Prophage-Dependent Neighbor Predation Fosters Horizontal Gene Transfer by Natural Transformation

Prophages are nearly ubiquitous in bacterial species. These integrated phage elements have previously been implicated in horizontal gene transfer (HGT) largely through their ability to carry out transduction (generalized or specialized). Here, we show that prophage-encoded viral particles promote neighbor predation leading to enhanced HGT by natural transformation in the waterborne pathogen Vibrio cholerae. Our findings contribute to a comprehensive understanding of the dynamic forces involved in prophage maintenance which ultimately drive the evolution of naturally competent bacteria in their natural environment.

IMPORTANCE Prophages are nearly ubiquitous in bacterial species. These integrated phage elements have previously been implicated in horizontal gene transfer (HGT) largely through their ability to carry out transduction (generalized or specialized). Here, we show that prophage-encoded viral particles promote neighbor predation leading to enhanced HGT by natural transformation in the waterborne pathogen Vibrio cholerae. Our findings contribute to a comprehensive understanding of the dynamic forces involved in prophage maintenance which ultimately drive the evolution of naturally competent bacteria in their natural environment. KEYWORDS HGT, Vibrio cholerae, bacteriophages, chitin, natural transformation, neighbor predation S everal bacterial species have evolved to capture DNA (natural competence) as a source of nutrients (1) or to incorporate it into their genome to speed their evolution via a process termed natural transformation (NT) (14)(15)(16). Vibrio cholerae is a genetically tractable and well-established model organism to study NT. This human pathogen is usually found in association with the chitinous carapaces of zooplankton in estuary and ocean waters (2). V. cholerae can utilize chitin as a major carbon and nitrogen source, and additionally, this polymer is a required signal for induction of NT (3). In estuarine chitin microcosms, it has been shown that this pathogen can take up multiple large DNA fragments when the exogenous DNA concentration is high (4,5). Much evidence points to NT and horizontal gene transfer (HGT) having contributed to the evolution of V. cholerae (6).
About 50% of V. cholerae clinical isolates carry the temperate kappa phage K139 (7). However, the role of K139 in the ecology of V. cholerae has been ill defined. Here, we explore whether the lytic replication of K139 affects the physiology of V. cholerae in chitin microcosms, which mimic the aquatic reservoir for this facultative pathogen.
Using the K139 lysogen E7946, an O1 El Tor V. cholerae strain, we first evaluated bacterial replication and production of viral particles over time in chitin microcosms. V. cholerae growth was slow, and bacterial numbers increased 100-fold by 24 h and 1,000-fold by 48 h of incubation, while K139 PFU increased 3 logs by 6 h, which is when bacteria were entering exponential growth (Fig. 1A). Interestingly, we found that insoluble chitin specifically increased K139 PFU compared to other carbon sources (see Fig. S1 in the supplemental material). Furthermore, addition of alternative carbon sources (e.g., glucose) to chitin microcosms reduced K139 PFU (Fig. S1 1B). This effect was dependent on K139 viral production, because an E7946 ΔK139 strain was unable to outcompete nonlysogens to the same extent as the WT strain. This is represented by an increase of the competitive index in these cocultures, where a CI of 1 is expected if the two strains compete equally (Fig. 1B).
Our results strongly suggest that K139 plays an ecological role by providing a competitive advantage for lysogenic strains in mixed populations containing nonlysogens. By excising within a small fraction of lysogenic cells, and efficiently killing the neighboring nonlysogenic cells via lytic replication, K139 may allow the remaining lysogenic population to successfully compete for resources in a nutrient-limited estuary environment. It has previously been shown that some phages can lyse bacterial cells and release intact DNA (8), while other phages degrade host DNA following lysis (9). Release of intact DNA could aid in promoting horizontal gene transfer by NT (8,10). Here, we have found that K139 promotes neighbor predation under the same chitin microcosm conditions that are required to induce NT. Therefore, we next wanted to test if K139-dependent neighbor predation could promote HGT in a chitin microcosm. To that end, we cocultured a mixture of strains where each carried an unlinked selection marker at a neutral site in the genome (5). The generation of strains with both selection markers indicated HGT. After 48 h of growth in chitin estuary microcosms, the coculture containing E7946 (lysogen, WT) and E7946 ΔK139 (nonlysogen), where neighbor predation is expected to occur, showed the highest number of transformants, which was ϳ100-fold higher than a coculture of two K139 lysogens or two nonlysogens where neighbor predation is not expected to occur (Fig. 1C). These results are consistent with neighbor predation promoting HGT.
Prophages have mainly been linked to HGT by transduction (10). However, in chitin microcosms, V. cholerae can also undergo HGT via NT. Thus, next we designed a strategy to determine if the HGT observed is attributable to phage transduction or to NT. To distinguish between these, we inactivated dprA in our V. cholerae strains, a gene that is essential for NT but is not required for transduction (11). After 48 h of growth in chitin estuary microcosms, the coculture containing an NT ϩ lysogen (E7946) and an NT Ϫ nonlysogen (E7946 ΔK139 ΔdprA) showed elevated rates of HGT similar to the coculture containing both NT ϩ strains (Fig. 1C). In contrast, a mixture containing an NT Ϫ lysogen and an NT ϩ nonlysogen showed the basal levels of HGT seen in cocultures where no phage predation occurs (Fig. 1C). This suggests that the HGT observed is due to the transfer of DNA from the nonlysogen to the lysogen. This is the opposite of what would be expected for phage transduction where a prophage excised from a lysogenic strain would transduce DNA to the nonlysogen. Together, our results point to an adaptative strategy used by lysogenic strains to induce prophage-dependent neighbor predation in order to thrive and also to capture released DNA for HGT in their aquatic environment.
To the best of our knowledge, our work is the first to establish an ecological role for K139 in enhancing V. cholerae fitness. We show that K139 can enhance survival of V. cholerae lysogens in chitin estuarine environments by neighbor predation of nonlysogens and by driving evolution via HGT. More broadly, our results suggest a novel mechanism by which prophages benefit their lysogenized hosts, which may contribute to the maintenance of these genetic elements in bacterial genomes.
Bacterial strains and growth conditions. See Table S1 for a list of all V. cholerae strains used in this study. All strains were routinely grown in Luria-Bertani Miller (LB) broth and agar at 30°C. Where necessary, medium was supplemented with erythromycin (10 g/ml) or trimethoprim (10 g/ml).
For chitin utilization experiments, cells were grown overnight in LB at 30°C with shaking. The next morning, cultures were washed and diluted in 0.7% Instant Ocean (Aquarium Systems) or M9 minimal medium. A 10 5 -CFU amount was inoculated in a 1% shrimp shell chitin (Sigma) suspension in 0.7% Instant Ocean (chitin microcosm) and was incubated up to 48 h at 30°C statically in 14-ml glass test tubes (Fisher Scientific). Lactate or glucose was added to a final concentration of 0.2% and 0.5% when required. GlcNAc sugars were added at 2 mM for pentasaccharides, 3.33 mM for trisaccharides, 5 mM for disaccharides, and 10 mM for monosaccharides. CFU counts were evaluated by serially diluting and plating on LB agar plates.
Bacteriophage assays. To assay phage titers, supernatants from bacterial cultures were filtered using 0.2-m filters (Costar). Filtered supernatants were serially diluted and tittered using E7946 ΔK139 (nonlysogenic, susceptible strain) as described in reference 12. Plates were incubated overnight at 37°C, and turbid plaques were counted the next morning.
HGT assays between strain pairs were conducted with differentially marked strains where each contained an antibiotic resistance (Ab r ) marker at a distinct neutral locus (ΔVC1807::Erm r or ΔVCA0692::Tm r ). As indicated, strains were mixed in a 1:1 ratio in chitin microcosms for 48 h at 30°C statically (5). After 48 h, a portion of each coculture was diluted in LB broth and outgrown for 2 h prior to plating. Reaction mixtures were plated for quantitative culture on trimethoprim (Tm) ϩ erythromycin (Erm) plates to quantify transformants, as well as on Erm and Tm alone to quantify the abundance of each strain within the coculture. Transformation frequency is expressed as CFU of transformants (Erm r ϩ Tm r double resistant)/(CFU of Erm r ϩ CFU of Tm r ).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. FIG S1, TIF file, 1.1 MB.