Regulatory elements involved in the expression of competence genes in naturally transformable Vibrio cholerae

Background The human pathogen Vibrio cholerae normally enters the developmental program of natural competence for transformation after colonizing chitinous surfaces. Natural competence is regulated by at least three pathways in this organism: chitin sensing/degradation, quorum sensing and carbon catabolite repression (CCR). The cyclic adenosine monophosphate (cAMP) receptor protein CRP, which is the global regulator of CCR, binds to regulatory DNA elements called CRP sites when in complex with cAMP. Previous studies in Haemophilus influenzae suggested that the CRP protein binds competence-specific CRP-S sites under competence-inducing conditions, most likely in concert with the master regulator of transformation Sxy/TfoX. Results In this study, we investigated the regulation of the competence genes qstR and comEA as an example of the complex process that controls competence gene activation in V. cholerae. We identified previously unrecognized putative CRP-S sites upstream of both genes. Deletion of these motifs significantly impaired natural transformability. Moreover, site-directed mutagenesis of these sites resulted in altered gene expression. This altered gene expression also correlated directly with protein levels, bacterial capacity for DNA uptake, and natural transformability. Conclusions Based on the data provided in this study we suggest that the identified sites are important for the expression of the competence genes qstR and comEA and therefore for natural transformability of V. cholerae even though the motifs might not reflect bona fide CRP-S sites.


Background
Vibrio cholerae is a Gram-negative bacterium that often lives in aquatic environments in association with the chitinous exoskeleton of zooplankton [1,2]. Chitin, a polymer of β-1,4-linked N-acetylglucosamine, is one of the most abundant biopolymers in nature [3]. In addition to its role as a nutrient source, chitin also induces natural competence for transformation in V. cholerae [4] and other Vibrio species (reviewed by [5]).
Natural competence is a mode of horizontal gene transfer, which is based on the ability of a bacterium to take up free DNA from the environment and recombine it with the bacterial genome resulting in natural transformation. In V. cholerae, chitin leads to the up-regulation of tfoX [4,6] (Figure 1). This gene encodes a protein that is the master regulator of transformation and a homolog of Sxy, which was first described in Haemophilus influenzae [7,8]. Indeed, tfoX expression is sufficient to induce natural competence and transformation in V. cholerae even in the absence of chitin [4,9,10]. In our current working model, the components of a type IV pilus combined with a few other competence proteins (such as ComEA, ComEC, and ComF) make up the majority of the DNA-uptake machinery. This machinery is responsible for binding to and pulling extracellular DNA into the periplasm of V. cholerae and subsequently, into the cytoplasm [5,[11][12][13][14] as previously suggested for other naturally competent bacteria (reviewed for example by [15][16][17][18][19]).
The third pathway involved in the regulation of natural competence for transformation is CCR [4,23]. This term indicates the mechanism by which, in the presence of a preferred carbon source such as glucose, the expression of genes necessary for the metabolism of other carbon sources is repressed [37]. The major players in CCR are the phosphoenolpyruvate-carbohydrate phosphotransferase system (PTS), adenylate cyclase (CyaA), the metabolite 3′,5′-cyclic adenosine monophosphate (cAMP) and the CRP protein. Unsaturated PTS transporters enhance the synthesis of cAMP by CyaA. High levels of cAMP trigger the formation of an active complex of CRP and cAMP, which binds the promoters of the target genes (e.g., those genes encoding proteins that are involved in the transport and utilization of alternative carbon sources). Conversely, when the PTS is saturated cAMP is not produced and the CRP-cAMP complex cannot form. Central metabolism and transport of the carbon sources are not the exclusive targets of CCR; cAMP and the CRP protein, as well as the PTS components (independent of cAMP), also control biofilm formation in V. cholerae [38][39][40]. With respect to natural competence for transformation, the presence of PTS sugars significantly decreases the transformability of V. cholerae; moreover, knockout strains for crp or cyaA are non-transformable [23].
The role and function of the CRP protein have been primarily studied in E. coli (reviewed by [41,42]). CRP, formerly known as catabolite activator protein (CAP), forms a dimer of two identical subunits. Each CRP subunit contains an N-terminal cAMP binding domain, a flexible hinge region and a C-terminal helix-turn-helix DNA binding motif. CRP recognizes and binds 22 bp-long symmetrical sequences called CRP sites. Under physiological conditions, CRP is likely present either as a free apo-CRP dimer (in the absence of cAMP) or as a dimer with each subunit bound to a molecule of cAMP. V. cholerae CRP and E. coli CRP (EcCRP) share 95% identity in amino acid sequence [43]. As with EcCRP, V. cholerae CRP displayed a biphasic dependence on cAMP levels in vitro. Moreover V. cholerae CRP is able to activate the transcription of E. coli gal promoters [44,45]. These findings strongly suggest that the CRP protein functions similarly in these two bacterial species.
In H. influenzae, the expression of the competence genes requires the CRP-cAMP complex [46,47] along with Figure 1 Proposed working model describing the transcriptional regulation of the competence genes qstR and comEA within the network of natural competence for transformation of V. cholerae. qstR and comEA are tightly regulated by three different pathways: quorum sensing, chitin-sensing/degradation, and carbon catabolite repression (signals indicated on top; from left to right). Upon growth on chitin or artifical induction, TfoX is produced within the cell. Furthermore, unsaturated PTS systems lead to the activation of the enzyme adenylate cyclase, which synthesizes cAMP. cAMP then forms a complex with CRP. Based on work performed in H. influenzae [20,21], we hypothesized that CRP-cAMP might bind to the putative CRP-S sites described here and that such binding would be dependent on TfoX. However, full expression of those genes requires a QS-dependent secondary activator (HapR and QstR, respectively). At high cell density the master regulator of QS, HapR, is produced and allows the expression of qstR by directly binding to its promoter (indicated by an asterisk marking a HapR binding site identified in vitro; [22]. The transcriptional regulator QstR, which might require a cofactor for its full activity (indicated by the triangle), positively regulates the comEA gene potentially by binding to a QstR-specific binding motif (indicated by the dashed box and so far unidentified). The question marks indicate the subjects addressed in this study. the master regulator of transformation Sxy [7] (TfoX in V. cholerae). The competence regulon of H. influenzae consists of genes characterized by the presence of competence regulatory elements (CRE) [48,49]. Due to their Sxy dependency, these specific competence-related CRP binding sites were later renamed CRP-S sites to distinguish them from the canonical Sxy-independent CRP-N sites [20]. Indeed, Cameron and Redfield suggested that in H. influenzae, and most likely in other competent Gram-negative bacteria, the induction of competence genes is under the control of CRP and Sxy/TfoX acting in concert at the CRP-S site [20]. Based on previously published expression data from V. cholerae [4,6] Cameron and Redfield also predicted a 22-bp CRP-S and CRP-N consensus motif for the Vibrionaceae family in silico, which was nnntTTnAAnTnnnTCGAAnnn for CRP-S and tnntGTGAnnnnnnTCACanan for CRP-N (the most common bases are indicated in upper case, the less likely bases in lower case; 'n' refers to any base even though minor preference might be valid for some of these positions; for details see [20]).
In this study, we tested the role of QstR as an activator of comEA. We demonstrated that overexpression of qstR was sufficient to increase the abundance of the comEA transcripts, although not to the same level observed under competence-inducing conditions. We therefore hypothesized that TfoX and CRP-cAMP were also involved in driving the expression of comEA and qstR ( Figure 1). In agreement with this hypothesis, we identified putative CRP-S sites in the promoter regions of qstR and comEA, which were not part of the in silico predicted Vibrionaceae CRP-S sites described above [20]. We investigated the importance of these motifs using site-directed mutagenesis, followed by the analysis of the respective mutants. Our results suggest that these sites are important for the transcriptional regulation of the respective competence genes but might not represent bona fide CRP-S sites.

Bacterial strains and plasmids
Vibrio cholerae strains and plasmids used in this study are listed in Table 1. Escherichia coli strains DH5α [50] and One Shot PIR1 or PIR2 (Invitrogen) were employed as hosts for cloning. E. coli strain S17-1λpir [51] was used as a mating donor for plasmid transfer between E. coli and V. cholerae.

Media and growth conditions
V. cholerae and E. coli strains were grown at either 30°C or at 37°C. Overnight cultures were grown in LB medium under aerobic conditions. Thiosulfate Citrate Bile Salts Sucrose (TCBS) agar plates were used to counterselect E. coli strains after triparental mating with V. cholerae strains. The TCBS agar plates were prepared following the manufacturer's instructions (Fluka). For plasmid maintenance or selection of transformants/transconjugants, antibiotics were added to the growth media at concentrations of 50 or 100 μg ml −1 for ampicillin, 75 μg ml −1 for kanamycin, and 50 μg ml −1 for gentamicin.

Construction of V. cholerae mutant strains
Chromosomally-encoded site-directed mutants were generated using the previously described TransFLP method [53,55,60]. The extended FRT scar was located downstream of the native comEA gene and upstream of the native qstR gene. In the latter case, a control strain was designed that was not modified in a site-directed manner but solely contained the integrated FRT scar upstream qstR (strain WT_qstR (FRT control)). This strain behaved as the WT with respect to the expression pattern, natural transformation, DNA uptake, and general growth behavior (described below).

Construction of plasmids
The majority of the plasmids carrying the comEA gene preceded by a certain length of its upstream region were made through inverse PCR, using the plasmid pBR-[own] comEA [22] as template and the oligonucleotides listed in Table 2. The plasmid pBR-[−134]comEA was constructed by cloning the PCR-amplified insert [−134]comEA into the NcoI site of pBR-Tet_MCSII [22].

Natural transformation assay (chitin-dependent and chitin-independent)
Natural transformation assays were performed as previously described, growing V. cholerae strains on chitin flakes with a medium change on day two [54], or in LB medium supplemented with 0.02% arabinose to express an inducible chromosomal copy of tfoX (preceeded by a P BAD promoter; [10]). The same growth conditions were used for in trans over-expression of the plasmid-encoded qstR gene. Notably, V. cholerae does not contain any obvious homolog of the low-affinity high-capacity arabinose transporter AraE, which is involved in the all-or-none induction of genes preceeded by the arabinose-inducible promoter P BAD in E. coli [61]. Statistical analyses of transformation data were carried out on log-transformed data [62] using a two-tailed Student's t-test.

SDS-PAGE and Western blotting
Proteins were separated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (SDS-PAGE) and then subjected to western blotting as previously described [10]. The primary antibody against ComEA (GP 1248; see below) and horseradish peroxidase (HRP)conjugated goat anti-rabbit secondary antibody (Sigma-Aldrich, Switzerland) were diluted at 1:5,000 and 1:20,000, respectively. Luminescent signals were produced and detected by Western Lightning-ECL (PerkinElmer) and chemiluminescence-detecting film (Amersham Hyperfilm ECL, GE Healthcare).

Generation of the antibodies against ComEA
Rabbit anti-ComEA antibodies were raised against synthetic peptides and produced by Eurogentec (Belgium). The antibody was tested in Western blot analysis against the comEA knockout strain to exclude potential crossreactions with proteins migrating towards the same position as the target protein.

Quantitative reverse transcription PCR (qRT-PCR)
V. cholerae strains were grown in LB medium supplemented with 0.02% arabinose to induce the qstR gene or natural competence. RNA preparation, DNase treatment, reverse transcription, and qPCR were performed as previously described [10,22,23].

5′ Rapid amplification of cDNA ends (5′RACE)
V. cholerae wild-type strain A1552-TntfoX was induced to competence as described above. Cell harvesting and RNA preparation were performed as previously published [22]. The 5′/3′ RACE Kit 2nd Generation (Roche) was used to identify the transcription start of the comEA gene. All steps were performed according to the manufacturer's protocol unless stated otherwise. Total RNA (2 μg) and the gene-specific primer comEA_284_rev were used to synthesize the first strand cDNA of comEA. The cDNA was then purified using the High Pure PCR Product Purification Kit (Roche, Switzerland). After addition of the Poly(A) tail to the 3′ end, the first strand cDNA was amplified by PCR with the gene-specific primer comEA_217_rev and oligo dT-Anchor primer. The PCR products were visualized on an agarose gel and purified. The double-stranded cDNA of comEA was further amplified by PCR with the primers F-EcoRI_Anchor_P and R-BamHI_comEA_217. The PCR products were cloned into the EcoRI/BamHI sites of the plasmid pBR-Tet_MCSI [10]. To determine the transcription start point of comEA, fifteen of those plasmids were sequenced. 12 out of those 15 sequences pointed to the C at position −24 bp upstream the comEA start codon as transcriptional start site.

Regulation of comEA by QstR and TfoX/CRP-cAMP
From previous studies it was known that the expression of comEA is dependent on a) the master regulator of transformation, TfoX; b) the CRP-cAMP complex; and c) the transcription factor QstR [4,22,23] (Figure 1). Indeed, we previously demonstrated that the QstR protein is required for proper expression of the two competence genes, comEA and comEC [22], both of which are essential for DNA uptake and natural transformation [12]. We also speculated that QstR might require a cofactor for proper binding to the respective promoter regions [22]. Here, we asked whether TfoX only regulates comEA indirectly through its influence on QstR, or whether it also has a role in directly regulating the expression of comEA (Figure 1). To do so, we measured the transcripts of comEA in the absence of natural competence induction (e.g., in the absence of a chitin surface and without artificial induction of the gene encoding the main regulator of transformation, TfoX) but with in trans overexpression of qstR. Using this approach, we observed an increased level of comEA expression, whereas expression of the upstream acting regulatory gene hapR and the competenceunrelated but HapR-activated gene hapA did not change as expected (Figure 2). However, the relative expression of comEA appeared lower than what was measured when natural competence was concomitantly induced even though the expression of qstR was significantly higher due to the multi-copy effect of the expression in trans (compare to [22] and data shown below). Next, we investigated the production of the ComEA protein in order to evaluate whether the transcript levels would also reflect the protein levels. Apart from the strains used to measure the comEA transcripts, we also included wild-type strains harboring an inducible copy  [59].
of tfoX on the chromosome as previously described [10] and grew the strains under both competence-inducing and non-inducing conditions. Total protein extracts were prepared from bacterial strains that had reached the high cell density state. As shown in Figure 2B, the ComEA protein was readily detectable in the competence-induced strain (lane 2) and absent in the same strain that was grown without competence induction (lane 1). Moreover, we detected low levels of ComEA in the wild-type strain when qstR was overexpressed but competence (e.g., tfoX) was not induced (lane 4), which was not the case for the vector control (lane 3). This result strengthened the evidence that QstR per se is able to drive the expression of comEA in the absence of competence induction. However, the protein was more abundant in the wild-type strain under competence-inducing conditions, suggesting that full comEA expression requires more than the QstR protein alone, which is consistent with the proposed dual regulation of comEA by TfoX / CRP-cAMP complex and QstR (Figure 1).

Narrowing down the promoter region driving the expression of comEA
To better understand how the expression of comEA is regulated, we first mapped the putative promoter region of this gene. To do so we constructed eight plasmids carrying the gene and its upstream region, which we incrementally shortened (Figure 3). The plasmids were originated using direct or inverse PCR and plasmid pBR-[own]comEA as template [22]; Table 1). This plasmid contains the comEA gene and a 900 bp sequence that is upstream of the start codon and therefore includes the promoter region [22]. All eight constructs were tested for their ability to restore the transformability of the V. cholerae comEA-minus strain (ΔcomEA) in trans after chitin-dependent induction of competence ( Figure 3). ΔcomEA strains harboring plasmid-encoded comEA and at least 134 bp of its upstream region were complemented and showed transformation frequencies comparable to the wild-type strain. Neither of the bacterial strains harboring either 100 bp or 40 bp of the comEA upstream region rescued transformability, nor did the strain carrying the empty vector ( Figure 3). We concluded that 134 bp of sequence upstream of the comEA start codon are sufficient to drive comEA expression.
Prediction of putative promoter elements within the comEA upstream region Next, we localized the transcription start site of the comEA transcript using 5′ RACE (Rapid amplification of cDNA ends). As schematized in Figure 4A, we could map the transcription start site at −24 bp upstream the start codon. In E. coli, the majority of the promoters consist of two conserved hexanucleotides, which are located at approximately −35 and −10 from the transcription start site. The consensus of a "typical" promoter of E. coli has been summarized as follows: TTGACA-(N [15][16][17][18][19] )-TATAAT-(N 5-7 )-start [63] (N = any nucleotide). The −35 and −10 regions are specifically recognized and bound by the σ subunit of the RNA polymerase (RNAP) holoenzyme [64], and the most commonly used σ factor in E. coli is σ 70 [65]. Comparing the sequence upstream of the transcription start site of comEA with the consensus of the E. coli σ 70 -activated promoter [65], we identified the putative −35 and −10 regions of the comEA promoter, TTTCCA-(N 16 )-TATCAT-(N 7 )start, as shown in Figure 4A. These regions were also predicted using the BPROM program, which works to recognize bacterial σ 70 promoters [66]. Moreover, we manually screened the upstream region of comEA and identified a motif similar to the in silico predicted consensus of the Vibrionaceae CRP-S site [20] ( Figure 4A and B). As described above, the CRP-S site is hypothesized to represent a CRP binding motif where the protein binds to in conjunction with TfoX. This putative CRP-S site is 22 bp in length and located between −79 bp and −58 bp upstream of the comEA start codon ( Figure 4A).

Investigation of the putative CRP-S site upstream of comEA
To understand the importance of this putative CRP-S site, we modified its sequence using site-directed mutagenesis at the original chromosomal locus. The sequences of the putative CRP-S site upstream of comEA, as well the site-directly changed CRP sites are represented in Figure 4B. Specifically, we either deleted the entire putative CRP-S site (ΔCRP-S), inverted it (CRP-S_inv), changed it to the consensus of the canonical CRP binding site, which works independently of TfoX (CRP-N), or to the in silico predicted CRP-N motif of the V. cholerae fumarate reductase gene [20] ([frdA] site), or we altered it to a motif that differed considerably from the original CRP-S consensus in the 3′ half of the motif (designated CRP-0) ( Figure 4B). When we tested these modified strains, we observed a perfect correlation between natural transformability ( Figure 5A) and the expression levels of comEA ( Figure 5B). The (non-) functionality of the CRP-S motifs in driving comEA expression was also confirmed by testing DNA uptake using a recently developed whole-cell duplex PCR assay [9,12,13] ( Figure 5C). Notably, all of these assays confirmed that the absence of the CRP-S site lowered the comEA expression level followed by reductions in ComEA-mediated DNA uptake ability and eventual transformation ( Figure 5). Interestingly, the mutant carrying the inverted CRP-S site turned out to be only mildly impaired in transformation ( Figure 5A) due to reduced expression of comEA ( Figure 5B) and reduced DNA uptake ( Figure 5C). Most importantly, the replacement of the CRP-S site by both CRP-N derivatives (CRP-N and [frdA] site) resulted in similar phenotypes as those seen with the CRP-S-minus strain. The CPR-0 motif mutant on the other hand displayed the highest expression level for comEA, which resulted in increased DNA uptake and transformation ( Figure 5).

The CRP-S site upstream of qstR
From the data presented above, we hypothesized that the putative CRP-S motif indeed plays a role in the regulation of comEA, but potentially not in the same way that would be expected from a bona fide CRP-S site. Notable, expression of comEA also requires QstR, which links the QS pathway with the competence genes [22]. Because qstR expression occurs in a TfoX-dependent manner [22], we speculated that the gene might also be preceded by a putative CRP-S site. Indeed, upon closer inspection we identified such a motif in the upstream region of qstR (−87 bp to −65 bp upstream from the ATG start codon of the ORF) ( Figure 4). To investigate the contribution of this site to qstR expression and natural transformability, we either deleted the putative CRP-S motif or used site-directed modifications as indicated in Figure 4. All of these variants were chromosomallyencoded and replaced the original putative CRP-S site. As a control for the genetic engineering method that we used to exchange the motif (TransFLP; [53,55,60], we also included a strain in which the FRT scar preceding the putative qstR promoter region was present without an altered CRP-S motif. This strain (WT_qstR (FRT control)) showed WT behavior with respect to all measured phenotypes (e.g., expression of qstR, natural transformation, and DNA uptake; Figure 6). Moreover, we also included a strain in which the in vitro-identified HapRbinding site [22] was deleted (strain ΔHapR-site_qstR).
Notably, this strain was severely impaired in transformation due to very low qstR and comEA expression, undetectable ComEA protein levels, and a DNA uptake ability, which was below the detection limit of the assay ( Figure 6). With respect to the putative CRP-S site, we observed similar phenotypes as those observed in the mutants that were modified in the upstream region of the comEA gene (described above). That is, both deleting the motif and changing it to an frdA-derived CRP-N site ([frdA] site_qstR) abolished qstR and comEA expression, ComEA production, DNA uptake and, consequently, natural transformability. Remarkably, the change to a CRP-0 site again slightly exaggerated those same phenotypes ( Figure 6).

Discussion
In bacteria, gene transcription begins only after 1) binding of the RNA polymerase (RNAP) holoenzyme to the promoter region and 2) formation of an open complex of the DNA. Promoters that are non-constitutively active (e.g., those of the competence genes) require one or several activating protein(s), such as the CRP protein, that directly interact with the RNAP and assist the holoenzyme in the steps preceding transcriptional initiation [67]. Using comEA and qstR as important genes of the competence regulon of V. cholerae, this study contributes to the understanding of the regulatory network driving natural competence. For comEA we showed that the region up to 134 bp upstream of the start codon is sufficient to drive comEA expression in this organism. But what initiates transcription of comEA? Trying to answer this question, we followed a common assumption, namely that the expression of the competence genes of V. cholerae is TfoX-and CRP-cAMP-dependent and that competence genes are preceeded by so-called CRP-S sites as previously suggested for H. influenzae [48,49]. And indeed, we demonstrated that the deletion of a newly identified motif with striking resemblance to the in silico predicted CRP-S sites [20] had a negative effect on natural transformability. Moreover, we also identified and investigated a putative CRP-S site upstream qstR, which likewise was required for qstR expression and natural transformability. Interestingly, our site-directed mutagenesis approach resulted in unexpected but interesting phenotypes. CRP-mediated activation at the CRP-S site is expected to be different from that at the canonical CRP binding site (CRP-N site; Sxy/TfoX-independent). We therefore assumed that exchanging the putative CRP-S site for a CRP-N motif would enhance competence gene expression. Notably, and in contrast to this assumption, the expression of comEA and qstR was significantly reduced when preceeded by a CRP-N site (Figures 5 and  6). Moreover, opposite results were observed for the strains carrying the CRP-0 motif variants. Despite the fact that this mutation affects the most highly conserved bases of the in silico predicted CRP-S consensus (those in the 3′ part of the motif), the comEA/qstR genes were expressed at higher levels, which correlated well with increased levels of ComEA protein, enhanced DNA uptake, and higher transformation frequencies (Figures 5 and 6). We therefore speculate that the CRP-0 mutation could either be a better binding site for the RNAP or favor the escape of the RNAP complex from the promoter in order to begin transcription. However, based on these unexpected Figure 5 Variants of the putative CRP-S motif preceding comEA are altered with respect to natural transformation. (A) Natural transformability of the indicated strains was tested under tfoX-inducing but chitin-independent conditions as previously described [10]. The frequencies are indicated on the Y-axis. Lane numbers are according to the legend shown on the right. The graph shows the average of at least three independent biological replicates (±SD as indicated by the error bar). Statistically significant differences were determined by Student's t-tests. **P < 0.01, ***P < 0.001, n.s. = not significant. (B) The same strains used in panel A were tested for the relative expression of tfoX (arabinose-induced), pilA, comEA, qstR, hapR, and comEC. The values are given on the Y-axis. Data are averages from at least three independent experiments ± SD. (C) The ability of the wild-type and mutant V. cholerae strains to take up DNA was tested using a whole-cell duplex PCR assay. The lower fragments reflect the quantity of acceptor bacteria and serve as internal controls. The upper band indicates internalized transforming DNA (tDNA). L, ladder.
phenotypes we concluded that the identified motifs indeed play a role in driving the expression of qstR and comEA but that they do not qualify as bona fide CRP-S sites. Upon visual inspection we did not identify other motifs resembling the CRP-S site within the putative promoter region of comEA. A sequence (TGCGA-N6-AAGCA) centered at −115.5 from the transcription start point and located between −147 bp and −132 bp upstream the comEA start codon, was recently discussed (though never experimentally addressed) by Antonova et al. The authors suggested that this sequence serves as a potential CRE element (competence regulatory element [49]; the former name for CRP-S sites) [68]. Although we cannot exclude that this sequence is indeed a CRP binding motif, this site is not essential for the transcription of comEA, as it was not part of the construct that in trans complemented the respective knockout strain (e.g. comEA preceded by 134 bp of its upstream region; Figure 3). Furthermore, the localization of this putative CRP-S site within the open reading frame of the adjacent gene (VC1918) also leads to Figure 6 Altering the promoter region of qstR affects natural transformation. Variants lacking either the HapR-binding site or the putative CRP-S site or containing site-directly changed CRP-S motifs were tested for natural transformability, relative expression of the selected genes, ComEA protein production, and DNA uptake (panels A-D) as described for the previous figures. The tested strains are indicated in the legend, and all experiments were performed at least three independent times. Error bars represent the standard deviation. Statistically significant differences with respect to natural transformation were determined by Student's t-tests. **P < 0.01, ***P < 0.001, n.s. = not significant. The value of the detection limit was used for the statistics of non-transformable strains.
questions regarding its functionality with respect to the regulation of comEA.

Conclusion
Given the absence of any obvious alternative CRP-S site within the 134 bp upstream of comEA we raised the possibility that TfoX and CRP-cAMP only indirectly regulate comEA via the intermediate transcriptional regulator QstR. Indeed, we have previously shown that QstR is necessary for the expression of comEA and comEC but not of the DNA-uptake pilus-encoding genes [12,22]. However, upon artificial induction of qstR in trans only low levels of comEA transcript were measured in accordance with the production of low levels of the ComEA protein ( Figure 2). Two hypothesis are therefore possible: either the proteins TfoX and/or CRP-cAMP are somehow involved in the production of the previously suggested cofactor of QstR [22], or a direct regulation of comEA by TfoX/CRP-cAMP still occurs but involves a CRP-S binding site that significantly differs form the consensus that was in silico predicted for the Vibrionaceae [20]. Further studies involving chromatin immunoprecipitation followed by high-throughput DNA sequencing (ChIP-seq) will be required to ultimately establish the in vivo binding sites of CRP-cAMP, QstR, and potentially also of TfoX in naturally competent V. cholerae cells.