Quorum sensing contributes to natural transformation of Vibrio cholerae in a species-specific manner

Abstract

Although it is a human pathogen, Vibrio cholerae is a regular member of aquatic habitats, such as coastal regions and estuaries. Within these environments, V. cholerae often takes advantage of the abundance of zooplankton and their chitinous molts as a nutritious surface on which the bacteria can form biofilms. Chitin also induces the developmental program of natural competence for transformation in several species of the genus Vibrio. In this study, we show that V. cholerae does not distinguish between species-specific and non-species-specific DNA at the level of DNA uptake. This is in contrast to what has been shown for other Gram-negative bacteria, such as Neisseria gonorrhoeae and Haemophilus influenzae. However, species specificity with respect to natural transformation still occurs in V. cholerae. This is based on a positive correlation between quorum sensing and natural transformation. Using mutant-strain analysis, cross-feeding experiments, and synthetic cholera autoinducer-1 (CAI-1), we provide strong evidence that the species-specific signaling molecule CAI-1 plays a major role in natural competence for transformation. We suggest that CAI-1 can be considered a competence pheromone.

Fig. 1.

Model of how quorum sensing contributes to natural transformation. In V. cholerae, two autoinducers are produced by the synthases CqsA and LuxS: cholera autoinducer-1 (CAI-1) and autoinducer 2 (AI-2), respectively. These autoinducers are sensed by their respective receptors CqsS and LuxPQ. The latter are involved in a phosphotransfer relay (dotted arrows [simplified]) and act as kinases at low cell density (LCD) and as phosphatases at high cell density (HCD). Depending on the resulting phosphorylation state of LuxO, the synthesis of HapR protein may be posttranscriptionally inhibited at LCD (indirectly; shown by the dashed line), whereas this inhibition is absent at HCD. HapR positively contributes to natural competence and transformation by activating the competence gene comEA and by repressing the extracellular nuclease gene dns. The smaller font size of “LuxS,” “AI-2,” and “LuxPQ” indicates the lower significance of this quorum-sensing system toward HapR-regulated processes as described here and elsewhere (43). Strain A1552ΔhapR () is comparable to this LCD state. Strain A1552ΔluxO (★) is similar to an HCD state because it lacks the posttranscriptional inhibition of HapR synthesis. Abbreviations: ∼, phosphorylated; de∼, dephosphorylated. This scheme is based on findings in references , , and .

Fig. 2.

Uptake of non-species-specific DNA by naturally competent V. cholerae cells. (A) Scheme of acceptor and donor strain-specific gDNA regions. All V. cholerae acceptor strains used in this experiment were lacZ positive and therefore semiquantitatively detectable with lacZ-specific primers (LacZ-missing-fw/LacZ-missing-bw, indicated by #1 and #2 here and in Table 2). The donor gDNA of the V. cholerae strain A1552-LacZ-Kan (37) lacks this part of the lacZ gene due to replacement with a kanamycin resistance cassette (aph; primers #3 and #4). A chromosomal region specific to E. coli BL21(DE3) gDNA encodes the T7 DNA-directed RNA polymerase (CP001509 [29]; primers #5 and 6); for B. subtilis 168, the spoEIII gene served as a species-specific genome template (primers #7 and #8). The expected PCR fragment sizes are indicated (not to scale). (B) Control experiment for validation of primers in the duplex PCR. Duplex PCR with the oligonucleotide combinations #1/#2 (acceptor strain specific; lanes 1 to 6) and donor gDNA-specific primer pairs #3/#4 (lanes 1 and 2), #5/#6 (lanes 3 and 4), and #7/#8 (lanes 5 and 6) was performed. The different templates were 30 ng gDNA of the V. cholerae acceptor strain A1552 (acceptor strain; lanes 1, 3, and 5) and donor gDNA derived from the V. cholerae strain A1552-LacZ-Kan (lane 2), E. coli strain BL21(DE3) (lane 4), or B. subtilis strain 168 (lane 6). (C) Detection of DNase I-resistant donor gDNA after induction of natural competence on chitin surfaces. V. cholerae strains A1552 (wild type, lane 1), A1552ΔpilQ (lane 2), A1552ΔcomEC (lanes 3, 6, and 7), A1552ΔdprA (lane 4), and A1552ΔrecA (lane 5) were induced for natural competence by growth on chitin flakes. After 24 h from the provision of donor gDNA of strains, namely, V. cholerae A1552-LacZ-Kan (lanes 1 to 5) (V.c.), E. coli BL21(DE3) (lane 6) (E.c.), and B. subtilis 168 (lane 7) (B.s.), the cells were analyzed for uptake of donor DNA by whole-cell duplex PCR. The asterisk indicates the quantification of acceptor strains using the primer pair #1 and #2 (1:10 diluted), as indicated for panel A. Both images depict the same area, but the upper image has been uniformly enhanced for contrast and brightness. Primer combinations are indicated above the figure. (D) Semiquantitative DNA uptake assay following artificial competence induction in liquid medium. Chitin- and surface-independent DNA uptake was performed in liquid medium by artificial overexpression of tfoX. This method allowed for the quantification of bacterial cells by optical density measurements. V. cholerae strains tested were A1552Δdns/pBAD-tfoX-stop (lanes 1, 5, and 9), A1552ΔdnsΔpilQ/pBAD-tfoX-stop (lanes 2, 6, and 10), and A1552ΔdnsΔcomEC/pBAD-tfoX-stop (lanes 3, 7, and 11). As a negative control, strain A1552ΔdnsΔcomEC/pBAD-tfoX-stop was tested in the absence of the tfoX inducer (lanes 4, 8, and 12). Donor gDNAs were derived from V. cholerae A1552-LacZ-Kan (lanes 1 to 4), E. coli BL21(DE3) (lanes 5 to 8), and B. subtilis 168 (lanes 9 to 12). Primer combinations are indicated above the figure. Shown are the results of one representative experiment of at least three independent replicates. L, ladder (left, 1 = kb ladder; right, 100 = bp ladder [Invitrogen]).

Fig. 3.

Effect of quorum-sensing systems 1 and 2 on the natural transformability of V. cholerae. (A) Transformation frequencies of V. cholerae strains with defects in the quorum-sensing circuit. Strains tested were V. cholerae strain A1552 (wild type; lane 1), A1552ΔcqsA (CAI-1⁻ AI-2⁺; lane 2), A1552ΔluxS (CAI-1⁺ AI-2⁻; lane 3), A1552ΔcqsAΔluxS (CAI-1⁻ AI-2⁻; lane 4), and A1552ΔhapR (lane 5). Natural-transformation frequencies are indicated on the y axis. < d.l., below the detection limit. The detection limit varied between 3.2 × 10⁻⁸ and 1.2 × 10⁻⁶ (44 independent experiments) and 4.0 × 10⁻⁸ and 8.7 × 10⁻⁷ (11 independent experiments) for A1552ΔcqsAΔluxS and A1552ΔhapR, respectively. Data are averages of results from at least 11 independent experiments. #, for strain A1552ΔcqsA, the transformation frequency was below the detection limit in 12 out of 23 experiments. To allow the calculation of the average, the transformation frequency was set to the detection limit for these strains. Thus, the indicated average slightly overestimates the residual transformability of this strain. A statistically significant difference from the transformation frequency of the wild-type strain (lane 1) is indicated (**, P < 0.01, as determined by Student's t test of log-transformed data). (B) Complementation of the cqsA deletion in trans. Transformation frequencies were scored for V. cholerae wild-type strain A1552 (lane 1) or its cqsA-minus derivative (A1552ΔcqsA; lanes 2 to 4). Complete restoration of natural transformability was possible by providing cqsA in trans preceded solely by its endogenous promoter (A1552ΔcqsA/pBR-[own]-cqsA; lane 3) or in combination with the constitutive tet promoter provided by the plasmid backbone (A1552ΔcqsA/pBR-[Tet+own]-cqsA; lane 4). The vector control is shown in lane 2 (A1552ΔcqsA/pBR322). Averages of results from four independent experiments are shown. The detection limit was between 1.8 × 10⁻⁷ and 5.8 × 10⁻⁷ for strain A1552ΔcqsA/pBR322.

Fig. 4.

Epistasis analysis indicates that the transformation-negative phenotype of a cqsA-minus strain occurs via the canonical quorum-sensing cascade. (A) Complete restoration of natural transformability through the deletion of luxO in a ΔcqsA background. Natural-transformation frequencies, as scored on chitin flakes, are indicated on the y axis. Strains tested were A1552 (lane 1), A1552ΔluxO (lane 2), A1552ΔcqsA (lane 3), and A1552ΔcqsAΔluxO (lane 4). (B) Partial restoration of natural transformability through elimination of the extracellular nuclease Dns (dns⁻). Strains tested for natural transformability were A1552 (lane 1), A1552ΔcqsA (lane 2), A1552ΔcqsAΔluxS (lane 3), A1552ΔhapR (lane 4), A1552Δdns (lane 5), A1552ΔcqsAΔdns (lane 6), A1552ΔcqsAΔluxSΔdns (lane 7), A1552ΔhapRΔdns (lane 8), A1552ΔcomEAΔdns (lane 9), A1552ΔcqsAΔluxSΔdnsΔcomEA (lane 10), and A1552ΔhapRΔdnsΔcomEA (lane 11). #, for strain A1552ΔcqsA, the transformation frequency was below the detection limit in 3 out of 4 experiments (A) and 4 out of 8 experiments (B). To allow calculation of the average, the transformation frequency was set to the detection limit for these strains. Thus, the indicated average slightly overestimates the residual transformability of this strain. The averages of results from at least three independent experiments are shown. < d.l., below the detection limit, which varied between 4.4 × 10⁻⁸ and 8.3 × 10⁻⁶ for all indicated strains. Statistically significant differences from the transformation frequency of the wild-type strain (lane 1) are indicated in both panels (*, P < 0.05; **, P < 0.01) determined by Student's t test of log-transformed data.