|Home | About | Journals | Submit | Contact Us | Français|
Bacteriophage lambda has an archetypal immunity system, which prevents the superinfection of its Escherichia coli lysogens. It is now known that superinfection can occur with toxigenic lambda-like phages at a high frequency, and here we demonstrate that the superinfection of a lambda lysogen can lead to the acquisition of additional lambda genomes, which was confirmed by Southern hybridization and quantitative PCR. As many as eight integration events were observed but at a very low frequency (6.4 × 10−4) and always as multiple insertions at the established primary integration site in E. coli. Sequence analysis of the complete immunity region demonstrated that these multiply infected lysogens were not immunity mutants. In conclusion, although lambda superinfection immunity can be confounded, it is a rare event.
Bacteriophage λ is a temperate phage, which, upon infection of Escherichia coli, enters either the lytic or lysogenic replication pathway. The latter is defined by the integration of the λ prophage within the bacterial host genome, where it is quiescently inherited by daughter cells. The maintenance of λ in this prophage state is achieved by an archetypal genetic switch for which the CI repressor protein is pivotal, ensuring its own synthesis while simultaneously repressing the expression of all other phage genes required for integration into the host genome and progression into the lytic replication cycle. This includes the blocking of the expression of integrase from incoming λ bacteriophages such that only a single prophage can be incorporated into a host genome. This elegant and very-well-characterized system is known as the lambda immunity model (19). Bacteriophages that utilize this mechanism and possess the complement of genes arranged as described for λ are termed “lambdoid.” They include a group of converting Stx phages that have become notorious as potentiators of Shiga toxin production (1, 18) and are therefore responsible for the extreme virulence of enterohemorrhagic E. coli strains. Shiga-toxigenic E. coli (STEC) strains were unknown before the early 1980s but are now prevalent, and E. coli O157:H7 in particular is associated with life-threatening gastrointestinal infections (20). The assumption that Stx phages conform to the lambda immunity model was overturned when multiple lysogens of a differentially labeled, isogenic lambdoid Stx phage (Φ24B) were produced (2). In fact, at least three prophages can be harbored simultaneously at different sites (13), and this has serious implications for enhanced virulence and accelerated evolution via intracellular recombination (7). Significantly, the rate at which these multiple infections occur is extremely high, increasing in frequency with each successive infection (13). Multiple isogenic lysogens have subsequently been produced with various additional Stx phages, which emphasizes the potential ubiquity of this phenomenon among toxigenic phages (21).
The portfolio of molecular biological methods now available, in combination with the power of selection for antibiotic resistance used to reveal Stx phage multiple lysogeny, encouraged a reexamination of the capability of bacteriophage λ to direct immunity to superinfection. Historically, the ability of bacteriophage λ to produce multiple lysogens has been reported for defective immunity mutants, undefined mutants, and excisionase-defective strains unable to segregate out additional prophage copies (3, 8, 14, 23). Furthermore, only tandem insertions were identified, notwithstanding the presence of secondary insertion sites that can also be used by λ (22). Most of the early work investigating the carriage of two or more prophages was performed with host cells that acquired two prophages at the initial point of infection. However, in experiments performed by Arber, it was demonstrated that the frequency of lambda superinfection events of a lambda lysogen could be increased by the irradiation of the infecting phage (4), which is likely to have favored subsequent integration/recombination events through the activation of DNA repair mechanisms in the host cell (17). In the current study, the λ lom gene, which has no role in the bacteriophage life cycle (6), was replaced with a spectinomycin resistance cassette, aadA, via recombineering technology (11). This recombinant lambda phage was used to superinfect an existing λ lysogen as well as a naïve E. coli K-12 strain, MC1061 (9), at a multiplicity of infection of 0.1 to minimize the probability of a cell being infected by more than one phage, according to a method previously described (2). Spectinomycin-resistant colonies (n = 10) were randomly picked for the confirmation of lysogen superinfection by DNA amplification using lom-specific primers (Table (Table1).1). Infected lysogens were defined by the presence of both wild-type lom and lom::aadA amplification products (1.2 kb and 1.8 kb, respectively). The rate of lysogen formation for the superinfection of the λ lysogen was decreased by 95% compared to the infection of the naïve E. coli K-12 host (560 and 11,180 mean lysogen-forming units/ml, respectively). This is in stark contrast to the lambdoid Stx phage Φ24B, where the rate of lysogen formation actually increased with subsequent rounds of infection (13). A Southern hybridization was performed for three of the multiple lysogens by using a PCR-generated digoxigenin (DIG)-labeled λ integrase probe (Table (Table1).1). Two bands were revealed, with sizes corresponding to an in silico prediction for tandem integration into the primary insertion site used by λ (Fig. (Fig.1).1). Alternatively, it is feasible that the banding pattern could be produced by a single integrated lambda prophage plus the presence of unintegrated, circularized bacteriophage lambda DNA. To rule out this possibility, a pulsed-field gel electrophoresis (PFGE) Southern hybridization was performed by using the PacI restriction endonuclease, which does not cleave the lambda genome and would therefore permit the distinction of the integrated and circular forms. DNA was extracted from mid-exponential-growth-phase single- and putative double-lysogen cultures embedded inside 1% (wt/vol) low-melting-point agarose plugs by using a standard 16-h protocol (10). Restriction endonuclease digestions were then carried out for 2 h at 37°C (10). While the PFGE Southern blot yielded no evidence of circular lambda DNA, and a band of the expected size in the single λ lysogen control was detected, the band produced by the presumptive double λ lysogen was considerably greater than the expected size (Fig. (Fig.2).2). The band was approximately equivalent to at least an additional four integrated prophages. The presence of multiple additional prophages was corroborated by quantitative PCR (qPCR) analysis of a single and three putative double lambda lysogens. Purified lysogen DNA template was amplified by using lambda-specific Q (antiterminator) gene primers plus two endogenous host reference genes, gyrB and the 16S rRNA gene. Relative gene copy quantification by the 2−ΔΔCT method (16) confirmed the presence of not two but multiple lambda prophages (Fig. (Fig.33).
The simplest explanation for the detection of multiple integrated lambda prophages here is that spontaneous immunity mutants were selected by antibiotic resistance; however, sequence analysis demonstrated that this is not the case. A multiple λ lysogen was induced by the addition of norfloxacin to a mid-exponential-phase culture (optical density at 600 nm [OD600] of 0.5) (2). The filtered phage suspension was plaque assayed to achieve single plaques, 10 of which were randomly picked for use as PCR templates and amplified with lom-specific primers (Table (Table1).1). Four λ phages were selected for further analysis, two wild-type and two recombinant phages discriminated by the sizes of amplification products. The entire immunity region from the λ N to O genes was amplified for each of the selected plaques (Table (Table1).1). Sequence analysis of these amplified immunity regions revealed only a single-base deviation from the previously published lambda phage sequence. This silent point mutation, present within a predicted open reading frame (ORF) immediately downstream of the N gene, was present in all four sequences, i.e., both the resident and superinfecting bacteriophages (Fig. (Fig.4).4). Although a Rho-dependent transcription termination site, tL1, is located downstream of the N gene and is responsible for an 80% blockage of RNA polymerase transcription in the absence of the N protein (12), the mutation discovered here lies approximately 120 bp upstream (position 34944 of the published lambda sequence). This base pair substitution is outside the characterized termination region and therefore would not be expected to interfere with transcription termination (15).
These fresh observations of the detection of multiple bacteriophage lambda lysogens remain distinct from those described for the lambdoid Stx phage Φ24B, because only the primary insertion site for λ is involved in each infection event, subsequent infection is not due to defects in the λ immunity region, and the frequency of multiple infection events for λ is very low indeed. These disparities indicate that the mechanisms involved in the production of multiple lysogens differ between the lambdoid phage Φ24B and bacteriophage lambda. While Φ24B apparently actively promotes the integration of multiple prophages via integrase-mediated site-specific recombination, multiple λ lysogens seem to be produced by a low-frequency, non-integrase-dependent mechanism. It was suggested 50 years ago (4) that the prophages integrated within a lambda lysogen might actually be present in more than one copy, but it was not possible to address this experimentally at that time. The application of molecular biological techniques has established that an integrated prophage does not have to be present in a single copy. The most likely hypothesis to explain this mechanism for lambda is that the superinfecting phage DNA enters the cell, where it is promptly bound by the CI repressor produced by the incumbent prophage. As predicted by the lambda immunity model, in the vast majority of cases, this CI-repressed incoming DNA will not be able to integrate or replicate and, consequently, will be diluted away as the cells divide. However, at low frequencies, the superinfecting phage DNA may recombine with the resident prophage by homologous recombination and become integrated within it, a phenomenon that could well be enhanced by the Red system encoded on lambda (1). Selection for antibiotic resistance would then maintain these normally aberrant phages and may lead to the additional amplification of the number of prophages present, also through homologous recombination. In conclusion, while immunity to superinfection therefore remains a core attribute of bacteriophage lambda, it is not incontrovertible, nor does it necessarily apply to those bacteriophages defined genetically as lambdoid or lambda-like.
This research was supported by the Biotechnology and Biological Sciences Research Council (BBSRC) and the Department for Environment, Food and Rural Affairs (DEFRA), United Kingdom.
Published ahead of print on 7 April 2010.