Chitin Induces Natural Competence in Vibrio cholerae

See allHide authors and affiliations

Science  16 Dec 2005:
Vol. 310, Issue 5755, pp. 1824-1827
DOI: 10.1126/science.1120096


The mosaic-structured Vibrio cholerae genome points to the importance of horizontal gene transfer (HGT) in the evolution of this human pathogen. We showed that V. cholerae can acquire new genetic material by natural transformation during growth on chitin, a biopolymer that is abundant in aquatic habitats (e.g., from crustacean exoskeletons), where it lives as an autochthonous microbe. Transformation competence was found to require a type IV pilus assembly complex, a putative DNA binding protein, and three convergent regulatory cascades, which are activated by chitin, increasing cell density, and nutrient limitation, a decline in growth rate, or stress.

Rivers, estuaries, and coastal waters are the principal reservoir for Vibrio cholerae in nature. In habitats of this kind, V. cholerae is found as a planktonic organism in the water column, in the mucilaginous sheaths of blue-green algae, and on the chitinous exoskeletons and molts of copepods (1). Population structure studies of aquatic habitats typically disclose ecosystems containing multiple microbial strains and species and high concentrations of phage and free DNA (2, 3). These features, when combined with mechanisms for HGT, likely explain why the Vibrionaceae have developed high levels of genomic diversity (4, 5).

One microenvironment where HGT could occur between V. cholerae and other strains and species is within microbial assemblages on natural chitin surfaces. V. cholerae readily attaches to and degrades chitin, a polymer of β-1,4–linked N-acetylglucosamine (GlcNAc). Chitin induces the expression of a 41-gene regulon involved in chitin colonization, digestion, transport, and assimilation, including genes predicted to encode a type IV pilus assembly complex (6).

V. cholerae has never been shown to be competent for natural transformation, and thus, with respect to HGT events, its genome is presumed to have evolved by transduction (responsible for the acquisition of the ctx genes encoding cholera toxin) and conjugation (7, 8). However, the induction of type IV pilin by chitin and the association of type IV pili and competence in several other species led us to test if chitin might induce natural competence in V. cholerae (9). We grew V. cholerae O1 El Tor, strain A1552, in a liquid minimal medium containing 2.5 mM (GlcNAc)6, a soluble chitin oligosaccharide that induces the chitin regulon (6). Then, genomic DNA from the V. cholerae O1 El Tor strain VCXB21, which harbors a gene for kanamycin resistance on the chromosome, was added to the culture. After 18 hours of growth, the culture was plated onto antibiotic-free and kanamycin-containing LB agar; this yielded a transformation frequency [kanamycin-resistant (Knr) colony-forming units (CFU)/total CFU] of 2.7 × 10–5 (Table 1). In the absence of donor DNA or when deoxyribonuclease (DNase) and donor DNA were added simultaneously, no Knr colonies were detected. Of other carbohydrates tested, including the chitin monomer GlcNAc, which does not up-regulate the chitin regulon, only chitin induced the competence phenotype. When glucose was combined with (GlcNAc)6, competence was inhibited, which suggests that the competence phenotype is subject to catabolite repression. Chitin-induced natural transformation with genomic DNA from the prototroph strain VCXB21 also restored prototrophy to two amino acid auxotrophic mutants that had deletions in either the proC or hisD gene and, thus, were unable to synthesize proline or histidine (Table 1). That the deleted version of the hisD gene was replaced by the wild-type copy from the donor DNA was shown by polymerase chain reaction (PCR) (fig. S1) and is indicative of homologous recombination. Together, these experiments showed that the growth of V. cholerae O1 with a soluble chitin oligosaccharide induced transformation competence and the capacity to acquire different genetic markers.

Table 1.

Transformation of V. cholerae; data are the average of at least three experiments. Transformation frequency is Knr or Strr CFU/total CFU; <DL, below detection limit (for values in (A), ∼4.0 × 10–8; for (B), ∼3.0 × 10–9; for (C), ∼1.0 × 10–7; for (D), ∼4.0 × 10–8).

Donor DNA Recipient strain Medium Transformation frequency Range
A. Transformation in liquid medium
VCXB21 A1552 + (GlcNAc)6 2.7 × 10-5 1.4 to 6.8 × 10-5
VCXB21 A1552 + (GlcNAc)6 + DNase <DL
None A1552 + (GlcNAc)6 <DL
VCXB21 A1552 + (GlcNAc)6 + Glucose <DL
VCXB21 A1552 + Glucose <DL
VCXB21 A1552 + GlcNAc <DL
B. Transformation in liquid medium of His and Pro auxotrophs
VCXB21 A1552proC + (GlcNAc)6 2.7 × 10-6 4.0 × 10-7 to 1.1 × 10-5
None A1552proC + (GlcNAc)6 <DL
VCXB21 A1552hisD + (GlcNAc)6 6.8 × 10-6 4.0 × 10-8 to 1.7 × 10-5
None A1552hisD + (GlcNAc)6 <DL
C. Transformation of chitin surface—attached bacteria
VCXB21 A1552 Crab shell 3.5 × 10-5 5 × 10-6 to 6.9 × 10-5
None A1552 Crab shell <DL
N16961 A1552 Crab shell 1.8 × 10-5 5 × 10-6 to 4.4 × 10-5
VCXB21 A1552 Crab shell + DNase 3.7 × 10-7 <DL to 8.0 × 10-7
VCXB21 N16961 Crab shell <DL
VCXB21 C6706 Crab shell 2.8 × 10-6 8.0 × 10-7 to 5 × 10-6
VCXB21 0395 (classical) Crab shell <DL
D. Transformation in biofilm communities without exogenous DNA
None A1552 -Kn/-Str Crab shell 4.4 × 10-5 1.4 to 8 × 10-5
None A1552 -Kn/-Str Crab shell + DNase <1.2 × 10-7 <DL to 1.2 × 10-7

In nature, V. cholerae experiences chitin as an insoluble polymer provided as a structural component of copepod exoskeletons and molts on which it grows as surface-attached colonies or as a biofilm (1). To determine whether growth on a natural chitin surface induced the competence phenotype, strain A1552 was allowed to establish a surface-attached population on a sterile crab shell fragment in defined (artificial) seawater medium (DASW) for 24 hours. Then, the crab shell was immersed in fresh DASW containing antibiotic-marked genomic DNA from V. cholerae O1 strain VCXB21 (Knr) or strain N16961, which is streptomycin-resistant (Strr). These experiments yielded transformation frequencies of 3.5 × 10–5 and 1.8 × 10–5, respectively; in the absence of added DNA, no transformants were obtained (Table 1). The addition of DNase reduced transformation frequency ≥100-fold, but did not abolish it entirely, possibly because of the resistance of surface-adsorbed DNA to DNase (10) or to the use of conditions in the transformation assay that are not optimal for the activity of this enzyme. Thus, during growth on a chitin surface, V. cholerae becomes competent for transformation by genomic DNA from other V. cholerae strains.

Some naturally competent species release DNA to an extent that can lead to localized concentrations in a biofilm that exceed 100 μg/ml (11). Simultaneously, transformation competence develops among other members of the population (12, 13). To determine whether V. cholerae can be transformed when DNA is provided by other members of a surface-attached consortium, two variants of the same competent V. cholerae strain (A1552), with different antibiotic resistance markers on the chromosome (-Knr or -Strr), were propagated together on a crab shell fragment (9). Twenty-four hours later, planktonic bacteria were discarded and fresh DASW added; 24 hours thereafter, the bacteria were detached from the crab shell surface and plated onto LB agar and LB agar containing both Kn and Str, then the colonies were counted. Experimental replicates yielded transformants resistant to both antibiotics with an average frequency of 4.4 × 10–5 (Table 1). The addition of DNase to the crab shell culture reduced transformation efficiency by ≥100-fold (Table 1). Thus, naked DNA is apparently provided in situ by V. cholerae growing on a chitin surface and can be acquired by competent members of the consortium.

Twelve chitin-induced genes are predicted to encode components of a type IV pilus assembly complex (6). To determine whether a type IV pilus is required for competence, the following genes were disrupted in V. cholerae strain A1552: pilA (VC2423), predicted to encode a type IV pilin; pilQ (VC2630), encoding a homolog of the secretin protein family; and pilB (VC2424), which specifies a traffic nucleotide triphosphatase(NTPase) believed to energize assembly of the pilus filament. The wild-type parent and each of the mutants were tested for competence with the crab shell transformation assay and DNA from the Knr strain VCXB21. The wild-type parent exhibited a transformation frequency of 3.5 × 10–5; no Knr transformants were detected for any of the mutants (Table 2). Thus, at least three components of a putative type IV pilus assembly complex are required for competence. This finding prompted us to use transmission and scanning electron microscopy (TEM and SEM, respectively) to search for new or additional pilus filaments emanating from the surface of chitin-induced bacteria. None was identified, which indicates that this assembly complex likely directs the synthesis of a competence pseudopilus (14).

Table 2.

Transformation of V. cholerae O1 El Tor strains and mutants in a crab shell–associated community. Transformation with DNA from V. cholerae strain VCXB21; data are the average of at least three experiments. Transformation frequency is KnR CFU/total CFU; <DL: below detection limit (1.0 × 10–7). Similar results were obtained using liquid medium with (GlcNAc)6 as carbon source.

Strain Genotype Transformation frequency
A1552 O1 El Tor, wt 3.5 × 10-5
A1552pilA A1552ΔVC2423 <DL
A1552pilB A1552ΔVC2424 <DL
A1552pilQ A1552ΔVC2630 <DL
A1552tfoX A1552ΔVC1153 <DL
A1552VC1917 A1552ΔVC1917 <DL
FY1 A1552ΔrpoS <DL
ATN140 A1552ΔhapR <DL
ATN194 A1552ΔhapR::mTn7hapR 1.3 × 10-5
N16961 O1 El Tor <DL
N16961ChapR N16961::mTn7hapR 7.5 × 10-6

We used BLAST to search among the 41 chitin-induced genes for homologs of competence genes in other species. One of these, VC1153 (further named tfoXVC), was found to contain two TfoX domains, which are involved with competence regulation in Haemophilus influenzae (15, 16). To determine whether tfoXVc is required for competence, it was disrupted in V. cholerae strain A1552, and the mutant was tested for competence by using the crab shell transformation assay. No transformants were detected (Table 2). To determine whether tfoXVc expression would promote transformation competence in the absence of chitin, tfoXVc was placed under the control of an arabinose-inducible promoter, and then either the vector alone (pBAD) or the recombinant plasmid (pBAD-tfoX) was introduced into the A1552tfoX mutant. Growth of the complemented mutant in LB liquid medium containing arabinose (but no chitin) led to chitin-independent overexpression of tfoXVc. The addition of genomic DNA from V. cholerae strain VCXB21 to this culture yielded Knr transformants at a frequency of 2.9 × 10–6. By contrast, under the same growth conditions, neither the wild-type parent nor the A1552tfoX mutant was competent (Table 3).

Table 3.

Expression of tfoXVC allows transformation in the absence of chitin. Transformation with DNA from strain VCXB21 in LB media supplemented with ampicillin and 0.2% l-arabinose; data are the average of two independent experiments. Transformation frequency, Knr CFU/total CFU; <DL: below detection limit (2.0 × 10–8).

Strain Plasmid Transformation frequency
A1552 pBAD <DL
A1552tfoX pBAD <DL
A1152tfoX pBAD-tfoX 2.9 × 10-6

The demonstration that TfoXVc is required for competence and the role of its ortholog in H. influenzae as a regulator of competence prompted us to use microarray expression profiling to identify genes it might regulate in V. cholerae. Transcriptional profiles were obtained by using RNAs isolated from the A1552tfoX mutant described above, which harbors either pBAD (does not express tfoXVc) or pBAD-tfoX (arabinose-induced expression of tfoXVc) during mid log phase growth in LB liquid medium containing arabinose, but no chitin. Under these conditions, 99 genes were significantly and ≥2.5-fold induced in the culture expressing tfoXVc compared with the control culture (table S1). Among the TfoXVc-induced genes were 28 that were previously reported to be induced by chitin, including the three pilus assembly genes, pilA, pilB, and pilQ; genes encoding four chitinases (including ChiA-1 and ChiA-2); and a chitoporin gene. Thus, TfoXVc is induced by chitin and controls the expression of genes encoding proteins with two quite different functions: chitin degradation and chitin-induced competence.

The foregoing result encouraged us to search for other TfoXVc-regulated genes that might be required for competence. This led to the identification of VC1917, predicted to encode a protein with a signal peptide and a motif homologous to the DNA-binding helix-hairpin-helix domain found in the Bacillus subtilis ComEA protein (17). When VC1917 was disrupted and the mutant tested for competence, no transformants were obtained (Table 2). This shows that VC1917 is required for competence in V. cholerae.

During the course of this study, we tested a total of eight V. cholerae strains for transformation competence: four V. cholerae O1 strains and four recent V. cholerae non-O1 environmental isolates. Strain C6706, an El Tor biotype, and each of the environmental isolates exhibited chitin-dependent competence. By contrast, neither strain N16961, an O1 El Tor biotype for which a genome sequence is available, nor strain O395, a V. cholerae O1 classical biotype, was found to be competent under the same experimental condition (Table 1). Both of these nontransformable strains are reported to have a frameshift mutation in hapR (18), whose protein product coordinately down-regulates the expression of virulence determinants and biofilm formation and up-regulates hemagglutinin/protease (HA/protease) production in response to increasing cell density. To determine whether the reported frameshift mutation in hapR accounts for the competence-negative phenotype of strain N16961, it was complemented with a wild-type copy of hapR from the transformable strain A1552. When tested in the crab shell transformation assay, the complemented strain, N16961ChapR, showed a transformation frequency of 7.5 × 10–6 (Table 2). To further examine the requirement of hapR for the competence phenotype, it was disrupted in A1552. The resulting mutant was transformation-negative (Table 2). Together, these data show that HapR is required for transformation competence.

The expression of hapR is positively controlled by the alternative sigma factor RpoS (19, 20). To find out if it is also required for transformation competence, an rpoS mutant was tested using the crab shell transformation assay and found to have a transformation-negative phenotype (Table 2). Thus, RpoS likely modulates transformation efficiency through its effect on hapR expression. The natural effectors of increased RpoS abundance in a chitin-associated biofilm were not identified. However, plausible candidates include nutrient limitation as population density increases, growth deceleration, or stress (20).

At low cell densities, hapR expression is repressed by the quorum-sensing regulator phospho-LuxO; at high cell densities, LuxO is dephosphorylated, the repression is relieved, and HapR is synthesized (18, 21). Because HapR is required for competence, we tested the effect of cell density on the competence phenotype. luxO was disrupted in A1552 and the mutant compared with the wild-type strain using the crab shell transformation assay. The average transformation frequency of the luxO mutant was about five times that of the wild-type parent. We reasoned that the effect of a luxO mutation on transformation efficiency would be most evident at low cell densities. To correlate bacterial cell density with transformation frequency, genomic DNA from KnR strain VCXB21 was added at 0, 2, 4, or 6 hours after inoculation of the crab shell, time intervals that correspond to increasing cell densities on the crab shell surface. Two hours after the addition of genomic DNA, the assay was treated with DNase to degrade residual donor DNA. Then the culture was allowed to grow for a total of 24 hours, and the experiment was scored as transformation positive or negative (9). KnR transformants were first evident more than 4 hours after we inoculated the crab shell assay with the wild-type strain. By contrast, the luxO mutant was transformable at lower cell densities, with some transformants noted between 0 and 2 hours after inoculation; many more transformants were evident after 2 to 4 hours (Fig. 1). These results show that high cell density positively controls transformation competence by relieving LuxO-dependent repression of HapR synthesis.

Fig. 1.

Transformation frequency as a function of cell density. Transformation was scored as positive or negative for the crab shell–associated wild type (gray bars) or luxO mutant (black bars) of strain A1552 (9). Donor DNA from strain VCXB21 was present during the postinoculation time intervals indicated on the x axis. Transformation (y axis) is calculated as the percentage of experiments that were scored as transformation-positive, based on results from two replicates of six independent experiments.

The data presented in this report have led us to propose a model of the V. cholerae competence regulatory network that posits three controlling environmental determinants of this phenotype. These are the presence of chitin; increasing cell density; and, nutrient limitation, growth deceleration, or stress (fig. S2). Chitin, acting through TfoXVc, induces the expression of a competence pseudopilus, as well as genes required for the extracellular degradation and uptake of chitin. Increasing cell density, in combination with effectors of heightened RpoS abundance, strengthen the expression of hapR, shown here to be required for the positive regulation of competence. Remarkably, both HapR and TfoXVc are required for the expression of the VC1917 gene (tables S1 and S2) and thus the three environmental determinants of competence converge in the regulation of this and perhaps other genes.

This model highlights the importance of natural competence, occurring in chitin-attached bacterial communities in the aqua sphere, as a powerful driver of the evolution of V. cholerae. It further suggests that environmental events giving rise to copepod blooms likely foster the rapid evolution of this pathogen.

Supporting Online Material

Material and Methods

Figs. S1 and S2

Tables S1 to S3


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article