Correspondence: Brian K. Hammer, School of Biology, Institute of Bioengineering and Biosciences, Georgia Institute of Technology, Atlanta, GA 30332-0230, USA. Tel.: +1 404 385 7701; fax: +1 404 894 0519; e-mail: firstname.lastname@example.org
Vibrio cholerae, the causative agent of cholera and a natural inhabitant of aquatic environments, regulates numerous behaviors using a quorum-sensing (QS) system conserved among many members of the marine genus Vibrio. The Vibrio QS response is mediated by two extracellular autoinducer (AI) molecules: CAI-I, which is produced only by Vibrios, and AI-2, which is produced by many bacteria. In marine biofilms on chitinous surfaces, QS-proficient V. cholerae become naturally competent to take up extracellular DNA. Because the direct role of AIs in this environmental behavior had not been determined, we sought to define the contribution of CAI-1 and AI-2 in controlling transcription of the competence gene, comEA, and in DNA uptake. In this study we demonstrated that comEA transcription and the horizontal acquisition of DNA by V. cholerae are induced in response to purified CAI-1 and AI-2, and also by autoinducers derived from other Vibrios co-cultured with V. cholerae within a mixed-species biofilm. These results suggest that autoinducer communication within a consortium may promote DNA exchange among Vibrios, perhaps contributing to the evolution of these bacterial pathogens.
Vibrio cholerae, a common marine bacterium and the causative agent of the disease cholera, produces and then responds to extracellular small molecules called autoinducers (AIs) to collectively control gene expression and coordinate group behaviors, a process called quorum sensing (QS) (Fuqua et al., 1994; Ng & Bassler, 2009). Specifically, V. cholerae produces two autoinducers: CAI-I (the product of the CqsA synthase), which is restricted to Vibrios, and AI-2 (the product of the LuxS synthase), an interspecies autoinducer molecule produced by many bacteria (Chen et al., 2002; Xavier & Bassler, 2005; Higgins et al., 2007). At low cell density (low autoinducer levels) the phosphorylated response regulator LuxO activates transcription of multiple small RNAs that base-pair with and alter translation of several mRNAs, most notably repressing the translation of hapR, which encodes the master regulator of QS (Lenz et al., 2004; Hammer & Bassler, 2007; Svenningsen et al., 2009; Rutherford et al., 2011). At high cell density (high autoinducer levels), the binding of autoinducers to their cognate receptors results in dephosphorylation and inactivation of LuxO, leading to the production of HapR. HapR represses multiple genes, and also activates others, such as the gene coding for ComEA, a ssDNA-binding protein required for DNA uptake or horizontal gene transfer (HGT) (Meibom et al., 2005) (Fig. 1). Thus, wild-type (WT) V. cholerae strains are naturally competent at high cell density, a ΔhapR mutant does not take up DNA, and a ΔluxO strain that constitutively expresses HapR is capable of comEA-dependent DNA uptake even at low cell density (Meibom et al., 2005; Blokesch & Schoolnik, 2008). A V. cholerae-like QS pathway is well conserved in other Vibrio species, such as Vibrio harveyi, which also produces both CAI-1 and AI-2 (Hammer & Bassler, 2008).
Bacterial strains, plasmids, and culture conditions
The relevant genotypes of the Vibrio strains and plasmids used in the study are listed in Table 1. Vibrio cholerae and Vibrio parahaemolyticus strains were incubated at 37 °C on Luria–Bertani (LB) agar and in LB broth with shaking. In co-culture experiments with V. harveyi and Vibrio fischeri, the Vibrios were incubated at 30 and 28 °C, respectively, and the autoinducer donors were incubated on Luria–Marine (LM) agar for quantification, and in LM broth before co-culturing. The antibiotics (Fisher BioReagents) chloramphenicol (Cm), kanamycin (Kan), and streptomycin (Str) were used at concentrations of 10, 100, 5000 μg mL−1, respectively. Expression of the tfoX gene encoded on ptfoX was induced with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG; Fisher BioReagents).
Table 1. Bacterial strains and plasmids used in this study
Standard protocols were used for all DNA manipulations (Sambrook, 2001). Restriction enzymes, T4 DNA ligase (New England Biolabs), and Phusion DNA polymerase (Finnzymes) were used for cloning and PCR reactions. Standard methods were used to make deletion constructs (Skorupski & Taylor, 1996), as well as the KanRV. cholerae strain, which contained a copy of the KanR cassette from plasmid pEVS143 integrated at the lacZ site (Dunn et al., 2006). Genomic DNA from the V. choleraeΔlacZ∷KanR strain was extracted using a ZR Fungal/Bacterial DNA kit™ (Zymo Research) for experiments measuring the uptake of DNA. The luciferase-based reporter plasmid, pcomEA-lux, was constructed by PCR amplifying the promoter and a portion of the coding region of vc1917 from WT V. cholerae, and then cloning it into the pBBRlux vector (described in Lenz et al., 2004) by insertion into the SpeI and BamHI restriction sites. The IPTG-inducible ptfoX plasmid was constructed by amplifying the entire coding region of vc1153 and cloning it into the pEVS143 vector by insertion into the EcoRI and BamHI restriction sites. The primer sequences used for plasmid construction are available upon request.
Plasmid ptfoX was introduced by conjugation into V. cholerae strains carrying pcomEA-lux. For measurement of comEA-lux expression, V. cholerae strains carrying both plasmids were grown in LB with appropriate antibiotics at 37 °C overnight, diluted 1 : 1000 into fresh medium, and incubated for approximately 8 h. To measure comEA-lux expression in response to purified autoinducers, the V. cholerae autoinducer-deficient recipient was incubated as described above, but diluted 1 : 1000 into fresh medium containing purified CAI-1 alone, AI-2 alone, or both autoinducers at a final concentration of 10 μM, and incubated for 8 h. Purified autoinducers were prepared as described (Schauder et al., 2001; Higgins et al., 2007). Bioluminescence was measured using a Wallac model 1409 liquid scintillation counter as described previously (Hammer & Bassler, 2007). Relative light units (RLU) are defined as counts min−1 mL−1 OD600 nm−1. Single-time-point experiments were performed with triplicate samples.
Chitin-induced natural transformation assay
Chitin-induced transformation experiments were performed as described previously (Meibom et al., 2005). In transformation experiments with purified autoinducers, crab shells were inoculated with 2 mL of the V. cholerae autoinducer-deficient strain, and supplemented with purified autoinducers (each at 10 μM concentration) at the time of inoculation of the crab shells and again 24 h later along with 2 μg of genomic DNA marked with the KanR gene. In mixed-species transformation assays, crab shells were inoculated with the V. cholerae autoinducer-deficient recipient and the Vibrio autoinducer donor at a 1 : 1 ratio and incubated for 24 h. After addition of the marked genomic DNA, biofilms were grown for an additional 24 h before harvesting and plating to determine the transformation efficiency defined as KanR CFU mL−1 per total CFU mL−1 (as described previously in Meibom et al., 2005). In all mixed-species experiments, harvested cells were plated onto selective media to determine the total number of CFU and the number of transformants. Vibrio cholerae was selected on LB containing streptomycin. The HapR− (QS−) V. cholerae autoinducer donor strains (BH1543, EA093, EA094 and BH2104) used in the control co-culture experiments display a rugose colony morphology easily distinguishable from the V. cholerae autoinducer-recipient (Hammer & Bassler, 2003), and no KanR HapR− (rugose) colonies were detected in these transformation experiments. Because the V. harveyi, V. fischeri, and V. parahaemolyticus strains used are ampicillin resistant (AmpR) (and also StrS), these strains were selected on LM and LB containing Amp, respectively. For enumeration of transformants, cultures were plated onto LB medium containing kanamycin and streptomycin. Independent experiments were performed in triplicate.
Autoinducer-deficient mutants of V. cholerae are impaired in expression of the comEA gene and in DNA uptake
Previous studies with V. cholerae mutants (ΔhapR and ΔluxO) documented that in addition to the chitin controlled TfoX pathway, QS is required for the activation of comEA transcription (Meibom et al., 2005; Blokesch & Schoolnik, 2008) (Fig. 1). We introduced into V. cholerae strains a plasmid-borne transcriptional reporter gene fusion of comEA to the luciferase operon (pcomEA-lux), and an inducible tfoX plasmid (ptfoX) that alleviated the need for chitin in experiments monitoring comEA expression. As described previously, both WT V. cholerae and a ΔluxO mutant express comEA, while a ΔhapR mutant is ∼100-fold reduced in comEA expression (Fig. 2a). To define the role of autoinducer molecules in the regulation of the comEA gene, we next measured the expression of comEA-lux in V. cholerae mutants that produce only CAI-1 (ΔluxS), only AI-2 (ΔcqsA), or neither autoinducer (ΔcqsAΔluxS). The ΔluxS strain producing CAI-1 expressed pcomEA-lux at levels near WT, expression was further reduced for the ΔcqsA mutant that only produces AI-2, and the autoinducer-deficient mutant (ΔcqsAΔluxS) expressed comEA at levels similar to the QS-deficient ΔhapR mutant (Fig. 1). As expected, a ΔtfoX strain only activates comEA expression when induced to express TfoX from the plasmid, and the absence of TfoX induction reduced comEA expression in all strains to levels ∼100 lower than the ΔhapR mutant (Fig. 2a, white bars). Thus, TfoX is required for comEA transcription, and CqsA and LuxS together enhance expression ∼50-fold relative to the ΔhapR mutant. CAI-1 is the major autoinducer and AI-2 is the minor autoinducer for comEA transcription, as reported for V. cholerae virulence factor production in vivo (Higgins et al., 2007; Duan & March, 2010).
To quantify the contribution to DNA uptake of autoinducers produced by V. cholerae, we measured the transformation frequency of V. cholerae WT, ΔhapR, and ΔluxO strains using a crab-shell microcosm system described previously (Meibom et al., 2005). Transformation efficiency of WT and the ΔluxO mutant were maximal, and no transformants were detected with the ΔhapR mutant (Fig. 2b). The ΔluxS mutant, which produces CAI-1, was approximately fourfold impaired for transformation; however, both QS mutants (ΔcqsA and ΔcqsAΔluxS) that do not produce CAI-1 were severely compromised for transformation by ∼100-fold relative to WT (Fig. 2b). No transformants were obtained in the absence of extracellular KanR DNA, or when extracellular DNA was unmarked (data not shown), and in ΔtfoX (Fig. 2b), and ΔcomEA mutants, as described previously (Meibom et al., 2005). Thus, autoinducers produced by V. cholerae within a single-species biofilm promote DNA uptake. The discrepancy between the transformation frequency of the ΔcqsAΔluxS and the ΔhapR mutants may reflect that QS sRNAs constitutively expressed in the autoinducer-deficient strain do not completely eliminate all hapR mRNA (Bardill et al., 2011). Apparently, low levels of HapR protein can occasionally promote DNA uptake in this 24-h assay where rare transformation events may be amplified by replication. Alternatively, it is possible that the presence of chitin used for transformation measurements (Fig. 2b) may provide signaling information, in addition to CAI-1 and AI-2, that is different from conditions when comEA expression is measured without chitin in the presence of TfoX induction (Fig. 2a).
Purified autoinducer molecules activate the comEA gene and DNA uptake by V. cholerae
To test directly the role of autoinducers in comEA transcription and DNA uptake, purified CAI-1 and AI-2 where applied to the V. cholerae autoinducer-deficient ΔcqsAΔluxS mutant under the conditions described above. As shown for the V. cholerae autoinducer synthase mutants (Fig. 2a), the presence of both purified autoinducers (at saturating concentrations of 10 μM) resulted in maximal comEA expression by the autoinducer-deficient V. cholerae strain, and slightly lower levels were obtained when purified CAI-1 was provided alone (Fig. 3a). Expression was reduced further when only AI-2 was provided, and the lowest comEA transcription was observed when neither autoinducer was provided (Fig 3a). Likewise, a similar pattern was observed with the purified autoinducers in the crab-shell microcosm assay with the autoinducer-deficient V. choleraeΔcqsAΔluxS mutant. We suspect that the slightly lower levels of comEA expression observed when the autoinducers were produced by V. cholerae (Fig. 2a) compared with the results with purified autoinducers (Fig. 3a) may perhaps reflect lower levels of autoinducer synthesis and/or secretion in artificial sea water, conditions under which autoinducer production has not been quantified. Finally, by providing exogenous, purified CAI-1 and AI-2 to the chitinous biofilm (as described in Materials and methods), the autoinducer-deficient strain was capable of taking up DNA with a transformation efficiency similar to V. cholerae strains that produced their own autoinducers (Fig. 3b).
Autoinducers produced by other bacteria in a mixed-species biofilm activate comEA expression and DNA uptake by V. cholerae
Based on our results with the QS mutants and purified autoinducers (Figs 2 and 3), we hypothesized that V. cholerae might also sense and respond to autoinducers irrespective of their origin, including autoinducers derived from other Vibrios within in a mixed-species biofilm. We reasoned that a mixed-species consortium may more closely reflect conditions in environmental biofilms that are unlikely to be mono-species in composition (Hall-Stoodley et al., 2004; Wintermute & Silver, 2010). To demonstrate the feasibility of a mixed-species, crab-shell microcosm assay, the V. cholerae autoinducer-deficient recipient (ΔcqsAΔluxS) was co-cultured on chitinous crab shells with V. cholerae autoinducer-proficient donor strains that were HapR− (and thus QS−) but still capable of producing both autoinducers, only CAI-1, only AI-2, or neither autoinducer. The autoinducer-deficient V. cholerae recipient responded to both autoinducers derived from V. cholerae HapR− autoinducer donors within the biofilm and efficiently acquired extracellular DNA. Maximal transformation frequency occurred when the V. cholerae autoinducer recipient was provided with both autoinducers, while the response to only CAI-1 or only AI-2 was reduced. An autoinducer donor unable to produce either autoinducer promoted the lowest transformation frequency (Fig. 4). Similar results were obtained with several additional V. cholerae isolates that served as the CAI-1 and AI-2 donor (data not shown).
These results validated that in the crab-shell microcosm autoinducers derived from donor V. cholerae cells could promote comEA expression in a V. cholerae recipient; thus we monitored DNA uptake in the V. choleraeΔcqsAΔluxS autoinducer-deficient strain, co-cultured in a mixed biofilm with different Vibrio species serving as autoinducer donors. Indeed, in these mixed-species biofilms, V. cholerae acquired extracellular DNA in response to autoinducers secreted by bioluminescent V. harveyi (Fig. 4), which encodes a V. cholerae QS pathway (Hammer & Bassler, 2008). As with V. cholerae, the maximal transformation frequency occurred with the WT V. harveyi strain, which produces both CAI-1 and AI-2. Transformation decreases when only CAI-1 or AI-2 was provided, and was most impaired in the absence of either autoinducer (Fig. 4). We also measured transformation frequency of V. cholerae autoinducer-deficient recipient in response to WT V. parahaemolyticus and V. fischeri autoinducer donors. Transformation efficiency of these Vibrio strains followed a pattern of comEA-lux expression that matched the corresponding donor strains; the V. parahaemolyticus strain used produces both CAI-1 and AI-2 and promoted transformation with a frequency similar to V. harveyi. The V. fischeri strain tested (and another sequenced V. fischeri strain, data not shown) only encode for luxS (and not cqsA), and thus produce AI-2, but not CAI-1. Vibrio fischeri poorly promoted DNA uptake by the V. cholerae recipient (Fig. 4), consistent with AI-2 playing a minor role in natural transformation. Taken together, these observations support a model that V. cholerae can switch to the competent state and acquire DNA horizontally in a chitinous environmental biofilm by responding to autoinducer signals derived from members of the multispecies consortium.
Induction of the competence program in V. cholerae requires the chitin-responsive TfoX pathway and the autoinducer-responsive QS pathway. When both systems are functional, DNA uptake machinery facilitates the transport of extracellular DNA into the bacterial cell, where it may be incorporated into the genome by homologous recombination (Hamilton & Dillard, 2006). Many Vibrios encode for chitin utilization and competence genes (Hunt et al., 2008; Gulig et al., 2009; Ng & Bassler, 2009; Pollack-Berti et al., 2010), which suggests the possibility that natural transformation may be a conserved mechanism for both pathogenic and nonpathogenic Vibrios to horizontally acquire virulence and other genes within a community. Recognizing that many Vibrios possess V. cholerae-like QS circuits and produce CAI-1 and AI-2, we examined the relationship between autoinducers production and DNA uptake. Specifically, we showed that (1) V. cholerae efficiently activated a comEA-lux reporter in response to self-produced autoinducers as well as purified autoinducers and (2) a V. cholerae autoinducer-deficient strain readily acquires DNA when co-cultured with purified autoinducers and also with autoinducers produced by other Vibrios within a chitinous mixed-species biofilm. These results support a model that V. cholerae can switch to the competent state in a chitinous environmental biofilm by responding to autoinducer molecules derived from members of the multispecies consortium.
Communication via Vibrio autoinducer molecules has been studied in many laboratory systems that relied exclusively on cell-free culture fluids or monocultures (Bassler et al., 1997; Miller & Bassler, 2001; Henke & Bassler, 2004a), single-species co-cultures (Hammer & Bassler, 2007), or co-cultures of Vibrios with other bacteria unlikely to occupy the same environmental niches (Xavier & Bassler, 2005). These studies were not designed to reflect natural environmental setting that Vibrios typically encounter, such as the chitinous surfaces of animals (Lipp et al., 2002). However, mutants of V. cholerae (ΔhapR and ΔluxO), which regulate QS-controlled genes irrespective of autoinducer accumulation, provided the first demonstration of the role of QS in an animal model of cholera (Zhu et al., 2002), but do not directly demonstrate the role of extracellular autoinducer molecules. Only recently has secreted CAI-1 been shown to repress virulence in vivo (Duan & March, 2010). In a similar manner, we show here for the first time that extracellular CAI-1 and AI-2 molecules directly activate DNA uptake within a mixed-species environmental biofilm. Vibrio-specific CAI-1 appears to play a major role and interspecies AI-2 a minor role, suggesting that induction of DNA uptake may not be restricted exclusively to a response to autoinducers produced by Vibrio species, but that HGT may also be promoted by AI-2 derived from non-Vibrio members of a biofilm. Addition studies will be necessary to determine whether the behavior described here is cooperative ‘cross-talk’ between bacteria or whether V. cholerae simply uses the autoinducer molecules derived from others as a cue to alter gene expression (Diggle et al., 2007). It will also be interesting to determine whether additional chitinous materials that support growth of Vibrios and other bacteria in marine environments (Kaneko & Colwell, 1975; Sochard et al., 1979; Davis & Sizemore, 1982; Huq et al., 1983; Bartlett & Azam, 2005; Lyons et al., 2007) also stimulate autoinducer-induced DNA uptake (Bartlett & Azam, 2005).
Recent genomic comparison studies of multiple V. cholerae isolates suggest that substantial HGT events among Vibrio species may account for the presence of large ‘genomic islands’ of transferred DNA (Chun et al., 2009). Transduction of the cholera toxin genes encoded within a filamentous phage (CTXΦ) permits exchange of virulence factors among V. cholerae (Waldor & Mekalanos, 1996). In laboratory microcosms, DNA encoding antigenic determinants and also carrying CTXΦ occurs via chitin-induced HGT (Blokesch & Schoolnik, 2007; Udden et al., 2008) between V. cholerae. It is proposed that HGT among Vibrio species likely explains the current genome structures, but it has yet to be demonstrated whether chitin-induced HGT can promote DNA exchange among different Vibrios in environmental microcosms. We are currently performing experiments to test a model that autoinducers may promote interspecies HGT and emergence of genetic diversity in Vibrios.
The major restraint to genetic exchange between species occurs at the level of homologous recombination between the donor extracellular DNA and recipient genomic DNA (Heinemann, 1991). Recombination between partially homologous DNA depends on the extent and degree of DNA homology, which is monitored by the mismatch repair system (MMR) (Schofield & Hsieh, 2003). Genomic comparisons indicate that naturally occurring MMR-deficient environmental ‘mutator’ strains of V. parahaemolyticus have increased genetic and phenotypic diversity relative to clinical isolates, suggesting that such mutator strains are also ‘promiscuous’ for interspecies DNA uptake (Hazen et al., 2009). Inactivation of the MMR gene, mutS, enhances HGT between Escherichia coli and Salmonella typhimurium by up to three orders of magnitude (Rayssiguier et al., 1989); likewise a V. choleraeΔmutS strain we constructed was indeed capable of interspecies DNA uptake (data not shown). We are currently characterizing collections of environmental V. cholerae isolates for MMR, QS, and transformation proficiency to determine the role of autoinducer molecules in the emergence of genetic diversity of these marine bacteria.
We thank E. Stabb for V. fischeri and B. Bassler for purified CAI-1 and AI-2. We also thank the Hammer lab for discussions and critical manuscript review. This study was supported by a National Science Foundation grant (MCB-0919821) to B.K.H.