Prophage 01 spans 16,875 bp and consists of genes encoding a myovirus-like tail, holin and lysozyme lytic genes, a putative chitinase gene (PFL_1213), and genes for a repressor protein (PFL_1210) and a leptin binding protein-like bacteriocin, LlpA1 (PFL_1229) (Fig. 1, see Additional file 1). The tail assembly gene cluster encodes tail sheath (PFL_1216), DNA circulation (PFL_1216), baseplate assembly (PFL_1222), and tail fiber (PFL_1226) proteins that closely resemble their counterparts from the long contractile tail of the serotype-converting bacteriophage SfV of Shigella flexneri 16.

We (D.V.M. and L.S.T.) previously identified a highly similar prophage element during a study focused on genetic traits contributing to colonization of the plant rhizosphere by P. fluorescens. In that project 17, we applied genomic subtractive hybridization to two strains of P. fluorescens, Q8r1-96 and Q2-87, which differ in their ability to colonize wheat roots. Among 32 recovered Q8r1-96-specific loci was a clone dubbed ssh6, which proved to constitute part of a 22-kb prophage element that closely resembles prophage 01 of strain Pf-5 (Figs. 1 and 2; see Additional file 2). Like its counterpart, the ssh6 prophage from Q8r1-96 carries genes for a myovirus-like tail (orf10 through orf21), the lytic enzymes holin (hol) and endolysin (lys), and a Cro/CI-like repressor protein (prtR) (Fig. 1; see Additional file 2). Genes in the Q8r1-96 cluster that are not present in Pf-5 encode a colicin M-like bacteriocin (cma), a tail collar protein (orf23), and putative tail fiber proteins (orf22 and orf25). Interestingly, the colicin M-like ORF from the ssh6 prophage of Q8r1-96 also encodes an enzymatically active protein although the range of microorganisms sensitive to this bacteriocin is currently unknown (Dr. Dominique Mengin-Lecreulx, Institut de Biochimie et Biophysique Moléculaire et Cellulaire, Université Paris-Sud, Orsay, France; personal communication).

The homologous prophage elements from Pf-5 and Q8r1-96 have simple overall organization, lack integrase and head morphogenesis genes, and carry conserved regulatory, lytic and lambda-like tail morphogenesis genes also found in phage SfV of Shigella flexneri (Fig. 1). Taken together, the results of sequence analyses suggest that these regions are not simple prophage remnants but rather, are similar to F-type pyocins. F-type pyocins were first discovered in P. aeruginosa and represent a class of high molecular weight protease- and nuclease-resistant bacteriocins that resemble flexible and non-contractile tails of bacteriophages 1819. This notion is further supported by the fact that the putative lytic genes found within Pf-5 prophage 01 (Fig. 3) and Q8r1-96 (data not shown) seem to be fully functional. In non-filamentous bacteriophages and bacteriophage tail-like bacteriocins, the lytic activity is provided by the combined action of the small membrane protein holin and a cytoplamic muralytic enzyme, endolysin 1920. During phage-mediated cell lysis, holin permeabilizes the cytoplasmic membrane and allows endolysin, which lacks a secretory signal sequence, to gain access to peptidoglycan. To confirm that the prophage 01-like loci indeed encode functional holin and endolysin, we cloned genes PFL_1211 and PFL_1227 from Pf-5 and their counterparts from Q8r1-96 (Fig. 1) in Escherichia coli under the control of an inducible T7 promotor. As shown in Fig. 3, induction of both of the putative holin and endolysin genes by IPTG had a strong lethal effect on the host, resulting in rapid cell lysis. In accordance with the current model of action of holin and endolysin, the lethal effect of the endolysin encoded by PFL_1227 was not apparent unless the cytoplasmic membrane was destabilized by addition of small amount of chloroform to the induced E. coli culture (Fig. 3B). Gene induction experiments carried out with putative holin and endolysin genes from the ssh6 locus of Q8r1-96 had a similar lytic effect on E. coli (data not shown).

Interestingly, a prophage element found in the identical spot (between mutS and cinA) in the genome of P. fluorescens SBW25 http://www.sanger.ac.uk/Projects/P_fluorescens has a similar overall organization but contains a P2-like bacteriophage tail cluster (orf5 through orf18) similar to that in phage CTX (Fig. 1), thus resembling another class of phage tail-like bacteriocins, the R-type pyocins of P. aeruginosa 19. Furthermore, a homologous region from P. fluorescens Pf0-1 (CP000094) contains both the lambda-like and P2-like tail clusters (Fig. 1), making it similar to the hybrid R2/F2 pyocin locus from P. aeruginosa PAO1 19. The differences in organization of the putative phage tail-like pyocins among these prophages clearly indicate that the corresponding loci are subject to extensive recombination, with a possible recombination hotspot between two highly conserved DNA segments comprised of the phage repressor (prtR) and holin (hol) genes, and the endolysin (lys) gene (Fig. 1).

In strains Pf-5 and Q8r1-96, the putative prophage 01-like pyocins are integrated between mutS and the cinA-recA-recX genes (Fig. 1), which suggests that these elements might be activated during the SOS response, as is the putative prophage gene cluster integrated into the mutS/cinA region of P. fluorescens DC206 21. The mutS/cinA region is syntenic in several Gram-negative bacteria 22, and comparisons reveal that prophage 01-like elements occupy the same site in the genomes of P. fluorescens Pf0-1, P. fluorescens SBW25, and P. entomophila L48 23, whereas unrelated prophages reside upstream of cinA in P. putida F1 (GenBank CP000712) and P. syringae pv. tomato DC3000 24. The latter strain, as well as P. putida KT2440 25, carry SfV-like bacteriophage tail assembly clusters elsewhere in the genome.

The putative F- and R-pyocins appear to be ubiquitously distributed among strains of P. fluorescens as illustrated by a screening experiment (Fig. 4) in which genomic DNA of different biocontrol strains was hybridized to probes targeting the lambda-like and P2-like bacteriophage tail gene clusters of Q8r1-96 and SBW25, respectively. The F- and R-pyocin-specific probes each strongly hybridized to DNA from 12 of 34 P. fluorescens strains, while the remaining 22 strains tested positive with both probes.

A second large prophage, prophage 03, spans 33.5 kb (Fig. 5A; see Additional file 3) of the Pf-5 genome. Closely related prophages exist in the genomes of P. putida KT2440 25 and P. syringae pv. tomato DC3000 24 (Fig. 2) but were not found in P. fluorescens strains Pf0-1 or SBW25. Prophage 03 is a chimeric element that contains a siphovirus head morphogenesis region and a myovirus-like tail assembly region (Fig. 5A). The prophage also carries a putative integrase gene (PFL_1976) that encodes an enzyme similar to shufflon recombinases such as the Rci recombinase from plasmid R64 26, a gene involved in DNA modification (PFL_1978), and a gene for a cytosine-specific methylase (PFL_1979). Genes encoding a LexA-like repressor (PFL_1986), a putative single strand binding protein (PFL_1989), and two genes (PFL_1976 and PFL_1982) with similarity to the pyocin transcriptional activator prtN also are present in this region. Holin (PFL_1991) and endolysin (PFL_2018) genes flank a region containing DNA packaging and head morphogenesis and tail assembly genes. The P2-like tail assembly region closely resembles the R2-specific part of R2/F2 pyocin locus of P. aeruginosa PA01 19 (Fig. 5A) and includes genes encoding a tail sheath protein (PFL_2009), a tape measure protein (PFL_2013), a major tail tube protein (PFL_2010), baseplate assembly proteins (PFL_2002 and PFL_2003), and a tail fiber protein (PFL_2007). This region also contains genes involved in head morphogenesis (PFL_1993–1998) that are not present in the R-part of R2/F2-type pyocin cluster of P. aeruginosa PA01. Therefore, prophage 03 may represent the genome of a temperate bacteriophage rather than an R-type pyocin.

Prophage 06 is the largest prophage region of P. fluorescens Pf-5 and encodes a 56-kb temperate lambdoid phage integrated into tRNA^Ser(see Additional file 4). It is mosaic in nature with no homologues present in strains Pf0-1 or SBW25. P. fluorescens Pf-5 carries four genomic copies of tRNA^Ser, of which tRNA^Ser(2) and tRNA^Ser(3) are associated with prophages carrying integrases of different specificity (see Additional file 5). The anticodon, V- and T-loops of tRNA^Ser(2) are parts of the 104-bp putative attB site of prophage 06, whereas the T-loop of tRNA^Ser(3) forms part of the 60-bp putative attachment site of prophage 02. The latter is a prophage remnant that spans 8.4 kb and consists of a gene encoding an ATP-dependent nuclease (PFL_1842) and a phage integrase gene with two internal frameshift mutations (see Additional file 6). The mobility of prophage 06 probably is mediated by a lambda-type integrase encoded by PFL_3794, which resides adjacent to the putative attR site. Prophage 06 contains gene modules that are involved in head morphogenesis (capsid proteins PFL_3764 and PFL_3765), DNA packaging (terminase PFL_3766), DNA recombination (a NinG-like protein, PFL_3773 and a putative NHN-endonuclease, Orf1) and tail morphogenesis (tail tip fiber proteins PFL_3744 and PFL_3751, tail length tape measure protein PFL_3753, and minor tail proteins PFL_3749, PFL_3750, and PFL_3752). The tail assembly module resembles the corresponding region from Burkholderia thailandensis bacteriophage φE125 27, although in prophage 06 the module is split by the integration of four extra genes (Fig. 5B). Prophage 06 also contains a regulatory circuit with genes for a Cro/C1 repressor protein (PFL_3780) and two putative antirepressor proteins (PFL_3747 and PFL_3746); a gene for a putative cytosine C5-specific methylase (PFL_3792); and lysis genes encoding holin (PFL_3770) and endolysin (PFL_3798). However, since the endolysin gene is localized beyond the putative attR site it is not clear whether it represents part of the prophage 06 genome or a remnant from integration of a different phage (see Additional file 4). Finally, prophage 06 contains two genes, PFL_3740 and PFL_3796, which probably arose through gene duplication and encode putative conserved phage-related proteins that are 88% identical to one another.

Prophages 04 and 05 are prophage remnants with reduced size and/or complexity that carry several mutated phage-related genes (Tables 1, see Additional files 7 and 8). Prophage 04 (13.5-kb) has an average G+C content of 56.3% and contains a putative phage integrase gene adjacent to tRNA^Pro (1). It also carries two genes (PFL_2122 and PFL_2123) that encode minor tail assembly proteins, a gene encoding Cro/C1 repressor, and the bacteriocin gene llpA1 (PFL_2127). Interestingly, the repressor gene and llpA1 are highly similar to their counterparts from prophage 01, suggesting that they arose via gene duplication. Prophage 05, a 2.6-kb prophage remnant, has a G+C content of 55.3% and carries genes encoding a truncated phage integrase and a putative phage tail protein (PFL_3464) (see Additional file 8). The region is flanked by 84-bp direct repeats, one of which probably represents the attB site and partially overlaps with the anticodon and T loops of tRNA^Cys.

Genomic island 02, or PFGI-2, spans 16.8 kb and has an average G+C content of 51.5%. It is flanked by imperfect 51-bp direct repeats, one of which partially overlaps with tRNA^Leu(6) and probably represents the attB site (see Additional file 10). Although P. fluorescens Pf-5 does not have a type III protein secretion pathway, approximately half of PFGI-2 (i.e. an 8.1-kb DNA segment spanning genes PFL_4977 to PFL_4980) closely resembles a gene cluster found in the exchangeable effector locus (EEL) of a

t

pa

i

Unlike the genomes of other Pseudomonas spp., that of P. fluorescens Pf-5 is devoid of IS elements and contains only one CDS (PFL_2698) that appears to encode a full-length transposase. Three other transposase-like CDSs (PFL_1553, PFL_3795, and PFL_2699) found in the Pf-5 genome contain frameshifts or encode truncated proteins. PFL_2698 and PFL_2699 encode IS66-like transposases and are found within a large cluster (PFL_2662 through PFL_2716) of conserved hypothetical genes. Corrupted transposases encoded by PFL_1553 and PFL_3795 belong to the IS5 family and are associated with gene clusters encoding a putative filamentous hemagglutinin and prophage 06, respectively.

Wild type variants of P. fluorescens Pf-5 5, P. fluorescens SBW25 73, and P. fluorescens Q8r1-96 74 were used in the study. Pseudomonas strains were grown at 28°C in King's B medium 75, while E. coli strains were grown in LB 76 or 2xYT 76 at 25°C or 37°C. When appropriate, antibiotic supplements were used at the following concentrations: tetracycline, 12.5 μg/ml; chloramphenicol, 35 μg/ml; and ampicillin, 100 μg/ml.

Plasmid DNA isolation, restriction enzyme digestion, agarose gel electrophoresis, ligation, and transformation were carried out using standard protocols 76. All primers were developed with Oligo 6.65 Software (Molecular Biology Insights, West Cascade, Colo.), and routine PCR amplifications were performed with Taq DNA polymerase (Promega, Madison, Wisc.) according to the manufacturer's recommendations.

Sequencing of prophage 01 from P. fluorescens Q8r1-96 was carried out essentially as described by Mavrodi et al. 77. Briefly, the Q8r1-96 gene library was screened by PCR with oligonucleotide primers col1 (5' GCT GCT GGG CAA TGG TAA CAC 3') and col2 (5' CTG CCG ACT GCT CAC CTA TC 3') and a positive cosmid clone was shotgun sequenced by using the EZ::TN™ <Kan-2> transposition system (Epicentre Technologies, Madison, Wisc.). DNA sequencing was carried out by using an ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, Calif.), and sequence data were compiled and analyzed with Vector NTI 9.1.0 (Invitrogen Corp., Carlsbad, Calif.) and OMIGA 2.0 (Accelrys, San Diego, Calif.) software packages. Database searches for similar protein sequences were performed using the NCBI's BLAST network service, and searches against PROSITE, Profile, HAMAP, and Pfam collections of protein motifs and domains were carried out by using the MyHits Internet engine 78. Signal peptides were predicted with SignalP v. 3.0. 79. The nucleotide sequence of prophage 01 from P. fluorescens Q8r1-96 has been deposited in GenBank under accession number EU982300.

The 3.12-kb prophage 01 probe was amplified by PCR from P. fluorescens Q8r1-96 genomic DNA with the oligonucleotide primers orf11-1 (5' CAT TCG TGT GCC GCT GTT CTA 3') and orf14-2 (5' TGA CCA GGC GAA CAG CGT CTG 3'). The 1.79-kb P. fluorescens SBW25-specific prophage 01 probe was amplified from genomic DNA of SBW25 with oligonucleotides SBW3 (5' GAA CTC ACC AGC GTC CTT AAC 3') and SBW4 (5' GGG CAG CTC CTT GGT GAA GTA 3'). Amplification was carried out with Expand Long DNA polymerase (Roche Applied Science, Indianapolis, IN) according to manufacturer's recommendations. The cycling program included a 2-min initial denaturation at 94°C followed by 25 cycles of 94°C for 10 sec, 60°C for 30 sec, and 68°C for 3 min. Amplified DNA was gel-purified with a QIAEX II gel extraction kit (Qiagen, Santa Clarita, Calif.) and labeled with the Biotin High Prime System (Roche Applied Science). Genomic DNA was purified using a cetyltrimethylammonium bromide miniprep protocol 76, digested with EcoRI and PstI, separated by electrophoresis in a 0.8% agarose gel, and transferred onto a BrightStar-Plus nylon membrane (Ambion, Inc., Austin, TX) in 0.4 M NaOH. The membranes were pre-hybridized and hybridized at 58°C in a solution containing 5 × SSC 80, 4 × Denhardt's solution 80, 0.1% SDS, and 300 μg per ml of denatured salmon sperm DNA (Sigma). After hybridization, the membranes were washed twice in 2 × SSC, 0.1% SDS at room temperature, twice in 0.2 × SSC, 0.1% SDS at room temperature, and once in 0.2 × SSC, 0.1% SDS at 60°C.

Full-length hol genes from strains Pf-5 and Q8r1-96 were amplified by using KOD Hot Start DNA polymerase (Novagen, Inc.) and oligonucleotide primer pairs holupPf5 (5' AGG GAC CTC TAG AAA CAT CGT TA 3') – holowPf5 (5' TTT TGG ATC CGG TGA GTC AAG GCT G 3') and hol-xba (5' GAC CAG TCT AGA CAT GCT CAT CA 3') – hol-low (5' TTT TGG ATC CGC GGT ATC GCT T 3'), respectively. Full-length lys genes from Pf-5 and Q8r1-96 were amplified by using primer sets lysupPf5 (5' CGC CAT TCT AGA TTA CTG AAC AA 3') – lyslowPf5 (5' TTT TGG ATC CGC AGG ACC TTC AGA C 3') and lysQ8-up (5' CGG ACA TCT AGA ATC ATG CAC TTG 3') – tail13 (5' GCC GCT TGG GTG ATT TGA TT 3'), respectively. The cycling program included a 2-min initial denaturation at 94°C followed by 35 cycles of 94°C for 15 sec, 59°C for 30 sec, and 68°C for 1 min, and a final extension at 68°C for 3 min. PCR products were gel-purified and cloned into the SmaI site of the plasmid vector pCR-Blunt (Invitrogen) under the control of the T7 promoter. The resultant plasmids were single-pass sequenced to confirm the integrity of cloned genes and electroporated into E. coli Rosetta/pLysS (Novagen) with a Gene Pulser II system (Bio-Rad Laboratories, Hercules, Calif.). Plasmid-bearing E. coli clones were selected overnight on LB agar supplemented with ampicillin and chloramphenicol, and suspended in 2xYT broth supplemented with antibiotics to give an OD₆₀₀ of 0.1. After incubation with shaking for one hour at room temperature, gene expression was induced in the broth cultures with 3 mM IPTG. The induced cultures were incubated with shaking for another 5 hours and the cell density was monitored by measuring OD₆₀₀ every 30 min. To disrupt cell membranes in endolysin-expressing cultures, a drop of chloroform was added after four hours of induction. Two independent repetitions were performed with each strain.