Insertion Site Occupancy by stx₂ Bacteriophages Depends on the Locus Availability of the Host Strain Chromosome

Abstract

Shiga toxin-producing Escherichia coli (STEC) is an emergent pathogen characterized by the expression of Shiga toxins, which are encoded in the genomes of lambdoid phages. These phages are infectious for other members of the Enterobacteriaceae and establish lysogeny when they integrate into the host chromosome. Five insertion sites, used mainly by these prophages, have been described to date. In the present study, the insertion of stx₂ prophages in these sites was analyzed in 168 STEC strains isolated from cattle. Additionally, insertion sites were determined for stx₂ phages which (i) converted diverse laboratory host strains, (ii) coexisted with another stx₂ prophage, and (iii) infected a recombinant host strain lacking the most commonly used insertion site. Results show that depending on the host strain, phages preferentially use one insertion site. For the most part, yehV is occupied in STEC strains while wrbA is preferentially selected by the same stx phages in E. coli laboratory strains. If this primary insertion site is unavailable, then a secondary insertion site is selected. It can be concluded that insertion site occupancy by stx phages depends on the host strain and on the availability of the preferred locus in the host strain.

Shiga toxin-producing Escherichia coli (STEC) is an emergent pathogen causing bloody diarrhea and hemolytic uremic syndrome (7). This group comprises different serotypes; the best-known serotype, which was one of the earliest determined to belong to this group, is O157:H7 (25). Cattle are a major reservoir for STEC. Ruminant animals such as cattle, sheep, and goats are naturally colonized with STEC and release these organisms into the environment with their feces (6, 35).

One of the main virulence factors in STEC is the expression of Shiga toxins (Stx). To date, two Shiga toxins, Stx₁ and Stx₂, and several variants (15, 37) have been described. These toxins are encoded in the genomes of lambdoid phages, which establish lysogeny in STEC bacteria (13). stx bacteriophages can convert different E. coli strains, as well as some other members of the Enterobacteriaceae, into Stx-producing bacteria (1, 2, 13, 19, 29, 34).

A successful lysogenic event can occur only if the phage genome is integrated into the bacterial chromosome. To achieve this, phage integrase, which is regulated by gene products activated in the lysogenic pathway, and host recombinases (e.g., integration host factors), which are not regulated by any phage gene, must recognize homologous sequences present in the phage DNA and in the bacterial genome (attP and attB, respectively) (14, 26).

Several studies have attempted to characterize phage integrases (3, 16, 23, 26, 27, 33). Although different families of this enzyme have been established, most of them conserve certain motifs in their protein architecture that allow attachment site recognition and phage genome-bacterial genome recombination.

Shiga toxin prophages can convert the host strain because they integrate their genome by using specific insertion sites. To date, five target sites have been described: wrbA, which codes for a NADH: quinone oxidoreductase (21a). yehV, which codes for a transcriptional regulator (36); sbcB, which produces an exonuclease (21); yecE, whose function remains unknown (24); and Z2577, which encodes an oxidoreductase (17).

Most STEC strains tested appear to have more than one site occupied (5, 31) by a phage (carrying the stx gene or not). In addition, a considerable number of these strains, isolated from environmental or clinical samples, seem to carry more than one stx gene copy (2, 12, 20, 32). As the origin of Stx must be an stx phage, the existence of more than one stx gene suggests that these STEC strains carry more than one functional or defective stx prophage. This means that these strains should have more than one insertion site occupied by different prophages (5, 31).

The gene acquisition and exchange mediated by stx phages constitutes a mechanism that not only increases the pathogenicity of STEC but also plays a critical role in bacterial evolution (8). In this study, we attempted to evaluate whether the insertion of stx phages in a given chromosomal site is highly specific. To this end, the insertion site occupancy in numerous STEC strains was evaluated, and then the insertion site occupancy by stx phages in single and double lysogens was examined.

MATERIALS AND METHODS

stx₂ bacteriophages and bacterial strains.

Phages φA9, φA75, φ312, φ534, φ549, φ557, and φVTB55 were induced from STEC strains isolated from cattle (19) and were used to generate single lysogens. These phages, as well as phage 933W, were modified with a cat or a tet gene inserted into the stx gene (30). The following modified phages were used in the double lysogen experiments: φ3538, modified with a cat gene inserted into the stx gene (29), which thereby confers resistance to chloramphenicol (Cm), and φ24_B, modified with the aph gene inserted into the stx gene (2), which thereby confers resistance to kanamycin (Km).

E. coli strains DH5α and C600 and a clinical isolate of Shigella sonnei, strain 866 (19), were used as hosts for lysogen generation.

A total of 168 STEC strains producing Shiga toxin 2 were also analyzed for insertion site usage. Of these, 165 were isolated from cattle (151 from calves and cows and 14 from beef meat) belonging to separate herds, while 3 strains were isolated from sheep. Strains were isolated and characterized as STEC as previously described (19).

Plasmids.

Plasmid pKD46 (GenBank no. AY048746), used for the expression of Red recombinase, was used to insert DNA fragments into the E. coli chromosome (10). Plasmid pKD3 (GenBank no. AY048742) (10) was used to obtain the chloramphenicol acetyltransferase gene (cat), which confers resistance to Cm. Plasmid pACYC184 (GenBank no. X06403) was used to obtain the tetracycline (Tc) resistance gene (tet) (30). All vectors were purified with a QIAGEN plasmid Midi purification kit (QIAGEN Inc., Valencia, CA).

Media and antibiotics.

Trypticase soy broth with 5 mM CaCl₂ was used in the phage induction experiments. Bacterial strains were grown in Luria-Bertani (LB) broth and on LB agar. The LB medium was supplemented with Km (20 μg ml⁻¹), Cm (10 and 5 μg ml⁻¹), and Tc (5 μg ml⁻¹), when required.

PCR techniques.

PCRs were performed with a GeneAmp PCR 2400 system (Perkin-Elmer, PE Applied Biosystems, Barcelona, Spain.). The DNA template was prepared directly from two colonies of each strain suspended in 50 μl of double-distilled water and heated to 96°C for 10 min prior to addition of the reaction mixture. Purified bacterial or phage DNA was diluted 1:20 in double-distilled water. The oligonucleotides used in this study are described in Table 1.

TABLE 1.

Oligonucleotides used in this study

Oligonucleotide	Sequencea	Gene	Length (bp)	Reference
SbcB1	CATGATCTGTTGCCACTCG	sbcB	1,768	21
SbcB2	AGGTCTGTCCGTTTCCACTC
YecE-F	CGAAGACGCCTGTAGTGCC	yecE	1,371	24
YecE-R	CGCAGGGAGAAAACCTC
wrbA2-lp	CGAATCGCTACGGAATAGAGA	wrbA left junction	506	31
wrbA2-LJ	CCGACCTTTGTACGGATGTAA
wrbA1-up	AGGAAGGTACGCATTTGACC
wrbA1-RJ	ATCGTTCGCAAGAATCACAA	wrbA right junction	592
YehVB	AACAGATGTGTGGTGAGTGTCTG	yehV left junction	824	31
YehV-F-LJ	CACCGGAAGGACAATTCATC
YehVA	AAGTGGTGCTTTGTGAT
YehV-E-RJ	GATGCACAATAGGCACTACGC	yehV right junction	702
Z2577F	AACCCCATTGATGCTCAGGCTC	Z2577	909	17
Z2577R	TTCCCATTTTACACTTCCTCCG
wrbCm5	GAAGCAGCTCCAGCCTACACAACACCTACCCGCTTTGGCAAC	5′ wrbA fragment	314	This study
Wrb2up	TGGCGTGAACCGTCACCG
wrbCm3	CTAAGGAGGATATTCATATGGTACATGGAATAATAAAGCACC	3′ wrbA fragment	189	This study
Wrb2lp	ATTACGTGCTTGTGAAAAG
Extwrb	TTAGCCGTTAAGTTTAAC	External primer wrbA		This study
UP378	GCGTTTTGACCATCTTCGT	stx2 fragment	378	18
LP378	ACAGGAGCAGTTTCAGACAG
S2Aup	ATGAAGTGTATATTATTTA	stx2AB subunit	1,241	20; this study
S2Blp	TCAGTCATTATTAAACTGCACTT
Cm5	TGTGTAGGCTGGAGCTGCTTC	cat cassette	1,015	30
Cm3	CATATGAATATCCTCCTTAG
Tc5	TCAGCCCCATACGATATAAG	tet cassette	1,344	30
Tc3	TGGAGTGGTGAATCCGTTAG
Km5	GTCAGCGTAATGCTCTGC	aph cassette	897	This study
Km3	GTCTGCTTACATAAACAG
intup	TCATCCTCCTGGTCACTT	Phage 933W int	1,309	This study
intlp	TGCTCTTGGACGCAGGAG
Lambda1F	GTTACMGGGCARMGAGTHGG	λ integrase	378	3
Lambda1R	ATGCCCGAGAAGAYGTTGAGC
P21-2F	GTTACTGGWCARCGKTTAGG	P21 integrase	642	3
P31-2R	GATCATCATKRTAWCGRTCGGT
P27-4F	CTBGCMTGGGARGATATHGA	P27 integrase	392	3
P27-4R	GMCCAGCABGCATARGTRTG

^aBold type indicates primers used to amplify the left junction; italic type indicates primers used to amplify the right junction; underlining indicates the overlapping region of the PCR fragments used to generate DH5α Δwrb::cat.

For amplification of stx₂, stx₁, and integrase (int) genes, a normal PCR procedure was conducted. For insertion site analyses, two different PCR procedures were carried out.

Single PCR.

For single PCR, the described insertion locus was simply amplified with the specific primers (Table 1). If the amplification was positive, the locus was assumed to be intact. The positive amplicons were confirmed in some cases by PCR using primers to amplify the right and left sides of the phage genome-bacterial genome junctions.

If the PCR revealed no amplicon, it might indicate that a prophage was occupying that particular locus (the total length of the phage genome inserted between the two primers did not allow PCR amplification). This type of PCR was carried out for sbcB, Z2577, and yecE insertion sites (17, 21, 24, 31). To confirm that the lack of amplicon is caused by integration of the phage in these loci, the left and right junctions of the phage were amplified, combining the upper and lower primers of each gene. LJ and RJ primer pairs were used to amplify left and right attP junctions of wrbA and yehV, combined with sbc, yecE, and Z2577 upper and lower primers (Table 1). For example, we amplified using primers wrbA2LJ/SbcB1, wrbA2RJ/SbcB2, and so on.

Double PCR.

For double PCR, two amplifications were performed for the right and left junctions of the attachment between the phage and bacterial genomes (attP and attB, respectively). Positive amplification was assumed to mean that the phage occupied that locus. This was the procedure employed for wrbA and yehV loci (22, 31, 36) (Table 1).

Standard DNA techniques.

Chromosomal DNA was prepared from 40 ml cultures as previously described (19). Chromosomal DNA was digested with XbaI restriction endonuclease (Promega Co., Madison, WI). This endonuclease cuts only once in the genome of the stx phages used in this study (including 933W) and does not cut the stx, tet, or cat gene. Restriction fragments were analyzed by separation on 0.7% agarose gels in Tris-borate-EDTA buffer, stained with ethidium bromide, and transferred to a nylon membrane for Southern blotting. PCR products were purified by using a PCR purification kit (QIAGEN Inc., Valencia, CA).

Preparation of digoxigenin (DIG)-labeled stxA₂-, tet-, aph-, and cat-specific gene probes.

A DNA fragment of the stx_2A gene obtained with primers UP378/LP378 was labeled with DIG and used as a probe. Other probes were designed for the tet gene (1,340 bp), the aph gene (Km resistance gene, 897 bp), and the cat gene (1,015 bp). The resulting fragments represent the outcome of the amplification with the respective primers (Table 1).

The probes were labeled by incorporating DIG-11-deoxyuridine-triphosphate (Roche Diagnostics, Barcelona, Spain) during the PCR, as described by Muniesa et al. (19).

Hybridization techniques.

Colony hybridization and Southern blotting were performed as previously described (12). The membranes were washed in a solution consisting of 1% sodium dodecyl sulfate-2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) for 5 min at 50°C. They were then washed in 0.1% sodium dodecyl sulfate-2× SSC for 5 min at room temperature and finally in 2× SSC for 5 min at room temperature.

Membranes were prehybridized with standard prehybridization solution at 68°C for 2 h. Stringent hybridization was achieved with a DIG DNA labeling and detection kit (Roche Diagnostics, Barcelona, Spain), according to the manufacturer's instructions. The DIG-labeled probes were prepared as described above.

Electroporation.

Electrocompetent cells were prepared from 50 ml of culture in SOB medium (1 × SOB is 20 g liter⁻¹ tryptone, 5 g liter⁻¹ yeast extract, 0.5 g liter⁻¹ NaCl, 2.5 mM KCl, and 10 mM MgCl₂) with l-arabinose 2% and concentrated by centrifugation at 3,000 × g for 5 min. They were then washed in 4 ml of ice-cold double-distilled water. After four washing steps, the cells were suspended in 100 μl of ice-cold double-distilled water. The cells were mixed with the corresponding amount of DNA (plasmid or PCR amplified, approximately 0.5 μg) in an ice-cold microcentrifuge tube and transferred to a 0.2-cm electroporation cuvette (Bio-Rad, Inc.). The cells were electroporated at 2.5 kV with 25 F and 200 Ω resistance. After electroporation, 1 ml of SOC medium (1 × SOC is 20 g liter⁻¹ tryptone, 5 g liter⁻¹ yeast extract, 0.5 g liter⁻¹ NaCl, and 2.5 mM KCl) (28) was added to the cuvette. The cells were transferred to a 17- by 100-mm polypropylene tube and incubated in SOC medium for 1 to 4 h at either 30°C (for temperature-sensitive plasmids) or 37°C, without shaking. Cells were concentrated 10× from a 1-ml culture before plating on selective media.

Generation of single lysogens.

Laboratory lysogens were obtained with E. coli strains DH5α and C600 and S. sonnei strain 866. Most of the single lysogens were generated as follows: the suitable host strain was grown in LB broth to an optical density at 600 nm of 0.5; it was then mixed with 2.5 ml of LB soft agar (0.7% agar-agar) and poured onto LB agar plates. Phage suspension was obtained by adding mitomycin C (0.5 μg μl⁻¹) to the Trypticase soy broth lysogen culture with an optical density at 600 nm of 0.5. Ten μl of the phage suspension and 10-fold dilutions of this suspension were spotted onto the agar layer containing the laboratory strain and incubated at 37°C. After overnight incubation, colonies present in the spot area were regarded as possible lysogens and subcultured in LB agar plates. The presence of the stx₂ gene was confirmed by colony blotting and specific PCR (30) as described above.

Generation of double lysogens.

To generate double lysogens, the same stx phages were genetically modified to facilitate selection. To this end, a Tc (tet) or Cm (cat) resistance gene was placed within the stx₂ gene using the Red recombinase system (30). The modified phages Δstx₂::tet and Δstx₂::cat and phages φ3538 and φ24_B were used to infect single lysogens carrying the stx₂ prophages. Briefly, 0.3 ml of an overnight culture of the single lysogen grown in LB broth was infected with 0.5 ml of a suspension containing modified stx phages, obtained as described above. The mixture was then incubated in 5 ml of LB broth for 4 h at 37°C. One ml of this culture was used to inoculate 3 ml of LB with Cm, Km, or Tc and further incubated at 37°C for 24 to 48 h. Serial dilutions of this culture were plated onto LB agar with the respective antibiotic.

Single lysogens of modified phages, which were used as controls, were also prepared using this procedure.

In the case of double-lysogen generation, colonies suspected of being double lysogens were confirmed via colony blotting and PCR by using primers that allowed us to see different amplicon sizes (Fig. 1 and Table 1). In Fig. 1, primers S2Aup and S2Blp were used in a single PCR to amplify the two prophages present in the double lysogens. The antibiotic resistance cassette of the modified phages was placed within the stx gene, and the primers annealed at the beginning and the end of the stx gene. Both the stx phages and the cat or tet phages could then be amplified using the same primer pair, although due to the presence of the antibiotic resistance gene, the amplicon obtained for the modified phages was different from that obtained for the stx phages (1,241 bp for stx phages and 1,701 bp for tet phages or 1,526 for cat phages). Simultaneous amplification was performed with the double lysogens, and two bands, each corresponding to one phage, were observed (Fig. 1).

FIG. 1.

Confirmation of double lysogens (stx phage/Δstx₂::tet or stx phage/Δstx₂::cat) by single PCR. Lanes: 1, stx₂ control; 2, amplified fragment of the stx₂ gene truncated with the tet or cat gene, respectively; 3, double lysogens obtained with phage 933WTc or φ3538; 4 to 10, double lysogens obtained with phages φA9, φA75, φA312, φA534, φA549, φA557, and φVTB55 with the stx₂ gene truncated with the tet or cat gene, respectively.

The insertion site occupancy of phages present in single and double lysogens in the different host strains was evaluated by PCR, as described above.

Construction of an E. coli DH5α Δwrb cat mutant.

wrbA was the insertion site most commonly used by phages in E. coli DH5α. For that reason, a mutant of E. coli DH5α lacking the wrbA site, used by stx₂ phages to integrate within the host chromosome, was constructed. This mutant was constructed by inserting a cat gene in the wrbA site, using the Red recombinase method described by Datsenko and Wanner (10) (Fig. 2).

FIG. 2.

Schematic diagram indicating the positions of homologous recombination between the wrbA gene and the wrbA::cat fragment, generating E. coli DH5α ΔwrbA::cat via the Red recombinase system.

The primer pairs Wrb2up/wrbCm5 and Wrb2lp/wrbCm3 (Table 1) were used for 5′ and 3′ homologous-region amplifications (producing two fragments of 314 and 189 bp, respectively). Cm5 and Cm3 oligonucleotides were used to amplify the cat gene. Amplicons of the 5′ fragment, the cat gene, and the 3′ fragment were annealed at their overlapping region and used as templates in a new PCR with primers Wrb2up/Wrb2lp. The generated amplicon (1,477 bp) carried the cat gene and was flanked by the two regions homologous to the wrbA gene. This fragment was electroporated as described previously for E. coli DH5α carrying the plasmid pKD46 (10). Mutant colonies were selected in LB medium with Cm. The mutant was verified by PCR amplification using a primer placed in an external sequence of wrbA (primer Ext wrb) and primer Cm3 in the 3′ end of the cat gene. Further confirmation was done by sequencing the corresponding amplification product.

E. coli DH5α ΔwrbA::cat was lysogenized with the Δstx₂::tet phages to determine where these phages are inserted when the primary insertion site is not available. The mutant strain E. coli DH5α ΔwrbA::cat was lysogenized with the recombinant Δstx₂::tet phages as described previously but using an LB culture with Cm. The lysogens were selected with Tc and Cm for the modified phages and with Km and Cm for those generated with phage φ24_B. Phage φ3538 was not used to lysogenize the recombinant strain, since the phage and the mutant strain carried the same resistance gene.

Sequencing.

Sequencing was performed with the ABI PRISM Big Dye III.1 Terminator cycle sequencing Ready Reaction kit (Perkin-Elmer, PE Applied Biosystems, Barcelona, Spain) in an ABI Prism 3700 DNA analyzer (Perkin Elmer), according to the manufacturer's instructions. All sequencing was performed in duplicate.

Nucleotide sequence analysis searches for homologous DNA sequences in the EMBL and GenBank database libraries were carried out using Wisconsin Package version 10.2 (Genetics Computer Group, Madison, WI). WiscBlast analyses were performed with tools available on the National Institutes of Health web page: http://www.ncbi.nlm.nih.gov.

RESULTS

Insertion site occupancy in STEC.

Insertion site occupancy was determined in a total of 168 STEC strains; yehV was the most commonly occupied locus in serotype O157:H7 (Table 2). yehV is occupied by a prophage in 98% of the O157:H7 strains. Although distribution among the other serotypes could not be elucidated due to the lower number of strains for each serotype studied, yehV remains one of the most occupied loci in non-O157 strains, although in lower numbers, being occupied only in a 28% of the non-O157:H7 strains. Locus Z2577 appeared to be the most occupied site in the non-O157:H7 strains, being occupied in 38% of the strains. Forty STEC strains had more than one insertion site occupied, with 21 of these belonging to serotype O157:H7. Moreover, there were three strains, belonging to serotypes O113, O15, and ONT, that had three insertion sites occupied and one in which four insertion sites were used simultaneously (serotype O157:H7). Twenty-four strains had no insertion at any of the tested sites. Thus, either these strains carry a prophage at a different site that may not have been described or they are not lysogenic. Only one of these strains belongs to the O157:H7 serotype.

TABLE 2.

Insertion sites occupied in STEC strains of different serotypes used in this study

STEC serotype	No. of strains
	Total	Carrying stx1	With insertion locus					With more than one insertion site occupied	With no detectable insertion site
			wrbA	sbcB	yehV	yecE	Z2577
O157:H7	97	21	9	9	95	5	4	21	1
O6	2			1	1		2	2	0
O15	3		1			1	1	1	2
O116	3			1	1		1	0	0
O136	2				1	1	1	1	0
O165	2		1		2		2	2	0
O2	3		1			1	1	1	1
O22	2	1					2	0	0
O163	2							0	2
O88	2			1		1		1	1
ONT	3		1	2			2	2	1
O8	4				3			0	1
O162	3				2			0	1
O156	3		1		2		2	1	0
O174	3						2	0	1
O91	3	1			1		2	0	0
O171	2					2		0	0
O113	3		1		3		3	3	0
O39	2				1		2	1	0
Othera	24	1	5	3	3	1	4	4	13
Total	168	24	20	17	115	12	31	40	24

^aThe 24 strains each belong to a different serotype.

The occupancy of the sbcB, yecE and Z2577 sites in strains showing no amplicons was confirmed by amplification of the right and left junctions, as described above. Twenty percent of the O157:H7 strains showing positive amplicons in these genes were also tested.

Identification of the int gene.

We attempted to amplify the integrase gene present in the 10 phages used in this study, in order to correlate the integrase and the insertion loci used by the different phages. Amplification of the int gene using primers designed from the 933W sequence showed that phages 933W, φ24_B, and φVTB55 harbor an identical integrase gene, which was confirmed by sequencing the fragment. The remaining phages were not amplified using this primer pair.

To amplify the integrases of the remaining phages, we used degenerated primers, originally designed to amplify the catalytic domains of integrase of prophages found in the environment (3) (Table 1). Results showed that all phages harbored an integrase belonging to the tyrosine recombinase family. Integrases of phage φ549 are amplifies with primers designed for group 1 integrases (3), including phage λ. Phages φA9, φ534, and φ3538 harbor integrases amplified with primers of group 2 (3). All three integrases were identical in the sequenced fragment. Included in group 2 was the sequence of the integrase of prophage e14 (accession number U00096) (3), which is also identical to three phages in the amplified fragment. Finally, phages φ312, φ557, and φA75 harbor integrases amplified with primers of group 4 (3), which also includes phage 933W. However, the integrases of phages φ312, φ557, and φA75 were similar, but not identical, to 933W integrase and were not amplified with the first set of primers designed from 933W sequence described above. The sequenced fragments of the integrases of phages φ312 and φ557 were identical to each other but different from the sequence obtained for the phage φA75 integrase.

Insertion site occupancy among different host strains.

In light of the previous results and to determine whether selection of a given insertion site was strain or phage dependent, the insertion site occupancy by a given stx₂ phage when converting different host strains was examined. To this end, 10 stx₂ phages were used to lysogenize two E. coli K-12 derived strains (DH5α and C600) as well as S. sonnei 866. The last strain was chosen since it yielded good results in terms of phage infectivity of bacterial cells (19) and lysogenization by stx phages (data not shown). All 10 phages used were heteroimmune, as indicated by their ability to infect the same strain (data not shown).

Prior to the lysogenization experiments, the presence of the five insertion sites in the host strains was evaluated by PCR and sequencing. In both E. coli strains, the five insertion loci remained accessible and unoccupied, while S. sonnei 866 lacked the wrbA attachment site.

To confirm that single lysogens carried only one copy of the genome of each phage type, we performed Southern blot hybridization with the chromosomal DNA of each lysogen restricted with XbaI. This endonuclease cuts only once in the genome of the stx phages used in this study and has no cutting sites in the stx gene. Only one band hybridized with the probe, indicating that only one copy of each phage was present in the genome of the lysogen.

Once the three host strains were lysogenized with the bacteriophages, the insertion sites occupied by the 10 phages were evaluated. Our results (Table 3) showed that the same phage is integrated into different insertion sites when it lysogenizes different host strains. A great majority of stx₂ phages occupied wrbA when infecting E. coli laboratory strains. In contrast, they occupied yehV when infecting S. sonnei 866 or the STEC strain from which the phage was induced. Since S. sonnei lacks the wrbA locus, these results suggest that the availability of the insertion site determines phage occupancy and that phages can occupy a secondary locus when the primary locus is not available. However, the question of what happens when a host strain is infected by two phages arises.

TABLE 3.

Insertions sites used by the same phages in different host strains^a

Phage	Insertion site used by:
	STECb	E. coli K-12		S. sonnei 866
		DH5α	C600
φA9	yehV	yecE	wrbA	yehV
φA75	yehV	wrbA	wrbA	yehV
φ312	yehV	wrbA	sbcB	yehV
φ534	yehV	wrbA	wrbA	yehV
φ549	yehV	wrbA	sbcB	yehV
φ557	yehV	wrbA	yehV	yehV
φVTB55	yehV	wrbA	yehV	yehV
933W	wrbA	wrbA	wrbA	—
φ3538	—	wrbA	wrbA	yehV
φ24B	—	wrbA	wrbA	—

^aEach result was confirmed after five independent replicates. —, not tested.

^bThe STEC strain is the original strain from which the phage was induced.

Use of insertion sites in double lysogens.

The next experiments addressed how phages integrate in double lysogens when, as can be seen in Table 3, the majority of these prophages preferentially occupy the same loci in the same host strain. Since the wrbA locus seemed to be among the most commonly occupied in E. coli strains, we generated double lysogens for phages that utilized the wrbA locus by using the recombinant Δstx₂::tet phages. This also allowed us to observe whether phages select a secondary site.

The presence of one copy of the two phage genomes in the double lysogens was confirmed by Southern blot hybridization of the chromosomal DNA restricted with XbaI and hybridized with the stx probe to detect the first phage and a cat, aph, or tet probe to detect the second phage. Results indicated that only one band hybridized with each probe, suggesting that only one copy of the stx phage genome and one copy of the modified phage genome were present in the double lysogens (Fig. 3).

FIG. 3.

Southern blot analysis of chromosomal DNA from single and double lysogens hybridized with the respective stx_2A and cat probes. (A) Southern blot of chromosomal DNA from C600(933W), C600(φ3538), and C600(933W/φ3538) digested with XbaI and hybridized with the stx probe for the detection of 933W (A) or with the cat probe for detection of φ3538 (B). Lanes 1, C600 (933W); lanes 2, C600(φ3538); lanes 3, C600(933W/φ3538).

In all cases, the presence of a second prophage resulted in the occupancy of a secondary insertion site (Table 4). If there is no competence, each phage occupies its preferred site [for example, C600(φ312/933WTc)]. Alternatively, two phages may occupy the same loci in their single lysogens; for instance, phages 933W and φ3538 used wrbA in their single lysogens, C600(933W) and C600(φ3538), respectively (Table 3). In that case, in the double lysogen E. coli C600(933W/φ3538), one prophage continually used wrbA while the other inserted itself in yehV. This same situation was seen in double lysogens obtained with phage φ3538 or phage φ24_B, as well as with the remaining phages. Moreover, for other loci, such as scbB and yecE, a similar situation occurs, and the second phage inserts in wrbA. The same results were obtained when two identical phages lysogenized the same strain (933W/933WTc). To sum up, one prophage used wrbA while the second primarily utilized yehV. The only exception was the double lysogen C600(φA9/φ24_B), in which Z2577 was the second site employed.

TABLE 4.

Insertion sites occupied by one phage in single lysogens and by two phages in double lysogens^a

Single lysogen	Primary insertion site	Second phage	Double lysogen	Secondary insertion site
C600(φA9)	wrbA	φ24B	C600(φA9/φ24B)	Z2577
DH5α(φ312)	wrbA	φ24B	DH5α(φ312/φ24B)	yehV
DH5α(φ534)	wrbA	φ24B	DH5α(φ534/φ24B)	yehV
DH5α(φ549)	wrbA	φ24B	DH5α(φ549/φ24B)	yehV
C600(933W)	wrbA	φ24B	C600(933W/φ24B)	yehV
C600(φA9)	wrbA	φ3538	C600(φA9/φ3538)	yehV
DH5α(φA75)	wrbA	φ3538	DH5α(φA75/φ3538)	yehV
DH5α(φ312)	wrbA	φ3538	DH5α(φ312/φ3538)	yehV
DH5α(φ534)	wrbA	φ3538	DH5α(φ534/φ3538)	yehV
DH5α(φ557)	wrbA	φ3538	DH5α(φ557/φ3538)	yehV
DH5α(φVTB55)	wrbA	φ3538	DH5α(φVTB55/φ3538)	yehV
C600(933W)	wrbA	φ3538	C600(933W/φ3538)	yehV
C600(φA9)	wrbA	933WTc	C600(φA9/933WTc)	yehV
DH5α(φA9)	yecE	933WTc	DH5α(φA9/933WTc)	wrbA
DH5α(φA75)	wrbA	933WTc	DH5α(φA75/933WTc)	yehV
DH5α(φ534)	wrbA	933WTc	DH5α(φ534/933WTc)	yehV
DH5α(933W)	wrbA	933WTc	DH5α(933W/933WTc)	sbcB
C600(933W)	wrbA	933WTc	C600(933W/933WTc)	yehV
C600(φ312)	sbcB	933WTc	C600(φ312/933WTc)	wrbA
C600(φ549)	sbcB	933WTc	C600(φ549/933WTc)	wrbA
C600(φ557)	yehV	933WTc	C600(φ557/933WTc)	wrbA
C600(φVTB55)	yehV	933WTc	C600(φVTB55/933WTc)	wrbA
C600(933W)	wrbA	φ312Tc	C600(933W/φ312)	yehV
	wrbA	φ534Tc	C600(933W/φ534Tc)	yehV
	wrbA	φ557Tc	C600(933W/φ557Tc)	yehV
	wrbA	φA9Tc	C600(933W/φA9Tc)	yehV
	wrbA	φ549Tc	C600(933W/φ549Tc)	yehV
	wrbA	φVTB55Tc	C600(933W/φVTB55Tc)	yehV
DH5α(φA9)	yecE	φA9Tc	DH5α(φA9/φA9Tc)	wrbA
DH5α(φ534)	wrbA	φ534Tc	DH5α(φ534/φ534Tc)	yehV
C600(φ312)	sbcB	φ549Tc	C600(φ312/φ549Tc)	wrbA

^aEach result obtained was confirmed after five independent replicates. The second modified phage was used to infect the single lysogen, and the double lysogen was screened for occupancy of the five insertion sites.

The sizes of the bands hybridizing with the stx probe in the Southern blots of chromosomal DNA from single and double lysogens were apparently the same, suggesting that the first phage does not change its integration site in the bacterial chromosome. Accordingly, the band corresponding to the second phage (hybridized with the cat, tet, or aph probe, respectively) was different between the single and the double lysogens, suggesting that the second phage used to lysogenize the double lysogens moved to another insertion locus (Fig. 3). As an example, Fig. 3 shows the Southern blot hybridization of chromosomal DNA of C600(933W), C600(φ3538), and the double lysogen C600(933W/φ3538). The size of the band detected by the stx probe in the single lysogen C600(933W) was approximately 25 kb and was apparently the same as that observed with C600(933W/φ3538). This indicated that the phage was in a similar location in the bacterial chromosome, the wrbA locus. However, the band detected with the cat probe in C600(φ3538) (25 kb) did not correspond to the band detected with the cat probe in C600(933W/φ3538) (27 kb.), suggesting that the second phage changed its previous insertion site (wrbA) for a new site, in this case yehV. This approach was done with several lysogens to confirm that the second phage occupied a new locus, and in all occasions the results were similar.

Use of secondary insertion sites depends on the genetic locus availability.

The conclusion that the availability of a given insertion locus appears to be the determining factor in its occupancy was based on two observations: (i) stx₂ phages seem to use different sites in different host strains, and (ii) when a locus is not available, the second phage must be inserted into a secondary insertion site. We designed a mutant lacking the most commonly used insertion site, wrbA (Table 3). This mutant, generated in E. coli DH5α, was used as a host for lysogenization with the phages.

Our results showed that the stx₂::tet phages were integrated into another insertion site (Table 5). The new integration site occupied in the mutant was generally the same as that observed in double lysogens, except in three cases: phages φA75, φ549, and φVTB55. These occupied yehV in the double lysogens and sbcB or Z2577 in the Δwrb::cat mutant. While φ549 showed its preference for the sbcB locus when infecting strain C600, φA75 did not. Phage φVTB55 selected a new locus of integration that had been occupied only in C600(φA9/φ24_B) (Table 4).

TABLE 5.

Insertion sites used in the mutant strain of E. coli DH5α ΔwrbA::cat, lacking the wrbA insertion site^a

Modified phage	Insertion site occupied in E. coli strain
	DH5α wild type	DH5α ΔwrbA::cat
φA9Tc	yecE	yecE
φA75Tc	wrbA	sbcB
φ312Tc	wrbA	yehV
φ534Tc	wrbA	yehV
φ549Tc	wrbA	sbcB
φ557Tc	wrbA	yehV
φVTB55Tc	wrbA	Z2577
933WTc	wrbA	yehV
φ24B	wrbA	yehV

^aPhage φA9, originally inserted in yecE in DH5α, was used as a control. Each result was confirmed after three independent replicates.

All phages tested occupied a secondary insertion site already described, and occupancy of a new insertion locus was not detected. The experiments were independently performed at least three times. Ten percent of the colonies generated in each experiment were tested for phage integration in the five insertion loci by PCR. The integrations of phages in yecE, sbcB, and Z2577 were confirmed by analysis of the right and left sides of the phage genome-bacterial genome junctions as described in Materials and Methods. No differences were observed within the group of lysogenized colonies obtained with the same phage.

DISCUSSION

STEC is an emergent pathogen that causes serious diseases. One of its virulence factors is Shiga toxins, which are encoded by lambdoid phages. stx phages establish lysogeny when they become integrated into the host chromosome. We first checked for insertion site usage in 168 STEC strains, mostly isolated from cattle. yehV was among the most frequently occupied loci, either alone or combination with others. This is in accordance with reports published by other authors (5, 31). Although distribution of the occupied loci varies depending on the serotype, the number of non-O157 serotypes was insufficient to provide a clear picture of their distribution. It is noteworthy that wrbA and yehV are occupied in a variety of non-O157:H7 serotypes, contrary to previous findings (31). However, most of the non-O157 organisms reported in Table 2 are not in the enterohemorrhagic E. coli 1 group, and none express the non-O157 O antigens, unlike those in which wrbA and yehV have been previously studied (31). Perhaps the isolates analyzed in this study are in different lineages than those previously reported.

Only 14.3% of the stx strains studied lacked occupancy at any of the five insertion sites tested. This indicates that their hypothetical prophages must occupy another locus not identified yet. Only one O157:H7 strain lacked occupancy at any of the five insertion sites tested. In contrast, a higher number (24%) of the STEC strains exhibited more than one occupied site (21 strains of serotype O157:H7). Moreover, one O157:H7 strain was found to have four simultaneously occupied sites. This was not surprising, since other studies have reported that some STEC are carriers of multiple prophages (2, 12, 20, 32). For example, some isolates presented another stx copy (stx₁) (19), indicating that they carried more than one stx prophage, either defective or functional. This implies that the temperate phages were inserted into a different locus when they lysogenized the bacterial strain.

An important factor that increases genetic variability in bacteria is horizontal gene transfer caused by bacteriophage conversion (8). In this context, phages should be able to integrate and maintain their genes inside the bacterial chromosome. The possibility of integration in a new host chromosome can be a limiting factor in the bacterial strain's acquisition of new genes. To fully evaluate the possibilities of integration in a given host strain, stx₂ phages were transduced from STEC to laboratory strains. In this way, the insertion site preference of a single stx₂ phage could be analyzed.

The seven phages available to our laboratory (19), together with phages 933W, φ3538, and φ24_B, were used for this set of experiments. These phages are genetically heterogeneous and are heteroimmune, based on previous results (19; unpublished observations). Nevertheless, most of them integrate at the same site. A determination of site occupancy in the different host strains showed that the most commonly used insertion sites were wrbA in E. coli and yehV in S. sonnei 866, since the latter strain does not possess the wrbA gene. This confirms that the same integrase can produce integration of the same phage genome at two different sites, if they are available, and that no apparent differences were observed in the number of lysogens generated with both strains. This conclusion is supported by the observation that this strain of S. sonnei has proven to be a very effective host for stx phage infection (19).

Analysis of the phage integrase genes indicated that the 10 phages can be grouped into at least three integrase types. The differences or similarities observed between the int genes do not appear to be related only to integration site selection, since phages carrying different integrase genes (e.g., φ549 and φ312) behaved in the same way, while those harboring identical integrase genes (e.g., φVTB55 and 933W) sometimes presented different insertion sites. Previous studies of site-specific recombination indicate that insertion specificity would depend not only on the properties of Int but also on the sequence of the phage attachment site (27, 33). We confirmed this hypothesis by comparison of sequences of the attL and attR amplimer of the yehV insertion site generated with phages φVTB55 and 933W. The attL sites differed in four nucleotides (data not shown). Since the attL region is a hybrid of attB and attP, these differences indicated that the attP regions of the two phages differ. The differences in the attL region could cause different selection of secondary insertion sites, although their integrase sequences were identical. In contrast, comparison of phage 933W and phage φ24_B, which select identical secondary loci, showed identical attL fragment sequences. Toth et al. (34) showed that phage φ3538, used in this study, failed to integrate in wrbA, scbB, yecA, or yehV loci when infecting a E. coli O45 EPEC strain, although these loci were intact in this strain. Another preferred locus, not identified yet, may be present in this strain, resulting in φ3538 integration elsewhere in the O45 chromosome.

We do not provide enough information on the attP regions and on the insertion mechanisms here to confirm that the system relates to the well-established model for phage tyrosine recombinases (26). However, given the results, we hypothesize that phage integration of stx phages probably follows the same model.

Some studies with P2-like coliphages (16) demonstrated that conversion of a lysogen with a second P2-like phage, one that shares the same integration site as the first, results in the second phage being integrated either in tandem with the first prophage or into a secondary site. That is, the second phage is integrated together with the first temperate phage or into another integration site, but not via the same attachment site (4, 16). Rutkai et al. (27) demonstrated the same phenomenon with their own model, in which lambdoid phages changed their insertion specificity by adapting their integrases to sequences found in secondary attachment sites.

These findings are confirmed by the observation of a change in the phage integration site in double lysogens. In fact, all phages acting as single prophages integrated into wrbA in E. coli laboratory strains but moved to another site, usually yehV, when wrbA was already occupied. Results showed that when the main target site was inaccessible, either due to the presence of another prophage or because of an induced mutation, prophages integrated within a secondary site located in the host chromosome. This finding is consistent with other published reports (4, 16).

For instance, when two phages which occupy the same site when infecting a host strain are used to generate a double lysogen, the first phage occupies its primary insertion site while the second phage, since its primary site is already occupied, is able to integrate into a secondary site. This was the case with both phages 933W and φ3538, whereas E. coli C600 used wrbA. In the double lysogen C600(933W/φ3538), the first phage (933W) integrated in wrbA and the second phage (φ3538) integrated into another insertion site, namely, yehV. This situation was observed with the other double lysogens. Moreover, two copies of the same phage can produce a double lysogen, showing that lysogens of stx phages are not immune to superinfection (Table 4). This observation is not new (2; unpublished observations), and would contradict the well-known immunity theory described for lambda-related phages. However, not all phages were able to superinfect their own lysogens. Some of them produced substitution of the first phage by the second (a phenomenon known as “heteroimmune curing”). In that situation, the second phage replaces the first one in the same insertion site (data not shown).

Other unpublished studies conducted with these double lysogens indicate that the occurrence of recombination events is very unlikely, since the two phages present in the double lysogens can generate isolated plaques of lysis and can lysogenize new host strains separately. In our assays, no evidence of recombination was found, and RecA-mediated recombination events can be excluded, since superinfection has also been observed in E. coli DH5α, a RecA-negative strain. Our results indicate that the repression caused by the prophage repressor in the superinfection is probably a very infrequent occurrence. Here, the frequency was increased due to antibiotic selection, and the superinfection by the same phage could be detected. Whether the integration of the second phage is caused by low constitutive levels of the phage integrase or whether another recombinase present in the host strain is the cause of integration of the second phage cannot be determined here. Therefore, we believe that double lysogens of these stx phages were generated by single-phage-directed integration in the host chromosome. Then, when two copies of the same phage infected the host, a secondary site was occupied by the second copy.

Even when the two phages integrated in the opposite order, the results remained unchanged. In C600(φA9/933W), for example, φA9 was the first prophage integrated. The first phage used wrbA, the second, yehV. In C600(933W/φA9), the order was the opposite: phage 933W integrated within wrbA, while φA9 integrated within yehV (Table 4). The possibility that phage φA9 remained in wrbA while 933W moved to yehV seems unlikely, given the results of Southern blotting. Although some recombination events cannot be excluded, it seems plausible to postulate that the first phage remains in the first locus while the second moves to a second insertion site. In any case, two different loci were occupied. With this in mind, we conclude that chromosomal locus selection results not from prophage interaction but from the insertion site's accessibility.

The results obtained with double lysogens showed that the principal sites used in DH5α were wrbA and yehV. The first site has been described as the preferred locus for stx₂ phages, while the second is reportedly used by stx₁ phages (22, 31, 36). As shown by our results, however, some stx₂ phages occupy yehV, primarily when the preferred insertion site is not available, but in some instances as a first choice. It is possible to find some other examples, such as a virulent sorbitol-fermenting E. coli O157:H⁻ strain isolated in Germany. In this strain, yehV and wrbA were present and intact, while the organism carried stx₂ (31). These results indicated that exceptions regarding the bacteriophage insertion sites can be found.

To directly confirm the hypothesis on site accessibility, we decided to render the wrbA junction site used for the stx₂ phage DNA integration in E. coli DH5α inaccessible. Not unexpectedly, the stx phages used secondary sites. However, they sometimes used different secondary insertion loci in comparison with the double lysogens. Three phages selected a different locus as their secondary site compared with that used in double lysogens. Phage φ549 has already been shown to integrate in the sbcB locus in E. coli C600, although this is not the case with φA75 or φVTB55. This discrepancy has several possible explanations. Phages may have a similar occupancy frequency for the different secondary sites, for instance yehV, sbcB, or Z2577, when the primary site is not available. Previous studies indicate that the integrases can evolve to recognize two or more secondary sites with similar preferences (26). This can also explain the variability observed in the STEC strains. Another plausible explanation is that in double lysogens, homologous recombination between the two phages can serve as the causative agent for the selection of two different sites (16). However, since in our experiments conducted with the mutant strain there was only one phage, homologous recombination between two prophages was not possible.

It is noteworthy that no new insertion sites were detected in these experiments; however, the occupation of other insertion loci in the E. coli chromosome by stx phages cannot be ruled out.

In light of our results, we can conclude that stx phages insert in different sites in STEC, with preferential insertion in yehV for O157:H7 strains. We determined that the preferred site can be different depending on the host strain. Moreover, as previously described, phages can select a secondary site if the primary site is not available. In our study, two different stx phages were integrated in the same host, generating a double lysogen. Those phages integrated in two different loci, one preferred and one secondary. No recombination between the two phages occurred.

As with the primary site, the selection of the secondary site differs depending on the host and the availability of the preferred site in the host. This differential selection of the insertion site would produce a different evolution of the bacterial hosts, which can exhibit a higher or lower rate of phage insertion depending on the availability of potential integration loci. Moreover, since the study of integration sites is used to establish a model of STEC evolution, our findings can help to elucidate how stx phages have been sequentially incorporated in the host chromosome and can help to identify some unknown steps in the evolution of STEC strains. The different selection of integration sites or previous occupancy of a preferred locus will cause interstrain variations that should be considered when restriction fragment length polymorphism patterns of related STEC strains in epidemiologic studies are analyzed.

Nevertheless, bacteriophages must adapt to their circumstances, since insertion site accessibility remains essential for stx₂ phage integration and survival. Instances in which phages infect bacteria already harboring a similar phage or lacking an available insertion site are common. Therefore, phages would need to evolve in order to have integrases with more flexible attachment site recognition. In this way, integrases could ensure their integration and stability at secondary sites inside the cell. Integrase flexibility should confer an evolutionary advantage on these phages, which otherwise cannot integrate within the chromosome and which ultimately face elimination. In these cases, a few conservative motifs at the new attachment site would allow efficient recognition of the integrase and thereby solve the problem of phage competition for these sites (16). In addition, new attachment sites may be generated due to recombination events occurring between the two phages that are, or have been, simultaneously present within the same cell.

In general, bacteria are exposed to several phage infections, some of which result in lysogeny. Most studies have demonstrated that some sequences present in the E. coli chromosomes have a phage origin (8). This means phage traffic, which could lead to chromosome modifications. The presence of a mixture of prophage DNA within the bacterial genome carries an important evolutionary implication: namely, that bacteria might become even more virulent. Indeed, they are likely become more aggressive and to grow rapidly under extreme or special conditions (9, 11).