pGIL01, a linear tectiviral plasmid prophage originating from Bacillus thuringiensis serovar israelensis

Céline Verheust1, Gert Jensen2 and Jacques Mahillon1

1 Université Catholique de Louvain, Place Croix du Sud, 2/12, B-1348 Louvain-la-Neuve, Belgium
2 National Institute of Occupational Health, Lersø Parkallé, DK-2100, Copenhagen, Denmark

Correspondence
Jacques Mahillon
Mahillon{at}mbla.ucl.ac.be


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacillus thuringiensis serovar israelensis harbours, in addition to several circular plasmids, a small linear molecule of about 15 kb. Sequence analysis of this molecule, named pGIL01, showed the presence of at least 30 ORFs, five of which displayed similarity with proteins involved in phage systems: a B-type family DNA polymerase, a LexA-like repressor, two potential muramidases and a DNA-packaging protein (distantly related to the P9 protein of the tectiviral phage PRD1). Experimental evidence confirmed that pGIL01 indeed corresponds to the linear prophage of a temperate phage. This bacteriophage, named GIL01, produces small turbid plaques and is sensitive to organic solvents, which suggests the presence of lipid components in its capsid. Experiments using proteases and exonucleases also revealed that proteins are linked to the genomes of both pGIL01 prophage and GIL01 phage at their 5' extremities. Altogether, these features are reminiscent of those of phages found in the Tectiviridae family, and more specifically of those of PRD1, a broad-host-range phage of Gram-negative bacteria. Dot-blot hybridization, PFGE, PCR and RFLP analyses also showed the presence of pGIL01 variants in the Bacillus cereus group.


Abbreviations: sv., serovar

The GenBank accession number for the sequence of pGIL01/GIL01 is AJ536073.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacillus thuringiensis, the most widely used entomopathogenic bacterium, belongs to the Bacillus cereus sensu lato group. This cluster also includes B. cereus sensu stricto, an opportunist organism implicated in food poisoning (Granum & Lund, 1997), Bacillus anthracis, a human and animal pathogen, Bacillus mycoides and Bacillus pseudomycoides, characterized by their rhizoid growth (Nakamura, 1998), and the psychrotolerant Bacillus weihenstephanensis (Lechner et al., 1998). Despite their broad virulence, these bacteria are genetically closely related (their 16S RNA gene sequences share more than 99 % identity) and could be regarded as pertaining to a single species (Ash et al., 1991; Helgason et al., 2000).

B. thuringiensis strains produce, during their sporulation, crystal toxins (delta-endotoxins) that are highly toxic to a number of insect larvae belonging to the orders Lepidoptera, Diptera and Coleoptera, but harmless to vertebrates. Classically, the numerous entomopathogenic B. thuringiensis strains, which have their own specific insecticidal activity, have been classified in different serotypes on the basis of their flagellar antigens. B. thuringiensis serovar (sv.) israelensis is active against dipteran species and is therefore one of the bioinsecticides of choice to control black flies and mosquitoes, both important vectors of human and animal diseases (for a recent review, see Glare & O'Callaghan, 2000).

B. thuringiensis sv. israelensis strain H14 has been reported to contain at least eight DNA molecules, including three small (5·4, 6·7 and 7·6 kb) and four large (128, 145, 240 and 350 kb) circular plasmids, and one linear molecule (G. Jensen & L. Andrup, unpublished results). The pathogenicity of this strain only depends on the presence of the 128 kb plasmid which encodes the Cry and Cyt toxins (González & Carlton, 1984). This megaplasmid, named pBtoxis, has been sequenced, allowing the discovery of new important toxic factors potentially involved in the insecticidal activity (Berry et al., 2002). The 350 kb molecule, named pXO16, is a conjugative plasmid also observed in other strains of B. thuringiensis. It is responsible for an aggregation-mediated conjugation system, leading to a very high frequency of transfer (~100 %) (Jensen et al., 1995). The sequence analysis of the three small circular plasmids, named pTX14-1, pTX14-2 and pTX14-3, revealed the presence of genes implicated in plasmid replication (rep) and mobilization (mob) (Madsen et al., 1993). Interestingly, they each harbour a specific gene containing repetitive elements similar to collagen. However, the function of these genes, termed bcol for Bacillus-collagen-like genes, has not been established (Andrup et al., 2003).

B. thuringiensis strains are naturally associated with a number of virulent (Colasito & Rogoff, 1969b) or temperate phages (Ackermann & Smirnoff, 1978; Colasito & Rogoff, 1969a). It has been shown that, during their lysogenic state, temperate phages are characterized by their integration into the host genome (chromosome or plasmids) (Kanda et al., 1998). However, until now, no prophage existing as an autonomous plasmid in the host cell has been observed in B. thuringiensis.

In this study, we report the complete DNA sequence of a linear molecule originating from B. thuringiensis sv. israelensis strain H14. This 15 kb linear element, named pGIL01, actually corresponds to the prophage form of a temperate bacteriophage, GIL01, and this establishes the first molecular characterization of a linear prophage originating from B. thuringiensis.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and plasmids.
A list of the most relevant Bacillus strains used in this study is given in Table 1. The complete list can be obtained from the authors upon request. Strain AND508 is a mutant of the commercial strain NB31 of B. thuringiensis sv. israelensis cured of its three small circular plasmids (Andrup et al., 1993). GBJ002 (Jensen et al., 1996), a B. thuringiensis sv. israelensis 4Q2-72-derived strain cured of all its plasmids, was used as an indicator strain for phage experiments. The Bacillus strains were grown on LB medium (Sambrook & Russel, 2001) at 30 °C with moderate agitation. Escherichia coli TOP10 (Invitrogen) was grown in LB medium at 37 °C and was used as a bacterial host for DNA manipulation.


View this table:
[in this window]
[in a new window]
 
Table 1. Strains used in this study

 
DNA manipulation.
General techniques for molecular manipulations were performed as described by Sambrook & Russel (2001). The antibiotic kanamycin was used at concentration of 50 µg ml-1. For PFGE, the preparation of intact genomic DNA in agarose plugs was performed as described previously (Léonard et al., 1998). The method developed for the isolation of large B. thuringiensis plasmids was used for the isolation of linear plasmids, as described by Jensen et al. (1995).

Cloning strategy.
A combination of different strategies was used to sequence the entire pGIL01 molecule. Restriction fragments as well as PCR products were cloned into positive selection vectors [pZErO-2, pCR4-TOPO and pCR-XL-TOPO (Invitrogen)]. Sequences derived from these PCR products were confirmed by direct sequencing on phage GIL01 DNA. The remaining sequences of pGIL01/GIL01 were obtained by ‘primer-walking’ along the phage DNA, including ‘run-off sequencing’ of the extremities. Poly(dT) tailing of GIL01 used for the confirmation of the terminal sequences was performed with the calf thymus terminal transferase (Roche) according to the manufacturer's recommendations (Polo et al., 1998). Both extremities of the tailed DNA were amplified using poly(dA)35 in combination with internal primers. The detailed strategy can be obtained from the authors upon request.

Proteinase K, exonuclease III and lambda exonuclease treatments of GIL01.
DNA preparations of GIL01 (see below), performed without proteinase K pretreatment, were incubated with variable amounts of proteinase K (0·01 and 0·1 mg ml-1) for 4 h at 37 °C. The effect of the protease was analysed by running the GIL01 DNA preparation on a 0·8 % agarose gel (0·5xTAE; TAE, 40 mM Tris/acetate, 1 mM EDTA). GIL01 preparations were treated with exonuclease III (MBI, Fermentas) or lambda exonuclease (Roche), according to the manufacturer's recommendations. After incubation at 37 °C for 10–90 min, the reaction was stopped by chilling on ice. The samples were analysed by electrophoresis on a 0·8 % agarose gel (0·5xTAE).

Phage DNA extraction.
DNA extraction from phage particles was performed using the protocol described by Santos (1991) and modified as follows. To each millilitre of phage suspension treated with DNase (5 µg ml-1) and RNase (10 µg ml-1) for 30 min at 37 °C, 20 µl of a filtered sterilized 2 M solution of ZnCl2 was added and the solution was incubated for 5 min at 37 °C. After centrifugation for 1 min at 10 000 r.p.m., the supernatant was removed and the pellet was resuspended in 500 µl TES buffer (0·1 M Tris/HCl, pH 8; 0·1 M EDTA; 0·3 % SDS) and incubated at 60 °C for 15 min. After incubation for 90 min at 37 °C with 20 µl proteinase K (stock solution of 20 mg ml-1), 60 µl of a 3 M potassium acetate solution (pH 5·2) was added (incubation on ice for 10–15 min). This solution was first treated with phenol/chloroform/isoamyl alcohol (25 : 24 : 1, by vol.) and then with chloroform/isoamyl alcohol (24 : 1, v/v), was precipitated with 2-propanol, and finally washed with 70 % ethanol, dried and recovered in 10–20 µl distilled water.

Treatment of phage GIL01 with organic solvents.
The GIL01 suspension (2 ml) was mixed with various amounts of chloroform (10, 40 and 80 µl) or ether (10, 40, 80, 160 and 320 µl) and incubated at room temperature for 15 min. The aqueous phase was then titrated for surviving phages. A control sample without solvent was treated simultaneously.

Mitomycin C, nalidixic acid and UV treatments.
An exponential culture (20 ml) was centrifuged and the pellet was resuspended in 5 ml of 0·01 M MgSO4. Mitomycin C and nalidixic acid were added to the solution at final concentrations of 1 and 40 µg ml-1, respectively. The solution was incubated for 30 min at 30 °C, and then washed twice with 5 ml of 0·01 M MgSO4. The bacterial cells were resuspended in 20 ml fresh LB medium and incubated for 4 h at 30 °C. After centrifugation, the filtered supernatant (0·45 µm filter) was analysed by spot test and/or by titration.

For UV treatment, the suspension was placed in a 9 cm glass Petri dish and irradiated with 254 nm UV light, for 5–20 s. This treatment produced between 90 and 95 % cell lethality. The irradiated suspension was then incubated for 15 min at 30 °C, and centrifuged. The bacterial cells were resuspended in 20 ml fresh LB medium, incubated for 4 h at 30 °C and, after centrifugation, the titre of the filtered supernatant was analysed.

PCR amplification from phage plaques.
Samples of top agar (~1 mm in diameter) corresponding to a phage plaque, or a bacterial lawn for the negative control, were recovered from a titration plate, and diluted in 20 µl of 0·9 % NaCl. PCRs were then performed on 2 µl of those samples using primers matching pGIL01. Primer 3 (5'-CGGTTGCTTTGCCGTATG-3') and primer 4 (5'-GGCGAACAACATGCTTTGGG-3') allowed the amplification of a 1·4 kb fragment.

Bioinformatics.
Analyses of DNA and protein sequences were performed with either the GCG (genetics computer group) or the EMBOSS packages at the Belgian EMBL Node (BEN). The sequence of pGIL01/GIL01 has been deposited in the reference databases under accession number AJ536073.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strain H14 of B. thuringiensis sv. israelensis harbours four large (pBtoxis, pXO16 and the 145 and 240 kb elements) and three small circular plasmids (pTX14-1, pTX14-2 and pTX14-3), and a linear molecule we named pGIL01. To facilitate the isolation and characterization of this linear element, strain AND508, cured of the three small plasmids, was used (Andrup et al., 1993).

pGIL01 is a 14 931 bp linear molecule
Several complementary approaches were combined to determine the entire DNA sequence of pGIL01 (see Methods). The size of pGIL01 was determined to be 14 931 bp. As illustrated in Fig. 1(b), its genome is flanked by 73 bp terminal inverted repeats sharing more than 75 % identity. Detailed analysis of the pGIL01 sequence also revealed the existence of 30 potential ORFs. The main features of these ORFs are listed in Table 2. Interestingly, most ORFs were relatively short, all pointing in the same direction, with small intergenic spaces (Fig. 1a).



View larger version (17K):
[in this window]
[in a new window]
 
Fig. 1. Genetic organization of pGIL01/GIL01. (a) Thirty ORFs were reported, six of which showed similarity with known proteins (indicated by black arrows). (Exc), Excisionase; pol, DNA polymerase; Lex-A, Lex-A-like repressor; DNA-pack, DNA-packaging protein; Mur A and B, muramidases. The full coordinates of these ORFs are reported in Table 2. (b) The 73 bp Terminal Inverted Repeats present at both pGIL01 extremities shared more than 75 % identity.

 

View this table:
[in this window]
[in a new window]
 
Table 2. Potential protein-coding ORFs of pGIL01

All ORFs were in the positive frame.

 
Among the 30 putative ORFs, only six showed homology to known proteins (Table 2 and Fig. 1a). The 735 residues of ORF5 showed similarity to several fungal DNA polymerases, but also to several phages' polymerases (38, 37 and 24 % identity with the DNA polymerases of linear plasmids of Neurospora crassa, Flammulina velutipes and Kluyveromyces lactis, respectively, and 30, 29, 30 and 24 % identity with those of phages M2, PZA and PRD1, respectively). All these DNA polymerases belong to the B-type family, which is characterized by three main conserved domains thought to be implicated in nucleotide binding (Meijer et al., 2001). This analysis also revealed that the pGIL01 putative polymerase contained the three conserved domains described for this family of DNA polymerases (Fig. 2a).



View larger version (68K):
[in this window]
[in a new window]
 
Fig. 2. (a) Comparison between the potential polypeptide encoded by pGIL01 ORF5 and various DNA polymerases of the B-type family. The sequences are from phages PZA, {pi}29, M2 and PRD1, and plasmids maranhar (Neurospora crassa), pGKL2 (Kluyveromyces lactis) and pFV (Flammulina velutipes). The three conserved domains typical of the B-type DNA polymerase are also observed in ORF5. Specific residues involved in the protein-priming replication process are indicated by asterisks. Amino acids residues of the pGIL01 polymerase conserved in at least two other sequences are indicated by black boxes. (b) Comparison of the pGIL01 putative protein encoded by ORF13 with the DNA-packaging protein P9 of PR4 and PRD1. Numbers in front of the sequences refer to the protein coordinates. Residues of the ATP-binding consensus motif (GxxGxGKxxxxxxxL) are indicated by asterisks.

 
ORF6 is related to the N-terminal part of several LexA-like repressors (35–38 % identity), which negatively regulate various bacterial genes involved in DNA repair and synthesis, or repress lytic functions in phages. The N-terminal domain of these repressors is thought to be implicated in the DNA recognition (Dubnau & Lovett, 2002).

ORF13 displays 23 % identity (Fig. 2b) to the DNA-packaging protein of PR4 and PRD1 (P9 protein), two closely related bacteriophages belonging to the Tectiviridae family. These proteins are necessary for filling new virus particles with DNA (Mindich et al., 1982).

The amino acid sequences of ORF25 and ORF30 are related to various lytic enzymes (muramidase and autolysin precursors, and N-acetylmuramidases) found in bacteria (B. thuringiensis, Lactobacillus plantarum, Streptococcus mutans, Streptococcus pyogenes and Clostridium tetani) and phages (bacteriophage Bastille, bacteriophage phi-gle and bacteriophage phi-105) (32–38 % identity for ORF25 and up to 53 % identity for ORF30). The similarity is mainly observed in the N-terminal region of the enzymes, which, most probably, contains the active site (Buist et al., 1995). In bacteria, these lytic enzymes are implicated in peptidoglycan recycling, cell separation, formation of flagella, or sporulation (Buist et al., 1995; Heidrich et al., 2001). They are also found in several phages where they allow the phage particle to enter or to escape the cell through (partial) wall lysis (Young et al., 2000).

Interestingly, ORF1 showed distant similarity (31–40 % identity) to putative excisionases (Mycobacterium tuberculosis, Corynebacterium glutamicum) and DNA-binding proteins (Mycobacterium leprae) reported in bacterial genomes (data not shown). Where documented, these proteins are known to help phage integrases in the process of prophage excision from the bacterial chromosome. However, in the case of pGIL01, no integrase-like ORF could be identified.

pGIL01 corresponds to the prophage state of a temperate phage, GIL01
Based on the pGIL01 sequence analysis indicating relevant homologies to different phage proteins, it was postulated that pGIL01 could correspond to the prophage form of a lysogenic phage. To verify the presence of particles of this putative phage (named GIL01), titration analysis was performed on strain GBJ002, a H-14 derivative cured of all plasmids, including pGIL01 (and therefore supposedly sensitive to GIL01 infections). As predicted, typical small and turbid plaques were observed on plate cultures of GBJ002, suggesting the presence of phage particles in the AND508 culture supernatant. Furthermore, no plaques were observed when strain AND508 was plated out, presumably because of phage immunity. The titre of phage suspensions, corresponding to the supernatant of an exponential culture of AND508, was estimated to be 8x105 p.f.u. ml-1.

To identify these virus particles, PCRs, using primers specific to pGIL01, were performed on individual plaques, as well as on the GBJ002 bacterial lawn. A 1·4 kb fragment was observed for both PCRs performed on the phage plaques and on the AND508 DNA preparation. However, the GBJ002 DNA preparation did not produce any PCR product (data not shown).

Finally, potential GBJ002 lysogens were isolated from the turbid plaques and tested for the presence of the pGIL01 prophage. When PCRs were performed on the DNA extracted from these bacteria, they showed the same 1·4 kb fragment, supporting the transfer of pGIL01 from strain AND508 to strain GBJ002 (data not shown).

The sensitivity of GIL01 to organic solvents was then investigated. GIL01 turned out to be very sensitive to chloroform. For a ratio of 25 : 1 (phage suspension versus solvent), all the GIL01 present in the phage suspension was inactivated. Conversely, ether had a more limited effect on GIL01: with a ratio of 6·25 : 1, less than 10 % of the phage particles were inactivated.

UV irradiation, mitomycin C and nalidixic acid induce GIL01 phages
Different DNA-damaging treatments, known to induce the lytic cycle of temperate phages, were tested on strain AND508 (see Methods). The addition of mitomycin C and nalidixic acid, at final concentrations of 1 and 40 µg ml-1, respectively, increased the GIL01 titre of a culture supernatant of AND508 by a factor of 102–103. GIL01 was also induced by UV irradiation since a 104-fold increase of the titre could be obtained with UV doses producing between 90 and 95 % cell lethality (data not shown). These experiments suggested the existence of a specific correlation between the bacterial SOS response and GIL01 induction.

pGIL01 and GIL01 DNA are protected by proteins at their 5'' extremities
Several Gram-positive linear phages (e.g. the {phi}29-like family) belong to the group of molecules called ‘invertrons’, which have a protein at their 5' extremities (Salas, 1991). These terminal proteins are essential for the replication process. Similar protein–DNA structures are also observed in adenoviruses, in the Tectiviridae virus PRD1 and in several bacterial and fungal linear plasmids. Experiments using proteases and exonucleases were therefore set up to characterize the structure of the GIL01 extremities.

Proteins present at the termini of a linear molecule generally prevent its migration through an electrophoresis gel. Their removal by proteolysis during DNA extraction is expected to restore its migration ability. The fact that GIL01 DNA migration was dependent on proteinase K treatment (Fig. 3, lane 2) suggested that proteins were bound to the DNA molecule. Furthermore, subsequent additions of proteinase K (0·01 and 0·1 mg ml-1) to GIL01 DNA preparations obtained without protease pretreatment restored its migration (Fig. 3, lanes 3 and 4).



View larger version (75K):
[in this window]
[in a new window]
 
Fig. 3. Effect of proteinase K on the migration of GIL01 DNA. Lanes: 1, smart ladder used as standard marker; 2, DNA preparation performed without proteinase K; 3, addition of 0·01 mg ml-1 proteinase K; 4, addition of 0·1 mg ml-1 proteinase K.

 
Linear DNA molecules with their extremities linked to proteins are insensitive to specific exonuclease digestion, even after proteinase K treatment (Polo et al., 1998). Exonuclease III catalyses the removal of 5' mononucleotides from the 3'-hydroxyl termini of dsDNA, whereas lambda exonuclease is a 5' exonuclease which attacks dsDNA in the 5' to 3' direction. Proteinase K-treated GIL01 DNA was incubated with both exonucleases (Fig. 4). This experiment showed that the GIL01 DNA is degraded by exonuclease III, whereas no significant decrease is observed when lambda exonuclease is used, suggesting a protection at the 5' termini. The same protease and nuclease experiments were also performed on the prophage DNA (obtained by plasmid extraction), and gave the same results (data not shown). They strongly suggested that a terminal protein tightly binds the 5' extremities of the pGIL01 and GIL01 genomes, presumably via a covalent link.



View larger version (51K):
[in this window]
[in a new window]
 
Fig. 4. Exonuclease treatments on GIL01 DNA. (a) Exonuclease III. Lane 2, GIL01 incubated with exonuclease III buffer for 90 min. Lanes 3–5, GIL01 DNA incubated with exonuclease III for 10, 30 and 90 min, respectively. (b) Lambda exonuclease. Lane 2, GIL01 incubated with lambda exonuclease buffer for 90 min. Lanes 3–5, GIL01 DNA incubated with lambda exonuclease for 10, 30 and 90 min, respectively. Lane 1, smart ladder used as size standard.

 
pGIL01-related molecules are present in other B. cereus sensu lato strains
One hundred and eight strains, representative of five members of the B. cereus sensu lato group (51 B. cereus sensu stricto, 16 B. mycoides, 5 B. pseudomycoides, 32 B. thuringiensis and 5 B. weihenstephanensis) were analysed by dot-blotting for the presence of pGIL01-like DNA. This preliminary analysis revealed that pGIL01-related molecules apparently occurred in about 10 % (12/108) of the strains examined (data not shown).

To complete this analysis, a subset of five positive strains was further investigated by PCR, using primer pairs defining a series of overlapping sections of pGIL01. However, some of the pGIL01-related PCR fragments obtained for strains B16, B23, DBT012, BGSC4D14 and Bt5 displayed different sizes than their corresponding pGIL01 fragments (data not shown), suggesting that these molecules were slightly different from pGIL01. To gather more insights into these pGIL01-related molecules, additional experiments were undertaken on strain B16. This strain supernatant produced small and turbid plaques when tested on the GIL01-sensitive strain GBJ002, and PCR analysis confirmed that the corresponding viral particles, named GIL16, were GIL01-related. Comparative RFLP analyses were undertaken on the B16 and AND508 phages using DNA preparations from phage mutants producing clear plaques (named cpGIL16 and cpGIL01, respectively). As shown in Fig. 5, although both phage DNA had a similar size (lanes 1 and 2), their restriction patterns indicated the presence of important variations. Whereas the restriction profile of the GIL01 DNA perfectly matched that deduced from the pGIL01 sequence, that of GIL16 turned out to be different.



View larger version (56K):
[in this window]
[in a new window]
 
Fig. 5. RFLP analysis of phage DNA. The DNA preparations of cpGIL01 (lane 1) and cpGIL16 (lane 2) were cleaved by EcoRI (lanes 4 and 5), by EcoRV (lanes 7 and 8) and by MluI (lanes 10 and 11). Lanes 3, 6, 9 and 12, smart ladder as size standard.

 
Two other B. thuringiensis H14-derived strains (T14035 and T14477), not tested by dot-blotting, were analysed by PCR. All the amplified fragments had the same size as for AND508, indicating the presence of a molecule closely related to pGIL01 in both strains.

Linear molecules of apparently the same size as pGIL01 were also reported in the type strain of B. cereus (ATCC 14579T) (Carlson et al., 1992) and in strain SCE2 of Paenibacillus polymyxa (Rosado & Seldin, 1993). However, hybridization, using pGIL01 as probe, performed on these strains displayed no cross-reaction between these plasmids, suggesting that they are not closely related (data not shown).

GIL01 host range
The activity of GIL01 was tested on 44 strains of the B. cereus sensu lato group (30 B. cereus sensu stricto, 3 B. thuringiensis, 3 B. weihenstephanensis, 7 B. mycoides and 1 B. pseudomycoides) which did not harbour pGIL01-related elements, as indicated by hybridization and PCR experiments. Surprisingly, none of the strains tested was sensitive to GIL01, and, under these experimental conditions at least, its host range seemed to be restricted to H14-derived strains cured of pGIL01.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The first account of a linear DNA molecule in B. thuringiensis dates back to 1984. Working on strain H14 of B. thuringiensis sv. israelensis, González & Carlton (1984) reported the presence of a small DNA molecule with migration features of a linear plasmid. The present study revealed that this linear DNA molecule, named pGIL01, actually corresponds to the prophage form of a temperate phage, called GIL01. Bacteriophage GIL01 possesses a 14 931 bp linear dsDNA delineated by 73 bp terminal inverted repeats with proteins linked to its 5' extremities. The phage particles are sensitive to organic solvents, suggesting the presence of lipid components in the capsid. These specific features are reminiscent of those of PRD1, a Tectiviridae phage found in Gram-negative bacteria (Bamford et al., 1995).

One of the most notable characteristics of tectiviruses resides in their double-layer capsid: the dsDNA is located within a lipid-containing membrane vesicle covered by a rigid protein capsid (Bamford & Mindich, 1982). This lipid layer, derived from the host phospholipid pool (Davis et al., 1982), plays a key role during infection. It is transformed into a tubular structure that penetrates the host cell, allowing the injection of the viral genome (Bamford & Mindich, 1982; Grahn et al., 2002).

Although members of the Tectiviridae family are found in both Gram-negative (PRD1, PR3, PR4, PR5, PR722 and L17) and Gram-positive hosts (AP50 and Bam35, see below), the best-studied members are those infecting Gram-negative bacteria. These phages are very closely related to each other and PRD1 is considered as the family model (Bamford et al., 1981). PRD1 can infect a large variety of Gram-negative hosts, including E. coli and Salmonella typhimurium, harbouring conjugative plasmids of the P, N and W incompatibility groups, which code for the phage receptor (Bradley & Rutherford, 1975). The genome of PRD1 itself consists of a 14 925 bp linear dsDNA protected by proteins covalently attached at the 5' ends (Bamford et al., 1983). As for adenoviruses (Challberg et al., 1980) and the Bacillus phage {phi}29 family (Meijer et al., 2001), the DNA polymerase catalyses the formation of a covalent bond between the first 5' nucleotide and the OH group of a specific amino acid of the terminal protein, which will then serve as primer (de Jong & van der Vliet, 1999; Salas, 1991).

In Bacillus spp., two potential Tectiviridae phages have been reported. Phage AP50 (Nagy, 1974) and Bam35 (Ackermann et al., 1978) were isolated from B. anthracis and from strain Nr.35 of B. thuringiensis sv. alesti, respectively. These bacteriophages were only partially characterized. While AP50 produced turbid plaques and had an outer diameter of 80 nm, Bam35 made clear plaques and measured 63 nm in diameter. They were both sensitive to lipid solvents, such as chloroform and ether, and were highly resistant to UV.

pGIL01/GIL01 harboured six ORFs showing significant similarity to known phage proteins. The 735 residues of ORF5 showed homology to several DNA polymerases of the B-type family requiring the terminal protein as primer to initiate the replication process. Analysis of the pGIL01 nucleotide sequence revealed that, in addition to the presence of the three regions conserved in the B-type family, this putative polymerase also contains the specific residues involved in the protein-priming mechanism (Fig. 2a) (Meijer et al., 2001), suggesting that its replication process may occur via a similar mechanism to those of the PRD1 and {phi}29 phages.

By analogy to the genetic organization of PRD1, {phi}29 and other linear molecules, ORF4, located upstream of the DNA polymerase gene, is a good candidate for encoding the terminal protein of pGIL01. Moreover, the size of this potential protein (245 aa) is similar to that of the PRD1 and {phi}29 terminal protein (259 and 266 aa, respectively). However, ORF4 did not share any significant homology with any other terminal protein. Experimental evidence is thus required to ascertain this putative function.

As in the case of other lysogenic phages, it is likely that pGIL01 ensures its cell immunity by expressing a phage repressor during its lysogenic cycle. The sequence derived from ORF6 showed homology to several LexA repressors. In the case of UV irradiation or other DNA-damaging treatments, RecA acts as a co-protease and cleaves the LexA repressor, resulting in acceleration of the replication and repair process. It has been shown that, for several prophages, the induction mechanism appears to be mediated by DNA-damaging treatments, in the same way as for the bacterial SOS response. A very similar SOS regulatory network to that of E. coli has been identified for the Bacillus group. In Bacillus subtilis, for instance, the induction of the SOS responses is dependent on the coordinate expression of more than 20 din (damage inducible) genes (Dubnau & Lovett, 2002). In the case of GIL01, experiments using mitomycin C, nalidixic acid and UV irradiation caused an increase in the phage titre by 50- to 1000-fold. Whether or not ORF6 corresponds to the bona fide phage repressor participating in the bacterial SOS induction remains to be confirmed.

ORFs 25 and 30 displayed homologies to different peptidoglycan lytic enzymes. For PRD1, two lytic enzymes have been reported. Protein P15, a 1,4-{beta}-N-acetylmuramidase, which is connected to the phage membrane, causes the host cell lysis and the release of the viruses (Caldentey et al., 1994). Protein P7 carries a conserved transglycosylase domain and is the main muralytic enzyme involved in PRD1 entry. However, P7 is not absolutely essential for infectivity, probably due to the presence of another lytic enzyme (Grahn et al., 2002; Rydman & Bamford, 2002). In the case of phage {phi}29, the membrane protein holin introduces pores in the cell membrane to permeabilize it, allowing the peptidoglycan hydrolase to degrade the cell wall (Meijer et al., 2001). In other cases where no muralytic enzyme is encoded, the lytic enzyme acts as a cell wall antibiotic inhibiting the synthesis of the peptidoglycan (Bernhardt et al., 2001). In GIL01, no ORF sequences displayed similarity to known holin proteins, nor to inhibitors of peptidoglycan synthesis.

ORF13 shared similarity to the P9 protein of PRD1, which corresponds to the DNA-packaging protein implicated in maturation of PRD1 particles. The presence of an ATP-binding consensus motif (GxxGxGKxxxxxxxL) on gene IX of PRD1, which is also observed on ORF13 of GIL01, suggests that protein P9 may be the packaging ATPase providing the energy required for DNA encapsidation (Bamford et al., 1991; Mindich et al., 1982).

The role of the putative excisionase (ORF1) in the translocation and/or structural rearrangements of the pGIL01/GIL01 genomes remains very speculative. Indeed, no site-specific recombinase (tyrosine or serine recombinases) could be identified in pGIL01. Such proteins are generally associated with excisionases in other plasmid or phage integration/excision processes.

GIL01 exhibited an extremely narrow host specificity since it seemed to be restricted to strains of serotype H14 of B. thuringiensis cured of pGIL01, such as GBJ002. Despite this GIL01-limited host range, pGIL01 variants were observed in five other B. cereus sensu lato strains (B16, B23, DBT012, BGSC4D14 and Bt5). It is therefore plausible that pGIL01 used alternative ways for transfer.

The PRD1 host spectrum depends on the presence in its Gram-negative hosts of conjugative plasmids from the P, N and W incompatibility groups (such as the broad-host-range IncP-type RP4 plasmid). The plasmid functions are only required in the early stage of infection, the phage DNA entry. In fact, the plasmid protein complex involved in the mating pair formation (Mpf, a type IV secretion system) is a membrane-associated structure facilitating the DNA transfer through the membranes into the recipient cell (Grahn et al., 2000). It remains to be seen whether GIL01 also takes advantage of a Gram-positive cell wall- and/or a membrane-associated complex involved in genetic transfer or bacterial pathogenesis.

No hybridization between pGIL01 and the linear molecules present in the type strain of B. cereus sensu stricto (ATCC 14579T) (Carlson et al., 1992) and in P. polymyxa (Rosado & Seldin, 1993) could be observed. Although the nature of these extrachromosomal DNA remains uncertain, it is possible that they also belong to the Tectiviridae family, which would further extend the variation spectrum of this interesting family of bacteriophages. Finally, it is also important to note that, in addition to pGIL01, the coliphage N15 is the only prophage existing as an autonomous linear plasmid reported so far (Rybchin & Svarchevsky, 1999). However, the occurrence of extrachromosomal linear molecules has been systematically underestimated due to the use of inappropriate protocols for plasmid DNA extraction. A more thorough search for linear extrachromosomal molecules might provide new linear prophage candidates, both in Gram-negative and Gram-positive bacteria.


   ACKNOWLEDGEMENTS
 
We are grateful to Axelle Loriot and Murielle Herman for their contribution and support in the early stages of this research. We are deeply indebted to Lars Andrup for his comments on the manuscript and for his continual encouragement. We would also like to thank Laurence Van Melderen and Geneviève Maenhaut-Michel for their advice on the phage induction experiments. This work was supported by grants from the National Fund for Scientific Research (FNRS, Belgium) from which J. M. was a Research Associate. C. V. holds a research fellowship from FRIA (Fonds pour la Formation à la Recherche dans l'Industrie et dans l'Agriculture). Finally, we are also grateful to the EU-concerted action on ‘Mobile elements' contribution to bacterial adaptability and diversity’ (MECBAD network) for its financial support and for the incentive provided by its members.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Ackermann, H. W. & Smirnoff, W. A. (1978). Study of lysogeny in Bacillus thuringiensis and B. cereus. Can J Microbiol 24, 818–826 (in French).[Medline]

Ackermann, H. W., Roy, R., Martin, M., Murthy, M. R. & Smirnoff, W. A. (1978). Partial characterization of a cubic Bacillus phage. Can J Microbiol 24, 986–993.[Medline]

Andrup, L., Damgaard, J. & Wassermann, K. (1993). Mobilization of small plasmids in Bacillus thuringiensis subsp. israelensis is accompanied by specific aggregation. J Bacteriol 175, 6530–6536.[Abstract]

Andrup, L., Jensen, G. B., Wilcks, A., Smidt, L., Hoflack, L. & Mahillon, J. (2003). The patchwork nature of rolling-circle plasmids: comparison of six plasmids from two distinct Bacillus thuringiensis serotypes. Plasmid 49, 205–232.

Ash, C., Farrow, J. A. E., Dorsch, M., Stackebrandt, E. & Collins, M. D. (1991). Comparative analysis of Bacillus anthracis, Bacillus cereus, and related species on the basis of reverse transcriptase sequencing of 16S rRNA. Int J Syst Bacteriol 41, 343–346.[Abstract]

Bamford, D. & Mindich, L. (1982). Structure of the lipid-containing bacteriophage PRD1: disruption of wild-type and nonsense mutant phage particles with guanidine hydrochloride. J Virol 44, 1031–1038.[Medline]

Bamford, D., McGraw, T., MacKenzie, G. & Mindich, L. (1983). Identification of a protein bound to the termini of bacteriophage PRD1 DNA. J Virol 47, 311–316.[Medline]

Bamford, D. H., Rouhiainen, L., Takkinen, K. & Soderlund, H. (1981). Comparison of the lipid-containing bacteriophages PRD1, PR3, PR4, PR5 and L17. J Gen Virol 57, 365–373.[Abstract]

Bamford, D. H., Caldentey, J. & Bamford, J. K. (1995). Bacteriophage PRD1: a broad host range DSDNA tectivirus with an internal membrane. Adv Virus Res 45, 281–319.[Medline]

Bamford, J. K., Hänninen, A. L., Pakula, T. M., Ojala, P. M., Kalkkinen, N., Frilander, M. & Bamford, D. H. (1991). Genome organization of membrane-containing bacteriophage PRD1. Virology 183, 658–676.[Medline]

Bell, J. A. & Friedman, S. B. (1994). Genetic structure and diversity within local populations of Bacillus mycoides. Evolution 48, 1698–1714.

Bernhardt, T. G., Wang, I. N., Struck, D. K. & Young, R. (2001). A protein antibiotic in the phage Q{beta} virion: diversity in lysis targets. Science 292, 2326–2329.[Abstract/Free Full Text]

Berry, C., O'Neil, S., Ben-Dov, E. & 7 other authors (2002). Complete sequence and organization of pBtoxis, the toxin-coding plasmid of Bacillus thuringiensis subsp. israelensis. Appl Environ Microbiol 68, 5082–5095.[Abstract/Free Full Text]

Bradley, D. E. & Rutherford, E. L. (1975). Basic characterization of a lipid-containing bacteriophage specific for plasmids of the P, N, and W compatibility groups. Can J Microbiol 21, 152–163.[Medline]

Buist, G., Kok, J., Leenhouts, K. J., Dabrowska, M., Venema, G. & Haandrikman, A. J. (1995). Molecular cloning and nucleotide sequence of the gene encoding the major peptidoglycan hydrolase of Lactococcus lactis, a muramidase needed for cell separation. J Bacteriol 177, 1554–1563.[Abstract]

Caldentey, J., Hänninen, A. L. & Bamford, D. H. (1994). Gene XV of bacteriophage PRD1 encodes a lytic enzyme with muramidase activity. Eur J Biochem 225, 341–346.[Abstract]

Carlson, C. R., Grønstad, A. & Kolstø, A. B. (1992). Physical maps of the genomes of three Bacillus cereus strains. J Bacteriol 174, 3750–3756.[Abstract]

Challberg, M. D., Desiderio, S. V. & Kelly, T. J., Jr (1980). Adenovirus DNA replication in vitro: characterization of a protein covalently linked to nascent DNA strands. Proc Natl Acad Sci U S A 77, 5105–5109.[Abstract]

Colasito, D. J. & Rogoff, M. H. (1969a). Characterization of temperate bacteriophages of Bacillus thuringiensis. J Gen Virol 5, 275–281.[Medline]

Colasito, D. J. & Rogoff, M. H. (1969b). Characterization of lytic bacteriophages of Bacillus thuringiensis. J Gen Virol 5, 267–274.[Medline]

Davis, T. N., Muller, E. D. & Cronan, J. E., Jr (1982). The virion of the lipid-containing bacteriophage PR4. Virology 120, 287–306.[Medline]

de Jong, R. N. & van der Vliet, P. C. (1999). Mechanism of DNA replication in eukaryotic cells: cellular host factors stimulating adenovirus DNA replication. Gene 236, 1–12.[CrossRef][Medline]

Dubnau, D. & Lovett, M. J. (2002). Transformation and recombination. In Bacillus subtilis and its Closest Relatives: from Genes to Cells, pp. 453–471. Edited by A. L. Sonenshein, J. A. Hoch & R. Losick. Washington, DC: American Society for Microbiology.

Glare, T. R. & O'Callaghan, M. (2000). Bacillus thuringiensis: Biology, Ecology and Safety. Chichester: Wiley.

González, J. M., Jr & Carlton, B. C. (1984). A large transmissible plasmid is required for crystal toxin production in Bacillus thuringiensis variety israelensis. Plasmid 11, 28–38.[Medline]

Grahn, A. M., Haase, J., Bamford, D. H. & Lanka, E. (2000). Components of the RP4 conjugative transfer apparatus form an envelope structure bridging inner and outer membranes of donor cells: implications for related macromolecule transport systems. J Bacteriol 182, 1564–1574.[Abstract/Free Full Text]

Grahn, A. M., Daugelavicius, R. & Bamford, D. H. (2002). Sequential model of phage PRD1 DNA delivery: active involvement of the viral membrane. Mol Microbiol 46, 1199–1209.[CrossRef][Medline]

Granum, P. E. & Lund, T. (1997). Bacillus cereus and its food poisoning toxins. FEMS Microbiol Lett 157, 223–228.[CrossRef][Medline]

Heidrich, C., Templin, M. F., Ursinus, A., Merdanovic, M., Berger, J., Schwarz, H., de Pedro, M. A. & Höltje, J. V. (2001). Involvement of N-acetylmuramyl-L-alanine amidases in cell separation and antibiotic-induced autolysis of Escherichia coli. Mol Microbiol 41, 167–178.[CrossRef][Medline]

Helgason, E., Økstad, O. A., Caugant, D. A., Johansen, H. A., Fouet, A., Mock, M., Hegna, I. & Kolstø, A. B. (2000). Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis – one species on the basis of genetic evidence. Appl Environ Microbiol 66, 2627–2630.[Abstract/Free Full Text]

Jensen, G. B., Wilcks, A., Petersen, S. S., Damgaard, J., Baum, J. A. & Andrup, L. (1995). The genetic basis of the aggregation system in Bacillus thuringiensis subsp. israelensis is located on the large conjugative plasmid pXO16. J Bacteriol 177, 2914–2917.[Abstract]

Jensen, G. B., Andrup, L., Wilcks, A., Smidt, L. & Poulsen, O. M. (1996). The aggregation-mediated conjugation system of Bacillus thuringiensis subsp. israelensis: host range and kinetics of transfer. Curr Microbiol 33, 228–236.[CrossRef][Medline]

Kanda, K., Kitajima, Y., Moriyama, Y., Kato, F. & Murata, A. (1998). Association of plasmid integrative J7W-1 prophage with Bacillus thuringiensis strains. Acta Virol 42, 315–318.[Medline]

Lechner, S., Mayr, R., Francis, K. P., Prüß, B. M., Kaplan, T., Wießner-Gunkel, E., Stewart, G. S. A. B. & Scherer, S. (1998). Bacillus weihenstephanensis sp. nov. is a new psychrotolerant species of the Bacillus cereus group. Int J Syst Bacteriol 48, 1373–1382.[Abstract/Free Full Text]

Léonard, C., Zekri, O. & Mahillon, J. (1998). Integrated physical and genetic mapping of Bacillus cereus and other gram-positive bacteria based on IS231A transposition vectors. Infect Immun 66, 2163–2169.[Abstract/Free Full Text]

Madsen, S. M., Andrup, L. & Boe, L. (1993). Fine mapping and DNA sequence of replication functions of Bacillus thuringiensis plasmid pTX14-3. Plasmid 30, 119–130.[CrossRef][Medline]

Meijer, W. J., Horcajadas, J. A. & Salas, M. (2001). Phi29 family of phages. Microbiol Mol Biol Rev 65, 261–287.[Abstract/Free Full Text]

Mindich, L., Bamford, D., McGraw, T. & Mackenzie, G. (1982). Assembly of bacteriophage PRD1: particle formation with wild-type and mutant viruses. J Virol 44, 1021–1030.[Medline]

Nagy, E. (1974). A highly specific phage attacking Bacillus anthracis strain Sterne. Acta Microbiol Acad Sci Hung 21, 257–263.[Medline]

Nakamura, L. K. (1998). Bacillus pseudomycoides sp. nov. Int J Syst Bacteriol 48, 1031–1035.[Abstract/Free Full Text]

Polo, S., Guerini, O., Sosio, M. & Dehò, G. (1998). Identification of two linear plasmids in the actinomycete Planobispora rosea. Microbiology 144, 2819–2825.[Abstract]

Rosado, A. & Seldin, L. (1993). Isolation and partial characterization of a new linear DNA plasmid isolated from Bacillus polymyxa SCE2. J Gen Microbiol 139, 1277–1282.

Rybchin, V. N. & Svarchevsky, A. N. (1999). The plasmid prophage N15: a linear DNA with covalently closed ends. Mol Microbiol 33, 895–903.[CrossRef][Medline]

Rydman, P. S. & Bamford, D. H. (2002). The lytic enzyme of bacteriophage PRD1 is associated with the viral membrane. J Bacteriol 184, 104–110.[Abstract/Free Full Text]

Salas, M. (1991). Protein-priming of DNA replication. Annu Rev Biochem 60, 39–71.[CrossRef][Medline]

Sambrook, J. & Russell, D. (2001). Molecular Cloning: a Laboratory Manual, 3rd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Santos, M. A. (1991). An improved method for the small scale preparation of bacteriophage DNA based on phage precipitation by zinc chloride. Nucleic Acids Res 19, 5442.[Medline]

Smith, N. R., Gibson, T., Gordon, R. E. & Sneath, P. H. A. (1964). Type cultures and proposed neotype cultures of some species in the genus Bacillus. J Gen Microbiol 34, 269–272.[Medline]

Young, I., Wang, I. & Roof, W. D. (2000). Phages will out: strategies of host cell lysis. Trends Microbiol 8, 120–128.[CrossRef][Medline]

Received 18 February 2003; revised 11 April 2003; accepted 28 April 2003.