The type VI secretion system of Vibrio cholerae fosters horizontal gene transfer

See allHide authors and affiliations

Science  02 Jan 2015:
Vol. 347, Issue 6217, pp. 63-67
DOI: 10.1126/science.1260064

Killing, sex, and gene swaps in bacteria

The bacterial type VI secretion system (T6SS) is used by bacteria to inject toxins into neighboring cells to eliminate competition. This molecular machine is thus considered to be a mechanism by which bacteria can exert social control in complex microbial communities. Borgeaud et al. have discovered that in Vibrio cholerae, T6SS genes are co-regulated with genes involved in DNA uptake. Hence, T6SS-dependent killing of other bacteria is directed at neighboring cells, which release their DNA to be taken up by the killer, which can then integrate valuable genes and rapidly evolve, leading to antibiotic resistance or virulence.

Science, this issue p. 63


Natural competence for transformation is a common mode of horizontal gene transfer and contributes to bacterial evolution. Transformation occurs through the uptake of external DNA and its integration into the genome. Here we show that the type VI secretion system (T6SS), which serves as a predatory killing device, is part of the competence regulon in the naturally transformable pathogen Vibrio cholerae. The T6SS-encoding gene cluster is under the positive control of the competence regulators TfoX and QstR and is induced by growth on chitinous surfaces. Live-cell imaging revealed that deliberate killing of nonimmune cells via competence-mediated induction of T6SS releases DNA and makes it accessible for horizontal gene transfer in V. cholerae.

Vibrio cholerae is a well-studied human pathogen that causes severe and potentially fatal diarrhea in humans. V. cholerae is primarily an aquatic bacterium that is often found in association with zooplankton (1). The molted exoskeletons of planktonic crustaceans are primarily composed of the polymer chitin. When growing on chitinous surfaces, V. cholerae initiates a developmental program known as natural competence (2, 3), which allows the bacterium to take up free DNA from the environment (4) using a competence-specific DNA uptake machinery (5, 6). The competence program is dependent on the regulatory protein TfoX, which is produced in the presence of chitin and chitin degradation products (4, 79) (fig. S1). Natural competence is also co-regulated by carbon catabolite repression (10) and quorum sensing (QS) (7). QS requires autoinducers [cholera autoinducer 1 (CAI-1) and autoinducer 2] and a master regulator (HapR) (11). We recently demonstrated that only a subset of the known competence genes (e.g., comEA and comEC) (fig. S1) are co-regulated by QS and in a CAI-1–dependent manner (12, 13), and we suggested that CAI-1 acts as a competence pheromone (12). The QS and TfoX-dependent regulator QstR links QS and TfoX activity in the induction of competence genes (14) (fig. S1). In this study, we demonstrate that the type VI secretion system (T6SS) (15, 16) of V. cholerae is part of the competence regulon and is expressed when the bacterium grows on chitinous surfaces. As a consequence, any nonimmune neighboring cells (1618) are killed, and the released DNA serves as transforming material that enhances horizontal gene transfer (HGT).

To understand the extent of the TfoX-dependent competence regulon, we used an RNA sequencing (RNA-seq) approach to enable an accurate assessment of the bacterial transcriptome (19). We studied a variant of the pandemic V. cholerae O1 El Tor isolate A1552 (table S1), which carries an arabinose (ara)–inducible copy of tfoX on its chromosome, as the wild-type (WT) strain and grew the bacteria in the absence or presence of arabinose to simulate chitin-induced expression of TfoX (5, 13). Upon the induction of tfoX, we observed substantial up-regulation of the three T6SS-encoding gene clusters (Fig. 1A, fig. S2, and table S2): the major or large gene cluster, which encodes the structural components of the T6SS, and two auxiliary clusters (17, 20), which encode haemolysin co-regulated (Hcp) proteins. All three clusters also encode valine-glycine repeat G proteins (VgrG-1, VgrG-2, VgrG-3) and various effector-immunity modules (17). The T6SS machinery structurally and functionally resembles intracellular and membrane-attached phage tails (15, 21) and forms a tubular structure composed of the two sheath proteins VipA and VipB (22). Upon contraction, the sheath propels an inner tube composed of Hcp proteins and capped by a complex between the VgrG proteins and a proline-alanine-alanine-arginine (PAAR) repeat–containing spike protein, together with the effectors, into neighboring bacterial or eukaryotic cells (1518, 20, 21, 23, 24). The predatory population itself is protected against self-destruction by the simultaneous production of effector-compatible immunity proteins (1618).

Fig. 1 The gene cluster encoding the type VI secretion system (T6SS) is activated by the transformation regulator TfoX.

(A) RNA-seq data showing the coverage of cDNA reads over the major T6SS gene cluster (shown on top; VCA0105 to VCA0123 as indicated) on the small chromosome of V. cholerae. Strains tested: WT, V. cholerae strain A1552; ΔhapR, A1552ΔhapR; and ΔqstR, A1552ΔqstR. All strains contain the TntfoX transposon (13) on their chromosome. “-” and “+” indicate the absence and presence of tfoX induction, respectively. Scale of the y axis for each strain: 0 to 500. (B) Verification of the RNA-seq data by qRT-PCR. The same strains as indicated in (A) were grown in the absence (-) or presence (+) of tfoX induction, and the relative expression of the indicated genes was evaluated through qRT-PCR (y axis). The data represent the mean of three independent biological replicates (± SD, as indicated by the error bars).

The regulation of the T6SS in V. cholerae is not well understood because the function and dynamics of the T6SS have primarily been studied in nonpandemic isolates with constitutive T6SS activity (e.g., strains V52 and 2740-80) (17, 20, 21, 23, 24). In strain V52, the enhancer-binding protein VasH, which is encoded by the major T6SS gene cluster (fig. S2), acts as an activator of the sigma factor RpoN (20). VasH has no effect on the structural genes of the major T6SS gene cluster but solely controls the auxiliary clusters and vgrG3 (25). Although high osmolarity promotes Hcp secretion in strain A1552, the main T6SS gene cluster is induced by only ~1.5-fold compared with low-osmolarity conditions (26). Therefore, Ho et al. proposed that there is an additional environmental signal that triggers the transcription of the major cluster (including vasH) and that VasH can then activate the expression of the auxiliary clusters in an RpoN-dependent manner (15). On the basis of our RNA-seq data, we suggest that the competence regulatory protein TfoX initiates the transcription of the major T6SS cluster and, thus, the auxiliary clusters (Fig. 1A, fig. S2, and table S2).

TfoX regulates the majority of the competence genes that encode components of the DNA uptake machinery of V. cholerae (47, 13). Notably, a subset of the so-far identified competence genes is co-regulated by QS in a HapR- and QstR-dependent manner (4, 13, 14) (fig. S1). We therefore repeated the RNA-seq experiment in tfoX-expressing hapR- and qstR-negative strains. In these mutants, TfoX-dependent up-regulation of the T6SS genes was lost (Fig. 1A, fig. S2, and tables S3 and S4). We used quantitative reverse-transcription polymerase chain reaction (qRT-PCR) to confirm the TfoX-, HapR-, and QstR-dependent regulation of the T6SS (Fig. 1B). Moreover, as the regulatory circuits often differ among strains, we tested the TfoX (and HapR)–dependent expression of the T6SS genes in five different V. cholerae O1 El Tor isolates (table S1) from South America, Asia, and Africa (fig. S3), which were all naturally transformable on tfoX induction (table S5).

The transcription and translation of native tfoX requires the presence of a chitinous substrate (4, 79), and we confirmed that the expression of the T6SS genes was elevated after growth of V. cholerae on chitin flakes (27) (fig. S4). We also employed a chitin colonization assay (10) to visualize the T6SS. We generated a translational fusion between superfolder green fluorescent protein (sfGFP) and the main sheath protein VipA, as previously described (23), and used the vipA-sfgfp allele to replace the indigenous copy of vipA. Next we verified that VipA-sfGFP was produced under competence-inducing conditions (e.g., in a TfoX- and high cell density–dependent manner) concomitantly with the periplasmic competence protein ComEA (fig. S5; details below). In accordance with earlier studies using the T6SS hyperactive strain 2740-80 (23), we observed both extended and contracted sheath structures after tfoX induction and dynamic T6SS behavior (fig. S5). Moreover, upon growth on chitin beads, we detected VipA-sfGFP forming extended and contracted sheath structures, indicating that the full T6SS gene cluster was expressed (Fig. 2). Such chitin-induced production of T6SS sheath structures was also observed in several other V. cholerae O1 and non-O1 strains (fig. S6) and was dependent on the regulator QstR and on other structural components of the T6SS (fig. S7).

Fig. 2 The T6SS is induced and assembled upon growth on chitin.

Fluorescence light microscopy of chitin bead–colonizing cells. The V. cholerae strain was engineered to carry translational fusions for both the competence protein ComEA-mCherry and the T6SS sheath protein VipA-sfGFP. From left to right (upper row): phase contrast image (Ph), phase contrast image overlaid with the signal from the green fluorescence channel (Ph+GFP), and merged signal from the two fluorescence channels (GFP+mCherry). Zoomed images of the boxed region are shown in the lower row. Assembled and contracted VipA-sfGFP sheath structures are indicated by the white arrows. Scale bars: 10 μm (upper row); 2 μm (lower row).

Having established that chitin and TfoX induce the T6SS gene cluster in V. cholerae, we assessed its functionality in an interspecies killing assay. Strains containing the inducible copy of tfoX exhibited significant killing behavior toward Escherichia coli in a T6SS-dependent manner when grown in the presence of the inducer (Fig. 3A and fig. S8).

Fig. 3 TfoX-mediated expression of the T6SS leads to bacterial killing and natural transformation.

(A) Quantification of E. coli TOP10 recovery [colony-forming units per milliliter (CFU/ml), as indicated on the y axis] after coculturing with V. cholerae or E. coli cells on plain LB agar plates (-ara) or LB agar plates supplemented with arabinose (+ara). The tested strains are indicated below the graph and contain the arabinose-inducible copy of tfoX (TntfoX) where indicated. The averages of three independent experiments (± SD, error bars) are shown. Asterisks denote statistically significant differences (P < 0.001); n.s., not significant. (B) TfoX-dependent expression of the T6SS enhances natural transformation. The indicated predatory strains are derivatives of V. cholerae O1 El Tor A1552 (RifR) and carry TntfoX where indicated. The predator was incubated without any prey (-) or with prey strains (+) that are derived from the environmental isolate Sa5Y (28) and contain a KanR cassette integrated into lacZlacZ; lanes 3 and 4; T6SS-proficient) or vipAvipA; lanes 6 to 16; T6SS-defective). The strains were cocultured on LB agar plates containing 0, 0.02, or 0.2% of the tfoX inducer arabinose (as indicated) before the selection of transformants on LB plates containing both antibiotics (Rif+Kan). “#” denotes strains that were killed by the T6SS-proficient Sa5Y strain and are thus nontransformable. The horizontal transfer of the KanR cassette from the prey to the predator was confirmed for randomly picked transformants (fig. S10).

To investigate whether competence-induced T6SS-mediated killing affects transformation, we cocultured the WT strain as the predator (without and with tfoX induction) with an environmental V. cholerae isolate as the prey [Sa5Y (28) (table S1)] in an intraspecies mixed-community assay. Comparative genomic hybridization data for strain Sa5Y (29) confirmed that this strain lacks the common O1 El Tor T6SS immunity genes tsiV1, tsiV2, and tsiV3 (30). To avoid premature killing of the predator strain by the prey before it could reach high cell density (which is required for competence- and QS-mediated induction of the T6SS), we inactivated the T6SS of strain Sa5Y (table S1).

Natural transformants were readily obtained upon TfoX-induction in predator cells (Fig. 3B and figs. S9 and S10). These natural transformation events were fully dependent on the competence co-regulators HapR and QstR and on ComEA (Fig. 3B and fig. S9), as these proteins are also required for natural transformation when purified genomic DNA serves as the transforming material (4, 1214) (fig. S11). Although the activation of the T6SS system seemed to be negligible in the latter case (fig. S11), transformants were undetectable or only rarely detectable in T6SS-defective strains when grown in a mixed community with strain Sa5Y (Fig. 3B and fig. S9). A similar T6SS-dependent increase in the transformation frequency was observed when the predator and prey strains were cocultured on chitinous surfaces (without artificial tfoX induction) (fig. S12). We concluded that upon competence induction, V. cholerae induced the T6SS and thus led to the killing of neighboring nonimmune bacteria. Lysis of neighboring bacteria causes release of genomic DNA that then transforms competent predatory cells. The enhancement of HGT in V. cholerae by the T6SS-mediated killing of nonimmune cells resembles bacterial fratricide described for naturally competent Streptococcus pneumoniae (31). However, in contrast to fratricide, which also promotes the lysis of noncompetent sibling cells (3, 31), competence-induced T6SS-mediated killing by V. cholerae appears to primarily target strains containing noncompatible effector-immunity modules and is contact-dependent. Earlier studies showed that the V. cholerae O1 El Tor strain C6706 is T6SS-silent (30) and unable to kill E. coli cells under standard laboratory conditions (32) (as shown in Fig. 3 for tfoX-uninduced conditions) owing to its inability to produce structural components of the T6SS (32). However, the tested strain was fully resistant against the T6SS-active strain V52, indicating that immunity is maintained even in the absence of T6SS activity (17, 32).

Our final goal was to visualize prey lysis and subsequent transfer of genetic material by live-cell fluorescence microscopy imaging. Thus, we used a translational fusion between the competence protein ComEA and the fluorescent protein mCherry (6). We previously demonstrated that ComEA of V. cholerae is a periplasmic protein that is highly mobile within this compartment (6). Moreover, ComEA is strictly required for DNA translocation across the outer membrane of competent cells and most likely contributes to the DNA uptake process by acting as a Brownian ratchet and compacting the incoming DNA within the periplasm (6). We therefore combined the comEA-mCherry and vipA-sfGFP alleles and incubated the resulting predator strain (after tfoX induction) with a gfp-tagged prey strain. Under those conditions, we observed high T6SS activity in the predator cells and, as a consequence, cell rounding and lysis of the prey (Fig. 4 and fig. S13). We also observed competent bacteria in close proximity to lysed cells exhibiting the distinctive focus formation of the ComEA-mCherry protein, which is indicative of DNA translocation into the periplasm (6) of a predator cell (Fig. 4 and figs. S13 and S14). Similar gene-transfer events were never or only rarely observed in the T6SS-negative strain; in the presence of extracellular deoxyribonuclease; if the predator lacked the outer membrane secretin protein PilQ, which is required for efficient DNA uptake (5, 6); or in the absence of any prey (figs. S14 and S15).

Fig. 4 TfoX-induced T6SS leads to the killing of neighboring cells followed by horizontal gene transfer.

This figure shows a time-lapse microscopy image series showing the T6SS-dependent attack of nonimmune prey cells followed by their rounding and lysis and the uptake of their DNA by neighboring competent predatory cells. A vipA-sfgfp– and comEA-mCherry–carrying V. cholerae strain (predator) was grown to high cell density under tfoX-inducing conditions, mixed with the gfp-labeled V. cholerae strain Sa5YΔvipA gfp (prey), and spotted onto agarose pads for imaging. The white arrows and arrowheads denote prey cells before and after attack (e.g., cell rounding), respectively, and the black arrows indicate cell debris after lysis [visualized in the phase contrast images (Ph; top row)]. DNA released by the prey lysis is taken up by the competent predator, as indicated by the gray arrowheads highlighting the redistribution of the periplasmic competence protein ComEA-mCherry (lower row). Accumulations of the VipA-sfGFP protein and assembled and contracted T6SS sheath structures are visible in the green channel (GFP). Scale bar: 2 μm (indicated in the upper left image and valid for all images).

Our findings indicated that the T6SS of V. cholerae is part of the competence regulon and is induced on chitinous surfaces in a TfoX-, HapR-, and QstR-dependent manner, thereby enhancing HGT (Fig. 5). HGT plays a major role in bacterial evolution and contributes to the spread of antibiotic resistance cassettes and pathogenicity islands. Moreover, because chitinous zooplankton are thought to play an important role in cholera transmission in endemic regions (1), chitin-induced expression of the T6SS might also enhance the virulence potential of the pathogen due to the killing of commensal bacteria within the human gut.

Fig. 5 Schematic summarizing the main findings of this study.

In its natural environment, V. cholerae often colonizes the exoskeleton molts of zooplankton. Upon initial attachment (I), the bacteria form a three-dimensional biofilm (II) on the chitinous surface, thereby reaching high cell density (HCD). Within these biofilms, different V. cholerae strains carrying diverse and noncompatible effector-immunity modules (shown by the different colors of bacteria) most likely form mixed communities (III). Under such chitin-dependent and HCD conditions, V. cholerae induces its natural competence program. In this study, we showed that the T6SS (indicated by the red arrow) is part of the competence regulon and induced upon the growth of V. cholerae on chitin substrates. The T6SS is active against nonimmune neighboring cells (IV), leading to their lysis (V). The DNA released by the lysed cells can be taken up by a competent predator cell (VI), leading to its natural transformation and evolution in the case that new DNA material is incorporated (VII).

Supplementary Materials

Materials and Methods

Figs. S1 to S15

Tables S1 to S5

References (3354)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank A. Boehm, M. Miller, members of the Institut National de Recherche Biomédicale of the Democratic Republic of the Congo, and M. Lo Scrudato for providing V. cholerae strains. We also acknowledge the service provided by Microsynth and members of the Blokesch laboratory for scientific discussions. This work was supported by the Swiss National Science Foundation (grant 31003A_143356) and the European Research Council (grant 309064-VIR4ENV). Supporting data are provided in the supplementary materials.
View Abstract

Navigate This Article