seeds per pool). The inoculated seeds were germinated and cultivated in 5-ml scintillation vials containing non-sterile vermiculite. After 14 days of selection, bacteria were recovered from the shoot (the photosynthetic parts) and rhizosphere (roots with attached vermiculite) and putative plant inducible fusions (Lac⁻) were selected on M9 plates containing tetracycline, lysine, DAP, CFC and X-Gal. White colonies were checked for auxotrophy on minimal M9 medium. In order to confirm that the genes are induced in the phytosphere, the ability of the fusion strains to colonise sugar beet seedlings was determined by a plant competitive colonisation assay. Each fusion strain was inoculated onto sugar beet seeds together with P. fluorescens SBW25-Sm (pQBR103) at a ratio of 1:1 (∼10³ per seed). Bacteria were recovered after 14 days and the IVET fusion strain was counted on M9 agar with tetracycline, lysine, DAP, CFC and X-Gal. The total count of the competitor strain SBW25-Sm (pQBR103) was determined on M9 plates with streptomycin and CFC and the values were similar (∼10⁷ cells per rhizosphere) in all treatments. Pre-selected Lac⁺ (PBR391) and Lac⁻ fusions (PBR393), which were randomly chosen from libraries not subjected to plant selection, were included as positive and negative controls, respectively. Fusions that reached a population density of more than 10⁵ per rhizosphere were considered to be plant inducible.

Confirmation of the plant-inducibility of candidate fusions was obtained by comparing the level of lacZ expression in the plant environment with the level of expression (of the same fusion) after growth in LB broth[7]. β-Galactosidase was measured using 4-methylumbelliferyl-β-d-galactoside as the substrate. The product, 7-hydroxy-4-methylcoumarin (4MU) was detected using the Hoefer DyNA Quant 200 fluorometer (Amersham Pharmacia Biotech, San Francisco, USA) and the reaction was monitored at 460 nm with 365 nm excitation wavelength.

The fusion plasmids integrated onto pQBR103 were recovered by a modified method of conjugative cloning [7,19]. Plasmids were prepared from E. coli DH5αλ_pir and digested by Eco RI/Bam HI to check for siblings. Fusions were directly sequenced using the primers Pdap and Pbla, matching sequences located on the 5′ end of the dapB gene and the 5′ end of the bla gene, respectively[5].

Phylogenetic analysis of the helicase-like genes

Alignments of the deduced amino acid sequences of helA, helB and their homologues were performed using ClusterX[20], and the same programme was used to construct the neighbour-joining trees. Positions with gaps were excluded from the Bootstrap analysis. The tree was displayed in TreeView[21]. Helicase or helicase-like sequences were collected from the NCBI database (accession numbers in parentheses): Magnetococcus sp. MC-1 (ZP_00043295); Coxiella burnetii RSA 493 (NP_820225); Nitrosomonas europaea ATCC19718 (NP_842170); Microbulbifer degradans 2–40 (ZP_00065250); Bordetella parapertussis (NP_883535); pCAR1 (NP_758710); pND6–1 (AAP44264); Homo sapiens (NP_835363); Schizosaccharomyces pombe (T40239); Agrobacterium tumefaciens (NP_353063); Mesorhizobium loti (NP_105807); Pseudomonas fluorescens PfO-1 (ZP_00085208). The putative helicase sequences (PP5352, PP3150, and PP2565) were derived from the P. putida KT2440 genome database (http://www.tigr.org). The P. fluorescens SBW25 homologue was obtained by BLASTP searching of the un-annotated SBW25 genome sequence (http://www.sanger.ac.uk/Projects/P_fluorescens/).

Deletion of plasmid-encoded helA

A BLASTN search of the pQBR103 sequences maintained in the P. fluorescens SBW25 encyclopaedia (http://genomics.nox.ac.uk) identified a recombinant plasmid p8-C2, which contains a 2.88 kb helA gene fragment in vector pUC18 and has a unique Sma I restriction site in the middle of the insert. A tetracycline resistance omega cassette was derived from pHP45Ω-Tc[22] by Sma I restriction and cloned into p8-C2. Then, the whole insert was amplified by PCR, using general M13 forward and reverse primers incorporating Spe I sites: the product was ligated to the integration vector pUIC3. The resultant plasmid (pUIC-helAΩTc) was mobilised into P. fluorescens SBW25 (pQBR103) by conjugation with the help of pRK2013 (tra⁺). The β-galactosidase-deficient allelic exchange mutant was selected on LB agar containing X-Gal, as previously described by Gal et al.[5]. The helA inactivation was further confirmed by PCR amplification using primers F2/R2 (Table 1).

Ultraviolet irradiation

Tolerances to UV radiation were compared between P. fluorescens SBW25-Sm and SBW25-Sm carrying a selection of naturally occurring pQBR plasmids. The bacterial strains were grown in LB broth with streptomycin to the late-log phase. One ml of the cultures was pelleted and washed once in quarter-strength Ringer's solution (Oxoid). The optical density (A = 600) was adjusted to 0.1 and kept the same for the two strains. Then, 50 μl of the cell suspension was spread onto half of a LB plate. Strains with and without plasmid were inoculated as pairs in one Petri dish to ensure that they received the same amount of UV-radiation. The cells were irradiated with UV wavelengths (peak at 302 nm) by placing the Petri-dishes underneath an UV Transilluminator from Ultra-Violet Products Ltd (Cambridge UK, model M-26, Hi, peak of wavelengths at 302 nm) with a fixed distance of 20 cm. Exposure times varied from 30 s to 4 min, with 30 s intervals. Colonies arising from surviving cells were compared after two days incubation in the dark. The results were repeated by two independent experiments.

Results and discussion

Isolation of plant-inducible genes from pQBR103

The DAP-based promoter trap was used to identify plant-inducible genes from pQBR103 and has been described in detail previously[5]. The promoter trapping technique consists of two components: (1) a suicide integration vector pIVETD, containing a promoterless version of ‘dapB and ‘lacZ and (2) a dapB mutant strain of P. fluorescens SBW25 (SBW25ΔdapB) containing pQBR103, which is defective in rhizosphere colonization but can be cultivated in laboratory media supplemented with lysine and DAP. Genomic libraries (contained in pIVETD) made from pQBR103 DNA were screened for transcriptionally active fusions by allowing small pools of fusions to colonise sugar beet seedlings. Growth of fusion strains on sugar beet seedlings only occurs in those instances where promoterless ‘dapB is fused to a pQBR103 promoter that is active during seedling colonisation. Fusion strains with no active promoter cannot grow: they fail to colonise the plant and are thus eliminated. Growth-based enrichment selects for two kinds of promoters: constitutive and inducible. To identify plant-inducible fusions, strains were recovered from the rhizosphere and phyllosphere of sugar beet seedlings after two weeks, and plated onto minimal medium containing DAP, lysine, tetracycline, HgCl₂ and X-Gal. White fusion strains, which contain fusions to putative sugar beet induced genes, were stored for subsequent analysis. Blue colonies (which contain fusions to constitutive promoters) were discarded.

After two weeks of selection in the sugar beet seedling environment approximately 40% of the fusions recovered exhibited white colonies in the presence of X-Gal. These white colonies were reassessed for their inability to grow on minimal plates in the absence of lysine and DAP, their ability to colonise seedlings and their ability to express β-galactosidase in the rhizosphere. Sixty fusion strains matched the criteria and were designated as ppi (plasmid-derived plant inducible) gene fusions and used for further genetic and phenotypic evaluation.

Sequence analysis of pQBR103 genes induced in the sugar beet phytosphere

In order to genetically characterise the 60 IVET-selected ppi genes, the pQBR103-integrated fusion plasmids in P. fluorescens SBW25ΔdapB had to be recovered into E. coli. The environmental plasmid pQBR103 has a very narrow host range and can only replicate in Pseudomonas (not in E. coli). Therefore, the modified single-step conjugative cloning technique, developed in the recovery of chromosomally integrated IVET fusions [7,19], was used to recover the pIVETD fusions integrated onto pQBR103. Restriction analysis of the recovered fusions by Bam HI/Eco RI showed that the same fusions were frequently isolated from different fusion libraries, different plants or different parts of a plant. Thirty-seven unique fusion plasmids were eventually identified from the 60 tested. Among the 37 plant-inducible fusions, 14 were originally isolated from the shoot and 14 from the rhizosphere, while nine fusions were from both. However, when we retested the 37 fusions for expression in the plant environment all fusions were induced in both shoot and root, indicating little evidence of niche-specific expression.

The fusion joint-points were sequenced using primers to the bla and dapB genes, which are located at both ends of the inserted fragments. All sequences are currently available at the P. fluorescens genome encyclopaedia (http://genomics.nox.ac.uk). Searches of DNA databases (http://www.ncbi.nlm.nih.gov/blast) with the fusion sequences resulted in the identification of three genes encoding proteins with a predicted helicase function, while all others have no homologues in the database (expected value <0.1).

Plasmid pQBR103 carries three helicase-like genes showing elevated levels of transcription on plant surface

The three helicase-like genes are the only IVET-identified fusions that show significant homology to genes in the public database. They were designated as helA, helB and helC, and have been subjected to further investigation. In order to verify the genetic context of these helicase genes, additional sequencing was done by primer-walking and searching pQBR103 shotgun sequences. As shown in Fig. 1, the entire coding region was sequenced for helA and helB, while the partial helC sequence represents the 3′ end of a putative helicase. These sequences are available under accession numbers AJ617289 to AJ617291. IVET isolates ppi-4, ppi-16, and ppi-17 have the promoterless ‘dapB gene fused to the 3′ end coding regions of helA, helB and helC, respectively (Fig. 1).

Results of the plant competitive colonisation assay of the three fusion isolates (ppi-4, ppi-16, and ppi-17) versus SBW25-Sm (pQBR103) are shown in Fig. 2. ANOVA revealed a significant difference among the treatments (P < 0.001): all three helicase fusions and a dapB constitutive positive control significantly increased in abundance relative to the negative control (Tukey's HSD P < 0.05); the negative control strain (PBR393) contained a fusion between promoterless dapB and a DNA fragment with no detectable promoter activity. Given that the IVET strains ppi-4, ppi-16 and ppi-17 are unable to grow in M9 medium without addition of lysine and DAP, their ability to grow in the phytosphere of sugar beet demonstrates that the helA, helB and helC genes are induced in the plant environment.

A complete genomic sequence of plasmid pQBR103 is currently not available. In order to confirm that the helABC genes identified in this study are derived from pQBR103 and not from contaminated chromosomal DNA, three primer pairs were designed (Fig. 1) to amplify ∼600 bp of the helA, helB and helC genes, respectively. Results showed that successful amplification of DNA fragments with the expected sizes was achieved from P. fluorescens SBW25 (pQBR103), while no products were obtained by using the plasmid-free strain of P. fluorescens SBW25 as a template (data not shown). Location of the helABC genes was also confirmed by conjugation experiments, in which the related IVET fusions (ppi-4, ppi-16 and ppi-17) were transferred as part of pQBR103 when mobilised by conjugation to P. fluorescens PF145. If these IVET fusions were located on the chromosome, they could not have been transferred by conjugation. Further proof of all three helicases being present on pQBR103 but not in the genome of P. fluorescens SBW25, was obtained by searching the complete genome sequence of SBW25.

Phylogenetic and functional analysis of the HelA, HelB and HelC genes

The helA gene is predicted to encode a protein of 1121 amino acids. A BLAST analysis of the conserved domain database (rpsblast, NCBI) revealed that it has full-length homology with RecB (COG1074, 98.4% aligned, expected value 2e⁻⁶²) and partial homology with UvrD (COG0210, 73.6% aligned, expected value 3e⁻³⁷). The RecB and UvrD proteins are involved in DNA recombination and DNA repair, respectively. The phylogenetic relationships of HelA and those identified by BLASTP search (NCBI) representing different levels of homology are shown in Fig. 3(a). The full length HelA is most similar to putative helicases derived from five genome-sequenced bacterial species across alpha, beta and gamma subdivisions of proteobacteria (Magnetococcus, Coxiella, Nitrosomonas, Microbulbifer and Bordetella). However, it is distinct from its closest Pseudomonas homologues (Fig. 3(a)), showing only 23–24 % amino acid identity among the aligned region (Table 3). This suggest that helA might have been acquired from other proteobacteria by horizontal gene transfer, or represents a deep and distinct evolutionary line from the chromosome.

The putative helB gene encodes a protein of 646 amino acids. Its closest homologues are probable helicase genes carried on two Pseudomonas plasmids pCAR1 and pND6-1, respectively. Full-length homologues were also found in the chromosome of P. fluorescens SBW25, PfO-1 and P. putida KT2440, but with only 25–34% amino acid identities (Table 3). However, no homologues were found in P. aeruginosa PAO1 or P. syringae DC3000, with expected value less than 0.001. A phylogenetic tree based on the amino acid sequences within the conserved UvrD domain is shown in Fig. 3(b). Interestingly, the group of helicase genes carried on the Pseudomonas chromosome or plasmids show homology with the helicase motifs of mammalian F-box proteins, which are components of modular E3 ubiquitin protein ligases (SCFs) involved in cell cycle regulation and signal transduction[23]. However, they do not contain the F-box consensus sequence.

The partially sequenced helC gene shows homology (E-value < 0.1) only with putative helicases from Pseudomonas (Table 3), suggesting that the helC might be either originated from the chromosome of Pseudomonas or might have co-evolved. Altogether, these data show that the three plasmid-derived helicase-like genes are phylogenetically distinct and thus may have different functions.

It is proposed that helicases unwind double-stranded DNA to single-stranded DNA by utilizing energy derived from the hydrolysis of nucleotide triphosphates. They play essential roles in various DNA transactions including replication, repair, recombination and plasmid conjugation[24]. The bacterial genome contains a large number of helicases or helicase-like genes. The possible functional overlapping makes it difficult to determine the special roles of a given helicase. It has been reported that some proteins contain well-conserved helicase motifs while no helicase function has been demonstrated[25]. This suggests that the helicase-like proteins might have functions not involved in DNA unwinding.

In an effort to investigate the functionalities of the present IVET-identified helicase-like genes, a helA knockout mutant was constructed by inserting an omega cassette carrying a tetracycline resistance gene into the Sma I site located in the middle of the helA gene (Fig. 1). The obtained helA mutant plasmid (pQBR103helA::ΩTc) is still self-transmissible and no significant reduction of transfer frequency (∼2 × 10⁻⁶ per recipient) was observed when the plasmid was transferred from P. fluorescens SBW25-Sm to SBW25 EeZY6KX.

In order to determine whether helA participates in DNA repair, the relative sensitivity of P. fluorescens SBW25, SBW25 carrying pQBR103 and SBW25 carrying pQBR103helA::ΩTc to ultraviolet radiation (UVR) was assessed. The results show that both pQBR103 and pQBR103 helA::ΩTc increased the UV tolerance of the host bacteria, but no significant differences were found between them (data not shown).

Plasmids exogenously isolated from the same field as pQBR103 were characterized into five distinct groups (Groups I–V) by whole plasmid restriction fragment length polymorphism (RFLP) analysis, and pQBR103 represents the dominant group of “Group I”[10]. In order to determine the distribution of the three plant-inducible helicase-like genes in natural Pseudomonas plasmid populations, 12 representative plasmids, including pQBR103, were selected for PCR amplification of the helA, helB or helC genes, using three primer pairs (Fig. 1). In addition, the UV-tolerance of P. fluorescens SBW25 carrying these plasmids was assessed by comparison to the plasmid-free strain of SBW25. The hel genes were detected only in Group I plasmids and only four of these plasmids, including pQBR103, carry all three genes (Table 2). Increased UV-tolerance was found only in Group I plasmids and it was not correlated with the presence of any of the helicase-like genes. For example, plasmid pQBR110 shows increased UV-tolerance but does not carry helA, helB or helC. These results again suggest that the helicase-like genes are not directly involved in DNA repair induced by UV-radiation.

IVET analysis of plant-inducible genes from P. fluorescens SBW25 (pQBR103)

According to the local adaptation hypothesis[26] most phenotypes associated with plasmids represent adaptations to local variations of environmental conditions. Being an indigenous plasmid active in the sugar beet phytosphere, we hypothesized that pQBR103 contains a relatively larger proportion of genes responsive to factors in the plant environments than the chromosomal DNA. This hypothesis was tested by IVET screening of plant-inducible genes from the genomic DNA of P. fluorescens SBW25 (pQBR103), other than the pQBR103 DNA described above. IVET fusions to plasmid pQBR103 genes were determined by their ability to conjugatively transfer to other Pseudomonas via the transmission of pQBR103, while the chromosomally integrated fusions were not transmissible.

In order to estimate the ratio of plasmid- and chromosome-derived plant-inducible genes, an IVET library was constructed using genomic DNA of P. fluorescens SBW25 (pQBR103). One hundred fusions strains were identified that contained fusions to plant-inducible loci either in the chromosome or in the plasmid pQBR103. To identify the plasmid-derived fusions, conjugation experiments were carried out between these Pseudomonas fusion strains (donors) and the P. fluorescens PF145 recipient strain. We observed that 13 of the 100 IVET-selected fusions gave rise to kanamycin and tetracycline transconjugants, indicative of pQBR103-based fusions. The estimated ratio of plant-inducible genes carried on pQBR103 and the chromosome (13:87, ∼ 1:6.7) is much higher than the ratio relative to their genome sizes (400 kb verses 6.71 Mb, 1:16.8) (http://www.sanger.ac.uk). It is not clear whether the significance is due to multiple copies of pQBR103 in the cell, which is currently unknown, or whether the plasmid carries a higher proportion of ppi genes. However, the result of direct assessment is supported by the comparison of 37 ppi fusions identified here from pQBR103 and 195 known in the chromosome (Rainey, unpublished data). Altogether, the data do suggest that the plasmid carries relatively more plant-inducible genes than the chromosome.

In this report, we describe the application of a promoter trapping strategy to the functional analysis of the large naturally occurring plasmid pQBR103. Our data show that pQBR103 carries genes whose expression levels are elevated in the plant environment. Thirty-seven different fusions were isolated and their plant inducing patterns were analysed in laboratory media and in the plant environment. Genome sequencing of large indigenous plasmids is very limited and the results available are generally by-products of genome sequencing efforts[27]. A homology search of the IVET fusion sequences revealed that most of them have no homologues in the database; the exceptions being three genes predicted to be helicases. The actual contributions of these plasmid-derived genes to the fitness of the host bacteria are currently unknown. Further functional characterisation of these plant-inducible genes is likely to contribute to the understanding of the ecological roles of plasmid DNA in the natural environment.

Acknowledgements

We are grateful to Lena Ciric for assistance with DNA sequencing, Sarah Turner for helpful discussions, Robert Jackson for critical reading and Andrew Spiers and Dawn Field for continued maintenance of the P. fluorescens SBW25 genome encyclopaedia. This work was supported by NERC (UK).

References