Isolation and Characterization of NaCl-Sensitive Mutants of Caulobacter crescentus

An attempt to characterize Caulobacter crescentus genes importantfor the response to high concentrations of NaCl was initiatedby the isolation of mutants defective in survival in the presenceof 85 mM NaCl. A transposon Tn5 library was screened, and fivestrains which contained different genes disrupted by the transposonwere isolated. Three of the mutants had the Tn5 in genes involvedin lipopolysaccharide biosynthesis, one had the Tn5 in the nhaAgene, which encodes a Na⁺/H⁺ antiporter, and one had the Tn5in the ppiD gene, which encodes a peptidyl-prolyl cis-transisomerase. All the mutant strains showed severe growth arrestin the presence of 85 mM NaCl, but only the nhaA mutant showeddecreased viability under these conditions. All the mutantsexcept the nhaA mutant showed a slightly reduced viability inthe presence of 40 mM KCl, but all the strains showed a moresevere reduction in viability in the presence of 150 mM sucrose,suggesting that they are defective in responding to osmoticshock. The promoter regions of each disrupted gene were clonedin lacZ reporter vectors, and the pattern of expression in responseto NaCl and sucrose was determined; this showed that both agentsinduced ppiD and nhaA gene expression but did not induce theother genes. Furthermore, the ppiD gene was not induced by heatshock, indicating that it does not belong to the ${sigma}$ ³² regulon,as opposed to what was observed for its Escherichia coli homolog.

INTRODUCTION

Top

Introduction

Free-living bacteria must survive dramatic changes in extracellularosmolality. To respond to osmotic stress, they have developedsome very specific mechanisms that allow the immediate restorationof the osmotic balance. When exposed to osmotic upshifts, bacteriarespond in three overlapping phases: dehydration, adjustmentof cytoplasmic solvent composition and rehydration, and cellularremodeling (49).

The cytoplasmic membrane is highly permeable to water, and thereare also water-specific porins (aquaporins) and other sortsof membrane solute transporters that play a major role in theadaptation to high osmolality (8, 11, 46, 50). In Escherichiacoli, the immediate response to osmotic stress is K⁺ uptakethrough the Kdp and Trk systems (5, 42) followed by uptake ofcompatible solutes, which also depends on outer membrane porins(16).

Bacteria also have to cope with variable levels of environmentalsolutes that, besides causing osmotic stress, are also toxicto the bacterial cell, as is the case of Na⁺ ions. The Na⁺ intracellularconcentration has to be kept at tolerable levels, and bacteriause several Na⁺ efflux systems to attain sodium homeostasis.These systems may carry out active transport of the ion by usingenergy driven by ATP, coupled to metabolic enzyme transportor to respiration, or by catalyzing the efflux of Na⁺ in exchangefor external H⁺ (reviewed in references 12 and 37).

When E. coli cells are exposed to high osmolarity, the generalstress factor ${sigma}$ ^s is highly induced, driving the transcriptionof many genes involved in osmoprotection, including ostBA, treA(osmA), osmB, proU, and proP (23). Furthermore, osmotic shocktransiently induces heat shock genes belonging to the ${sigma}$ ³² and ${sigma}$ ^E regulons that are necessary for protein folding in the cytoplasmand cell envelope (7, 20, 33).

In Bacillus subtilis, the ${sigma}$ ^B regulon mediates the general stressresponse, recognizing environmental signals such as those resultingfrom salt stress, heat shock, or ethanol (1, 22). Besides thegeneral response, a more specific response may be achieved viaother sigma factors. Horsburgh and Moir showed that the extracytoplasmic-function(ECF) sigma factor ${sigma}$ ^M is required for growth and survival aftersalt stress (24). Although both ${sigma}$ ^B and ${sigma}$ ^M are important for thesalt stress response in Bacillus, their major contributionsare made in different phases of the cell cycle. Expression of ${sigma}$ ^M is highest during early exponential and mid-exponential growth,and its transcription is reduced in post-exponential growth,whereas ${sigma}$ ^B activity rises at the end of exponential growth beforefalling slowly as the cells enter the stationary phase (24,47).

Caulobacter crescentus is one of the major models for studyingcellular differentiation in prokaryotes, and its molecular mechanismshave been extensively characterized (27, 35). The complete genomeof C. crescentus is now available (34), which allows the evaluationof most of its metabolic capabilities. However, the responsesof C. crescentus to environmental stimuli are poorly characterized,except for the heat shock response (3, 6). The presence of asignificant number of sigma factors that belong to the extracytoplasmicfamily and the absence of a sigma S homolog suggests that thestress responses in this bacterium are probably more specificthan in E. coli (34).

As an aquatic bacterium, Caulobacter faces different stressesin its environment and must maintain an adequate intracellularsteady state at all situations. To investigate how Caulobacterresponds to osmotic stress, particularly to high NaCl concentrations,we have identified and characterized five genetic loci thatare important for its growth and survival under these conditions.Moreover, we have determined that NaCl and sucrose induce transcriptionof the ppiD and nhaA genes. This is a first step toward understandingthe regulatory network involved in the osmotic stress responsein Caulobacter.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains, growth conditions and genetic procedures.
C. crescentus NA1000 (15) was used as the wild type for allexperiments. The bacteria were grown at 30°C in peptoneyeast extract (PYE) medium or minimal M2 glucose medium (14)supplemented with kanamycin (5 µg/ml), tetracycline (1µg/ml), or nalidixic acid (25 µg/ml) as necessary.E. coli DH5 ${alpha}$ (Life Technologies) was used in the cloning procedures.E. coli was grown at 37°C in Luria-Bertani medium supplementedwith ampicillin (100 µg/ml) or tetracycline (12.5 µg/ml)as necessary. Plasmids were introduced into Caulobacter by conjugationwith E. coli strain S17-1 (44).

Isolation of NaCl-sensitive strains from a C. crescentus transposition library.
We used a previously constructed C. crescentus Tn5(Kan) transpositionlibrary of 5,000 clones (25). Screening was performed by growingthe cells in 96-well plates at 30°C followed by platingin PYE plus kanamycin containing 85 mM NaCl. Five mutant strainsthat showed no growth and were also unable to grow in platescontaining 50 mM NaCl, clones SP1503, SP1803, SP4104, SP4202,and SP4703, were selected.

Cell growth and survival tests.
The strains were grown in PYE plus kanamycin to mid-log phase;the cultures were then divided into 5-ml aliquots, and one ofthe following agents was added to an aliquot: 85 mM NaCl, 40mM KCl, and 150 mM sucrose. The cells were further incubatedat 30°C with agitation, and the growth of each culture wasdetermined by measuring the optical density at 600 nm at differenttimes during the incubation. A survival test was performed foreach osmotic agent tested, by adding the agent to exponentialcultures and taking aliquots at different times. The cells weresubjected to serial dilutions in PYE without addition of osmoticagents and plated in PYE medium to determine the number of CFU.

Determination of the Tn5 insertion sites.
The Tn5 insertion sites of the strains were determined by reversePCR as described previously (25). The sizes of the bands obtainedfor each strain were as follows: SP1503, 1.0 kb; SP1803, 1.6kb; SP4104, 0.7 kb; SP4202, 1.2 kb; and SP4703, two bands (1.3and 1.0 kb). The DNA fragments were gel purified and used forautomatic DNA sequencing with BigDye terminators on an ABI Prism377 sequencer (AP Biosystems).

Cloning of the promoter regions and gene expression analysis.
The complete sequence of the regions flanking the disruptedgenes was obtained from the C. crescentus genome sequence (34).Oligonucleotides 15-3AN (CAATGGATCCATGGAAGGGGCAGTGG) and 15-3A2(CAATGAATTCTAGGCGGACAGGAAGTAG) were used for PCR amplificationof the putative promoter region in strain SP4703, generatinga 675-bp fragment. Oligonucleotides 15-3A1 (CAATGGATCCGATCGATATTGAGCACAC)and 15-3A2 were used for PCR amplification of the putative promoterregion in strain SP1503, generating a 450-bp fragment. Oligonucleotides41-4E1 (CAATGAATTCTGGCTCAGACGATGCAAC) and 41-4E2 (CAATGGATCCTCGATCGCGCTCATACTG)were used for PCR amplification of the putative promoter regionin strain SP4104, generating a 470-bp fragment. Those fragmentswere cloned into pBluescript SK (Stratagene). To amplify theputative promoter of the disrupted gene in strain SP4202, oligonucleotides42-2F1 (CAATGGATCCTCGGAATTTGCGATCACC) and 42-2F2 (CAATGAATTCCGTCATACTCCGAGGGCC)were used, and for strain SP1803 oligonucleotides 18-3B1 (CAATGAATTCGGGCTGGTCAATGAACCG)and 18-3B2 (CAATGGATCCAGTTCAGCGGTATAGAAC) were used, generating980- and 1,065-bp fragments respectively. The amplified fragmentswere cloned into the TOPO vector from the TOPO TA cloning kitfor sequencing (Invitrogen) and confirmed by DNA sequencing.

PCRs were carried out with 1.5 µg of C. crescentus chromosomalDNA, 50 pmol of each set of oligonucleotides (described above),0.2 mM each deoxynucleoside triphosphate, 1.5 mM MgCl₂, 2.5U of Taq DNA polymerase (Invitrogen) and 1x PCR buffer (suppliedwith the enzyme). The PCR conditions were 5 min at 94°Cfollowed by 40 cycles of 90 s at 94°C, 1 min at 50°C,and 1 min at 72°C, with a final cycle of 7 min at 72°C.

All the fragments obtained were cloned into pRKlacZ290 (17),and promoter activities were determined by immunoprecipitationassays with monoclonal anti-ß-galactosidase antibody(Sigma) and by ß-galactosidase activity assays. Cellswere grown to mid-log phase in M2 glucose medium supplementedwith tetracycline (1 µg/ml); the cultures were them dividedinto aliquots, and one of the following agents was added toan aliquot; 85 mM NaCl and 150 mM sucrose. For the immunoprecipitationexperiments, cells were incubated further at 30°C with agitationand proteins were pulse-labeled with 10 µCi of [³⁵S]methionine(NEN) per ml for 5 min at different times during the incubation.Equal counts of labeled proteins (10⁶ cpm) were immunoprecipitatedas described previously (18). For the ß-galactosidaseactivity assays, the cultures were incubated at 30°C afteraddition of the agents and aliquots were taken at several timepoints to determine enzyme activity as described by Miller (31).

For the heat shock experiment, the NA1000 cells harboring constructppiD/lacZ lacZ were incubated either at 30 or at 42°C andaliquots were pulse-labeled with [³⁵S]methionine as above. Theproteins were then immunoprecipitated with antiserums anti-ß-galactosidaseand anti-GroEL (6).

RESULTS AND DISCUSSION

Top

Results and Discussion

Isolation of mutants sensitive to NaCl.
To identify genes necessary for the adaptation of C. crescentusto high salt concentrations, we screened a Tn5 insertion libraryof 5,000 clones for mutants that showed no growth in platescontaining 85 mM added NaCl. Previous experiments showed thatthis is the maximal concentration of NaCl that Caulobacter canwithstand without affecting growth or viability (data not shown).

These mutants isolated were further tested for growing in platescontaining 50 mM added NaCl. The presence of Tn5 was confirmedin five strains that were unable to grow even at this lowerconcentration of NaCl (Table 1). It should be noticed that transposoninsertion was found only once in each gene, suggesting thatthe genome is not saturated with mutations, and therefore therestill could be other genes involved in adaptation to high NaClconcentration that were not identified.

View this table:
[in this window]
[in a new window]
TABLE 1. Identification of the Tn5 insertion sites and similarity to ORF sequences from the database

Interestingly, three of these mutants were disrupted in genesprobably involved in lipopolysaccharide (LPS) biosynthesis (gmhB,rfa, and manA). Analysis of each Tn5 insertion region showedthat the manA and rfa are probably not cotranscribed with anyother downstream gene. The rfa gene product belongs to the heptosyltransferasefamily and has a putative function of LPS core biosynthesis.In E. coli, the two heptoses are linked to the 3-deoxy- ${alpha}$ -D-manno-oct-2-ulopyranosonicacid (KDO) by the action of two heptosyltransferases, WaaC (RfaC)and WaaF (RfaF), and heptose III is transferred to heptose IIby WaaQ (RfaQ) (19). The disrupted gene in SP4703 (manA) encodesa mannose-6-phosphate isomerase and is transcribed divergentlyto the gene disrupted in mutant SP1503. The manA gene productis involved in the conversion of fructose-6-phosphate to mannose-6-phosphatein a reversible reaction, and mannose is then incorporated intothe outer core of the Caulobacter LPS structure (38, 45). ThegmhB gene encodes a conserved hypothetical protein, which has36% identity to a D-glycero-D-manno-heptose 1,7-bisphosphatephosphatase, and its last codon overlaps the first codon ofa downstream gene encoding a dTDP-4-dehydroramnose reductase.The gmhB gene product in Aneurinibacillus thermoaerophilus isnecessary in the conversion of D- ${alpha}$ -D-heptose 1,7-bisphosphateto D- ${alpha}$ -D-heptose 1-phosphate (29).

The LPS structure of C. crescentus consists of an inner coreregion containing three molecules of KDO, two ${alpha}$ -L-glycero-D-mannoheptoseand one ${alpha}$ -D-glycero-D-mannoheptose; an outer core region of ${alpha}$ -D-mannose, ${alpha}$ -D-galactose, and ${alpha}$ -D-glucose (probably phosphorylated); anda lipid A moiety containing 3-OH-dodecanoic acid attached toa backbone with an undetermined structure (40). The structureof LPS O antigen is not yet completely clarified, but in a recentpaper Awram and Smit proposed that it is composed of a single4,6-dideoxy-4-aminohexose-N-acetylperosamine (4).

The LPS structures produced by these three strains are probablydefective in the core sugar composition. The inner core sugarsseem to be essential for the correct trimerization of porinsin the outer membrane of Salmonella enterica serovar Typhimurium,since a deep rough LPS has a very low percentage of assemblyof trimeric OmpF porin (43). In the mutants isolated in thiswork, the LPS would have only the two initial ${alpha}$ -L-glycero-D-mannoheptoses(gmhB and rfa), or extend further to the ${alpha}$ -D-glycero-D-mannoheptose(manA), and therefore they are probably affected in the assemblyof outer membrane protein complexes.

Strain SP1803 is disrupted in an open reading frame (ORF) thatencodes a homolog of the E. coli NhaA protein, involved in theadaptation to Na⁺ and alkaline pH (in the presence of Na⁺).This class of antiporters catalyzes efflux of Na⁺ in exchangefor external H⁺ and plays a major role in Na⁺ resistance. Despitethe coexistence of several antiporter systems in the bacterialcell, a single mutation or deletion in one of these antiporter-encodinggenes may produce a sodium-sensitive phenotype (9, 21, 26, 36,48).

The gene disrupted in SP4104 (ppiD) encodes a putative rotamasewith 26% identity to PpiD from Agrobacterium tumefaciens, anddownstream of it lies the trpE gene, which codes for anthranylatesynthase component I and probably forms an operon with ppiD.PpiD is a peptidyl-prolyl cis-trans isomerase of the parvulinclass, which catalyze cis-trans isomerization around X-Pro peptidebonds and are present both in the cytoplasm and in the periplasmicspace of the cell. There are two characterized enzymes of theparvulin family in the periplasm, SurA and PpiD. SurA helpsin the maturation and folding process of outer membrane proteinmonomers, and surA mutants have a defective cell envelope aswell as increased sensibility to hydrophobic agents (30, 32,41). ppiD is a suppressor of surA mutants in E. coli, and itis able to restore the ability of outer membrane protein foldingwhen overexpressed (10). Although C. crescentus also has a homologof the SurA protein, the ppiD knockout in strain SP4104 causedthe cells to be highly sensitive to osmotic stress, indicatingthat a deficient periplasmic protein folding prevents a properresponse.

Effect of osmotic agents in growth and viability of the mutants.
The inability of the mutants to grow in plates containing NaClcould be an effect of either growth impairment or loss of viability.To assess if NaCl causes growth arrest in the strains selected,growth curves were determined for cultures to which NaCl wasadded to yield a final NaCl concentration of 85 mM (Fig. 1).As can be seen, all the cultures showed growth inhibition afteraddition of the salt, with a more pronounced effect after 6h, except for mutant SP1803, which stopped growing immediatelywhen exposed to NaCl.

View larger version (31K):
[in this window]
[in a new window]
FIG. 1. Growth of the parental and mutant strains under different osmotic challenges. Strains NA1000, SP1503, SP1803, SP4104, SP4202, and SP4703 were grown in PYE medium to mid-log phase, and after 3 h the culture was divided into aliquots and the following solutes were added to the respective final concentrations: 85 mM NaCl (black squares), 40 mM KCl (black triangles), and 150 mM sucrose (white triangles). As a control, one aliquot was left without additions (white squares). Growth was determined by measuring the absorbance at 600 nm.

To determine whether the addition of NaCl was also affectingthe viability of the cells, a viability test was performed bydetermining the number of CFU after incubation with the salt(Fig. 2A). The experiment showed that viability was not affectedin four of the mutants, but the nhaA strain suffered a drasticreduction in viability after 3 h in the presence of NaCl, whichagrees with the growth arrest observed for this mutant. Thisloss of viability is not accompanied by lysis of the cells,as judged by microscopy and absorbance of the nhaA strain after24 h in the presence of NaCl. These results indicate that theother four mutants, despite being affected, could still resumegrowth after removal of the salt (when plated for the CFU counts),suggesting that in these mutants the internal Na⁺ concentrationhas not irreversibly damaged any vital functions.

View larger version (22K):
[in this window]
[in a new window]
FIG. 2. Viability test of the wild-type and mutant strains to osmotic agents. Strains NA1000, SP1503, SP1803, SP4104, SP4202, and SP4703 were grown in PYE medium to mid-log phase, and the following agents were added to the cultures at time zero: 85 mM NaCl (A), 40 mM KCl (B), or 150 mM sucrose (C). At intervals after addition of the salt, aliquots were removed and plated for counting colonies, and survival was determined relative to the culture without addition (time zero). The results are the average of three independent results. The key for each curve depicted in panel A is the same for all graphs.

The mutant strains were then evaluated for their ability torespond to other osmotic agents, such as KCl and sucrose. Ascan be seen in Fig. 1, 40 mM KCl caused a more severe reductionin growth of the ppiD and manA strains while nhaA, gmhB, andrfa showed a less strong effect. Accordingly, the ppiD, manA,gmhB, and rfa strains showed a more significant decrease inviability while nhaA was not affected, as expected (Fig. 2B).These results suggested that with exception of nhaA, the deficientresponse to NaCl could be a result of a broader defect in thecells, possibly involving a deficiency in the osmotic stressresponse. To test this hypothesis, we tested the effect of anonionic high-osmolarity shock, by growing the cells in thepresence of 150 mM sucrose (Fig. 1 and 2C). The results showeda decrease in absorbance in the presence of high sucrose concentrationsfor all mutants except the nhaA strain. However, this strainshowed a similar decrease in viability in the presence of sucroseto that shown by the others, which suggested that the increasein absorbance was a result of an alteration in the cell volumeinstead of an increase in cell number. Microscopy observationshowed that this was really the case; the cells of the nhaAstrain become filamentous in the presence of sucrose, causingan increase in absorbance. The reason for the loss of viabilityof the nhaA strain in 150 mM sucrose is not clear, but it couldbe that in the absence of Na⁺ efflux the cell cannot achieveosmotic balance under high-osmolarity conditions.

The reasons for the impaired osmotic stress response of thedefective-LPS strains and the ppiD mutant could be that theproteins responsible for the cell adaptation to high osmolarityare not properly folded in the outer compartment of the cell.Further data that corroborate this idea are that overproductionof PpiD complements the defects of the htrM (rfaD) mutation,where the gene htrM encodes an ADP-1-glycol-D-mannoheptose-6epimerase, which is needed for LPS biogenesis (39).

Transcriptional regulation in response to osmotic stress.
To analyze whether the disrupted genes could be regulated byosmotic upshift, the putative promoter regions of each genewere cloned into a lacZ reporter vector and introduced intoC. crescentus NA1000 strain. The cells were incubated with either85 mM NaCl or 150 mM sucrose, and transcription was measuredboth by pulse-labeling the proteins with [³⁵S]methionine for5 min at several time points and immunoprecipitating the ß-galactosidaseand by performing ß-galactosidase activity assays.The promoter fusions of manA, gmhB and rfa showed no detectableinduction in the presence of either osmotic agent (not shown).

Figure 3A shows that transcription driven from the promoterof ppiD is increased from 60 min after addition of NaCl andfrom 30 min after addition of sucrose, indicating that thisgene is induced by high-osmolarity conditions. This higher levelof expression of ppiD in the presence of the osmotic agentsis maintained for at least 5 h, showing a 2-fold induction byNaCl and a 2.2-fold induction by sucrose. The Caulobacter nhaAhomolog showed a small induction by NaCl in ß-galactosidaseactivity assays (1.6-fold induction [Fig. 3C]), in contrastto the E. coli nhaA gene, whose transcription is highly inducedby Na⁺ and Li⁺ (5- to 10-fold after 60 minutes) but not by increasesin osmolarity and ionic strength (13, 28). A very small increasein ß-galactosidase activity (1.2-fold) was observedafter incubation with 150 mM sucrose, but whether this facthas any physiological relevance remains to be determined.

View larger version (87K):
[in this window]
[in a new window]
FIG. 3. Determination of promoter activity of the transcriptional fusions. The promoter regions of the genes disrupted in mutants SP4104 and SP1803 were cloned in the transcriptional reporter plasmid pRKlacZ290 (17), generating fusions ppiD-lacZ (A and B) and nhaA-lacZ (C and D), respectively. Mid-log-phase cells were incubated with either 85 mM NaCl (A and C) or 150 mM sucrose (B and D), and ß-galactosidase (ß-gal) activity was assayed at different time points after addition. The background value of ß-galactosidase activity given by pRKlacZ290 was subtracted from the results. (E) Expression of the ppiD-lacZ fusion under heat shock. The proteins were labeled with [³⁵S]methionine for 5 min at several time points after incubation at 42°C, equal counts were immunoprecipitated with both anti-ß-galactosidase and anti-GroEL antisera, and the proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

In E. coli, ppiD is induced by high temperature, and its transcriptionis dependent on ${sigma}$ ³²-RNA polymerase. PpiD is so far the only periplasmicchaperone that has shown regulation by the heat shock sigmafactor ${sigma}$ ³², and is also regulated by the two-component systemCpxR-CpxA (10). ${sigma}$ ³² from Caulobacter has a well-known consensus-bindingsite, and it regulates specific heat shock genes (2, 3, 6).The putative promoter region of ppiD in Caulobacter does nothave any sequence resembling the ${sigma}$ ³² consensus, and transcriptionis not induced by heat shock (Fig. 3E). These results indicatethat in C. crescentus, ppiD transcription is dependent on otherregulatory mechanisms, being induced by osmotic upshift butnot by heat shock. Since Caulobacter has 13 sigma factors ofthe ECF family (34), one of these alternative factors couldregulate ppiD expression, possibly together with a two-componentregulatory system. In any case, further investigation of theregulatory regions involved in the osmotic shock induction willallow us to determine the pathways used by Caulobacter to respondto this stress condition.

ACKNOWLEDGMENTS

This work was supported by Fundação de Amparoà Pesquisa do Estado de São Paulo (FAPESP grant2000/02245-0). During the course of this work, L.F.G.Z. andV.C.S.I. were supported by fellowships from FAPESP.

We thank Suely L. Gomes for the anti-GroEL antiserum and forcritical reading of the manuscript.

FOOTNOTES

* Corresponding author. Mailing address: Depto. de Microbiologia, Instituto de Ciências Biomédicas, Universidade de São Paulo, Av. Prof. Lineu Prestes 1374, 05508-900 São Paulo SP, Brazil. Phone: 55-11-3091-7299. Fax: 55-11-3091-7354. E-mail: mvmarque{at}usp.br.

REFERENCES

Top