VpsT Is a Transcriptional Regulator Required for Expression of vps Biosynthesis Genes and the Development of Rugose Colonial Morphology in Vibrio cholerae O1 El Tor

Vibrio cholerae switches between smooth and rugose colonialvariants. The rugose variant produces more vibrio polysaccharides(VPS^{El Tor}) and forms well-developed biofilms. Both phenotypesdepend on expression of vps biosynthesis genes. We identifieda positive transcriptional regulator of vps gene expression,VpsT, which is homologous to response regulators of two-componentregulatory systems. Disruption of vpsT in the rugose variantyields smooth colonies, prevents formation of mature biofilms,and decreases vps gene expression. The interaction between VpsTand VpsR, a previously identified positive regulator of vpsgenes, was also investigated.

INTRODUCTION

Top

Introduction

Vibrio cholerae, the causative agent of the disease cholera,is a natural inhabitant of aquatic ecosystems. The pathogencauses periodic, seasonal cholera outbreaks in regions wherethe disease is endemic and can spread worldwide in pandemics(10). The ability of V. cholerae to cause epidemics is linkedto its survival in aquatic habitats.

During its life cycle, V. cholerae undergoes phase variationwhich results in the generation of two morphologically differentvariants termed smooth and rugose (23). Compared to smooth variants,rugose variants have increased capacity to produce exopolysaccharideVPS^{El Tor}, which enables them to form well-developed biofilmsand to better resist environmental stresses (14, 21-23, 27).Exopolysaccharide VPS^{El Tor} production depends on transcriptionof the vibrio polysaccharide synthesis (vps) genes (27). vpsgenes are clustered in two regions in the V. cholerae chromosome.One cluster harbors genes vpsA through vpsK, and the other oneharbors genes vpsL through vpsQ. Rugose variants lacking vpsAor vpsL do not produce VPS^{El Tor} and exhibit a smooth colonialmorphology (27). One positive regulator of vps genes, VpsR (25),and two negative regulators, HapR (5, 8, 28) and CytR (6), havebeen identified. In this communication, we report the identificationof a second positive regulator of vps genes, designated VpsT,which is required for the formation of a corrugated colonialmorphology, biofilm formation, and vps gene expression in therugose variant. We further show that VpsT and VpsR positivelyautoregulate their own expression and also form a complex regulatorynetwork by positively regulating each other's expression.

Identification of vpsT.

Top

Identification of vpst.

Whole-genome expression profiling of exponentially grown smoothand rugose variants revealed that expression of gene VCA0952(TIGR annotation; now named vpsT) was fourfold elevated in therugose variant compared to that in the smooth variant (F. H.Yildiz et al., submitted for publication). The vpsT gene productis 671 bp long and is predicted to encode a 224-amino-acid,25.8-kDa protein that is similar to proteins that belong tothe UhpA (FixJ) family of transcriptional response regulators(12, 15). VpsT is homologous (44% homology and 65% similarity)to the transcriptional regulators CsgD and AgfD from Escherichiacoli and Salmonella enterica, respectively. CsgD and AgfD arerequired for the production of extracellular matrix components,cellulose and curli fimbriae, which are important for the developmentof wrinkled colonies and biofilm formation in these bacteria(3, 4, 16, 17).

VpsT affects colony morphology.

Top

Vpst affects colony morphology.

In order to determine the role of vpsT in maintaining the rugosecolonial morphology and associated phenotypes, we deleted vpsTin the V. cholerae O1 El Tor rugose variant (FY_Vc_0004) (Table2), designated the R ${Delta}$ vpsT mutant (FY_Vc_0005). Deletions weredone according to the modified method of Horton (2, 7, 9), usingthe VCA0952 primer set listed in Table 1. The R ${Delta}$ vpsT mutant exhibiteda smooth colonial morphology on Luria-Bertani (LB) agar plates(Fig. 1), indicating that the vpsT gene product is requiredfor the formation of a rugose colonial morphology. To verifythat the smooth colony phenotype of the R ${Delta}$ vpsT mutant was causedby the vpsT deletion, we amplified the wild-type copy of vpsT,including 559 bp upstream and 380 bp downstream, and clonedit into the low-copy-number plasmid pACYC177, generating pCC17.Introduction of pCC17 into the R ${Delta}$ vpsT mutant resulted in conversionof the smooth colonial morphology to the wild-type rugose colonialmorphology (Fig. 1). We also found that vpsT cloned from eitherthe smooth (pCC14) or the rugose (pCC16) variant complementedthe R ${Delta}$ vpsT mutant, indicating that this gene is not physicallyaltered during phase variation. Introduction of the cloningvector alone did not result in complementation. Deletion ofthe previously identified transcriptional regulator VpsR (usingprimer set VC0665) (Table 1) in the rugose variant also resultedin formation of a smooth colonial morphology (25). To determineany epistasis between the two positive transcriptional regulators,we generated a ${Delta}$ vpsR ${Delta}$ vpsT double mutant in the rugose variantand observed that the colonial morphologies of the single anddouble mutants did not differ in their characteristics (Fig.1). Taken together, these results indicate that both VpsT andVpsR are involved in the formation of corrugated colonies.

View this table:
[in this window]
[in a new window]
TABLE 2. Plasmids and strains used in this study

View this table:
[in this window]
[in a new window]
TABLE 1. Primers used in this study

View larger version (20K):
[in this window]
[in a new window]
FIG. 1. VpsT is required for the rugose colonial morphology. Colonial morphology of the smooth and rugose variants and the indicated deletion mutants formed on LB plates after 2 days of incubation at 30°C is shown.

VpsT affects biofilm formation.

Top

Vpst affects biofilm formation.

In the rugose variant, formation of corrugated colonies andwell-developed biofilms depend on VPS^{El Tor} production. To determinewhether VpsT influences biofilm formation, we compared quantitativeand qualitative differences in biofilms of the R ${Delta}$ vpsT mutantto those of the rugose and smooth variants. For quantitativeanalysis, biofilms were formed on polyvinyl microtiter plates.After 8 h of growth in LB medium at 30°C under static conditions,biofilms were quantified by crystal violet staining (1). Theresults presented in Fig. 2A show that under the conditionstested the rugose variant formed six times more biofilm thanthe smooth variant. The biofilm-forming capacity of the R ${Delta}$ vpsTmutant was reduced to the level of that of the smooth variant.Complementation of the R ${Delta}$ vpsT mutant by pCC17 restored biofilmformation to wild-type levels. The cloning vector alone didnot result in complementation. The biofilm formation of theR ${Delta}$ vpsR mutant and of the R ${Delta}$ vpsR ${Delta}$ vpsT double mutant was similarto that of the R ${Delta}$ vpsT mutant and also to that of the smooth variant(Fig. 2A). The growth rates of the strains were similar (datanot shown), indicating that the differences in biofilm formationwere not due to different growth rates.

View larger version (52K):
[in this window]
[in a new window]
FIG. 2. Biofilm formation of vpsT and vpsR mutants. (A) Biofilm biomass measured with crystal violet-staining assay. The figure shows the average and standard deviation of three independent measurements for each mutant and for the wild-type parents. (B) Confocal scanning laser microscopic images of biofilms formed on borosilicate cover glass chambers. Top panels, horizontal projections; bottom panels, vertical projections.

To compare the biofilm morphologies of the wild-type variantsand the mutants, we introduced plasmid pV25, harboring a geneconstitutively expressing the green fluorescent protein (20),into each strain. Biofilms were formed on a borosilicate coverglass in LB medium at 30°C under static conditions and analyzedafter 12 h by confocal scanning laser microscopy. Horizontalprojected views (Fig. 2B) show that the rugose variant formedbiofilms with distinct islands, whereas the smooth variant mainlyattached as evenly spread-out single cells. Vertical views ofthe same biofilms show that the rugose biofilm (approximately60 µm) was about five times thicker than the smooth biofilmand displayed distinctive structures. Surface colonization ofthe R ${Delta}$ vpsR and R ${Delta}$ vpsT mutants and the R ${Delta}$ vpsR ${Delta}$ vpsT double mutantwas similar to that of the smooth variant. Together these resultssuggest that VpsR and VpsT both contribute to biofilm formationin the rugose variant.

VpsT induces vps gene expression.

Top

Vpst induces vps gene...

The R ${Delta}$ vpsT mutant is affected in its capacity to form rugosecolony morphology and biofilms, which both depend on vps geneexpression (6, 25). As VpsT is homologous to response regulatorsof two-component signal transduction systems, we examined whetherVpsT affects vps gene expression. For the expression analysis,we chose vpsA (VC0917) and vpsL (VC0934), which are the firstgenes of the two vps operons (27), respectively. We constructedtranscriptional fusions of the upstream regulatory sequencesof vpsA (574 bp upstream) and vpsL (565 bp upstream) to theß-galactosidase (lacZ) gene. To this end, vpsAp andvpsLp were amplified and cloned upstream of promoterless lacZin vector pRS415, yielding plasmids pCC11 and pCC12, respectively.Both plasmids and the parent vector (pRS415) (19) were introducedinto the smooth, rugose R ${Delta}$ vpsT and R ${Delta}$ vpsR mutants and the R ${Delta}$ vpsR ${Delta}$ vpsT double mutant. Transcription was measured by determiningß-galactosidase activity (13) of cultures grown tomid-exponential phase (optical density at 600 nm, 0.3 to 0.4)in LB medium at 30°C by shaking. Figure 3A shows that vpsAtranscription was five times higher in the rugose variant thanin the smooth variant. Deletion of the vpsT gene in the rugosevariant resulted in a significant reduction of vpsA gene transcription.Deletion of vpsR resulted in a similar decrease in vpsA genetranscription. The R ${Delta}$ vpsR ${Delta}$ vpsT double mutant had low vpsA geneexpression, similar to that for the individual deletion mutants.Strains transformed with vector pRS415 had no ß-galactosidaseactivity (data not shown). vpsL transcription was forty timeshigher in the rugose variant than in the smooth variant. Transcriptionof vpsL was markedly (13-fold) decreased in the R ${Delta}$ vpsT mutant.In contrast, in the R ${Delta}$ vpsR mutant and the R ${Delta}$ vpsR ${Delta}$ vpsT double mutant,vpsL transcription was below the detection level. The resultsindicate that VpsT and VpsR are both required for maximal transcriptionof the vpsA and vpsL promoters in the rugose variant duringthe logarithmic growth phase. Transcription of vpsL was higherthan that of vpsA in all tested strains. The relative effectsof VpsT were similar on vpsA and vpsL transcription. VpsR deletion,on the other hand, had a stronger effect on vpsL transcriptionthan on vpsA transcription. The results suggest that the regulationof these two genes is different in the rugose variant. Interestingly,there is a computationally identified VpsR binding site upstreamof vpsL but not upstream of vpsA (Yildiz et al., submitted).It remains to be determined if the actions of VpsT and VpsRare indeed mediated by the direct binding to the vpsL and vpsApromoter regions or through other regulatory proteins. Complexregulation of exopolysaccharide biosynthesis genes is a commonphenomenon. In the alginate biosynthesis pathway of Pseudomonasaeruginosa, response regulators AlgB and AlgR are both requiredas positive regulators of algD, the first gene of the alginatebiosynthetic operon (11, 24). Furthermore, activation of theeps operon, which harbors genes required for exopolysaccharideI production in Ralstonia solanacearum, is mediated by two responseregulators that are themselves under the control of a complexregulatory network (18).

View larger version (30K):
[in this window]
[in a new window]
FIG. 3. vps gene expression in logarithmic, planktonic, and biofilm growth states. (A) Expression of vpsA, vpsL, vpsT, and vpsR in logarithmically grown wild-type variants and mutants. (B) vpsL transcription in planktonic cells (stippled bars) and in biofilm cells (grey bars) in the indicated mutants. Error bars indicate the standard deviation. Note the different y axis scales in all panels.

To determine a possible interaction between the two positiveregulators, we analyzed vpsT and vpsR transcription in the describedstrains (Fig. 3). To this end, we constructed vpsRp-lacZ andvpsTp-lacZ transcriptional fusions by amplifying the upstreamregulatory sequences of vpsT (primers VCA0952_C and rev_prom)and vpsR (primers vpsR_prom_5' and vpsR_prom_3') and cloningthem into pRS415. Transcription of vpsT and vpsR was determinedduring exponential growth (optical density at 600 nm, 0.3 to0.4) in LB medium at 30°C by measuring ß-galactosidaseactivity. The results revealed that vpsT transcription was 45times higher in the rugose variant than in the smooth variant,confirming the trend of the initial microarray experiment (Yildizet al., submitted). Furthermore, vpsT transcription was 2.5-foldlower in the R ${Delta}$ vpsT mutant than in the rugose variant, indicatingthat VpsT positively regulates its own expression. Deletionof vpsR from the rugose variant caused a 50-fold reduction invpsT expression. vpsT expression in the R ${Delta}$ vpsR R ${Delta}$ vpsT double mutantwas similar to that in the R ${Delta}$ vpsR mutant.

We also determined vpsR transcription in wild-type phase variantsand the mutant strains. Figure 3 shows that vpsR transcriptionwas fourfold higher in the rugose variant than in the smoothvariant. The experiments also showed that, compared to thatfor the rugose variant, vpsR transcription was decreased twofoldin the R ${Delta}$ vpsT mutant and ninefold in the R ${Delta}$ vpsR mutant and inthe R ${Delta}$ vpsR ${Delta}$ vpsT double mutant. The results indicate that VpsTand VpsR positively regulate vpsR expression and that VpsR hada more dramatic effect on its own expression.

Next we examined whether vpsA and vpsL transcription differin planktonic and biofilm cells. For these measurements, overnight-growncultures were diluted in LB medium, inoculated into polystyrenepetri plates, and incubated at 30°C under static conditionsfor 12 h. ß-Galactosidase activities of planktonicand attached bacteria were compared for each of the strains.In the rugose variant, vpsL transcription was three times higherin the biofilm cells than in the planktonic cells (Fig. 3B).This result is similar to measurements in the V. cholerae O139strain (6). The smooth variant formed less-developed biofilms.Significantly, biofilm cells of the smooth variant did not havean increased vpsL transcription compared to that of planktoniccells. vpsL transcription in smooth planktonic and biofilm cellswas 30- and 100-fold lower, respectively, than expression inthe corresponding rugose cells. When vpsT was deleted from therugose variant, the biofilm growth-dependent vpsL inductiondid not occur. In addition, vpsL transcription in the R ${Delta}$ vpsTmutant planktonic and biofilm cells was 5- and 25-fold lowerthan that for the respective rugose variant cells.

Deletion of vpsR in the rugose variant or in the double mutantprevented vpsL transcription altogether, similar to the resultsobtained from logarithmically grown cells. vpsA transcriptionin the rugose variant was increased 25% in the biofilm cellscompared to that in the planktonic cells (data not shown). Incontrast, this induction was not observed in the other strains.

In summary, we have identified VpsT, a positive regulator ofvps gene expression. VpsT and the previously identified regulatorVpsR (25) are both necessary for maximal vps transcription inthe rugose variant. VpsT and VpsR both influence vpsA and vpsLexpression and positively regulate their own and each other'sexpression.

VpsR and VpsT are homologous to response regulators of two-componentregulatory systems. Response regulators usually act togetherwith a sensor histidine kinase. Sensor histidine kinase(s) thatregulates expression of vps genes and in turn the developmentof the rugose colonial morphology of V. cholerae O1 El Tor hasnot been identified thus far and is under investigation.

V. cholerae occupies different niches during its life cycleand is likely to be exposed to fluctuating environmental conditions(26). Two-component signal transduction systems are involvedin sensing and responding to environmental stimuli. Future workwill focus on the identification of environmental signals sensedby signal transduction systems involving VpsR and VpsT and onthe importance of the processes regulated by these two regulatorsin the adaptation responses of the pathogen.

ACKNOWLEDGMENTS

This work was supported by grants from UC TSR&TP, the EllisonMedical Foundation, and NIH (5R01AI055987-02). F.H.Y. is a newscholar in the Ellison Medical Foundation Global InfectiousDiseases Program.

We thank Uyen Tram and William Sullivan for help with the confocalmicroscopy and Sofie Salama and Karen Ottemann for criticalreading of the manuscript.

FOOTNOTES

* Corresponding author. Mailing address: Department of Environmental Toxicology, Jack Baskin Engineering, Room 269, University of California, Santa Cruz, Santa Cruz, CA 95064. Phone: (831) 459-1588. Fax: (831) 459-3524. E-mail: yildiz{at}etox.ucsc.edu.

REFERENCES

Top