Control of Virulence by the Two-Component System CiaR/H Is Mediated via HtrA, a Major Virulence Factor of Streptococcus pneumoniae

The CiaR/H two-component system is involved in regulating virulenceand competence in Streptococcus pneumoniae. The system is knownto regulate many genes, including that for high-temperaturerequirement A (HtrA). This gene has been implicated in the abilityof the pneumococcus to colonize the nasopharynx of infant rats.We reported previously that deletion of the gene for HtrA madethe pneumococcal strains much less virulent in mouse models,less able to grow at higher temperatures, and more sensitiveto oxidative stress. In this report, we show that the growthphenotype as well as sensitivity to oxidative stress of ${Delta}$ ciaRmutant was very similar to that of a ${Delta}$ htrA mutant and that theexpression of the HtrA protein was reduced in a ciaR-null mutant.Both the in vitro phenotype and the reduced virulence of ${Delta}$ ciaRmutant could be restored by increasing the expression of HtrA.

INTRODUCTION

Top

Introduction

Streptococcus pneumoniae (the pneumococcus) is an importanthuman pathogen. This gram-positive organism is a major causeof a variety of diseases such as pneumonia, bacteremia, meningitis,otitis media, and sinusitis in both adults and children allover the world (40, 43). The nasopharynx is the major reservoirof pneumococci, from which they can spread to other sites suchas bloodstream or lung tissues (44).

Pathogenic bacteria encounter a number of environmental stressesduring their life cycle such as temperature shifts, variationsin osmolarity, changes in pH, and nutrient deprivation (45).The pneumococcus responds to these stresses, particularly heatstress, by mediating a cascade of events leading to the synthesisof a unique group of proteins called heat shock proteins (HSPs)(7). The production of these proteins represents a protectivecellular response to cope with the stress-induced damage ofproteins (26). HSPs are molecular chaperones or proteases thattake part in protein quality control during normal growth andunder stress-inducing conditions (6, 12).

Recent studies have demonstrated the role of bacterial two-componentsignal-transducing systems in mediating adaptive responses toenvironmental signals (30, 38). The pneumococcal CiaR/H two-componentsystem consists of a sensor histidine kinase, CiaH, anchoredin the cell membrane and a cytoplasmic response regulator, CiaR,which is a DNA-binding protein involved in the regulation ofgenes in response to environmental signals sensed by CiaH (15).This system was identified as 1 of 13 two-component signal-transducingsystems in two genomic screens (24, 42). Pleotropic effectscaused by cia mutations in the pneumococcus include sensitivityto cefotaxime, an ability to form protoplast, susceptibilityto lysis by deoxycholate (11), and a tendency to early lysis(24). CiaH mutants also have transformation deficiency (8, 11,16). In vivo, CiaR/H has been shown to contribute to colonizationof the mouse lung (42) and the nasopharynx of infant rats (35)and to be involved in systemic infection in mice (27). Previousstudies have shown that the CiaR/H regulon contains many genes,including the high-temperature requirement A gene (29, 35).

High-temperature requirement A (HtrA), also known as DegP orDO protease (36), is a stress-induced serine protease that manifestsboth general molecular chaperone and proteolytic activitiesand switches from chaperone to protease in a temperature-dependentmanner (37), the protease activity being most apparent at elevatedtemperature. It is involved in the degradation of periplasmicmisfolded proteins in Escherichia coli (39). The evidence pointsto a major role for this protease in helping organisms to surviveenvironmental stresses such as elevated temperature and oxidativeand osmotic stresses (13, 31, 46). HtrA is known to be involvedin the virulence of many gram-negative bacteria such as Salmonellaenterica serovar Typhimurium (2), Brucella abortus (9), andYersinia enterocolitica (25). This protease is also requiredfor full virulence of the gram-positive bacterium Streptococcuspyogenes (20). An HtrA homologue has also been identified inS. pneumoniae (10) and is regulated by the CiaR/H two-componentsystem (29, 35). These studies demonstrated that CiaR positivelyregulates HtrA and that this regulation is probably direct,as the CiaR was shown to bind to DNA upstream of the htrA gene.An S. pneumoniae strain lacking the htrA gene showed decreasedfitness in a competitive model of colonization (35). HtrA wasidentified as a virulence factor of the pneumococcus in a signature-taggedmutagenesis screen (17).

We previously reported that HtrA plays a crucial role in virulenceof the pneumococcus (18), as HtrA-deficient strains are attenuatedin both pneumonia and bacteremia models of infection. HtrA isinvolved in resistance to temperature and oxidative stress andalso in genetic competence. Here we present evidence that thecontribution of the CiaR/H system to thermotolerance, oxidativestress tolerance, and virulence of S. pneumoniae is mediatedthrough HtrA.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains, primers, and growth conditions. Bacterial strains and oligonucleotide primers used in this studyare listed in Table 1. S. pneumoniae strains were grown routinelyon blood agar base 2 (Oxoid, Basingstoke, United Kingdom) supplementedwith 5% (vol/vol) defibrinated horse blood (BAB; E&O Laboratories,Bonnybridge, United Kingdom) and in brain heart infusion (BHI)broth at 37°C. E. coli ultracompetent cells (Stratagene)used for cloning were grown on Luria-Bertani broth or Luria-Bertaniagar plates. Where appropriate, antibiotics were added to thegrowth medium at the following concentrations: ampicillin at50 µg/ml, erythromycin at 1 mg/ml for E. coli or 1 µg/mlfor S. pneumoniae, and spectinomycin at 100 µg/ml.

View this table:
[in this window]
[in a new window]
TABLE 1. List of bacterial strains and primers used in this study

DNA techniques and transformation. Chromosomal DNA was prepared from S. pneumoniae as describedelsewhere (4). PCR, restriction endonuclease digestions, DNAligation, and DNA electrophoresis were performed according tostandard protocols (34). Kits from QIAGEN were used for DNApurification and plasmid preparations according to the manufacturer'sinstructions. Transformation of E. coli with plasmid DNA wascarried out according to the manufacturer's instructions (Stratagene).Transformation of S. pneumoniae D39 (serotype 2) was done byusing a modification of the method of Lacks and Hotchkiss (23).Competent cells of D39 were prepared as follows: 200 µlof a glycerol stock of cells (previously frozen down at an opticaldensity at 600 nm [OD₆₀₀] of $~$ 0.6) was used to inoculate 10 mlof CAT medium (32) supplemented with 20% glucose and 0.5 M K₂HPO₄(CAT/GP medium) and incubated at 37°C until they reachedan OD₆₀₀ of 0.3 to 0.4. Cells were then harvested by centrifugationat 5,000 x g for 10 min at 4°C, and cell pellets were resuspendedin CAT/GP medium containing 20% glycerol and frozen down at–80°C as 100-µl aliquots. An aliquot of competentcells was thawed and diluted 1/10 in CAT/GP medium supplementedwith 4% bovine serum albumin and 0.1 M CaCl₂ (CTM medium), and100 ng of CSP1/ml was added (33). Different concentrations oftransforming DNA (0.5 to 2 µg) were then added, and thecells were incubated at 37°C for 10 min and then at 30°Cfor 20 min. Transformed cells were selected on BAB plates withappropriate antibiotic selection.

Construction of the plasmid expressing the htrA gene in S. pneumoniae. htrA was expressed in pAL2 plasmid (3) from the constitutivepromoter of an S. pneumoniae aminopterin resistance operon (ami)(1). The construction of the pAL2-HtrA plasmid was describedelsewhere (18), and a schematic representation of it is shownin Fig. 1. To test whether the effect of the presence of thepAL2-HtrA plasmid was solely due to htrA insertion, the EcoRI-digestedpAL2 fragment containing the ami promoter was self-ligated tocreate the empty plasmid pAL2YI. This plasmid was used to transformthe D39 wild type to generate the D39/pAL2YI strain.

View larger version (15K):
[in this window]
[in a new window]
FIG. 1. A schematic representation of pAL2-HtrA plasmid used for expression of the htrA gene in S. pneumoniae. The pAL2-HtrA plasmid was modified from pAL2 plasmid (3) as explained in Materials and Methods. Arrows indicate the locations and orientations of open reading frames. The nucleotide sequence shows the region between the ami promoter and the start codon of htrA. The gram-positive ribosome-binding site is shown in bold and uppercase characters, the htrA start codon is shown in uppercase characters, and restriction sites are underlined. Em-R, erythromycin resistance gene; Cm-R, chloramphenicol resistance gene.

Construction of ${Delta}$ htrA and ${Delta}$ ciaR mutants and complementation with HtrA in serotype 2 background. The ${Delta}$ htrA mutant was constructed as described elsewhere (18).The ${Delta}$ ciaR mutant strain cia spc 136b (28) was used to move the ${Delta}$ ciaR mutation into D39 by PCR amplification of a 6-kb fragmentcorresponding to the ${Delta}$ ciaR mutation by the use of primer pairMP144 and MP145 (28); the amplified fragment was used directlyto transform the D39 wild type to create D39 ${Delta}$ ciaR. Transformantcells were selected on spectinomycin. The pAL2-HtrA plasmid,which carries an erythromycin-resistant cassette, was used totransform D39 ${Delta}$ htrA and D39 ${Delta}$ ciaR mutants to generate the complementedstrains D39 ${Delta}$ htrA/phtrA⁺ and D39 ${Delta}$ ciaR/phtrA⁺, in which HtrA isexpressed from the plasmid.

Western immunoblotting. Total cellular proteins were extracted from cultures grown tomid-log phase as static cultures in BHI broth. Cell pelletswere collected by centrifugation at 5,000 x g for 15 min andwere resuspended in 1 ml of phosphate-buffered saline (PBS).Cell suspensions were sonicated four times at 30 s each timewith a 13-mm probe (Sonicator, Vibra cell; Sonics & MaterialsInc.) with a power output of 36 W. The tubes containing thesamples were kept in crushed ice during sonication. The celldebris was removed by centrifugation at 13,000 x g for 10 min.Concentrations of total proteins were determined by the Bradfordassay, with the use of bovine serum albumin as a standard (5).For Western blot analysis, 15 µg of total proteins fromeach strain was separated on sodium dodecyl sulfate-10% polyacrylamidegel electrophoresis, electroblotted onto nitrocellulose membranes(Amersham), and reacted with specific antiserum against HtrAby a standard protocol (34). To ensure that nitrocellulose wasnot saturated with bound HtrA, a dilution series was performed;the results confirmed that using this amount of total cell protein,band intensity was related to the amount of HtrA.

In vitro experiments. To compare the abilities of ${Delta}$ ciaR of both type 2 and type 3 pneumococciand the complemented strain D39 ${Delta}$ ciaR/phtrA⁺ mutants to grow atnormal and elevated temperatures to those of their wild types,the same number of viable cells (10⁶ CFU/ml) was used to inoculateBHI broth prewarmed at 37 and 40°C. At 1-h intervals, sampleswere withdrawn to measure the OD₆₀₀ and viable counts. The sensitivityof D39 ${Delta}$ ciaR and D39 ${Delta}$ ciaR/phtrA⁺ strains to H₂O₂ was tested byexposing aliquots of cultures grown to an OD of $~$ 0.3 to 40 mMof H₂O₂ for 5, 10, and 15 min at room temperature. The viablecells were counted by plating onto blood agar plates beforeand after the exposure to H₂O₂, and the result was expressedas percent survival (19).

Mice and infections. Female outbred MF1 mice (25 to 30 g of body weight) were purchasedfrom Harlan Olac, Bicester, United Kingdom. They were used atthe age of 9 weeks. For intranasal infection, mice were lightlyanesthetized with 1.5% (vol/vol) halothane and a 50-µlinfectious dose (10⁶ CFU/mouse for type 2 strains or 10⁷ CFU/mousefor type 3 strains) was administered to the nostrils of miceheld vertically (21). For intraperitoneal infection, the infectiousdose (2 x 10⁵ CFU/mouse) was resuspended in 200 µl ofsterile PBS and administered into the peritoneal cavity of mice.After the infectious dose was administered, mice were observedand the development of symptoms was recorded.

Bacteriological investigation. At prechosen intervals after infection, groups of mice weresacrificed by cervical dislocation and a blood sample was removedvia cardiac puncture. Lungs were removed and homogenized in5 ml of PBS with a glass handheld tissue homogenizer (Jencons,Leighton Buzzard, United Kingdom). Viable bacteria in lung andblood samples were counted by plating out serial 10-fold dilutionson BAB (22).

Competition experiments. Competitive analysis of virulence was examined using a slightmodification of the method of Hava and Camilli (17). Essentially,challenge doses of wild-type and mutant strains were mixed ina 1:1 ratio and inoculated intranasally at 2 x 10⁶ and 2 x 10⁵bacteria for intraperitoneal infections. After 24 and 48 h,bacteria were recovered from the lungs and blood. The numberof wild-type versus mutant bacteria recovered was determinedby plating on nonselective and selective media, and competitiveindices were calculated as the ratio of mutant to wild-typebacteria recovered from each animal.

Statistical analysis. Statistical analyses were carried out using StatView 4.1 (AbacusConcept). Survival times were analyzed by using nonparametricMann-Whitney U analysis. Bacteriology results are expressedas geometric means ± standard errors of the means (SEM).Comparisons of bacterial loads in the time course bacteriologyexperiment were performed using unpaired t tests. In all analyses,a P value of <0.05 was considered statistically significant.

RESULTS

Top

Results

Introduction of a ${Delta}$ ciaR mutation into S. pneumoniae. A ${Delta}$ ciaR mutation in strain 0100993 was made by allelic replacementwith an erythromycin resistance marker and was kindly providedby Martin Burnham (42). To perform the complementation experimentswith the pAL2-HtrA plasmid (erythromycin selection) we neededto use another antibiotic marker to disrupt the ciaR gene. Wetherefore constructed a ${Delta}$ ciaR strain of D39 with resistance tospectinomycin by transformation with a 6-kb fragment amplifiedby PCR from strain ciaspec136b (28). This mutation was confirmedby PCR and sequencing (data not shown).

HtrA Western immunoblot analysis. As CiaR is involved in virulence and regulates htrA, which wehave shown previously to be a crucial virulence factor (18),we asked whether the CiaR phenotype was associated with thedown-regulation of htrA. To test this we constructed a plasmidto allow the constitutive expression of HtrA in the pneumococcus.Plasmid pAL2-HtrA (Fig. 1) allows the expression of HtrA fromthe pneumococcal ami promoter (1). Western blotting was usedto examine the levels of HtrA expressed by pneumococcal strains.The level of expression of HtrA was shown to be higher at 40°Cthan at 37°C in the wild-type D39 but not in the D39 ${Delta}$ ciaRmutant (Fig. 2A), confirming HtrA to be an HSP of the pneumococcusand to be up-regulated by CiaR in response to heat shock. Asexpected, HtrA expression was abolished in D39 ${Delta}$ htrA. We couldrestore levels of expression of HtrA in the D39 ${Delta}$ htrA mutant tolevels similar to wild-type levels with the plasmid pAL2-HtrA.The level of HtrA was reduced but not abolished in ${Delta}$ ciaR, andintroduction of pAL2-HtrA into the ${Delta}$ ciaR strain resulted in expressionof levels of HtrA similar to wild-type levels (Fig. 2B). Westernblotting is semiquantitative, but a dilution series of samplesallowed us to confirm that we were loading amounts of HtrA thatwere not saturating for the nitrocellulose membrane. Visualexamination of these blots suggests that deletion of CiaR resultsin an approximately fourfold reduction in HtrA. Expression ofHtrA in the ${Delta}$ ciaR strain from the plasmid restored the levelof HtrA to that of the wild type. Expression of HtrA from theplasmid in the wild-type strain did not increase total levelsof HtrA (data not shown).

View larger version (31K):
[in this window]
[in a new window]
FIG. 2. Western immunoblot analysis. (A) Expression of HtrA at 37 and 40°C in wild-type (WT) strain D39 and strain D39 ${Delta}$ ciaR. (B) Levels of HtrA in strain D39 ${Delta}$ htrA and strain D39 ${Delta}$ ciaR and their complemented strains grown at 37°C.

In vitro growth phenotype of ${Delta}$ ciaR mutants. We have shown before that HtrA is involved in the ability ofthe type 2 pneumococci to grow at elevated temperature and thatthe effect of htrA deletion was to slow growth rather than preventit (18). We also found that the D39 ${Delta}$ htrA mutant has a decreasedrate of autolysis compared to its parent strain at 40°C,as indicated by viable counts at the stationary and declinephases of growth (data not shown). We also studied the growthphenotype of D39 ${Delta}$ ciaR mutant at normal and elevated temperatures.In similarity to the results seen with D39 ${Delta}$ htrA, the growth ofthe D39 ${Delta}$ ciaR mutant was less than that of the wild type at 40°C,as judged by OD (Fig. 3A). The D39 ${Delta}$ ciaR strain showed a decreasedrate of autolysis after reaching the stationary phase of growthat both 37 and 40°C compared to the D39 wild type (Fig.3B). This growth phenotype of strain D39 ${Delta}$ ciaR could be restoredby increasing HtrA levels in the complemented strain D39 ${Delta}$ ciaR/phtrA⁺(Fig. 3). Introduction of the pAL2YI vector had no effect onthe growth of strain D39 at either 37 or 40°C (data notshown). Also, introduction of the pAL2-HtrA plasmid into theD39 wild-type strain did not result in altered growth at thesetemperatures.

View larger version (27K):
[in this window]
[in a new window]
FIG. 3. Growth curves of D39 wild-type (WT), D39 ${Delta}$ ciaR, and D39 ${Delta}$ ciaR/phtrA⁺ strains at 37 and 40°C; results are represented as OD₆₀₀ values in panel A and as viable counts after reaching the stationary phase of growth in panel B. A total of 10⁶ CFU/ml of each strain was used to inoculate BHI broth prewarmed at the indicated temperatures, and samples were withdrawn at 1-h intervals to measure the OD₆₀₀ values and viable counts.

The effect of ciaR mutation in type 3 strain 0100993 was moredramatic at elevated temperature when compared to type 2 results.Growth of the 0100993 ${Delta}$ ciaR mutant was similar to that of thewild type and a similar rate of autolysis was seen at 37°C,whereas growth at 40°C was prevented by the mutation (Fig.4).

View larger version (16K):
[in this window]
[in a new window]
FIG. 4. Growth curves of the type 3 0100993 wild-type (WT) strain and its ${Delta}$ ciaR mutant at 37 and 40°C; results are represented by OD₆₀₀ values in panel A and by viable counts in panel B. The growth of the 0100993 ${Delta}$ ciaR mutant was impaired at 40°C and showed a rate of autolysis similar to that of the wild type at 37°C.

Sensitivity of the D39 ${Delta}$ ciaR strain to oxidative stress is similar to that of the D39 ${Delta}$ htrA mutant and can be restored by complementation with HtrA. We have shown previously that HtrA is involved in the abilityof D39 to resist oxidative stress (18). To determine whetherreduced expression of HtrA in strain D39 ${Delta}$ ciaR resulted in sensitivityto oxidative stress we compared D39 ${Delta}$ ciaR to the wild type andthe complemented strain D39 ${Delta}$ ciaR/phtrA⁺ for sensitivity to hydrogenperoxide. The D39 ${Delta}$ ciaR strain was significantly more sensitiveto peroxide than the parent strain D39 after 10 and 15 min ofexposure to hydrogen peroxide. There was no statistically significantdifference 5 min after peroxide treatment, probably becausestrain D39 ${Delta}$ ciaR still expresses some HtrA. By restoring HtrAto a level similar to that of the wild type in D39 ${Delta}$ ciaR/phtrA⁺,the strain was again identical to the wild type in responseto oxidative stress (Fig. 5). The sensitivity of strain D39 ${Delta}$ ciaRto oxidative stress is therefore similar to what we reportedpreviously for strain D39 ${Delta}$ htrA (18) and could be explained bythe down-regulation of HtrA in the CiaR-null mutant.

View larger version (21K):
[in this window]
[in a new window]
FIG. 5. H₂O₂ sensitivity assay for D39 wild-type (WT), D39 ${Delta}$ ciaR, and D39 ${Delta}$ ciaR/phtrA⁺ strains. H₂O₂ (40 mM) was added to 1-ml aliquots of culture grown to an OD₆₀₀ of $~$ 0.3. Viable counts were performed on BAB plates before and after the addition of peroxide, and the percentages of survival were calculated. Values expressed are the means (± SEM) of three independent experiments. *, P < 0.05 for lower survival for the D39 ${Delta}$ ciaR mutant than the wild type.

The ${Delta}$ ciaR mutant displays attenuated virulence in vivo. The CiaR/H system has previously been reported to play a rolein mouse lung colonization (42) and colonization of the nasopharynxof infant rats (35) and has also been identified as playinga role in a mouse model of systemic disease (27). We confirmedand further investigated the role of CiaR in our model of infection.Initially we compared a ${Delta}$ ciaR mutant on a serotype 3 backgroundwith its wild-type parent (strain 0100993). This is the samemutant as that used by Throup and coworkers (42) and Sebertand coworkers (35) in the studies described above. All animalschallenged intranasally with 10⁷ CFU of the parent strain weremoribund by 144 h postinfection. In contrast, the 0100993 ${Delta}$ ciaRmutant only caused the moribund state in 25% of the animals(Fig. 6A). Analysis of bacteriology showed that the 0100993 ${Delta}$ ciaR strain grew to a viable count of approximately 10³ CFUin lung compared to 10⁶ CFU for the wild-type organism. Thelevel of bacteremia caused by the 0100993 ${Delta}$ ciaR strain was belowthe limit of detection of the assay (approximately 100 organismsper ml) (Fig. 6B). Competition experiments showed that CiaRplays a role in both the lung and systemic infection models(Table 2). We have also confirmed that CiaR plays a role inthe virulence of strain D39 (see below).

View larger version (29K):
[in this window]
[in a new window]
FIG. 6. Effect of CiaR on virulence of strain 0100993. (A). Survival of animals given 10⁷ CFU of wild-type or ${Delta}$ ciaR bacteria by the intranasal route; n = 10. (B). Mean (SEM) bacteriology in lung (left) and blood (right) during infection, n = 5. The broken line represents the limit of detection of the assay. (A) *, P < 0.05 for shorter survival times for the wild-type (WT) strain than for the ${Delta}$ ciaR mutant; (B) *, P < 0.05 for lower bacterial loads for the ${Delta}$ ciaR strain than for the wild-type strain.

View this table:
[in this window]
[in a new window]
TABLE 2. Competitive index for 0100993 wild type versus 0100993 ${Delta}$ ciaR in intranasal and intraperitoneal infections^a

HtrA can complement the CiaR mutation and restore virulence. As the growth phenotype and sensitivity to peroxide of the D39 ${Delta}$ ciaRmutant could be restored by complementation with HtrA, we investigatedthe virulence phenotype of the knockout and complemented strainsin our mouse pneumonia model. For these studies, all mutantsused were on a D39 background and a dose of 10⁶ CFU was givenintranasally. For the D39 wild-type strain, all animals succumbedto the infection by 72 h whereas no animals became sick whenchallenged with strain D39 ${Delta}$ htrA (Fig. 7). A ${Delta}$ ciaR mutant of D39showed attenuation in the model, with 40% of the animals becomingmoribund during the experiment. When the pAL2-HtrA plasmid wasintroduced into the D39 ${Delta}$ htrA strain it was restored to full virulence.Moreover, when the level of HtrA expression was corrected inthe ${Delta}$ ciaR mutant this was also restored to full virulence.

View larger version (14K):
[in this window]
[in a new window]
FIG. 7. Effect of complementation with HtrA on virulence. Strains D39 ${Delta}$ htrA and D39 ${Delta}$ ciaR show reduced virulence, as judged by survival time. Complementation with HtrA (from plasmid pAL2-HtrA) causes these strains to revert to full virulence. WT, wild type.

DISCUSSION

Top

Discussion

The CiaR/H two-component system is known to be important ina number of aspects of pneumococcal biology. Mutations in thehistidine protein kinase ciaH conferred resistance to beta-lactamantibiotics, suggesting that the system may control genes importantin cell wall metabolism (14). The system is also involved inthe control of bacterial lysis (11, 16, 24) and the regulationof genetic competence (8, 14, 28). Previous studies have shownthat this system plays a role in colonization of the lungs ofmice (42) and the nasopharynx of infant rats (35). We have usedthe same mutant of S. pneumoniae (kindly provided by MartinBurnham of SmithKlineBeecham [now GlaxoSmithKline]) used inthese two studies to show that CiaR/H also plays a major rolein the causation of disease in the mouse. The ${Delta}$ ciaR/H mutantof strain 0100993 was much reduced in its overall virulence(as judged by survival times) when given intranasally. The mutantorganism grew to about 10³ CFU in lungs compared to 10⁶ forthe wild-type organism (a reduction of 3 logs). A large reductionin growth in the lung was also observed by Throup et al. (42),who observed a reduction of approximately 5 logs in growth inthe lung due to disruption of the CiaR/H operon. In studiesof colonization of the rat nasopharynx (35), inoculation ofwild-type organisms resulted in maximum colonization levelsof about 10 ⁶ CFU/ml of nasal wash whereas the ${Delta}$ ciaR/H mutantdid not colonize the nasopharynx. In the experiments reportedhere, the mutant organism only invaded the bloodstream to verylow levels (at the limit of detection of 100 organisms per ml).This two-component system therefore seems to be very importantfor colonization and growth in the lungs and also for invasionof the bloodstream.

Low-level bacteremia following intranasal challenge could becaused either by poor invasion from the lungs or by poor growthin the blood. Competition experiments using wild-type and mutantstrains showed that the mutant was much less able to colonizethe lungs and also less able to grow systemically. We have confirmedthat CiaR also plays a role in the virulence of type 2 strainD39. It is important to investigate the role of these systemsin more than one strain, as we have recently shown that thecontribution of two-component systems to virulence can be straindependent (4). A role of CiaR/H in the systemic virulence ofD39 was also shown by Marra and coworkers (27).

As CiaR/H plays a role in virulence, the next question is howdoes this system mediate pathogenesis of infection? The ciaregulon has already been defined and includes genes importantfor the synthesis and modification of cell wall polymers, peptidepheromone and bacteriocin production, and the htrA-spoJ region(29). Because it is already known that CiaR/H regulates thegene for HtrA (29, 35) and that deletion of the htrA gene reducesthe ability of the pneumococcus to colonize the nasopharynx(35) and attenuates a type 2 pneumococcal strain in both pneumoniaand bacteremia models of infection (18), we attempted to definethe contribution of HtrA to the overall CiaR/H phenotype.

Both CiaR-null and HtrA-null mutants have reduced virulence,and the CiaR/H system is known to regulate HtrA expression.We therefore investigated whether the reduced virulence of a ${Delta}$ ciaR mutant could be explained by a down-regulation of HtrAexpression. To do this we made a ${Delta}$ ciaR version of strain D39and confirmed that this was less virulent. The reduction inlung colonization was of the order of several orders of magnitude,a finding similar to that of previous studies (42). Westernblotting indicated that strain D39 ${Delta}$ ciaR expressed approximatelyfourfold less HtrA protein than wild-type D39. When HtrA wasexpressed from a plasmid in the D39 ${Delta}$ ciaR strain, Western blottingshowed the level of HtrA was restored to that of the wild type.We also found that when HtrA was expressed from the plasmidin the wild-type strain, there was no increase in the levelof HtrA protein. This finding is surprising and suggests thatthere is a regulatory mechanism for controlling maximal levelsof HtrA and that this mechanism may not be dependent on thepresence of CiaR/H. Restoring the level of HtrA to wild-typelevels in strain D39 ${Delta}$ ciaR produced a strain that was again fullyvirulent in mice. This shows that the virulence phenotype ofD39 ${Delta}$ ciaR can be completely explained by the reduced expressionof HtrA. Interestingly, D39 ${Delta}$ ciaR still expresses low levelsof HtrA; this may explain why it is more virulent than D39 ${Delta}$ htrA.

We have shown in this study that HtrA behaves as a typical HSP,with levels of the protein increasing after exposure of bacteriato a temperature of 40°C. Interestingly, we also demonstratedthat the increased expression of HtrA at high temperature isdependent on the presence of CiaR/H. Based on these findingswe predicted that strain D39 ${Delta}$ ciaR would be defective in growthat 40°C (due to decreased expression of HtrA). Analysisin vitro indicates that the growth of the D39 ${Delta}$ ciaR mutant wasless than that of the wild type at elevated temperature. TheCiaR/H system is also involved in the control of bacterial lysis(11, 16, 24, 29); in particular, Lange and coworkers (24) reportedthat a ciaR mutant of the strain R6 had an increased rate ofautolysis when grown in Todd-Hewitt medium at 37°C. In contrastto this report, our D39 ${Delta}$ ciaR mutant showed a decreased rate ofautolysis, as judged by viable counts at stationary and declinephases of growth at both normal and elevated temperatures. Thismay reflect differences in the genetic content of R6 and D39,which are now known to differ at several loci (41). We alsostudied the in vitro growth phenotype of the serotype 3 ciaRmutant. The effect of ciaR deletion in this strain was moredramatic, as the growth of this mutant was prevented at elevatedtemperature; however, it showed a rate of autolysis similarto that of the parent strain at 37°C. These findings highlightthe differences of the effects of mutations in different strains.The growth phenotype of strain D39 ${Delta}$ ciaR was restored to thatof the wild type by introducing pAL2-HtrA plasmid in the complementedstrain, suggesting that the strain D39 ${Delta}$ ciaR growth phenotypeis due to the low level of HtrA expressed by this strain.

As we have previously shown that deletion of HtrA is associatedwith increased sensitivity to oxidative stress (18), we predictedthat reduced levels of HtrA in strain D39 ${Delta}$ ciaR would also resultin reduced resistance to oxidative stress. The sensitivity tooxidative stress of D39 ${Delta}$ ciaR was increased, although not to thesame extent as found previously for the D39 ${Delta}$ htrA mutant. After5 min of exposure to 40 mM hydrogen peroxide, the survival ofstrain D39 ${Delta}$ ciaR was similar to that of the wild type (Fig. 5)whereas the survival of strain D39 ${Delta}$ htrA was significantly reducedat this time (18). This probably reflects the expression oflow levels of HtrA in strain D39 ${Delta}$ ciaR. Resistance of strain D39 ${Delta}$ ciaRto hydrogen peroxide could be restored to wild-type levels bycomplementation with HtrA, which again confirms that the D39 ${Delta}$ ciaRphenotype could be explained by down-regulation of HtrA in thismutant.

In summary, we have confirmed that the CiaR/H system is involvedin controlling the levels of HtrA within the cell and have shownthat up-regulation of HtrA in response to heat is dependenton the CiaR/H system. As the response regulator CiaR has beenshown to physically bind to DNA in the region of HtrA, thisregulation is proposed to be direct (29). We have shown thata number of the phenotypes associated with deficiency in CiaR/Hare due to the alterations in levels of HtrA. Thus, the resistanceof cells to autolysis, the increased sensitivity to oxidativestress, and the decreased virulence of strain D39 ${Delta}$ ciaR can allbe explained by alterations in levels of HtrA. We have alsoshown that transformation efficiency is also altered by HtrA(18). While the CiaR/H regulon is obviously very complex, manyof the phenotypes observed with mutants of this regulon canbe explained by changes in HtrA. The exact mechanism by whichHtrA controls these processes is still unclear and is the subjectof ongoing studies.

ACKNOWLEDGMENTS

Yasser Musa Ibrahim is the recipient of a scholarship from theEgyptian government.

We thank J.-P. Claverys and Bernard Martin (CNRS-UniversitéPaul Sabatier, Toulouse, France) for provision of strain R6carrying a mariner mutation in CiaR and Mark Roberts (Universityof Glasgow) for provision of HtrA antiserum. We thank V. Salisbury(University of the West of England) for providing plasmid pAL2.We are also grateful to Martin Burnham (GlaxoSmithKline) forprovision of type 3 strain 0100993 with a CiaR mutation.

FOOTNOTES

* Corresponding author. Mailing address: Division of Infection and Immunity, Institute of Biomedical and Life Sciences, University of Glasgow, Joseph Black Building, Glasgow G12-8QQ, United Kingdom. Phone: 44 141 330 3749. Fax: 44 141 330 3727. E-mail: t.mitchell{at}bio.gla.ac.uk.

REFERENCES

Top