![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
Department of Environmental Toxicology, University of California, Santa Cruz, Santa Cruz, California 95064
Received 12 June 2006/ Accepted 13 November 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Most of the vps genes are clustered on the large chromosome of V. cholerae O1 El Tor and organized into the vpsI (vpsA to -K) and vpsII regions (vpsL to -Q) (61, 63). Two positive transcriptional regulators, VpsR and VpsT, both of which exhibit homology to response regulators, are required for vps gene expression and, in turn, for VPS production and corrugated colony formation. In contrast, the HapR transcriptional regulator, which controls quorum-sensing responses, negatively regulates the expression of vps genes (20, 61, 64). Recently, it was shown that proteins responsible for the turnover of a novel nucleotide, bis-(3'-5')-cyclic dimeric GMP (c-di-GMP), regulate the expression of the genes of two vps clusters, genes located between the vps clusters, vpsR, and vpsT (3, 4, 28, 34, 55). However, the molecular mechanisms of vps gene regulation by these proteins remain unknown.
Cyclic nucleotides are cell signaling molecules that play diverse roles in the biology of microorganisms. c-di-GMP is a ubiquitous second messenger whose production and degradation are controlled by proteins containing the GGDEF and EAL domains, respectively. The GGDEF and EAL domains, named for their common consensus sequence, were recently shown to harbor intrinsic diguanylate cyclase (DGC) and c-di-GMP phosphodiesterase (PDEA) activity, respectively (8, 41, 48, 49, 54).
c-di-GMP was initially identified in Gluconacetobacter xylinus as an allosteric activator of cellulose synthase (46). However, recent studies have recognized the involvement of c-di-GMP in regulation of metabolic processes, cell differentiation, and modulation of the cell surface properties of microorganisms. More specifically, intracellular levels of c-di-GMP have been shown to regulate the formation of rugose colony morphology, intracellular aggregation, exopolysaccharide synthesis, biofilm formation, and motility in several species, such as Pseudomonas aeruginosa, Escherichia coli, Salmonella enterica serovar Typhimurium, and V. cholerae (10, 16, 26, 44, 45). In the past several years, data have emerged on the role of c-di-GMP in regulating the pathogenic capacity of bacterial pathogens and the production of virulence factors. For example, in Bordetella pertussis, an EAL domain protein called BvgR was found to regulate the expression of virulence factors and other genes of unknown function, thereby contributing to respiratory infection in mice (37). Through an in vivo screen to identify genes contributing to the response of Salmonella enterica serovar Typhimurium to host defense mechanisms, Hisert et al. identified CdgR, an EAL domain protein (23). CdgR was shown to promote resistance to hydrogen peroxide and to affect intracellular c-di-GMP levels, thereby suppressing the killing of pathogens by macrophage (23). Recently, mutational analysis of all Pseudomonas aeruginosa genes coding for proteins with a GGDEF and/or an EAL domain revealed that intracellular c-di-GMP levels regulate virulence-related traits, such as type III secretion system-mediated cytotoxicity (32). Additionally, two studies showed that c-di-GMP levels in the cell regulate the transcription of bacterial virulence factors. In one of them, Mendez-Ortiz et al. monitored the genome-wide transcriptional profile of E. coli in response to high levels of c-di-GMP and identified a set of genes related to virulence (36). In the other, Tischler and Camilli showed that VieA, an EAL domain protein from V. cholerae, regulates the transcription of toxT, the most downstream regulator of V. cholerae virulence factors, thereby regulating the expression of the ctxA and ctxB genes, encoding cholera toxin (CT) (56). This transcriptional regulation occurred through the EAL domain's control of c-di-GMP levels in the cell (56).
There are 53 genes encoding proteins with GGDEF and/or EAL domains in the V. cholerae genome: 31 encode GGDEF proteins, 12 encode EAL proteins, and 10 encode proteins with both a GGDEF and an EAL domain on the same polypeptide (17). At present, we have a limited understanding of the functions of these proteins in V. cholerae, as only a few of them have been studied in some detail (4, 28, 30, 42, 55-57). Previously, we identified and characterized 2 of the 10 genes coding for proteins containing both GGDEF and EAL domains, cdgC and mbaA. Expression levels of both genes were higher in the rugose variant than in the smooth variant (34, 61). Phenotypic analysis classified CdgC and MbaA as negative regulators of rugose colony morphology, biofilm formation, and vpsL transcription through the known signal transduction pathway involving the response regulators VpsR and VpsT and the quorum-sensing transcriptional regulator HapR (61). In addition, epistasis analysis revealed that CdgC and MbaA regulate rugose colony morphology in a nonredundant manner in V. cholerae (34).
The goal of
the current study was to determine additional biological processes
controlled by these two proteins. To this end, we determined the
transcriptional profile of cdgC deletion mutants generated in
both the rugose (R
cdgC) and the smooth
(ScdgC) phase variants and of an mbaA
deletion in the rugose (RmbaA) genetic background. We
found that in addition to regulating the expression of genes that can
potentially alter cell surface properties, CdgC regulates expression of
genes involved in extracellular protein secretion, flagellar
biosynthesis, and virulence factor production. We also observed that
CdgC regulates gene expression in a phase variant-dependent manner. We
carried out gene expression profile comparisons of
RcdgC and RmbaA during exponential
and stationary phases of growth to understand the different molecular
functions of the two genes. We observed that the expression profiles of
RcdgC and RmbaA overlap in the
regulation of certain processes but that the timing of their regulation
is unique.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and
CC118
pir were used for the maintenance of plasmids.
E. coli strain S17-1pir was used to deliver
plasmids to V. cholerae by conjugation. V. cholerae
and E. coli strains were grown in Luria-Bertani (LB; 1%
tryptone, 1% NaCl, 0.5% yeast extract, 1.5% Difco granulated agar)
broth with aeration at 30°C unless otherwise noted. LB medium
was supplemented with the following concentrations of antibiotics:
ampicillin, 100 µg ml–1, and rifampin, 100
µg ml–1. Agar plates consisting of LB agar
with 0.3% Difco agar were used to measure motility. LB broth without
NaCl and with 10% sucrose was used for counterselection with
sacB-containing plasmids. For virulence factor-inducing
conditions, strains were cultured in AKI medium as described previously
(25).
|
In-frame deletions for all genes were
carried out using the same general strategy, as previously described
(14). Briefly, a
600-bp fragment 5' of the gene and including several
nucleotides of the gene was amplified by PCR with primers A and B
(described in Table S1 in the supplemental material). A similar
fragment was also amplified from the 3' end of the gene using
primers C and D (see Table S1 in the supplemental material). Purified
PCR fragments from these reactions were allowed to anneal to sequences
in primers B and C and amplified in a second PCR. The resulting
1,200-bp fragment was then amplified with primers A and D,
creating the in-frame deletion construct. The purified PCR fragment was
digested with two of the three restriction enzymes SacI, NcoI, and XbaI
and ligated into plasmid pGP704-sacB28 digested with the same
enzyme. All of the clones were sequenced to ensure that mutations were
not introduced during the manipulation procedures. The in-frame
deletion constructs are listed in Table
1.
Construction of overexpression plasmids. The overexpression constructs containing cdgC and cdgC-AAL were generated in plasmid pBAD/Myc-His B. For construction of pcdgC (pFY_469), VCA0785_pBAD_A and VCA0785-pBAD-D primers (see Table S1 in the supplemental material) were used to amplify the cdgC gene. PCR products were digested with restriction enzymes XhoI and XbaI and ligated into similarly digested pBAD/Myc-His B plasmid. For construction of pcdgC-AAL (pFY_470), VCA0785_pBAD_A and VCA0785_AAL_B primers (see Table S1 in the supplemental material) were used to amplify a 1,219-bp 5' region of the cdgC gene plus 1 bp of upstream flanking region, and VCA0785_pBAD_D and VCA0785_AAL_C primers (see Table S1 in the supplemental material) were used to amplify a 660-bp 3' region of the cdgC gene plus 2 bp of downstream flanking region. These two fragments were joined using the splicing overlap extension technique (24, 33), and the resulting PCR product, harboring new coding sequence for the substituted amino acid residue (E407A), was digested with restriction enzymes XhoI and XbaI and ligated into similarly digested pBAD/Myc-His B plasmid. The PCR products were sequenced (UC Berkeley DNA Sequencing Facility) to ensure that no errors were introduced during PCR amplification. The overexpression constructs are listed in Table 1. Overexpression constructs were electroporated in the V. cholerae smooth wild-type strain.
Construction of lacZ transcriptional fusions.
The lacZ transcriptional
fusions were generated in plasmid pRS415. Promoters of genes of
interest were cloned in a nonpolar manner upstream of the promoterless
lacZ gene in pRS415. Promoters were cloned using primers with
EcoRI and BamHI sites that annealed about 500 base pairs upstream and
about 20 base pairs downstream, respectively, of the start codon. The
promoter of flaA was cloned as cited in the work of Correa and
Klose and included 500 base pairs upstream and 500 base pairs
downstream of the start codon
(9). The primers used for
cloning the promoters are listed in Table S1 in the supplemental
material. All clones were sequenced to ensure that no mutations were
introduced during the manipulation procedures. lacZ
transcriptional fusions were electroporated in V. cholerae
strains harboring a lacZ
deletion.
Construction of V. cholerae knockout mutants. Strain construction with pGP704-sacB28 was performed as shown previously (14) with modifications described by Lim et al. (34).
RNA isolation. Collection of RNA from V. cholerae in the exponential phase of growth was done as previously described (61). Briefly, cultures were grown overnight in LB medium at 30°C with shaking. Overnight-grown cultures were diluted 1:200 in fresh LB medium grown to exponential phase and then diluted again to 1:200 and grown to an optical density at 600 nm (OD600) of 0.4 to ensure homogeneity. Aliquots (2 ml) were collected when cultures reached exponential phase at an OD600 of 0.2, 0.35, and 0.6; centrifuged for 2 min at room temperature; resuspended in 1 ml of TRIzol (Invitrogen); and stored at –80°C. Total RNA from the pellets was isolated according to the manufacturer's instructions. To remove contaminating DNA, total RNA was incubated with RNase-free DNase I (Ambion), and the RNeasy Mini kit (QIAGEN) was used to clean up RNA after DNase digestion.
To obtain RNA from
stationary-phase cultures, overnight-grown cultures were diluted 1:200
in fresh LB medium and incubated at 30°C with shaking (200 rpm)
for 8 h to an OD600 of 2.0. Cell
harvesting and RNA isolation procedures were performed as described
above.
cDNA synthesis, microarray hybridization, and data analysis. Whole-genome expression profiling analysis was performed using a common reference RNA that contained equal amounts of total RNA isolated from exponentially grown smooth and rugose cells. RNA from test and reference samples was used in a reverse transcription reaction. cDNA synthesis and microarray hybridization were done as described previously (3). The raw microarray data were obtained by using the software package GenePix 4.1 (Axon). Normalized signal ratios were obtained with LOWESS print-tip normalization using the Bioconductor packages (http://www.bioconductor.org) in the R environment (19). Differentially regulated genes were determined (with at least two biological and two technical replicates for each data point) with the Significance Analysis of Microarrays package (58) using 1.5-fold differences in gene expression and a 3% false-positive discovery rate as cutoff values.
Semiquantitative RT-PCR. Semiquantitative reverse transcription-PCR (RT-PCR) was performed in one step using QIAGEN's One Step RT-PCR kit as outlined by the manufacturer. For cDNA synthesis, 1 ng of V. cholerae RNA served as template. PCR (25-µl reaction volume) was performed as follows: 50°C for 30 min; 95°C for 15 min; and 25 cycles of 94°C for 30 s, 53°C for 30 s, and 72°C for 1 min. For visualization, 8 µl of the resulting PCR was subjected to agarose gel electrophoresis and stained with ethidium bromide. Gel band intensities were quantified using ImageQuant TL software (GE Healthcare Life Sciences).
Analysis of virulence factor production. For toxin-coregulated pilus (TCP) detection, equivalent protein concentrations of total-cell lysates, as determined with Coomassie blue Bradford assay reagent (Pierce), were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and probed with TcpA antibody antiserum using an ECL detection reagent from Pierce. CT was measured by a GM1-ganglioside enzyme-linked immunosorbent assay, as described previously (18).
ß-Galactosidase assays. ß-Galactosidase activity was determined using a protocol similar to that described previously (38). Modifications to the procedure were carried out as described previously (34).
Motility assays. Ampicillin-containing LB soft agar plates (0.3% agar) were used to determine the motility of bacterial strains. The diameter of the motility zone was measured after 22 h of incubation at 30°C with or without addition of arabinose.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
cdgC and
ScdgC mutants with respect to their wild-type
parental background. Gene expression data were analyzed using the
Significance Analysis of Microarrays program
(58). We applied the
following criteria to define significantly regulated genes: 
3%
false-positive discovery rate and 
1.5-fold transcript
abundance differences between the samples.
The transcriptomes of
RcdgC and ScdgC were compared to
their corresponding parental strain using total RNA isolated from cells
during either the exponential (OD600, 0.3 to 0.4) or the
stationary (OD600, 2.0) phase of growth in liquid LB medium.
Using the selection criteria given above, a total of 490 genes were
found to be differentially regulated by at least 1.5-fold in
RcdgC compared to the wild-type rugose strain during
exponential phase. Of these genes, 167 were induced and 323 repressed
in RcdgC compared to wild type (see Tables S2 and S3
in the supplemental material). During the stationary phase, 238 genes
were differentially regulated in RcdgC compared to
wild type, with the majority being induced (224 genes) and a small
number repressed (14 genes) (see Tables S2 and S3 in the supplemental
material). In the ScdgC mutant 94 genes were
differentially expressed during exponential phase, with 71 being
upregulated and 23 being downregulated (see Tables S4 and S5 in the
supplemental material). Four genes were differentially regulated in the
ScdgC mutant during stationary
phase.
CdgC represses expression of vpsR, vpsT, and vps region genes in both the rugose and smooth phase variants.
We recently showed that deletion of
cdgC in the rugose background led to an increase in colony
corrugation (34). Based
on this result, we expected that the RcdgC mutant
would have an induced expression of the vps region genes,
VC0916, VC0917 to VC0927 (vpsI), and VC0934 to VC0939
(vpsII), compared to wild type. Expression profiling showed
that mRNAs of these genes were increased by 1.5- to 7.8-fold relative
to wild type (Fig.
2A) during exponential and stationary phases of growth. This finding
corroborates our previous study, where we showed that
RcdgC carrying a
vpsL::lacZ fusion exhibited
increased ß-galactosidase activity compared to a wild-type
rugose strain during both exponential and stationary
phases.
|
cdgC mutant was 1.9- and
4.5-fold higher than that in the wild type during exponential and
stationary phases of growth, respectively (Fig.
2A; see also Table S3 in
the supplemental material). Expression of vpsR increased
slightly, by 1.4-fold, during the exponential phase and did not change
significantly during the stationary phase (Fig.
2A; see also Table S3 in
the supplemental material). These results were further confirmed by
measuring ß-galactosidase activity in RcdgC
mutant and wild-type rugose strains carrying a
vpsT::lacZ or
vpsR::lacZ transcriptional fusion
construct. During the exponential phase, RcdgC
exhibited a 3.5-and 3.1-fold increase in
ß-galactosidase activity, compared to wild type, from the
vpsT::lacZ and
vpsR::lacZ fusions, respectively
(Fig.
2B). During the
stationary phase, RcdgC carrying a
vpsT::lacZ fusion had 3.2-fold
more ß-galactosidase activity than the wild type (Fig.
2C). On the other hand,
there was no significant difference in ß-galactosidase
activities between RcdgC and the wild-type rugose
strain carrying a vpsR::lacZ
fusion during the stationary phase.
Our transcriptome analysis
also revealed a 2.7-fold decrease in the expression of the
hapR gene in the RcdgC mutant relative to
wild type during the exponential phase (see Table S3 in the
supplemental material). To determine the extent to which CdgC regulates
hapR expression, we performed semiquantitative RT-PCR to
detect the amount of hapR transcript in both wild-type rugose
and RcdgC strains during the early exponential
(OD600, 0.2), mid-exponential (OD600, 0.35), late
exponential (OD600, 0.6), and stationary (OD600,
2.0) phases of growth. Quantification of band intensities with
ImageQuant software revealed that RcdgC exhibited
decreased levels of hapR transcripts relative to the rugose
variant in the early, mid-, and late exponential phases, by 1.9-, 1.3-,
and 1.3-fold, respectively, but not in the stationary phase (Fig.
2D), confirming the
microarray results. It should also be noted that there were no changes
in the gyrA message abundance between the
RcdgC and the rugose variant.
Expression
profiling of the wild-type smooth variant and ScdgC
mutant revealed that transcript levels for the vps genes were
higher in ScdgC (by 1.5- to 5.0-fold; Fig.
2A; see also Table S5 in
the supplemental material) during the exponential phase. Similar to the
rugose genetic background, expression levels of vpsT and
vpsR increased by 2.5-and 2.1-fold, respectively, in the
ScdgC mutant relative to wild type during the
exponential phase (see Table S5 in the supplemental material).
Increased vps expression in the ScdgC mutant
was confirmed by comparing the ß-galactosidase activities of
vpsA::lacZ and
vpsL::lacZ transcriptional fusion
constructs in the ScdgC mutant and smooth wild-type
variant. The ScdgC mutant showed an increase in
transcriptional activity for vpsA and vpsL of 3.1-
and 5.3-fold, respectively (Fig.
2E).
Expression of eps, purine/pyrimidine biosynthesis, flagellar biogenesis, and chemotaxis genes is altered in the RcdgC mutant.
The expression of
eps (extracellular protein
secretion) genes, located in a 12-gene operon containing genes
epsC to epsN, increased by 1.4- to 1.9-fold in the
RcdgC mutant relative to rugose wild type
during the exponential phase (Fig.
3A; see also Table S5 in the supplemental material). The products of the
EPS pathway (encoded by epsD and epsE) are required
for the formation of the rugose colonial morphotype
(1). Thus, expression of
eps genes has been shown to be consistently higher in the
rugose strain than in the wild-type smooth strain
(61). The EPS secretion
system is predicted to be responsible for either secretion of VPS or
secretion of protein(s) involved in the transport/assembly of VPS
(61). Furthermore, it was
recently shown that the expression of vps and eps
genes is positively regulated by an increase in cellular c-di-GMP
levels (3). The observed
increased expression of eps genes in the
RcdgC mutant relative to rugose wild type may be due
to a minimal, but consistent, increase in c-di-GMP levels
(34).
|
cdgC mutant relative
to wild type during exponential phase (Fig.
3B; see also Table S1 in
the supplemental material). RcdgC produces higher
levels of cellular c-di-GMP than the wild-type rugose variant
(34); these results
indicate that RcdgC may be maximizing its guanine
nucleotide pool to increase c-di-GMP production and/or may be reacting
to increased cellular levels of guanine nucleotides.
Comparison
of expression profiles from the RcdgC mutant and
rugose wild type during exponential phase revealed that transcription
of 24 genes required for flagellum biosynthesis was decreased in
RcdgC (Fig.
3C; see also Table S3 in
the supplemental material), suggesting that CdgC positively regulates
expression of flagellar genes. Deletion of flaA, which encodes
the major flagellin component of the flagellar filament in V.
cholerae, leads to a nonmotile phenotype
(29). To confirm the
expression profile data, we engineered a
flaA::lacZ transcriptional fusion
construct and measured ß-galactosidase activity in the
RcdgC mutant and rugose wild type grown under the
same conditions. The RcdgC mutant exhibited 4.4- and
5.3-fold decreases in ß-galactosidase activity during
exponential and stationary phases, respectively (Fig.
3E). In addition, as
previously reported, there was a concomitant decrease in the motility
of the RcdgC mutant
(34).
Additionally,
expression of several genes predicted to function in
chemotaxis (VCA0068, VCA0663, VCA0864, VCA1031, VC1313, VC1413, VC1859,
VC1898, VC2059, VC2062, VC2063, VC2064, and VC2161) was differentially
regulated (see Table S3 in the supplemental material). Of these genes,
nine were methyl-accepting chemotaxis proteins and
four were chemotaxis proteins (Che). We have yet to determine the
relative contributions of flagellar biosynthesis genes and chemotaxis
genes in the RcdgC mutant to the observed decrease in
motility.
Expression of RTX genes is increased in the RcdgC mutant.
The expression of genes required for
producing the RTX toxin, which causes rounding of epithelial cells via
its action on actin polymerization
(15,
35,
50), was higher in the
RcdgC mutant. The genes responsible for the RTX
production (rtxA), transport (rtxB, rtxD,
and rtxE), and activation (rtxC) are organized into
two divergently transcribed operons on the large chromosome of V.
cholerae O1 El Tor. Expression of the RTX
genes—rtxA, rtxB, rtxC, and
rtxD—was increased in the RcdgC
mutant during the exponential phase of growth by 1.7-, 1.9-, 2.6-, and
1.8-fold, respectively, relative to wild type (Fig.
3D; see also Table S3 in
the supplemental material).
CdgC regulates expression of virulence factors in V. cholerae. The major virulence factors of V. cholerae are CT and the TCP colonization factor, encoded by ctxA/ctxB and tcpA, respectively. A complex transcriptional regulatory cascade controls CT and TCP production, whereby transcription of ctxA/ctxB and tcpA is regulated by the regulatory proteins ToxR and TcpP, respectively. Two additional proteins, ToxS and TcpH, seem to enhance transcriptional activity of ToxR and TcpP, respectively (6, 12). These two regulatory units control the expression of toxT, the most downstream regulator of virulence factors in V. cholerae (21, 39). In addition, expression of tcpP and tcpH is under the control of two other regulatory proteins, AphA and AphB. The quorum-sensing transcriptional regulator HapR acts as a negative regulator of ctxA/ctxB and tcpA expression by negatively regulating expression of aphA (31).
Expression
analysis of the RcdgC mutant during stationary phase
revealed a 2.0-fold increase in the transcript abundance of the
tcpA gene compared to the rugose wild-type variant (see Table
S3 in the supplemental material). This result was confirmed by
determining the ß-galactosidase activity of both rugose wild
type and RcdgC mutant carrying transcriptional fusion
constructs. During stationary phase at 30°C in LB medium,
RcdgC carrying a
ctxA::lacZ or
tcpA::lacZ transcriptional fusion
exhibited 1.9- and 2.8-fold increases in ß-galactosidase
activity, respectively, compared to the wild type (Fig.
4A). High expression levels of virulence factors among V.
cholerae El Tor strains have been observed under specific in vitro
growth conditions, termed AKI growth conditions. Compared to the
wild-type rugose strain, TcpA production was higher in the
RcdgC mutant grown under AKI conditions (Fig.
4B). Similarly, CT
production was also greatly increased in the RcdgC
mutant, whereby culture supernatants of RcdgC yielded
a 48-fold-higher amount of CT than did the wild type (Fig.
4B).
|
cdgC
mutant, transcriptome analysis revealed a 2.6-fold increase in the
transcript abundance of the aphA gene compared to the rugose
wild-type variant during exponential phase (see Table S3 in the
supplemental material). In addition, transcription of the tcpP
regulatory gene, which is activated by aphA
(53), was induced
1.7-fold during stationary phase in the RcdgC mutant
relative to rugose wild type (see Table S3 in the supplemental
material). We hypothesized that CdgC negatively regulates
ctxA/ctxB and tcpA, at least in part through
tcpP/tcpH transcriptional regulation. We confirmed that the
transcription level of tcpP was increased in
RcdgC relative to rugose wild type by using a
tcpP::lacZ fusion construct.
During stationary phase (OD600, 2.0),
RcdgC containing the
tcpP::lacZ fusion
exhibited approximately 1.9-fold-more ß-galactosidase activity
than wild-type rugose (Fig.
4C). To determine
whether CdgC regulates tcpA expression primarily by
regulating tcpP expression, we compared
the ß-galactosidase activities of the
tcpA::lacZ transcriptional fusion
in rugose wild type, RcdgC, RtcpP,
and the double deletion mutant
RcdgCtcpP. If CdgC modulates
tcpA transcription solely by regulating tcpP
expression, ß-galactosidase measurements in the
RtcpP and
RcdgCtcpP strains should be
similar. We found that deletion of tcpP in the rugose
background decreased tcpA expression by 2.0-fold (Fig.
4D). However, the double
deletion mutant RcdgCtcpP exhibited
ß-galactosidase measurements similar to those of the rugose
wild-type strain, suggesting that tcpP is not fully epistatic
to cdgC. In other words, CdgC regulates tcpA
expression through additional mechanisms or through a combination of
the two (Fig. 4D). ToxT is
believed to be the direct transcriptional activator of the
ctxA and tcpA gene promoters
(12). We did not,
however, observe any significant differences in the transcription of
toxT between mutant and wild type by either microarray
analysis or the toxT::lacZ fusion
(see Table S3 in the supplemental material; data not
shown).
During the exponential phase of growth, we found that
expression of tcpA and toxT was decreased 1.8- and
1.6-fold, respectively, in the ScdgC mutant relative
to the smooth wild-type strain (see Table S5 in the supplemental
material). To confirm these results and test whether deletion of
cdgC in the smooth background selectively regulated
tcpA expression, we examined the ß-galactosidase
activities of the ctxA::lacZ and
tcpA::lacZ transcriptional fusion
constructs in the smooth and ScdgC strains grown at
30°C to exponential phase. ScdgC carrying a
ctxA::lacZ fusion had
ß-galactosidase activity similar to that of the smooth
wild-type variant (Fig.
4E). However,
ScdgC carrying a
tcpA::lacZ fusion showed a
1.4-fold decrease in ß-galactosidase activity compared to the
smooth variant (Fig. 4E).
In contrast, the RcdgC mutant grown at 30°C
to exponential phase exhibited a 3.7-fold increase in transcriptional
activity from the tcpA::lacZ
fusion compared to the rugose wild type (data not shown). Altogether,
these data indicate that the regulation of virulence factors by CdgC
differs between the V. cholerae rugose and smooth
variants.
Virulence factor expression is different between the smooth and rugose phase variants in V. cholerae El Tor A1552. CdgC regulates the expression of virulence genes in an opposing manner in the two phase variants. To determine possible differences in the regulation of ctxA and tcpA transcription between the smooth and rugose variants, we examined the expression of ctxA and tcpA using the ctxA::lacZ or tcpA::lacZ fusion. Rugose and smooth variants carrying the fusions were grown at 30°C to exponential and stationary phases and also under AKI conditions. During the exponential phase, the rugose variant harboring the ctxA::lacZ fusion showed a minimal 1.4-fold increase in ß-galactosidase activity compared to the smooth strain. No significant difference in tcpA levels between the phase variants was observed (Fig. 5A). The rugose variant carrying the ctxA::lacZ transcriptional fusion exhibited a 2.7-fold increase in ß-galactosidase activity compared to the smooth variant both during the stationary phase and under AKI conditions (Fig. 5B and Fig. 5C). Similarly, the rugose variant carrying the tcpA::lacZ fusion showed a 4.3- and 3.2-fold increase in ß-galactosidase activity in the stationary phase and under AKI conditions, respectively, compared to the smooth strain (Fig. 5B and Fig. 5C).
|
mbaA mutant to
understand whether mbaA and cdgC regulate a similar
set of genes and cellular processes. We compared gene expression
patterns of the RmbaA mutant with those of the
wild-type rugose strain during both exponential and stationary growth
phases. Using the selection criteria described above, 192 genes were
differentially expressed by at least 1.5-fold during exponential phase
in liquid LB medium. Of these genes, 73 were induced and 119 were
repressed (see Tables S6 and S7 in the supplemental material). During
the stationary phase, 120 genes were differentially regulated; the
majority were induced (112 genes) and a small number were repressed
(eight genes) (see Tables S6 and S7 in the supplemental
material).
We initially identified mbaA as a negative
regulator of rugose colony development. As predicted, we saw
that expression of many of the genes located in the vpsI
cluster (VC0917 to VC0927), in the vpsII cluster (VC0934 to
VC0939), and between the vps clusters—VC0928, VC0929,
VC0930, VC0932, and VC0933—was significantly induced by 1.8- to
5.1-fold in the RmbaA mutant compared to rugose wild
type, primarily during stationary phase (see Table S7 in the
supplemental material). This increase in gene expression may be
responsible, at least in part, for the observed increase in the
biofilm-forming capacity of the RmbaA mutant during
later phases of growth (4,
28,
34). The induction of
vps gene expression in RmbaA may be due to
increased transcription of either vpsR or vpsT, or
both. Our results showed no increase in vpsR gene expression
in the RmbaA mutant; however, expression of
vpsT was induced 2.0- and 5.1-fold relative to the wild-type
rugose variant during exponential and stationary phase, respectively
(see Table S7 in the supplemental material). We confirmed our array
results by measuring ß-galactosidase activities of
vpsR::lacZ and
vpsT::lacZ fusion constructs in
the RmbaA mutant and rugose wild type. Whereas there
was no significant difference in ß-galactosidase activity for
the vpsR::lacZ fusion, we
observed a 1.5- and 1.4-fold increase in ß-galactosidase
activity in the RmbaA mutant carrying the
vpsT::lacZ transcriptional fusion
during exponential and stationary phases, respectively (Fig.
6A and 6B). Additionally,
work from Karatan et al. identified a set of genes regulated by MbaA in
the smooth variant of the O139 strain MO10
(28). This study also
observed an increase in expression of vps genes and the
vpsT gene in the mbaA mutant relative to wild type.
Altogether, these results suggest that the increase in vps
gene transcription can, in part, be due to the action of VpsT. In fact,
we previously identified vpsT as being epistatic to
mbaA in terms of rugose colony morphology
(34). However, while
vpsT expression was increased during both exponential and
stationary phases in the RmbaA mutant, we saw a
significant induction of the expression of the vps genes only
during the stationary phase (see Table S7 in the supplemental material)
(34). These results
indicate that mbaA regulates vps gene expression
temporally, through a yet-to-be-identified mechanism.
|
mbaA mutant showed a 1.6-fold
repression of hapR expression during the exponential phase.
This effect may be partly responsible for the induction of
vpsT in the RmbaA mutant but cannot explain
the similarity in vpsR gene expression levels between
RmbaA and the rugose wild-type variant during
exponential and stationary phases (Fig.
6A and
6B). Taken together, these
results suggest that additional mechanisms are responsible for the
regulation of vps transcription by mbaA and reveal
the complexity of c-di-GMP signaling in V. cholerae cellular
processes.
We also observed an increase (by 1.3- to 1.8-fold) in
the expression of several eps genes (epsG,
epsI, epsJ, and epsL) in the
RmbaA mutant relative to wild type during exponential
and stationary growth phases (see Table S7 in the supplemental
material; data not shown). On the other hand, expression of several
flagellar biosynthesis genes was 1.5- to 4.2-fold lower in the mutant
during exponential (flaB, flaC, and flaG
genes) and stationary (flaA, flaC, flaD, and
flgB genes) phases. Consistent with these findings, we
previously showed that the RmbaA mutant exhibited
reduced motility compared to the wild-type rugose variant
(34).
Comparison of CdgC and MbaA regulons during exponential and stationary growth phases.
Both CdgC and MbaA
are negative regulators of rugose colony development
(34). They contain a
GGDEF and an EAL domain but differ primarily in the N-terminal domain
sequence. The expression of cdgC is higher in the rugose than
in the smooth variant during exponential and stationary phases, whereas
expression of mbaA is higher in the rugose variant in the
stationary growth phase only
(61; unpublished data).
These findings suggest either some redundancy in the functions of
mbaA and cdgC or increased complexity in their output
signal. To better understand the functions of CdgC and MbaA expression
during exponential and stationary phases, we compared the transcription
profiles of the RcdgC mutant and the rugose wild type
and of RmbaA and the rugose wild type. Of the 165
genes found to be differentially regulated in the two mutants during
the exponential phase, 15% were involved in metabolism, 8% were
involved in nucleotide biosynthesis, 15% coded for transport and
binding proteins, and 21% coded for hypothetical proteins. Of
particular interest are the 8% of the common differentially regulated
genes that have a regulatory function, including hapR and
vpsT (whose expression is repressed and induced,
respectively). However, during the exponential growth phase,
RcdgC alone exhibited increased vps
(vpsI and vpsII clusters) and vpsR gene
transcription and a significant decrease in expression of the genes
encoding flagellar and chemotaxis machinery components.
Of the 52
genes differentially regulated in both RcdgC and
RmbaA mutants during the stationary phase, 31% of
them consisted of vps region genes (from both clusters), genes
located between the vps clusters, and vpsT. This
result indicates that both CdgC and MbaA function to regulate
vps transcription during stationary phase. These results all
together indicate that the cdgC and mbaA regulons
overlap based on several molecular phenotypes but that the timing of
their transcriptional regulation
differs.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cdgC and ScdgC mutants to those of
wild-type rugose and smooth variants, respectively, during exponential
and stationary phases of growth. The analysis indicated that a large
number of processes are regulated by cdgC in V.
cholerae in a phase variant-specific manner. All physiological
processes previously identified
(34) as being modulated
by cdgC—such as rugose colony development, biofilm
formation and architecture, and motility behavior—could be
explained by the genes being differentially regulated in the
RcdgC and ScdgC mutants.
Specifically, the increase in rugosity and biofilm-forming capacity of
the RcdgC and ScdgC mutants
correlated with the observed increase in the expression of vps
and eps genes. Additionally, the decrease in motility in the
RcdgC mutant corresponded with the repression of
genes responsible for flagellar biosynthesis and chemotaxis (Fig.
7).
|
cdgC expression analysis suggests that CdgC may
negatively regulate expression of vps indirectly, by
repressing vpsR and vpsT or increasing hapR
expression. It was shown that VpsR and VpsT are required for the
transcription of vps genes
(7,
60). HapR negatively
regulates vps expression either indirectly, by decreasing
vpsR or vpsT expression, or directly, by physically
binding upstream of the vpsI cluster
(61). As HapR was shown
to positively regulate expression of flagellar biosynthesis genes, the
decrease in hapR transcription in the RcdgC
mutant may be responsible for the decrease in flagellar gene
expression. Consistent with this idea, in silico studies indicate the
presence of a putative HapR binding site upstream of flagellar
biosynthesis genes, such as VC2120, VC2129, VC2130, VC2142, VC2187,
VC2195, and VC2196 (61).
As several genes predicted to have regulatory functions are
differentially expressed in the RcdgC mutant compared
to wild type, we cannot rule out the possibility that CdgC may regulate
vps gene expression through additional mechanisms. In
addition, the regulation of a large number of differentially expressed
genes in the RcdgC and the ScdgC
mutants may not be directly affected by CdgC binding to their promoters
but rather through the control of known regulators, either at the
transcriptional or at the posttranslational level.
In this study,
we observed that CdgC negatively regulates the expression of the
transcriptional regulators vpsR and vpsT (Fig.
7). In contrast, CdgC
positively regulates hapR expression, as the amount of
hapR message was decreased in the RcdgC
mutant compared to rugose wild type (Fig.
7). Intriguingly, in a
study designed to identify the VpsR and HapR regulons, we found
cdgC to be part of both regulons and its promoter region to
contain potential binding sites for vpsR and hapR
(61). Now, in this study,
we show that regulatory proteins and GGDEF/EAL-containing proteins
comprise an interrelated network that controls several cellular
processes in V. cholerae. In V. cholerae, c-di-GMP
and predicted DGCs and PDEAs have been implicated in the regulation of
diverse cellular processes, such as vps synthesis, biofilm
formation, motile behavior, virulence factor production, and phenotypic
switching between the known phase variants
(3,
34,
42,
55,
56). Interestingly, these
physiological processes are regulated by genes that play a role in
quorum sensing, which is the ability of a cell to monitor its own
population as well as the abundance of other microbial populations in a
bacterial community. Specifically, the quorum-sensing transcriptional
regulator hapR regulates expression of several DGCs;
furthermore, several DGC and PDEA genes are predicted to have
hapR binding sites in their promoter regions
(61,
64). As first suggested
by Camilli and Bassler, extracellular quorum-sensing signaling and
intracellular c-di-GMP concentrations may be linked, based on the
overlap of the physiological processes that these two systems regulate
(5). In this study, we
reveal that DGCs and/or PDEAs can also regulate the expression of
quorum-sensing regulators. Increased complexity of regulatory networks
involving the c-di-GMP second messenger is likely to provide the
necessary fine-tuning of cellular processes involved in environmental
survival and in cellular communication.
In the current study, we
observed that CdgC represses ctxA, tcpA,
aphA, and tcpP expression in the rugose genetic
background (Fig. 7). We
previously observed that RcdgC exhibited a slight
increase in cellular c-di-GMP levels compared to the rugose wild type
(34). It is possible that
such an increase could induce the expression of virulence genes.
Alternatively, a decrease in the hapR transcript, as observed
in the RcdgC mutant, could lead to an increase in
virulence gene transcription (Fig.
7). We have a limited
understanding of the role that c-di-GMP plays in regulating virulence
factor production in V. cholerae. Previous work by Tischler
and Camilli showed that increased cellular levels of c-di-GMP (due to a
mutation in the response regulator VieA, an EAL domain protein) in the
V. cholerae classical strain O395 led to a decrease in
ctxA expression possibly through the repression of the
toxT transcript
(56). In contrast, the
same vieA mutation in the El Tor biotype does not cause a
decrease in virulence gene expression under the same growth conditions,
suggesting that c-di-GMP regulates virulence gene expression in a
biotype-specific manner
(2,
56). Furthermore,
whole-genome expression analysis of the V. cholerae classical
and El Tor strains (grown at 30°C in morpholinepropanesulfonic
acid minimal medium supplemented with 0.5% glycerol, 25 mM asparagine,
arginine, glutamate, and serine to an OD600 of 0.3) also
showed that virulence gene expression did not significantly decrease in
either classical or El Tor biotypes 30 min after an increase in
intracellular c-di-GMP levels
(3). Taken together, these
studies indicate that there are significant gaps in our knowledge of
the mechanism by which c-di-GMP regulates virulence gene expression in
V. cholerae.
We compared the transcriptomes of
cdgC and mbaA, which were previously shown to
regulate vps expression
(4,
28,
34). Our transcriptome
comparisons indicated that CdgC represses the expression of
vps genes during both exponential and stationary phases, while
MbaA represses vps gene transcription primarily in the
stationary phase. Expression of mbaA is upregulated during the
stationary phase in the rugose variant compared to the smooth variant,
indicating that the cell temporally restricts the transcription of this
GGDEF/EAL protein. Therefore, both proteins regulate vps
transcription in the rugose background but in a unique
temporal manner. We additionally showed that the magnitude of the
increase in the expression of the vpsR and vpsT genes
varies between the RcdgC and RmbaA
mutants, suggesting that the mechanism by which cdgC and
mbaA regulate gene expression may also be unique. This result
brings into question whether the two proteins have a redundant function
or an overlapping output signal activated under different cellular or
environmental signals. Multiple sequence alignments of CdgC and MbaA
have shown that the predicted EAL domain is enzymatically active while
the GGDEF domain is predicted to be inactive, suggesting that both of
the proteins function as phosphodiesterases of c-di-GMP
(34). In this study, we
were also able to show that the EAL domain of CdgC is active and
required for a well-known c-di-GMP-associated phenotype, flagellar
motility. It is also yet to be determined whether the regulatory motifs
found in the N-terminal regions of CdgC and MbaA (the GAF and HAMP
domains, respectively) restrict signal production to certain
environmental or cellular states. It should also be noted that
regulation of vps expression by MbaA could be further
modulated by the presence of norspermidine in the environment
(28).
Although c-di-GMP was first discovered to allosterically activate cellulose synthase in G. xylinus (46), it is becoming increasingly clear that c-di-GMP exerts its effects at levels other than the posttranslational one. Recent studies show that proteins that function as DGCs and PDEAs and proteins containing an HD-GYP domain play diverse roles in many bacterial organisms from exopolysaccharide production to the control of motility to virulence factor production and host survival (10, 26, 44, 45, 47). In order to understand the complexities of this novel signaling pathway, it will be important to further tease apart the c-di-GMP signal transduction pathways and to identify both posttranscriptional and posttranslational mechanisms by which c-di-GMP can regulate a wide range of cellular processes.
|
ACKNOWLEDGMENTS |
|---|
We thank Lindsay Stuart Odell for creation of pFY-222 and strain FY_Vc_576 and Ron Taylor for TcpA antibody.
|
FOOTNOTES |
|---|
Published
ahead of print on 22 November 2006.
Supplemental material for this article may be found at
http://jb.asm.org/.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
