VGJɸ integration and excision mechanisms contribute to the genetic diversity of Vibrio cholerae epidemic strains

Abstract

Most strains of Vibrio cholerae are not pathogenic or cause only local outbreaks of gastroenteritis. Acquisition of the capacity to produce the cholera toxin results from a lysogenic conversion event due to a filamentous bacteriophage, CTXɸ. Two V. cholerae tyrosine recombinases that normally serve to resolve chromosome dimers, XerC and XerD, promote CTXɸ integration by directly recombining the ssDNA genome of the phage with the dimer resolution site of either or both V. cholerae chromosomes. This smart mechanism renders the process irreversible. Many other filamentous vibriophages seem to attach to chromosome dimer resolution sites and participate in the rapid and continuous evolution of toxigenic V. cholerae strains. We analyzed the molecular mechanism of integration of VGJɸ, a representative of the largest family of these phages. We found that XerC and XerD promote the integration of VGJɸ into a specific chromosome dimer resolution site, and that the dsDNA replicative form of the phage is recombined. We show that XerC and XerD can promote excision of the integrated prophage, and that this participates in the production of new extrachromosomal copies of the phage genome. We further show how hybrid molecules harboring the concatenated genomes of CTXɸ and VGJɸ can be produced efficiently. Finally, we discuss how the integration and excision mechanisms of VGJɸ can explain the origin of recent epidemic V. cholerae strains.

Most strains of Vibrio cholerae, the causative agent of the deadly human disease cholera, are not pathogenic or cause only local outbreaks of gastroenteritis. Indeed, to date epidemic strains have been found in only 2 out of more than 200 possible somatic O-antigen serogroups (1). However, comparative genomics has revealed that cholera epidemic strains are highly diversified and that they evolve continuously and rapidly, mainly via horizontal gene transfer (2). This is particularly well illustrated by the mode of acquisition of the capacity to produce the cholera toxin, which causes the lethal diarrhea associated with the disease (3). Cholera toxin is encoded in the genome of a lysogenic filamentous bacteriophage, CTXɸ (4). Several variants of this phage exist and have been found to be integrated at one or both V. cholerae chromosomes, as a single copy or in multiple tandemly arrayed copies and in conjunction or not with numerous other mobile elements. These variants contribute to the extreme diversity of V. cholerae strains; indeed, most V. cholerae epidemic strains display a different arrangement of genetic elements in the CTXɸ attachment region of each of their two chromosomes (2). In addition, no correlation has been observed between any arrangement and any phyletic lineage, further indicating that horizontal gene transfer drives the rapid evolution of these regions (2). The two above observations sparked considerable interest in understanding the subjacent molecular mechanisms of acquisition.

Previous studies have concentrated on the mechanism of CTXɸ integration. This integration is mediated by XerC and XerD (5), two host-encoded tyrosine recombinases that normally serve to resolve chromosome dimers (6). Each of the two V. cholerae chromosomes harbors a site dedicated to the resolution of chromosome dimers (6). The dif sites are composed of binding sites for XerC and XerD, separated by a 6-bp overlap region (Fig. 1A). The dsDNA replicative form (RF) of CTXɸ harbors two different dif-like sequences in inverted orientations, attP1 and attP2 (Fig. 1A and ref. 7). In the (+) ssDNA genome of the phage, these sequences fold back to form the stem of a double-hairpin structure, CTXɸ attP(+), which serves as phage attachment site (Fig. 1B and refs. 8 and 9). The integration reaction occurs within a nucleoprotein complex consisting of CTXɸ attP(+), a dif site, and one pair each of the two Xer recombinases (Fig. 1C, synapsis). Within this complex, the XerC recombinases catalyze the cleavage of a specific pair of strands immediately 3′ to their binding sites (Fig. 1B, black triangle), which generates recombinase–DNA covalent phosphotyrosyl linkages on one side of the cleaved strands and free 5′-hydroxyl extremities on the other side (Fig. 1C, cleavage). After cleavage, a few bases from each of the two liberated 5′-hydroxyl extremities melt from their complementary strand and attack the recombinase–DNA covalent phosphotyrosyl linkage of their recombining partner (Fig. 1C, strand exchange). Subsequent ligation (Fig. 1C, ligation) requires the stabilization of strand invasion by Watson–Crick or wobble base-pairing interactions, which explains why the three different dif sites identified in the V. cholerae strains sequenced so far (dif1, dif2, and difG) are targeted by specific CTXɸ variants (Fig. 1A and ref. 8). CTXɸ integration depends on the exchange of a single pair of strands. As a consequence, homology between the overlap regions of the attachment site of the phage and of the target chromosome dimer resolution site is required only next to the XerC points of cleavage (Fig. 1B and refs. 8 and 9). The pseudo-Holliday junction (pseudo-HJ) resulting from the exchange of strands catalyzed by XerC is probably converted into product by replication (Fig. 1C, host DNA replication; ref. 9). This process renders integration irreversible, because the phage attachment site is masked in the dsDNA prophage (9). As a consequence, no extrachromosomal copies of the phage genome can be created by excision. However, multiple tandem copies of CTXɸ are often integrated. The rolling circle machinery of CTXɸ can then drive the production of the (+) ssDNA genome of the phage replication by initiating replication on a first integrated copy and terminating it on a second copy (10).

Fig. 1.

Two different types of lysogenic phages that hijack the dimmer resolution machinery of their host. (A) The attachment region of the genome of CTXɸ-related phages contains two dif-like sites, whereas the attachment region of the genome of VGJɸ-related phages contains a single dif-like site. The genome of filamentous phages known to integrate in the vicinity of chromosome dimer resolution sites was scanned for the presence of dif-like sites. The corresponding regions were then aligned to the sequences of the (+) strand of El Tor CTXɸ attP or to the putative VGJɸ attP region. Dots indicate identical bases. Vc, V. cholerae; Vf, V. fischeri; Vm, V. mimicus; Yp, Yersinia pestis; Ec, E. coli; Vp, V. parahaemolyticus; El Tor, El Tor variants of CTXɸ; classical, classical variants of CTXɸ; GGT, G variants of CTXɸ; TAT, CTXɸ variants found in V. fisheri and V. mimicus. References: El Tor, classical and GGT, (8); TAT (23, 24); Ypfɸ and CUS-1ɸ (9); VGJɸ (15); VEJɸ (12); VSKK (AF452449); VSK (AF453500); fs2 (25); f237 (AP000581); VfO3K6 (AB043678); Vf33 (AB012573); and Vf12 (AB012574). (B) Double-stranded scheme of V. cholerae dif1 and of the related sites found in the RF of El Tor CTXɸ (attP1-attP2), in the (+) ssDNA genome of El Tor CTXɸ [attP(+)], and in the RF of VGJɸ (VGJɸ attP). XerC cleavage positions are indicated by triangles. The bases that are involved in the stabilization of the strand exchanges catalyzed by XerC between dif1 and El Tor CTXɸ attP(+) are shown in color, with those of the strand liberated on XerC-mediated cleavage in red and those of the complementary partner strand in blue. Homologous bases found in the overlap regions of El Tor CTXɸ attP1 and attP2 and VGJɸ attP are highlighted similarly. El Tor CTXɸ attP1 and attP2 are connected by a 90-bp DNA sequence, which is indicated as a dotted line. (C) Scheme of the XerC- and XerD-dependent mechanism of lysogenic conversion of V. cholerae by CTXɸ. Chromosomal DNA strands are shown as continuous lines, and the (+) ssDNA genome of the phage is represented by a dotted line. The strands of the bacterial attachment site, dif1, and of the phage attachment site, CTXɸ attP(+), which are exchanged after XerC cleavage, are indicated in red and blue, respectively. XerD and XerC are represented as light-gray and dark-gray figures, respectively. The XerC/DNA phosphotyrosyl linkage is depicted by a black dot.

Among the numerous other mobile DNA elements found integrated next to the dimer resolution sites of the two V. cholerae chromosomes, filamentous phages merit particular attention, because they can be rapidly propagated in the environment and thus participate in the rapid diversification of their host (11). In addition, several reports indicate that the genome of different CTXɸ variants can be packaged by the morphogenesis machinery of such phages and thus be transmitted to strains that do not express the receptor of CTXɸ (TCP), which could participate in the dissemination of cholera toxin genes in the environment (12–14). Sequence data suggest that dif1 could be the attachment site of the largest family of these phages, which includes VGJɸ, VEJɸ, VSK, VSKK, fs2, f237, and Vf33 (Fig. 1A); however, no experimental proof was obtained. In addition, the genome of these phages was found to carry a single potential Xer recombination site, raising questions about the integration mechanism (Fig. 1A). Finally, phages of this family were found to integrate as a single copy, raising questions on the mechanism allowing for the production of new phage DNA from the host genome. This prompted us to study the molecular mechanism subjacent to the integration of a representative member of this family, VGJɸ (15).

Results and Discussion

VGJɸ Hijacks XerC and XerD to Specifically Integrate at dif1.

We previously described a colorimetric screen used to monitor the specificity and efficiency of integration events into the V. cholerae dif sites (8). In brief, the sites were inserted in the coding region of the Escherichia coli lacZ gene in such a manner that the cognate peptides might retain their β-galactosidase activity. The lacZ-dif alleles then replaced dif I in a V. cholerae strain in which dif II and the endogenous lacZ gene had been deleted. The resulting lacZ-dif strains yielded blue colonies on X-gal media, but disruption of the ORF of the lacZ-dif allele by specific phage integration events led to the appearance of white sectors. We infected lacZ-dif1 and lacZ-dif2 V. cholerae cells with a version of VGJɸ tagged with a kanamycin-resistance gene (14). Infection of lacZ-dif1 cells led to the appearance of white sectors in 23.8% of the resulting kanamycin-resistant colonies (Table 1). In contrast, not a single integration event was observed in lacZ-dif2 cells. Deletion of recA did not suppress VGJɸ integration, indicating that integration did not depend on homologous recombination or SOS induction. In contrast, VGJɸ integration was completely abolished in the absence of XerC. Taken together, these observations formally demonstrate that VGJɸ specifically integrates at dif1 via Xer recombination.

Table 1.

In vivo integration of VGJɸ and El Tor RS2

Phage machinery	Host	AttB	Integration, %	Screened colonies
VGJɸ	Δdif2	dif1	23.8	1,827
VGJɸ	Δdif1	dif2	<0.13	724
VGJɸ	Δdif2 ΔxerC	dif1	<0.17	608
VGJɸ	Δdif2 ΔrecA	dif1	19.2	1,064
El Tor RS2	Δdif1 Δdif2	attPVGJ	1.58	631
VGJɸ	Δdif1 Δdif2	attPVGJ	0.29	681
El Tor RS2	Δdif2	dif1	100	852
El Tor RS2	Δdif1	dif2	<0.2	495

Data are from at least three independent experiments.

dsDNA Replicative Form of VGJɸ Is Used as a Substrate for Integration.

The genome of VGJɸ harbors a dif-like site, attP^VGJ (Fig. 1B and refs. 12 and 15). Three base pairs of this site's overlap region are identical to those of dif1 (Fig. 1B) and are immediately adjacent to the XerC-binding site, suggesting that XerC can directly catalyze the exchange of a pair of strands between attP^VGJ and dif1 (Fig. 2A). We checked this in vitro using purified V. cholerae XerC and XerD and annealed synthetic oligonucleotides that mimic dif1, dif2, and attP^VGJ (Fig. 2B). Cleavage of each of the recombining strands and their subsequent ligation to the opposite partner strand was monitored by labeling their 3′ extremities. Strand cleavage led to the appearance of a shorter migration product on a sequencing gel, whereas ligation to a partner strand harboring a longer extension on the 5′ side of the XerC-binding site led to the appearance of a longer recombinant product. Ligation products were detected when attP^VGJ was reacted against dif1 (Fig. 2B, Upper). No ligation products were found when attP^VGJ was reacted against dif2 (Fig. 2B, Upper), even though XerC cleavage was still detected (Fig. 2B, Lower). Thus, XerC and XerD promote the specific exchange of one pair of strands between attP^VGJ and dif1 in the absence of any other host or phage factors.

Fig. 2.

XerC- and XerD-dependent recombination of VGJɸ attP with dif1 and CTXɸ attP(+). (A) Schemes of the Watson–Crick base-pair interactions that could stabilize the strand exchange catalyzed by XerC between the overlap regions of the two dimer resolution sites of N16961, dif1 and dif2, and the dif-like site found in the RF of VGJɸ, attP^VGJ. The strand cleaved by XerC on dif1 and dif2 is shown in red, and the equivalent strain in attP^VGJ is shown in blue. The positions of the the XerC/DNA phosphotyrosyl linkages are indicated by black dots. Pairing interactions are indicated by the proximity of the bases. Numbers indicate the length of the overlap region of each of the four strands of the pseudo-HJ intermediates. (B) V. cholerae XerC- and XerD-mediated recombination of attP^VGJ with dif1 and dif2. A short radioactively labeled attP substrate was reacted with a longer cold dimer resolution substrate (Left), and short radioactively labeled dimer resolution substrates were reacted with a longer cold attP substrate (Right). Schemes of substrate and products are indicated on the side of each panel. A black triangle indicates the position of cleavage of V. cholerae XerC. A star indicates the position of the radioactive label on the probes. (C) Scheme of the XerC- and XerD-dependent mechanism of lintegration of VGJɸ into dif1. The legend is as in Fig. 1B. (D) Scheme of the Watson–Crick base-pair interactions that could stabilize the strand exchange catalyzed by XerC between the overlap regions of the attachment sites of El Tor CTXɸ attP^ET(+) and of VGJɸ attP and attP^VGJ. The legend is as in A. (E) V. cholerae XerC- and XerD-mediated recombination of attP^VGJ with attP^ET(+). Schemes of the products and of the substrates are indicated as in B.

These results suggest that the dsDNA RF of VGJɸ is used as a substrate for integration. To demonstrate that this is the case in vivo, we cloned attP^VGJ in two orientations on a suicide conjugative vector so that either the (+) or (−) strand of attP^VGJ would be specifically transferred during conjugation. The same frequency of integration was observed when either the (+) or (−) strands of attP^VGJ were transferred in lacZ-dif1 cells, in contrast to the exclusive integration of suicide vectors carrying the (+) strand of the attachment region of El Tor CTXɸ (Table 2). No integration was observed in strains lacking dif1, confirming the specificity of integration of VGJɸ (Δdif1 Δdif2). We conclude that the Xer recombination site present on the dsDNA form of VGJɸ, attP^VGJ, directly serves as a phage attachment site. It is probable that the pseudo-HJ resulting from the strand exchange catalyzed by XerC between attP^VGJ and dif1 is converted into product by replication (Fig. 2C). Interestingly, this mechanism was initially proposed for the integration of CTXɸ (5, 7). It is intriguing to observe that the overlap region of attP^VGJ is 7 bp long, like the overlap region of attP2, and that its orientation compared with the RCR machinery of the phage is identical to the one of attP2 (Fig. 1 A and B). This could suggest an evolutionary relationship between VGJɸ-like and CTXɸ-like phages.

Table 2.

Frequency of integration (×10⁻⁶) of nonreplicative plasmids carrying VGJɸ or CTXɸ attP sites

	VGJɸ (+)	VGJɸ (−)	CTXɸ (+)	CTXɸ (−)
Δdif1 Δdif2	<0.2	<0.2	<0.6	<0.6
lacZ-dif1 Δdif2	148 (± 28)	123 (± 20)	1,320 (± 448)	<0.014

Data are from three independent experiments. The origin of the conjugated attP site is indicated at the head of the columns. The (+) and (−) symbols indicate whether or not the strand transferred during conjugation corresponds to the (+) or (−) strand of the phage DNA. The relevant genotype of the recipient strains is indicated on the left of each line. The frequency of integration into dif1 was calculated as the ratio of white transconjugants expressing the nonreplicative plasmid antibiotic-resistance marker (Cm^R) to the number of clones obtained when conjugating a plasmid harboring the El Tor RS2 replication machinery.

CTXɸ Can Integrate into the Replicative Form of VGJɸ.

It was previously reported that the conjunct infection of cells by CTXɸ and VGJɸ led to the formation of molecules harboring the concatenated genomes of the two phages and that this allowed for the efficient propagation of the cholera toxin genes to TCP⁻ strains via VGJɸ particles packaging (12, 14). This prompted us to check how such molecules could be formed. The homologies between the overlap regions of CTXɸ attP(+) and attP^VGJ (Fig. 1A) suggested that XerC could catalyze the exchange of a pair strands between these two sites (Fig. 2D). This was checked in vitro using annealed synthetic oligonucleotides that mimic CTXɸ attP(+) (Fig. 2E). Ligation products were detected when attP^VGJ was reacted against CTXɸ attP(+), demonstrating that XerC and XerD can form a pseudo-HJ between CTXɸ attP(+) and attP^VGJ in the absence of any other host or phage factors (Fig. 2E, Upper). To demonstrate that such a reaction could occur in vivo and that the resulting pseudo-HJ could be converted into product, attP^VGJ was cloned in the E. coli lacZ gene in such a manner that the corresponding peptide might retain its β-galactosidase activity and the resulting allele was used to replace dif1 in ΔlacZ Δdif2 V. cholerae cells. We then used conjugation to directly deliver circular DNA molecules carrying the replication and integration region of an El Tor variant of CTXɸ, El Tor RS2, and a chloramphenicol-resistant cassette. This led to the appearance of white sectors in 1.58% of the resulting chloramphenical-resistant colonies (Table 1). However, the low frequency of integration of El Tor RS2 into attP^VGJ contrasted with the 100% efficiency of integration of El Tor RS2 into dif1, suggesting that the direct recombination of CTXɸ attP(+) and attP^VGJ might not be the primary mechanism of formation of concatenated CTXɸ and VGJɸ genomes. Likewise, we observed that the integration of VGJɸ into attP^VGJ was far less efficient than in dif1 (Table 1). This is likely explained by the fact that the overlap regions of attP^VGJ and CTXɸ attP(+) are larger than the 6-bp canonical overlap region of dif sites (Fig. 1B).

VGJɸ-Integrated Copies Are Unstable.

An important feature of the life cycle of CTXɸ is that it cannot be excised from the genome of its host (Table 3, El Tor RS2; ref. 16). This is due to the incompatibility of the dif-like sites flanking the integrated prophage. There is only 1 bp of homology between the overlap region of attP2 and the V. cholerae dif sites; the 12-bp overlap region of attP1 is too large to permit cooperative interactions between XerC and XerD (9). In contrast, VGJɸ-integrated copies are flanked by two compatible Xer recombination sites, almost identical to dif1 and attP^VGJ (Fig. 2C). Indeed, we found blue sectors and/or papilla in the colonies formed by cells in which a kanamycin-tagged version of VGJɸ had been integrated in lacZ-dif1 (Fig. 3A, Left). After 24 h of growth, >5% of the cells isolated from a fully white clone in which VGJɸ had been integrated were lac+ (Table 3, second row). Deletion of xerC abolished prophage excision, further indicating that the instability was due to Xer recombination (Fig. 3A, Right, and Table 3, third row). Thus, the most likely explanation for the unstability of the prophage is that XerC and XerD can create a pseudo-HJ between the two sites flanking the prophage (Fig. 3B). This corresponds to the “excision” of one strand of the phage genome. Replication of the host DNA will then drive the formation of two nonidentical chromosomes, with the prophage excised from only one of these (Fig. 3B).

Table 3.

Instability of the integrated copies of VGJɸ and VGJɸ El Tor RS2 hybrids

Integrated phage	Host	Loss, %	Screened colonies
El Tor RS2	Δdif2	<0.13	754
VGJɸ	Δdif2	5.62	1,642
VGJɸ	Δdif2 ΔxerC	<0.09	1,078
VGJɸ	Δdif2 ΔrecA	5.82	630
VGJɸ El Tor RS2	Δdif2 ΔmshA	5.35	765

Data are from at least three independent experiments.

Fig. 3.

XerC and XerD-dependent excision of the VGJɸ prophage and de novo XerC- and XerD-dependent production of its RF. (A) β-galactosidase production in cells harboring a copy of VGJɸ integrated at lacZ-dif1. (Left) Xer⁺ cells. (Right) Xer⁻ cells. (B) Scheme of the XerC- and XerD-dependent mechanism of excision of VGJɸ. The legend is as in Fig. 1B. (C) Phage DNA was extracted from strains and analyzed by agaraose gel electrophoresis and ethidium bromide staining. The genotype of the strains is indicated above the pictures of the gels. xerC and recA were successively deleted using an integration/excision strategy. The presence or absence of the endogenous xerC gene is indicated by a “+” or “−” above each lane. Complementation with a functional copy of xerC is indicated by a “+” or “−” above each lane. Integration by homologous recombination of a suicide vector was used to reintroduce xerC into the recA⁺ strain. In the recA⁻ strain, xerC had to be harbored on a replicative plasmid. VGJɸ, RF of the phage; pXerC, pSC101 vector harboring a copy of xerC under its own promoter.

Production of VGJɸ Relies on Xer Recombination.

For most lysogenic phages, such as phage λ, excision of the prophage from the host genome is a crucial step in the life cycle. It serves to produce new extrachromosomal copies of the phage genome and is generally induced when the cell experiences DNA damage due to a high-stress environment and starts to undergo the SOS response. Likewise, the SOS response regulates production of new extrachromosomal copies of the CTXɸ genome (17). In contrast, we found that VGJɸ excision did not depend on SOS induction (Table 3, fourth row). In addition, we could not get rid of cells of the extrachromosomal dsDNA RF of VGJɸ even when a copy of the phage was integrated. Taken together, these observations raise questions as to the importance of lysogeny in the life cycle of VGJɸ. However, we noticed that cells in which Xer recombination was abolished by deletion of xerC were easily cured from the VGJɸ RF (Fig. 3C). Furthermore, the VGJɸ RF reappeared once a functional copy of xerC was reintroduced in such cells, and the de novo production of the RF of the phage did not depend on SOS induction or homologous recombination, given that it was observed in recA⁻ cells (Fig. 3C). Taken together, these observations indicate that Xer-dependent integration and excision of VGJɸ participates in the steady-state production of extrachromosmal copies of its genome.

Tandem Integration of VGJɸ and CTXɸ Leads to Hybrid Phage Production but Makes the dif1-Integrated Copies of CTXɸ Unstable.

Our observation that new dsDNA replicative copies of VGJɸ could be produced from an integrated copy of the phage genome suggests an alternative mechanism for the production of hybrid phages harboring the concatenated genomes of CTXɸ and VGJɸ: When CTXɸ integrates in the new dif1-like chromosome dimer resolution site resulting from the integration of a phage of the VGJɸ family, the tandemly arrayed phage genomes are necessarily flanked by a dif1-like site and a attP^VGJ-like site (Fig. 2C). Such a situation has been observed in the O1 El Tor strain MAK757. In such cases, Xer recombination between the new dif site and the attP^VGJ-like site would be expected to create a pseudo-HJ, which would lead to the production of concatenated phage molecules and/or excision of the prophages, as depicted in Fig. 3B.

To investigate this possibility, we successively integrated VGJɸ and El Tor RS2 into lacZ-dif1, yielding a white strain. In the absence of antibiotic selective pressure, blue colonies reappeared on X-gal plates as frequently as in the case of the single integration of VGJɸ (Table 3). Thus, the increased distance separating the new dif1-like chromosome dimer resolution site and the attP^VGJ-like site did not affect the efficiency of formation of the pseudo-HJ. This observation suggests that arrays of mobile DNA elements integrated at dif1 can be lost in a single Xer recombination step without affecting the dimer resolution site of the chromosome. Such a scenario could explain how the Mozambique strain of V. cholerae arose. This strain derives from a seventh pandemic strain similar to N16961 in which the classical variant of CTXɸ integrated in dif2, the dimer resolution site of chromosome II, and in which no copies of CTXɸ were left in dif1, the dimer resolution site of chromosome I (18). Sequencing of the genome of the Mozambique strain revealed that dif1 is still present on chromosome I (2). Even though classical CTXɸ can target dif2, its integration into dif1 is much more efficient (8). Why, then, did we find no copies of the El Tor or of the classical variant of CTXɸ on chromosome I in this strain? We propose that the Mozambique strain derives from a strain similar to O1 El Tor MAK757. In such a strain, the integration of new copies of the cholera toxin genes on chromosome II via classical CTXɸ integration would remove the selective pressure for maintenance of the cholera toxin copies present on chromosome I for epidemicity. The mobile DNA elements integrated on chromosome I could then be eliminated by Xer recombination without affecting dif1, as depicted in Fig. 3B.

Finally, we found that the tandem integration of VGJɸ and RS2 led to the production of hybrid molecules harboring the concatenated genomes of VGJɸ and RS2 (Fig. S1), as expected given the mode of production of new extrachromosomal copies of VGJɸ (Fig. 3B). Indeed, this is probably the major pathway for the production of hybrid phage molecules, because the efficiency of the excision reaction (Table 3) contrasts with the inefficiency with which CTXɸ attP(+) integrates into attP^VGJ (Table 1). Nevertheless, we wish to emphasize that hybrid phage molecules could be purified only in strains in which mshA, the gene encoding for the receptor of VGJɸ, was deleted, as reported previously (12, 14). As a consequence, we think that the contribution made by the formation of hybrid phage genomes to the dissemination of the cholera toxin gene in the environment might be less important than the contribution made by the direct packaging of the genome of CTXɸ into particles of the VGJɸ family of phages.

In conclusion, our study of VGJɸ unraveled a new mechanism by which filamentous phages can hijack the chromosome dimer resolution system of their bacterial host, which help understand how such phages participate in the rapid evolution of epidemic strains of V. cholerae. Likewise, we believe that the study of other mobile DNA elements that are associated to chromosome dimer resolution sites will be valuable for our understanding of bacterial evolution. In particular, future work will need to address how elements derived from the TLC satellite phage integrates into V. cholerae strains harboring defective dif sites and how this enables toxigenic conversion by CTXɸ (19).

Methods

Strains and Plasmids.

Relevant strains and plasmids are described in Tables S1 and S2, respectively. Both E. coli and V. cholerae cells were grown in LB at 37 °C with shaking. Unless indicated otherwise, cognate antibiotics were used at the following concentrations: streptomycin (St), 100 μg/mL; spectinomycin (Sp), 100 μg/mL; chloramphenicol (Cm), 34 μg/mL for E. coli and 3 μg/mL for V. cholerae; and rifampicin (Rif), 100 μg/mL for E. coli and 2 μg/mL for V. cholerae. All V. cholerae reporter strains were constructed by allele-exchange methods using derivatives of suicide vectors carrying either sacB or rpsL as a counter-selectable marker (20, 21). Engineered strains were confirmed by PCR and sequencing. For long storage, cells were maintained at –70 °C in LB containing 20% glycerol. The El Tor phage replication and integration machinery (RS2) was amplified using the genomic DNA of the N16961 V. cholerae strain as a template. The amplicons were cloned into the suicide vector pSW23T (22). The recombinant suicide vectors carrying the functional lacZ::attP^VGJ allele flanking by the chromosomal fragments of V. cholerae were constructed by cloning 29-bp attP^VGJ site in the natural ClaI site of the E. coli lacZ gene. Unless stated otherwise, all experiments were performed using a derivative of VGJɸ tagged with a kanamycin-resistance marker, VGJɸ-Kn (15). In brief, the RF of this phage carries the R6K replication origin and a kanamycin-resistance cassette integrated into a unique Xba1 site present on the WT phage RF. This insertion does not interrupt any of the identified ORFs of VGJɸ, and the corresponding phage was shown to have the same transduction and integration as VGJɸ (15).

In Vivo Integration Assays.

Transduction.

VGJɸ particles were isolated by serial filtration as described previously (15). In brief, VGJɸ-infected N16961 cells were grown with shaking at 37 °C in LB until they reached an OD of 2 at 600 nm. Bacteria were pelleted by centrifugation, and 1.5 mL of the culture supernatant was serially filtered through 0.45- and 0.22-μm pore filters. The sterility of the filtered supernatant was checked by plating 10-μL aliquots on LB agar plates. For infection, 10 μL of the supernatant were mixed with 100 μL of a culture of the receptor strains that had reached an OD of 1.5 at 600 nm (∼10⁸ cells). The mixture was incubated at 37 °C for 30 min to allow infection before plating on appropriate selection plates.

Conjugation.

For conjugation, both donor [diaminopimelic acid (DAP) auxotroph E. coli] and recipient (V. cholerae) strains were grown separately in LB to an OD of ∼0.3 at 600 nm. Bacteria were pelleted by centrifugation, washed, and mixed in 1/4 of the initial volume in fresh LB supplemented with 0.3 mM DAP. The mix was then dropped onto sterile filter paper on top of an LB-agar plate supplemented with DAP. After 4 h of incubation at 37 °C, bacterial cells were resuspended and plated on LB plates containing X-gal isopropyl-β-D-thiogalactopyranoside (IPTG) and cognate antibiotics. Transconjugants carrying an integrated or a RF of phage machinery were monitored after 36 h of growth at 37 °C.

Protein Purification.

The XerC and XerD ORFs were amplified by PCR from the N16961 genome and cloned into the pTYB-11 expression vector (New England Biolabs) using SapI and PstI restriction sites. Proteins were produced at 30 °C in BL21Gold cells (Stratagene). XerD-producing cells were grown for 2 h in the presence of 0.1 mM IPTG. XerC-producing cells were grown for 2 h in the presence of 0.2% glucose and 0.5 mM IPTG. Cells were collected and resuspended in buffer A [25 mM TrisHCl (pH 8), 1 M NaCl, 10% glycerol], frozen in dry ice, and lysed with a Carver press. Lysats were centrifuged for 1 h at 25, 000 × g. The supernatants were loaded on a chitin bead column and washed extensively with buffer A. Intein tag cleavage was performed in buffer A adjusted to 0.5 M NaCl and supplemented with 50 mM DTT at 7 °C for 16 h. Untagged XerC and XerD were eluted, and small aliquots were frozen and stored at −70 °C. Protein concentrations were evaluated by the Bradford method using BSA as a standard.

In Vitro Recombination Assays.

The synthetic oligos used to mimic dif1, dif2, attP^ET, and attP^VGJ are listed in Table S3. Recombination reactions were performed in a 20-μL volume in the presence of 25 mM Tris-HCl (pH 7.4), 100 mM NaCl, 1 mM EDTA, 0.1 μg/mL BSA, 40% glycerol, and 5 nM each of the cold and radioactively labeled recombination substrates. XerC and XerD were used at final concentrations of 150 nM and 100 nM, respectively. Reactions were incubated for 3 h at 37 °C, precipitated with ethanol, and analyzed by PAGE using a 10% acrylamide-urea gel. Dried gels were exposed on phosphor screens. Signals were detected with a Typhoon storage phosphor imaging system and quantified using IQT 7.0 software (GE Healthcare).